blob: 5042df95fb233ebf82fc79a4cfde2ab1111f49ce [file] [log] [blame]
Willy Tarreau6a06a402007-07-15 20:15:28 +02001 ----------------------
Willy Tarreau8317b282014-04-23 01:49:41 +02002 HAProxy
Willy Tarreau6a06a402007-07-15 20:15:28 +02003 Configuration Manual
4 ----------------------
Willy Tarreaub3066502017-11-26 19:50:17 +01005 version 1.9
Willy Tarreau6a06a402007-07-15 20:15:28 +02006 willy tarreau
Willy Tarreau253006d2018-09-12 18:59:48 +02007 2018/09/12
Willy Tarreau6a06a402007-07-15 20:15:28 +02008
9
10This document covers the configuration language as implemented in the version
Davor Ocelice9ed2812017-12-25 17:49:28 +010011specified above. It does not provide any hints, examples, or advice. For such
Willy Tarreau0ba27502007-12-24 16:55:16 +010012documentation, please refer to the Reference Manual or the Architecture Manual.
Davor Ocelice9ed2812017-12-25 17:49:28 +010013The summary below is meant to help you find sections by name and navigate
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014through the document.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016Note to documentation contributors :
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017 This document is formatted with 80 columns per line, with even number of
Willy Tarreauc57f0e22009-05-10 13:12:33 +020018 spaces for indentation and without tabs. Please follow these rules strictly
19 so that it remains easily printable everywhere. If a line needs to be
20 printed verbatim and does not fit, please end each line with a backslash
Willy Tarreau62a36c42010-08-17 15:53:10 +020021 ('\') and continue on next line, indented by two characters. It is also
Davor Ocelice9ed2812017-12-25 17:49:28 +010022 sometimes useful to prefix all output lines (logs, console outputs) with 3
23 closing angle brackets ('>>>') in order to emphasize the difference between
24 inputs and outputs when they may be ambiguous. If you add sections,
Willy Tarreau62a36c42010-08-17 15:53:10 +020025 please update the summary below for easier searching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020026
27
28Summary
29-------
30
311. Quick reminder about HTTP
321.1. The HTTP transaction model
331.2. HTTP request
Davor Ocelice9ed2812017-12-25 17:49:28 +0100341.2.1. The request line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200351.2.2. The request headers
361.3. HTTP response
Davor Ocelice9ed2812017-12-25 17:49:28 +0100371.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200381.3.2. The response headers
39
402. Configuring HAProxy
412.1. Configuration file format
William Lallemandf9873ba2015-05-05 17:37:14 +0200422.2. Quoting and escaping
William Lallemandb2f07452015-05-12 14:27:13 +0200432.3. Environment variables
442.4. Time format
452.5. Examples
Willy Tarreauc57f0e22009-05-10 13:12:33 +020046
473. Global parameters
483.1. Process management and security
493.2. Performance tuning
503.3. Debugging
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +0100513.4. Userlists
Cyril Bontédc4d9032012-04-08 21:57:39 +0200523.5. Peers
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200533.6. Mailers
Willy Tarreauc57f0e22009-05-10 13:12:33 +020054
554. Proxies
564.1. Proxy keywords matrix
574.2. Alphabetically sorted keywords reference
58
Davor Ocelice9ed2812017-12-25 17:49:28 +0100595. Bind and server options
Willy Tarreau086fbf52012-09-24 20:34:51 +0200605.1. Bind options
615.2. Server and default-server options
Baptiste Assmann1fa66662015-04-14 00:28:47 +0200625.3. Server DNS resolution
635.3.1. Global overview
645.3.2. The resolvers section
Willy Tarreauc57f0e22009-05-10 13:12:33 +020065
666. HTTP header manipulation
67
Willy Tarreau74ca5042013-06-11 23:12:07 +0200687. Using ACLs and fetching samples
697.1. ACL basics
707.1.1. Matching booleans
717.1.2. Matching integers
727.1.3. Matching strings
737.1.4. Matching regular expressions (regexes)
747.1.5. Matching arbitrary data blocks
757.1.6. Matching IPv4 and IPv6 addresses
767.2. Using ACLs to form conditions
777.3. Fetching samples
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200787.3.1. Converters
797.3.2. Fetching samples from internal states
807.3.3. Fetching samples at Layer 4
817.3.4. Fetching samples at Layer 5
827.3.5. Fetching samples from buffer contents (Layer 6)
837.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +0200847.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020085
868. Logging
878.1. Log levels
888.2. Log formats
898.2.1. Default log format
908.2.2. TCP log format
918.2.3. HTTP log format
William Lallemand48940402012-01-30 16:47:22 +0100928.2.4. Custom log format
Willy Tarreau5f51e1a2012-12-03 18:40:10 +0100938.2.5. Error log format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200948.3. Advanced logging options
958.3.1. Disabling logging of external tests
968.3.2. Logging before waiting for the session to terminate
978.3.3. Raising log level upon errors
988.3.4. Disabling logging of successful connections
998.4. Timing events
1008.5. Session state at disconnection
1018.6. Non-printable characters
1028.7. Capturing HTTP cookies
1038.8. Capturing HTTP headers
1048.9. Examples of logs
105
Christopher Fauletc3fe5332016-04-07 15:30:10 +02001069. Supported filters
1079.1. Trace
1089.2. HTTP compression
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +02001099.3. Stream Processing Offload Engine (SPOE)
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200110
William Lallemand86d0df02017-11-24 21:36:45 +010011110. Cache
Cyril Bonté7b888f12017-11-26 22:24:31 +010011210.1. Limitation
11310.2. Setup
11410.2.1. Cache section
11510.2.2. Proxy section
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200116
1171. Quick reminder about HTTP
118----------------------------
119
Davor Ocelice9ed2812017-12-25 17:49:28 +0100120When HAProxy is running in HTTP mode, both the request and the response are
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200121fully analyzed and indexed, thus it becomes possible to build matching criteria
122on almost anything found in the contents.
123
124However, it is important to understand how HTTP requests and responses are
125formed, and how HAProxy decomposes them. It will then become easier to write
126correct rules and to debug existing configurations.
127
128
1291.1. The HTTP transaction model
130-------------------------------
131
132The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100133to one and only one response. Traditionally, a TCP connection is established
Davor Ocelice9ed2812017-12-25 17:49:28 +0100134from the client to the server, a request is sent by the client through the
135connection, the server responds, and the connection is closed. A new request
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200136will involve a new connection :
137
138 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
139
140In this mode, called the "HTTP close" mode, there are as many connection
141establishments as there are HTTP transactions. Since the connection is closed
142by the server after the response, the client does not need to know the content
143length.
144
145Due to the transactional nature of the protocol, it was possible to improve it
146to avoid closing a connection between two subsequent transactions. In this mode
147however, it is mandatory that the server indicates the content length for each
148response so that the client does not wait indefinitely. For this, a special
149header is used: "Content-length". This mode is called the "keep-alive" mode :
150
151 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
152
153Its advantages are a reduced latency between transactions, and less processing
154power required on the server side. It is generally better than the close mode,
155but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200156a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200157
Willy Tarreau95c4e142017-11-26 12:18:55 +0100158Another improvement in the communications is the pipelining mode. It still uses
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200159keep-alive, but the client does not wait for the first response to send the
160second request. This is useful for fetching large number of images composing a
161page :
162
163 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
164
165This can obviously have a tremendous benefit on performance because the network
166latency is eliminated between subsequent requests. Many HTTP agents do not
167correctly support pipelining since there is no way to associate a response with
168the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100169server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200170
Willy Tarreau95c4e142017-11-26 12:18:55 +0100171The next improvement is the multiplexed mode, as implemented in HTTP/2. This
172time, each transaction is assigned a single stream identifier, and all streams
173are multiplexed over an existing connection. Many requests can be sent in
174parallel by the client, and responses can arrive in any order since they also
175carry the stream identifier.
176
Willy Tarreau70dffda2014-01-30 03:07:23 +0100177By default HAProxy operates in keep-alive mode with regards to persistent
178connections: for each connection it processes each request and response, and
179leaves the connection idle on both sides between the end of a response and the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100180start of a new request. When it receives HTTP/2 connections from a client, it
181processes all the requests in parallel and leaves the connection idling,
182waiting for new requests, just as if it was a keep-alive HTTP connection.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200183
Willy Tarreau70dffda2014-01-30 03:07:23 +0100184HAProxy supports 5 connection modes :
185 - keep alive : all requests and responses are processed (default)
186 - tunnel : only the first request and response are processed,
187 everything else is forwarded with no analysis.
188 - passive close : tunnel with "Connection: close" added in both directions.
189 - server close : the server-facing connection is closed after the response.
190 - forced close : the connection is actively closed after end of response.
191
Davor Ocelice9ed2812017-12-25 17:49:28 +0100192For HTTP/2, the connection mode resembles more the "server close" mode : given
193the independence of all streams, there is currently no place to hook the idle
Willy Tarreau95c4e142017-11-26 12:18:55 +0100194server connection after a response, so it is closed after the response. HTTP/2
195is only supported for incoming connections, not on connections going to
196servers.
197
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200198
1991.2. HTTP request
200-----------------
201
202First, let's consider this HTTP request :
203
204 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100205 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
207 2 Host: www.mydomain.com
208 3 User-agent: my small browser
209 4 Accept: image/jpeg, image/gif
210 5 Accept: image/png
211
212
2131.2.1. The Request line
214-----------------------
215
216Line 1 is the "request line". It is always composed of 3 fields :
217
218 - a METHOD : GET
219 - a URI : /serv/login.php?lang=en&profile=2
220 - a version tag : HTTP/1.1
221
222All of them are delimited by what the standard calls LWS (linear white spaces),
223which are commonly spaces, but can also be tabs or line feeds/carriage returns
224followed by spaces/tabs. The method itself cannot contain any colon (':') and
225is limited to alphabetic letters. All those various combinations make it
226desirable that HAProxy performs the splitting itself rather than leaving it to
227the user to write a complex or inaccurate regular expression.
228
229The URI itself can have several forms :
230
231 - A "relative URI" :
232
233 /serv/login.php?lang=en&profile=2
234
235 It is a complete URL without the host part. This is generally what is
236 received by servers, reverse proxies and transparent proxies.
237
238 - An "absolute URI", also called a "URL" :
239
240 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
241
242 It is composed of a "scheme" (the protocol name followed by '://'), a host
243 name or address, optionally a colon (':') followed by a port number, then
244 a relative URI beginning at the first slash ('/') after the address part.
245 This is generally what proxies receive, but a server supporting HTTP/1.1
246 must accept this form too.
247
248 - a star ('*') : this form is only accepted in association with the OPTIONS
249 method and is not relayable. It is used to inquiry a next hop's
250 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100251
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200252 - an address:port combination : 192.168.0.12:80
253 This is used with the CONNECT method, which is used to establish TCP
254 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
255 other protocols too.
256
257In a relative URI, two sub-parts are identified. The part before the question
258mark is called the "path". It is typically the relative path to static objects
259on the server. The part after the question mark is called the "query string".
260It is mostly used with GET requests sent to dynamic scripts and is very
261specific to the language, framework or application in use.
262
Willy Tarreau95c4e142017-11-26 12:18:55 +0100263HTTP/2 doesn't convey a version information with the request, so the version is
Davor Ocelice9ed2812017-12-25 17:49:28 +0100264assumed to be the same as the one of the underlying protocol (i.e. "HTTP/2").
Willy Tarreau95c4e142017-11-26 12:18:55 +0100265However, haproxy natively processes HTTP/1.x requests and headers, so requests
266received over an HTTP/2 connection are transcoded to HTTP/1.1 before being
267processed. This explains why they still appear as "HTTP/1.1" in haproxy's logs
268as well as in server logs.
269
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200270
2711.2.2. The request headers
272--------------------------
273
274The headers start at the second line. They are composed of a name at the
275beginning of the line, immediately followed by a colon (':'). Traditionally,
276an LWS is added after the colon but that's not required. Then come the values.
277Multiple identical headers may be folded into one single line, delimiting the
278values with commas, provided that their order is respected. This is commonly
279encountered in the "Cookie:" field. A header may span over multiple lines if
280the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
281define a total of 3 values for the "Accept:" header.
282
Davor Ocelice9ed2812017-12-25 17:49:28 +0100283Contrary to a common misconception, header names are not case-sensitive, and
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200284their values are not either if they refer to other header names (such as the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100285"Connection:" header). In HTTP/2, header names are always sent in lower case,
286as can be seen when running in debug mode.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200287
288The end of the headers is indicated by the first empty line. People often say
289that it's a double line feed, which is not exact, even if a double line feed
290is one valid form of empty line.
291
292Fortunately, HAProxy takes care of all these complex combinations when indexing
293headers, checking values and counting them, so there is no reason to worry
294about the way they could be written, but it is important not to accuse an
295application of being buggy if it does unusual, valid things.
296
297Important note:
Lukas Tribus23953682017-04-28 13:24:30 +0000298 As suggested by RFC7231, HAProxy normalizes headers by replacing line breaks
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200299 in the middle of headers by LWS in order to join multi-line headers. This
300 is necessary for proper analysis and helps less capable HTTP parsers to work
301 correctly and not to be fooled by such complex constructs.
302
303
3041.3. HTTP response
305------------------
306
307An HTTP response looks very much like an HTTP request. Both are called HTTP
308messages. Let's consider this HTTP response :
309
310 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100311 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200312 1 HTTP/1.1 200 OK
313 2 Content-length: 350
314 3 Content-Type: text/html
315
Willy Tarreau816b9792009-09-15 21:25:21 +0200316As a special case, HTTP supports so called "Informational responses" as status
317codes 1xx. These messages are special in that they don't convey any part of the
318response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100319continue to post its request for instance. In the case of a status 100 response
320the requested information will be carried by the next non-100 response message
321following the informational one. This implies that multiple responses may be
322sent to a single request, and that this only works when keep-alive is enabled
323(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
324correctly forward and skip them, and only process the next non-100 response. As
325such, these messages are neither logged nor transformed, unless explicitly
326state otherwise. Status 101 messages indicate that the protocol is changing
327over the same connection and that haproxy must switch to tunnel mode, just as
328if a CONNECT had occurred. Then the Upgrade header would contain additional
329information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200330
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200331
Davor Ocelice9ed2812017-12-25 17:49:28 +01003321.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200333------------------------
334
335Line 1 is the "response line". It is always composed of 3 fields :
336
337 - a version tag : HTTP/1.1
338 - a status code : 200
339 - a reason : OK
340
341The status code is always 3-digit. The first digit indicates a general status :
Davor Ocelice9ed2812017-12-25 17:49:28 +0100342 - 1xx = informational message to be skipped (e.g. 100, 101)
343 - 2xx = OK, content is following (e.g. 200, 206)
344 - 3xx = OK, no content following (e.g. 302, 304)
345 - 4xx = error caused by the client (e.g. 401, 403, 404)
346 - 5xx = error caused by the server (e.g. 500, 502, 503)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200347
Lukas Tribus23953682017-04-28 13:24:30 +0000348Please refer to RFC7231 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100349"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200350found there, but it's a common practice to respect the well-established
351messages. It can be composed of one or multiple words, such as "OK", "Found",
352or "Authentication Required".
353
Davor Ocelice9ed2812017-12-25 17:49:28 +0100354HAProxy may emit the following status codes by itself :
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200355
356 Code When / reason
357 200 access to stats page, and when replying to monitoring requests
358 301 when performing a redirection, depending on the configured code
359 302 when performing a redirection, depending on the configured code
360 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100361 307 when performing a redirection, depending on the configured code
362 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200363 400 for an invalid or too large request
364 401 when an authentication is required to perform the action (when
365 accessing the stats page)
366 403 when a request is forbidden by a "block" ACL or "reqdeny" filter
367 408 when the request timeout strikes before the request is complete
368 500 when haproxy encounters an unrecoverable internal error, such as a
369 memory allocation failure, which should never happen
370 502 when the server returns an empty, invalid or incomplete response, or
371 when an "rspdeny" filter blocks the response.
372 503 when no server was available to handle the request, or in response to
373 monitoring requests which match the "monitor fail" condition
374 504 when the response timeout strikes before the server responds
375
376The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3774.2).
378
379
3801.3.2. The response headers
381---------------------------
382
383Response headers work exactly like request headers, and as such, HAProxy uses
384the same parsing function for both. Please refer to paragraph 1.2.2 for more
385details.
386
387
3882. Configuring HAProxy
389----------------------
390
3912.1. Configuration file format
392------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200393
394HAProxy's configuration process involves 3 major sources of parameters :
395
396 - the arguments from the command-line, which always take precedence
397 - the "global" section, which sets process-wide parameters
398 - the proxies sections which can take form of "defaults", "listen",
399 "frontend" and "backend".
400
Willy Tarreau0ba27502007-12-24 16:55:16 +0100401The configuration file syntax consists in lines beginning with a keyword
402referenced in this manual, optionally followed by one or several parameters
William Lallemandf9873ba2015-05-05 17:37:14 +0200403delimited by spaces.
Willy Tarreau0ba27502007-12-24 16:55:16 +0100404
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200405
William Lallemandf9873ba2015-05-05 17:37:14 +02004062.2. Quoting and escaping
407-------------------------
408
409HAProxy's configuration introduces a quoting and escaping system similar to
410many programming languages. The configuration file supports 3 types: escaping
411with a backslash, weak quoting with double quotes, and strong quoting with
412single quotes.
413
414If spaces have to be entered in strings, then they must be escaped by preceding
415them by a backslash ('\') or by quoting them. Backslashes also have to be
416escaped by doubling or strong quoting them.
417
418Escaping is achieved by preceding a special character by a backslash ('\'):
419
420 \ to mark a space and differentiate it from a delimiter
421 \# to mark a hash and differentiate it from a comment
422 \\ to use a backslash
423 \' to use a single quote and differentiate it from strong quoting
424 \" to use a double quote and differentiate it from weak quoting
425
426Weak quoting is achieved by using double quotes (""). Weak quoting prevents
427the interpretation of:
428
429 space as a parameter separator
430 ' single quote as a strong quoting delimiter
431 # hash as a comment start
432
William Lallemandb2f07452015-05-12 14:27:13 +0200433Weak quoting permits the interpretation of variables, if you want to use a non
434-interpreted dollar within a double quoted string, you should escape it with a
435backslash ("\$"), it does not work outside weak quoting.
436
437Interpretation of escaping and special characters are not prevented by weak
William Lallemandf9873ba2015-05-05 17:37:14 +0200438quoting.
439
440Strong quoting is achieved by using single quotes (''). Inside single quotes,
441nothing is interpreted, it's the efficient way to quote regexes.
442
443Quoted and escaped strings are replaced in memory by their interpreted
444equivalent, it allows you to perform concatenation.
445
446 Example:
447 # those are equivalents:
448 log-format %{+Q}o\ %t\ %s\ %{-Q}r
449 log-format "%{+Q}o %t %s %{-Q}r"
450 log-format '%{+Q}o %t %s %{-Q}r'
451 log-format "%{+Q}o %t"' %s %{-Q}r'
452 log-format "%{+Q}o %t"' %s'\ %{-Q}r
453
454 # those are equivalents:
455 reqrep "^([^\ :]*)\ /static/(.*)" \1\ /\2
456 reqrep "^([^ :]*)\ /static/(.*)" '\1 /\2'
457 reqrep "^([^ :]*)\ /static/(.*)" "\1 /\2"
458 reqrep "^([^ :]*)\ /static/(.*)" "\1\ /\2"
459
460
William Lallemandb2f07452015-05-12 14:27:13 +02004612.3. Environment variables
462--------------------------
463
464HAProxy's configuration supports environment variables. Those variables are
465interpreted only within double quotes. Variables are expanded during the
466configuration parsing. Variable names must be preceded by a dollar ("$") and
467optionally enclosed with braces ("{}") similarly to what is done in Bourne
468shell. Variable names can contain alphanumerical characters or the character
469underscore ("_") but should not start with a digit.
470
471 Example:
472
473 bind "fd@${FD_APP1}"
474
475 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
476
477 user "$HAPROXY_USER"
478
William Lallemanddaf4cd22018-04-17 16:46:13 +0200479A special variable $HAPROXY_LOCALPEER is defined at the startup of the process
480which contains the name of the local peer. (See "-L" in the management guide.)
481
William Lallemandb2f07452015-05-12 14:27:13 +0200482
4832.4. Time format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200484----------------
485
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100486Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100487values are generally expressed in milliseconds (unless explicitly stated
488otherwise) but may be expressed in any other unit by suffixing the unit to the
489numeric value. It is important to consider this because it will not be repeated
490for every keyword. Supported units are :
491
492 - us : microseconds. 1 microsecond = 1/1000000 second
493 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
494 - s : seconds. 1s = 1000ms
495 - m : minutes. 1m = 60s = 60000ms
496 - h : hours. 1h = 60m = 3600s = 3600000ms
497 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
498
499
Lukas Tribusaa83a312017-03-21 09:25:09 +00005002.5. Examples
Patrick Mezard35da19c2010-06-12 17:02:47 +0200501-------------
502
503 # Simple configuration for an HTTP proxy listening on port 80 on all
504 # interfaces and forwarding requests to a single backend "servers" with a
505 # single server "server1" listening on 127.0.0.1:8000
506 global
507 daemon
508 maxconn 256
509
510 defaults
511 mode http
512 timeout connect 5000ms
513 timeout client 50000ms
514 timeout server 50000ms
515
516 frontend http-in
517 bind *:80
518 default_backend servers
519
520 backend servers
521 server server1 127.0.0.1:8000 maxconn 32
522
523
524 # The same configuration defined with a single listen block. Shorter but
525 # less expressive, especially in HTTP mode.
526 global
527 daemon
528 maxconn 256
529
530 defaults
531 mode http
532 timeout connect 5000ms
533 timeout client 50000ms
534 timeout server 50000ms
535
536 listen http-in
537 bind *:80
538 server server1 127.0.0.1:8000 maxconn 32
539
540
541Assuming haproxy is in $PATH, test these configurations in a shell with:
542
Willy Tarreauccb289d2010-12-11 20:19:38 +0100543 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200544
545
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005463. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200547--------------------
548
549Parameters in the "global" section are process-wide and often OS-specific. They
550are generally set once for all and do not need being changed once correct. Some
551of them have command-line equivalents.
552
553The following keywords are supported in the "global" section :
554
555 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200556 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200557 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200558 - crt-base
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200559 - cpu-map
Willy Tarreau6a06a402007-07-15 20:15:28 +0200560 - daemon
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200561 - description
562 - deviceatlas-json-file
563 - deviceatlas-log-level
564 - deviceatlas-separator
565 - deviceatlas-properties-cookie
Simon Horman98637e52014-06-20 12:30:16 +0900566 - external-check
Willy Tarreau6a06a402007-07-15 20:15:28 +0200567 - gid
568 - group
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100569 - hard-stop-after
Willy Tarreau6a06a402007-07-15 20:15:28 +0200570 - log
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200571 - log-tag
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100572 - log-send-hostname
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200573 - lua-load
Willy Tarreau6a06a402007-07-15 20:15:28 +0200574 - nbproc
Christopher Fauletbe0faa22017-08-29 15:37:10 +0200575 - nbthread
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200576 - node
Willy Tarreau6a06a402007-07-15 20:15:28 +0200577 - pidfile
Willy Tarreau1d549722016-02-16 12:41:57 +0100578 - presetenv
579 - resetenv
Willy Tarreau6a06a402007-07-15 20:15:28 +0200580 - uid
581 - ulimit-n
582 - user
Willy Tarreau1d549722016-02-16 12:41:57 +0100583 - setenv
Willy Tarreaufbee7132007-10-18 13:53:22 +0200584 - stats
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200585 - ssl-default-bind-ciphers
586 - ssl-default-bind-options
587 - ssl-default-server-ciphers
588 - ssl-default-server-options
589 - ssl-dh-param-file
Emeric Brun850efd52014-01-29 12:24:34 +0100590 - ssl-server-verify
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100591 - unix-bind
Willy Tarreau1d549722016-02-16 12:41:57 +0100592 - unsetenv
Thomas Holmesdb04f192015-05-18 13:21:39 +0100593 - 51degrees-data-file
594 - 51degrees-property-name-list
Dragan Dosen93b38d92015-06-29 16:43:25 +0200595 - 51degrees-property-separator
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200596 - 51degrees-cache-size
scientiamobiled0027ed2016-11-04 10:55:08 +0100597 - wurfl-data-file
598 - wurfl-information-list
599 - wurfl-information-list-separator
600 - wurfl-engine-mode
601 - wurfl-cache-size
602 - wurfl-useragent-priority
Willy Tarreaud72758d2010-01-12 10:42:19 +0100603
Willy Tarreau6a06a402007-07-15 20:15:28 +0200604 * Performance tuning
Willy Tarreau1746eec2014-04-25 10:46:47 +0200605 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +0200606 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +0200607 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +0100608 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +0100609 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100610 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +0200611 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +0200612 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +0200613 - maxsslrate
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200614 - maxzlibmem
Willy Tarreau6a06a402007-07-15 20:15:28 +0200615 - noepoll
616 - nokqueue
617 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100618 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300619 - nogetaddrinfo
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +0000620 - noreuseport
Willy Tarreaufe255b72007-10-14 23:09:26 +0200621 - spread-checks
Baptiste Assmann5626f482015-08-23 10:00:10 +0200622 - server-state-base
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200623 - server-state-file
Grant Zhang872f9c22017-01-21 01:10:18 +0000624 - ssl-engine
Grant Zhangfa6c7ee2017-01-14 01:42:15 +0000625 - ssl-mode-async
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200626 - tune.buffers.limit
627 - tune.buffers.reserve
Willy Tarreau27a674e2009-08-17 07:23:33 +0200628 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +0200629 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +0100630 - tune.comp.maxlevel
Willy Tarreaufe20e5b2017-07-27 11:42:14 +0200631 - tune.h2.header-table-size
Willy Tarreaue6baec02017-07-27 11:45:11 +0200632 - tune.h2.initial-window-size
Willy Tarreau5242ef82017-07-27 11:47:28 +0200633 - tune.h2.max-concurrent-streams
Willy Tarreau193b8c62012-11-22 00:17:38 +0100634 - tune.http.cookielen
Stéphane Cottin23e9e932017-05-18 08:58:41 +0200635 - tune.http.logurilen
Willy Tarreauac1932d2011-10-24 19:14:41 +0200636 - tune.http.maxhdr
Willy Tarreau7e312732014-02-12 16:35:14 +0100637 - tune.idletimer
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100638 - tune.lua.forced-yield
Willy Tarreau32f61e22015-03-18 17:54:59 +0100639 - tune.lua.maxmem
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100640 - tune.lua.session-timeout
641 - tune.lua.task-timeout
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +0200642 - tune.lua.service-timeout
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100643 - tune.maxaccept
644 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +0200645 - tune.maxrewrite
Willy Tarreauf3045d22015-04-29 16:24:50 +0200646 - tune.pattern.cache-size
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200647 - tune.pipesize
Willy Tarreaue803de22010-01-21 17:43:04 +0100648 - tune.rcvbuf.client
649 - tune.rcvbuf.server
Willy Tarreaub22fc302015-12-14 12:04:35 +0100650 - tune.recv_enough
Olivier Houchard1599b802018-05-24 18:59:04 +0200651 - tune.runqueue-depth
Willy Tarreaue803de22010-01-21 17:43:04 +0100652 - tune.sndbuf.client
653 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +0100654 - tune.ssl.cachesize
Willy Tarreaubfd59462013-02-21 07:46:09 +0100655 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +0200656 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +0100657 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +0200658 - tune.ssl.default-dh-param
Christopher Faulet31af49d2015-06-09 17:29:50 +0200659 - tune.ssl.ssl-ctx-cache-size
Thierry FOURNIER5bf77322017-02-25 12:45:22 +0100660 - tune.ssl.capture-cipherlist-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200661 - tune.vars.global-max-size
Christopher Fauletff2613e2016-11-09 11:36:17 +0100662 - tune.vars.proc-max-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200663 - tune.vars.reqres-max-size
664 - tune.vars.sess-max-size
665 - tune.vars.txn-max-size
William Lallemanda509e4c2012-11-07 16:54:34 +0100666 - tune.zlib.memlevel
667 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +0100668
Willy Tarreau6a06a402007-07-15 20:15:28 +0200669 * Debugging
670 - debug
671 - quiet
Willy Tarreau6a06a402007-07-15 20:15:28 +0200672
673
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006743.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +0200675------------------------------------
676
Emeric Brunc8e8d122012-10-02 18:42:10 +0200677ca-base <dir>
678 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emeric Brunfd33a262012-10-11 16:28:27 +0200679 relative path is used with "ca-file" or "crl-file" directives. Absolute
680 locations specified in "ca-file" and "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +0200681
Willy Tarreau6a06a402007-07-15 20:15:28 +0200682chroot <jail dir>
683 Changes current directory to <jail dir> and performs a chroot() there before
684 dropping privileges. This increases the security level in case an unknown
685 vulnerability would be exploited, since it would make it very hard for the
686 attacker to exploit the system. This only works when the process is started
687 with superuser privileges. It is important to ensure that <jail_dir> is both
Davor Ocelice9ed2812017-12-25 17:49:28 +0100688 empty and non-writable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100689
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100690cpu-map [auto:]<process-set>[/<thread-set>] <cpu-set>...
691 On Linux 2.6 and above, it is possible to bind a process or a thread to a
692 specific CPU set. This means that the process or the thread will never run on
693 other CPUs. The "cpu-map" directive specifies CPU sets for process or thread
694 sets. The first argument is a process set, eventually followed by a thread
695 set. These sets have the format
696
697 all | odd | even | number[-[number]]
698
699 <number>> must be a number between 1 and 32 or 64, depending on the machine's
Davor Ocelice9ed2812017-12-25 17:49:28 +0100700 word size. Any process IDs above nbproc and any thread IDs above nbthread are
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100701 ignored. It is possible to specify a range with two such number delimited by
702 a dash ('-'). It also is possible to specify all processes at once using
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +0100703 "all", only odd numbers using "odd" or even numbers using "even", just like
704 with the "bind-process" directive. The second and forthcoming arguments are
Davor Ocelice9ed2812017-12-25 17:49:28 +0100705 CPU sets. Each CPU set is either a unique number between 0 and 31 or 63 or a
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +0100706 range with two such numbers delimited by a dash ('-'). Multiple CPU numbers
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100707 or ranges may be specified, and the processes or threads will be allowed to
Davor Ocelice9ed2812017-12-25 17:49:28 +0100708 bind to all of them. Obviously, multiple "cpu-map" directives may be
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100709 specified. Each "cpu-map" directive will replace the previous ones when they
710 overlap. A thread will be bound on the intersection of its mapping and the
711 one of the process on which it is attached. If the intersection is null, no
712 specific binding will be set for the thread.
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100713
Christopher Fauletff4121f2017-11-22 16:38:49 +0100714 Ranges can be partially defined. The higher bound can be omitted. In such
715 case, it is replaced by the corresponding maximum value, 32 or 64 depending
716 on the machine's word size.
717
Christopher Faulet26028f62017-11-22 15:01:51 +0100718 The prefix "auto:" can be added before the process set to let HAProxy
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100719 automatically bind a process or a thread to a CPU by incrementing
720 process/thread and CPU sets. To be valid, both sets must have the same
721 size. No matter the declaration order of the CPU sets, it will be bound from
722 the lowest to the highest bound. Having a process and a thread range with the
723 "auto:" prefix is not supported. Only one range is supported, the other one
724 must be a fixed number.
Christopher Faulet26028f62017-11-22 15:01:51 +0100725
726 Examples:
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100727 cpu-map 1-4 0-3 # bind processes 1 to 4 on the first 4 CPUs
728
729 cpu-map 1/all 0-3 # bind all threads of the first process on the
730 # first 4 CPUs
731
732 cpu-map 1- 0- # will be replaced by "cpu-map 1-64 0-63"
733 # or "cpu-map 1-32 0-31" depending on the machine's
734 # word size.
735
Christopher Faulet26028f62017-11-22 15:01:51 +0100736 # all these lines bind the process 1 to the cpu 0, the process 2 to cpu 1
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100737 # and so on.
Christopher Faulet26028f62017-11-22 15:01:51 +0100738 cpu-map auto:1-4 0-3
739 cpu-map auto:1-4 0-1 2-3
740 cpu-map auto:1-4 3 2 1 0
741
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100742 # all these lines bind the thread 1 to the cpu 0, the thread 2 to cpu 1
743 # and so on.
744 cpu-map auto:1/1-4 0-3
745 cpu-map auto:1/1-4 0-1 2-3
746 cpu-map auto:1/1-4 3 2 1 0
747
Davor Ocelice9ed2812017-12-25 17:49:28 +0100748 # bind each process to exactly one CPU using all/odd/even keyword
Christopher Faulet26028f62017-11-22 15:01:51 +0100749 cpu-map auto:all 0-63
750 cpu-map auto:even 0-31
751 cpu-map auto:odd 32-63
752
753 # invalid cpu-map because process and CPU sets have different sizes.
754 cpu-map auto:1-4 0 # invalid
755 cpu-map auto:1 0-3 # invalid
756
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100757 # invalid cpu-map because automatic binding is used with a process range
758 # and a thread range.
759 cpu-map auto:all/all 0 # invalid
760 cpu-map auto:all/1-4 0 # invalid
761 cpu-map auto:1-4/all 0 # invalid
762
Emeric Brunc8e8d122012-10-02 18:42:10 +0200763crt-base <dir>
764 Assigns a default directory to fetch SSL certificates from when a relative
765 path is used with "crtfile" directives. Absolute locations specified after
766 "crtfile" prevail and ignore "crt-base".
767
Willy Tarreau6a06a402007-07-15 20:15:28 +0200768daemon
769 Makes the process fork into background. This is the recommended mode of
770 operation. It is equivalent to the command line "-D" argument. It can be
Lukas Tribusf46bf952017-11-21 12:39:34 +0100771 disabled by the command line "-db" argument. This option is ignored in
772 systemd mode.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200773
David Carlier8167f302015-06-01 13:50:06 +0200774deviceatlas-json-file <path>
775 Sets the path of the DeviceAtlas JSON data file to be loaded by the API.
Davor Ocelice9ed2812017-12-25 17:49:28 +0100776 The path must be a valid JSON data file and accessible by HAProxy process.
David Carlier8167f302015-06-01 13:50:06 +0200777
778deviceatlas-log-level <value>
Davor Ocelice9ed2812017-12-25 17:49:28 +0100779 Sets the level of information returned by the API. This directive is
David Carlier8167f302015-06-01 13:50:06 +0200780 optional and set to 0 by default if not set.
781
782deviceatlas-separator <char>
783 Sets the character separator for the API properties results. This directive
784 is optional and set to | by default if not set.
785
Cyril Bonté0306c4a2015-10-26 22:37:38 +0100786deviceatlas-properties-cookie <name>
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200787 Sets the client cookie's name used for the detection if the DeviceAtlas
788 Client-side component was used during the request. This directive is optional
789 and set to DAPROPS by default if not set.
David Carlier29b3ca32015-09-25 14:09:21 +0100790
Simon Horman98637e52014-06-20 12:30:16 +0900791external-check
792 Allows the use of an external agent to perform health checks.
793 This is disabled by default as a security precaution.
794 See "option external-check".
795
Willy Tarreau6a06a402007-07-15 20:15:28 +0200796gid <number>
797 Changes the process' group ID to <number>. It is recommended that the group
798 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
799 be started with a user belonging to this group, or with superuser privileges.
Michael Schererab012dd2013-01-12 18:35:19 +0100800 Note that if haproxy is started from a user having supplementary groups, it
801 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200802 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100803
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100804hard-stop-after <time>
805 Defines the maximum time allowed to perform a clean soft-stop.
806
807 Arguments :
808 <time> is the maximum time (by default in milliseconds) for which the
809 instance will remain alive when a soft-stop is received via the
810 SIGUSR1 signal.
811
812 This may be used to ensure that the instance will quit even if connections
813 remain opened during a soft-stop (for example with long timeouts for a proxy
814 in tcp mode). It applies both in TCP and HTTP mode.
815
816 Example:
817 global
818 hard-stop-after 30s
819
Willy Tarreau6a06a402007-07-15 20:15:28 +0200820group <group name>
821 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
822 See also "gid" and "user".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100823
Dragan Dosen7ad31542015-09-28 17:16:47 +0200824log <address> [len <length>] [format <format>] <facility> [max level [min level]]
Cyril Bonté3e954872018-03-20 23:30:27 +0100825 Adds a global syslog server. Several global servers can be defined. They
Davor Ocelice9ed2812017-12-25 17:49:28 +0100826 will receive logs for starts and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +0100827 configured with "log global".
828
829 <address> can be one of:
830
Willy Tarreau2769aa02007-12-27 18:26:09 +0100831 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +0100832 no port is specified, 514 is used by default (the standard syslog
833 port).
834
David du Colombier24bb5f52011-03-17 10:40:23 +0100835 - An IPv6 address followed by a colon and optionally a UDP port. If
836 no port is specified, 514 is used by default (the standard syslog
837 port).
838
Robert Tsai81ae1952007-12-05 10:47:29 +0100839 - A filesystem path to a UNIX domain socket, keeping in mind
840 considerations for chroot (be sure the path is accessible inside
841 the chroot) and uid/gid (be sure the path is appropriately
Davor Ocelice9ed2812017-12-25 17:49:28 +0100842 writable).
Robert Tsai81ae1952007-12-05 10:47:29 +0100843
William Lallemandb2f07452015-05-12 14:27:13 +0200844 You may want to reference some environment variables in the address
845 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +0100846
Willy Tarreau18324f52014-06-27 18:10:07 +0200847 <length> is an optional maximum line length. Log lines larger than this value
848 will be truncated before being sent. The reason is that syslog
849 servers act differently on log line length. All servers support the
850 default value of 1024, but some servers simply drop larger lines
851 while others do log them. If a server supports long lines, it may
852 make sense to set this value here in order to avoid truncating long
853 lines. Similarly, if a server drops long lines, it is preferable to
854 truncate them before sending them. Accepted values are 80 to 65535
855 inclusive. The default value of 1024 is generally fine for all
856 standard usages. Some specific cases of long captures or
Davor Ocelice9ed2812017-12-25 17:49:28 +0100857 JSON-formatted logs may require larger values. You may also need to
858 increase "tune.http.logurilen" if your request URIs are truncated.
Willy Tarreau18324f52014-06-27 18:10:07 +0200859
Dragan Dosen7ad31542015-09-28 17:16:47 +0200860 <format> is the log format used when generating syslog messages. It may be
861 one of the following :
862
863 rfc3164 The RFC3164 syslog message format. This is the default.
864 (https://tools.ietf.org/html/rfc3164)
865
866 rfc5424 The RFC5424 syslog message format.
867 (https://tools.ietf.org/html/rfc5424)
868
Robert Tsai81ae1952007-12-05 10:47:29 +0100869 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200870
871 kern user mail daemon auth syslog lpr news
872 uucp cron auth2 ftp ntp audit alert cron2
873 local0 local1 local2 local3 local4 local5 local6 local7
874
875 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +0200876 all messages are sent. If a maximum level is specified, only messages with a
877 severity at least as important as this level will be sent. An optional minimum
878 level can be specified. If it is set, logs emitted with a more severe level
879 than this one will be capped to this level. This is used to avoid sending
880 "emerg" messages on all terminals on some default syslog configurations.
881 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200882
Cyril Bontédc4d9032012-04-08 21:57:39 +0200883 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +0200884
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100885log-send-hostname [<string>]
886 Sets the hostname field in the syslog header. If optional "string" parameter
887 is set the header is set to the string contents, otherwise uses the hostname
888 of the system. Generally used if one is not relaying logs through an
889 intermediate syslog server or for simply customizing the hostname printed in
890 the logs.
891
Kevinm48936af2010-12-22 16:08:21 +0000892log-tag <string>
893 Sets the tag field in the syslog header to this string. It defaults to the
894 program name as launched from the command line, which usually is "haproxy".
895 Sometimes it can be useful to differentiate between multiple processes
Willy Tarreau094af4e2015-01-07 15:03:42 +0100896 running on the same host. See also the per-proxy "log-tag" directive.
Kevinm48936af2010-12-22 16:08:21 +0000897
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100898lua-load <file>
899 This global directive loads and executes a Lua file. This directive can be
900 used multiple times.
901
William Lallemand4cfede82017-11-24 22:02:34 +0100902master-worker [no-exit-on-failure]
William Lallemande202b1e2017-06-01 17:38:56 +0200903 Master-worker mode. It is equivalent to the command line "-W" argument.
904 This mode will launch a "master" which will monitor the "workers". Using
905 this mode, you can reload HAProxy directly by sending a SIGUSR2 signal to
Davor Ocelice9ed2812017-12-25 17:49:28 +0100906 the master. The master-worker mode is compatible either with the foreground
William Lallemande202b1e2017-06-01 17:38:56 +0200907 or daemon mode. It is recommended to use this mode with multiprocess and
908 systemd.
William Lallemand4cfede82017-11-24 22:02:34 +0100909 By default, if a worker exits with a bad return code, in the case of a
910 segfault for example, all workers will be killed, and the master will leave.
911 It is convenient to combine this behavior with Restart=on-failure in a
912 systemd unit file in order to relaunch the whole process. If you don't want
913 this behavior, you must use the keyword "no-exit-on-failure".
William Lallemande202b1e2017-06-01 17:38:56 +0200914
William Lallemand4cfede82017-11-24 22:02:34 +0100915 See also "-W" in the management guide.
William Lallemande202b1e2017-06-01 17:38:56 +0200916
Willy Tarreau6a06a402007-07-15 20:15:28 +0200917nbproc <number>
918 Creates <number> processes when going daemon. This requires the "daemon"
919 mode. By default, only one process is created, which is the recommended mode
920 of operation. For systems limited to small sets of file descriptors per
921 process, it may be needed to fork multiple daemons. USING MULTIPLE PROCESSES
922 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. See also "daemon".
923
Christopher Fauletbe0faa22017-08-29 15:37:10 +0200924nbthread <number>
925 This setting is only available when support for threads was built in. It
926 creates <number> threads for each created processes. It means if HAProxy is
927 started in foreground, it only creates <number> threads for the first
928 process. FOR NOW, THREADS SUPPORT IN HAPROXY IS HIGHLY EXPERIMENTAL AND IT
929 MUST BE ENABLED WITH CAUTION AND AT YOUR OWN RISK. See also "nbproc".
930
Willy Tarreau6a06a402007-07-15 20:15:28 +0200931pidfile <pidfile>
Davor Ocelice9ed2812017-12-25 17:49:28 +0100932 Writes PIDs of all daemons into file <pidfile>. This option is equivalent to
Willy Tarreau6a06a402007-07-15 20:15:28 +0200933 the "-p" command line argument. The file must be accessible to the user
934 starting the process. See also "daemon".
935
Willy Tarreau1d549722016-02-16 12:41:57 +0100936presetenv <name> <value>
937 Sets environment variable <name> to value <value>. If the variable exists, it
938 is NOT overwritten. The changes immediately take effect so that the next line
939 in the configuration file sees the new value. See also "setenv", "resetenv",
940 and "unsetenv".
941
942resetenv [<name> ...]
943 Removes all environment variables except the ones specified in argument. It
944 allows to use a clean controlled environment before setting new values with
945 setenv or unsetenv. Please note that some internal functions may make use of
946 some environment variables, such as time manipulation functions, but also
947 OpenSSL or even external checks. This must be used with extreme care and only
948 after complete validation. The changes immediately take effect so that the
949 next line in the configuration file sees the new environment. See also
950 "setenv", "presetenv", and "unsetenv".
951
Christopher Fauletff4121f2017-11-22 16:38:49 +0100952stats bind-process [ all | odd | even | <process_num>[-[process_num>]] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +0200953 Limits the stats socket to a certain set of processes numbers. By default the
954 stats socket is bound to all processes, causing a warning to be emitted when
955 nbproc is greater than 1 because there is no way to select the target process
956 when connecting. However, by using this setting, it becomes possible to pin
957 the stats socket to a specific set of processes, typically the first one. The
958 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100959 the number of processes used. The maximum process ID depends on the machine's
Christopher Fauletff4121f2017-11-22 16:38:49 +0100960 word size (32 or 64). Ranges can be partially defined. The higher bound can
961 be omitted. In such case, it is replaced by the corresponding maximum
962 value. A better option consists in using the "process" setting of the "stats
963 socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +0200964
Baptiste Assmann5626f482015-08-23 10:00:10 +0200965server-state-base <directory>
966 Specifies the directory prefix to be prepended in front of all servers state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +0200967 file names which do not start with a '/'. See also "server-state-file",
968 "load-server-state-from-file" and "server-state-file-name".
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200969
970server-state-file <file>
971 Specifies the path to the file containing state of servers. If the path starts
972 with a slash ('/'), it is considered absolute, otherwise it is considered
973 relative to the directory specified using "server-state-base" (if set) or to
974 the current directory. Before reloading HAProxy, it is possible to save the
975 servers' current state using the stats command "show servers state". The
976 output of this command must be written in the file pointed by <file>. When
977 starting up, before handling traffic, HAProxy will read, load and apply state
978 for each server found in the file and available in its current running
Baptiste Assmann01c6cc32015-08-23 11:45:29 +0200979 configuration. See also "server-state-base" and "show servers state",
980 "load-server-state-from-file" and "server-state-file-name"
Baptiste Assmann5626f482015-08-23 10:00:10 +0200981
Willy Tarreau1d549722016-02-16 12:41:57 +0100982setenv <name> <value>
983 Sets environment variable <name> to value <value>. If the variable exists, it
984 is overwritten. The changes immediately take effect so that the next line in
985 the configuration file sees the new value. See also "presetenv", "resetenv",
986 and "unsetenv".
987
Willy Tarreau610f04b2014-02-13 11:36:41 +0100988ssl-default-bind-ciphers <ciphers>
989 This setting is only available when support for OpenSSL was built in. It sets
990 the default string describing the list of cipher algorithms ("cipher suite")
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300991 that are negotiated during the SSL/TLS handshake for all "bind" lines which
Willy Tarreau610f04b2014-02-13 11:36:41 +0100992 do not explicitly define theirs. The format of the string is defined in
993 "man 1 ciphers" from OpenSSL man pages, and can be for instance a string such
994 as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes). Please check the
995 "bind" keyword for more information.
996
Emeric Brun2c86cbf2014-10-30 15:56:50 +0100997ssl-default-bind-options [<option>]...
998 This setting is only available when support for OpenSSL was built in. It sets
999 default ssl-options to force on all "bind" lines. Please check the "bind"
1000 keyword to see available options.
1001
1002 Example:
1003 global
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +02001004 ssl-default-bind-options ssl-min-ver TLSv1.0 no-tls-tickets
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001005
Willy Tarreau610f04b2014-02-13 11:36:41 +01001006ssl-default-server-ciphers <ciphers>
1007 This setting is only available when support for OpenSSL was built in. It
1008 sets the default string describing the list of cipher algorithms that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001009 negotiated during the SSL/TLS handshake with the server, for all "server"
Willy Tarreau610f04b2014-02-13 11:36:41 +01001010 lines which do not explicitly define theirs. The format of the string is
1011 defined in "man 1 ciphers". Please check the "server" keyword for more
1012 information.
1013
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001014ssl-default-server-options [<option>]...
1015 This setting is only available when support for OpenSSL was built in. It sets
1016 default ssl-options to force on all "server" lines. Please check the "server"
1017 keyword to see available options.
1018
Remi Gacogne47783ef2015-05-29 15:53:22 +02001019ssl-dh-param-file <file>
1020 This setting is only available when support for OpenSSL was built in. It sets
1021 the default DH parameters that are used during the SSL/TLS handshake when
1022 ephemeral Diffie-Hellman (DHE) key exchange is used, for all "bind" lines
Davor Ocelice9ed2812017-12-25 17:49:28 +01001023 which do not explicitly define theirs. It will be overridden by custom DH
Remi Gacogne47783ef2015-05-29 15:53:22 +02001024 parameters found in a bind certificate file if any. If custom DH parameters
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001025 are not specified either by using ssl-dh-param-file or by setting them
1026 directly in the certificate file, pre-generated DH parameters of the size
1027 specified by tune.ssl.default-dh-param will be used. Custom parameters are
1028 known to be more secure and therefore their use is recommended.
Remi Gacogne47783ef2015-05-29 15:53:22 +02001029 Custom DH parameters may be generated by using the OpenSSL command
1030 "openssl dhparam <size>", where size should be at least 2048, as 1024-bit DH
1031 parameters should not be considered secure anymore.
1032
Emeric Brun850efd52014-01-29 12:24:34 +01001033ssl-server-verify [none|required]
1034 The default behavior for SSL verify on servers side. If specified to 'none',
1035 servers certificates are not verified. The default is 'required' except if
1036 forced using cmdline option '-dV'.
1037
Willy Tarreauabb175f2012-09-24 12:43:26 +02001038stats socket [<address:port>|<path>] [param*]
1039 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
1040 Connections to this socket will return various statistics outputs and even
1041 allow some commands to be issued to change some runtime settings. Please
Willy Tarreau1af20c72017-06-23 16:01:14 +02001042 consult section 9.3 "Unix Socket commands" of Management Guide for more
Kevin Decherf949c7202015-10-13 23:26:44 +02001043 details.
Willy Tarreau6162db22009-10-10 17:13:00 +02001044
Willy Tarreauabb175f2012-09-24 12:43:26 +02001045 All parameters supported by "bind" lines are supported, for instance to
1046 restrict access to some users or their access rights. Please consult
1047 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001048
1049stats timeout <timeout, in milliseconds>
1050 The default timeout on the stats socket is set to 10 seconds. It is possible
1051 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +01001052 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001053
1054stats maxconn <connections>
1055 By default, the stats socket is limited to 10 concurrent connections. It is
1056 possible to change this value with "stats maxconn".
1057
Willy Tarreau6a06a402007-07-15 20:15:28 +02001058uid <number>
1059 Changes the process' user ID to <number>. It is recommended that the user ID
1060 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
1061 be started with superuser privileges in order to be able to switch to another
1062 one. See also "gid" and "user".
1063
1064ulimit-n <number>
1065 Sets the maximum number of per-process file-descriptors to <number>. By
1066 default, it is automatically computed, so it is recommended not to use this
1067 option.
1068
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001069unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
1070 [ group <group> ] [ gid <gid> ]
1071
1072 Fixes common settings to UNIX listening sockets declared in "bind" statements.
1073 This is mainly used to simplify declaration of those UNIX sockets and reduce
1074 the risk of errors, since those settings are most commonly required but are
1075 also process-specific. The <prefix> setting can be used to force all socket
1076 path to be relative to that directory. This might be needed to access another
1077 component's chroot. Note that those paths are resolved before haproxy chroots
1078 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
1079 all have the same meaning as their homonyms used by the "bind" statement. If
1080 both are specified, the "bind" statement has priority, meaning that the
1081 "unix-bind" settings may be seen as process-wide default settings.
1082
Willy Tarreau1d549722016-02-16 12:41:57 +01001083unsetenv [<name> ...]
1084 Removes environment variables specified in arguments. This can be useful to
1085 hide some sensitive information that are occasionally inherited from the
1086 user's environment during some operations. Variables which did not exist are
1087 silently ignored so that after the operation, it is certain that none of
1088 these variables remain. The changes immediately take effect so that the next
1089 line in the configuration file will not see these variables. See also
1090 "setenv", "presetenv", and "resetenv".
1091
Willy Tarreau6a06a402007-07-15 20:15:28 +02001092user <user name>
1093 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
1094 See also "uid" and "group".
1095
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02001096node <name>
1097 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
1098
1099 This statement is useful in HA configurations where two or more processes or
1100 servers share the same IP address. By setting a different node-name on all
1101 nodes, it becomes easy to immediately spot what server is handling the
1102 traffic.
1103
1104description <text>
1105 Add a text that describes the instance.
1106
1107 Please note that it is required to escape certain characters (# for example)
1108 and this text is inserted into a html page so you should avoid using
1109 "<" and ">" characters.
1110
Thomas Holmesdb04f192015-05-18 13:21:39 +0100111151degrees-data-file <file path>
1112 The path of the 51Degrees data file to provide device detection services. The
Davor Ocelice9ed2812017-12-25 17:49:28 +01001113 file should be unzipped and accessible by HAProxy with relevant permissions.
Thomas Holmesdb04f192015-05-18 13:21:39 +01001114
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001115 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001116 compiled with USE_51DEGREES.
1117
Ben Shillitof25e8e52016-12-02 14:25:37 +0000111851degrees-property-name-list [<string> ...]
Thomas Holmesdb04f192015-05-18 13:21:39 +01001119 A list of 51Degrees property names to be load from the dataset. A full list
1120 of names is available on the 51Degrees website:
1121 https://51degrees.com/resources/property-dictionary
1122
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001123 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001124 compiled with USE_51DEGREES.
1125
Dragan Dosen93b38d92015-06-29 16:43:25 +0200112651degrees-property-separator <char>
Thomas Holmesdb04f192015-05-18 13:21:39 +01001127 A char that will be appended to every property value in a response header
1128 containing 51Degrees results. If not set that will be set as ','.
1129
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001130 Please note that this option is only available when haproxy has been
1131 compiled with USE_51DEGREES.
1132
113351degrees-cache-size <number>
1134 Sets the size of the 51Degrees converter cache to <number> entries. This
1135 is an LRU cache which reminds previous device detections and their results.
1136 By default, this cache is disabled.
1137
1138 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001139 compiled with USE_51DEGREES.
1140
scientiamobiled0027ed2016-11-04 10:55:08 +01001141wurfl-data-file <file path>
1142 The path of the WURFL data file to provide device detection services. The
1143 file should be accessible by HAProxy with relevant permissions.
1144
1145 Please note that this option is only available when haproxy has been compiled
1146 with USE_WURFL=1.
1147
1148wurfl-information-list [<capability>]*
1149 A space-delimited list of WURFL capabilities, virtual capabilities, property
1150 names we plan to use in injected headers. A full list of capability and
1151 virtual capability names is available on the Scientiamobile website :
1152
1153 https://www.scientiamobile.com/wurflCapability
1154
1155 Valid WURFL properties are:
1156 - wurfl_id Contains the device ID of the matched device.
1157
1158 - wurfl_root_id Contains the device root ID of the matched
1159 device.
1160
1161 - wurfl_isdevroot Tells if the matched device is a root device.
1162 Possible values are "TRUE" or "FALSE".
1163
1164 - wurfl_useragent The original useragent coming with this
1165 particular web request.
1166
1167 - wurfl_api_version Contains a string representing the currently
1168 used Libwurfl API version.
1169
1170 - wurfl_engine_target Contains a string representing the currently
1171 set WURFL Engine Target. Possible values are
1172 "HIGH_ACCURACY", "HIGH_PERFORMANCE", "INVALID".
1173
1174 - wurfl_info A string containing information on the parsed
1175 wurfl.xml and its full path.
1176
1177 - wurfl_last_load_time Contains the UNIX timestamp of the last time
1178 WURFL has been loaded successfully.
1179
1180 - wurfl_normalized_useragent The normalized useragent.
1181
1182 - wurfl_useragent_priority The user agent priority used by WURFL.
1183
1184 Please note that this option is only available when haproxy has been compiled
1185 with USE_WURFL=1.
1186
1187wurfl-information-list-separator <char>
1188 A char that will be used to separate values in a response header containing
1189 WURFL results. If not set that a comma (',') will be used by default.
1190
1191 Please note that this option is only available when haproxy has been compiled
1192 with USE_WURFL=1.
1193
1194wurfl-patch-file [<file path>]
1195 A list of WURFL patch file paths. Note that patches are loaded during startup
1196 thus before the chroot.
1197
1198 Please note that this option is only available when haproxy has been compiled
1199 with USE_WURFL=1.
1200
1201wurfl-engine-mode { accuracy | performance }
1202 Sets the WURFL engine target. You can choose between 'accuracy' or
1203 'performance' targets. In performance mode, desktop web browser detection is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001204 done programmatically without referencing the WURFL data. As a result, most
scientiamobiled0027ed2016-11-04 10:55:08 +01001205 desktop web browsers are returned as generic_web_browser WURFL ID for
Davor Ocelice9ed2812017-12-25 17:49:28 +01001206 performance. If either performance or accuracy are not defined, performance
scientiamobiled0027ed2016-11-04 10:55:08 +01001207 mode is enabled by default.
1208
1209 Please note that this option is only available when haproxy has been compiled
1210 with USE_WURFL=1.
1211
1212wurfl-cache-size <U>[,<D>]
1213 Sets the WURFL caching strategy. Here <U> is the Useragent cache size, and
1214 <D> is the internal device cache size. There are three possibilities here :
1215 - "0" : no cache is used.
1216 - <U> : the Single LRU cache is used, the size is expressed in elements.
1217 - <U>,<D> : the Double LRU cache is used, both sizes are in elements. This is
1218 the highest performing option.
1219
1220 Please note that this option is only available when haproxy has been compiled
1221 with USE_WURFL=1.
1222
1223wurfl-useragent-priority { plain | sideloaded_browser }
1224 Tells WURFL if it should prioritize use of the plain user agent ('plain')
1225 over the default sideloaded browser user agent ('sideloaded_browser').
1226
1227 Please note that this option is only available when haproxy has been compiled
1228 with USE_WURFL=1.
1229
Willy Tarreau6a06a402007-07-15 20:15:28 +02001230
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012313.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +02001232-----------------------
1233
Willy Tarreau1746eec2014-04-25 10:46:47 +02001234max-spread-checks <delay in milliseconds>
1235 By default, haproxy tries to spread the start of health checks across the
1236 smallest health check interval of all the servers in a farm. The principle is
1237 to avoid hammering services running on the same server. But when using large
1238 check intervals (10 seconds or more), the last servers in the farm take some
1239 time before starting to be tested, which can be a problem. This parameter is
1240 used to enforce an upper bound on delay between the first and the last check,
1241 even if the servers' check intervals are larger. When servers run with
1242 shorter intervals, their intervals will be respected though.
1243
Willy Tarreau6a06a402007-07-15 20:15:28 +02001244maxconn <number>
1245 Sets the maximum per-process number of concurrent connections to <number>. It
1246 is equivalent to the command-line argument "-n". Proxies will stop accepting
1247 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +02001248 automatically adjusted according to this value. See also "ulimit-n". Note:
1249 the "select" poller cannot reliably use more than 1024 file descriptors on
1250 some platforms. If your platform only supports select and reports "select
1251 FAILED" on startup, you need to reduce maxconn until it works (slightly
Willy Tarreaud0256482015-01-15 21:45:22 +01001252 below 500 in general). If this value is not set, it will default to the value
1253 set in DEFAULT_MAXCONN at build time (reported in haproxy -vv) if no memory
1254 limit is enforced, or will be computed based on the memory limit, the buffer
1255 size, memory allocated to compression, SSL cache size, and use or not of SSL
1256 and the associated maxsslconn (which can also be automatic).
Willy Tarreau6a06a402007-07-15 20:15:28 +02001257
Willy Tarreau81c25d02011-09-07 15:17:21 +02001258maxconnrate <number>
1259 Sets the maximum per-process number of connections per second to <number>.
1260 Proxies will stop accepting connections when this limit is reached. It can be
1261 used to limit the global capacity regardless of each frontend capacity. It is
1262 important to note that this can only be used as a service protection measure,
1263 as there will not necessarily be a fair share between frontends when the
1264 limit is reached, so it's a good idea to also limit each frontend to some
1265 value close to its expected share. Also, lowering tune.maxaccept can improve
1266 fairness.
1267
William Lallemandd85f9172012-11-09 17:05:39 +01001268maxcomprate <number>
1269 Sets the maximum per-process input compression rate to <number> kilobytes
Davor Ocelice9ed2812017-12-25 17:49:28 +01001270 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +01001271 level will be decreased during the session. If the maximum is reached at the
1272 beginning of a session, the session will not compress at all. If the maximum
1273 is not reached, the compression level will be increased up to
Davor Ocelice9ed2812017-12-25 17:49:28 +01001274 tune.comp.maxlevel. A value of zero means there is no limit, this is the
William Lallemandd85f9172012-11-09 17:05:39 +01001275 default value.
1276
William Lallemand072a2bf2012-11-20 17:01:01 +01001277maxcompcpuusage <number>
1278 Sets the maximum CPU usage HAProxy can reach before stopping the compression
1279 for new requests or decreasing the compression level of current requests.
1280 It works like 'maxcomprate' but measures CPU usage instead of incoming data
1281 bandwidth. The value is expressed in percent of the CPU used by haproxy. In
1282 case of multiple processes (nbproc > 1), each process manages its individual
1283 usage. A value of 100 disable the limit. The default value is 100. Setting
1284 a lower value will prevent the compression work from slowing the whole
1285 process down and from introducing high latencies.
1286
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001287maxpipes <number>
1288 Sets the maximum per-process number of pipes to <number>. Currently, pipes
1289 are only used by kernel-based tcp splicing. Since a pipe contains two file
1290 descriptors, the "ulimit-n" value will be increased accordingly. The default
1291 value is maxconn/4, which seems to be more than enough for most heavy usages.
1292 The splice code dynamically allocates and releases pipes, and can fall back
1293 to standard copy, so setting this value too low may only impact performance.
1294
Willy Tarreau93e7c002013-10-07 18:51:07 +02001295maxsessrate <number>
1296 Sets the maximum per-process number of sessions per second to <number>.
1297 Proxies will stop accepting connections when this limit is reached. It can be
1298 used to limit the global capacity regardless of each frontend capacity. It is
1299 important to note that this can only be used as a service protection measure,
1300 as there will not necessarily be a fair share between frontends when the
1301 limit is reached, so it's a good idea to also limit each frontend to some
1302 value close to its expected share. Also, lowering tune.maxaccept can improve
1303 fairness.
1304
Willy Tarreau403edff2012-09-06 11:58:37 +02001305maxsslconn <number>
1306 Sets the maximum per-process number of concurrent SSL connections to
1307 <number>. By default there is no SSL-specific limit, which means that the
1308 global maxconn setting will apply to all connections. Setting this limit
1309 avoids having openssl use too much memory and crash when malloc returns NULL
1310 (since it unfortunately does not reliably check for such conditions). Note
1311 that the limit applies both to incoming and outgoing connections, so one
1312 connection which is deciphered then ciphered accounts for 2 SSL connections.
Willy Tarreaud0256482015-01-15 21:45:22 +01001313 If this value is not set, but a memory limit is enforced, this value will be
1314 automatically computed based on the memory limit, maxconn, the buffer size,
1315 memory allocated to compression, SSL cache size, and use of SSL in either
1316 frontends, backends or both. If neither maxconn nor maxsslconn are specified
1317 when there is a memory limit, haproxy will automatically adjust these values
1318 so that 100% of the connections can be made over SSL with no risk, and will
1319 consider the sides where it is enabled (frontend, backend, both).
Willy Tarreau403edff2012-09-06 11:58:37 +02001320
Willy Tarreaue43d5322013-10-07 20:01:52 +02001321maxsslrate <number>
1322 Sets the maximum per-process number of SSL sessions per second to <number>.
1323 SSL listeners will stop accepting connections when this limit is reached. It
1324 can be used to limit the global SSL CPU usage regardless of each frontend
1325 capacity. It is important to note that this can only be used as a service
1326 protection measure, as there will not necessarily be a fair share between
1327 frontends when the limit is reached, so it's a good idea to also limit each
1328 frontend to some value close to its expected share. It is also important to
1329 note that the sessions are accounted before they enter the SSL stack and not
1330 after, which also protects the stack against bad handshakes. Also, lowering
1331 tune.maxaccept can improve fairness.
1332
William Lallemand9d5f5482012-11-07 16:12:57 +01001333maxzlibmem <number>
1334 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
1335 When the maximum amount is reached, future sessions will not compress as long
1336 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +01001337 The default value is 0. The value is available in bytes on the UNIX socket
1338 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
1339 "ZlibMemUsage" in bytes.
1340
Willy Tarreau6a06a402007-07-15 20:15:28 +02001341noepoll
1342 Disables the use of the "epoll" event polling system on Linux. It is
1343 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +01001344 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001345
1346nokqueue
1347 Disables the use of the "kqueue" event polling system on BSD. It is
1348 equivalent to the command-line argument "-dk". The next polling system
1349 used will generally be "poll". See also "nopoll".
1350
1351nopoll
1352 Disables the use of the "poll" event polling system. It is equivalent to the
1353 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001354 It should never be needed to disable "poll" since it's available on all
Willy Tarreaue9f49e72012-11-11 17:42:00 +01001355 platforms supported by HAProxy. See also "nokqueue" and "noepoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001356
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001357nosplice
1358 Disables the use of kernel tcp splicing between sockets on Linux. It is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001359 equivalent to the command line argument "-dS". Data will then be copied
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001360 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001361 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001362 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
1363 be used. This option makes it easier to globally disable kernel splicing in
1364 case of doubt. See also "option splice-auto", "option splice-request" and
1365 "option splice-response".
1366
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001367nogetaddrinfo
1368 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
1369 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
1370
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00001371noreuseport
1372 Disables the use of SO_REUSEPORT - see socket(7). It is equivalent to the
1373 command line argument "-dR".
1374
Willy Tarreaufe255b72007-10-14 23:09:26 +02001375spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +09001376 Sometimes it is desirable to avoid sending agent and health checks to
1377 servers at exact intervals, for instance when many logical servers are
1378 located on the same physical server. With the help of this parameter, it
1379 becomes possible to add some randomness in the check interval between 0
1380 and +/- 50%. A value between 2 and 5 seems to show good results. The
1381 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +02001382
Davor Ocelice9ed2812017-12-25 17:49:28 +01001383ssl-engine <name> [algo <comma-separated list of algorithms>]
Grant Zhang872f9c22017-01-21 01:10:18 +00001384 Sets the OpenSSL engine to <name>. List of valid values for <name> may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01001385 obtained using the command "openssl engine". This statement may be used
Grant Zhang872f9c22017-01-21 01:10:18 +00001386 multiple times, it will simply enable multiple crypto engines. Referencing an
1387 unsupported engine will prevent haproxy from starting. Note that many engines
1388 will lead to lower HTTPS performance than pure software with recent
1389 processors. The optional command "algo" sets the default algorithms an ENGINE
1390 will supply using the OPENSSL function ENGINE_set_default_string(). A value
Davor Ocelice9ed2812017-12-25 17:49:28 +01001391 of "ALL" uses the engine for all cryptographic operations. If no list of
1392 algo is specified then the value of "ALL" is used. A comma-separated list
Grant Zhang872f9c22017-01-21 01:10:18 +00001393 of different algorithms may be specified, including: RSA, DSA, DH, EC, RAND,
1394 CIPHERS, DIGESTS, PKEY, PKEY_CRYPTO, PKEY_ASN1. This is the same format that
1395 openssl configuration file uses:
1396 https://www.openssl.org/docs/man1.0.2/apps/config.html
1397
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00001398ssl-mode-async
1399 Adds SSL_MODE_ASYNC mode to the SSL context. This enables asynchronous TLS
Emeric Brun3854e012017-05-17 20:42:48 +02001400 I/O operations if asynchronous capable SSL engines are used. The current
Emeric Brunb5e42a82017-06-06 12:35:14 +00001401 implementation supports a maximum of 32 engines. The Openssl ASYNC API
1402 doesn't support moving read/write buffers and is not compliant with
1403 haproxy's buffer management. So the asynchronous mode is disabled on
1404 read/write operations (it is only enabled during initial and reneg
1405 handshakes).
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00001406
Willy Tarreau33cb0652014-12-23 22:52:37 +01001407tune.buffers.limit <number>
1408 Sets a hard limit on the number of buffers which may be allocated per process.
1409 The default value is zero which means unlimited. The minimum non-zero value
1410 will always be greater than "tune.buffers.reserve" and should ideally always
1411 be about twice as large. Forcing this value can be particularly useful to
1412 limit the amount of memory a process may take, while retaining a sane
Davor Ocelice9ed2812017-12-25 17:49:28 +01001413 behavior. When this limit is reached, sessions which need a buffer wait for
Willy Tarreau33cb0652014-12-23 22:52:37 +01001414 another one to be released by another session. Since buffers are dynamically
1415 allocated and released, the waiting time is very short and not perceptible
1416 provided that limits remain reasonable. In fact sometimes reducing the limit
1417 may even increase performance by increasing the CPU cache's efficiency. Tests
1418 have shown good results on average HTTP traffic with a limit to 1/10 of the
1419 expected global maxconn setting, which also significantly reduces memory
1420 usage. The memory savings come from the fact that a number of connections
1421 will not allocate 2*tune.bufsize. It is best not to touch this value unless
1422 advised to do so by an haproxy core developer.
1423
Willy Tarreau1058ae72014-12-23 22:40:40 +01001424tune.buffers.reserve <number>
1425 Sets the number of buffers which are pre-allocated and reserved for use only
1426 during memory shortage conditions resulting in failed memory allocations. The
1427 minimum value is 2 and is also the default. There is no reason a user would
1428 want to change this value, it's mostly aimed at haproxy core developers.
1429
Willy Tarreau27a674e2009-08-17 07:23:33 +02001430tune.bufsize <number>
1431 Sets the buffer size to this size (in bytes). Lower values allow more
1432 sessions to coexist in the same amount of RAM, and higher values allow some
1433 applications with very large cookies to work. The default value is 16384 and
1434 can be changed at build time. It is strongly recommended not to change this
1435 from the default value, as very low values will break some services such as
1436 statistics, and values larger than default size will increase memory usage,
1437 possibly causing the system to run out of memory. At least the global maxconn
Willy Tarreau45a66cc2017-11-24 11:28:00 +01001438 parameter should be decreased by the same factor as this one is increased. In
1439 addition, use of HTTP/2 mandates that this value must be 16384 or more. If an
1440 HTTP request is larger than (tune.bufsize - tune.maxrewrite), haproxy will
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04001441 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
1442 than this size, haproxy will return HTTP 502 (Bad Gateway).
Willy Tarreau27a674e2009-08-17 07:23:33 +02001443
Willy Tarreau43961d52010-10-04 20:39:20 +02001444tune.chksize <number>
1445 Sets the check buffer size to this size (in bytes). Higher values may help
1446 find string or regex patterns in very large pages, though doing so may imply
1447 more memory and CPU usage. The default value is 16384 and can be changed at
1448 build time. It is not recommended to change this value, but to use better
1449 checks whenever possible.
1450
William Lallemandf3747832012-11-09 12:33:10 +01001451tune.comp.maxlevel <number>
1452 Sets the maximum compression level. The compression level affects CPU
1453 usage during compression. This value affects CPU usage during compression.
1454 Each session using compression initializes the compression algorithm with
1455 this value. The default value is 1.
1456
Willy Tarreaufe20e5b2017-07-27 11:42:14 +02001457tune.h2.header-table-size <number>
1458 Sets the HTTP/2 dynamic header table size. It defaults to 4096 bytes and
1459 cannot be larger than 65536 bytes. A larger value may help certain clients
1460 send more compact requests, depending on their capabilities. This amount of
1461 memory is consumed for each HTTP/2 connection. It is recommended not to
1462 change it.
1463
Willy Tarreaue6baec02017-07-27 11:45:11 +02001464tune.h2.initial-window-size <number>
1465 Sets the HTTP/2 initial window size, which is the number of bytes the client
Davor Ocelice9ed2812017-12-25 17:49:28 +01001466 can upload before waiting for an acknowledgment from haproxy. This setting
1467 only affects payload contents (i.e. the body of POST requests), not headers.
Willy Tarreaue6baec02017-07-27 11:45:11 +02001468 The default value is 65535, which roughly allows up to 5 Mbps of upload
1469 bandwidth per client over a network showing a 100 ms ping time, or 500 Mbps
1470 over a 1-ms local network. It can make sense to increase this value to allow
1471 faster uploads, or to reduce it to increase fairness when dealing with many
1472 clients. It doesn't affect resource usage.
1473
Willy Tarreau5242ef82017-07-27 11:47:28 +02001474tune.h2.max-concurrent-streams <number>
1475 Sets the HTTP/2 maximum number of concurrent streams per connection (ie the
1476 number of outstanding requests on a single connection). The default value is
1477 100. A larger one may slightly improve page load time for complex sites when
1478 visited over high latency networks, but increases the amount of resources a
1479 single client may allocate. A value of zero disables the limit so a single
1480 client may create as many streams as allocatable by haproxy. It is highly
1481 recommended not to change this value.
1482
Willy Tarreau193b8c62012-11-22 00:17:38 +01001483tune.http.cookielen <number>
1484 Sets the maximum length of captured cookies. This is the maximum value that
1485 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
1486 will automatically be truncated to this one. It is important not to set too
1487 high a value because all cookie captures still allocate this size whatever
1488 their configured value (they share a same pool). This value is per request
1489 per response, so the memory allocated is twice this value per connection.
1490 When not specified, the limit is set to 63 characters. It is recommended not
1491 to change this value.
1492
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001493tune.http.logurilen <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001494 Sets the maximum length of request URI in logs. This prevents truncating long
1495 request URIs with valuable query strings in log lines. This is not related
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001496 to syslog limits. If you increase this limit, you may also increase the
Davor Ocelice9ed2812017-12-25 17:49:28 +01001497 'log ... len yyy' parameter. Your syslog daemon may also need specific
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001498 configuration directives too.
1499 The default value is 1024.
1500
Willy Tarreauac1932d2011-10-24 19:14:41 +02001501tune.http.maxhdr <number>
1502 Sets the maximum number of headers in a request. When a request comes with a
1503 number of headers greater than this value (including the first line), it is
1504 rejected with a "400 Bad Request" status code. Similarly, too large responses
1505 are blocked with "502 Bad Gateway". The default value is 101, which is enough
1506 for all usages, considering that the widely deployed Apache server uses the
1507 same limit. It can be useful to push this limit further to temporarily allow
Christopher Faulet50174f32017-06-21 16:31:35 +02001508 a buggy application to work by the time it gets fixed. The accepted range is
1509 1..32767. Keep in mind that each new header consumes 32bits of memory for
1510 each session, so don't push this limit too high.
Willy Tarreauac1932d2011-10-24 19:14:41 +02001511
Willy Tarreau7e312732014-02-12 16:35:14 +01001512tune.idletimer <timeout>
1513 Sets the duration after which haproxy will consider that an empty buffer is
1514 probably associated with an idle stream. This is used to optimally adjust
1515 some packet sizes while forwarding large and small data alternatively. The
1516 decision to use splice() or to send large buffers in SSL is modulated by this
1517 parameter. The value is in milliseconds between 0 and 65535. A value of zero
1518 means that haproxy will not try to detect idle streams. The default is 1000,
Davor Ocelice9ed2812017-12-25 17:49:28 +01001519 which seems to correctly detect end user pauses (e.g. read a page before
Willy Tarreau7e312732014-02-12 16:35:14 +01001520 clicking). There should be not reason for changing this value. Please check
1521 tune.ssl.maxrecord below.
1522
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001523tune.lua.forced-yield <number>
1524 This directive forces the Lua engine to execute a yield each <number> of
Tim Düsterhus4896c442016-11-29 02:15:19 +01001525 instructions executed. This permits interrupting a long script and allows the
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001526 HAProxy scheduler to process other tasks like accepting connections or
1527 forwarding traffic. The default value is 10000 instructions. If HAProxy often
Davor Ocelice9ed2812017-12-25 17:49:28 +01001528 executes some Lua code but more responsiveness is required, this value can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001529 lowered. If the Lua code is quite long and its result is absolutely required
1530 to process the data, the <number> can be increased.
1531
Willy Tarreau32f61e22015-03-18 17:54:59 +01001532tune.lua.maxmem
1533 Sets the maximum amount of RAM in megabytes per process usable by Lua. By
1534 default it is zero which means unlimited. It is important to set a limit to
1535 ensure that a bug in a script will not result in the system running out of
1536 memory.
1537
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001538tune.lua.session-timeout <timeout>
1539 This is the execution timeout for the Lua sessions. This is useful for
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001540 preventing infinite loops or spending too much time in Lua. This timeout
1541 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001542 not taken in account. The default timeout is 4s.
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001543
1544tune.lua.task-timeout <timeout>
1545 Purpose is the same as "tune.lua.session-timeout", but this timeout is
1546 dedicated to the tasks. By default, this timeout isn't set because a task may
1547 remain alive during of the lifetime of HAProxy. For example, a task used to
1548 check servers.
1549
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001550tune.lua.service-timeout <timeout>
1551 This is the execution timeout for the Lua services. This is useful for
1552 preventing infinite loops or spending too much time in Lua. This timeout
1553 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001554 not taken in account. The default timeout is 4s.
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001555
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001556tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +01001557 Sets the maximum number of consecutive connections a process may accept in a
1558 row before switching to other work. In single process mode, higher numbers
1559 give better performance at high connection rates. However in multi-process
1560 modes, keeping a bit of fairness between processes generally is better to
1561 increase performance. This value applies individually to each listener, so
1562 that the number of processes a listener is bound to is taken into account.
1563 This value defaults to 64. In multi-process mode, it is divided by twice
1564 the number of processes the listener is bound to. Setting this value to -1
1565 completely disables the limitation. It should normally not be needed to tweak
1566 this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001567
1568tune.maxpollevents <number>
1569 Sets the maximum amount of events that can be processed at once in a call to
1570 the polling system. The default value is adapted to the operating system. It
1571 has been noticed that reducing it below 200 tends to slightly decrease
1572 latency at the expense of network bandwidth, and increasing it above 200
1573 tends to trade latency for slightly increased bandwidth.
1574
Willy Tarreau27a674e2009-08-17 07:23:33 +02001575tune.maxrewrite <number>
1576 Sets the reserved buffer space to this size in bytes. The reserved space is
1577 used for header rewriting or appending. The first reads on sockets will never
1578 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
1579 bufsize, though that does not make much sense since there are rarely large
1580 numbers of headers to add. Setting it too high prevents processing of large
1581 requests or responses. Setting it too low prevents addition of new headers
1582 to already large requests or to POST requests. It is generally wise to set it
1583 to about 1024. It is automatically readjusted to half of bufsize if it is
1584 larger than that. This means you don't have to worry about it when changing
1585 bufsize.
1586
Willy Tarreauf3045d22015-04-29 16:24:50 +02001587tune.pattern.cache-size <number>
1588 Sets the size of the pattern lookup cache to <number> entries. This is an LRU
1589 cache which reminds previous lookups and their results. It is used by ACLs
1590 and maps on slow pattern lookups, namely the ones using the "sub", "reg",
1591 "dir", "dom", "end", "bin" match methods as well as the case-insensitive
1592 strings. It applies to pattern expressions which means that it will be able
1593 to memorize the result of a lookup among all the patterns specified on a
1594 configuration line (including all those loaded from files). It automatically
1595 invalidates entries which are updated using HTTP actions or on the CLI. The
1596 default cache size is set to 10000 entries, which limits its footprint to
1597 about 5 MB on 32-bit systems and 8 MB on 64-bit systems. There is a very low
1598 risk of collision in this cache, which is in the order of the size of the
1599 cache divided by 2^64. Typically, at 10000 requests per second with the
1600 default cache size of 10000 entries, there's 1% chance that a brute force
1601 attack could cause a single collision after 60 years, or 0.1% after 6 years.
1602 This is considered much lower than the risk of a memory corruption caused by
1603 aging components. If this is not acceptable, the cache can be disabled by
1604 setting this parameter to 0.
1605
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02001606tune.pipesize <number>
1607 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
1608 are the default size for the system. But sometimes when using TCP splicing,
1609 it can improve performance to increase pipe sizes, especially if it is
1610 suspected that pipes are not filled and that many calls to splice() are
1611 performed. This has an impact on the kernel's memory footprint, so this must
1612 not be changed if impacts are not understood.
1613
Willy Tarreaue803de22010-01-21 17:43:04 +01001614tune.rcvbuf.client <number>
1615tune.rcvbuf.server <number>
1616 Forces the kernel socket receive buffer size on the client or the server side
1617 to the specified value in bytes. This value applies to all TCP/HTTP frontends
1618 and backends. It should normally never be set, and the default size (0) lets
1619 the kernel autotune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001620 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01001621 order to save kernel memory by preventing it from buffering too large amounts
1622 of received data. Lower values will significantly increase CPU usage though.
1623
Willy Tarreaub22fc302015-12-14 12:04:35 +01001624tune.recv_enough <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001625 HAProxy uses some hints to detect that a short read indicates the end of the
Willy Tarreaub22fc302015-12-14 12:04:35 +01001626 socket buffers. One of them is that a read returns more than <recv_enough>
1627 bytes, which defaults to 10136 (7 segments of 1448 each). This default value
1628 may be changed by this setting to better deal with workloads involving lots
1629 of short messages such as telnet or SSH sessions.
1630
Olivier Houchard1599b802018-05-24 18:59:04 +02001631tune.runqueue-depth <number>
1632 Sets the maxinum amount of task that can be processed at once when running
1633 tasks. The default value is 200. Increasing it may incur latency when
1634 dealing with I/Os, making it too small can incur extra overhead.
1635
Willy Tarreaue803de22010-01-21 17:43:04 +01001636tune.sndbuf.client <number>
1637tune.sndbuf.server <number>
1638 Forces the kernel socket send buffer size on the client or the server side to
1639 the specified value in bytes. This value applies to all TCP/HTTP frontends
1640 and backends. It should normally never be set, and the default size (0) lets
1641 the kernel autotune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001642 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01001643 order to save kernel memory by preventing it from buffering too large amounts
1644 of received data. Lower values will significantly increase CPU usage though.
1645 Another use case is to prevent write timeouts with extremely slow clients due
1646 to the kernel waiting for a large part of the buffer to be read before
1647 notifying haproxy again.
1648
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001649tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +01001650 Sets the size of the global SSL session cache, in a number of blocks. A block
1651 is large enough to contain an encoded session without peer certificate.
1652 An encoded session with peer certificate is stored in multiple blocks
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001653 depending on the size of the peer certificate. A block uses approximately
Emeric Brunaf9619d2012-11-28 18:47:52 +01001654 200 bytes of memory. The default value may be forced at build time, otherwise
Davor Ocelice9ed2812017-12-25 17:49:28 +01001655 defaults to 20000. When the cache is full, the most idle entries are purged
Emeric Brunaf9619d2012-11-28 18:47:52 +01001656 and reassigned. Higher values reduce the occurrence of such a purge, hence
1657 the number of CPU-intensive SSL handshakes by ensuring that all users keep
1658 their session as long as possible. All entries are pre-allocated upon startup
Emeric Brun22890a12012-12-28 14:41:32 +01001659 and are shared between all processes if "nbproc" is greater than 1. Setting
1660 this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001661
Emeric Brun8dc60392014-05-09 13:52:00 +02001662tune.ssl.force-private-cache
1663 This boolean disables SSL session cache sharing between all processes. It
1664 should normally not be used since it will force many renegotiations due to
1665 clients hitting a random process. But it may be required on some operating
1666 systems where none of the SSL cache synchronization method may be used. In
1667 this case, adding a first layer of hash-based load balancing before the SSL
1668 layer might limit the impact of the lack of session sharing.
1669
Emeric Brun4f65bff2012-11-16 15:11:00 +01001670tune.ssl.lifetime <timeout>
1671 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001672 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01001673 does not guarantee that sessions will last that long, because if the cache is
1674 full, the longest idle sessions will be purged despite their configured
1675 lifetime. The real usefulness of this setting is to prevent sessions from
1676 being used for too long.
1677
Willy Tarreaubfd59462013-02-21 07:46:09 +01001678tune.ssl.maxrecord <number>
1679 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
1680 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
1681 data only once it has received a full record. With large records, it means
1682 that clients might have to download up to 16kB of data before starting to
1683 process them. Limiting the value can improve page load times on browsers
1684 located over high latency or low bandwidth networks. It is suggested to find
1685 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
1686 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
1687 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
1688 2859 gave good results during tests. Use "strace -e trace=write" to find the
Davor Ocelice9ed2812017-12-25 17:49:28 +01001689 best value. HAProxy will automatically switch to this setting after an idle
Willy Tarreau7e312732014-02-12 16:35:14 +01001690 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01001691
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001692tune.ssl.default-dh-param <number>
1693 Sets the maximum size of the Diffie-Hellman parameters used for generating
1694 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
1695 final size will try to match the size of the server's RSA (or DSA) key (e.g,
1696 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
1697 this maximum value. Default value if 1024. Only 1024 or higher values are
1698 allowed. Higher values will increase the CPU load, and values greater than
1699 1024 bits are not supported by Java 7 and earlier clients. This value is not
Remi Gacogne47783ef2015-05-29 15:53:22 +02001700 used if static Diffie-Hellman parameters are supplied either directly
1701 in the certificate file or by using the ssl-dh-param-file parameter.
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001702
Christopher Faulet31af49d2015-06-09 17:29:50 +02001703tune.ssl.ssl-ctx-cache-size <number>
1704 Sets the size of the cache used to store generated certificates to <number>
1705 entries. This is a LRU cache. Because generating a SSL certificate
1706 dynamically is expensive, they are cached. The default cache size is set to
1707 1000 entries.
1708
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01001709tune.ssl.capture-cipherlist-size <number>
1710 Sets the maximum size of the buffer used for capturing client-hello cipher
1711 list. If the value is 0 (default value) the capture is disabled, otherwise
1712 a buffer is allocated for each SSL/TLS connection.
1713
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001714tune.vars.global-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001715tune.vars.proc-max-size <size>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001716tune.vars.reqres-max-size <size>
1717tune.vars.sess-max-size <size>
1718tune.vars.txn-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001719 These five tunes help to manage the maximum amount of memory used by the
1720 variables system. "global" limits the overall amount of memory available for
1721 all scopes. "proc" limits the memory for the process scope, "sess" limits the
1722 memory for the session scope, "txn" for the transaction scope, and "reqres"
1723 limits the memory for each request or response processing.
1724 Memory accounting is hierarchical, meaning more coarse grained limits include
1725 the finer grained ones: "proc" includes "sess", "sess" includes "txn", and
1726 "txn" includes "reqres".
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001727
Daniel Schneller0b547052016-03-21 20:46:57 +01001728 For example, when "tune.vars.sess-max-size" is limited to 100,
1729 "tune.vars.txn-max-size" and "tune.vars.reqres-max-size" cannot exceed
1730 100 either. If we create a variable "txn.var" that contains 100 bytes,
1731 all available space is consumed.
1732 Notice that exceeding the limits at runtime will not result in an error
1733 message, but values might be cut off or corrupted. So make sure to accurately
1734 plan for the amount of space needed to store all your variables.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001735
William Lallemanda509e4c2012-11-07 16:54:34 +01001736tune.zlib.memlevel <number>
1737 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001738 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01001739 state. A value of 1 uses minimum memory but is slow and reduces compression
Davor Ocelice9ed2812017-12-25 17:49:28 +01001740 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
William Lallemanda509e4c2012-11-07 16:54:34 +01001741 between 1 and 9. The default value is 8.
1742
1743tune.zlib.windowsize <number>
1744 Sets the window size (the size of the history buffer) as a parameter of the
1745 zlib initialization for each session. Larger values of this parameter result
Davor Ocelice9ed2812017-12-25 17:49:28 +01001746 in better compression at the expense of memory usage. Can be a value between
1747 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001748
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017493.3. Debugging
1750--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02001751
1752debug
1753 Enables debug mode which dumps to stdout all exchanges, and disables forking
1754 into background. It is the equivalent of the command-line argument "-d". It
1755 should never be used in a production configuration since it may prevent full
1756 system startup.
1757
1758quiet
1759 Do not display any message during startup. It is equivalent to the command-
1760 line argument "-q".
1761
Emeric Brunf099e792010-09-27 12:05:28 +02001762
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010017633.4. Userlists
1764--------------
1765It is possible to control access to frontend/backend/listen sections or to
1766http stats by allowing only authenticated and authorized users. To do this,
1767it is required to create at least one userlist and to define users.
1768
1769userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01001770 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001771 used to store authentication & authorization data for independent customers.
1772
1773group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01001774 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001775 attach users to this group by using a comma separated list of names
1776 proceeded by "users" keyword.
1777
Cyril Bontéf0c60612010-02-06 14:44:47 +01001778user <username> [password|insecure-password <password>]
1779 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001780 Adds user <username> to the current userlist. Both secure (encrypted) and
1781 insecure (unencrypted) passwords can be used. Encrypted passwords are
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01001782 evaluated using the crypt(3) function, so depending on the system's
1783 capabilities, different algorithms are supported. For example, modern Glibc
1784 based Linux systems support MD5, SHA-256, SHA-512, and, of course, the
1785 classic DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001786
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01001787 Attention: Be aware that using encrypted passwords might cause significantly
1788 increased CPU usage, depending on the number of requests, and the algorithm
1789 used. For any of the hashed variants, the password for each request must
1790 be processed through the chosen algorithm, before it can be compared to the
1791 value specified in the config file. Most current algorithms are deliberately
1792 designed to be expensive to compute to achieve resistance against brute
1793 force attacks. They do not simply salt/hash the clear text password once,
1794 but thousands of times. This can quickly become a major factor in haproxy's
1795 overall CPU consumption!
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001796
1797 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01001798 userlist L1
1799 group G1 users tiger,scott
1800 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001801
Cyril Bontéf0c60612010-02-06 14:44:47 +01001802 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
1803 user scott insecure-password elgato
1804 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001805
Cyril Bontéf0c60612010-02-06 14:44:47 +01001806 userlist L2
1807 group G1
1808 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001809
Cyril Bontéf0c60612010-02-06 14:44:47 +01001810 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
1811 user scott insecure-password elgato groups G1,G2
1812 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001813
1814 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001815
Emeric Brunf099e792010-09-27 12:05:28 +02001816
18173.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02001818----------
Emeric Brun94900952015-06-11 18:25:54 +02001819It is possible to propagate entries of any data-types in stick-tables between
1820several haproxy instances over TCP connections in a multi-master fashion. Each
1821instance pushes its local updates and insertions to remote peers. The pushed
1822values overwrite remote ones without aggregation. Interrupted exchanges are
1823automatically detected and recovered from the last known point.
1824In addition, during a soft restart, the old process connects to the new one
1825using such a TCP connection to push all its entries before the new process
1826tries to connect to other peers. That ensures very fast replication during a
1827reload, it typically takes a fraction of a second even for large tables.
1828Note that Server IDs are used to identify servers remotely, so it is important
1829that configurations look similar or at least that the same IDs are forced on
1830each server on all participants.
Emeric Brunf099e792010-09-27 12:05:28 +02001831
1832peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001833 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02001834 which is referenced by one or more stick-tables.
1835
Willy Tarreau77e4bd12015-05-01 20:02:17 +02001836disabled
1837 Disables a peers section. It disables both listening and any synchronization
1838 related to this section. This is provided to disable synchronization of stick
1839 tables without having to comment out all "peers" references.
1840
1841enable
1842 This re-enables a disabled peers section which was previously disabled.
1843
Emeric Brunf099e792010-09-27 12:05:28 +02001844peer <peername> <ip>:<port>
1845 Defines a peer inside a peers section.
1846 If <peername> is set to the local peer name (by default hostname, or forced
1847 using "-L" command line option), haproxy will listen for incoming remote peer
1848 connection on <ip>:<port>. Otherwise, <ip>:<port> defines where to connect to
1849 to join the remote peer, and <peername> is used at the protocol level to
1850 identify and validate the remote peer on the server side.
1851
1852 During a soft restart, local peer <ip>:<port> is used by the old instance to
1853 connect the new one and initiate a complete replication (teaching process).
1854
1855 It is strongly recommended to have the exact same peers declaration on all
1856 peers and to only rely on the "-L" command line argument to change the local
1857 peer name. This makes it easier to maintain coherent configuration files
1858 across all peers.
1859
William Lallemandb2f07452015-05-12 14:27:13 +02001860 You may want to reference some environment variables in the address
1861 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01001862
Cyril Bontédc4d9032012-04-08 21:57:39 +02001863 Example:
Emeric Brunf099e792010-09-27 12:05:28 +02001864 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001865 peer haproxy1 192.168.0.1:1024
1866 peer haproxy2 192.168.0.2:1024
1867 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02001868
1869 backend mybackend
1870 mode tcp
1871 balance roundrobin
1872 stick-table type ip size 20k peers mypeers
1873 stick on src
1874
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001875 server srv1 192.168.0.30:80
1876 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02001877
1878
Simon Horman51a1cf62015-02-03 13:00:44 +090018793.6. Mailers
1880------------
1881It is possible to send email alerts when the state of servers changes.
1882If configured email alerts are sent to each mailer that is configured
1883in a mailers section. Email is sent to mailers using SMTP.
1884
Pieter Baauw386a1272015-08-16 15:26:24 +02001885mailers <mailersect>
Simon Horman51a1cf62015-02-03 13:00:44 +09001886 Creates a new mailer list with the name <mailersect>. It is an
1887 independent section which is referenced by one or more proxies.
1888
1889mailer <mailername> <ip>:<port>
1890 Defines a mailer inside a mailers section.
1891
1892 Example:
1893 mailers mymailers
1894 mailer smtp1 192.168.0.1:587
1895 mailer smtp2 192.168.0.2:587
1896
1897 backend mybackend
1898 mode tcp
1899 balance roundrobin
1900
1901 email-alert mailers mymailers
1902 email-alert from test1@horms.org
1903 email-alert to test2@horms.org
1904
1905 server srv1 192.168.0.30:80
1906 server srv2 192.168.0.31:80
1907
Pieter Baauw235fcfc2016-02-13 15:33:40 +01001908timeout mail <time>
1909 Defines the time available for a mail/connection to be made and send to
1910 the mail-server. If not defined the default value is 10 seconds. To allow
1911 for at least two SYN-ACK packets to be send during initial TCP handshake it
1912 is advised to keep this value above 4 seconds.
1913
1914 Example:
1915 mailers mymailers
1916 timeout mail 20s
1917 mailer smtp1 192.168.0.1:587
Simon Horman51a1cf62015-02-03 13:00:44 +09001918
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019194. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02001920----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001921
Willy Tarreau6a06a402007-07-15 20:15:28 +02001922Proxy configuration can be located in a set of sections :
William Lallemand6e62fb62015-04-28 16:55:23 +02001923 - defaults [<name>]
Willy Tarreau6a06a402007-07-15 20:15:28 +02001924 - frontend <name>
1925 - backend <name>
1926 - listen <name>
1927
1928A "defaults" section sets default parameters for all other sections following
1929its declaration. Those default parameters are reset by the next "defaults"
1930section. See below for the list of parameters which can be set in a "defaults"
Willy Tarreau0ba27502007-12-24 16:55:16 +01001931section. The name is optional but its use is encouraged for better readability.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001932
1933A "frontend" section describes a set of listening sockets accepting client
1934connections.
1935
1936A "backend" section describes a set of servers to which the proxy will connect
1937to forward incoming connections.
1938
1939A "listen" section defines a complete proxy with its frontend and backend
1940parts combined in one section. It is generally useful for TCP-only traffic.
1941
Willy Tarreau0ba27502007-12-24 16:55:16 +01001942All proxy names must be formed from upper and lower case letters, digits,
1943'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
1944case-sensitive, which means that "www" and "WWW" are two different proxies.
1945
1946Historically, all proxy names could overlap, it just caused troubles in the
1947logs. Since the introduction of content switching, it is mandatory that two
1948proxies with overlapping capabilities (frontend/backend) have different names.
1949However, it is still permitted that a frontend and a backend share the same
1950name, as this configuration seems to be commonly encountered.
1951
1952Right now, two major proxy modes are supported : "tcp", also known as layer 4,
1953and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001954bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01001955protocol, and can interact with it by allowing, blocking, switching, adding,
1956modifying, or removing arbitrary contents in requests or responses, based on
1957arbitrary criteria.
1958
Willy Tarreau70dffda2014-01-30 03:07:23 +01001959In HTTP mode, the processing applied to requests and responses flowing over
1960a connection depends in the combination of the frontend's HTTP options and
1961the backend's. HAProxy supports 5 connection modes :
1962
1963 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
1964 requests and responses are processed, and connections remain open but idle
1965 between responses and new requests.
1966
1967 - TUN: tunnel ("option http-tunnel") : this was the default mode for versions
1968 1.0 to 1.5-dev21 : only the first request and response are processed, and
1969 everything else is forwarded with no analysis at all. This mode should not
1970 be used as it creates lots of trouble with logging and HTTP processing.
1971
1972 - PCL: passive close ("option httpclose") : exactly the same as tunnel mode,
1973 but with "Connection: close" appended in both directions to try to make
1974 both ends close after the first request/response exchange.
1975
1976 - SCL: server close ("option http-server-close") : the server-facing
1977 connection is closed after the end of the response is received, but the
1978 client-facing connection remains open.
1979
1980 - FCL: forced close ("option forceclose") : the connection is actively closed
1981 after the end of the response.
1982
1983The effective mode that will be applied to a connection passing through a
1984frontend and a backend can be determined by both proxy modes according to the
1985following matrix, but in short, the modes are symmetric, keep-alive is the
1986weakest option and force close is the strongest.
1987
1988 Backend mode
1989
1990 | KAL | TUN | PCL | SCL | FCL
1991 ----+-----+-----+-----+-----+----
1992 KAL | KAL | TUN | PCL | SCL | FCL
1993 ----+-----+-----+-----+-----+----
1994 TUN | TUN | TUN | PCL | SCL | FCL
1995 Frontend ----+-----+-----+-----+-----+----
1996 mode PCL | PCL | PCL | PCL | FCL | FCL
1997 ----+-----+-----+-----+-----+----
1998 SCL | SCL | SCL | FCL | SCL | FCL
1999 ----+-----+-----+-----+-----+----
2000 FCL | FCL | FCL | FCL | FCL | FCL
2001
Willy Tarreau0ba27502007-12-24 16:55:16 +01002002
Willy Tarreau70dffda2014-01-30 03:07:23 +01002003
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020044.1. Proxy keywords matrix
2005--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01002006
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002007The following list of keywords is supported. Most of them may only be used in a
2008limited set of section types. Some of them are marked as "deprecated" because
2009they are inherited from an old syntax which may be confusing or functionally
2010limited, and there are new recommended keywords to replace them. Keywords
Davor Ocelice9ed2812017-12-25 17:49:28 +01002011marked with "(*)" can be optionally inverted using the "no" prefix, e.g. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002012option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02002013and must be disabled for a specific instance. Such options may also be prefixed
2014with "default" in order to restore default settings regardless of what has been
2015specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002016
Willy Tarreau6a06a402007-07-15 20:15:28 +02002017
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002018 keyword defaults frontend listen backend
2019------------------------------------+----------+----------+---------+---------
2020acl - X X X
Willy Tarreau294d0f02015-08-10 19:40:12 +02002021appsession - - - -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002022backlog X X X -
2023balance X - X X
2024bind - X X -
2025bind-process X X X X
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02002026block (deprecated) - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002027capture cookie - X X -
2028capture request header - X X -
2029capture response header - X X -
2030clitimeout (deprecated) X X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02002031compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002032contimeout (deprecated) X - X X
2033cookie X - X X
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02002034declare capture - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002035default-server X - X X
2036default_backend X X X -
2037description - X X X
2038disabled X X X X
2039dispatch - - X X
Simon Horman51a1cf62015-02-03 13:00:44 +09002040email-alert from X X X X
Simon Horman64e34162015-02-06 11:11:57 +09002041email-alert level X X X X
Simon Horman51a1cf62015-02-03 13:00:44 +09002042email-alert mailers X X X X
2043email-alert myhostname X X X X
2044email-alert to X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002045enabled X X X X
2046errorfile X X X X
2047errorloc X X X X
2048errorloc302 X X X X
2049-- keyword -------------------------- defaults - frontend - listen -- backend -
2050errorloc303 X X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01002051force-persist - - X X
Christopher Fauletc3fe5332016-04-07 15:30:10 +02002052filter - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002053fullconn X - X X
2054grace X X X X
2055hash-type X - X X
2056http-check disable-on-404 X - X X
Willy Tarreaubd741542010-03-16 18:46:54 +01002057http-check expect - - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02002058http-check send-state X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002059http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02002060http-response - X X X
Willy Tarreau30631952015-08-06 15:05:24 +02002061http-reuse X - X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02002062http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002063id - X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01002064ignore-persist - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02002065load-server-state-from-file X - X X
William Lallemand0f99e342011-10-12 17:50:54 +02002066log (*) X X X X
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01002067log-format X X X -
Dragan Dosen7ad31542015-09-28 17:16:47 +02002068log-format-sd X X X -
Willy Tarreau094af4e2015-01-07 15:03:42 +01002069log-tag X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02002070max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002071maxconn X X X -
2072mode X X X X
2073monitor fail - X X -
2074monitor-net X X X -
2075monitor-uri X X X -
2076option abortonclose (*) X - X X
2077option accept-invalid-http-request (*) X X X -
2078option accept-invalid-http-response (*) X - X X
2079option allbackups (*) X - X X
2080option checkcache (*) X - X X
2081option clitcpka (*) X X X -
2082option contstats (*) X X X -
2083option dontlog-normal (*) X X X -
2084option dontlognull (*) X X X -
2085option forceclose (*) X X X X
2086-- keyword -------------------------- defaults - frontend - listen -- backend -
2087option forwardfor X X X X
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02002088option http-buffer-request (*) X X X X
Willy Tarreau82649f92015-05-01 22:40:51 +02002089option http-ignore-probes (*) X X X -
Willy Tarreau16bfb022010-01-16 19:48:41 +01002090option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02002091option http-no-delay (*) X X X X
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02002092option http-pretend-keepalive (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002093option http-server-close (*) X X X X
Willy Tarreau02bce8b2014-01-30 00:15:28 +01002094option http-tunnel (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002095option http-use-proxy-header (*) X X X -
2096option httpchk X - X X
2097option httpclose (*) X X X X
2098option httplog X X X X
2099option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002100option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02002101option ldap-check X - X X
Simon Horman98637e52014-06-20 12:30:16 +09002102option external-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002103option log-health-checks (*) X - X X
2104option log-separate-errors (*) X X X -
2105option logasap (*) X X X -
2106option mysql-check X - X X
2107option nolinger (*) X X X X
2108option originalto X X X X
2109option persist (*) X - X X
Baptiste Assmann809e22a2015-10-12 20:22:55 +02002110option pgsql-check X - X X
2111option prefer-last-server (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002112option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02002113option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002114option smtpchk X - X X
2115option socket-stats (*) X X X -
2116option splice-auto (*) X X X X
2117option splice-request (*) X X X X
2118option splice-response (*) X X X X
Christopher Fauletba7bc162016-11-07 21:07:38 +01002119option spop-check - - - X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002120option srvtcpka (*) X - X X
2121option ssl-hello-chk X - X X
2122-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01002123option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002124option tcp-smart-accept (*) X X X -
2125option tcp-smart-connect (*) X - X X
2126option tcpka X X X X
2127option tcplog X X X X
2128option transparent (*) X - X X
Simon Horman98637e52014-06-20 12:30:16 +09002129external-check command X - X X
2130external-check path X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002131persist rdp-cookie X - X X
2132rate-limit sessions X X X -
2133redirect - X X X
2134redisp (deprecated) X - X X
2135redispatch (deprecated) X - X X
2136reqadd - X X X
2137reqallow - X X X
2138reqdel - X X X
2139reqdeny - X X X
2140reqiallow - X X X
2141reqidel - X X X
2142reqideny - X X X
2143reqipass - X X X
2144reqirep - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002145reqitarpit - X X X
2146reqpass - X X X
2147reqrep - X X X
2148-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002149reqtarpit - X X X
2150retries X - X X
2151rspadd - X X X
2152rspdel - X X X
2153rspdeny - X X X
2154rspidel - X X X
2155rspideny - X X X
2156rspirep - X X X
2157rsprep - X X X
2158server - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02002159server-state-file-name X - X X
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02002160server-template - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002161source X - X X
2162srvtimeout (deprecated) X - X X
Baptiste Assmann5a549212015-10-12 20:30:24 +02002163stats admin - X X X
2164stats auth X X X X
2165stats enable X X X X
2166stats hide-version X X X X
2167stats http-request - X X X
2168stats realm X X X X
2169stats refresh X X X X
2170stats scope X X X X
2171stats show-desc X X X X
2172stats show-legends X X X X
2173stats show-node X X X X
2174stats uri X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002175-- keyword -------------------------- defaults - frontend - listen -- backend -
2176stick match - - X X
2177stick on - - X X
2178stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02002179stick store-response - - X X
Adam Spiers68af3c12017-04-06 16:31:39 +01002180stick-table - X X X
Willy Tarreau938c7fe2014-04-25 14:21:39 +02002181tcp-check connect - - X X
2182tcp-check expect - - X X
2183tcp-check send - - X X
2184tcp-check send-binary - - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02002185tcp-request connection - X X -
2186tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02002187tcp-request inspect-delay - X X X
Willy Tarreau4f614292016-10-21 17:49:36 +02002188tcp-request session - X X -
Emeric Brun0a3b67f2010-09-24 15:34:53 +02002189tcp-response content - - X X
2190tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002191timeout check X - X X
2192timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02002193timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002194timeout clitimeout (deprecated) X X X -
2195timeout connect X - X X
2196timeout contimeout (deprecated) X - X X
2197timeout http-keep-alive X X X X
2198timeout http-request X X X X
2199timeout queue X - X X
2200timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02002201timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002202timeout srvtimeout (deprecated) X - X X
2203timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02002204timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002205transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01002206unique-id-format X X X -
2207unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002208use_backend - X X -
Willy Tarreau4a5cade2012-04-05 21:09:48 +02002209use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002210------------------------------------+----------+----------+---------+---------
2211 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02002212
Willy Tarreau0ba27502007-12-24 16:55:16 +01002213
Willy Tarreauc57f0e22009-05-10 13:12:33 +020022144.2. Alphabetically sorted keywords reference
2215---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01002216
2217This section provides a description of each keyword and its usage.
2218
2219
2220acl <aclname> <criterion> [flags] [operator] <value> ...
2221 Declare or complete an access list.
2222 May be used in sections : defaults | frontend | listen | backend
2223 no | yes | yes | yes
2224 Example:
2225 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
2226 acl invalid_src src_port 0:1023
2227 acl local_dst hdr(host) -i localhost
2228
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002229 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002230
2231
Cyril Bontéb21570a2009-11-29 20:04:48 +01002232appsession <cookie> len <length> timeout <holdtime>
2233 [request-learn] [prefix] [mode <path-parameters|query-string>]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002234 Define session stickiness on an existing application cookie.
2235 May be used in sections : defaults | frontend | listen | backend
2236 no | no | yes | yes
2237 Arguments :
2238 <cookie> this is the name of the cookie used by the application and which
2239 HAProxy will have to learn for each new session.
2240
Cyril Bontéb21570a2009-11-29 20:04:48 +01002241 <length> this is the max number of characters that will be memorized and
Willy Tarreau0ba27502007-12-24 16:55:16 +01002242 checked in each cookie value.
2243
2244 <holdtime> this is the time after which the cookie will be removed from
2245 memory if unused. If no unit is specified, this time is in
2246 milliseconds.
2247
Cyril Bontébf47aeb2009-10-15 00:15:40 +02002248 request-learn
2249 If this option is specified, then haproxy will be able to learn
2250 the cookie found in the request in case the server does not
2251 specify any in response. This is typically what happens with
2252 PHPSESSID cookies, or when haproxy's session expires before
2253 the application's session and the correct server is selected.
2254 It is recommended to specify this option to improve reliability.
2255
Cyril Bontéb21570a2009-11-29 20:04:48 +01002256 prefix When this option is specified, haproxy will match on the cookie
2257 prefix (or URL parameter prefix). The appsession value is the
2258 data following this prefix.
2259
2260 Example :
2261 appsession ASPSESSIONID len 64 timeout 3h prefix
2262
Davor Ocelice9ed2812017-12-25 17:49:28 +01002263 This will match the cookie ASPSESSIONIDXXX=XXXX,
2264 the appsession value will be XXX=XXXX.
Cyril Bontéb21570a2009-11-29 20:04:48 +01002265
2266 mode This option allows to change the URL parser mode.
2267 2 modes are currently supported :
2268 - path-parameters :
2269 The parser looks for the appsession in the path parameters
2270 part (each parameter is separated by a semi-colon), which is
2271 convenient for JSESSIONID for example.
2272 This is the default mode if the option is not set.
2273 - query-string :
2274 In this mode, the parser will look for the appsession in the
2275 query string.
2276
Willy Tarreau294d0f02015-08-10 19:40:12 +02002277 As of version 1.6, appsessions was removed. It is more flexible and more
2278 convenient to use stick-tables instead, and stick-tables support multi-master
2279 replication and data conservation across reloads, which appsessions did not.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002280
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01002281 See also : "cookie", "capture cookie", "balance", "stick", "stick-table",
2282 "ignore-persist", "nbproc" and "bind-process".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002283
2284
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002285backlog <conns>
2286 Give hints to the system about the approximate listen backlog desired size
2287 May be used in sections : defaults | frontend | listen | backend
2288 yes | yes | yes | no
2289 Arguments :
2290 <conns> is the number of pending connections. Depending on the operating
2291 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02002292 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002293
2294 In order to protect against SYN flood attacks, one solution is to increase
2295 the system's SYN backlog size. Depending on the system, sometimes it is just
2296 tunable via a system parameter, sometimes it is not adjustable at all, and
2297 sometimes the system relies on hints given by the application at the time of
2298 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
2299 to the listen() syscall. On systems which can make use of this value, it can
2300 sometimes be useful to be able to specify a different value, hence this
2301 backlog parameter.
2302
2303 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
2304 used as a hint and the system accepts up to the smallest greater power of
2305 two, and never more than some limits (usually 32768).
2306
2307 See also : "maxconn" and the target operating system's tuning guide.
2308
2309
Willy Tarreau0ba27502007-12-24 16:55:16 +01002310balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02002311balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002312 Define the load balancing algorithm to be used in a backend.
2313 May be used in sections : defaults | frontend | listen | backend
2314 yes | no | yes | yes
2315 Arguments :
2316 <algorithm> is the algorithm used to select a server when doing load
2317 balancing. This only applies when no persistence information
2318 is available, or when a connection is redispatched to another
2319 server. <algorithm> may be one of the following :
2320
2321 roundrobin Each server is used in turns, according to their weights.
2322 This is the smoothest and fairest algorithm when the server's
2323 processing time remains equally distributed. This algorithm
2324 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02002325 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08002326 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02002327 large farms, when a server becomes up after having been down
2328 for a very short time, it may sometimes take a few hundreds
2329 requests for it to be re-integrated into the farm and start
2330 receiving traffic. This is normal, though very rare. It is
2331 indicated here in case you would have the chance to observe
2332 it, so that you don't worry.
2333
2334 static-rr Each server is used in turns, according to their weights.
2335 This algorithm is as similar to roundrobin except that it is
2336 static, which means that changing a server's weight on the
2337 fly will have no effect. On the other hand, it has no design
2338 limitation on the number of servers, and when a server goes
2339 up, it is always immediately reintroduced into the farm, once
2340 the full map is recomputed. It also uses slightly less CPU to
2341 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01002342
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01002343 leastconn The server with the lowest number of connections receives the
2344 connection. Round-robin is performed within groups of servers
2345 of the same load to ensure that all servers will be used. Use
2346 of this algorithm is recommended where very long sessions are
2347 expected, such as LDAP, SQL, TSE, etc... but is not very well
2348 suited for protocols using short sessions such as HTTP. This
2349 algorithm is dynamic, which means that server weights may be
2350 adjusted on the fly for slow starts for instance.
2351
Willy Tarreauf09c6602012-02-13 17:12:08 +01002352 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002353 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01002354 identifier to the highest (see server parameter "id"), which
2355 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02002356 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01002357 not make sense to use this algorithm without setting maxconn.
2358 The purpose of this algorithm is to always use the smallest
2359 number of servers so that extra servers can be powered off
2360 during non-intensive hours. This algorithm ignores the server
2361 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02002362 or IMAP than HTTP, though it can be useful there too. In
2363 order to use this algorithm efficiently, it is recommended
2364 that a cloud controller regularly checks server usage to turn
2365 them off when unused, and regularly checks backend queue to
2366 turn new servers on when the queue inflates. Alternatively,
2367 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01002368
Willy Tarreau0ba27502007-12-24 16:55:16 +01002369 source The source IP address is hashed and divided by the total
2370 weight of the running servers to designate which server will
2371 receive the request. This ensures that the same client IP
2372 address will always reach the same server as long as no
2373 server goes down or up. If the hash result changes due to the
2374 number of running servers changing, many clients will be
2375 directed to a different server. This algorithm is generally
2376 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002377 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01002378 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002379 static by default, which means that changing a server's
2380 weight on the fly will have no effect, but this can be
2381 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002382
Oskar Stolc8dc41842012-05-19 10:19:54 +01002383 uri This algorithm hashes either the left part of the URI (before
2384 the question mark) or the whole URI (if the "whole" parameter
2385 is present) and divides the hash value by the total weight of
2386 the running servers. The result designates which server will
2387 receive the request. This ensures that the same URI will
2388 always be directed to the same server as long as no server
2389 goes up or down. This is used with proxy caches and
2390 anti-virus proxies in order to maximize the cache hit rate.
2391 Note that this algorithm may only be used in an HTTP backend.
2392 This algorithm is static by default, which means that
2393 changing a server's weight on the fly will have no effect,
2394 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002395
Oskar Stolc8dc41842012-05-19 10:19:54 +01002396 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002397 "depth", both followed by a positive integer number. These
2398 options may be helpful when it is needed to balance servers
2399 based on the beginning of the URI only. The "len" parameter
2400 indicates that the algorithm should only consider that many
2401 characters at the beginning of the URI to compute the hash.
2402 Note that having "len" set to 1 rarely makes sense since most
2403 URIs start with a leading "/".
2404
2405 The "depth" parameter indicates the maximum directory depth
2406 to be used to compute the hash. One level is counted for each
2407 slash in the request. If both parameters are specified, the
2408 evaluation stops when either is reached.
2409
Willy Tarreau0ba27502007-12-24 16:55:16 +01002410 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002411 the query string of each HTTP GET request.
2412
2413 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02002414 request entity will be searched for the parameter argument,
2415 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02002416 ('?') in the URL. The message body will only start to be
2417 analyzed once either the advertised amount of data has been
2418 received or the request buffer is full. In the unlikely event
2419 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02002420 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02002421 be randomly balanced if at all. This keyword used to support
2422 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002423
2424 If the parameter is found followed by an equal sign ('=') and
2425 a value, then the value is hashed and divided by the total
2426 weight of the running servers. The result designates which
2427 server will receive the request.
2428
2429 This is used to track user identifiers in requests and ensure
2430 that a same user ID will always be sent to the same server as
2431 long as no server goes up or down. If no value is found or if
2432 the parameter is not found, then a round robin algorithm is
2433 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002434 backend. This algorithm is static by default, which means
2435 that changing a server's weight on the fly will have no
2436 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002437
Cyril Bontédc4d9032012-04-08 21:57:39 +02002438 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
2439 request. Just as with the equivalent ACL 'hdr()' function,
2440 the header name in parenthesis is not case sensitive. If the
2441 header is absent or if it does not contain any value, the
2442 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01002443
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002444 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01002445 reducing the hash algorithm to the main domain part with some
2446 specific headers such as 'Host'. For instance, in the Host
2447 value "haproxy.1wt.eu", only "1wt" will be considered.
2448
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002449 This algorithm is static by default, which means that
2450 changing a server's weight on the fly will have no effect,
2451 but this can be changed using "hash-type".
2452
Willy Tarreau760e81d2018-05-03 07:20:40 +02002453 random A random number will be used as the key for the consistent
2454 hashing function. This means that the servers' weights are
2455 respected, dynamic weight changes immediately take effect, as
2456 well as new server additions. Random load balancing can be
2457 useful with large farms or when servers are frequently added
2458 or removed. The hash-balance-factor directive can be used to
2459 further improve fairness of the load balancing, especially
2460 in situations where servers show highly variable response
2461 times.
2462
Emeric Brun736aa232009-06-30 17:56:00 +02002463 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02002464 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02002465 The RDP cookie <name> (or "mstshash" if omitted) will be
2466 looked up and hashed for each incoming TCP request. Just as
2467 with the equivalent ACL 'req_rdp_cookie()' function, the name
2468 is not case-sensitive. This mechanism is useful as a degraded
2469 persistence mode, as it makes it possible to always send the
2470 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002471 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02002472 used instead.
2473
2474 Note that for this to work, the frontend must ensure that an
2475 RDP cookie is already present in the request buffer. For this
2476 you must use 'tcp-request content accept' rule combined with
2477 a 'req_rdp_cookie_cnt' ACL.
2478
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002479 This algorithm is static by default, which means that
2480 changing a server's weight on the fly will have no effect,
2481 but this can be changed using "hash-type".
2482
Cyril Bontédc4d9032012-04-08 21:57:39 +02002483 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09002484
Willy Tarreau0ba27502007-12-24 16:55:16 +01002485 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002486 algorithms. Right now, only "url_param" and "uri" support an
2487 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002488
Willy Tarreau3cd9af22009-03-15 14:06:41 +01002489 The load balancing algorithm of a backend is set to roundrobin when no other
2490 algorithm, mode nor option have been set. The algorithm may only be set once
2491 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002492
2493 Examples :
2494 balance roundrobin
2495 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002496 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01002497 balance hdr(User-Agent)
2498 balance hdr(host)
2499 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002500
2501 Note: the following caveats and limitations on using the "check_post"
2502 extension with "url_param" must be considered :
2503
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002504 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002505 to determine if the parameters will be found in the body or entity which
2506 may contain binary data. Therefore another method may be required to
2507 restrict consideration of POST requests that have no URL parameters in
2508 the body. (see acl reqideny http_end)
2509
2510 - using a <max_wait> value larger than the request buffer size does not
2511 make sense and is useless. The buffer size is set at build time, and
2512 defaults to 16 kB.
2513
2514 - Content-Encoding is not supported, the parameter search will probably
2515 fail; and load balancing will fall back to Round Robin.
2516
2517 - Expect: 100-continue is not supported, load balancing will fall back to
2518 Round Robin.
2519
Lukas Tribus23953682017-04-28 13:24:30 +00002520 - Transfer-Encoding (RFC7230 3.3.1) is only supported in the first chunk.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002521 If the entire parameter value is not present in the first chunk, the
2522 selection of server is undefined (actually, defined by how little
2523 actually appeared in the first chunk).
2524
2525 - This feature does not support generation of a 100, 411 or 501 response.
2526
2527 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002528 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002529 white space or control characters are found, indicating the end of what
2530 might be a URL parameter list. This is probably not a concern with SGML
2531 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002532
Willy Tarreau294d0f02015-08-10 19:40:12 +02002533 See also : "dispatch", "cookie", "transparent", "hash-type" and "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002534
2535
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002536bind [<address>]:<port_range> [, ...] [param*]
2537bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002538 Define one or several listening addresses and/or ports in a frontend.
2539 May be used in sections : defaults | frontend | listen | backend
2540 no | yes | yes | no
2541 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002542 <address> is optional and can be a host name, an IPv4 address, an IPv6
2543 address, or '*'. It designates the address the frontend will
2544 listen on. If unset, all IPv4 addresses of the system will be
2545 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01002546 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01002547 Optionally, an address family prefix may be used before the
2548 address to force the family regardless of the address format,
2549 which can be useful to specify a path to a unix socket with
2550 no slash ('/'). Currently supported prefixes are :
2551 - 'ipv4@' -> address is always IPv4
2552 - 'ipv6@' -> address is always IPv6
2553 - 'unix@' -> address is a path to a local unix socket
Willy Tarreau70f72e02014-07-08 00:37:50 +02002554 - 'abns@' -> address is in abstract namespace (Linux only).
2555 Note: since abstract sockets are not "rebindable", they
2556 do not cope well with multi-process mode during
2557 soft-restart, so it is better to avoid them if
2558 nbproc is greater than 1. The effect is that if the
2559 new process fails to start, only one of the old ones
2560 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01002561 - 'fd@<n>' -> use file descriptor <n> inherited from the
2562 parent. The fd must be bound and may or may not already
2563 be listening.
William Lallemand2fe7dd02018-09-11 16:51:29 +02002564 - 'sockpair@<n>'-> like fd@ but you must use the fd of a
2565 connected unix socket or of a socketpair. The bind waits
2566 to receive a FD over the unix socket and uses it as if it
2567 was the FD of an accept(). Should be used carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02002568 You may want to reference some environment variables in the
2569 address parameter, see section 2.3 about environment
2570 variables.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002571
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002572 <port_range> is either a unique TCP port, or a port range for which the
2573 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002574 above. The port is mandatory for TCP listeners. Note that in
2575 the case of an IPv6 address, the port is always the number
2576 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002577 - a numerical port (ex: '80')
2578 - a dash-delimited ports range explicitly stating the lower
2579 and upper bounds (ex: '2000-2100') which are included in
2580 the range.
2581
2582 Particular care must be taken against port ranges, because
2583 every <address:port> couple consumes one socket (= a file
2584 descriptor), so it's easy to consume lots of descriptors
2585 with a simple range, and to run out of sockets. Also, each
2586 <address:port> couple must be used only once among all
2587 instances running on a same system. Please note that binding
2588 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002589 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002590 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002591
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002592 <path> is a UNIX socket path beginning with a slash ('/'). This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002593 alternative to the TCP listening port. HAProxy will then
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002594 receive UNIX connections on the socket located at this place.
2595 The path must begin with a slash and by default is absolute.
2596 It can be relative to the prefix defined by "unix-bind" in
2597 the global section. Note that the total length of the prefix
2598 followed by the socket path cannot exceed some system limits
2599 for UNIX sockets, which commonly are set to 107 characters.
2600
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002601 <param*> is a list of parameters common to all sockets declared on the
2602 same line. These numerous parameters depend on OS and build
2603 options and have a complete section dedicated to them. Please
2604 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002605
Willy Tarreau0ba27502007-12-24 16:55:16 +01002606 It is possible to specify a list of address:port combinations delimited by
2607 commas. The frontend will then listen on all of these addresses. There is no
2608 fixed limit to the number of addresses and ports which can be listened on in
2609 a frontend, as well as there is no limit to the number of "bind" statements
2610 in a frontend.
2611
2612 Example :
2613 listen http_proxy
2614 bind :80,:443
2615 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002616 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01002617
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002618 listen http_https_proxy
2619 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02002620 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002621
Willy Tarreau24709282013-03-10 21:32:12 +01002622 listen http_https_proxy_explicit
2623 bind ipv6@:80
2624 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
2625 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
2626
Willy Tarreaudad36a32013-03-11 01:20:04 +01002627 listen external_bind_app1
William Lallemandb2f07452015-05-12 14:27:13 +02002628 bind "fd@${FD_APP1}"
Willy Tarreaudad36a32013-03-11 01:20:04 +01002629
Willy Tarreau55dcaf62015-09-27 15:03:15 +02002630 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
2631 sun_path length is used for the address length. Some other programs
2632 such as socat use the string length only by default. Pass the option
2633 ",unix-tightsocklen=0" to any abstract socket definition in socat to
2634 make it compatible with HAProxy's.
2635
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002636 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002637 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002638
2639
Christopher Fauletff4121f2017-11-22 16:38:49 +01002640bind-process [ all | odd | even | <process_num>[-[<process_num>]] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002641 Limit visibility of an instance to a certain set of processes numbers.
2642 May be used in sections : defaults | frontend | listen | backend
2643 yes | yes | yes | yes
2644 Arguments :
2645 all All process will see this instance. This is the default. It
2646 may be used to override a default value.
2647
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002648 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002649 option may be combined with other numbers.
2650
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002651 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002652 option may be combined with other numbers. Do not use it
2653 with less than 2 processes otherwise some instances might be
2654 missing from all processes.
2655
Christopher Fauletff4121f2017-11-22 16:38:49 +01002656 process_num The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002657 whose values must all be between 1 and 32 or 64 depending on
Christopher Fauletff4121f2017-11-22 16:38:49 +01002658 the machine's word size. Ranges can be partially defined. The
2659 higher bound can be omitted. In such case, it is replaced by
2660 the corresponding maximum value. If a proxy is bound to
2661 process numbers greater than the configured global.nbproc, it
2662 will either be forced to process #1 if a single process was
Willy Tarreau102df612014-05-07 23:56:38 +02002663 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002664
2665 This keyword limits binding of certain instances to certain processes. This
2666 is useful in order not to have too many processes listening to the same
2667 ports. For instance, on a dual-core machine, it might make sense to set
2668 'nbproc 2' in the global section, then distributes the listeners among 'odd'
2669 and 'even' instances.
2670
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002671 At the moment, it is not possible to reference more than 32 or 64 processes
2672 using this keyword, but this should be more than enough for most setups.
2673 Please note that 'all' really means all processes regardless of the machine's
2674 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002675
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002676 Each "bind" line may further be limited to a subset of the proxy's processes,
2677 please consult the "process" bind keyword in section 5.1.
2678
Willy Tarreaub369a042014-09-16 13:21:03 +02002679 When a frontend has no explicit "bind-process" line, it tries to bind to all
2680 the processes referenced by its "bind" lines. That means that frontends can
2681 easily adapt to their listeners' processes.
2682
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002683 If some backends are referenced by frontends bound to other processes, the
2684 backend automatically inherits the frontend's processes.
2685
2686 Example :
2687 listen app_ip1
2688 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002689 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002690
2691 listen app_ip2
2692 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002693 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002694
2695 listen management
2696 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002697 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002698
Willy Tarreau110ecc12012-11-15 17:50:01 +01002699 listen management
2700 bind 10.0.0.4:80
2701 bind-process 1-4
2702
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002703 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002704
2705
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02002706block { if | unless } <condition> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002707 Block a layer 7 request if/unless a condition is matched
2708 May be used in sections : defaults | frontend | listen | backend
2709 no | yes | yes | yes
2710
2711 The HTTP request will be blocked very early in the layer 7 processing
2712 if/unless <condition> is matched. A 403 error will be returned if the request
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002713 is blocked. The condition has to reference ACLs (see section 7). This is
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02002714 typically used to deny access to certain sensitive resources if some
Willy Tarreau0ba27502007-12-24 16:55:16 +01002715 conditions are met or not met. There is no fixed limit to the number of
Jarno Huuskonen95b012b2017-04-06 13:59:14 +03002716 "block" statements per instance. To block connections at layer 4 (without
2717 sending a 403 error) see "tcp-request connection reject" and
2718 "tcp-request content reject" rules.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002719
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02002720 This form is deprecated, do not use it in any new configuration, use the new
2721 "http-request deny" instead.
2722
Willy Tarreau0ba27502007-12-24 16:55:16 +01002723 Example:
2724 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
2725 acl invalid_src src_port 0:1023
2726 acl local_dst hdr(host) -i localhost
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +03002727 # block is deprecated. Use http-request deny instead:
2728 #block if invalid_src || local_dst
2729 http-request deny if invalid_src || local_dst
Willy Tarreau0ba27502007-12-24 16:55:16 +01002730
Jarno Huuskonen95b012b2017-04-06 13:59:14 +03002731 See also : section 7 about ACL usage, "http-request deny",
2732 "http-response deny", "tcp-request connection reject" and
2733 "tcp-request content reject".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002734
2735capture cookie <name> len <length>
2736 Capture and log a cookie in the request and in the response.
2737 May be used in sections : defaults | frontend | listen | backend
2738 no | yes | yes | no
2739 Arguments :
2740 <name> is the beginning of the name of the cookie to capture. In order
2741 to match the exact name, simply suffix the name with an equal
2742 sign ('='). The full name will appear in the logs, which is
2743 useful with application servers which adjust both the cookie name
Davor Ocelice9ed2812017-12-25 17:49:28 +01002744 and value (e.g. ASPSESSIONXXX).
Willy Tarreau0ba27502007-12-24 16:55:16 +01002745
2746 <length> is the maximum number of characters to report in the logs, which
2747 include the cookie name, the equal sign and the value, all in the
2748 standard "name=value" form. The string will be truncated on the
2749 right if it exceeds <length>.
2750
2751 Only the first cookie is captured. Both the "cookie" request headers and the
2752 "set-cookie" response headers are monitored. This is particularly useful to
2753 check for application bugs causing session crossing or stealing between
2754 users, because generally the user's cookies can only change on a login page.
2755
2756 When the cookie was not presented by the client, the associated log column
2757 will report "-". When a request does not cause a cookie to be assigned by the
2758 server, a "-" is reported in the response column.
2759
2760 The capture is performed in the frontend only because it is necessary that
2761 the log format does not change for a given frontend depending on the
2762 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01002763 "capture cookie" statement in a frontend. The maximum capture length is set
2764 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
2765 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002766
2767 Example:
2768 capture cookie ASPSESSION len 32
2769
2770 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002771 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002772
2773
2774capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01002775 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002776 May be used in sections : defaults | frontend | listen | backend
2777 no | yes | yes | no
2778 Arguments :
2779 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002780 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01002781 appear in the requests, with the first letter of each word in
2782 upper case. The header name will not appear in the logs, only the
2783 value is reported, but the position in the logs is respected.
2784
2785 <length> is the maximum number of characters to extract from the value and
2786 report in the logs. The string will be truncated on the right if
2787 it exceeds <length>.
2788
Willy Tarreau4460d032012-11-21 23:37:37 +01002789 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002790 value will be added to the logs between braces ('{}'). If multiple headers
2791 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002792 in the same order they were declared in the configuration. Non-existent
2793 headers will be logged just as an empty string. Common uses for request
2794 header captures include the "Host" field in virtual hosting environments, the
2795 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002796 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002797 environments to find where the request came from.
2798
2799 Note that when capturing headers such as "User-agent", some spaces may be
2800 logged, making the log analysis more difficult. Thus be careful about what
2801 you log if you know your log parser is not smart enough to rely on the
2802 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002803
Willy Tarreau0900abb2012-11-22 00:21:46 +01002804 There is no limit to the number of captured request headers nor to their
2805 length, though it is wise to keep them low to limit memory usage per session.
2806 In order to keep log format consistent for a same frontend, header captures
2807 can only be declared in a frontend. It is not possible to specify a capture
2808 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002809
2810 Example:
2811 capture request header Host len 15
2812 capture request header X-Forwarded-For len 15
Cyril Bontéd1b0f7c2015-10-26 22:37:39 +01002813 capture request header Referer len 15
Willy Tarreau0ba27502007-12-24 16:55:16 +01002814
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002815 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002816 about logging.
2817
2818
2819capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01002820 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002821 May be used in sections : defaults | frontend | listen | backend
2822 no | yes | yes | no
2823 Arguments :
2824 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002825 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01002826 appear in the response, with the first letter of each word in
2827 upper case. The header name will not appear in the logs, only the
2828 value is reported, but the position in the logs is respected.
2829
2830 <length> is the maximum number of characters to extract from the value and
2831 report in the logs. The string will be truncated on the right if
2832 it exceeds <length>.
2833
Willy Tarreau4460d032012-11-21 23:37:37 +01002834 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002835 result will be added to the logs between braces ('{}') after the captured
2836 request headers. If multiple headers are captured, they will be delimited by
2837 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002838 the configuration. Non-existent headers will be logged just as an empty
2839 string. Common uses for response header captures include the "Content-length"
2840 header which indicates how many bytes are expected to be returned, the
2841 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002842
Willy Tarreau0900abb2012-11-22 00:21:46 +01002843 There is no limit to the number of captured response headers nor to their
2844 length, though it is wise to keep them low to limit memory usage per session.
2845 In order to keep log format consistent for a same frontend, header captures
2846 can only be declared in a frontend. It is not possible to specify a capture
2847 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002848
2849 Example:
2850 capture response header Content-length len 9
2851 capture response header Location len 15
2852
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002853 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002854 about logging.
2855
2856
Cyril Bontéf0c60612010-02-06 14:44:47 +01002857clitimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002858 Set the maximum inactivity time on the client side.
2859 May be used in sections : defaults | frontend | listen | backend
2860 yes | yes | yes | no
2861 Arguments :
2862 <timeout> is the timeout value is specified in milliseconds by default, but
2863 can be in any other unit if the number is suffixed by the unit,
2864 as explained at the top of this document.
2865
2866 The inactivity timeout applies when the client is expected to acknowledge or
2867 send data. In HTTP mode, this timeout is particularly important to consider
2868 during the first phase, when the client sends the request, and during the
2869 response while it is reading data sent by the server. The value is specified
2870 in milliseconds by default, but can be in any other unit if the number is
2871 suffixed by the unit, as specified at the top of this document. In TCP mode
2872 (and to a lesser extent, in HTTP mode), it is highly recommended that the
2873 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002874 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01002875 losses by specifying timeouts that are slightly above multiples of 3 seconds
Davor Ocelice9ed2812017-12-25 17:49:28 +01002876 (e.g. 4 or 5 seconds).
Willy Tarreau0ba27502007-12-24 16:55:16 +01002877
2878 This parameter is specific to frontends, but can be specified once for all in
2879 "defaults" sections. This is in fact one of the easiest solutions not to
2880 forget about it. An unspecified timeout results in an infinite timeout, which
2881 is not recommended. Such a usage is accepted and works but reports a warning
2882 during startup because it may results in accumulation of expired sessions in
2883 the system if the system's timeouts are not configured either.
2884
2885 This parameter is provided for compatibility but is currently deprecated.
2886 Please use "timeout client" instead.
2887
Willy Tarreau036fae02008-01-06 13:24:40 +01002888 See also : "timeout client", "timeout http-request", "timeout server", and
2889 "srvtimeout".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002890
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002891compression algo <algorithm> ...
2892compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02002893compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02002894 Enable HTTP compression.
2895 May be used in sections : defaults | frontend | listen | backend
2896 yes | yes | yes | yes
2897 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002898 algo is followed by the list of supported compression algorithms.
2899 type is followed by the list of MIME types that will be compressed.
2900 offload makes haproxy work as a compression offloader only (see notes).
2901
2902 The currently supported algorithms are :
Willy Tarreauc91840a2015-03-28 17:00:39 +01002903 identity this is mostly for debugging, and it was useful for developing
2904 the compression feature. Identity does not apply any change on
2905 data.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002906
Willy Tarreauc91840a2015-03-28 17:00:39 +01002907 gzip applies gzip compression. This setting is only available when
Baptiste Assmannf085d632015-12-21 17:57:32 +01002908 support for zlib or libslz was built in.
Willy Tarreauc91840a2015-03-28 17:00:39 +01002909
2910 deflate same as "gzip", but with deflate algorithm and zlib format.
2911 Note that this algorithm has ambiguous support on many
2912 browsers and no support at all from recent ones. It is
2913 strongly recommended not to use it for anything else than
2914 experimentation. This setting is only available when support
Baptiste Assmannf085d632015-12-21 17:57:32 +01002915 for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002916
Willy Tarreauc91840a2015-03-28 17:00:39 +01002917 raw-deflate same as "deflate" without the zlib wrapper, and used as an
2918 alternative when the browser wants "deflate". All major
2919 browsers understand it and despite violating the standards,
2920 it is known to work better than "deflate", at least on MSIE
2921 and some versions of Safari. Do not use it in conjunction
2922 with "deflate", use either one or the other since both react
2923 to the same Accept-Encoding token. This setting is only
Baptiste Assmannf085d632015-12-21 17:57:32 +01002924 available when support for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002925
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04002926 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002927 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002928 If backend servers support HTTP compression, these directives
2929 will be no-op: haproxy will see the compressed response and will not
2930 compress again. If backend servers do not support HTTP compression and
2931 there is Accept-Encoding header in request, haproxy will compress the
2932 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02002933
2934 The "offload" setting makes haproxy remove the Accept-Encoding header to
2935 prevent backend servers from compressing responses. It is strongly
2936 recommended not to do this because this means that all the compression work
2937 will be done on the single point where haproxy is located. However in some
2938 deployment scenarios, haproxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002939 with broken HTTP compression implementation which can't be turned off.
2940 In that case haproxy can be used to prevent that gateway from emitting
2941 invalid payloads. In this case, simply removing the header in the
2942 configuration does not work because it applies before the header is parsed,
2943 so that prevents haproxy from compressing. The "offload" setting should
Willy Tarreauffea9fd2014-07-12 16:37:02 +02002944 then be used for such scenarios. Note: for now, the "offload" setting is
2945 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02002946
William Lallemand05097442012-11-20 12:14:28 +01002947 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002948 * the request does not advertise a supported compression algorithm in the
2949 "Accept-Encoding" header
2950 * the response message is not HTTP/1.1
William Lallemandd3002612012-11-26 14:34:47 +01002951 * HTTP status code is not 200
William Lallemand8bb4e342013-12-10 17:28:48 +01002952 * response header "Transfer-Encoding" contains "chunked" (Temporary
2953 Workaround)
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002954 * response contain neither a "Content-Length" header nor a
2955 "Transfer-Encoding" whose last value is "chunked"
2956 * response contains a "Content-Type" header whose first value starts with
2957 "multipart"
2958 * the response contains the "no-transform" value in the "Cache-control"
2959 header
2960 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
2961 and later
2962 * The response contains a "Content-Encoding" header, indicating that the
2963 response is already compressed (see compression offload)
William Lallemand05097442012-11-20 12:14:28 +01002964
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002965 Note: The compression does not rewrite Etag headers, and does not emit the
2966 Warning header.
William Lallemand05097442012-11-20 12:14:28 +01002967
William Lallemand82fe75c2012-10-23 10:25:10 +02002968 Examples :
2969 compression algo gzip
2970 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01002971
Christopher Fauletc3fe5332016-04-07 15:30:10 +02002972
Cyril Bontéf0c60612010-02-06 14:44:47 +01002973contimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002974 Set the maximum time to wait for a connection attempt to a server to succeed.
2975 May be used in sections : defaults | frontend | listen | backend
2976 yes | no | yes | yes
2977 Arguments :
2978 <timeout> is the timeout value is specified in milliseconds by default, but
2979 can be in any other unit if the number is suffixed by the unit,
2980 as explained at the top of this document.
2981
2982 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002983 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +01002984 cover one or several TCP packet losses by specifying timeouts that are
Davor Ocelice9ed2812017-12-25 17:49:28 +01002985 slightly above multiples of 3 seconds (e.g. 4 or 5 seconds). By default, the
Willy Tarreau0ba27502007-12-24 16:55:16 +01002986 connect timeout also presets the queue timeout to the same value if this one
2987 has not been specified. Historically, the contimeout was also used to set the
2988 tarpit timeout in a listen section, which is not possible in a pure frontend.
2989
2990 This parameter is specific to backends, but can be specified once for all in
2991 "defaults" sections. This is in fact one of the easiest solutions not to
2992 forget about it. An unspecified timeout results in an infinite timeout, which
2993 is not recommended. Such a usage is accepted and works but reports a warning
2994 during startup because it may results in accumulation of failed sessions in
2995 the system if the system's timeouts are not configured either.
2996
2997 This parameter is provided for backwards compatibility but is currently
2998 deprecated. Please use "timeout connect", "timeout queue" or "timeout tarpit"
2999 instead.
3000
3001 See also : "timeout connect", "timeout queue", "timeout tarpit",
3002 "timeout server", "contimeout".
3003
3004
Willy Tarreau55165fe2009-05-10 12:02:55 +02003005cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02003006 [ postonly ] [ preserve ] [ httponly ] [ secure ]
3007 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Olivier Houchard4e694042017-03-14 20:01:29 +01003008 [ dynamic ]
Willy Tarreau0ba27502007-12-24 16:55:16 +01003009 Enable cookie-based persistence in a backend.
3010 May be used in sections : defaults | frontend | listen | backend
3011 yes | no | yes | yes
3012 Arguments :
3013 <name> is the name of the cookie which will be monitored, modified or
3014 inserted in order to bring persistence. This cookie is sent to
3015 the client via a "Set-Cookie" header in the response, and is
3016 brought back by the client in a "Cookie" header in all requests.
3017 Special care should be taken to choose a name which does not
3018 conflict with any likely application cookie. Also, if the same
Davor Ocelice9ed2812017-12-25 17:49:28 +01003019 backends are subject to be used by the same clients (e.g.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003020 HTTP/HTTPS), care should be taken to use different cookie names
3021 between all backends if persistence between them is not desired.
3022
3023 rewrite This keyword indicates that the cookie will be provided by the
3024 server and that haproxy will have to modify its value to set the
3025 server's identifier in it. This mode is handy when the management
3026 of complex combinations of "Set-cookie" and "Cache-control"
3027 headers is left to the application. The application can then
3028 decide whether or not it is appropriate to emit a persistence
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003029 cookie. Since all responses should be monitored, this mode
3030 doesn't work in HTTP tunnel mode. Unless the application
Davor Ocelice9ed2812017-12-25 17:49:28 +01003031 behavior is very complex and/or broken, it is advised not to
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003032 start with this mode for new deployments. This keyword is
3033 incompatible with "insert" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003034
3035 insert This keyword indicates that the persistence cookie will have to
Willy Tarreaua79094d2010-08-31 22:54:15 +02003036 be inserted by haproxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003037
Willy Tarreaua79094d2010-08-31 22:54:15 +02003038 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003039 server. When used without the "preserve" option, if the server
3040 emits a cookie with the same name, it will be remove before
Davor Ocelice9ed2812017-12-25 17:49:28 +01003041 processing. For this reason, this mode can be used to upgrade
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003042 existing configurations running in the "rewrite" mode. The cookie
3043 will only be a session cookie and will not be stored on the
3044 client's disk. By default, unless the "indirect" option is added,
3045 the server will see the cookies emitted by the client. Due to
3046 caching effects, it is generally wise to add the "nocache" or
3047 "postonly" keywords (see below). The "insert" keyword is not
3048 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003049
3050 prefix This keyword indicates that instead of relying on a dedicated
3051 cookie for the persistence, an existing one will be completed.
3052 This may be needed in some specific environments where the client
3053 does not support more than one single cookie and the application
3054 already needs it. In this case, whenever the server sets a cookie
3055 named <name>, it will be prefixed with the server's identifier
3056 and a delimiter. The prefix will be removed from all client
3057 requests so that the server still finds the cookie it emitted.
3058 Since all requests and responses are subject to being modified,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003059 this mode doesn't work with tunnel mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02003060 not compatible with "rewrite" and "insert". Note: it is highly
3061 recommended not to use "indirect" with "prefix", otherwise server
3062 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003063
Willy Tarreaua79094d2010-08-31 22:54:15 +02003064 indirect When this option is specified, no cookie will be emitted to a
3065 client which already has a valid one for the server which has
3066 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003067 it will be removed, unless the "preserve" option is also set. In
3068 "insert" mode, this will additionally remove cookies from the
3069 requests transmitted to the server, making the persistence
3070 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02003071 Note: it is highly recommended not to use "indirect" with
3072 "prefix", otherwise server cookie updates would not be sent to
3073 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003074
3075 nocache This option is recommended in conjunction with the insert mode
3076 when there is a cache between the client and HAProxy, as it
3077 ensures that a cacheable response will be tagged non-cacheable if
3078 a cookie needs to be inserted. This is important because if all
3079 persistence cookies are added on a cacheable home page for
3080 instance, then all customers will then fetch the page from an
3081 outer cache and will all share the same persistence cookie,
3082 leading to one server receiving much more traffic than others.
3083 See also the "insert" and "postonly" options.
3084
3085 postonly This option ensures that cookie insertion will only be performed
3086 on responses to POST requests. It is an alternative to the
3087 "nocache" option, because POST responses are not cacheable, so
3088 this ensures that the persistence cookie will never get cached.
3089 Since most sites do not need any sort of persistence before the
3090 first POST which generally is a login request, this is a very
3091 efficient method to optimize caching without risking to find a
3092 persistence cookie in the cache.
3093 See also the "insert" and "nocache" options.
3094
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003095 preserve This option may only be used with "insert" and/or "indirect". It
3096 allows the server to emit the persistence cookie itself. In this
3097 case, if a cookie is found in the response, haproxy will leave it
3098 untouched. This is useful in order to end persistence after a
3099 logout request for instance. For this, the server just has to
Davor Ocelice9ed2812017-12-25 17:49:28 +01003100 emit a cookie with an invalid value (e.g. empty) or with a date in
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003101 the past. By combining this mechanism with the "disable-on-404"
3102 check option, it is possible to perform a completely graceful
3103 shutdown because users will definitely leave the server after
3104 they logout.
3105
Willy Tarreau4992dd22012-05-31 21:02:17 +02003106 httponly This option tells haproxy to add an "HttpOnly" cookie attribute
3107 when a cookie is inserted. This attribute is used so that a
3108 user agent doesn't share the cookie with non-HTTP components.
3109 Please check RFC6265 for more information on this attribute.
3110
3111 secure This option tells haproxy to add a "Secure" cookie attribute when
3112 a cookie is inserted. This attribute is used so that a user agent
3113 never emits this cookie over non-secure channels, which means
3114 that a cookie learned with this flag will be presented only over
3115 SSL/TLS connections. Please check RFC6265 for more information on
3116 this attribute.
3117
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02003118 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003119 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01003120 name. If the domain begins with a dot, the browser is allowed to
3121 use it for any host ending with that name. It is also possible to
3122 specify several domain names by invoking this option multiple
3123 times. Some browsers might have small limits on the number of
3124 domains, so be careful when doing that. For the record, sending
3125 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02003126
Willy Tarreau996a92c2010-10-13 19:30:47 +02003127 maxidle This option allows inserted cookies to be ignored after some idle
3128 time. It only works with insert-mode cookies. When a cookie is
3129 sent to the client, the date this cookie was emitted is sent too.
3130 Upon further presentations of this cookie, if the date is older
3131 than the delay indicated by the parameter (in seconds), it will
3132 be ignored. Otherwise, it will be refreshed if needed when the
3133 response is sent to the client. This is particularly useful to
3134 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01003135 too long on the same server (e.g. after a farm size change). When
Willy Tarreau996a92c2010-10-13 19:30:47 +02003136 this option is set and a cookie has no date, it is always
3137 accepted, but gets refreshed in the response. This maintains the
3138 ability for admins to access their sites. Cookies that have a
3139 date in the future further than 24 hours are ignored. Doing so
3140 lets admins fix timezone issues without risking kicking users off
3141 the site.
3142
3143 maxlife This option allows inserted cookies to be ignored after some life
3144 time, whether they're in use or not. It only works with insert
3145 mode cookies. When a cookie is first sent to the client, the date
3146 this cookie was emitted is sent too. Upon further presentations
3147 of this cookie, if the date is older than the delay indicated by
3148 the parameter (in seconds), it will be ignored. If the cookie in
3149 the request has no date, it is accepted and a date will be set.
3150 Cookies that have a date in the future further than 24 hours are
3151 ignored. Doing so lets admins fix timezone issues without risking
3152 kicking users off the site. Contrary to maxidle, this value is
3153 not refreshed, only the first visit date counts. Both maxidle and
3154 maxlife may be used at the time. This is particularly useful to
3155 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01003156 too long on the same server (e.g. after a farm size change). This
Willy Tarreau996a92c2010-10-13 19:30:47 +02003157 is stronger than the maxidle method in that it forces a
3158 redispatch after some absolute delay.
3159
Olivier Houchard4e694042017-03-14 20:01:29 +01003160 dynamic Activate dynamic cookies. When used, a session cookie is
3161 dynamically created for each server, based on the IP and port
3162 of the server, and a secret key, specified in the
3163 "dynamic-cookie-key" backend directive.
3164 The cookie will be regenerated each time the IP address change,
3165 and is only generated for IPv4/IPv6.
3166
Willy Tarreau0ba27502007-12-24 16:55:16 +01003167 There can be only one persistence cookie per HTTP backend, and it can be
3168 declared in a defaults section. The value of the cookie will be the value
3169 indicated after the "cookie" keyword in a "server" statement. If no cookie
3170 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02003171
Willy Tarreau0ba27502007-12-24 16:55:16 +01003172 Examples :
3173 cookie JSESSIONID prefix
3174 cookie SRV insert indirect nocache
3175 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02003176 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01003177
Willy Tarreau294d0f02015-08-10 19:40:12 +02003178 See also : "balance source", "capture cookie", "server" and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003179
Willy Tarreau983e01e2010-01-11 18:42:06 +01003180
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003181declare capture [ request | response ] len <length>
3182 Declares a capture slot.
3183 May be used in sections : defaults | frontend | listen | backend
3184 no | yes | yes | no
3185 Arguments:
3186 <length> is the length allowed for the capture.
3187
3188 This declaration is only available in the frontend or listen section, but the
3189 reserved slot can be used in the backends. The "request" keyword allocates a
3190 capture slot for use in the request, and "response" allocates a capture slot
3191 for use in the response.
3192
3193 See also: "capture-req", "capture-res" (sample converters),
Baptiste Assmann5ac425c2015-10-21 23:13:46 +02003194 "capture.req.hdr", "capture.res.hdr" (sample fetches),
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003195 "http-request capture" and "http-response capture".
3196
3197
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003198default-server [param*]
3199 Change default options for a server in a backend
3200 May be used in sections : defaults | frontend | listen | backend
3201 yes | no | yes | yes
3202 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01003203 <param*> is a list of parameters for this server. The "default-server"
3204 keyword accepts an important number of options and has a complete
3205 section dedicated to it. Please refer to section 5 for more
3206 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003207
Willy Tarreau983e01e2010-01-11 18:42:06 +01003208 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003209 default-server inter 1000 weight 13
3210
3211 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01003212
Willy Tarreau983e01e2010-01-11 18:42:06 +01003213
Willy Tarreau0ba27502007-12-24 16:55:16 +01003214default_backend <backend>
3215 Specify the backend to use when no "use_backend" rule has been matched.
3216 May be used in sections : defaults | frontend | listen | backend
3217 yes | yes | yes | no
3218 Arguments :
3219 <backend> is the name of the backend to use.
3220
3221 When doing content-switching between frontend and backends using the
3222 "use_backend" keyword, it is often useful to indicate which backend will be
3223 used when no rule has matched. It generally is the dynamic backend which
3224 will catch all undetermined requests.
3225
Willy Tarreau0ba27502007-12-24 16:55:16 +01003226 Example :
3227
3228 use_backend dynamic if url_dyn
3229 use_backend static if url_css url_img extension_img
3230 default_backend dynamic
3231
Willy Tarreau98d04852015-05-26 12:18:29 +02003232 See also : "use_backend"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003233
Willy Tarreau0ba27502007-12-24 16:55:16 +01003234
Baptiste Assmann27f51342013-10-09 06:51:49 +02003235description <string>
3236 Describe a listen, frontend or backend.
3237 May be used in sections : defaults | frontend | listen | backend
3238 no | yes | yes | yes
3239 Arguments : string
3240
3241 Allows to add a sentence to describe the related object in the HAProxy HTML
3242 stats page. The description will be printed on the right of the object name
3243 it describes.
3244 No need to backslash spaces in the <string> arguments.
3245
3246
Willy Tarreau0ba27502007-12-24 16:55:16 +01003247disabled
3248 Disable a proxy, frontend or backend.
3249 May be used in sections : defaults | frontend | listen | backend
3250 yes | yes | yes | yes
3251 Arguments : none
3252
3253 The "disabled" keyword is used to disable an instance, mainly in order to
3254 liberate a listening port or to temporarily disable a service. The instance
3255 will still be created and its configuration will be checked, but it will be
3256 created in the "stopped" state and will appear as such in the statistics. It
3257 will not receive any traffic nor will it send any health-checks or logs. It
3258 is possible to disable many instances at once by adding the "disabled"
3259 keyword in a "defaults" section.
3260
3261 See also : "enabled"
3262
3263
Willy Tarreau5ce94572010-06-07 14:35:41 +02003264dispatch <address>:<port>
3265 Set a default server address
3266 May be used in sections : defaults | frontend | listen | backend
3267 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003268 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02003269
3270 <address> is the IPv4 address of the default server. Alternatively, a
3271 resolvable hostname is supported, but this name will be resolved
3272 during start-up.
3273
3274 <ports> is a mandatory port specification. All connections will be sent
3275 to this port, and it is not permitted to use port offsets as is
3276 possible with normal servers.
3277
Willy Tarreau787aed52011-04-15 06:45:37 +02003278 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02003279 server can take the connection. In the past it was used to forward non
3280 persistent connections to an auxiliary load balancer. Due to its simple
3281 syntax, it has also been used for simple TCP relays. It is recommended not to
3282 use it for more clarity, and to use the "server" directive instead.
3283
3284 See also : "server"
3285
Olivier Houchard4e694042017-03-14 20:01:29 +01003286
3287dynamic-cookie-key <string>
3288 Set the dynamic cookie secret key for a backend.
3289 May be used in sections : defaults | frontend | listen | backend
3290 yes | no | yes | yes
3291 Arguments : The secret key to be used.
3292
3293 When dynamic cookies are enabled (see the "dynamic" directive for cookie),
Davor Ocelice9ed2812017-12-25 17:49:28 +01003294 a dynamic cookie is created for each server (unless one is explicitly
Olivier Houchard4e694042017-03-14 20:01:29 +01003295 specified on the "server" line), using a hash of the IP address of the
3296 server, the TCP port, and the secret key.
Davor Ocelice9ed2812017-12-25 17:49:28 +01003297 That way, we can ensure session persistence across multiple load-balancers,
Olivier Houchard4e694042017-03-14 20:01:29 +01003298 even if servers are dynamically added or removed.
Willy Tarreau5ce94572010-06-07 14:35:41 +02003299
Willy Tarreau0ba27502007-12-24 16:55:16 +01003300enabled
3301 Enable a proxy, frontend or backend.
3302 May be used in sections : defaults | frontend | listen | backend
3303 yes | yes | yes | yes
3304 Arguments : none
3305
3306 The "enabled" keyword is used to explicitly enable an instance, when the
3307 defaults has been set to "disabled". This is very rarely used.
3308
3309 See also : "disabled"
3310
3311
3312errorfile <code> <file>
3313 Return a file contents instead of errors generated by HAProxy
3314 May be used in sections : defaults | frontend | listen | backend
3315 yes | yes | yes | yes
3316 Arguments :
3317 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003318 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3319 503, and 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003320
3321 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003322 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01003323 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01003324 error pages, and to use absolute paths, since files are read
3325 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003326
3327 It is important to understand that this keyword is not meant to rewrite
3328 errors returned by the server, but errors detected and returned by HAProxy.
3329 This is why the list of supported errors is limited to a small set.
3330
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003331 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3332
Willy Tarreau0ba27502007-12-24 16:55:16 +01003333 The files are returned verbatim on the TCP socket. This allows any trick such
3334 as redirections to another URL or site, as well as tricks to clean cookies,
3335 force enable or disable caching, etc... The package provides default error
3336 files returning the same contents as default errors.
3337
Willy Tarreau59140a22009-02-22 12:02:30 +01003338 The files should not exceed the configured buffer size (BUFSIZE), which
3339 generally is 8 or 16 kB, otherwise they will be truncated. It is also wise
Davor Ocelice9ed2812017-12-25 17:49:28 +01003340 not to put any reference to local contents (e.g. images) in order to avoid
Willy Tarreau59140a22009-02-22 12:02:30 +01003341 loops between the client and HAProxy when all servers are down, causing an
3342 error to be returned instead of an image. For better HTTP compliance, it is
3343 recommended that all header lines end with CR-LF and not LF alone.
3344
Willy Tarreau0ba27502007-12-24 16:55:16 +01003345 The files are read at the same time as the configuration and kept in memory.
3346 For this reason, the errors continue to be returned even when the process is
3347 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01003348 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01003349 403 status code and interrogating a blocked URL.
3350
3351 See also : "errorloc", "errorloc302", "errorloc303"
3352
Willy Tarreau59140a22009-02-22 12:02:30 +01003353 Example :
3354 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau989222a2016-01-15 10:26:26 +01003355 errorfile 408 /dev/null # work around Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01003356 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
3357 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
3358
Willy Tarreau2769aa02007-12-27 18:26:09 +01003359
3360errorloc <code> <url>
3361errorloc302 <code> <url>
3362 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3363 May be used in sections : defaults | frontend | listen | backend
3364 yes | yes | yes | yes
3365 Arguments :
3366 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003367 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3368 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003369
3370 <url> it is the exact contents of the "Location" header. It may contain
3371 either a relative URI to an error page hosted on the same site,
3372 or an absolute URI designating an error page on another site.
3373 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01003374 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01003375
3376 It is important to understand that this keyword is not meant to rewrite
3377 errors returned by the server, but errors detected and returned by HAProxy.
3378 This is why the list of supported errors is limited to a small set.
3379
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003380 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3381
Willy Tarreau2769aa02007-12-27 18:26:09 +01003382 Note that both keyword return the HTTP 302 status code, which tells the
3383 client to fetch the designated URL using the same HTTP method. This can be
3384 quite problematic in case of non-GET methods such as POST, because the URL
3385 sent to the client might not be allowed for something other than GET. To
Willy Tarreau989222a2016-01-15 10:26:26 +01003386 work around this problem, please use "errorloc303" which send the HTTP 303
Willy Tarreau2769aa02007-12-27 18:26:09 +01003387 status code, indicating to the client that the URL must be fetched with a GET
3388 request.
3389
3390 See also : "errorfile", "errorloc303"
3391
3392
3393errorloc303 <code> <url>
3394 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3395 May be used in sections : defaults | frontend | listen | backend
3396 yes | yes | yes | yes
3397 Arguments :
3398 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003399 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3400 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003401
3402 <url> it is the exact contents of the "Location" header. It may contain
3403 either a relative URI to an error page hosted on the same site,
3404 or an absolute URI designating an error page on another site.
3405 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01003406 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01003407
3408 It is important to understand that this keyword is not meant to rewrite
3409 errors returned by the server, but errors detected and returned by HAProxy.
3410 This is why the list of supported errors is limited to a small set.
3411
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003412 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3413
Willy Tarreau2769aa02007-12-27 18:26:09 +01003414 Note that both keyword return the HTTP 303 status code, which tells the
3415 client to fetch the designated URL using the same HTTP GET method. This
3416 solves the usual problems associated with "errorloc" and the 302 code. It is
3417 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003418 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003419
3420 See also : "errorfile", "errorloc", "errorloc302"
3421
3422
Simon Horman51a1cf62015-02-03 13:00:44 +09003423email-alert from <emailaddr>
3424 Declare the from email address to be used in both the envelope and header
Davor Ocelice9ed2812017-12-25 17:49:28 +01003425 of email alerts. This is the address that email alerts are sent from.
Simon Horman51a1cf62015-02-03 13:00:44 +09003426 May be used in sections: defaults | frontend | listen | backend
3427 yes | yes | yes | yes
3428
3429 Arguments :
3430
3431 <emailaddr> is the from email address to use when sending email alerts
3432
3433 Also requires "email-alert mailers" and "email-alert to" to be set
3434 and if so sending email alerts is enabled for the proxy.
3435
Simon Horman64e34162015-02-06 11:11:57 +09003436 See also : "email-alert level", "email-alert mailers",
Cyril Bonté307ee1e2015-09-28 23:16:06 +02003437 "email-alert myhostname", "email-alert to", section 3.6 about
3438 mailers.
Simon Horman64e34162015-02-06 11:11:57 +09003439
3440
3441email-alert level <level>
3442 Declare the maximum log level of messages for which email alerts will be
3443 sent. This acts as a filter on the sending of email alerts.
3444 May be used in sections: defaults | frontend | listen | backend
3445 yes | yes | yes | yes
3446
3447 Arguments :
3448
3449 <level> One of the 8 syslog levels:
3450 emerg alert crit err warning notice info debug
3451 The above syslog levels are ordered from lowest to highest.
3452
3453 By default level is alert
3454
3455 Also requires "email-alert from", "email-alert mailers" and
3456 "email-alert to" to be set and if so sending email alerts is enabled
3457 for the proxy.
3458
Simon Horman1421e212015-04-30 13:10:35 +09003459 Alerts are sent when :
3460
3461 * An un-paused server is marked as down and <level> is alert or lower
3462 * A paused server is marked as down and <level> is notice or lower
3463 * A server is marked as up or enters the drain state and <level>
3464 is notice or lower
3465 * "option log-health-checks" is enabled, <level> is info or lower,
3466 and a health check status update occurs
3467
Simon Horman64e34162015-02-06 11:11:57 +09003468 See also : "email-alert from", "email-alert mailers",
3469 "email-alert myhostname", "email-alert to",
Simon Horman51a1cf62015-02-03 13:00:44 +09003470 section 3.6 about mailers.
3471
3472
3473email-alert mailers <mailersect>
3474 Declare the mailers to be used when sending email alerts
3475 May be used in sections: defaults | frontend | listen | backend
3476 yes | yes | yes | yes
3477
3478 Arguments :
3479
3480 <mailersect> is the name of the mailers section to send email alerts.
3481
3482 Also requires "email-alert from" and "email-alert to" to be set
3483 and if so sending email alerts is enabled for the proxy.
3484
Simon Horman64e34162015-02-06 11:11:57 +09003485 See also : "email-alert from", "email-alert level", "email-alert myhostname",
3486 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003487
3488
3489email-alert myhostname <hostname>
3490 Declare the to hostname address to be used when communicating with
3491 mailers.
3492 May be used in sections: defaults | frontend | listen | backend
3493 yes | yes | yes | yes
3494
3495 Arguments :
3496
Baptiste Assmann738bad92015-12-21 15:27:53 +01003497 <hostname> is the hostname to use when communicating with mailers
Simon Horman51a1cf62015-02-03 13:00:44 +09003498
3499 By default the systems hostname is used.
3500
3501 Also requires "email-alert from", "email-alert mailers" and
3502 "email-alert to" to be set and if so sending email alerts is enabled
3503 for the proxy.
3504
Simon Horman64e34162015-02-06 11:11:57 +09003505 See also : "email-alert from", "email-alert level", "email-alert mailers",
3506 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003507
3508
3509email-alert to <emailaddr>
Davor Ocelice9ed2812017-12-25 17:49:28 +01003510 Declare both the recipient address in the envelope and to address in the
Simon Horman51a1cf62015-02-03 13:00:44 +09003511 header of email alerts. This is the address that email alerts are sent to.
3512 May be used in sections: defaults | frontend | listen | backend
3513 yes | yes | yes | yes
3514
3515 Arguments :
3516
3517 <emailaddr> is the to email address to use when sending email alerts
3518
3519 Also requires "email-alert mailers" and "email-alert to" to be set
3520 and if so sending email alerts is enabled for the proxy.
3521
Simon Horman64e34162015-02-06 11:11:57 +09003522 See also : "email-alert from", "email-alert level", "email-alert mailers",
Simon Horman51a1cf62015-02-03 13:00:44 +09003523 "email-alert myhostname", section 3.6 about mailers.
3524
3525
Willy Tarreau4de91492010-01-22 19:10:05 +01003526force-persist { if | unless } <condition>
3527 Declare a condition to force persistence on down servers
3528 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003529 no | no | yes | yes
Willy Tarreau4de91492010-01-22 19:10:05 +01003530
3531 By default, requests are not dispatched to down servers. It is possible to
3532 force this using "option persist", but it is unconditional and redispatches
3533 to a valid server if "option redispatch" is set. That leaves with very little
3534 possibilities to force some requests to reach a server which is artificially
3535 marked down for maintenance operations.
3536
3537 The "force-persist" statement allows one to declare various ACL-based
3538 conditions which, when met, will cause a request to ignore the down status of
3539 a server and still try to connect to it. That makes it possible to start a
3540 server, still replying an error to the health checks, and run a specially
3541 configured browser to test the service. Among the handy methods, one could
3542 use a specific source IP address, or a specific cookie. The cookie also has
3543 the advantage that it can easily be added/removed on the browser from a test
3544 page. Once the service is validated, it is then possible to open the service
3545 to the world by returning a valid response to health checks.
3546
3547 The forced persistence is enabled when an "if" condition is met, or unless an
3548 "unless" condition is met. The final redispatch is always disabled when this
3549 is used.
3550
Cyril Bonté0d4bf012010-04-25 23:21:46 +02003551 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02003552 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01003553
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003554
3555filter <name> [param*]
3556 Add the filter <name> in the filter list attached to the proxy.
3557 May be used in sections : defaults | frontend | listen | backend
3558 no | yes | yes | yes
3559 Arguments :
3560 <name> is the name of the filter. Officially supported filters are
3561 referenced in section 9.
3562
Tim Düsterhus4896c442016-11-29 02:15:19 +01003563 <param*> is a list of parameters accepted by the filter <name>. The
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003564 parsing of these parameters are the responsibility of the
Tim Düsterhus4896c442016-11-29 02:15:19 +01003565 filter. Please refer to the documentation of the corresponding
3566 filter (section 9) for all details on the supported parameters.
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003567
3568 Multiple occurrences of the filter line can be used for the same proxy. The
3569 same filter can be referenced many times if needed.
3570
3571 Example:
3572 listen
3573 bind *:80
3574
3575 filter trace name BEFORE-HTTP-COMP
3576 filter compression
3577 filter trace name AFTER-HTTP-COMP
3578
3579 compression algo gzip
3580 compression offload
3581
3582 server srv1 192.168.0.1:80
3583
3584 See also : section 9.
3585
Willy Tarreau4de91492010-01-22 19:10:05 +01003586
Willy Tarreau2769aa02007-12-27 18:26:09 +01003587fullconn <conns>
3588 Specify at what backend load the servers will reach their maxconn
3589 May be used in sections : defaults | frontend | listen | backend
3590 yes | no | yes | yes
3591 Arguments :
3592 <conns> is the number of connections on the backend which will make the
3593 servers use the maximal number of connections.
3594
Willy Tarreau198a7442008-01-17 12:05:32 +01003595 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01003596 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01003597 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01003598 load. The server will then always accept at least <minconn> connections,
3599 never more than <maxconn>, and the limit will be on the ramp between both
3600 values when the backend has less than <conns> concurrent connections. This
3601 makes it possible to limit the load on the servers during normal loads, but
3602 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003603 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003604
Willy Tarreaufbb78422011-06-05 15:38:35 +02003605 Since it's hard to get this value right, haproxy automatically sets it to
3606 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01003607 backend (based on "use_backend" and "default_backend" rules). That way it's
3608 safe to leave it unset. However, "use_backend" involving dynamic names are
3609 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02003610
Willy Tarreau2769aa02007-12-27 18:26:09 +01003611 Example :
3612 # The servers will accept between 100 and 1000 concurrent connections each
3613 # and the maximum of 1000 will be reached when the backend reaches 10000
3614 # connections.
3615 backend dynamic
3616 fullconn 10000
3617 server srv1 dyn1:80 minconn 100 maxconn 1000
3618 server srv2 dyn2:80 minconn 100 maxconn 1000
3619
3620 See also : "maxconn", "server"
3621
3622
3623grace <time>
3624 Maintain a proxy operational for some time after a soft stop
3625 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01003626 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01003627 Arguments :
3628 <time> is the time (by default in milliseconds) for which the instance
3629 will remain operational with the frontend sockets still listening
3630 when a soft-stop is received via the SIGUSR1 signal.
3631
3632 This may be used to ensure that the services disappear in a certain order.
3633 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003634 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01003635 needed by the equipment to detect the failure.
3636
3637 Note that currently, there is very little benefit in using this parameter,
3638 and it may in fact complicate the soft-reconfiguration process more than
3639 simplify it.
3640
Willy Tarreau0ba27502007-12-24 16:55:16 +01003641
Andrew Rodland17be45e2016-10-25 17:04:12 -04003642hash-balance-factor <factor>
3643 Specify the balancing factor for bounded-load consistent hashing
3644 May be used in sections : defaults | frontend | listen | backend
3645 yes | no | no | yes
3646 Arguments :
3647 <factor> is the control for the maximum number of concurrent requests to
3648 send to a server, expressed as a percentage of the average number
3649 of concurrent requests across all of the active servers.
3650
3651 Specifying a "hash-balance-factor" for a server with "hash-type consistent"
3652 enables an algorithm that prevents any one server from getting too many
3653 requests at once, even if some hash buckets receive many more requests than
3654 others. Setting <factor> to 0 (the default) disables the feature. Otherwise,
3655 <factor> is a percentage greater than 100. For example, if <factor> is 150,
3656 then no server will be allowed to have a load more than 1.5 times the average.
3657 If server weights are used, they will be respected.
3658
3659 If the first-choice server is disqualified, the algorithm will choose another
3660 server based on the request hash, until a server with additional capacity is
3661 found. A higher <factor> allows more imbalance between the servers, while a
3662 lower <factor> means that more servers will be checked on average, affecting
3663 performance. Reasonable values are from 125 to 200.
3664
Willy Tarreau760e81d2018-05-03 07:20:40 +02003665 This setting is also used by "balance random" which internally relies on the
3666 consistent hashing mechanism.
3667
Andrew Rodland17be45e2016-10-25 17:04:12 -04003668 See also : "balance" and "hash-type".
3669
3670
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003671hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003672 Specify a method to use for mapping hashes to servers
3673 May be used in sections : defaults | frontend | listen | backend
3674 yes | no | yes | yes
3675 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04003676 <method> is the method used to select a server from the hash computed by
3677 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003678
Bhaskar98634f02013-10-29 23:30:51 -04003679 map-based the hash table is a static array containing all alive servers.
3680 The hashes will be very smooth, will consider weights, but
3681 will be static in that weight changes while a server is up
3682 will be ignored. This means that there will be no slow start.
3683 Also, since a server is selected by its position in the array,
3684 most mappings are changed when the server count changes. This
3685 means that when a server goes up or down, or when a server is
3686 added to a farm, most connections will be redistributed to
3687 different servers. This can be inconvenient with caches for
3688 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01003689
Bhaskar98634f02013-10-29 23:30:51 -04003690 consistent the hash table is a tree filled with many occurrences of each
3691 server. The hash key is looked up in the tree and the closest
3692 server is chosen. This hash is dynamic, it supports changing
3693 weights while the servers are up, so it is compatible with the
3694 slow start feature. It has the advantage that when a server
3695 goes up or down, only its associations are moved. When a
3696 server is added to the farm, only a few part of the mappings
3697 are redistributed, making it an ideal method for caches.
3698 However, due to its principle, the distribution will never be
3699 very smooth and it may sometimes be necessary to adjust a
3700 server's weight or its ID to get a more balanced distribution.
3701 In order to get the same distribution on multiple load
3702 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003703 same IDs. Note: consistent hash uses sdbm and avalanche if no
3704 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04003705
3706 <function> is the hash function to be used :
3707
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003708 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04003709 reimplementation of ndbm) database library. It was found to do
3710 well in scrambling bits, causing better distribution of the keys
3711 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003712 function with good distribution, unless the total server weight
3713 is a multiple of 64, in which case applying the avalanche
3714 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04003715
3716 djb2 this function was first proposed by Dan Bernstein many years ago
3717 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003718 function provides a better distribution than sdbm. It generally
3719 works well with text-based inputs though it can perform extremely
3720 poorly with numeric-only input or when the total server weight is
3721 a multiple of 33, unless the avalanche modifier is also used.
3722
Willy Tarreaua0f42712013-11-14 14:30:35 +01003723 wt6 this function was designed for haproxy while testing other
3724 functions in the past. It is not as smooth as the other ones, but
3725 is much less sensible to the input data set or to the number of
3726 servers. It can make sense as an alternative to sdbm+avalanche or
3727 djb2+avalanche for consistent hashing or when hashing on numeric
3728 data such as a source IP address or a visitor identifier in a URL
3729 parameter.
3730
Willy Tarreau324f07f2015-01-20 19:44:50 +01003731 crc32 this is the most common CRC32 implementation as used in Ethernet,
3732 gzip, PNG, etc. It is slower than the other ones but may provide
3733 a better distribution or less predictable results especially when
3734 used on strings.
3735
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003736 <modifier> indicates an optional method applied after hashing the key :
3737
3738 avalanche This directive indicates that the result from the hash
3739 function above should not be used in its raw form but that
3740 a 4-byte full avalanche hash must be applied first. The
3741 purpose of this step is to mix the resulting bits from the
3742 previous hash in order to avoid any undesired effect when
3743 the input contains some limited values or when the number of
3744 servers is a multiple of one of the hash's components (64
3745 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
3746 result less predictable, but it's also not as smooth as when
3747 using the original function. Some testing might be needed
3748 with some workloads. This hash is one of the many proposed
3749 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003750
Bhaskar98634f02013-10-29 23:30:51 -04003751 The default hash type is "map-based" and is recommended for most usages. The
3752 default function is "sdbm", the selection of a function should be based on
3753 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003754
Andrew Rodland17be45e2016-10-25 17:04:12 -04003755 See also : "balance", "hash-balance-factor", "server"
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003756
3757
Willy Tarreau0ba27502007-12-24 16:55:16 +01003758http-check disable-on-404
3759 Enable a maintenance mode upon HTTP/404 response to health-checks
3760 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01003761 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01003762 Arguments : none
3763
3764 When this option is set, a server which returns an HTTP code 404 will be
3765 excluded from further load-balancing, but will still receive persistent
3766 connections. This provides a very convenient method for Web administrators
3767 to perform a graceful shutdown of their servers. It is also important to note
3768 that a server which is detected as failed while it was in this mode will not
3769 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
3770 will immediately be reinserted into the farm. The status on the stats page
3771 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01003772 option only works in conjunction with the "httpchk" option. If this option
3773 is used with "http-check expect", then it has precedence over it so that 404
3774 responses will still be considered as soft-stop.
3775
3776 See also : "option httpchk", "http-check expect"
3777
3778
3779http-check expect [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003780 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01003781 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02003782 yes | no | yes | yes
Willy Tarreaubd741542010-03-16 18:46:54 +01003783 Arguments :
3784 <match> is a keyword indicating how to look for a specific pattern in the
3785 response. The keyword may be one of "status", "rstatus",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003786 "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01003787 exclamation mark ("!") to negate the match. Spaces are allowed
3788 between the exclamation mark and the keyword. See below for more
3789 details on the supported keywords.
3790
3791 <pattern> is the pattern to look for. It may be a string or a regular
3792 expression. If the pattern contains spaces, they must be escaped
3793 with the usual backslash ('\').
3794
3795 By default, "option httpchk" considers that response statuses 2xx and 3xx
3796 are valid, and that others are invalid. When "http-check expect" is used,
3797 it defines what is considered valid or invalid. Only one "http-check"
3798 statement is supported in a backend. If a server fails to respond or times
3799 out, the check obviously fails. The available matches are :
3800
3801 status <string> : test the exact string match for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003802 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003803 response's status code is exactly this string. If the
3804 "status" keyword is prefixed with "!", then the response
3805 will be considered invalid if the status code matches.
3806
3807 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003808 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003809 response's status code matches the expression. If the
3810 "rstatus" keyword is prefixed with "!", then the response
3811 will be considered invalid if the status code matches.
3812 This is mostly used to check for multiple codes.
3813
3814 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003815 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003816 response's body contains this exact string. If the
3817 "string" keyword is prefixed with "!", then the response
3818 will be considered invalid if the body contains this
3819 string. This can be used to look for a mandatory word at
3820 the end of a dynamic page, or to detect a failure when a
Davor Ocelice9ed2812017-12-25 17:49:28 +01003821 specific error appears on the check page (e.g. a stack
Willy Tarreaubd741542010-03-16 18:46:54 +01003822 trace).
3823
3824 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003825 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003826 response's body matches this expression. If the "rstring"
3827 keyword is prefixed with "!", then the response will be
3828 considered invalid if the body matches the expression.
3829 This can be used to look for a mandatory word at the end
3830 of a dynamic page, or to detect a failure when a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01003831 error appears on the check page (e.g. a stack trace).
Willy Tarreaubd741542010-03-16 18:46:54 +01003832
3833 It is important to note that the responses will be limited to a certain size
3834 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
3835 Thus, too large responses may not contain the mandatory pattern when using
3836 "string" or "rstring". If a large response is absolutely required, it is
3837 possible to change the default max size by setting the global variable.
3838 However, it is worth keeping in mind that parsing very large responses can
3839 waste some CPU cycles, especially when regular expressions are used, and that
3840 it is always better to focus the checks on smaller resources.
3841
Cyril Bonté32602d22015-01-30 00:07:07 +01003842 Also "http-check expect" doesn't support HTTP keep-alive. Keep in mind that it
3843 will automatically append a "Connection: close" header, meaning that this
3844 header should not be present in the request provided by "option httpchk".
3845
Willy Tarreaubd741542010-03-16 18:46:54 +01003846 Last, if "http-check expect" is combined with "http-check disable-on-404",
3847 then this last one has precedence when the server responds with 404.
3848
3849 Examples :
3850 # only accept status 200 as valid
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003851 http-check expect status 200
Willy Tarreaubd741542010-03-16 18:46:54 +01003852
3853 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003854 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01003855
3856 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003857 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01003858
3859 # check that we have a correct hexadecimal tag before /html
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03003860 http-check expect rstring <!--tag:[0-9a-f]*--></html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01003861
Willy Tarreaubd741542010-03-16 18:46:54 +01003862 See also : "option httpchk", "http-check disable-on-404"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003863
3864
Willy Tarreauef781042010-01-27 11:53:01 +01003865http-check send-state
3866 Enable emission of a state header with HTTP health checks
3867 May be used in sections : defaults | frontend | listen | backend
3868 yes | no | yes | yes
3869 Arguments : none
3870
3871 When this option is set, haproxy will systematically send a special header
3872 "X-Haproxy-Server-State" with a list of parameters indicating to each server
3873 how they are seen by haproxy. This can be used for instance when a server is
3874 manipulated without access to haproxy and the operator needs to know whether
3875 haproxy still sees it up or not, or if the server is the last one in a farm.
3876
3877 The header is composed of fields delimited by semi-colons, the first of which
3878 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
3879 checks on the total number before transition, just as appears in the stats
3880 interface. Next headers are in the form "<variable>=<value>", indicating in
3881 no specific order some values available in the stats interface :
Joseph Lynch514061c2015-01-15 17:52:59 -08003882 - a variable "address", containing the address of the backend server.
3883 This corresponds to the <address> field in the server declaration. For
3884 unix domain sockets, it will read "unix".
3885
3886 - a variable "port", containing the port of the backend server. This
3887 corresponds to the <port> field in the server declaration. For unix
3888 domain sockets, it will read "unix".
3889
Willy Tarreauef781042010-01-27 11:53:01 +01003890 - a variable "name", containing the name of the backend followed by a slash
3891 ("/") then the name of the server. This can be used when a server is
3892 checked in multiple backends.
3893
3894 - a variable "node" containing the name of the haproxy node, as set in the
3895 global "node" variable, otherwise the system's hostname if unspecified.
3896
3897 - a variable "weight" indicating the weight of the server, a slash ("/")
3898 and the total weight of the farm (just counting usable servers). This
3899 helps to know if other servers are available to handle the load when this
3900 one fails.
3901
3902 - a variable "scur" indicating the current number of concurrent connections
3903 on the server, followed by a slash ("/") then the total number of
3904 connections on all servers of the same backend.
3905
3906 - a variable "qcur" indicating the current number of requests in the
3907 server's queue.
3908
3909 Example of a header received by the application server :
3910 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
3911 scur=13/22; qcur=0
3912
3913 See also : "option httpchk", "http-check disable-on-404"
3914
Willy Tarreau53275e82017-11-24 07:52:01 +01003915http-request { allow | auth [realm <realm>] | redirect <rule> | reject |
Jarno Huuskonen800d1762017-03-06 14:56:36 +02003916 tarpit [deny_status <status>] | deny [deny_status <status>] |
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003917 add-header <name> <fmt> | set-header <name> <fmt> |
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02003918 capture <sample> [ len <length> | id <id> ] |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02003919 del-header <name> | set-nice <nice> | set-log-level <level> |
Sasha Pachev218f0642014-06-16 12:05:59 -06003920 replace-header <name> <match-regex> <replace-fmt> |
3921 replace-value <name> <match-regex> <replace-fmt> |
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01003922 set-method <fmt> | set-path <fmt> | set-query <fmt> |
3923 set-uri <fmt> | set-tos <tos> | set-mark <mark> |
Patrick Hemmer268a7072018-05-11 12:52:31 -04003924 set-priority-class <expr> | set-priority-offset <expr>
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003925 add-acl(<file name>) <key fmt> |
3926 del-acl(<file name>) <key fmt> |
3927 del-map(<file name>) <key fmt> |
Baptiste Assmannbb7e86a2014-09-03 18:29:47 +02003928 set-map(<file name>) <key fmt> <value fmt> |
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003929 set-var(<var name>) <expr> |
Christopher Faulet85d79c92016-11-09 16:54:56 +01003930 unset-var(<var name>) |
Thierry FOURNIER90da1912015-03-05 11:17:06 +01003931 { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] |
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02003932 sc-inc-gpc0(<sc-id>) |
Frédéric Lécaille6778b272018-01-29 15:22:53 +01003933 sc-inc-gpc1(<sc-id>) |
Thierry FOURNIER236657b2015-08-19 08:25:14 +02003934 sc-set-gpt0(<sc-id>) <int> |
Willy Tarreau2d392c22015-08-24 01:43:45 +02003935 silent-drop |
William Lallemand86d0df02017-11-24 21:36:45 +01003936 send-spoe-group |
3937 cache-use
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003938 }
Cyril Bontéf0c60612010-02-06 14:44:47 +01003939 [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003940 Access control for Layer 7 requests
3941
3942 May be used in sections: defaults | frontend | listen | backend
3943 no | yes | yes | yes
3944
Willy Tarreau20b0de52012-12-24 15:45:22 +01003945 The http-request statement defines a set of rules which apply to layer 7
3946 processing. The rules are evaluated in their declaration order when they are
3947 met in a frontend, listen or backend section. Any rule may optionally be
3948 followed by an ACL-based condition, in which case it will only be evaluated
3949 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003950
Willy Tarreau20b0de52012-12-24 15:45:22 +01003951 The first keyword is the rule's action. Currently supported actions include :
3952 - "allow" : this stops the evaluation of the rules and lets the request
3953 pass the check. No further "http-request" rules are evaluated.
3954
3955 - "deny" : this stops the evaluation of the rules and immediately rejects
Willy Tarreaube1d34d2016-06-26 19:37:59 +02003956 the request and emits an HTTP 403 error, or optionally the status code
3957 specified as an argument to "deny_status". The list of permitted status
3958 codes is limited to those that can be overridden by the "errorfile"
3959 directive. No further "http-request" rules are evaluated.
Willy Tarreau20b0de52012-12-24 15:45:22 +01003960
Willy Tarreau53275e82017-11-24 07:52:01 +01003961 - "reject" : this stops the evaluation of the rules and immediately closes
3962 the connection without sending any response. It acts similarly to the
3963 "tcp-request content reject" rules. It can be useful to force an
3964 immediate connection closure on HTTP/2 connections.
3965
Willy Tarreauccbcc372012-12-27 12:37:57 +01003966 - "tarpit" : this stops the evaluation of the rules and immediately blocks
3967 the request without responding for a delay specified by "timeout tarpit"
3968 or "timeout connect" if the former is not set. After that delay, if the
Jarno Huuskonen800d1762017-03-06 14:56:36 +02003969 client is still connected, an HTTP error 500 (or optionally the status
3970 code specified as an argument to "deny_status") is returned so that the
Willy Tarreauccbcc372012-12-27 12:37:57 +01003971 client does not suspect it has been tarpitted. Logs will report the flags
3972 "PT". The goal of the tarpit rule is to slow down robots during an attack
3973 when they're limited on the number of concurrent requests. It can be very
3974 efficient against very dumb robots, and will significantly reduce the
3975 load on firewalls compared to a "deny" rule. But when facing "correctly"
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003976 developed robots, it can make things worse by forcing haproxy and the
Willy Tarreau2d392c22015-08-24 01:43:45 +02003977 front firewall to support insane number of concurrent connections. See
3978 also the "silent-drop" action below.
Willy Tarreauccbcc372012-12-27 12:37:57 +01003979
Willy Tarreau20b0de52012-12-24 15:45:22 +01003980 - "auth" : this stops the evaluation of the rules and immediately responds
3981 with an HTTP 401 or 407 error code to invite the user to present a valid
3982 user name and password. No further "http-request" rules are evaluated. An
3983 optional "realm" parameter is supported, it sets the authentication realm
3984 that is returned with the response (typically the application's name).
3985
Willy Tarreau81499eb2012-12-27 12:19:02 +01003986 - "redirect" : this performs an HTTP redirection based on a redirect rule.
3987 This is exactly the same as the "redirect" statement except that it
3988 inserts a redirect rule which can be processed in the middle of other
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01003989 "http-request" rules and that these rules use the "log-format" strings.
3990 See the "redirect" keyword for the rule's syntax.
Willy Tarreau81499eb2012-12-27 12:19:02 +01003991
Willy Tarreau20b0de52012-12-24 15:45:22 +01003992 - "add-header" appends an HTTP header field whose name is specified in
3993 <name> and whose value is defined by <fmt> which follows the log-format
3994 rules (see Custom Log Format in section 8.2.4). This is particularly
Davor Ocelice9ed2812017-12-25 17:49:28 +01003995 useful to pass connection-specific information to the server (e.g. the
Willy Tarreau20b0de52012-12-24 15:45:22 +01003996 client's SSL certificate), or to combine several headers into one. This
3997 rule is not final, so it is possible to add other similar rules. Note
3998 that header addition is performed immediately, so one rule might reuse
3999 the resulting header from a previous rule.
4000
4001 - "set-header" does the same as "add-header" except that the header name
4002 is first removed if it existed. This is useful when passing security
4003 information to the server, where the header must not be manipulated by
Willy Tarreau85603282015-01-21 20:39:27 +01004004 external users. Note that the new value is computed before the removal so
4005 it is possible to concatenate a value to an existing header.
Willy Tarreau20b0de52012-12-24 15:45:22 +01004006
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004007 - "del-header" removes all HTTP header fields whose name is specified in
4008 <name>.
4009
Sasha Pachev218f0642014-06-16 12:05:59 -06004010 - "replace-header" matches the regular expression in all occurrences of
4011 header field <name> according to <match-regex>, and replaces them with
4012 the <replace-fmt> argument. Format characters are allowed in replace-fmt
4013 and work like in <fmt> arguments in "add-header". The match is only
4014 case-sensitive. It is important to understand that this action only
4015 considers whole header lines, regardless of the number of values they
4016 may contain. This usage is suited to headers naturally containing commas
4017 in their value, such as If-Modified-Since and so on.
4018
4019 Example:
4020
4021 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
4022
4023 applied to:
4024
4025 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
4026
4027 outputs:
4028
4029 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
4030
4031 assuming the backend IP is 192.168.1.20
4032
4033 - "replace-value" works like "replace-header" except that it matches the
4034 regex against every comma-delimited value of the header field <name>
4035 instead of the entire header. This is suited for all headers which are
4036 allowed to carry more than one value. An example could be the Accept
4037 header.
4038
4039 Example:
4040
4041 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
4042
4043 applied to:
4044
4045 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
4046
4047 outputs:
4048
4049 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
4050
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004051 - "set-method" rewrites the request method with the result of the
4052 evaluation of format string <fmt>. There should be very few valid reasons
4053 for having to do so as this is more likely to break something than to fix
4054 it.
4055
4056 - "set-path" rewrites the request path with the result of the evaluation of
4057 format string <fmt>. The query string, if any, is left intact. If a
4058 scheme and authority is found before the path, they are left intact as
4059 well. If the request doesn't have a path ("*"), this one is replaced with
4060 the format. This can be used to prepend a directory component in front of
4061 a path for example. See also "set-query" and "set-uri".
4062
4063 Example :
4064 # prepend the host name before the path
4065 http-request set-path /%[hdr(host)]%[path]
4066
4067 - "set-query" rewrites the request's query string which appears after the
4068 first question mark ("?") with the result of the evaluation of format
4069 string <fmt>. The part prior to the question mark is left intact. If the
4070 request doesn't contain a question mark and the new value is not empty,
4071 then one is added at the end of the URI, followed by the new value. If
4072 a question mark was present, it will never be removed even if the value
4073 is empty. This can be used to add or remove parameters from the query
4074 string. See also "set-query" and "set-uri".
4075
4076 Example :
4077 # replace "%3D" with "=" in the query string
4078 http-request set-query %[query,regsub(%3D,=,g)]
4079
4080 - "set-uri" rewrites the request URI with the result of the evaluation of
4081 format string <fmt>. The scheme, authority, path and query string are all
4082 replaced at once. This can be used to rewrite hosts in front of proxies,
4083 or to perform complex modifications to the URI such as moving parts
4084 between the path and the query string. See also "set-path" and
4085 "set-query".
4086
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004087 - "set-nice" sets the "nice" factor of the current request being processed.
4088 It only has effect against the other requests being processed at the same
4089 time. The default value is 0, unless altered by the "nice" setting on the
4090 "bind" line. The accepted range is -1024..1024. The higher the value, the
4091 nicest the request will be. Lower values will make the request more
4092 important than other ones. This can be useful to improve the speed of
4093 some requests, or lower the priority of non-important requests. Using
4094 this setting without prior experimentation can cause some major slowdown.
4095
Willy Tarreau9a355ec2013-06-11 17:45:46 +02004096 - "set-log-level" is used to change the log level of the current request
4097 when a certain condition is met. Valid levels are the 8 syslog levels
4098 (see the "log" keyword) plus the special level "silent" which disables
4099 logging for this request. This rule is not final so the last matching
4100 rule wins. This rule can be useful to disable health checks coming from
4101 another equipment.
4102
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004103 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
4104 the client to the value passed in <tos> on platforms which support this.
4105 This value represents the whole 8 bits of the IP TOS field, and can be
4106 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
4107 that only the 6 higher bits are used in DSCP or TOS, and the two lower
Davor Ocelice9ed2812017-12-25 17:49:28 +01004108 bits are always 0. This can be used to adjust some routing behavior on
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004109 border routers based on some information from the request. See RFC 2474,
4110 2597, 3260 and 4594 for more information.
4111
Willy Tarreau51347ed2013-06-11 19:34:13 +02004112 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
4113 client to the value passed in <mark> on platforms which support it. This
4114 value is an unsigned 32 bit value which can be matched by netfilter and
4115 by the routing table. It can be expressed both in decimal or hexadecimal
4116 format (prefixed by "0x"). This can be useful to force certain packets to
4117 take a different route (for example a cheaper network path for bulk
4118 downloads). This works on Linux kernels 2.6.32 and above and requires
4119 admin privileges.
4120
Patrick Hemmer268a7072018-05-11 12:52:31 -04004121 - "set-priority-class" is used to set the queue priority class of the
4122 current request. The value must be a sample expression which converts to
4123 an integer in the range -2047..2047. Results outside this range will be
4124 truncated. The priority class determines the order in which queued
4125 requests are processed. Lower values have higher priority.
4126
4127 - "set-priority-offset" is used to set the queue priority timestamp offset
4128 of the current request. The value must be a sample expression which
4129 converts to an integer in the range -524287..524287. Results outside this
4130 range will be truncated. When a request is queued, it is ordered first by
4131 the priority class, then by the current timestamp adjusted by the given
4132 offset in milliseconds. Lower values have higher priority.
4133 Note that the resulting timestamp is is only tracked with enough precision
4134 for 524,287ms (8m44s287ms). If the request is queued long enough to where
4135 the adjusted timestamp exceeds this value, it will be misidentified as
4136 highest priority. Thus it is important to set "timeout queue" to a value,
4137 where when combined with the offset, does not exceed this limit.
4138
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004139 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
4140 from a file (even a dummy empty file). The file name of the ACL to be
4141 updated is passed between parentheses. It takes one argument: <key fmt>,
4142 which follows log-format rules, to collect content of the new entry. It
4143 performs a lookup in the ACL before insertion, to avoid duplicated (or
4144 more) values. This lookup is done by a linear search and can be expensive
4145 with large lists! It is the equivalent of the "add acl" command from the
4146 stats socket, but can be triggered by an HTTP request.
4147
4148 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
4149 from a file (even a dummy empty file). The file name of the ACL to be
4150 updated is passed between parentheses. It takes one argument: <key fmt>,
4151 which follows log-format rules, to collect content of the entry to delete.
4152 It is the equivalent of the "del acl" command from the stats socket, but
4153 can be triggered by an HTTP request.
4154
4155 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
4156 from a file (even a dummy empty file). The file name of the MAP to be
4157 updated is passed between parentheses. It takes one argument: <key fmt>,
4158 which follows log-format rules, to collect content of the entry to delete.
4159 It takes one argument: "file name" It is the equivalent of the "del map"
4160 command from the stats socket, but can be triggered by an HTTP request.
4161
4162 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
4163 from a file (even a dummy empty file). The file name of the MAP to be
4164 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
4165 which follows log-format rules, used to collect MAP key, and <value fmt>,
4166 which follows log-format rules, used to collect content for the new entry.
4167 It performs a lookup in the MAP before insertion, to avoid duplicated (or
4168 more) values. This lookup is done by a linear search and can be expensive
4169 with large lists! It is the equivalent of the "set map" command from the
4170 stats socket, but can be triggered by an HTTP request.
4171
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02004172 - capture <sample> [ len <length> | id <id> ] :
Willy Tarreaua9083d02015-05-08 15:27:59 +02004173 captures sample expression <sample> from the request buffer, and converts
4174 it to a string of at most <len> characters. The resulting string is
4175 stored into the next request "capture" slot, so it will possibly appear
4176 next to some captured HTTP headers. It will then automatically appear in
4177 the logs, and it will be possible to extract it using sample fetch rules
4178 to feed it into headers or anything. The length should be limited given
4179 that this size will be allocated for each capture during the whole
4180 session life. Please check section 7.3 (Fetching samples) and "capture
4181 request header" for more information.
4182
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02004183 If the keyword "id" is used instead of "len", the action tries to store
4184 the captured string in a previously declared capture slot. This is useful
4185 to run captures in backends. The slot id can be declared by a previous
4186 directive "http-request capture" or with the "declare capture" keyword.
Baptiste Assmanne9544932015-11-03 23:31:35 +01004187 If the slot <id> doesn't exist, then HAProxy fails parsing the
4188 configuration to prevent unexpected behavior at run time.
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02004189
William Lallemand86d0df02017-11-24 21:36:45 +01004190 - cache-use <name> :
4191 See section 10.2 about cache setup.
4192
Willy Tarreau09448f72014-06-25 18:12:15 +02004193 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
4194 enables tracking of sticky counters from current request. These rules
4195 do not stop evaluation and do not change default action. Three sets of
4196 counters may be simultaneously tracked by the same connection. The first
4197 "track-sc0" rule executed enables tracking of the counters of the
4198 specified table as the first set. The first "track-sc1" rule executed
4199 enables tracking of the counters of the specified table as the second
4200 set. The first "track-sc2" rule executed enables tracking of the
4201 counters of the specified table as the third set. It is a recommended
4202 practice to use the first set of counters for the per-frontend counters
4203 and the second set for the per-backend ones. But this is just a
4204 guideline, all may be used everywhere.
4205
4206 These actions take one or two arguments :
4207 <key> is mandatory, and is a sample expression rule as described
4208 in section 7.3. It describes what elements of the incoming
Davor Ocelice9ed2812017-12-25 17:49:28 +01004209 request or connection will be analyzed, extracted, combined,
Willy Tarreau09448f72014-06-25 18:12:15 +02004210 and used to select which table entry to update the counters.
4211
4212 <table> is an optional table to be used instead of the default one,
4213 which is the stick-table declared in the current proxy. All
4214 the counters for the matches and updates for the key will
4215 then be performed in that table until the session ends.
4216
4217 Once a "track-sc*" rule is executed, the key is looked up in the table
4218 and if it is not found, an entry is allocated for it. Then a pointer to
4219 that entry is kept during all the session's life, and this entry's
4220 counters are updated as often as possible, every time the session's
4221 counters are updated, and also systematically when the session ends.
4222 Counters are only updated for events that happen after the tracking has
4223 been started. As an exception, connection counters and request counters
4224 are systematically updated so that they reflect useful information.
4225
4226 If the entry tracks concurrent connection counters, one connection is
4227 counted for as long as the entry is tracked, and the entry will not
4228 expire during that time. Tracking counters also provides a performance
4229 advantage over just checking the keys, because only one table lookup is
4230 performed for all ACL checks that make use of it.
4231
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004232 - sc-set-gpt0(<sc-id>) <int> :
4233 This action sets the GPT0 tag according to the sticky counter designated
4234 by <sc-id> and the value of <int>. The expected result is a boolean. If
4235 an error occurs, this action silently fails and the actions evaluation
4236 continues.
4237
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004238 - sc-inc-gpc0(<sc-id>):
4239 This action increments the GPC0 counter according with the sticky counter
4240 designated by <sc-id>. If an error occurs, this action silently fails and
4241 the actions evaluation continues.
4242
Frédéric Lécaille6778b272018-01-29 15:22:53 +01004243 - sc-inc-gpc1(<sc-id>):
4244 Same as "sc-inc-gpc0" action above but for GPC1 counter.
4245
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004246 - set-var(<var-name>) <expr> :
4247 Is used to set the contents of a variable. The variable is declared
4248 inline.
4249
Daniel Schneller0b547052016-03-21 20:46:57 +01004250 <var-name> The name of the variable starts with an indication about
4251 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01004252 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01004253 "sess" : the variable is shared with the whole session
4254 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004255 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01004256 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004257 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01004258 "res" : the variable is shared only during response
4259 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004260 This prefix is followed by a name. The separator is a '.'.
Daniel Schneller0b547052016-03-21 20:46:57 +01004261 The name may only contain characters 'a-z', 'A-Z', '0-9'
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004262 and '_'.
4263
4264 <expr> Is a standard HAProxy expression formed by a sample-fetch
4265 followed by some converters.
4266
4267 Example:
4268
4269 http-request set-var(req.my_var) req.fhdr(user-agent),lower
4270
Christopher Faulet85d79c92016-11-09 16:54:56 +01004271 - unset-var(<var-name>) :
4272 Is used to unset a variable. See above for details about <var-name>.
4273
4274 Example:
4275
4276 http-request unset-var(req.my_var)
4277
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004278 - set-src <expr> :
4279 Is used to set the source IP address to the value of specified
4280 expression. Useful when a proxy in front of HAProxy rewrites source IP,
4281 but provides the correct IP in a HTTP header; or you want to mask
4282 source IP for privacy.
4283
4284 <expr> Is a standard HAProxy expression formed by a sample-fetch
4285 followed by some converters.
4286
4287 Example:
4288
4289 http-request set-src hdr(x-forwarded-for)
4290 http-request set-src src,ipmask(24)
4291
Willy Tarreau00005ce2016-10-21 15:07:45 +02004292 When possible, set-src preserves the original source port as long as the
4293 address family allows it, otherwise the source port is set to 0.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004294
William Lallemand44be6402016-05-25 01:51:35 +02004295 - set-src-port <expr> :
4296 Is used to set the source port address to the value of specified
4297 expression.
4298
4299 <expr> Is a standard HAProxy expression formed by a sample-fetch
4300 followed by some converters.
4301
4302 Example:
4303
4304 http-request set-src-port hdr(x-port)
4305 http-request set-src-port int(4000)
4306
Willy Tarreau00005ce2016-10-21 15:07:45 +02004307 When possible, set-src-port preserves the original source address as long
4308 as the address family supports a port, otherwise it forces the source
4309 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02004310
William Lallemand13e9b0c2016-05-25 02:34:07 +02004311 - set-dst <expr> :
4312 Is used to set the destination IP address to the value of specified
4313 expression. Useful when a proxy in front of HAProxy rewrites destination
4314 IP, but provides the correct IP in a HTTP header; or you want to mask
4315 the IP for privacy. If you want to connect to the new address/port, use
4316 '0.0.0.0:0' as a server address in the backend.
4317
4318 <expr> Is a standard HAProxy expression formed by a sample-fetch
4319 followed by some converters.
4320
4321 Example:
4322
4323 http-request set-dst hdr(x-dst)
4324 http-request set-dst dst,ipmask(24)
4325
Willy Tarreau00005ce2016-10-21 15:07:45 +02004326 When possible, set-dst preserves the original destination port as long as
4327 the address family allows it, otherwise the destination port is set to 0.
4328
William Lallemand13e9b0c2016-05-25 02:34:07 +02004329 - set-dst-port <expr> :
4330 Is used to set the destination port address to the value of specified
4331 expression. If you want to connect to the new address/port, use
4332 '0.0.0.0:0' as a server address in the backend.
4333
4334 <expr> Is a standard HAProxy expression formed by a sample-fetch
4335 followed by some converters.
4336
4337 Example:
4338
4339 http-request set-dst-port hdr(x-port)
4340 http-request set-dst-port int(4000)
4341
Willy Tarreau00005ce2016-10-21 15:07:45 +02004342 When possible, set-dst-port preserves the original destination address as
4343 long as the address family supports a port, otherwise it forces the
4344 destination address to IPv4 "0.0.0.0" before rewriting the port.
4345
Willy Tarreau2d392c22015-08-24 01:43:45 +02004346 - "silent-drop" : this stops the evaluation of the rules and makes the
Davor Ocelice9ed2812017-12-25 17:49:28 +01004347 client-facing connection suddenly disappear using a system-dependent way
Willy Tarreau2d392c22015-08-24 01:43:45 +02004348 that tries to prevent the client from being notified. The effect it then
4349 that the client still sees an established connection while there's none
4350 on HAProxy. The purpose is to achieve a comparable effect to "tarpit"
4351 except that it doesn't use any local resource at all on the machine
4352 running HAProxy. It can resist much higher loads than "tarpit", and slow
Davor Ocelice9ed2812017-12-25 17:49:28 +01004353 down stronger attackers. It is important to understand the impact of using
4354 this mechanism. All stateful equipment placed between the client and
Willy Tarreau2d392c22015-08-24 01:43:45 +02004355 HAProxy (firewalls, proxies, load balancers) will also keep the
4356 established connection for a long time and may suffer from this action.
4357 On modern Linux systems running with enough privileges, the TCP_REPAIR
4358 socket option is used to block the emission of a TCP reset. On other
4359 systems, the socket's TTL is reduced to 1 so that the TCP reset doesn't
4360 pass the first router, though it's still delivered to local networks. Do
4361 not use it unless you fully understand how it works.
4362
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004363
Olivier Houchardccaa7de2017-10-02 11:51:03 +02004364 - "wait-for-handshake" : this will delay the processing of the request
4365 until the SSL handshake happened. This is mostly useful to delay
4366 processing early data until we're sure they are valid.
4367
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004368 - send-spoe-group <engine-name> <group-name> :
4369 This action is used to trigger sending of a group of SPOE messages. To do
4370 so, the SPOE engine used to send messages must be defined, as well as the
4371 SPOE group to send. Of course, the SPOE engine must refer to an existing
4372 SPOE filter. If not engine name is provided on the SPOE filter line, the
4373 SPOE agent name must be used.
4374
4375 <engine-name> The SPOE engine name.
4376
4377 <group-name> The SPOE group name as specified in the engine
4378 configuration.
4379
Willy Tarreau20b0de52012-12-24 15:45:22 +01004380 There is no limit to the number of http-request statements per instance.
4381
4382 It is important to know that http-request rules are processed very early in
4383 the HTTP processing, just after "block" rules and before "reqdel" or "reqrep"
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004384 or "reqadd" rules. That way, headers added by "add-header"/"set-header" are
4385 visible by almost all further ACL rules.
4386
4387 Using "reqadd"/"reqdel"/"reqrep" to manipulate request headers is discouraged
4388 in newer versions (>= 1.5). But if you need to use regular expression to
4389 delete headers, you can still use "reqdel". Also please use
4390 "http-request deny/allow/tarpit" instead of "reqdeny"/"reqpass"/"reqtarpit".
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004391
4392 Example:
Cyril Bonté78caf842010-03-10 22:41:43 +01004393 acl nagios src 192.168.129.3
4394 acl local_net src 192.168.0.0/16
4395 acl auth_ok http_auth(L1)
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004396
Cyril Bonté78caf842010-03-10 22:41:43 +01004397 http-request allow if nagios
4398 http-request allow if local_net auth_ok
4399 http-request auth realm Gimme if local_net auth_ok
4400 http-request deny
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004401
Cyril Bonté78caf842010-03-10 22:41:43 +01004402 Example:
4403 acl auth_ok http_auth_group(L1) G1
Cyril Bonté78caf842010-03-10 22:41:43 +01004404 http-request auth unless auth_ok
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004405
Willy Tarreau20b0de52012-12-24 15:45:22 +01004406 Example:
4407 http-request set-header X-Haproxy-Current-Date %T
4408 http-request set-header X-SSL %[ssl_fc]
Willy Tarreaufca42612015-08-27 17:15:05 +02004409 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id,hex]
Willy Tarreau20b0de52012-12-24 15:45:22 +01004410 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
4411 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
4412 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
4413 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
4414 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
4415 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
4416
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004417 Example:
4418 acl key req.hdr(X-Add-Acl-Key) -m found
4419 acl add path /addacl
4420 acl del path /delacl
4421
4422 acl myhost hdr(Host) -f myhost.lst
4423
4424 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
4425 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
4426
4427 Example:
4428 acl value req.hdr(X-Value) -m found
4429 acl setmap path /setmap
4430 acl delmap path /delmap
4431
4432 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
4433
4434 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
4435 http-request del-map(map.lst) %[src] if delmap
4436
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02004437 See also : "stats http-request", section 3.4 about userlists and section 7
4438 about ACL usage.
Willy Tarreauef781042010-01-27 11:53:01 +01004439
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004440http-response { allow | deny | add-header <name> <fmt> | set-nice <nice> |
Willy Tarreau51d861a2015-05-22 17:30:48 +02004441 capture <sample> id <id> | redirect <rule> |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004442 set-header <name> <fmt> | del-header <name> |
Sasha Pachev218f0642014-06-16 12:05:59 -06004443 replace-header <name> <regex-match> <replace-fmt> |
4444 replace-value <name> <regex-match> <replace-fmt> |
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004445 set-status <status> [reason <str>] |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004446 set-log-level <level> | set-mark <mark> | set-tos <tos> |
4447 add-acl(<file name>) <key fmt> |
4448 del-acl(<file name>) <key fmt> |
4449 del-map(<file name>) <key fmt> |
Thierry FOURNIER90da1912015-03-05 11:17:06 +01004450 set-map(<file name>) <key fmt> <value fmt> |
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004451 set-var(<var-name>) <expr> |
Christopher Faulet85d79c92016-11-09 16:54:56 +01004452 unset-var(<var-name>) |
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004453 { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] |
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004454 sc-inc-gpc0(<sc-id>) |
Frédéric Lécaille6778b272018-01-29 15:22:53 +01004455 sc-inc-gpc1(<sc-id>) |
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004456 sc-set-gpt0(<sc-id>) <int> |
Willy Tarreau2d392c22015-08-24 01:43:45 +02004457 silent-drop |
William Lallemand86d0df02017-11-24 21:36:45 +01004458 send-spoe-group |
4459 cache-store
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004460 }
Lukas Tribus2dd1d1a2013-06-19 23:34:41 +02004461 [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004462 Access control for Layer 7 responses
4463
4464 May be used in sections: defaults | frontend | listen | backend
4465 no | yes | yes | yes
4466
4467 The http-response statement defines a set of rules which apply to layer 7
4468 processing. The rules are evaluated in their declaration order when they are
4469 met in a frontend, listen or backend section. Any rule may optionally be
4470 followed by an ACL-based condition, in which case it will only be evaluated
4471 if the condition is true. Since these rules apply on responses, the backend
4472 rules are applied first, followed by the frontend's rules.
4473
4474 The first keyword is the rule's action. Currently supported actions include :
4475 - "allow" : this stops the evaluation of the rules and lets the response
4476 pass the check. No further "http-response" rules are evaluated for the
4477 current section.
4478
4479 - "deny" : this stops the evaluation of the rules and immediately rejects
4480 the response and emits an HTTP 502 error. No further "http-response"
4481 rules are evaluated.
4482
4483 - "add-header" appends an HTTP header field whose name is specified in
4484 <name> and whose value is defined by <fmt> which follows the log-format
4485 rules (see Custom Log Format in section 8.2.4). This may be used to send
4486 a cookie to a client for example, or to pass some internal information.
4487 This rule is not final, so it is possible to add other similar rules.
4488 Note that header addition is performed immediately, so one rule might
4489 reuse the resulting header from a previous rule.
4490
4491 - "set-header" does the same as "add-header" except that the header name
4492 is first removed if it existed. This is useful when passing security
4493 information to the server, where the header must not be manipulated by
4494 external users.
4495
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004496 - "del-header" removes all HTTP header fields whose name is specified in
4497 <name>.
4498
Sasha Pachev218f0642014-06-16 12:05:59 -06004499 - "replace-header" matches the regular expression in all occurrences of
4500 header field <name> according to <match-regex>, and replaces them with
4501 the <replace-fmt> argument. Format characters are allowed in replace-fmt
4502 and work like in <fmt> arguments in "add-header". The match is only
4503 case-sensitive. It is important to understand that this action only
4504 considers whole header lines, regardless of the number of values they
4505 may contain. This usage is suited to headers naturally containing commas
4506 in their value, such as Set-Cookie, Expires and so on.
4507
4508 Example:
4509
4510 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
4511
4512 applied to:
4513
4514 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
4515
4516 outputs:
4517
4518 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
4519
4520 assuming the backend IP is 192.168.1.20.
4521
4522 - "replace-value" works like "replace-header" except that it matches the
4523 regex against every comma-delimited value of the header field <name>
4524 instead of the entire header. This is suited for all headers which are
4525 allowed to carry more than one value. An example could be the Accept
4526 header.
4527
4528 Example:
4529
4530 http-response replace-value Cache-control ^public$ private
4531
4532 applied to:
4533
4534 Cache-Control: max-age=3600, public
4535
4536 outputs:
4537
4538 Cache-Control: max-age=3600, private
4539
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004540 - "set-status" replaces the response status code with <status> which must
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004541 be an integer between 100 and 999. Optionally, a custom reason text can be
4542 provided defined by <str>, or the default reason for the specified code
4543 will be used as a fallback.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004544
4545 Example:
4546
4547 # return "431 Request Header Fields Too Large"
4548 http-response set-status 431
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004549 # return "503 Slow Down", custom reason
4550 http-response set-status 503 reason "Slow Down".
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004551
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004552 - "set-nice" sets the "nice" factor of the current request being processed.
4553 It only has effect against the other requests being processed at the same
4554 time. The default value is 0, unless altered by the "nice" setting on the
4555 "bind" line. The accepted range is -1024..1024. The higher the value, the
4556 nicest the request will be. Lower values will make the request more
4557 important than other ones. This can be useful to improve the speed of
4558 some requests, or lower the priority of non-important requests. Using
4559 this setting without prior experimentation can cause some major slowdown.
4560
Willy Tarreau9a355ec2013-06-11 17:45:46 +02004561 - "set-log-level" is used to change the log level of the current request
4562 when a certain condition is met. Valid levels are the 8 syslog levels
4563 (see the "log" keyword) plus the special level "silent" which disables
4564 logging for this request. This rule is not final so the last matching
4565 rule wins. This rule can be useful to disable health checks coming from
4566 another equipment.
4567
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004568 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
4569 the client to the value passed in <tos> on platforms which support this.
4570 This value represents the whole 8 bits of the IP TOS field, and can be
4571 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
4572 that only the 6 higher bits are used in DSCP or TOS, and the two lower
Davor Ocelice9ed2812017-12-25 17:49:28 +01004573 bits are always 0. This can be used to adjust some routing behavior on
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004574 border routers based on some information from the request. See RFC 2474,
4575 2597, 3260 and 4594 for more information.
4576
Willy Tarreau51347ed2013-06-11 19:34:13 +02004577 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
4578 client to the value passed in <mark> on platforms which support it. This
4579 value is an unsigned 32 bit value which can be matched by netfilter and
4580 by the routing table. It can be expressed both in decimal or hexadecimal
4581 format (prefixed by "0x"). This can be useful to force certain packets to
4582 take a different route (for example a cheaper network path for bulk
4583 downloads). This works on Linux kernels 2.6.32 and above and requires
4584 admin privileges.
4585
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004586 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
4587 from a file (even a dummy empty file). The file name of the ACL to be
4588 updated is passed between parentheses. It takes one argument: <key fmt>,
4589 which follows log-format rules, to collect content of the new entry. It
4590 performs a lookup in the ACL before insertion, to avoid duplicated (or
4591 more) values. This lookup is done by a linear search and can be expensive
4592 with large lists! It is the equivalent of the "add acl" command from the
4593 stats socket, but can be triggered by an HTTP response.
4594
4595 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
4596 from a file (even a dummy empty file). The file name of the ACL to be
4597 updated is passed between parentheses. It takes one argument: <key fmt>,
4598 which follows log-format rules, to collect content of the entry to delete.
4599 It is the equivalent of the "del acl" command from the stats socket, but
4600 can be triggered by an HTTP response.
4601
4602 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
4603 from a file (even a dummy empty file). The file name of the MAP to be
4604 updated is passed between parentheses. It takes one argument: <key fmt>,
4605 which follows log-format rules, to collect content of the entry to delete.
4606 It takes one argument: "file name" It is the equivalent of the "del map"
4607 command from the stats socket, but can be triggered by an HTTP response.
4608
4609 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
4610 from a file (even a dummy empty file). The file name of the MAP to be
4611 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
4612 which follows log-format rules, used to collect MAP key, and <value fmt>,
4613 which follows log-format rules, used to collect content for the new entry.
4614 It performs a lookup in the MAP before insertion, to avoid duplicated (or
4615 more) values. This lookup is done by a linear search and can be expensive
4616 with large lists! It is the equivalent of the "set map" command from the
4617 stats socket, but can be triggered by an HTTP response.
4618
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004619 - capture <sample> id <id> :
4620 captures sample expression <sample> from the response buffer, and converts
4621 it to a string. The resulting string is stored into the next request
4622 "capture" slot, so it will possibly appear next to some captured HTTP
4623 headers. It will then automatically appear in the logs, and it will be
4624 possible to extract it using sample fetch rules to feed it into headers or
4625 anything. Please check section 7.3 (Fetching samples) and "capture
4626 response header" for more information.
4627
4628 The keyword "id" is the id of the capture slot which is used for storing
4629 the string. The capture slot must be defined in an associated frontend.
4630 This is useful to run captures in backends. The slot id can be declared by
4631 a previous directive "http-response capture" or with the "declare capture"
4632 keyword.
Baptiste Assmanne9544932015-11-03 23:31:35 +01004633 If the slot <id> doesn't exist, then HAProxy fails parsing the
4634 configuration to prevent unexpected behavior at run time.
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004635
William Lallemand86d0df02017-11-24 21:36:45 +01004636 - cache-store <name> :
4637 See section 10.2 about cache setup.
4638
Willy Tarreau51d861a2015-05-22 17:30:48 +02004639 - "redirect" : this performs an HTTP redirection based on a redirect rule.
4640 This supports a format string similarly to "http-request redirect" rules,
4641 with the exception that only the "location" type of redirect is possible
4642 on the response. See the "redirect" keyword for the rule's syntax. When
4643 a redirect rule is applied during a response, connections to the server
4644 are closed so that no data can be forwarded from the server to the client.
4645
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004646 - set-var(<var-name>) expr:
4647 Is used to set the contents of a variable. The variable is declared
4648 inline.
4649
Daniel Schneller0b547052016-03-21 20:46:57 +01004650 <var-name> The name of the variable starts with an indication about
4651 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01004652 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01004653 "sess" : the variable is shared with the whole session
4654 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004655 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01004656 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004657 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01004658 "res" : the variable is shared only during response
4659 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004660 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01004661 The name may only contain characters 'a-z', 'A-Z', '0-9',
4662 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004663
4664 <expr> Is a standard HAProxy expression formed by a sample-fetch
4665 followed by some converters.
4666
4667 Example:
4668
4669 http-response set-var(sess.last_redir) res.hdr(location)
4670
Christopher Faulet85d79c92016-11-09 16:54:56 +01004671 - unset-var(<var-name>) :
4672 Is used to unset a variable. See above for details about <var-name>.
4673
4674 Example:
4675
4676 http-response unset-var(sess.last_redir)
4677
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004678 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
4679 enables tracking of sticky counters from current response. Please refer to
4680 "http-request track-sc" for a complete description. The only difference
4681 from "http-request track-sc" is the <key> sample expression can only make
Davor Ocelice9ed2812017-12-25 17:49:28 +01004682 use of samples in response (e.g. res.*, status etc.) and samples below
4683 Layer 6 (e.g. SSL-related samples, see section 7.3.4). If the sample is
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004684 not supported, haproxy will fail and warn while parsing the config.
4685
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004686 - sc-set-gpt0(<sc-id>) <int> :
4687 This action sets the GPT0 tag according to the sticky counter designated
4688 by <sc-id> and the value of <int>. The expected result is a boolean. If
4689 an error occurs, this action silently fails and the actions evaluation
4690 continues.
4691
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004692 - sc-inc-gpc0(<sc-id>):
4693 This action increments the GPC0 counter according with the sticky counter
4694 designated by <sc-id>. If an error occurs, this action silently fails and
4695 the actions evaluation continues.
4696
Frédéric Lécaille6778b272018-01-29 15:22:53 +01004697 - sc-inc-gpc1(<sc-id>):
4698 Same as "sc-inc-gpc0" action above but for GPC1 counter.
4699
Willy Tarreau2d392c22015-08-24 01:43:45 +02004700 - "silent-drop" : this stops the evaluation of the rules and makes the
Davor Ocelice9ed2812017-12-25 17:49:28 +01004701 client-facing connection suddenly disappear using a system-dependent way
Willy Tarreau2d392c22015-08-24 01:43:45 +02004702 that tries to prevent the client from being notified. The effect it then
4703 that the client still sees an established connection while there's none
4704 on HAProxy. The purpose is to achieve a comparable effect to "tarpit"
4705 except that it doesn't use any local resource at all on the machine
4706 running HAProxy. It can resist much higher loads than "tarpit", and slow
Davor Ocelice9ed2812017-12-25 17:49:28 +01004707 down stronger attackers. It is important to understand the impact of using
4708 this mechanism. All stateful equipment placed between the client and
Willy Tarreau2d392c22015-08-24 01:43:45 +02004709 HAProxy (firewalls, proxies, load balancers) will also keep the
4710 established connection for a long time and may suffer from this action.
4711 On modern Linux systems running with enough privileges, the TCP_REPAIR
4712 socket option is used to block the emission of a TCP reset. On other
4713 systems, the socket's TTL is reduced to 1 so that the TCP reset doesn't
4714 pass the first router, though it's still delivered to local networks. Do
4715 not use it unless you fully understand how it works.
4716
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004717 - send-spoe-group <engine-name> <group-name> :
4718 This action is used to trigger sending of a group of SPOE messages. To do
4719 so, the SPOE engine used to send messages must be defined, as well as the
4720 SPOE group to send. Of course, the SPOE engine must refer to an existing
4721 SPOE filter. If not engine name is provided on the SPOE filter line, the
4722 SPOE agent name must be used.
4723
4724 <engine-name> The SPOE engine name.
4725
4726 <group-name> The SPOE group name as specified in the engine
4727 configuration.
4728
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004729 There is no limit to the number of http-response statements per instance.
4730
Godbach09250262013-07-02 01:19:15 +08004731 It is important to know that http-response rules are processed very early in
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004732 the HTTP processing, before "rspdel" or "rsprep" or "rspadd" rules. That way,
Davor Ocelice9ed2812017-12-25 17:49:28 +01004733 headers added by "add-header"/"set-header" are visible by almost all further
4734 ACL rules.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004735
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004736 Using "rspadd"/"rspdel"/"rsprep" to manipulate request headers is discouraged
4737 in newer versions (>= 1.5). But if you need to use regular expression to
4738 delete headers, you can still use "rspdel". Also please use
4739 "http-response deny" instead of "rspdeny".
4740
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004741 Example:
4742 acl key_acl res.hdr(X-Acl-Key) -m found
4743
4744 acl myhost hdr(Host) -f myhost.lst
4745
4746 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
4747 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
4748
4749 Example:
4750 acl value res.hdr(X-Value) -m found
4751
4752 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
4753
4754 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
4755 http-response del-map(map.lst) %[src] if ! value
4756
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004757 See also : "http-request", section 3.4 about userlists and section 7 about
4758 ACL usage.
4759
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02004760
Willy Tarreau30631952015-08-06 15:05:24 +02004761http-reuse { never | safe | aggressive | always }
4762 Declare how idle HTTP connections may be shared between requests
4763
4764 May be used in sections: defaults | frontend | listen | backend
4765 yes | no | yes | yes
4766
4767 By default, a connection established between haproxy and the backend server
4768 belongs to the session that initiated it. The downside is that between the
4769 response and the next request, the connection remains idle and is not used.
4770 In many cases for performance reasons it is desirable to make it possible to
4771 reuse these idle connections to serve other requests from different sessions.
Davor Ocelice9ed2812017-12-25 17:49:28 +01004772 This directive allows to tune this behavior.
Willy Tarreau30631952015-08-06 15:05:24 +02004773
4774 The argument indicates the desired connection reuse strategy :
4775
4776 - "never" : idle connections are never shared between sessions. This is
4777 the default choice. It may be enforced to cancel a different
4778 strategy inherited from a defaults section or for
4779 troubleshooting. For example, if an old bogus application
4780 considers that multiple requests over the same connection come
4781 from the same client and it is not possible to fix the
4782 application, it may be desirable to disable connection sharing
4783 in a single backend. An example of such an application could
4784 be an old haproxy using cookie insertion in tunnel mode and
4785 not checking any request past the first one.
4786
4787 - "safe" : this is the recommended strategy. The first request of a
4788 session is always sent over its own connection, and only
4789 subsequent requests may be dispatched over other existing
4790 connections. This ensures that in case the server closes the
4791 connection when the request is being sent, the browser can
4792 decide to silently retry it. Since it is exactly equivalent to
4793 regular keep-alive, there should be no side effects.
4794
4795 - "aggressive" : this mode may be useful in webservices environments where
4796 all servers are not necessarily known and where it would be
4797 appreciable to deliver most first requests over existing
4798 connections. In this case, first requests are only delivered
4799 over existing connections that have been reused at least once,
4800 proving that the server correctly supports connection reuse.
4801 It should only be used when it's sure that the client can
4802 retry a failed request once in a while and where the benefit
4803 of aggressive connection reuse significantly outweights the
4804 downsides of rare connection failures.
4805
4806 - "always" : this mode is only recommended when the path to the server is
4807 known for never breaking existing connections quickly after
4808 releasing them. It allows the first request of a session to be
4809 sent to an existing connection. This can provide a significant
4810 performance increase over the "safe" strategy when the backend
4811 is a cache farm, since such components tend to show a
Davor Ocelice9ed2812017-12-25 17:49:28 +01004812 consistent behavior and will benefit from the connection
Willy Tarreau30631952015-08-06 15:05:24 +02004813 sharing. It is recommended that the "http-keep-alive" timeout
4814 remains low in this mode so that no dead connections remain
4815 usable. In most cases, this will lead to the same performance
4816 gains as "aggressive" but with more risks. It should only be
4817 used when it improves the situation over "aggressive".
4818
4819 When http connection sharing is enabled, a great care is taken to respect the
Davor Ocelice9ed2812017-12-25 17:49:28 +01004820 connection properties and compatibility. Specifically :
4821 - connections made with "usesrc" followed by a client-dependent value
4822 ("client", "clientip", "hdr_ip") are marked private and never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02004823
4824 - connections sent to a server with a TLS SNI extension are marked private
Davor Ocelice9ed2812017-12-25 17:49:28 +01004825 and are never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02004826
4827 - connections receiving a status code 401 or 407 expect some authentication
4828 to be sent in return. Due to certain bogus authentication schemes (such
4829 as NTLM) relying on the connection, these connections are marked private
Davor Ocelice9ed2812017-12-25 17:49:28 +01004830 and are never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02004831
4832 No connection pool is involved, once a session dies, the last idle connection
4833 it was attached to is deleted at the same time. This ensures that connections
4834 may not last after all sessions are closed.
4835
4836 Note: connection reuse improves the accuracy of the "server maxconn" setting,
4837 because almost no new connection will be established while idle connections
4838 remain available. This is particularly true with the "always" strategy.
4839
4840 See also : "option http-keep-alive", "server maxconn"
4841
4842
Mark Lamourinec2247f02012-01-04 13:02:01 -05004843http-send-name-header [<header>]
4844 Add the server name to a request. Use the header string given by <header>
4845
4846 May be used in sections: defaults | frontend | listen | backend
4847 yes | no | yes | yes
4848
4849 Arguments :
4850
4851 <header> The header string to use to send the server name
4852
4853 The "http-send-name-header" statement causes the name of the target
Davor Ocelice9ed2812017-12-25 17:49:28 +01004854 server to be added to the headers of an HTTP request. The name
Mark Lamourinec2247f02012-01-04 13:02:01 -05004855 is added with the header string proved.
4856
4857 See also : "server"
4858
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01004859id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02004860 Set a persistent ID to a proxy.
4861 May be used in sections : defaults | frontend | listen | backend
4862 no | yes | yes | yes
4863 Arguments : none
4864
4865 Set a persistent ID for the proxy. This ID must be unique and positive.
4866 An unused ID will automatically be assigned if unset. The first assigned
4867 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01004868
4869
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004870ignore-persist { if | unless } <condition>
4871 Declare a condition to ignore persistence
4872 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01004873 no | no | yes | yes
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004874
4875 By default, when cookie persistence is enabled, every requests containing
4876 the cookie are unconditionally persistent (assuming the target server is up
4877 and running).
4878
4879 The "ignore-persist" statement allows one to declare various ACL-based
4880 conditions which, when met, will cause a request to ignore persistence.
4881 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03004882 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004883 persistence for a specific User-Agent (for example, some web crawler bots).
4884
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004885 The persistence is ignored when an "if" condition is met, or unless an
4886 "unless" condition is met.
4887
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03004888 Example:
4889 acl url_static path_beg /static /images /img /css
4890 acl url_static path_end .gif .png .jpg .css .js
4891 ignore-persist if url_static
4892
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004893 See also : "force-persist", "cookie", and section 7 about ACL usage.
4894
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004895load-server-state-from-file { global | local | none }
4896 Allow seamless reload of HAProxy
4897 May be used in sections: defaults | frontend | listen | backend
4898 yes | no | yes | yes
4899
4900 This directive points HAProxy to a file where server state from previous
4901 running process has been saved. That way, when starting up, before handling
4902 traffic, the new process can apply old states to servers exactly has if no
Davor Ocelice9ed2812017-12-25 17:49:28 +01004903 reload occurred. The purpose of the "load-server-state-from-file" directive is
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004904 to tell haproxy which file to use. For now, only 2 arguments to either prevent
4905 loading state or load states from a file containing all backends and servers.
4906 The state file can be generated by running the command "show servers state"
4907 over the stats socket and redirect output.
4908
Davor Ocelice9ed2812017-12-25 17:49:28 +01004909 The format of the file is versioned and is very specific. To understand it,
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004910 please read the documentation of the "show servers state" command (chapter
Willy Tarreau1af20c72017-06-23 16:01:14 +02004911 9.3 of Management Guide).
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004912
4913 Arguments:
4914 global load the content of the file pointed by the global directive
4915 named "server-state-file".
4916
4917 local load the content of the file pointed by the directive
4918 "server-state-file-name" if set. If not set, then the backend
4919 name is used as a file name.
4920
4921 none don't load any stat for this backend
4922
4923 Notes:
Willy Tarreaue5a60682016-11-09 14:54:53 +01004924 - server's IP address is preserved across reloads by default, but the
4925 order can be changed thanks to the server's "init-addr" setting. This
4926 means that an IP address change performed on the CLI at run time will
Davor Ocelice9ed2812017-12-25 17:49:28 +01004927 be preserved, and that any change to the local resolver (e.g. /etc/hosts)
Willy Tarreaue5a60682016-11-09 14:54:53 +01004928 will possibly not have any effect if the state file is in use.
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004929
4930 - server's weight is applied from previous running process unless it has
4931 has changed between previous and new configuration files.
4932
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004933 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004934
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004935 global
4936 stats socket /tmp/socket
4937 server-state-file /tmp/server_state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004938
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004939 defaults
4940 load-server-state-from-file global
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004941
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004942 backend bk
4943 server s1 127.0.0.1:22 check weight 11
4944 server s2 127.0.0.1:22 check weight 12
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004945
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004946
4947 Then one can run :
4948
4949 socat /tmp/socket - <<< "show servers state" > /tmp/server_state
4950
4951 Content of the file /tmp/server_state would be like this:
4952
4953 1
4954 # <field names skipped for the doc example>
4955 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
4956 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
4957
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004958 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004959
4960 global
4961 stats socket /tmp/socket
4962 server-state-base /etc/haproxy/states
4963
4964 defaults
4965 load-server-state-from-file local
4966
4967 backend bk
4968 server s1 127.0.0.1:22 check weight 11
4969 server s2 127.0.0.1:22 check weight 12
4970
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004971
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004972 Then one can run :
4973
4974 socat /tmp/socket - <<< "show servers state bk" > /etc/haproxy/states/bk
4975
4976 Content of the file /etc/haproxy/states/bk would be like this:
4977
4978 1
4979 # <field names skipped for the doc example>
4980 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
4981 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
4982
4983 See also: "server-state-file", "server-state-file-name", and
4984 "show servers state"
4985
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004986
Willy Tarreau2769aa02007-12-27 18:26:09 +01004987log global
Willy Tarreau18324f52014-06-27 18:10:07 +02004988log <address> [len <length>] <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02004989no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01004990 Enable per-instance logging of events and traffic.
4991 May be used in sections : defaults | frontend | listen | backend
4992 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02004993
4994 Prefix :
4995 no should be used when the logger list must be flushed. For example,
4996 if you don't want to inherit from the default logger list. This
4997 prefix does not allow arguments.
4998
Willy Tarreau2769aa02007-12-27 18:26:09 +01004999 Arguments :
5000 global should be used when the instance's logging parameters are the
5001 same as the global ones. This is the most common usage. "global"
5002 replaces <address>, <facility> and <level> with those of the log
5003 entries found in the "global" section. Only one "log global"
5004 statement may be used per instance, and this form takes no other
5005 parameter.
5006
5007 <address> indicates where to send the logs. It takes the same format as
5008 for the "global" section's logs, and can be one of :
5009
5010 - An IPv4 address optionally followed by a colon (':') and a UDP
5011 port. If no port is specified, 514 is used by default (the
5012 standard syslog port).
5013
David du Colombier24bb5f52011-03-17 10:40:23 +01005014 - An IPv6 address followed by a colon (':') and optionally a UDP
5015 port. If no port is specified, 514 is used by default (the
5016 standard syslog port).
5017
Willy Tarreau2769aa02007-12-27 18:26:09 +01005018 - A filesystem path to a UNIX domain socket, keeping in mind
5019 considerations for chroot (be sure the path is accessible
5020 inside the chroot) and uid/gid (be sure the path is
Davor Ocelice9ed2812017-12-25 17:49:28 +01005021 appropriately writable).
Willy Tarreau2769aa02007-12-27 18:26:09 +01005022
William Lallemandb2f07452015-05-12 14:27:13 +02005023 You may want to reference some environment variables in the
5024 address parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01005025
Willy Tarreau18324f52014-06-27 18:10:07 +02005026 <length> is an optional maximum line length. Log lines larger than this
5027 value will be truncated before being sent. The reason is that
5028 syslog servers act differently on log line length. All servers
5029 support the default value of 1024, but some servers simply drop
5030 larger lines while others do log them. If a server supports long
5031 lines, it may make sense to set this value here in order to avoid
5032 truncating long lines. Similarly, if a server drops long lines,
5033 it is preferable to truncate them before sending them. Accepted
5034 values are 80 to 65535 inclusive. The default value of 1024 is
5035 generally fine for all standard usages. Some specific cases of
Davor Ocelice9ed2812017-12-25 17:49:28 +01005036 long captures or JSON-formatted logs may require larger values.
Willy Tarreau18324f52014-06-27 18:10:07 +02005037
Willy Tarreau2769aa02007-12-27 18:26:09 +01005038 <facility> must be one of the 24 standard syslog facilities :
5039
5040 kern user mail daemon auth syslog lpr news
5041 uucp cron auth2 ftp ntp audit alert cron2
5042 local0 local1 local2 local3 local4 local5 local6 local7
5043
5044 <level> is optional and can be specified to filter outgoing messages. By
5045 default, all messages are sent. If a level is specified, only
5046 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02005047 will be sent. An optional minimum level can be specified. If it
5048 is set, logs emitted with a more severe level than this one will
5049 be capped to this level. This is used to avoid sending "emerg"
5050 messages on all terminals on some default syslog configurations.
5051 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01005052
5053 emerg alert crit err warning notice info debug
5054
William Lallemand0f99e342011-10-12 17:50:54 +02005055 It is important to keep in mind that it is the frontend which decides what to
5056 log from a connection, and that in case of content switching, the log entries
5057 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01005058
5059 However, backend log declaration define how and where servers status changes
5060 will be logged. Level "notice" will be used to indicate a server going up,
5061 "warning" will be used for termination signals and definitive service
5062 termination, and "alert" will be used for when a server goes down.
5063
5064 Note : According to RFC3164, messages are truncated to 1024 bytes before
5065 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005066
5067 Example :
5068 log global
Willy Tarreauf7edefa2009-05-10 17:20:05 +02005069 log 127.0.0.1:514 local0 notice # only send important events
5070 log 127.0.0.1:514 local0 notice notice # same but limit output level
William Lallemandb2f07452015-05-12 14:27:13 +02005071 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
Willy Tarreaudad36a32013-03-11 01:20:04 +01005072
Willy Tarreau2769aa02007-12-27 18:26:09 +01005073
William Lallemand48940402012-01-30 16:47:22 +01005074log-format <string>
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01005075 Specifies the log format string to use for traffic logs
5076 May be used in sections: defaults | frontend | listen | backend
5077 yes | yes | yes | no
William Lallemand48940402012-01-30 16:47:22 +01005078
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01005079 This directive specifies the log format string that will be used for all logs
5080 resulting from traffic passing through the frontend using this line. If the
5081 directive is used in a defaults section, all subsequent frontends will use
5082 the same log format. Please see section 8.2.4 which covers the log format
5083 string in depth.
William Lallemand48940402012-01-30 16:47:22 +01005084
Guillaume de Lafond29f45602017-03-31 19:52:15 +02005085 "log-format" directive overrides previous "option tcplog", "log-format" and
5086 "option httplog" directives.
5087
Dragan Dosen7ad31542015-09-28 17:16:47 +02005088log-format-sd <string>
5089 Specifies the RFC5424 structured-data log format string
5090 May be used in sections: defaults | frontend | listen | backend
5091 yes | yes | yes | no
5092
5093 This directive specifies the RFC5424 structured-data log format string that
5094 will be used for all logs resulting from traffic passing through the frontend
5095 using this line. If the directive is used in a defaults section, all
5096 subsequent frontends will use the same log format. Please see section 8.2.4
5097 which covers the log format string in depth.
5098
5099 See https://tools.ietf.org/html/rfc5424#section-6.3 for more information
5100 about the RFC5424 structured-data part.
5101
5102 Note : This log format string will be used only for loggers that have set
5103 log format to "rfc5424".
5104
5105 Example :
5106 log-format-sd [exampleSDID@1234\ bytes=\"%B\"\ status=\"%ST\"]
5107
5108
Willy Tarreau094af4e2015-01-07 15:03:42 +01005109log-tag <string>
5110 Specifies the log tag to use for all outgoing logs
5111 May be used in sections: defaults | frontend | listen | backend
5112 yes | yes | yes | yes
5113
5114 Sets the tag field in the syslog header to this string. It defaults to the
5115 log-tag set in the global section, otherwise the program name as launched
5116 from the command line, which usually is "haproxy". Sometimes it can be useful
5117 to differentiate between multiple processes running on the same host, or to
5118 differentiate customer instances running in the same process. In the backend,
5119 logs about servers up/down will use this tag. As a hint, it can be convenient
5120 to set a log-tag related to a hosted customer in a defaults section then put
5121 all the frontends and backends for that customer, then start another customer
5122 in a new defaults section. See also the global "log-tag" directive.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005123
Willy Tarreauc35362a2014-04-25 13:58:37 +02005124max-keep-alive-queue <value>
5125 Set the maximum server queue size for maintaining keep-alive connections
5126 May be used in sections: defaults | frontend | listen | backend
5127 yes | no | yes | yes
5128
5129 HTTP keep-alive tries to reuse the same server connection whenever possible,
5130 but sometimes it can be counter-productive, for example if a server has a lot
5131 of connections while other ones are idle. This is especially true for static
5132 servers.
5133
5134 The purpose of this setting is to set a threshold on the number of queued
5135 connections at which haproxy stops trying to reuse the same server and prefers
5136 to find another one. The default value, -1, means there is no limit. A value
5137 of zero means that keep-alive requests will never be queued. For very close
5138 servers which can be reached with a low latency and which are not sensible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01005139 breaking keep-alive, a low value is recommended (e.g. local static server can
Willy Tarreauc35362a2014-04-25 13:58:37 +02005140 use a value of 10 or less). For remote servers suffering from a high latency,
5141 higher values might be needed to cover for the latency and/or the cost of
5142 picking a different server.
5143
5144 Note that this has no impact on responses which are maintained to the same
5145 server consecutively to a 401 response. They will still go to the same server
5146 even if they have to be queued.
5147
5148 See also : "option http-server-close", "option prefer-last-server", server
5149 "maxconn" and cookie persistence.
5150
5151
Willy Tarreau2769aa02007-12-27 18:26:09 +01005152maxconn <conns>
5153 Fix the maximum number of concurrent connections on a frontend
5154 May be used in sections : defaults | frontend | listen | backend
5155 yes | yes | yes | no
5156 Arguments :
5157 <conns> is the maximum number of concurrent connections the frontend will
5158 accept to serve. Excess connections will be queued by the system
5159 in the socket's listen queue and will be served once a connection
5160 closes.
5161
5162 If the system supports it, it can be useful on big sites to raise this limit
5163 very high so that haproxy manages connection queues, instead of leaving the
5164 clients with unanswered connection attempts. This value should not exceed the
5165 global maxconn. Also, keep in mind that a connection contains two buffers
Baptiste Assmann79fb45d2016-03-06 23:34:31 +01005166 of tune.bufsize (16kB by default) each, as well as some other data resulting
5167 in about 33 kB of RAM being consumed per established connection. That means
5168 that a medium system equipped with 1GB of RAM can withstand around
5169 20000-25000 concurrent connections if properly tuned.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005170
5171 Also, when <conns> is set to large values, it is possible that the servers
5172 are not sized to accept such loads, and for this reason it is generally wise
5173 to assign them some reasonable connection limits.
5174
Vincent Bernat6341be52012-06-27 17:18:30 +02005175 By default, this value is set to 2000.
5176
Willy Tarreau2769aa02007-12-27 18:26:09 +01005177 See also : "server", global section's "maxconn", "fullconn"
5178
5179
5180mode { tcp|http|health }
5181 Set the running mode or protocol of the instance
5182 May be used in sections : defaults | frontend | listen | backend
5183 yes | yes | yes | yes
5184 Arguments :
5185 tcp The instance will work in pure TCP mode. A full-duplex connection
5186 will be established between clients and servers, and no layer 7
5187 examination will be performed. This is the default mode. It
5188 should be used for SSL, SSH, SMTP, ...
5189
5190 http The instance will work in HTTP mode. The client request will be
5191 analyzed in depth before connecting to any server. Any request
5192 which is not RFC-compliant will be rejected. Layer 7 filtering,
5193 processing and switching will be possible. This is the mode which
5194 brings HAProxy most of its value.
5195
5196 health The instance will work in "health" mode. It will just reply "OK"
Willy Tarreau82569f92012-09-27 23:48:56 +02005197 to incoming connections and close the connection. Alternatively,
5198 If the "httpchk" option is set, "HTTP/1.0 200 OK" will be sent
5199 instead. Nothing will be logged in either case. This mode is used
5200 to reply to external components health checks. This mode is
5201 deprecated and should not be used anymore as it is possible to do
5202 the same and even better by combining TCP or HTTP modes with the
5203 "monitor" keyword.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005204
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005205 When doing content switching, it is mandatory that the frontend and the
5206 backend are in the same mode (generally HTTP), otherwise the configuration
5207 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005208
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005209 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01005210 defaults http_instances
5211 mode http
5212
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005213 See also : "monitor", "monitor-net"
Willy Tarreau2769aa02007-12-27 18:26:09 +01005214
Willy Tarreau0ba27502007-12-24 16:55:16 +01005215
Cyril Bontéf0c60612010-02-06 14:44:47 +01005216monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01005217 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005218 May be used in sections : defaults | frontend | listen | backend
5219 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01005220 Arguments :
5221 if <cond> the monitor request will fail if the condition is satisfied,
5222 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005223 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01005224 are met, for instance a low number of servers both in a
5225 backend and its backup.
5226
5227 unless <cond> the monitor request will succeed only if the condition is
5228 satisfied, and will fail otherwise. Such a condition may be
5229 based on a test on the presence of a minimum number of active
5230 servers in a list of backends.
5231
5232 This statement adds a condition which can force the response to a monitor
5233 request to report a failure. By default, when an external component queries
5234 the URI dedicated to monitoring, a 200 response is returned. When one of the
5235 conditions above is met, haproxy will return 503 instead of 200. This is
5236 very useful to report a site failure to an external component which may base
5237 routing advertisements between multiple sites on the availability reported by
5238 haproxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02005239 criterion. Note that "monitor fail" only works in HTTP mode. Both status
5240 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005241
5242 Example:
5243 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01005244 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01005245 acl site_dead nbsrv(dynamic) lt 2
5246 acl site_dead nbsrv(static) lt 2
5247 monitor-uri /site_alive
5248 monitor fail if site_dead
5249
Willy Tarreauae94d4d2011-05-11 16:28:49 +02005250 See also : "monitor-net", "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01005251
5252
5253monitor-net <source>
5254 Declare a source network which is limited to monitor requests
5255 May be used in sections : defaults | frontend | listen | backend
5256 yes | yes | yes | no
5257 Arguments :
5258 <source> is the source IPv4 address or network which will only be able to
5259 get monitor responses to any request. It can be either an IPv4
5260 address, a host name, or an address followed by a slash ('/')
5261 followed by a mask.
5262
5263 In TCP mode, any connection coming from a source matching <source> will cause
5264 the connection to be immediately closed without any log. This allows another
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005265 equipment to probe the port and verify that it is still listening, without
Willy Tarreau2769aa02007-12-27 18:26:09 +01005266 forwarding the connection to a remote server.
5267
5268 In HTTP mode, a connection coming from a source matching <source> will be
5269 accepted, the following response will be sent without waiting for a request,
5270 then the connection will be closed : "HTTP/1.0 200 OK". This is normally
5271 enough for any front-end HTTP probe to detect that the service is UP and
Willy Tarreau82569f92012-09-27 23:48:56 +02005272 running without forwarding the request to a backend server. Note that this
5273 response is sent in raw format, without any transformation. This is important
5274 as it means that it will not be SSL-encrypted on SSL listeners.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005275
Willy Tarreau82569f92012-09-27 23:48:56 +02005276 Monitor requests are processed very early, just after tcp-request connection
5277 ACLs which are the only ones able to block them. These connections are short
5278 lived and never wait for any data from the client. They cannot be logged, and
5279 it is the intended purpose. They are only used to report HAProxy's health to
5280 an upper component, nothing more. Please note that "monitor fail" rules do
5281 not apply to connections intercepted by "monitor-net".
Willy Tarreau2769aa02007-12-27 18:26:09 +01005282
Willy Tarreau95cd2832010-03-04 23:36:33 +01005283 Last, please note that only one "monitor-net" statement can be specified in
5284 a frontend. If more than one is found, only the last one will be considered.
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005285
Willy Tarreau2769aa02007-12-27 18:26:09 +01005286 Example :
5287 # addresses .252 and .253 are just probing us.
5288 frontend www
5289 monitor-net 192.168.0.252/31
5290
5291 See also : "monitor fail", "monitor-uri"
5292
5293
5294monitor-uri <uri>
5295 Intercept a URI used by external components' monitor requests
5296 May be used in sections : defaults | frontend | listen | backend
5297 yes | yes | yes | no
5298 Arguments :
5299 <uri> is the exact URI which we want to intercept to return HAProxy's
5300 health status instead of forwarding the request.
5301
5302 When an HTTP request referencing <uri> will be received on a frontend,
5303 HAProxy will not forward it nor log it, but instead will return either
5304 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
5305 conditions defined with "monitor fail". This is normally enough for any
5306 front-end HTTP probe to detect that the service is UP and running without
5307 forwarding the request to a backend server. Note that the HTTP method, the
5308 version and all headers are ignored, but the request must at least be valid
5309 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
5310
Willy Tarreau721d8e02017-12-01 18:25:08 +01005311 Monitor requests are processed very early, just after the request is parsed
5312 and even before any "http-request" or "block" rulesets. The only rulesets
5313 applied before are the tcp-request ones. They cannot be logged either, and it
5314 is the intended purpose. They are only used to report HAProxy's health to an
5315 upper component, nothing more. However, it is possible to add any number of
5316 conditions using "monitor fail" and ACLs so that the result can be adjusted
5317 to whatever check can be imagined (most often the number of available servers
5318 in a backend).
Willy Tarreau2769aa02007-12-27 18:26:09 +01005319
5320 Example :
5321 # Use /haproxy_test to report haproxy's status
5322 frontend www
5323 mode http
5324 monitor-uri /haproxy_test
5325
5326 See also : "monitor fail", "monitor-net"
5327
Willy Tarreau0ba27502007-12-24 16:55:16 +01005328
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005329option abortonclose
5330no option abortonclose
5331 Enable or disable early dropping of aborted requests pending in queues.
5332 May be used in sections : defaults | frontend | listen | backend
5333 yes | no | yes | yes
5334 Arguments : none
5335
5336 In presence of very high loads, the servers will take some time to respond.
5337 The per-instance connection queue will inflate, and the response time will
5338 increase respective to the size of the queue times the average per-session
5339 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01005340 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005341 the queue, and slowing down other users, and the servers as well, because the
5342 request will eventually be served, then aborted at the first error
5343 encountered while delivering the response.
5344
5345 As there is no way to distinguish between a full STOP and a simple output
5346 close on the client side, HTTP agents should be conservative and consider
5347 that the client might only have closed its output channel while waiting for
5348 the response. However, this introduces risks of congestion when lots of users
5349 do the same, and is completely useless nowadays because probably no client at
5350 all will close the session while waiting for the response. Some HTTP agents
Davor Ocelice9ed2812017-12-25 17:49:28 +01005351 support this behavior (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005352 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01005353 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005354 of being the single component to break rare but valid traffic is extremely
5355 low, which adds to the temptation to be able to abort a session early while
5356 still not served and not pollute the servers.
5357
Davor Ocelice9ed2812017-12-25 17:49:28 +01005358 In HAProxy, the user can choose the desired behavior using the option
5359 "abortonclose". By default (without the option) the behavior is HTTP
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005360 compliant and aborted requests will be served. But when the option is
5361 specified, a session with an incoming channel closed will be aborted while
5362 it is still possible, either pending in the queue for a connection slot, or
5363 during the connection establishment if the server has not yet acknowledged
5364 the connection request. This considerably reduces the queue size and the load
5365 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01005366 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005367
5368 If this option has been enabled in a "defaults" section, it can be disabled
5369 in a specific instance by prepending the "no" keyword before it.
5370
5371 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
5372
5373
Willy Tarreau4076a152009-04-02 15:18:36 +02005374option accept-invalid-http-request
5375no option accept-invalid-http-request
5376 Enable or disable relaxing of HTTP request parsing
5377 May be used in sections : defaults | frontend | listen | backend
5378 yes | yes | yes | no
5379 Arguments : none
5380
Willy Tarreau91852eb2015-05-01 13:26:00 +02005381 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005382 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01005383 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02005384 forbidden characters are essentially used to build attacks exploiting server
5385 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5386 server will emit invalid header names for whatever reason (configuration,
5387 implementation) and the issue will not be immediately fixed. In such a case,
5388 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01005389 even if that does not make sense, by specifying this option. Similarly, the
5390 list of characters allowed to appear in a URI is well defined by RFC3986, and
5391 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
5392 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
Davor Ocelice9ed2812017-12-25 17:49:28 +01005393 not allowed at all. HAProxy always blocks a number of them (0..32, 127). The
Willy Tarreau91852eb2015-05-01 13:26:00 +02005394 remaining ones are blocked by default unless this option is enabled. This
Willy Tarreau13317662015-05-01 13:47:08 +02005395 option also relaxes the test on the HTTP version, it allows HTTP/0.9 requests
5396 to pass through (no version specified) and multiple digits for both the major
5397 and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005398
5399 This option should never be enabled by default as it hides application bugs
5400 and open security breaches. It should only be deployed after a problem has
5401 been confirmed.
5402
5403 When this option is enabled, erroneous header names will still be accepted in
5404 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01005405 analysis using the "show errors" request on the UNIX stats socket. Similarly,
5406 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02005407 also helps confirming that the issue has been solved.
5408
5409 If this option has been enabled in a "defaults" section, it can be disabled
5410 in a specific instance by prepending the "no" keyword before it.
5411
5412 See also : "option accept-invalid-http-response" and "show errors" on the
5413 stats socket.
5414
5415
5416option accept-invalid-http-response
5417no option accept-invalid-http-response
5418 Enable or disable relaxing of HTTP response parsing
5419 May be used in sections : defaults | frontend | listen | backend
5420 yes | no | yes | yes
5421 Arguments : none
5422
Willy Tarreau91852eb2015-05-01 13:26:00 +02005423 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005424 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01005425 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02005426 forbidden characters are essentially used to build attacks exploiting server
5427 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5428 server will emit invalid header names for whatever reason (configuration,
5429 implementation) and the issue will not be immediately fixed. In such a case,
5430 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau91852eb2015-05-01 13:26:00 +02005431 even if that does not make sense, by specifying this option. This option also
5432 relaxes the test on the HTTP version format, it allows multiple digits for
5433 both the major and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005434
5435 This option should never be enabled by default as it hides application bugs
5436 and open security breaches. It should only be deployed after a problem has
5437 been confirmed.
5438
5439 When this option is enabled, erroneous header names will still be accepted in
5440 responses, but the complete response will be captured in order to permit
5441 later analysis using the "show errors" request on the UNIX stats socket.
5442 Doing this also helps confirming that the issue has been solved.
5443
5444 If this option has been enabled in a "defaults" section, it can be disabled
5445 in a specific instance by prepending the "no" keyword before it.
5446
5447 See also : "option accept-invalid-http-request" and "show errors" on the
5448 stats socket.
5449
5450
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005451option allbackups
5452no option allbackups
5453 Use either all backup servers at a time or only the first one
5454 May be used in sections : defaults | frontend | listen | backend
5455 yes | no | yes | yes
5456 Arguments : none
5457
5458 By default, the first operational backup server gets all traffic when normal
5459 servers are all down. Sometimes, it may be preferred to use multiple backups
5460 at once, because one will not be enough. When "option allbackups" is enabled,
5461 the load balancing will be performed among all backup servers when all normal
5462 ones are unavailable. The same load balancing algorithm will be used and the
5463 servers' weights will be respected. Thus, there will not be any priority
5464 order between the backup servers anymore.
5465
5466 This option is mostly used with static server farms dedicated to return a
5467 "sorry" page when an application is completely offline.
5468
5469 If this option has been enabled in a "defaults" section, it can be disabled
5470 in a specific instance by prepending the "no" keyword before it.
5471
5472
5473option checkcache
5474no option checkcache
Godbach7056a352013-12-11 20:01:07 +08005475 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005476 May be used in sections : defaults | frontend | listen | backend
5477 yes | no | yes | yes
5478 Arguments : none
5479
5480 Some high-level frameworks set application cookies everywhere and do not
5481 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005482 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005483 high risk of session crossing or stealing between users traversing the same
5484 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02005485 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005486
5487 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005488 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01005489 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005490 response to check if there's a risk of caching a cookie on a client-side
5491 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01005492 to the client are :
Davor Ocelice9ed2812017-12-25 17:49:28 +01005493 - all those without "Set-Cookie" header;
Willy Tarreauc55ddce2017-12-21 11:41:38 +01005494 - all those with a return code other than 200, 203, 204, 206, 300, 301,
5495 404, 405, 410, 414, 501, provided that the server has not set a
Davor Ocelice9ed2812017-12-25 17:49:28 +01005496 "Cache-control: public" header field;
Willy Tarreau24ea0bc2017-12-21 11:32:55 +01005497 - all those that result from a request using a method other than GET, HEAD,
5498 OPTIONS, TRACE, provided that the server has not set a 'Cache-Control:
Davor Ocelice9ed2812017-12-25 17:49:28 +01005499 public' header field;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005500 - those with a 'Pragma: no-cache' header
5501 - those with a 'Cache-control: private' header
5502 - those with a 'Cache-control: no-store' header
5503 - those with a 'Cache-control: max-age=0' header
5504 - those with a 'Cache-control: s-maxage=0' header
5505 - those with a 'Cache-control: no-cache' header
5506 - those with a 'Cache-control: no-cache="set-cookie"' header
5507 - those with a 'Cache-control: no-cache="set-cookie,' header
5508 (allowing other fields after set-cookie)
5509
5510 If a response doesn't respect these requirements, then it will be blocked
Willy Tarreau198a7442008-01-17 12:05:32 +01005511 just as if it was from an "rspdeny" filter, with an "HTTP 502 bad gateway".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005512 The session state shows "PH--" meaning that the proxy blocked the response
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005513 during headers processing. Additionally, an alert will be sent in the logs so
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005514 that admins are informed that there's something to be fixed.
5515
5516 Due to the high impact on the application, the application should be tested
5517 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01005518 good practice to always activate it during tests, even if it is not used in
Davor Ocelice9ed2812017-12-25 17:49:28 +01005519 production, as it will report potentially dangerous application behaviors.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005520
5521 If this option has been enabled in a "defaults" section, it can be disabled
5522 in a specific instance by prepending the "no" keyword before it.
5523
5524
5525option clitcpka
5526no option clitcpka
5527 Enable or disable the sending of TCP keepalive packets on the client side
5528 May be used in sections : defaults | frontend | listen | backend
5529 yes | yes | yes | no
5530 Arguments : none
5531
5532 When there is a firewall or any session-aware component between a client and
5533 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01005534 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005535 components decides to expire a session which has remained idle for too long.
5536
5537 Enabling socket-level TCP keep-alives makes the system regularly send packets
5538 to the other end of the connection, leaving it active. The delay between
5539 keep-alive probes is controlled by the system only and depends both on the
5540 operating system and its tuning parameters.
5541
5542 It is important to understand that keep-alive packets are neither emitted nor
5543 received at the application level. It is only the network stacks which sees
5544 them. For this reason, even if one side of the proxy already uses keep-alives
5545 to maintain its connection alive, those keep-alive packets will not be
5546 forwarded to the other side of the proxy.
5547
5548 Please note that this has nothing to do with HTTP keep-alive.
5549
5550 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
5551 client side of a connection, which should help when session expirations are
5552 noticed between HAProxy and a client.
5553
5554 If this option has been enabled in a "defaults" section, it can be disabled
5555 in a specific instance by prepending the "no" keyword before it.
5556
5557 See also : "option srvtcpka", "option tcpka"
5558
5559
Willy Tarreau0ba27502007-12-24 16:55:16 +01005560option contstats
5561 Enable continuous traffic statistics updates
5562 May be used in sections : defaults | frontend | listen | backend
5563 yes | yes | yes | no
5564 Arguments : none
5565
5566 By default, counters used for statistics calculation are incremented
5567 only when a session finishes. It works quite well when serving small
5568 objects, but with big ones (for example large images or archives) or
5569 with A/V streaming, a graph generated from haproxy counters looks like
Willy Tarreaudef0d222016-11-08 22:03:00 +01005570 a hedgehog. With this option enabled counters get incremented frequently
5571 along the session, typically every 5 seconds, which is often enough to
5572 produce clean graphs. Recounting touches a hotpath directly so it is not
5573 not enabled by default, as it can cause a lot of wakeups for very large
5574 session counts and cause a small performance drop.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005575
5576
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005577option dontlog-normal
5578no option dontlog-normal
5579 Enable or disable logging of normal, successful connections
5580 May be used in sections : defaults | frontend | listen | backend
5581 yes | yes | yes | no
5582 Arguments : none
5583
5584 There are large sites dealing with several thousand connections per second
5585 and for which logging is a major pain. Some of them are even forced to turn
5586 logs off and cannot debug production issues. Setting this option ensures that
5587 normal connections, those which experience no error, no timeout, no retry nor
5588 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
5589 mode, the response status code is checked and return codes 5xx will still be
5590 logged.
5591
5592 It is strongly discouraged to use this option as most of the time, the key to
5593 complex issues is in the normal logs which will not be logged here. If you
5594 need to separate logs, see the "log-separate-errors" option instead.
5595
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005596 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005597 logging.
5598
5599
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005600option dontlognull
5601no option dontlognull
5602 Enable or disable logging of null connections
5603 May be used in sections : defaults | frontend | listen | backend
5604 yes | yes | yes | no
5605 Arguments : none
5606
5607 In certain environments, there are components which will regularly connect to
5608 various systems to ensure that they are still alive. It can be the case from
5609 another load balancer as well as from monitoring systems. By default, even a
5610 simple port probe or scan will produce a log. If those connections pollute
5611 the logs too much, it is possible to enable option "dontlognull" to indicate
5612 that a connection on which no data has been transferred will not be logged,
Willy Tarreau0f228a02015-05-01 15:37:53 +02005613 which typically corresponds to those probes. Note that errors will still be
5614 returned to the client and accounted for in the stats. If this is not what is
5615 desired, option http-ignore-probes can be used instead.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005616
5617 It is generally recommended not to use this option in uncontrolled
Davor Ocelice9ed2812017-12-25 17:49:28 +01005618 environments (e.g. internet), otherwise scans and other malicious activities
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005619 would not be logged.
5620
5621 If this option has been enabled in a "defaults" section, it can be disabled
5622 in a specific instance by prepending the "no" keyword before it.
5623
Willy Tarreau0f228a02015-05-01 15:37:53 +02005624 See also : "log", "http-ignore-probes", "monitor-net", "monitor-uri", and
5625 section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005626
5627
5628option forceclose
5629no option forceclose
5630 Enable or disable active connection closing after response is transferred.
5631 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaua31e5df2009-12-30 01:10:35 +01005632 yes | yes | yes | yes
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005633 Arguments : none
5634
5635 Some HTTP servers do not necessarily close the connections when they receive
5636 the "Connection: close" set by "option httpclose", and if the client does not
5637 close either, then the connection remains open till the timeout expires. This
5638 causes high number of simultaneous connections on the servers and shows high
5639 global session times in the logs.
5640
5641 When this happens, it is possible to use "option forceclose". It will
Willy Tarreau82eeaf22009-12-29 12:09:05 +01005642 actively close the outgoing server channel as soon as the server has finished
Cyril Bonté653dcd62014-02-20 00:13:15 +01005643 to respond and release some resources earlier than with "option httpclose".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005644
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005645 This option may also be combined with "option http-pretend-keepalive", which
5646 will disable sending of the "Connection: close" header, but will still cause
5647 the connection to be closed once the whole response is received.
5648
Cyril Bonté653dcd62014-02-20 00:13:15 +01005649 This option disables and replaces any previous "option httpclose", "option
5650 http-server-close", "option http-keep-alive", or "option http-tunnel".
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005651
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005652 If this option has been enabled in a "defaults" section, it can be disabled
5653 in a specific instance by prepending the "no" keyword before it.
5654
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005655 See also : "option httpclose" and "option http-pretend-keepalive"
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005656
5657
Willy Tarreau87cf5142011-08-19 22:57:24 +02005658option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01005659 Enable insertion of the X-Forwarded-For header to requests sent to servers
5660 May be used in sections : defaults | frontend | listen | backend
5661 yes | yes | yes | yes
5662 Arguments :
5663 <network> is an optional argument used to disable this option for sources
5664 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02005665 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01005666 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005667
5668 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
5669 their client address. This is sometimes annoying when the client's IP address
5670 is expected in server logs. To solve this problem, the well-known HTTP header
5671 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
5672 This header contains a value representing the client's IP address. Since this
5673 header is always appended at the end of the existing header list, the server
5674 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02005675 the server's manual to find how to enable use of this standard header. Note
5676 that only the last occurrence of the header must be used, since it is really
5677 possible that the client has already brought one.
5678
Willy Tarreaud72758d2010-01-12 10:42:19 +01005679 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02005680 the default "X-Forwarded-For". This can be useful where you might already
Davor Ocelice9ed2812017-12-25 17:49:28 +01005681 have a "X-Forwarded-For" header from a different application (e.g. stunnel),
Willy Tarreaud72758d2010-01-12 10:42:19 +01005682 and you need preserve it. Also if your backend server doesn't use the
Davor Ocelice9ed2812017-12-25 17:49:28 +01005683 "X-Forwarded-For" header and requires different one (e.g. Zeus Web Servers
Ross Westaf72a1d2008-08-03 10:51:45 +02005684 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01005685
5686 Sometimes, a same HAProxy instance may be shared between a direct client
5687 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
5688 used to decrypt HTTPS traffic). It is possible to disable the addition of the
5689 header for a known source address or network by adding the "except" keyword
5690 followed by the network address. In this case, any source IP matching the
5691 network will not cause an addition of this header. Most common uses are with
5692 private networks or 127.0.0.1.
5693
Willy Tarreau87cf5142011-08-19 22:57:24 +02005694 Alternatively, the keyword "if-none" states that the header will only be
5695 added if it is not present. This should only be used in perfectly trusted
5696 environment, as this might cause a security issue if headers reaching haproxy
5697 are under the control of the end-user.
5698
Willy Tarreauc27debf2008-01-06 08:57:02 +01005699 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02005700 least one of them uses it, the header will be added. Note that the backend's
5701 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02005702 both are defined. In the case of the "if-none" argument, if at least one of
5703 the frontend or the backend does not specify it, it wants the addition to be
5704 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005705
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005706 Example :
Willy Tarreauc27debf2008-01-06 08:57:02 +01005707 # Public HTTP address also used by stunnel on the same machine
5708 frontend www
5709 mode http
5710 option forwardfor except 127.0.0.1 # stunnel already adds the header
5711
Ross Westaf72a1d2008-08-03 10:51:45 +02005712 # Those servers want the IP Address in X-Client
5713 backend www
5714 mode http
5715 option forwardfor header X-Client
5716
Willy Tarreau87cf5142011-08-19 22:57:24 +02005717 See also : "option httpclose", "option http-server-close",
Willy Tarreau70dffda2014-01-30 03:07:23 +01005718 "option forceclose", "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01005719
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005720
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005721option http-buffer-request
5722no option http-buffer-request
5723 Enable or disable waiting for whole HTTP request body before proceeding
5724 May be used in sections : defaults | frontend | listen | backend
5725 yes | yes | yes | yes
5726 Arguments : none
5727
5728 It is sometimes desirable to wait for the body of an HTTP request before
5729 taking a decision. This is what is being done by "balance url_param" for
5730 example. The first use case is to buffer requests from slow clients before
5731 connecting to the server. Another use case consists in taking the routing
5732 decision based on the request body's contents. This option placed in a
5733 frontend or backend forces the HTTP processing to wait until either the whole
5734 body is received, or the request buffer is full, or the first chunk is
5735 complete in case of chunked encoding. It can have undesired side effects with
Tim Düsterhus4896c442016-11-29 02:15:19 +01005736 some applications abusing HTTP by expecting unbuffered transmissions between
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005737 the frontend and the backend, so this should definitely not be used by
5738 default.
5739
Baptiste Assmanneccdf432015-10-28 13:49:01 +01005740 See also : "option http-no-delay", "timeout http-request"
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005741
5742
Willy Tarreau0f228a02015-05-01 15:37:53 +02005743option http-ignore-probes
5744no option http-ignore-probes
5745 Enable or disable logging of null connections and request timeouts
5746 May be used in sections : defaults | frontend | listen | backend
5747 yes | yes | yes | no
5748 Arguments : none
5749
5750 Recently some browsers started to implement a "pre-connect" feature
5751 consisting in speculatively connecting to some recently visited web sites
5752 just in case the user would like to visit them. This results in many
5753 connections being established to web sites, which end up in 408 Request
5754 Timeout if the timeout strikes first, or 400 Bad Request when the browser
5755 decides to close them first. These ones pollute the log and feed the error
5756 counters. There was already "option dontlognull" but it's insufficient in
5757 this case. Instead, this option does the following things :
5758 - prevent any 400/408 message from being sent to the client if nothing
Davor Ocelice9ed2812017-12-25 17:49:28 +01005759 was received over a connection before it was closed;
5760 - prevent any log from being emitted in this situation;
Willy Tarreau0f228a02015-05-01 15:37:53 +02005761 - prevent any error counter from being incremented
5762
5763 That way the empty connection is silently ignored. Note that it is better
5764 not to use this unless it is clear that it is needed, because it will hide
5765 real problems. The most common reason for not receiving a request and seeing
5766 a 408 is due to an MTU inconsistency between the client and an intermediary
5767 element such as a VPN, which blocks too large packets. These issues are
5768 generally seen with POST requests as well as GET with large cookies. The logs
5769 are often the only way to detect them.
5770
5771 If this option has been enabled in a "defaults" section, it can be disabled
5772 in a specific instance by prepending the "no" keyword before it.
5773
5774 See also : "log", "dontlognull", "errorfile", and section 8 about logging.
5775
5776
Willy Tarreau16bfb022010-01-16 19:48:41 +01005777option http-keep-alive
5778no option http-keep-alive
5779 Enable or disable HTTP keep-alive from client to server
5780 May be used in sections : defaults | frontend | listen | backend
5781 yes | yes | yes | yes
5782 Arguments : none
5783
Willy Tarreau70dffda2014-01-30 03:07:23 +01005784 By default HAProxy operates in keep-alive mode with regards to persistent
5785 connections: for each connection it processes each request and response, and
5786 leaves the connection idle on both sides between the end of a response and the
5787 start of a new request. This mode may be changed by several options such as
5788 "option http-server-close", "option forceclose", "option httpclose" or
5789 "option http-tunnel". This option allows to set back the keep-alive mode,
5790 which can be useful when another mode was used in a defaults section.
5791
5792 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
5793 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01005794 network) and the fastest session reuse on the server side at the expense
5795 of maintaining idle connections to the servers. In general, it is possible
5796 with this option to achieve approximately twice the request rate that the
5797 "http-server-close" option achieves on small objects. There are mainly two
5798 situations where this option may be useful :
5799
5800 - when the server is non-HTTP compliant and authenticates the connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01005801 instead of requests (e.g. NTLM authentication)
Willy Tarreau16bfb022010-01-16 19:48:41 +01005802
5803 - when the cost of establishing the connection to the server is significant
5804 compared to the cost of retrieving the associated object from the server.
5805
5806 This last case can happen when the server is a fast static server of cache.
5807 In this case, the server will need to be properly tuned to support high enough
5808 connection counts because connections will last until the client sends another
5809 request.
5810
5811 If the client request has to go to another backend or another server due to
5812 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01005813 immediately be closed and a new one re-opened. Option "prefer-last-server" is
5814 available to try optimize server selection so that if the server currently
5815 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01005816
Willy Tarreau16bfb022010-01-16 19:48:41 +01005817 At the moment, logs will not indicate whether requests came from the same
5818 session or not. The accept date reported in the logs corresponds to the end
5819 of the previous request, and the request time corresponds to the time spent
5820 waiting for a new request. The keep-alive request time is still bound to the
5821 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
5822 not set.
5823
Cyril Bonté653dcd62014-02-20 00:13:15 +01005824 This option disables and replaces any previous "option httpclose", "option
5825 http-server-close", "option forceclose" or "option http-tunnel". When backend
Willy Tarreau70dffda2014-01-30 03:07:23 +01005826 and frontend options differ, all of these 4 options have precedence over
Cyril Bonté653dcd62014-02-20 00:13:15 +01005827 "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01005828
5829 See also : "option forceclose", "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01005830 "option prefer-last-server", "option http-pretend-keepalive",
5831 "option httpclose", and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01005832
5833
Willy Tarreau96e31212011-05-30 18:10:30 +02005834option http-no-delay
5835no option http-no-delay
5836 Instruct the system to favor low interactive delays over performance in HTTP
5837 May be used in sections : defaults | frontend | listen | backend
5838 yes | yes | yes | yes
5839 Arguments : none
5840
5841 In HTTP, each payload is unidirectional and has no notion of interactivity.
5842 Any agent is expected to queue data somewhat for a reasonably low delay.
5843 There are some very rare server-to-server applications that abuse the HTTP
5844 protocol and expect the payload phase to be highly interactive, with many
5845 interleaved data chunks in both directions within a single request. This is
5846 absolutely not supported by the HTTP specification and will not work across
5847 most proxies or servers. When such applications attempt to do this through
5848 haproxy, it works but they will experience high delays due to the network
5849 optimizations which favor performance by instructing the system to wait for
5850 enough data to be available in order to only send full packets. Typical
5851 delays are around 200 ms per round trip. Note that this only happens with
5852 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
5853 affected.
5854
5855 When "option http-no-delay" is present in either the frontend or the backend
5856 used by a connection, all such optimizations will be disabled in order to
5857 make the exchanges as fast as possible. Of course this offers no guarantee on
5858 the functionality, as it may break at any other place. But if it works via
5859 HAProxy, it will work as fast as possible. This option should never be used
5860 by default, and should never be used at all unless such a buggy application
5861 is discovered. The impact of using this option is an increase of bandwidth
5862 usage and CPU usage, which may significantly lower performance in high
5863 latency environments.
5864
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005865 See also : "option http-buffer-request"
5866
Willy Tarreau96e31212011-05-30 18:10:30 +02005867
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005868option http-pretend-keepalive
5869no option http-pretend-keepalive
5870 Define whether haproxy will announce keepalive to the server or not
5871 May be used in sections : defaults | frontend | listen | backend
5872 yes | yes | yes | yes
5873 Arguments : none
5874
5875 When running with "option http-server-close" or "option forceclose", haproxy
5876 adds a "Connection: close" header to the request forwarded to the server.
5877 Unfortunately, when some servers see this header, they automatically refrain
5878 from using the chunked encoding for responses of unknown length, while this
5879 is totally unrelated. The immediate effect is that this prevents haproxy from
5880 maintaining the client connection alive. A second effect is that a client or
5881 a cache could receive an incomplete response without being aware of it, and
5882 consider the response complete.
5883
5884 By setting "option http-pretend-keepalive", haproxy will make the server
5885 believe it will keep the connection alive. The server will then not fall back
5886 to the abnormal undesired above. When haproxy gets the whole response, it
5887 will close the connection with the server just as it would do with the
5888 "forceclose" option. That way the client gets a normal response and the
5889 connection is correctly closed on the server side.
5890
5891 It is recommended not to enable this option by default, because most servers
5892 will more efficiently close the connection themselves after the last packet,
5893 and release its buffers slightly earlier. Also, the added packet on the
5894 network could slightly reduce the overall peak performance. However it is
5895 worth noting that when this option is enabled, haproxy will have slightly
5896 less work to do. So if haproxy is the bottleneck on the whole architecture,
5897 enabling this option might save a few CPU cycles.
5898
5899 This option may be set both in a frontend and in a backend. It is enabled if
5900 at least one of the frontend or backend holding a connection has it enabled.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005901 This option may be combined with "option httpclose", which will cause
Willy Tarreau22a95342010-09-29 14:31:41 +02005902 keepalive to be announced to the server and close to be announced to the
5903 client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005904
5905 If this option has been enabled in a "defaults" section, it can be disabled
5906 in a specific instance by prepending the "no" keyword before it.
5907
Willy Tarreau16bfb022010-01-16 19:48:41 +01005908 See also : "option forceclose", "option http-server-close", and
5909 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005910
Willy Tarreauc27debf2008-01-06 08:57:02 +01005911
Willy Tarreaub608feb2010-01-02 22:47:18 +01005912option http-server-close
5913no option http-server-close
5914 Enable or disable HTTP connection closing on the server side
5915 May be used in sections : defaults | frontend | listen | backend
5916 yes | yes | yes | yes
5917 Arguments : none
5918
Willy Tarreau70dffda2014-01-30 03:07:23 +01005919 By default HAProxy operates in keep-alive mode with regards to persistent
5920 connections: for each connection it processes each request and response, and
5921 leaves the connection idle on both sides between the end of a response and
5922 the start of a new request. This mode may be changed by several options such
5923 as "option http-server-close", "option forceclose", "option httpclose" or
5924 "option http-tunnel". Setting "option http-server-close" enables HTTP
5925 connection-close mode on the server side while keeping the ability to support
Davor Ocelice9ed2812017-12-25 17:49:28 +01005926 HTTP keep-alive and pipelining on the client side. This provides the lowest
Willy Tarreau70dffda2014-01-30 03:07:23 +01005927 latency on the client side (slow network) and the fastest session reuse on
5928 the server side to save server resources, similarly to "option forceclose".
5929 It also permits non-keepalive capable servers to be served in keep-alive mode
Lukas Tribus23953682017-04-28 13:24:30 +00005930 to the clients if they conform to the requirements of RFC7230. Please note
Willy Tarreau70dffda2014-01-30 03:07:23 +01005931 that some servers do not always conform to those requirements when they see
5932 "Connection: close" in the request. The effect will be that keep-alive will
5933 never be used. A workaround consists in enabling "option
5934 http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01005935
5936 At the moment, logs will not indicate whether requests came from the same
5937 session or not. The accept date reported in the logs corresponds to the end
5938 of the previous request, and the request time corresponds to the time spent
5939 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01005940 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
5941 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01005942
5943 This option may be set both in a frontend and in a backend. It is enabled if
5944 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01005945 It disables and replaces any previous "option httpclose", "option forceclose",
5946 "option http-tunnel" or "option http-keep-alive". Please check section 4
Willy Tarreau70dffda2014-01-30 03:07:23 +01005947 ("Proxies") to see how this option combines with others when frontend and
5948 backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01005949
5950 If this option has been enabled in a "defaults" section, it can be disabled
5951 in a specific instance by prepending the "no" keyword before it.
5952
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02005953 See also : "option forceclose", "option http-pretend-keepalive",
Willy Tarreau16bfb022010-01-16 19:48:41 +01005954 "option httpclose", "option http-keep-alive", and
5955 "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01005956
5957
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005958option http-tunnel
5959no option http-tunnel
5960 Disable or enable HTTP connection processing after first transaction
5961 May be used in sections : defaults | frontend | listen | backend
5962 yes | yes | yes | yes
5963 Arguments : none
5964
Willy Tarreau70dffda2014-01-30 03:07:23 +01005965 By default HAProxy operates in keep-alive mode with regards to persistent
5966 connections: for each connection it processes each request and response, and
5967 leaves the connection idle on both sides between the end of a response and
5968 the start of a new request. This mode may be changed by several options such
5969 as "option http-server-close", "option forceclose", "option httpclose" or
5970 "option http-tunnel".
5971
5972 Option "http-tunnel" disables any HTTP processing past the first request and
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005973 the first response. This is the mode which was used by default in versions
Willy Tarreau70dffda2014-01-30 03:07:23 +01005974 1.0 to 1.5-dev21. It is the mode with the lowest processing overhead, which
5975 is normally not needed anymore unless in very specific cases such as when
5976 using an in-house protocol that looks like HTTP but is not compatible, or
5977 just to log one request per client in order to reduce log size. Note that
5978 everything which works at the HTTP level, including header parsing/addition,
5979 cookie processing or content switching will only work for the first request
5980 and will be ignored after the first response.
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005981
5982 If this option has been enabled in a "defaults" section, it can be disabled
5983 in a specific instance by prepending the "no" keyword before it.
5984
5985 See also : "option forceclose", "option http-server-close",
5986 "option httpclose", "option http-keep-alive", and
5987 "1.1. The HTTP transaction model".
5988
5989
Willy Tarreau88d349d2010-01-25 12:15:43 +01005990option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01005991no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01005992 Make use of non-standard Proxy-Connection header instead of Connection
5993 May be used in sections : defaults | frontend | listen | backend
5994 yes | yes | yes | no
5995 Arguments : none
5996
Lukas Tribus23953682017-04-28 13:24:30 +00005997 While RFC7230 explicitly states that HTTP/1.1 agents must use the
Willy Tarreau88d349d2010-01-25 12:15:43 +01005998 Connection header to indicate their wish of persistent or non-persistent
5999 connections, both browsers and proxies ignore this header for proxied
6000 connections and make use of the undocumented, non-standard Proxy-Connection
6001 header instead. The issue begins when trying to put a load balancer between
6002 browsers and such proxies, because there will be a difference between what
6003 haproxy understands and what the client and the proxy agree on.
6004
6005 By setting this option in a frontend, haproxy can automatically switch to use
6006 that non-standard header if it sees proxied requests. A proxied request is
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01006007 defined here as one where the URI begins with neither a '/' nor a '*'. This
6008 is incompatible with the HTTP tunnel mode. Note that this option can only be
6009 specified in a frontend and will affect the request along its whole life.
Willy Tarreau88d349d2010-01-25 12:15:43 +01006010
Willy Tarreau844a7e72010-01-31 21:46:18 +01006011 Also, when this option is set, a request which requires authentication will
6012 automatically switch to use proxy authentication headers if it is itself a
6013 proxied request. That makes it possible to check or enforce authentication in
6014 front of an existing proxy.
6015
Willy Tarreau88d349d2010-01-25 12:15:43 +01006016 This option should normally never be used, except in front of a proxy.
6017
6018 See also : "option httpclose", "option forceclose" and "option
6019 http-server-close".
6020
6021
Willy Tarreaud63335a2010-02-26 12:56:52 +01006022option httpchk
6023option httpchk <uri>
6024option httpchk <method> <uri>
6025option httpchk <method> <uri> <version>
6026 Enable HTTP protocol to check on the servers health
6027 May be used in sections : defaults | frontend | listen | backend
6028 yes | no | yes | yes
6029 Arguments :
6030 <method> is the optional HTTP method used with the requests. When not set,
6031 the "OPTIONS" method is used, as it generally requires low server
6032 processing and is easy to filter out from the logs. Any method
6033 may be used, though it is not recommended to invent non-standard
6034 ones.
6035
6036 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
6037 which is accessible by default on almost any server, but may be
6038 changed to any other URI. Query strings are permitted.
6039
6040 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
6041 but some servers might behave incorrectly in HTTP 1.0, so turning
6042 it to HTTP/1.1 may sometimes help. Note that the Host field is
6043 mandatory in HTTP/1.1, and as a trick, it is possible to pass it
6044 after "\r\n" following the version string.
6045
6046 By default, server health checks only consist in trying to establish a TCP
6047 connection. When "option httpchk" is specified, a complete HTTP request is
6048 sent once the TCP connection is established, and responses 2xx and 3xx are
6049 considered valid, while all other ones indicate a server failure, including
6050 the lack of any response.
6051
6052 The port and interval are specified in the server configuration.
6053
6054 This option does not necessarily require an HTTP backend, it also works with
6055 plain TCP backends. This is particularly useful to check simple scripts bound
6056 to some dedicated ports using the inetd daemon.
6057
6058 Examples :
6059 # Relay HTTPS traffic to Apache instance and check service availability
6060 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
6061 backend https_relay
6062 mode tcp
6063 option httpchk OPTIONS * HTTP/1.1\r\nHost:\ www
6064 server apache1 192.168.1.1:443 check port 80
6065
Simon Hormanafc47ee2013-11-25 10:46:35 +09006066 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
6067 "option pgsql-check", "http-check" and the "check", "port" and
6068 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01006069
6070
Willy Tarreauc27debf2008-01-06 08:57:02 +01006071option httpclose
6072no option httpclose
6073 Enable or disable passive HTTP connection closing
6074 May be used in sections : defaults | frontend | listen | backend
6075 yes | yes | yes | yes
6076 Arguments : none
6077
Willy Tarreau70dffda2014-01-30 03:07:23 +01006078 By default HAProxy operates in keep-alive mode with regards to persistent
6079 connections: for each connection it processes each request and response, and
6080 leaves the connection idle on both sides between the end of a response and
6081 the start of a new request. This mode may be changed by several options such
Cyril Bonté653dcd62014-02-20 00:13:15 +01006082 as "option http-server-close", "option forceclose", "option httpclose" or
Willy Tarreau70dffda2014-01-30 03:07:23 +01006083 "option http-tunnel".
6084
6085 If "option httpclose" is set, HAProxy will work in HTTP tunnel mode and check
6086 if a "Connection: close" header is already set in each direction, and will
6087 add one if missing. Each end should react to this by actively closing the TCP
6088 connection after each transfer, thus resulting in a switch to the HTTP close
6089 mode. Any "Connection" header different from "close" will also be removed.
6090 Note that this option is deprecated since what it does is very cheap but not
6091 reliable. Using "option http-server-close" or "option forceclose" is strongly
6092 recommended instead.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006093
6094 It seldom happens that some servers incorrectly ignore this header and do not
Jamie Gloudonaaa21002012-08-25 00:18:33 -04006095 close the connection even though they reply "Connection: close". For this
Willy Tarreau0dfdf192010-01-05 11:33:11 +01006096 reason, they are not compatible with older HTTP 1.0 browsers. If this happens
6097 it is possible to use the "option forceclose" which actively closes the
6098 request connection once the server responds. Option "forceclose" also
6099 releases the server connection earlier because it does not have to wait for
6100 the client to acknowledge it.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006101
6102 This option may be set both in a frontend and in a backend. It is enabled if
6103 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01006104 It disables and replaces any previous "option http-server-close",
6105 "option forceclose", "option http-keep-alive" or "option http-tunnel". Please
Willy Tarreau70dffda2014-01-30 03:07:23 +01006106 check section 4 ("Proxies") to see how this option combines with others when
6107 frontend and backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006108
6109 If this option has been enabled in a "defaults" section, it can be disabled
6110 in a specific instance by prepending the "no" keyword before it.
6111
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02006112 See also : "option forceclose", "option http-server-close" and
6113 "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01006114
6115
Emeric Brun3a058f32009-06-30 18:26:00 +02006116option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01006117 Enable logging of HTTP request, session state and timers
6118 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01006119 yes | yes | yes | no
Emeric Brun3a058f32009-06-30 18:26:00 +02006120 Arguments :
6121 clf if the "clf" argument is added, then the output format will be
6122 the CLF format instead of HAProxy's default HTTP format. You can
6123 use this when you need to feed HAProxy's logs through a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01006124 log analyzer which only support the CLF format and which is not
Emeric Brun3a058f32009-06-30 18:26:00 +02006125 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006126
6127 By default, the log output format is very poor, as it only contains the
6128 source and destination addresses, and the instance name. By specifying
6129 "option httplog", each log line turns into a much richer format including,
6130 but not limited to, the HTTP request, the connection timers, the session
6131 status, the connections numbers, the captured headers and cookies, the
6132 frontend, backend and server name, and of course the source address and
6133 ports.
6134
PiBa-NLbd556bf2014-12-11 21:31:54 +01006135 Specifying only "option httplog" will automatically clear the 'clf' mode
6136 if it was set by default.
Emeric Brun3a058f32009-06-30 18:26:00 +02006137
Guillaume de Lafond29f45602017-03-31 19:52:15 +02006138 "option httplog" overrides any previous "log-format" directive.
6139
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006140 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006141
Willy Tarreau55165fe2009-05-10 12:02:55 +02006142
6143option http_proxy
6144no option http_proxy
6145 Enable or disable plain HTTP proxy mode
6146 May be used in sections : defaults | frontend | listen | backend
6147 yes | yes | yes | yes
6148 Arguments : none
6149
6150 It sometimes happens that people need a pure HTTP proxy which understands
6151 basic proxy requests without caching nor any fancy feature. In this case,
6152 it may be worth setting up an HAProxy instance with the "option http_proxy"
6153 set. In this mode, no server is declared, and the connection is forwarded to
6154 the IP address and port found in the URL after the "http://" scheme.
6155
6156 No host address resolution is performed, so this only works when pure IP
6157 addresses are passed. Since this option's usage perimeter is rather limited,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01006158 it will probably be used only by experts who know they need exactly it. This
6159 is incompatible with the HTTP tunnel mode.
Willy Tarreau55165fe2009-05-10 12:02:55 +02006160
6161 If this option has been enabled in a "defaults" section, it can be disabled
6162 in a specific instance by prepending the "no" keyword before it.
6163
6164 Example :
6165 # this backend understands HTTP proxy requests and forwards them directly.
6166 backend direct_forward
6167 option httpclose
6168 option http_proxy
6169
6170 See also : "option httpclose"
6171
Willy Tarreau211ad242009-10-03 21:45:07 +02006172
Jamie Gloudon801a0a32012-08-25 00:18:33 -04006173option independent-streams
6174no option independent-streams
6175 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006176 May be used in sections : defaults | frontend | listen | backend
6177 yes | yes | yes | yes
6178 Arguments : none
6179
6180 By default, when data is sent over a socket, both the write timeout and the
6181 read timeout for that socket are refreshed, because we consider that there is
6182 activity on that socket, and we have no other means of guessing if we should
6183 receive data or not.
6184
Davor Ocelice9ed2812017-12-25 17:49:28 +01006185 While this default behavior is desirable for almost all applications, there
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006186 exists a situation where it is desirable to disable it, and only refresh the
6187 read timeout if there are incoming data. This happens on sessions with large
6188 timeouts and low amounts of exchanged data such as telnet session. If the
6189 server suddenly disappears, the output data accumulates in the system's
6190 socket buffers, both timeouts are correctly refreshed, and there is no way
6191 to know the server does not receive them, so we don't timeout. However, when
6192 the underlying protocol always echoes sent data, it would be enough by itself
6193 to detect the issue using the read timeout. Note that this problem does not
6194 happen with more verbose protocols because data won't accumulate long in the
6195 socket buffers.
6196
6197 When this option is set on the frontend, it will disable read timeout updates
6198 on data sent to the client. There probably is little use of this case. When
6199 the option is set on the backend, it will disable read timeout updates on
6200 data sent to the server. Doing so will typically break large HTTP posts from
6201 slow lines, so use it with caution.
6202
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006203 Note: older versions used to call this setting "option independent-streams"
Jamie Gloudon801a0a32012-08-25 00:18:33 -04006204 with a spelling mistake. This spelling is still supported but
6205 deprecated.
6206
Willy Tarreauce887fd2012-05-12 12:50:00 +02006207 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006208
6209
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02006210option ldap-check
6211 Use LDAPv3 health checks for server testing
6212 May be used in sections : defaults | frontend | listen | backend
6213 yes | no | yes | yes
6214 Arguments : none
6215
6216 It is possible to test that the server correctly talks LDAPv3 instead of just
6217 testing that it accepts the TCP connection. When this option is set, an
6218 LDAPv3 anonymous simple bind message is sent to the server, and the response
6219 is analyzed to find an LDAPv3 bind response message.
6220
6221 The server is considered valid only when the LDAP response contains success
6222 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
6223
6224 Logging of bind requests is server dependent see your documentation how to
6225 configure it.
6226
6227 Example :
6228 option ldap-check
6229
6230 See also : "option httpchk"
6231
6232
Simon Horman98637e52014-06-20 12:30:16 +09006233option external-check
6234 Use external processes for server health checks
6235 May be used in sections : defaults | frontend | listen | backend
6236 yes | no | yes | yes
6237
6238 It is possible to test the health of a server using an external command.
6239 This is achieved by running the executable set using "external-check
6240 command".
6241
6242 Requires the "external-check" global to be set.
6243
6244 See also : "external-check", "external-check command", "external-check path"
6245
6246
Willy Tarreau211ad242009-10-03 21:45:07 +02006247option log-health-checks
6248no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02006249 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02006250 May be used in sections : defaults | frontend | listen | backend
6251 yes | no | yes | yes
6252 Arguments : none
6253
Willy Tarreaubef1b322014-05-13 21:01:39 +02006254 By default, failed health check are logged if server is UP and successful
6255 health checks are logged if server is DOWN, so the amount of additional
6256 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02006257
Willy Tarreaubef1b322014-05-13 21:01:39 +02006258 When this option is enabled, any change of the health check status or to
6259 the server's health will be logged, so that it becomes possible to know
6260 that a server was failing occasional checks before crashing, or exactly when
6261 it failed to respond a valid HTTP status, then when the port started to
6262 reject connections, then when the server stopped responding at all.
6263
Davor Ocelice9ed2812017-12-25 17:49:28 +01006264 Note that status changes not caused by health checks (e.g. enable/disable on
Willy Tarreaubef1b322014-05-13 21:01:39 +02006265 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02006266
Willy Tarreaubef1b322014-05-13 21:01:39 +02006267 See also: "option httpchk", "option ldap-check", "option mysql-check",
6268 "option pgsql-check", "option redis-check", "option smtpchk",
6269 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02006270
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02006271
6272option log-separate-errors
6273no option log-separate-errors
6274 Change log level for non-completely successful connections
6275 May be used in sections : defaults | frontend | listen | backend
6276 yes | yes | yes | no
6277 Arguments : none
6278
6279 Sometimes looking for errors in logs is not easy. This option makes haproxy
6280 raise the level of logs containing potentially interesting information such
6281 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
6282 level changes from "info" to "err". This makes it possible to log them
6283 separately to a different file with most syslog daemons. Be careful not to
6284 remove them from the original file, otherwise you would lose ordering which
6285 provides very important information.
6286
6287 Using this option, large sites dealing with several thousand connections per
6288 second may log normal traffic to a rotating buffer and only archive smaller
6289 error logs.
6290
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006291 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02006292 logging.
6293
Willy Tarreauc27debf2008-01-06 08:57:02 +01006294
6295option logasap
6296no option logasap
6297 Enable or disable early logging of HTTP requests
6298 May be used in sections : defaults | frontend | listen | backend
6299 yes | yes | yes | no
6300 Arguments : none
6301
6302 By default, HTTP requests are logged upon termination so that the total
6303 transfer time and the number of bytes appear in the logs. When large objects
6304 are being transferred, it may take a while before the request appears in the
6305 logs. Using "option logasap", the request gets logged as soon as the server
6306 sends the complete headers. The only missing information in the logs will be
6307 the total number of bytes which will indicate everything except the amount
6308 of data transferred, and the total time which will not take the transfer
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01006309 time into account. In such a situation, it's a good practice to capture the
Willy Tarreauc27debf2008-01-06 08:57:02 +01006310 "Content-Length" response header so that the logs at least indicate how many
6311 bytes are expected to be transferred.
6312
Willy Tarreaucc6c8912009-02-22 10:53:55 +01006313 Examples :
6314 listen http_proxy 0.0.0.0:80
6315 mode http
6316 option httplog
6317 option logasap
6318 log 192.168.2.200 local3
6319
6320 >>> Feb 6 12:14:14 localhost \
6321 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
6322 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
6323 "GET /image.iso HTTP/1.0"
6324
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006325 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01006326 logging.
6327
6328
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02006329option mysql-check [ user <username> [ post-41 ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006330 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006331 May be used in sections : defaults | frontend | listen | backend
6332 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006333 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02006334 <username> This is the username which will be used when connecting to MySQL
6335 server.
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02006336 post-41 Send post v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006337
6338 If you specify a username, the check consists of sending two MySQL packet,
6339 one Client Authentication packet, and one QUIT packet, to correctly close
Davor Ocelice9ed2812017-12-25 17:49:28 +01006340 MySQL session. We then parse the MySQL Handshake Initialization packet and/or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006341 Error packet. It is a basic but useful test which does not produce error nor
6342 aborted connect on the server. However, it requires adding an authorization
6343 in the MySQL table, like this :
6344
6345 USE mysql;
6346 INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');
6347 FLUSH PRIVILEGES;
6348
6349 If you don't specify a username (it is deprecated and not recommended), the
Davor Ocelice9ed2812017-12-25 17:49:28 +01006350 check only consists in parsing the Mysql Handshake Initialization packet or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006351 Error packet, we don't send anything in this mode. It was reported that it
6352 can generate lockout if check is too frequent and/or if there is not enough
6353 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
6354 value as if a connection is established successfully within fewer than MySQL
6355 "max_connect_errors" attempts after a previous connection was interrupted,
6356 the error count for the host is cleared to zero. If HAProxy's server get
6357 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
6358
6359 Remember that this does not check database presence nor database consistency.
6360 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006361
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02006362 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006363
6364 Most often, an incoming MySQL server needs to see the client's IP address for
6365 various purposes, including IP privilege matching and connection logging.
6366 When possible, it is often wise to masquerade the client's IP address when
6367 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02006368 which requires the transparent proxy feature to be compiled in, and the MySQL
6369 server to route the client via the machine hosting haproxy.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006370
6371 See also: "option httpchk"
6372
6373
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006374option nolinger
6375no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006376 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006377 May be used in sections: defaults | frontend | listen | backend
6378 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006379 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006380
Davor Ocelice9ed2812017-12-25 17:49:28 +01006381 When clients or servers abort connections in a dirty way (e.g. they are
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006382 physically disconnected), the session timeouts triggers and the session is
6383 closed. But it will remain in FIN_WAIT1 state for some time in the system,
6384 using some resources and possibly limiting the ability to establish newer
6385 connections.
6386
6387 When this happens, it is possible to activate "option nolinger" which forces
6388 the system to immediately remove any socket's pending data on close. Thus,
6389 the session is instantly purged from the system's tables. This usually has
6390 side effects such as increased number of TCP resets due to old retransmits
6391 getting immediately rejected. Some firewalls may sometimes complain about
6392 this too.
6393
6394 For this reason, it is not recommended to use this option when not absolutely
6395 needed. You know that you need it when you have thousands of FIN_WAIT1
6396 sessions on your system (TIME_WAIT ones do not count).
6397
6398 This option may be used both on frontends and backends, depending on the side
6399 where it is required. Use it on the frontend for clients, and on the backend
6400 for servers.
6401
6402 If this option has been enabled in a "defaults" section, it can be disabled
6403 in a specific instance by prepending the "no" keyword before it.
6404
6405
Willy Tarreau55165fe2009-05-10 12:02:55 +02006406option originalto [ except <network> ] [ header <name> ]
6407 Enable insertion of the X-Original-To header to requests sent to servers
6408 May be used in sections : defaults | frontend | listen | backend
6409 yes | yes | yes | yes
6410 Arguments :
6411 <network> is an optional argument used to disable this option for sources
6412 matching <network>
6413 <name> an optional argument to specify a different "X-Original-To"
6414 header name.
6415
6416 Since HAProxy can work in transparent mode, every request from a client can
6417 be redirected to the proxy and HAProxy itself can proxy every request to a
6418 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
6419 be lost. This is annoying when you want access rules based on destination ip
6420 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
6421 added by HAProxy to all requests sent to the server. This header contains a
6422 value representing the original destination IP address. Since this must be
6423 configured to always use the last occurrence of this header only. Note that
6424 only the last occurrence of the header must be used, since it is really
6425 possible that the client has already brought one.
6426
6427 The keyword "header" may be used to supply a different header name to replace
6428 the default "X-Original-To". This can be useful where you might already
6429 have a "X-Original-To" header from a different application, and you need
6430 preserve it. Also if your backend server doesn't use the "X-Original-To"
6431 header and requires different one.
6432
6433 Sometimes, a same HAProxy instance may be shared between a direct client
6434 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
6435 used to decrypt HTTPS traffic). It is possible to disable the addition of the
6436 header for a known source address or network by adding the "except" keyword
6437 followed by the network address. In this case, any source IP matching the
6438 network will not cause an addition of this header. Most common uses are with
6439 private networks or 127.0.0.1.
6440
6441 This option may be specified either in the frontend or in the backend. If at
6442 least one of them uses it, the header will be added. Note that the backend's
6443 setting of the header subargument takes precedence over the frontend's if
6444 both are defined.
6445
Willy Tarreau55165fe2009-05-10 12:02:55 +02006446 Examples :
6447 # Original Destination address
6448 frontend www
6449 mode http
6450 option originalto except 127.0.0.1
6451
6452 # Those servers want the IP Address in X-Client-Dst
6453 backend www
6454 mode http
6455 option originalto header X-Client-Dst
6456
Willy Tarreau87cf5142011-08-19 22:57:24 +02006457 See also : "option httpclose", "option http-server-close",
6458 "option forceclose"
Willy Tarreau55165fe2009-05-10 12:02:55 +02006459
6460
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006461option persist
6462no option persist
6463 Enable or disable forced persistence on down servers
6464 May be used in sections: defaults | frontend | listen | backend
6465 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006466 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006467
6468 When an HTTP request reaches a backend with a cookie which references a dead
6469 server, by default it is redispatched to another server. It is possible to
6470 force the request to be sent to the dead server first using "option persist"
6471 if absolutely needed. A common use case is when servers are under extreme
6472 load and spend their time flapping. In this case, the users would still be
6473 directed to the server they opened the session on, in the hope they would be
6474 correctly served. It is recommended to use "option redispatch" in conjunction
6475 with this option so that in the event it would not be possible to connect to
6476 the server at all (server definitely dead), the client would finally be
6477 redirected to another valid server.
6478
6479 If this option has been enabled in a "defaults" section, it can be disabled
6480 in a specific instance by prepending the "no" keyword before it.
6481
Willy Tarreau4de91492010-01-22 19:10:05 +01006482 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006483
6484
Willy Tarreau0c122822013-12-15 18:49:01 +01006485option pgsql-check [ user <username> ]
6486 Use PostgreSQL health checks for server testing
6487 May be used in sections : defaults | frontend | listen | backend
6488 yes | no | yes | yes
6489 Arguments :
6490 <username> This is the username which will be used when connecting to
6491 PostgreSQL server.
6492
6493 The check sends a PostgreSQL StartupMessage and waits for either
6494 Authentication request or ErrorResponse message. It is a basic but useful
6495 test which does not produce error nor aborted connect on the server.
6496 This check is identical with the "mysql-check".
6497
6498 See also: "option httpchk"
6499
6500
Willy Tarreau9420b122013-12-15 18:58:25 +01006501option prefer-last-server
6502no option prefer-last-server
6503 Allow multiple load balanced requests to remain on the same server
6504 May be used in sections: defaults | frontend | listen | backend
6505 yes | no | yes | yes
6506 Arguments : none
6507
6508 When the load balancing algorithm in use is not deterministic, and a previous
6509 request was sent to a server to which haproxy still holds a connection, it is
6510 sometimes desirable that subsequent requests on a same session go to the same
6511 server as much as possible. Note that this is different from persistence, as
6512 we only indicate a preference which haproxy tries to apply without any form
6513 of warranty. The real use is for keep-alive connections sent to servers. When
6514 this option is used, haproxy will try to reuse the same connection that is
6515 attached to the server instead of rebalancing to another server, causing a
6516 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01006517 not make much sense to use this in combination with hashing algorithms. Note,
6518 haproxy already automatically tries to stick to a server which sends a 401 or
6519 to a proxy which sends a 407 (authentication required). This is mandatory for
6520 use with the broken NTLM authentication challenge, and significantly helps in
6521 troubleshooting some faulty applications. Option prefer-last-server might be
6522 desirable in these environments as well, to avoid redistributing the traffic
6523 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01006524
6525 If this option has been enabled in a "defaults" section, it can be disabled
6526 in a specific instance by prepending the "no" keyword before it.
6527
6528 See also: "option http-keep-alive"
6529
6530
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006531option redispatch
Joseph Lynch726ab712015-05-11 23:25:34 -07006532option redispatch <interval>
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006533no option redispatch
6534 Enable or disable session redistribution in case of connection failure
6535 May be used in sections: defaults | frontend | listen | backend
6536 yes | no | yes | yes
Joseph Lynch726ab712015-05-11 23:25:34 -07006537 Arguments :
6538 <interval> The optional integer value that controls how often redispatches
6539 occur when retrying connections. Positive value P indicates a
6540 redispatch is desired on every Pth retry, and negative value
Davor Ocelice9ed2812017-12-25 17:49:28 +01006541 N indicate a redispatch is desired on the Nth retry prior to the
Joseph Lynch726ab712015-05-11 23:25:34 -07006542 last retry. For example, the default of -1 preserves the
Davor Ocelice9ed2812017-12-25 17:49:28 +01006543 historical behavior of redispatching on the last retry, a
Joseph Lynch726ab712015-05-11 23:25:34 -07006544 positive value of 1 would indicate a redispatch on every retry,
6545 and a positive value of 3 would indicate a redispatch on every
6546 third retry. You can disable redispatches with a value of 0.
6547
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006548
6549 In HTTP mode, if a server designated by a cookie is down, clients may
6550 definitely stick to it because they cannot flush the cookie, so they will not
6551 be able to access the service anymore.
6552
6553 Specifying "option redispatch" will allow the proxy to break their
6554 persistence and redistribute them to a working server.
6555
Joseph Lynch726ab712015-05-11 23:25:34 -07006556 It also allows to retry connections to another server in case of multiple
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006557 connection failures. Of course, it requires having "retries" set to a nonzero
6558 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01006559
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006560 This form is the preferred form, which replaces both the "redispatch" and
6561 "redisp" keywords.
6562
6563 If this option has been enabled in a "defaults" section, it can be disabled
6564 in a specific instance by prepending the "no" keyword before it.
6565
Willy Tarreau4de91492010-01-22 19:10:05 +01006566 See also : "redispatch", "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006567
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006568
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02006569option redis-check
6570 Use redis health checks for server testing
6571 May be used in sections : defaults | frontend | listen | backend
6572 yes | no | yes | yes
6573 Arguments : none
6574
6575 It is possible to test that the server correctly talks REDIS protocol instead
6576 of just testing that it accepts the TCP connection. When this option is set,
6577 a PING redis command is sent to the server, and the response is analyzed to
6578 find the "+PONG" response message.
6579
6580 Example :
6581 option redis-check
6582
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03006583 See also : "option httpchk", "option tcp-check", "tcp-check expect"
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02006584
6585
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006586option smtpchk
6587option smtpchk <hello> <domain>
6588 Use SMTP health checks for server testing
6589 May be used in sections : defaults | frontend | listen | backend
6590 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01006591 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006592 <hello> is an optional argument. It is the "hello" command to use. It can
6593 be either "HELO" (for SMTP) or "EHLO" (for ESTMP). All other
6594 values will be turned into the default command ("HELO").
6595
6596 <domain> is the domain name to present to the server. It may only be
6597 specified (and is mandatory) if the hello command has been
6598 specified. By default, "localhost" is used.
6599
6600 When "option smtpchk" is set, the health checks will consist in TCP
6601 connections followed by an SMTP command. By default, this command is
6602 "HELO localhost". The server's return code is analyzed and only return codes
6603 starting with a "2" will be considered as valid. All other responses,
6604 including a lack of response will constitute an error and will indicate a
6605 dead server.
6606
6607 This test is meant to be used with SMTP servers or relays. Depending on the
6608 request, it is possible that some servers do not log each connection attempt,
Davor Ocelice9ed2812017-12-25 17:49:28 +01006609 so you may want to experiment to improve the behavior. Using telnet on port
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006610 25 is often easier than adjusting the configuration.
6611
6612 Most often, an incoming SMTP server needs to see the client's IP address for
6613 various purposes, including spam filtering, anti-spoofing and logging. When
6614 possible, it is often wise to masquerade the client's IP address when
6615 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02006616 which requires the transparent proxy feature to be compiled in.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006617
6618 Example :
6619 option smtpchk HELO mydomain.org
6620
6621 See also : "option httpchk", "source"
6622
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006623
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02006624option socket-stats
6625no option socket-stats
6626
6627 Enable or disable collecting & providing separate statistics for each socket.
6628 May be used in sections : defaults | frontend | listen | backend
6629 yes | yes | yes | no
6630
6631 Arguments : none
6632
6633
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006634option splice-auto
6635no option splice-auto
6636 Enable or disable automatic kernel acceleration on sockets in both directions
6637 May be used in sections : defaults | frontend | listen | backend
6638 yes | yes | yes | yes
6639 Arguments : none
6640
6641 When this option is enabled either on a frontend or on a backend, haproxy
6642 will automatically evaluate the opportunity to use kernel tcp splicing to
Davor Ocelice9ed2812017-12-25 17:49:28 +01006643 forward data between the client and the server, in either direction. HAProxy
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006644 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006645 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006646 are not much aggressive in order to limit excessive use of splicing. This
6647 option requires splicing to be enabled at compile time, and may be globally
6648 disabled with the global option "nosplice". Since splice uses pipes, using it
6649 requires that there are enough spare pipes.
6650
6651 Important note: kernel-based TCP splicing is a Linux-specific feature which
6652 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
6653 transfer data between sockets without copying these data to user-space, thus
6654 providing noticeable performance gains and CPU cycles savings. Since many
6655 early implementations are buggy, corrupt data and/or are inefficient, this
6656 feature is not enabled by default, and it should be used with extreme care.
6657 While it is not possible to detect the correctness of an implementation,
6658 2.6.29 is the first version offering a properly working implementation. In
6659 case of doubt, splicing may be globally disabled using the global "nosplice"
6660 keyword.
6661
6662 Example :
6663 option splice-auto
6664
6665 If this option has been enabled in a "defaults" section, it can be disabled
6666 in a specific instance by prepending the "no" keyword before it.
6667
6668 See also : "option splice-request", "option splice-response", and global
6669 options "nosplice" and "maxpipes"
6670
6671
6672option splice-request
6673no option splice-request
6674 Enable or disable automatic kernel acceleration on sockets for requests
6675 May be used in sections : defaults | frontend | listen | backend
6676 yes | yes | yes | yes
6677 Arguments : none
6678
6679 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04006680 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006681 the client to the server. It might still use the recv/send scheme if there
6682 are no spare pipes left. This option requires splicing to be enabled at
6683 compile time, and may be globally disabled with the global option "nosplice".
6684 Since splice uses pipes, using it requires that there are enough spare pipes.
6685
6686 Important note: see "option splice-auto" for usage limitations.
6687
6688 Example :
6689 option splice-request
6690
6691 If this option has been enabled in a "defaults" section, it can be disabled
6692 in a specific instance by prepending the "no" keyword before it.
6693
6694 See also : "option splice-auto", "option splice-response", and global options
6695 "nosplice" and "maxpipes"
6696
6697
6698option splice-response
6699no option splice-response
6700 Enable or disable automatic kernel acceleration on sockets for responses
6701 May be used in sections : defaults | frontend | listen | backend
6702 yes | yes | yes | yes
6703 Arguments : none
6704
6705 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04006706 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006707 the server to the client. It might still use the recv/send scheme if there
6708 are no spare pipes left. This option requires splicing to be enabled at
6709 compile time, and may be globally disabled with the global option "nosplice".
6710 Since splice uses pipes, using it requires that there are enough spare pipes.
6711
6712 Important note: see "option splice-auto" for usage limitations.
6713
6714 Example :
6715 option splice-response
6716
6717 If this option has been enabled in a "defaults" section, it can be disabled
6718 in a specific instance by prepending the "no" keyword before it.
6719
6720 See also : "option splice-auto", "option splice-request", and global options
6721 "nosplice" and "maxpipes"
6722
6723
Christopher Fauletba7bc162016-11-07 21:07:38 +01006724option spop-check
6725 Use SPOP health checks for server testing
6726 May be used in sections : defaults | frontend | listen | backend
6727 no | no | no | yes
6728 Arguments : none
6729
6730 It is possible to test that the server correctly talks SPOP protocol instead
6731 of just testing that it accepts the TCP connection. When this option is set,
6732 a HELLO handshake is performed between HAProxy and the server, and the
6733 response is analyzed to check no error is reported.
6734
6735 Example :
6736 option spop-check
6737
6738 See also : "option httpchk"
6739
6740
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006741option srvtcpka
6742no option srvtcpka
6743 Enable or disable the sending of TCP keepalive packets on the server side
6744 May be used in sections : defaults | frontend | listen | backend
6745 yes | no | yes | yes
6746 Arguments : none
6747
6748 When there is a firewall or any session-aware component between a client and
6749 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01006750 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006751 components decides to expire a session which has remained idle for too long.
6752
6753 Enabling socket-level TCP keep-alives makes the system regularly send packets
6754 to the other end of the connection, leaving it active. The delay between
6755 keep-alive probes is controlled by the system only and depends both on the
6756 operating system and its tuning parameters.
6757
6758 It is important to understand that keep-alive packets are neither emitted nor
6759 received at the application level. It is only the network stacks which sees
6760 them. For this reason, even if one side of the proxy already uses keep-alives
6761 to maintain its connection alive, those keep-alive packets will not be
6762 forwarded to the other side of the proxy.
6763
6764 Please note that this has nothing to do with HTTP keep-alive.
6765
6766 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
6767 server side of a connection, which should help when session expirations are
6768 noticed between HAProxy and a server.
6769
6770 If this option has been enabled in a "defaults" section, it can be disabled
6771 in a specific instance by prepending the "no" keyword before it.
6772
6773 See also : "option clitcpka", "option tcpka"
6774
6775
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006776option ssl-hello-chk
6777 Use SSLv3 client hello health checks for server testing
6778 May be used in sections : defaults | frontend | listen | backend
6779 yes | no | yes | yes
6780 Arguments : none
6781
6782 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
6783 possible to test that the server correctly talks SSL instead of just testing
6784 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
6785 SSLv3 client hello messages are sent once the connection is established to
6786 the server, and the response is analyzed to find an SSL server hello message.
6787 The server is considered valid only when the response contains this server
6788 hello message.
6789
6790 All servers tested till there correctly reply to SSLv3 client hello messages,
6791 and most servers tested do not even log the requests containing only hello
6792 messages, which is appreciable.
6793
Willy Tarreau763a95b2012-10-04 23:15:39 +02006794 Note that this check works even when SSL support was not built into haproxy
6795 because it forges the SSL message. When SSL support is available, it is best
6796 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006797
Willy Tarreau763a95b2012-10-04 23:15:39 +02006798 See also: "option httpchk", "check-ssl"
6799
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006800
Willy Tarreaued179852013-12-16 01:07:00 +01006801option tcp-check
6802 Perform health checks using tcp-check send/expect sequences
6803 May be used in sections: defaults | frontend | listen | backend
6804 yes | no | yes | yes
6805
6806 This health check method is intended to be combined with "tcp-check" command
6807 lists in order to support send/expect types of health check sequences.
6808
6809 TCP checks currently support 4 modes of operations :
6810 - no "tcp-check" directive : the health check only consists in a connection
6811 attempt, which remains the default mode.
6812
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006813 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01006814 used to send a string along with a connection opening. With some
6815 protocols, it helps sending a "QUIT" message for example that prevents
6816 the server from logging a connection error for each health check. The
6817 check result will still be based on the ability to open the connection
6818 only.
6819
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006820 - "tcp-check expect" only is mentioned : this is used to test a banner.
Willy Tarreaued179852013-12-16 01:07:00 +01006821 The connection is opened and haproxy waits for the server to present some
6822 contents which must validate some rules. The check result will be based
6823 on the matching between the contents and the rules. This is suited for
6824 POP, IMAP, SMTP, FTP, SSH, TELNET.
6825
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006826 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Davor Ocelice9ed2812017-12-25 17:49:28 +01006827 used to test a hello-type protocol. HAProxy sends a message, the server
6828 responds and its response is analyzed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006829 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01006830 suited for protocols which require a binding or a request/response model.
6831 LDAP, MySQL, Redis and SSL are example of such protocols, though they
6832 already all have their dedicated checks with a deeper understanding of
6833 the respective protocols.
6834 In this mode, many questions may be sent and many answers may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01006835 analyzed.
Willy Tarreaued179852013-12-16 01:07:00 +01006836
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006837 A fifth mode can be used to insert comments in different steps of the
6838 script.
6839
6840 For each tcp-check rule you create, you can add a "comment" directive,
6841 followed by a string. This string will be reported in the log and stderr
6842 in debug mode. It is useful to make user-friendly error reporting.
6843 The "comment" is of course optional.
6844
6845
Willy Tarreaued179852013-12-16 01:07:00 +01006846 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01006847 # perform a POP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01006848 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006849 tcp-check expect string +OK\ POP3\ ready comment POP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01006850
Davor Ocelice9ed2812017-12-25 17:49:28 +01006851 # perform an IMAP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01006852 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006853 tcp-check expect string *\ OK\ IMAP4\ ready comment IMAP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01006854
6855 # look for the redis master server after ensuring it speaks well
6856 # redis protocol, then it exits properly.
Davor Ocelice9ed2812017-12-25 17:49:28 +01006857 # (send a command then analyze the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01006858 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006859 tcp-check comment PING\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01006860 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02006861 tcp-check expect string +PONG
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006862 tcp-check comment role\ check
Willy Tarreaued179852013-12-16 01:07:00 +01006863 tcp-check send info\ replication\r\n
6864 tcp-check expect string role:master
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006865 tcp-check comment QUIT\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01006866 tcp-check send QUIT\r\n
6867 tcp-check expect string +OK
6868
Davor Ocelice9ed2812017-12-25 17:49:28 +01006869 forge a HTTP request, then analyze the response
Willy Tarreaued179852013-12-16 01:07:00 +01006870 (send many headers before analyzing)
6871 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006872 tcp-check comment forge\ and\ send\ HTTP\ request
Willy Tarreaued179852013-12-16 01:07:00 +01006873 tcp-check send HEAD\ /\ HTTP/1.1\r\n
6874 tcp-check send Host:\ www.mydomain.com\r\n
6875 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
6876 tcp-check send \r\n
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006877 tcp-check expect rstring HTTP/1\..\ (2..|3..) comment check\ HTTP\ response
Willy Tarreaued179852013-12-16 01:07:00 +01006878
6879
6880 See also : "tcp-check expect", "tcp-check send"
6881
6882
Willy Tarreau9ea05a72009-06-14 12:07:01 +02006883option tcp-smart-accept
6884no option tcp-smart-accept
6885 Enable or disable the saving of one ACK packet during the accept sequence
6886 May be used in sections : defaults | frontend | listen | backend
6887 yes | yes | yes | no
6888 Arguments : none
6889
6890 When an HTTP connection request comes in, the system acknowledges it on
6891 behalf of HAProxy, then the client immediately sends its request, and the
6892 system acknowledges it too while it is notifying HAProxy about the new
6893 connection. HAProxy then reads the request and responds. This means that we
6894 have one TCP ACK sent by the system for nothing, because the request could
6895 very well be acknowledged by HAProxy when it sends its response.
6896
6897 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
6898 sending this useless ACK on platforms which support it (currently at least
6899 Linux). It must not cause any problem, because the system will send it anyway
6900 after 40 ms if the response takes more time than expected to come.
6901
6902 During complex network debugging sessions, it may be desirable to disable
6903 this optimization because delayed ACKs can make troubleshooting more complex
6904 when trying to identify where packets are delayed. It is then possible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01006905 fall back to normal behavior by specifying "no option tcp-smart-accept".
Willy Tarreau9ea05a72009-06-14 12:07:01 +02006906
6907 It is also possible to force it for non-HTTP proxies by simply specifying
6908 "option tcp-smart-accept". For instance, it can make sense with some services
6909 such as SMTP where the server speaks first.
6910
6911 It is recommended to avoid forcing this option in a defaults section. In case
6912 of doubt, consider setting it back to automatic values by prepending the
6913 "default" keyword before it, or disabling it using the "no" keyword.
6914
Willy Tarreaud88edf22009-06-14 15:48:17 +02006915 See also : "option tcp-smart-connect"
6916
6917
6918option tcp-smart-connect
6919no option tcp-smart-connect
6920 Enable or disable the saving of one ACK packet during the connect sequence
6921 May be used in sections : defaults | frontend | listen | backend
6922 yes | no | yes | yes
6923 Arguments : none
6924
6925 On certain systems (at least Linux), HAProxy can ask the kernel not to
6926 immediately send an empty ACK upon a connection request, but to directly
6927 send the buffer request instead. This saves one packet on the network and
6928 thus boosts performance. It can also be useful for some servers, because they
6929 immediately get the request along with the incoming connection.
6930
6931 This feature is enabled when "option tcp-smart-connect" is set in a backend.
6932 It is not enabled by default because it makes network troubleshooting more
6933 complex.
6934
6935 It only makes sense to enable it with protocols where the client speaks first
6936 such as HTTP. In other situations, if there is no data to send in place of
6937 the ACK, a normal ACK is sent.
6938
6939 If this option has been enabled in a "defaults" section, it can be disabled
6940 in a specific instance by prepending the "no" keyword before it.
6941
6942 See also : "option tcp-smart-accept"
6943
Willy Tarreau9ea05a72009-06-14 12:07:01 +02006944
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006945option tcpka
6946 Enable or disable the sending of TCP keepalive packets on both sides
6947 May be used in sections : defaults | frontend | listen | backend
6948 yes | yes | yes | yes
6949 Arguments : none
6950
6951 When there is a firewall or any session-aware component between a client and
6952 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01006953 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006954 components decides to expire a session which has remained idle for too long.
6955
6956 Enabling socket-level TCP keep-alives makes the system regularly send packets
6957 to the other end of the connection, leaving it active. The delay between
6958 keep-alive probes is controlled by the system only and depends both on the
6959 operating system and its tuning parameters.
6960
6961 It is important to understand that keep-alive packets are neither emitted nor
6962 received at the application level. It is only the network stacks which sees
6963 them. For this reason, even if one side of the proxy already uses keep-alives
6964 to maintain its connection alive, those keep-alive packets will not be
6965 forwarded to the other side of the proxy.
6966
6967 Please note that this has nothing to do with HTTP keep-alive.
6968
6969 Using option "tcpka" enables the emission of TCP keep-alive probes on both
6970 the client and server sides of a connection. Note that this is meaningful
6971 only in "defaults" or "listen" sections. If this option is used in a
6972 frontend, only the client side will get keep-alives, and if this option is
6973 used in a backend, only the server side will get keep-alives. For this
6974 reason, it is strongly recommended to explicitly use "option clitcpka" and
6975 "option srvtcpka" when the configuration is split between frontends and
6976 backends.
6977
6978 See also : "option clitcpka", "option srvtcpka"
6979
Willy Tarreau844e3c52008-01-11 16:28:18 +01006980
6981option tcplog
6982 Enable advanced logging of TCP connections with session state and timers
6983 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01006984 yes | yes | yes | no
Willy Tarreau844e3c52008-01-11 16:28:18 +01006985 Arguments : none
6986
6987 By default, the log output format is very poor, as it only contains the
6988 source and destination addresses, and the instance name. By specifying
6989 "option tcplog", each log line turns into a much richer format including, but
6990 not limited to, the connection timers, the session status, the connections
6991 numbers, the frontend, backend and server name, and of course the source
6992 address and ports. This option is useful for pure TCP proxies in order to
6993 find which of the client or server disconnects or times out. For normal HTTP
6994 proxies, it's better to use "option httplog" which is even more complete.
6995
Guillaume de Lafond29f45602017-03-31 19:52:15 +02006996 "option tcplog" overrides any previous "log-format" directive.
6997
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006998 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01006999
7000
Willy Tarreau844e3c52008-01-11 16:28:18 +01007001option transparent
7002no option transparent
7003 Enable client-side transparent proxying
7004 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01007005 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01007006 Arguments : none
7007
7008 This option was introduced in order to provide layer 7 persistence to layer 3
7009 load balancers. The idea is to use the OS's ability to redirect an incoming
7010 connection for a remote address to a local process (here HAProxy), and let
7011 this process know what address was initially requested. When this option is
7012 used, sessions without cookies will be forwarded to the original destination
7013 IP address of the incoming request (which should match that of another
7014 equipment), while requests with cookies will still be forwarded to the
7015 appropriate server.
7016
7017 Note that contrary to a common belief, this option does NOT make HAProxy
7018 present the client's IP to the server when establishing the connection.
7019
Willy Tarreaua1146052011-03-01 09:51:54 +01007020 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007021 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01007022
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007023
Simon Horman98637e52014-06-20 12:30:16 +09007024external-check command <command>
7025 Executable to run when performing an external-check
7026 May be used in sections : defaults | frontend | listen | backend
7027 yes | no | yes | yes
7028
7029 Arguments :
7030 <command> is the external command to run
7031
Simon Horman98637e52014-06-20 12:30:16 +09007032 The arguments passed to the to the command are:
7033
Cyril Bonté777be862014-12-02 21:21:35 +01007034 <proxy_address> <proxy_port> <server_address> <server_port>
Simon Horman98637e52014-06-20 12:30:16 +09007035
Cyril Bonté777be862014-12-02 21:21:35 +01007036 The <proxy_address> and <proxy_port> are derived from the first listener
7037 that is either IPv4, IPv6 or a UNIX socket. In the case of a UNIX socket
7038 listener the proxy_address will be the path of the socket and the
7039 <proxy_port> will be the string "NOT_USED". In a backend section, it's not
7040 possible to determine a listener, and both <proxy_address> and <proxy_port>
7041 will have the string value "NOT_USED".
Simon Horman98637e52014-06-20 12:30:16 +09007042
Cyril Bonté72cda2a2014-12-27 22:28:39 +01007043 Some values are also provided through environment variables.
7044
7045 Environment variables :
7046 HAPROXY_PROXY_ADDR The first bind address if available (or empty if not
7047 applicable, for example in a "backend" section).
7048
7049 HAPROXY_PROXY_ID The backend id.
7050
7051 HAPROXY_PROXY_NAME The backend name.
7052
7053 HAPROXY_PROXY_PORT The first bind port if available (or empty if not
7054 applicable, for example in a "backend" section or
7055 for a UNIX socket).
7056
7057 HAPROXY_SERVER_ADDR The server address.
7058
7059 HAPROXY_SERVER_CURCONN The current number of connections on the server.
7060
7061 HAPROXY_SERVER_ID The server id.
7062
7063 HAPROXY_SERVER_MAXCONN The server max connections.
7064
7065 HAPROXY_SERVER_NAME The server name.
7066
7067 HAPROXY_SERVER_PORT The server port if available (or empty for a UNIX
7068 socket).
7069
7070 PATH The PATH environment variable used when executing
7071 the command may be set using "external-check path".
7072
Simon Horman98637e52014-06-20 12:30:16 +09007073 If the command executed and exits with a zero status then the check is
7074 considered to have passed, otherwise the check is considered to have
7075 failed.
7076
7077 Example :
7078 external-check command /bin/true
7079
7080 See also : "external-check", "option external-check", "external-check path"
7081
7082
7083external-check path <path>
7084 The value of the PATH environment variable used when running an external-check
7085 May be used in sections : defaults | frontend | listen | backend
7086 yes | no | yes | yes
7087
7088 Arguments :
7089 <path> is the path used when executing external command to run
7090
7091 The default path is "".
7092
7093 Example :
7094 external-check path "/usr/bin:/bin"
7095
7096 See also : "external-check", "option external-check",
7097 "external-check command"
7098
7099
Emeric Brun647caf12009-06-30 17:57:00 +02007100persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02007101persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02007102 Enable RDP cookie-based persistence
7103 May be used in sections : defaults | frontend | listen | backend
7104 yes | no | yes | yes
7105 Arguments :
7106 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02007107 default cookie name "msts" will be used. There currently is no
7108 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02007109
7110 This statement enables persistence based on an RDP cookie. The RDP cookie
7111 contains all information required to find the server in the list of known
Davor Ocelice9ed2812017-12-25 17:49:28 +01007112 servers. So when this option is set in the backend, the request is analyzed
Emeric Brun647caf12009-06-30 17:57:00 +02007113 and if an RDP cookie is found, it is decoded. If it matches a known server
7114 which is still UP (or if "option persist" is set), then the connection is
7115 forwarded to this server.
7116
7117 Note that this only makes sense in a TCP backend, but for this to work, the
7118 frontend must have waited long enough to ensure that an RDP cookie is present
7119 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007120 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02007121 a single "listen" section.
7122
Willy Tarreau61e28f22010-05-16 22:31:05 +02007123 Also, it is important to understand that the terminal server will emit this
7124 RDP cookie only if it is configured for "token redirection mode", which means
7125 that the "IP address redirection" option is disabled.
7126
Emeric Brun647caf12009-06-30 17:57:00 +02007127 Example :
7128 listen tse-farm
7129 bind :3389
7130 # wait up to 5s for an RDP cookie in the request
7131 tcp-request inspect-delay 5s
7132 tcp-request content accept if RDP_COOKIE
7133 # apply RDP cookie persistence
7134 persist rdp-cookie
7135 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02007136 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02007137 balance rdp-cookie
7138 server srv1 1.1.1.1:3389
7139 server srv2 1.1.1.2:3389
7140
Simon Hormanab814e02011-06-24 14:50:20 +09007141 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
7142 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02007143
7144
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007145rate-limit sessions <rate>
7146 Set a limit on the number of new sessions accepted per second on a frontend
7147 May be used in sections : defaults | frontend | listen | backend
7148 yes | yes | yes | no
7149 Arguments :
7150 <rate> The <rate> parameter is an integer designating the maximum number
7151 of new sessions per second to accept on the frontend.
7152
7153 When the frontend reaches the specified number of new sessions per second, it
7154 stops accepting new connections until the rate drops below the limit again.
7155 During this time, the pending sessions will be kept in the socket's backlog
7156 (in system buffers) and haproxy will not even be aware that sessions are
7157 pending. When applying very low limit on a highly loaded service, it may make
7158 sense to increase the socket's backlog using the "backlog" keyword.
7159
7160 This feature is particularly efficient at blocking connection-based attacks
7161 or service abuse on fragile servers. Since the session rate is measured every
7162 millisecond, it is extremely accurate. Also, the limit applies immediately,
7163 no delay is needed at all to detect the threshold.
7164
7165 Example : limit the connection rate on SMTP to 10 per second max
7166 listen smtp
7167 mode tcp
7168 bind :25
7169 rate-limit sessions 10
Panagiotis Panagiotopoulos7282d8e2016-02-11 16:37:15 +02007170 server smtp1 127.0.0.1:1025
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007171
Willy Tarreaua17c2d92011-07-25 08:16:20 +02007172 Note : when the maximum rate is reached, the frontend's status is not changed
7173 but its sockets appear as "WAITING" in the statistics if the
7174 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007175
7176 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
7177
7178
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007179redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
7180redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
7181redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007182 Return an HTTP redirection if/unless a condition is matched
7183 May be used in sections : defaults | frontend | listen | backend
7184 no | yes | yes | yes
7185
7186 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +01007187 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007188
Willy Tarreau0140f252008-11-19 21:07:09 +01007189 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007190 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007191 the HTTP "Location" header. When used in an "http-request" rule,
7192 <loc> value follows the log-format rules and can include some
7193 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007194
7195 <pfx> With "redirect prefix", the "Location" header is built from the
7196 concatenation of <pfx> and the complete URI path, including the
7197 query string, unless the "drop-query" option is specified (see
7198 below). As a special case, if <pfx> equals exactly "/", then
7199 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007200 redirect to the same URL (for instance, to insert a cookie). When
7201 used in an "http-request" rule, <pfx> value follows the log-format
7202 rules and can include some dynamic values (see Custom Log Format
7203 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007204
7205 <sch> With "redirect scheme", then the "Location" header is built by
7206 concatenating <sch> with "://" then the first occurrence of the
7207 "Host" header, and then the URI path, including the query string
7208 unless the "drop-query" option is specified (see below). If no
7209 path is found or if the path is "*", then "/" is used instead. If
7210 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007211 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007212 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007213 HTTPS. When used in an "http-request" rule, <sch> value follows
7214 the log-format rules and can include some dynamic values (see
7215 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +01007216
7217 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +01007218 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
7219 with 302 used by default if no code is specified. 301 means
7220 "Moved permanently", and a browser may cache the Location. 302
Baptiste Assmannea849c02015-08-03 11:42:50 +02007221 means "Moved temporarily" and means that the browser should not
Willy Tarreaub67fdc42013-03-29 19:28:11 +01007222 cache the redirection. 303 is equivalent to 302 except that the
7223 browser will fetch the location with a GET method. 307 is just
7224 like 302 but makes it clear that the same method must be reused.
7225 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +01007226
7227 <option> There are several options which can be specified to adjust the
Davor Ocelice9ed2812017-12-25 17:49:28 +01007228 expected behavior of a redirection :
Willy Tarreau0140f252008-11-19 21:07:09 +01007229
7230 - "drop-query"
7231 When this keyword is used in a prefix-based redirection, then the
7232 location will be set without any possible query-string, which is useful
7233 for directing users to a non-secure page for instance. It has no effect
7234 with a location-type redirect.
7235
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01007236 - "append-slash"
7237 This keyword may be used in conjunction with "drop-query" to redirect
7238 users who use a URL not ending with a '/' to the same one with the '/'.
7239 It can be useful to ensure that search engines will only see one URL.
7240 For this, a return code 301 is preferred.
7241
Willy Tarreau0140f252008-11-19 21:07:09 +01007242 - "set-cookie NAME[=value]"
7243 A "Set-Cookie" header will be added with NAME (and optionally "=value")
7244 to the response. This is sometimes used to indicate that a user has
7245 been seen, for instance to protect against some types of DoS. No other
7246 cookie option is added, so the cookie will be a session cookie. Note
7247 that for a browser, a sole cookie name without an equal sign is
7248 different from a cookie with an equal sign.
7249
7250 - "clear-cookie NAME[=]"
7251 A "Set-Cookie" header will be added with NAME (and optionally "="), but
7252 with the "Max-Age" attribute set to zero. This will tell the browser to
7253 delete this cookie. It is useful for instance on logout pages. It is
7254 important to note that clearing the cookie "NAME" will not remove a
7255 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
7256 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007257
7258 Example: move the login URL only to HTTPS.
7259 acl clear dst_port 80
7260 acl secure dst_port 8080
7261 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +01007262 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +01007263 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +01007264 acl cookie_set hdr_sub(cookie) SEEN=1
7265
7266 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +01007267 redirect prefix https://mysite.com if login_page !secure
7268 redirect prefix http://mysite.com drop-query if login_page !uid_given
7269 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +01007270 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007271
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01007272 Example: send redirects for request for articles without a '/'.
7273 acl missing_slash path_reg ^/article/[^/]*$
7274 redirect code 301 prefix / drop-query append-slash if missing_slash
7275
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007276 Example: redirect all HTTP traffic to HTTPS when SSL is handled by haproxy.
David BERARDe7153042012-11-03 00:11:31 +01007277 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007278
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007279 Example: append 'www.' prefix in front of all hosts not having it
Coen Rosdorff596659b2016-04-11 11:33:49 +02007280 http-request redirect code 301 location \
7281 http://www.%[hdr(host)]%[capture.req.uri] \
7282 unless { hdr_beg(host) -i www }
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007283
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007284 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007285
7286
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007287redisp (deprecated)
7288redispatch (deprecated)
7289 Enable or disable session redistribution in case of connection failure
7290 May be used in sections: defaults | frontend | listen | backend
7291 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007292 Arguments : none
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007293
7294 In HTTP mode, if a server designated by a cookie is down, clients may
7295 definitely stick to it because they cannot flush the cookie, so they will not
7296 be able to access the service anymore.
7297
7298 Specifying "redispatch" will allow the proxy to break their persistence and
7299 redistribute them to a working server.
7300
7301 It also allows to retry last connection to another server in case of multiple
7302 connection failures. Of course, it requires having "retries" set to a nonzero
7303 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01007304
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007305 This form is deprecated, do not use it in any new configuration, use the new
7306 "option redispatch" instead.
7307
7308 See also : "option redispatch"
7309
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007310
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01007311reqadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01007312 Add a header at the end of the HTTP request
7313 May be used in sections : defaults | frontend | listen | backend
7314 no | yes | yes | yes
7315 Arguments :
7316 <string> is the complete line to be added. Any space or known delimiter
7317 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007318 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007319
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01007320 <cond> is an optional matching condition built from ACLs. It makes it
7321 possible to ignore this rule when other conditions are not met.
7322
Willy Tarreau303c0352008-01-17 19:01:39 +01007323 A new line consisting in <string> followed by a line feed will be added after
7324 the last header of an HTTP request.
7325
7326 Header transformations only apply to traffic which passes through HAProxy,
7327 and not to traffic generated by HAProxy, such as health-checks or error
7328 responses.
7329
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01007330 Example : add "X-Proto: SSL" to requests coming via port 81
7331 acl is-ssl dst_port 81
7332 reqadd X-Proto:\ SSL if is-ssl
7333
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007334 See also: "rspadd", "http-request", section 6 about HTTP header manipulation,
7335 and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007336
7337
Willy Tarreau5321c422010-01-28 20:35:13 +01007338reqallow <search> [{if | unless} <cond>]
7339reqiallow <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007340 Definitely allow an HTTP request if a line matches a regular expression
7341 May be used in sections : defaults | frontend | listen | backend
7342 no | yes | yes | yes
7343 Arguments :
7344 <search> is the regular expression applied to HTTP headers and to the
7345 request line. This is an extended regular expression. Parenthesis
7346 grouping is supported and no preliminary backslash is required.
7347 Any space or known delimiter must be escaped using a backslash
7348 ('\'). The pattern applies to a full line at a time. The
7349 "reqallow" keyword strictly matches case while "reqiallow"
7350 ignores case.
7351
Willy Tarreau5321c422010-01-28 20:35:13 +01007352 <cond> is an optional matching condition built from ACLs. It makes it
7353 possible to ignore this rule when other conditions are not met.
7354
Willy Tarreau303c0352008-01-17 19:01:39 +01007355 A request containing any line which matches extended regular expression
7356 <search> will mark the request as allowed, even if any later test would
7357 result in a deny. The test applies both to the request line and to request
7358 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007359 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007360
7361 It is easier, faster and more powerful to use ACLs to write access policies.
7362 Reqdeny, reqallow and reqpass should be avoided in new designs.
7363
7364 Example :
7365 # allow www.* but refuse *.local
7366 reqiallow ^Host:\ www\.
7367 reqideny ^Host:\ .*\.local
7368
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007369 See also: "reqdeny", "block", "http-request", section 6 about HTTP header
7370 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007371
7372
Willy Tarreau5321c422010-01-28 20:35:13 +01007373reqdel <search> [{if | unless} <cond>]
7374reqidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007375 Delete all headers matching a regular expression in an HTTP request
7376 May be used in sections : defaults | frontend | listen | backend
7377 no | yes | yes | yes
7378 Arguments :
7379 <search> is the regular expression applied to HTTP headers and to the
7380 request line. This is an extended regular expression. Parenthesis
7381 grouping is supported and no preliminary backslash is required.
7382 Any space or known delimiter must be escaped using a backslash
7383 ('\'). The pattern applies to a full line at a time. The "reqdel"
7384 keyword strictly matches case while "reqidel" ignores case.
7385
Willy Tarreau5321c422010-01-28 20:35:13 +01007386 <cond> is an optional matching condition built from ACLs. It makes it
7387 possible to ignore this rule when other conditions are not met.
7388
Willy Tarreau303c0352008-01-17 19:01:39 +01007389 Any header line matching extended regular expression <search> in the request
7390 will be completely deleted. Most common use of this is to remove unwanted
7391 and/or dangerous headers or cookies from a request before passing it to the
7392 next servers.
7393
7394 Header transformations only apply to traffic which passes through HAProxy,
7395 and not to traffic generated by HAProxy, such as health-checks or error
7396 responses. Keep in mind that header names are not case-sensitive.
7397
7398 Example :
7399 # remove X-Forwarded-For header and SERVER cookie
7400 reqidel ^X-Forwarded-For:.*
7401 reqidel ^Cookie:.*SERVER=
7402
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007403 See also: "reqadd", "reqrep", "rspdel", "http-request", section 6 about
7404 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007405
7406
Willy Tarreau5321c422010-01-28 20:35:13 +01007407reqdeny <search> [{if | unless} <cond>]
7408reqideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007409 Deny an HTTP request if a line matches a regular expression
7410 May be used in sections : defaults | frontend | listen | backend
7411 no | yes | yes | yes
7412 Arguments :
7413 <search> is the regular expression applied to HTTP headers and to the
7414 request line. This is an extended regular expression. Parenthesis
7415 grouping is supported and no preliminary backslash is required.
7416 Any space or known delimiter must be escaped using a backslash
7417 ('\'). The pattern applies to a full line at a time. The
7418 "reqdeny" keyword strictly matches case while "reqideny" ignores
7419 case.
7420
Willy Tarreau5321c422010-01-28 20:35:13 +01007421 <cond> is an optional matching condition built from ACLs. It makes it
7422 possible to ignore this rule when other conditions are not met.
7423
Willy Tarreau303c0352008-01-17 19:01:39 +01007424 A request containing any line which matches extended regular expression
7425 <search> will mark the request as denied, even if any later test would
7426 result in an allow. The test applies both to the request line and to request
7427 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007428 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007429
Willy Tarreauced27012008-01-17 20:35:34 +01007430 A denied request will generate an "HTTP 403 forbidden" response once the
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007431 complete request has been parsed. This is consistent with what is practiced
Willy Tarreaud72758d2010-01-12 10:42:19 +01007432 using ACLs.
Willy Tarreauced27012008-01-17 20:35:34 +01007433
Willy Tarreau303c0352008-01-17 19:01:39 +01007434 It is easier, faster and more powerful to use ACLs to write access policies.
7435 Reqdeny, reqallow and reqpass should be avoided in new designs.
7436
7437 Example :
7438 # refuse *.local, then allow www.*
7439 reqideny ^Host:\ .*\.local
7440 reqiallow ^Host:\ www\.
7441
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007442 See also: "reqallow", "rspdeny", "block", "http-request", section 6 about
7443 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007444
7445
Willy Tarreau5321c422010-01-28 20:35:13 +01007446reqpass <search> [{if | unless} <cond>]
7447reqipass <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007448 Ignore any HTTP request line matching a regular expression in next rules
7449 May be used in sections : defaults | frontend | listen | backend
7450 no | yes | yes | yes
7451 Arguments :
7452 <search> is the regular expression applied to HTTP headers and to the
7453 request line. This is an extended regular expression. Parenthesis
7454 grouping is supported and no preliminary backslash is required.
7455 Any space or known delimiter must be escaped using a backslash
7456 ('\'). The pattern applies to a full line at a time. The
7457 "reqpass" keyword strictly matches case while "reqipass" ignores
7458 case.
7459
Willy Tarreau5321c422010-01-28 20:35:13 +01007460 <cond> is an optional matching condition built from ACLs. It makes it
7461 possible to ignore this rule when other conditions are not met.
7462
Willy Tarreau303c0352008-01-17 19:01:39 +01007463 A request containing any line which matches extended regular expression
7464 <search> will skip next rules, without assigning any deny or allow verdict.
7465 The test applies both to the request line and to request headers. Keep in
7466 mind that URLs in request line are case-sensitive while header names are not.
7467
7468 It is easier, faster and more powerful to use ACLs to write access policies.
7469 Reqdeny, reqallow and reqpass should be avoided in new designs.
7470
7471 Example :
7472 # refuse *.local, then allow www.*, but ignore "www.private.local"
7473 reqipass ^Host:\ www.private\.local
7474 reqideny ^Host:\ .*\.local
7475 reqiallow ^Host:\ www\.
7476
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007477 See also: "reqallow", "reqdeny", "block", "http-request", section 6 about
7478 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007479
7480
Willy Tarreau5321c422010-01-28 20:35:13 +01007481reqrep <search> <string> [{if | unless} <cond>]
7482reqirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007483 Replace a regular expression with a string in an HTTP request line
7484 May be used in sections : defaults | frontend | listen | backend
7485 no | yes | yes | yes
7486 Arguments :
7487 <search> is the regular expression applied to HTTP headers and to the
7488 request line. This is an extended regular expression. Parenthesis
7489 grouping is supported and no preliminary backslash is required.
7490 Any space or known delimiter must be escaped using a backslash
7491 ('\'). The pattern applies to a full line at a time. The "reqrep"
7492 keyword strictly matches case while "reqirep" ignores case.
7493
7494 <string> is the complete line to be added. Any space or known delimiter
7495 must be escaped using a backslash ('\'). References to matched
7496 pattern groups are possible using the common \N form, with N
7497 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007498 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007499
Willy Tarreau5321c422010-01-28 20:35:13 +01007500 <cond> is an optional matching condition built from ACLs. It makes it
7501 possible to ignore this rule when other conditions are not met.
7502
Willy Tarreau303c0352008-01-17 19:01:39 +01007503 Any line matching extended regular expression <search> in the request (both
7504 the request line and header lines) will be completely replaced with <string>.
7505 Most common use of this is to rewrite URLs or domain names in "Host" headers.
7506
7507 Header transformations only apply to traffic which passes through HAProxy,
7508 and not to traffic generated by HAProxy, such as health-checks or error
7509 responses. Note that for increased readability, it is suggested to add enough
7510 spaces between the request and the response. Keep in mind that URLs in
7511 request line are case-sensitive while header names are not.
7512
7513 Example :
7514 # replace "/static/" with "/" at the beginning of any request path.
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04007515 reqrep ^([^\ :]*)\ /static/(.*) \1\ /\2
Willy Tarreau303c0352008-01-17 19:01:39 +01007516 # replace "www.mydomain.com" with "www" in the host name.
7517 reqirep ^Host:\ www.mydomain.com Host:\ www
7518
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007519 See also: "reqadd", "reqdel", "rsprep", "tune.bufsize", "http-request",
7520 section 6 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007521
7522
Willy Tarreau5321c422010-01-28 20:35:13 +01007523reqtarpit <search> [{if | unless} <cond>]
7524reqitarpit <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007525 Tarpit an HTTP request containing a line matching a regular expression
7526 May be used in sections : defaults | frontend | listen | backend
7527 no | yes | yes | yes
7528 Arguments :
7529 <search> is the regular expression applied to HTTP headers and to the
7530 request line. This is an extended regular expression. Parenthesis
7531 grouping is supported and no preliminary backslash is required.
7532 Any space or known delimiter must be escaped using a backslash
7533 ('\'). The pattern applies to a full line at a time. The
7534 "reqtarpit" keyword strictly matches case while "reqitarpit"
7535 ignores case.
7536
Willy Tarreau5321c422010-01-28 20:35:13 +01007537 <cond> is an optional matching condition built from ACLs. It makes it
7538 possible to ignore this rule when other conditions are not met.
7539
Willy Tarreau303c0352008-01-17 19:01:39 +01007540 A request containing any line which matches extended regular expression
7541 <search> will be tarpitted, which means that it will connect to nowhere, will
Willy Tarreauced27012008-01-17 20:35:34 +01007542 be kept open for a pre-defined time, then will return an HTTP error 500 so
7543 that the attacker does not suspect it has been tarpitted. The status 500 will
7544 be reported in the logs, but the completion flags will indicate "PT". The
Willy Tarreau303c0352008-01-17 19:01:39 +01007545 delay is defined by "timeout tarpit", or "timeout connect" if the former is
7546 not set.
7547
7548 The goal of the tarpit is to slow down robots attacking servers with
7549 identifiable requests. Many robots limit their outgoing number of connections
7550 and stay connected waiting for a reply which can take several minutes to
7551 come. Depending on the environment and attack, it may be particularly
7552 efficient at reducing the load on the network and firewalls.
7553
Willy Tarreau5321c422010-01-28 20:35:13 +01007554 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01007555 # ignore user-agents reporting any flavor of "Mozilla" or "MSIE", but
Willy Tarreau303c0352008-01-17 19:01:39 +01007556 # block all others.
7557 reqipass ^User-Agent:\.*(Mozilla|MSIE)
7558 reqitarpit ^User-Agent:
7559
Willy Tarreau5321c422010-01-28 20:35:13 +01007560 # block bad guys
7561 acl badguys src 10.1.0.3 172.16.13.20/28
7562 reqitarpit . if badguys
7563
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007564 See also: "reqallow", "reqdeny", "reqpass", "http-request", section 6
7565 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007566
7567
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007568retries <value>
7569 Set the number of retries to perform on a server after a connection failure
7570 May be used in sections: defaults | frontend | listen | backend
7571 yes | no | yes | yes
7572 Arguments :
7573 <value> is the number of times a connection attempt should be retried on
7574 a server when a connection either is refused or times out. The
7575 default value is 3.
7576
7577 It is important to understand that this value applies to the number of
7578 connection attempts, not full requests. When a connection has effectively
7579 been established to a server, there will be no more retry.
7580
7581 In order to avoid immediate reconnections to a server which is restarting,
Joseph Lynch726ab712015-05-11 23:25:34 -07007582 a turn-around timer of min("timeout connect", one second) is applied before
7583 a retry occurs.
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007584
7585 When "option redispatch" is set, the last retry may be performed on another
7586 server even if a cookie references a different server.
7587
7588 See also : "option redispatch"
7589
7590
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007591rspadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01007592 Add a header at the end of the HTTP response
7593 May be used in sections : defaults | frontend | listen | backend
7594 no | yes | yes | yes
7595 Arguments :
7596 <string> is the complete line to be added. Any space or known delimiter
7597 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007598 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007599
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007600 <cond> is an optional matching condition built from ACLs. It makes it
7601 possible to ignore this rule when other conditions are not met.
7602
Willy Tarreau303c0352008-01-17 19:01:39 +01007603 A new line consisting in <string> followed by a line feed will be added after
7604 the last header of an HTTP response.
7605
7606 Header transformations only apply to traffic which passes through HAProxy,
7607 and not to traffic generated by HAProxy, such as health-checks or error
7608 responses.
7609
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007610 See also: "rspdel" "reqadd", "http-response", section 6 about HTTP header
7611 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007612
7613
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007614rspdel <search> [{if | unless} <cond>]
7615rspidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007616 Delete all headers matching a regular expression in an HTTP response
7617 May be used in sections : defaults | frontend | listen | backend
7618 no | yes | yes | yes
7619 Arguments :
7620 <search> is the regular expression applied to HTTP headers and to the
7621 response line. This is an extended regular expression, so
7622 parenthesis grouping is supported and no preliminary backslash
7623 is required. Any space or known delimiter must be escaped using
7624 a backslash ('\'). The pattern applies to a full line at a time.
7625 The "rspdel" keyword strictly matches case while "rspidel"
7626 ignores case.
7627
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007628 <cond> is an optional matching condition built from ACLs. It makes it
7629 possible to ignore this rule when other conditions are not met.
7630
Willy Tarreau303c0352008-01-17 19:01:39 +01007631 Any header line matching extended regular expression <search> in the response
7632 will be completely deleted. Most common use of this is to remove unwanted
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02007633 and/or sensitive headers or cookies from a response before passing it to the
Willy Tarreau303c0352008-01-17 19:01:39 +01007634 client.
7635
7636 Header transformations only apply to traffic which passes through HAProxy,
7637 and not to traffic generated by HAProxy, such as health-checks or error
7638 responses. Keep in mind that header names are not case-sensitive.
7639
7640 Example :
7641 # remove the Server header from responses
Willy Tarreau5e80e022013-05-25 08:31:25 +02007642 rspidel ^Server:.*
Willy Tarreau303c0352008-01-17 19:01:39 +01007643
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007644 See also: "rspadd", "rsprep", "reqdel", "http-response", section 6 about
7645 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007646
7647
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007648rspdeny <search> [{if | unless} <cond>]
7649rspideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007650 Block an HTTP response if a line matches a regular expression
7651 May be used in sections : defaults | frontend | listen | backend
7652 no | yes | yes | yes
7653 Arguments :
7654 <search> is the regular expression applied to HTTP headers and to the
7655 response line. This is an extended regular expression, so
7656 parenthesis grouping is supported and no preliminary backslash
7657 is required. Any space or known delimiter must be escaped using
7658 a backslash ('\'). The pattern applies to a full line at a time.
7659 The "rspdeny" keyword strictly matches case while "rspideny"
7660 ignores case.
7661
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007662 <cond> is an optional matching condition built from ACLs. It makes it
7663 possible to ignore this rule when other conditions are not met.
7664
Willy Tarreau303c0352008-01-17 19:01:39 +01007665 A response containing any line which matches extended regular expression
7666 <search> will mark the request as denied. The test applies both to the
7667 response line and to response headers. Keep in mind that header names are not
7668 case-sensitive.
7669
7670 Main use of this keyword is to prevent sensitive information leak and to
Willy Tarreauced27012008-01-17 20:35:34 +01007671 block the response before it reaches the client. If a response is denied, it
7672 will be replaced with an HTTP 502 error so that the client never retrieves
7673 any sensitive data.
Willy Tarreau303c0352008-01-17 19:01:39 +01007674
7675 It is easier, faster and more powerful to use ACLs to write access policies.
7676 Rspdeny should be avoided in new designs.
7677
7678 Example :
7679 # Ensure that no content type matching ms-word will leak
7680 rspideny ^Content-type:\.*/ms-word
7681
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007682 See also: "reqdeny", "acl", "block", "http-response", section 6 about
7683 HTTP header manipulation and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007684
7685
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007686rsprep <search> <string> [{if | unless} <cond>]
7687rspirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007688 Replace a regular expression with a string in an HTTP response line
7689 May be used in sections : defaults | frontend | listen | backend
7690 no | yes | yes | yes
7691 Arguments :
7692 <search> is the regular expression applied to HTTP headers and to the
7693 response line. This is an extended regular expression, so
7694 parenthesis grouping is supported and no preliminary backslash
7695 is required. Any space or known delimiter must be escaped using
7696 a backslash ('\'). The pattern applies to a full line at a time.
7697 The "rsprep" keyword strictly matches case while "rspirep"
7698 ignores case.
7699
7700 <string> is the complete line to be added. Any space or known delimiter
7701 must be escaped using a backslash ('\'). References to matched
7702 pattern groups are possible using the common \N form, with N
7703 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007704 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007705
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007706 <cond> is an optional matching condition built from ACLs. It makes it
7707 possible to ignore this rule when other conditions are not met.
7708
Willy Tarreau303c0352008-01-17 19:01:39 +01007709 Any line matching extended regular expression <search> in the response (both
7710 the response line and header lines) will be completely replaced with
7711 <string>. Most common use of this is to rewrite Location headers.
7712
7713 Header transformations only apply to traffic which passes through HAProxy,
7714 and not to traffic generated by HAProxy, such as health-checks or error
7715 responses. Note that for increased readability, it is suggested to add enough
7716 spaces between the request and the response. Keep in mind that header names
7717 are not case-sensitive.
7718
7719 Example :
7720 # replace "Location: 127.0.0.1:8080" with "Location: www.mydomain.com"
7721 rspirep ^Location:\ 127.0.0.1:8080 Location:\ www.mydomain.com
7722
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007723 See also: "rspadd", "rspdel", "reqrep", "http-response", section 6 about
7724 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007725
7726
David du Colombier486df472011-03-17 10:40:26 +01007727server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007728 Declare a server in a backend
7729 May be used in sections : defaults | frontend | listen | backend
7730 no | no | yes | yes
7731 Arguments :
7732 <name> is the internal name assigned to this server. This name will
Davor Ocelice9ed2812017-12-25 17:49:28 +01007733 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -05007734 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007735
David du Colombier486df472011-03-17 10:40:26 +01007736 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
7737 resolvable hostname is supported, but this name will be resolved
7738 during start-up. Address "0.0.0.0" or "*" has a special meaning.
7739 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +02007740 address as the one from the client connection. This is useful in
7741 transparent proxy architectures where the client's connection is
7742 intercepted and haproxy must forward to the original destination
7743 address. This is more or less what the "transparent" keyword does
7744 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +01007745 to report statistics. Optionally, an address family prefix may be
7746 used before the address to force the family regardless of the
7747 address format, which can be useful to specify a path to a unix
7748 socket with no slash ('/'). Currently supported prefixes are :
7749 - 'ipv4@' -> address is always IPv4
7750 - 'ipv6@' -> address is always IPv6
7751 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02007752 - 'abns@' -> address is in abstract namespace (Linux only)
William Lallemand2fe7dd02018-09-11 16:51:29 +02007753 - 'sockpair@' -> address is the FD of a connected unix
7754 socket or of a socketpair. During a connection, the
7755 backend creates a pair of connected sockets, and passes
7756 one of them over the FD. The bind part will use the
7757 received socket as the client FD. Should be used
7758 carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02007759 You may want to reference some environment variables in the
7760 address parameter, see section 2.3 about environment
Willy Tarreau6a031d12016-11-07 19:42:35 +01007761 variables. The "init-addr" setting can be used to modify the way
7762 IP addresses should be resolved upon startup.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007763
Willy Tarreaub6205fd2012-09-24 12:27:33 +02007764 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007765 be sent to this port. If unset, the same port the client
7766 connected to will be used. The port may also be prefixed by a "+"
7767 or a "-". In this case, the server's port will be determined by
7768 adding this value to the client's port.
7769
7770 <param*> is a list of parameters for this server. The "server" keywords
7771 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007772 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007773
7774 Examples :
7775 server first 10.1.1.1:1080 cookie first check inter 1000
7776 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +01007777 server transp ipv4@
William Lallemandb2f07452015-05-12 14:27:13 +02007778 server backup "${SRV_BACKUP}:1080" backup
7779 server www1_dc1 "${LAN_DC1}.101:80"
7780 server www1_dc2 "${LAN_DC2}.101:80"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007781
Willy Tarreau55dcaf62015-09-27 15:03:15 +02007782 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
7783 sun_path length is used for the address length. Some other programs
7784 such as socat use the string length only by default. Pass the option
7785 ",unix-tightsocklen=0" to any abstract socket definition in socat to
7786 make it compatible with HAProxy's.
7787
Mark Lamourinec2247f02012-01-04 13:02:01 -05007788 See also: "default-server", "http-send-name-header" and section 5 about
7789 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007790
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007791server-state-file-name [<file>]
7792 Set the server state file to read, load and apply to servers available in
7793 this backend. It only applies when the directive "load-server-state-from-file"
7794 is set to "local". When <file> is not provided or if this directive is not
7795 set, then backend name is used. If <file> starts with a slash '/', then it is
7796 considered as an absolute path. Otherwise, <file> is concatenated to the
7797 global directive "server-state-file-base".
7798
7799 Example: the minimal configuration below would make HAProxy look for the
7800 state server file '/etc/haproxy/states/bk':
7801
7802 global
7803 server-state-file-base /etc/haproxy/states
7804
7805 backend bk
7806 load-server-state-from-file
7807
7808 See also: "server-state-file-base", "load-server-state-from-file", and
7809 "show servers state"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007810
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02007811server-template <prefix> <num | range> <fqdn>[:<port>] [params*]
7812 Set a template to initialize servers with shared parameters.
7813 The names of these servers are built from <prefix> and <num | range> parameters.
7814 May be used in sections : defaults | frontend | listen | backend
7815 no | no | yes | yes
7816
7817 Arguments:
7818 <prefix> A prefix for the server names to be built.
7819
7820 <num | range>
7821 If <num> is provided, this template initializes <num> servers
7822 with 1 up to <num> as server name suffixes. A range of numbers
7823 <num_low>-<num_high> may also be used to use <num_low> up to
7824 <num_high> as server name suffixes.
7825
7826 <fqdn> A FQDN for all the servers this template initializes.
7827
7828 <port> Same meaning as "server" <port> argument (see "server" keyword).
7829
7830 <params*>
7831 Remaining server parameters among all those supported by "server"
7832 keyword.
7833
7834 Examples:
7835 # Initializes 3 servers with srv1, srv2 and srv3 as names,
7836 # google.com as FQDN, and health-check enabled.
7837 server-template srv 1-3 google.com:80 check
7838
7839 # or
7840 server-template srv 3 google.com:80 check
7841
7842 # would be equivalent to:
7843 server srv1 google.com:80 check
7844 server srv2 google.com:80 check
7845 server srv3 google.com:80 check
7846
7847
7848
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007849source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02007850source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +01007851source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007852 Set the source address for outgoing connections
7853 May be used in sections : defaults | frontend | listen | backend
7854 yes | no | yes | yes
7855 Arguments :
7856 <addr> is the IPv4 address HAProxy will bind to before connecting to a
7857 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +01007858
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007859 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +01007860 the most appropriate address to reach its destination. Optionally
7861 an address family prefix may be used before the address to force
7862 the family regardless of the address format, which can be useful
7863 to specify a path to a unix socket with no slash ('/'). Currently
7864 supported prefixes are :
7865 - 'ipv4@' -> address is always IPv4
7866 - 'ipv6@' -> address is always IPv6
7867 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02007868 - 'abns@' -> address is in abstract namespace (Linux only)
Cyril Bonté307ee1e2015-09-28 23:16:06 +02007869 You may want to reference some environment variables in the
7870 address parameter, see section 2.3 about environment variables.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007871
7872 <port> is an optional port. It is normally not needed but may be useful
7873 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02007874 the system will select a free port. Note that port ranges are not
7875 supported in the backend. If you want to force port ranges, you
7876 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007877
7878 <addr2> is the IP address to present to the server when connections are
7879 forwarded in full transparent proxy mode. This is currently only
7880 supported on some patched Linux kernels. When this address is
7881 specified, clients connecting to the server will be presented
7882 with this address, while health checks will still use the address
7883 <addr>.
7884
7885 <port2> is the optional port to present to the server when connections
7886 are forwarded in full transparent proxy mode (see <addr2> above).
7887 The default value of zero means the system will select a free
7888 port.
7889
Willy Tarreaubce70882009-09-07 11:51:47 +02007890 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
7891 This is the name of a comma-separated header list which can
7892 contain multiple IP addresses. By default, the last occurrence is
7893 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +01007894 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +02007895 by previous proxy, typically Stunnel. In order to use another
7896 occurrence from the last one, please see the <occ> parameter
7897 below. When the header (or occurrence) is not found, no binding
7898 is performed so that the proxy's default IP address is used. Also
7899 keep in mind that the header name is case insensitive, as for any
7900 HTTP header.
7901
7902 <occ> is the occurrence number of a value to be used in a multi-value
7903 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007904 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +02007905 address. Positive values indicate a position from the first
7906 occurrence, 1 being the first one. Negative values indicate
7907 positions relative to the last one, -1 being the last one. This
7908 is helpful for situations where an X-Forwarded-For header is set
7909 at the entry point of an infrastructure and must be used several
7910 proxy layers away. When this value is not specified, -1 is
7911 assumed. Passing a zero here disables the feature.
7912
Willy Tarreaud53f96b2009-02-04 18:46:54 +01007913 <name> is an optional interface name to which to bind to for outgoing
7914 traffic. On systems supporting this features (currently, only
7915 Linux), this allows one to bind all traffic to the server to
7916 this interface even if it is not the one the system would select
7917 based on routing tables. This should be used with extreme care.
7918 Note that using this option requires root privileges.
7919
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007920 The "source" keyword is useful in complex environments where a specific
7921 address only is allowed to connect to the servers. It may be needed when a
7922 private address must be used through a public gateway for instance, and it is
7923 known that the system cannot determine the adequate source address by itself.
7924
7925 An extension which is available on certain patched Linux kernels may be used
7926 through the "usesrc" optional keyword. It makes it possible to connect to the
7927 servers with an IP address which does not belong to the system itself. This
7928 is called "full transparent proxy mode". For this to work, the destination
7929 servers have to route their traffic back to this address through the machine
7930 running HAProxy, and IP forwarding must generally be enabled on this machine.
7931
7932 In this "full transparent proxy" mode, it is possible to force a specific IP
7933 address to be presented to the servers. This is not much used in fact. A more
7934 common use is to tell HAProxy to present the client's IP address. For this,
7935 there are two methods :
7936
7937 - present the client's IP and port addresses. This is the most transparent
7938 mode, but it can cause problems when IP connection tracking is enabled on
7939 the machine, because a same connection may be seen twice with different
7940 states. However, this solution presents the huge advantage of not
7941 limiting the system to the 64k outgoing address+port couples, because all
7942 of the client ranges may be used.
7943
7944 - present only the client's IP address and select a spare port. This
7945 solution is still quite elegant but slightly less transparent (downstream
7946 firewalls logs will not match upstream's). It also presents the downside
7947 of limiting the number of concurrent connections to the usual 64k ports.
7948 However, since the upstream and downstream ports are different, local IP
7949 connection tracking on the machine will not be upset by the reuse of the
7950 same session.
7951
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007952 This option sets the default source for all servers in the backend. It may
7953 also be specified in a "defaults" section. Finer source address specification
7954 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007955 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007956
Baptiste Assmann91bd3372015-07-17 21:59:42 +02007957 In order to work, "usesrc" requires root privileges.
7958
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007959 Examples :
7960 backend private
7961 # Connect to the servers using our 192.168.1.200 source address
7962 source 192.168.1.200
7963
7964 backend transparent_ssl1
7965 # Connect to the SSL farm from the client's source address
7966 source 192.168.1.200 usesrc clientip
7967
7968 backend transparent_ssl2
7969 # Connect to the SSL farm from the client's source address and port
7970 # not recommended if IP conntrack is present on the local machine.
7971 source 192.168.1.200 usesrc client
7972
7973 backend transparent_ssl3
7974 # Connect to the SSL farm from the client's source address. It
7975 # is more conntrack-friendly.
7976 source 192.168.1.200 usesrc clientip
7977
7978 backend transparent_smtp
7979 # Connect to the SMTP farm from the client's source address/port
7980 # with Tproxy version 4.
7981 source 0.0.0.0 usesrc clientip
7982
Willy Tarreaubce70882009-09-07 11:51:47 +02007983 backend transparent_http
7984 # Connect to the servers using the client's IP as seen by previous
7985 # proxy.
7986 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
7987
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007988 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007989 the Linux kernel on www.balabit.com, the "bind" keyword.
7990
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007991
Willy Tarreau844e3c52008-01-11 16:28:18 +01007992srvtimeout <timeout> (deprecated)
7993 Set the maximum inactivity time on the server side.
7994 May be used in sections : defaults | frontend | listen | backend
7995 yes | no | yes | yes
7996 Arguments :
7997 <timeout> is the timeout value specified in milliseconds by default, but
7998 can be in any other unit if the number is suffixed by the unit,
7999 as explained at the top of this document.
8000
8001 The inactivity timeout applies when the server is expected to acknowledge or
8002 send data. In HTTP mode, this timeout is particularly important to consider
8003 during the first phase of the server's response, when it has to send the
8004 headers, as it directly represents the server's processing time for the
8005 request. To find out what value to put there, it's often good to start with
8006 what would be considered as unacceptable response times, then check the logs
8007 to observe the response time distribution, and adjust the value accordingly.
8008
8009 The value is specified in milliseconds by default, but can be in any other
8010 unit if the number is suffixed by the unit, as specified at the top of this
8011 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
8012 recommended that the client timeout remains equal to the server timeout in
8013 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01008014 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +01008015 packet losses by specifying timeouts that are slightly above multiples of 3
Davor Ocelice9ed2812017-12-25 17:49:28 +01008016 seconds (e.g. 4 or 5 seconds minimum).
Willy Tarreau844e3c52008-01-11 16:28:18 +01008017
8018 This parameter is specific to backends, but can be specified once for all in
8019 "defaults" sections. This is in fact one of the easiest solutions not to
8020 forget about it. An unspecified timeout results in an infinite timeout, which
8021 is not recommended. Such a usage is accepted and works but reports a warning
8022 during startup because it may results in accumulation of expired sessions in
8023 the system if the system's timeouts are not configured either.
8024
8025 This parameter is provided for compatibility but is currently deprecated.
8026 Please use "timeout server" instead.
8027
Willy Tarreauce887fd2012-05-12 12:50:00 +02008028 See also : "timeout server", "timeout tunnel", "timeout client" and
8029 "clitimeout".
Willy Tarreau844e3c52008-01-11 16:28:18 +01008030
8031
Cyril Bonté66c327d2010-10-12 00:14:37 +02008032stats admin { if | unless } <cond>
8033 Enable statistics admin level if/unless a condition is matched
8034 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008035 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +02008036
8037 This statement enables the statistics admin level if/unless a condition is
8038 matched.
8039
8040 The admin level allows to enable/disable servers from the web interface. By
8041 default, statistics page is read-only for security reasons.
8042
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008043 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8044 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008045 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008046
Cyril Bonté23b39d92011-02-10 22:54:44 +01008047 Currently, the POST request is limited to the buffer size minus the reserved
8048 buffer space, which means that if the list of servers is too long, the
8049 request won't be processed. It is recommended to alter few servers at a
8050 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +02008051
8052 Example :
8053 # statistics admin level only for localhost
8054 backend stats_localhost
8055 stats enable
8056 stats admin if LOCALHOST
8057
8058 Example :
8059 # statistics admin level always enabled because of the authentication
8060 backend stats_auth
8061 stats enable
8062 stats auth admin:AdMiN123
8063 stats admin if TRUE
8064
8065 Example :
8066 # statistics admin level depends on the authenticated user
8067 userlist stats-auth
8068 group admin users admin
8069 user admin insecure-password AdMiN123
8070 group readonly users haproxy
8071 user haproxy insecure-password haproxy
8072
8073 backend stats_auth
8074 stats enable
8075 acl AUTH http_auth(stats-auth)
8076 acl AUTH_ADMIN http_auth_group(stats-auth) admin
8077 stats http-request auth unless AUTH
8078 stats admin if AUTH_ADMIN
8079
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008080 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
8081 "bind-process", section 3.4 about userlists and section 7 about
8082 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +02008083
8084
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008085stats auth <user>:<passwd>
8086 Enable statistics with authentication and grant access to an account
8087 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008088 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008089 Arguments :
8090 <user> is a user name to grant access to
8091
8092 <passwd> is the cleartext password associated to this user
8093
8094 This statement enables statistics with default settings, and restricts access
8095 to declared users only. It may be repeated as many times as necessary to
8096 allow as many users as desired. When a user tries to access the statistics
8097 without a valid account, a "401 Forbidden" response will be returned so that
8098 the browser asks the user to provide a valid user and password. The real
8099 which will be returned to the browser is configurable using "stats realm".
8100
8101 Since the authentication method is HTTP Basic Authentication, the passwords
8102 circulate in cleartext on the network. Thus, it was decided that the
8103 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02008104 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008105
8106 It is also possible to reduce the scope of the proxies which appear in the
8107 report using "stats scope".
8108
8109 Though this statement alone is enough to enable statistics reporting, it is
8110 recommended to set all other settings in order to avoid relying on default
8111 unobvious parameters.
8112
8113 Example :
8114 # public access (limited to this backend only)
8115 backend public_www
8116 server srv1 192.168.0.1:80
8117 stats enable
8118 stats hide-version
8119 stats scope .
8120 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008121 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008122 stats auth admin1:AdMiN123
8123 stats auth admin2:AdMiN321
8124
8125 # internal monitoring access (unlimited)
8126 backend private_monitoring
8127 stats enable
8128 stats uri /admin?stats
8129 stats refresh 5s
8130
8131 See also : "stats enable", "stats realm", "stats scope", "stats uri"
8132
8133
8134stats enable
8135 Enable statistics reporting with default settings
8136 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008137 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008138 Arguments : none
8139
8140 This statement enables statistics reporting with default settings defined
8141 at build time. Unless stated otherwise, these settings are used :
8142 - stats uri : /haproxy?stats
8143 - stats realm : "HAProxy Statistics"
8144 - stats auth : no authentication
8145 - stats scope : no restriction
8146
8147 Though this statement alone is enough to enable statistics reporting, it is
8148 recommended to set all other settings in order to avoid relying on default
8149 unobvious parameters.
8150
8151 Example :
8152 # public access (limited to this backend only)
8153 backend public_www
8154 server srv1 192.168.0.1:80
8155 stats enable
8156 stats hide-version
8157 stats scope .
8158 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008159 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008160 stats auth admin1:AdMiN123
8161 stats auth admin2:AdMiN321
8162
8163 # internal monitoring access (unlimited)
8164 backend private_monitoring
8165 stats enable
8166 stats uri /admin?stats
8167 stats refresh 5s
8168
8169 See also : "stats auth", "stats realm", "stats uri"
8170
8171
Willy Tarreaud63335a2010-02-26 12:56:52 +01008172stats hide-version
8173 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008174 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008175 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008176 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008177
Willy Tarreaud63335a2010-02-26 12:56:52 +01008178 By default, the stats page reports some useful status information along with
8179 the statistics. Among them is HAProxy's version. However, it is generally
8180 considered dangerous to report precise version to anyone, as it can help them
8181 target known weaknesses with specific attacks. The "stats hide-version"
8182 statement removes the version from the statistics report. This is recommended
8183 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008184
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02008185 Though this statement alone is enough to enable statistics reporting, it is
8186 recommended to set all other settings in order to avoid relying on default
8187 unobvious parameters.
8188
Willy Tarreaud63335a2010-02-26 12:56:52 +01008189 Example :
8190 # public access (limited to this backend only)
8191 backend public_www
8192 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02008193 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01008194 stats hide-version
8195 stats scope .
8196 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008197 stats realm HAProxy\ Statistics
Willy Tarreaud63335a2010-02-26 12:56:52 +01008198 stats auth admin1:AdMiN123
8199 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008200
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008201 # internal monitoring access (unlimited)
8202 backend private_monitoring
8203 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01008204 stats uri /admin?stats
8205 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +01008206
Willy Tarreaud63335a2010-02-26 12:56:52 +01008207 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008208
Willy Tarreau983e01e2010-01-11 18:42:06 +01008209
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02008210stats http-request { allow | deny | auth [realm <realm>] }
8211 [ { if | unless } <condition> ]
8212 Access control for statistics
8213
8214 May be used in sections: defaults | frontend | listen | backend
8215 no | no | yes | yes
8216
8217 As "http-request", these set of options allow to fine control access to
8218 statistics. Each option may be followed by if/unless and acl.
8219 First option with matched condition (or option without condition) is final.
8220 For "deny" a 403 error will be returned, for "allow" normal processing is
8221 performed, for "auth" a 401/407 error code is returned so the client
8222 should be asked to enter a username and password.
8223
8224 There is no fixed limit to the number of http-request statements per
8225 instance.
8226
8227 See also : "http-request", section 3.4 about userlists and section 7
8228 about ACL usage.
8229
8230
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008231stats realm <realm>
8232 Enable statistics and set authentication realm
8233 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008234 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008235 Arguments :
8236 <realm> is the name of the HTTP Basic Authentication realm reported to
8237 the browser. The browser uses it to display it in the pop-up
8238 inviting the user to enter a valid username and password.
8239
8240 The realm is read as a single word, so any spaces in it should be escaped
8241 using a backslash ('\').
8242
8243 This statement is useful only in conjunction with "stats auth" since it is
8244 only related to authentication.
8245
8246 Though this statement alone is enough to enable statistics reporting, it is
8247 recommended to set all other settings in order to avoid relying on default
8248 unobvious parameters.
8249
8250 Example :
8251 # public access (limited to this backend only)
8252 backend public_www
8253 server srv1 192.168.0.1:80
8254 stats enable
8255 stats hide-version
8256 stats scope .
8257 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008258 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008259 stats auth admin1:AdMiN123
8260 stats auth admin2:AdMiN321
8261
8262 # internal monitoring access (unlimited)
8263 backend private_monitoring
8264 stats enable
8265 stats uri /admin?stats
8266 stats refresh 5s
8267
8268 See also : "stats auth", "stats enable", "stats uri"
8269
8270
8271stats refresh <delay>
8272 Enable statistics with automatic refresh
8273 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008274 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008275 Arguments :
8276 <delay> is the suggested refresh delay, specified in seconds, which will
8277 be returned to the browser consulting the report page. While the
8278 browser is free to apply any delay, it will generally respect it
8279 and refresh the page this every seconds. The refresh interval may
8280 be specified in any other non-default time unit, by suffixing the
8281 unit after the value, as explained at the top of this document.
8282
8283 This statement is useful on monitoring displays with a permanent page
8284 reporting the load balancer's activity. When set, the HTML report page will
8285 include a link "refresh"/"stop refresh" so that the user can select whether
8286 he wants automatic refresh of the page or not.
8287
8288 Though this statement alone is enough to enable statistics reporting, it is
8289 recommended to set all other settings in order to avoid relying on default
8290 unobvious parameters.
8291
8292 Example :
8293 # public access (limited to this backend only)
8294 backend public_www
8295 server srv1 192.168.0.1:80
8296 stats enable
8297 stats hide-version
8298 stats scope .
8299 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008300 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008301 stats auth admin1:AdMiN123
8302 stats auth admin2:AdMiN321
8303
8304 # internal monitoring access (unlimited)
8305 backend private_monitoring
8306 stats enable
8307 stats uri /admin?stats
8308 stats refresh 5s
8309
8310 See also : "stats auth", "stats enable", "stats realm", "stats uri"
8311
8312
8313stats scope { <name> | "." }
8314 Enable statistics and limit access scope
8315 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008316 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008317 Arguments :
8318 <name> is the name of a listen, frontend or backend section to be
8319 reported. The special name "." (a single dot) designates the
8320 section in which the statement appears.
8321
8322 When this statement is specified, only the sections enumerated with this
8323 statement will appear in the report. All other ones will be hidden. This
8324 statement may appear as many times as needed if multiple sections need to be
8325 reported. Please note that the name checking is performed as simple string
8326 comparisons, and that it is never checked that a give section name really
8327 exists.
8328
8329 Though this statement alone is enough to enable statistics reporting, it is
8330 recommended to set all other settings in order to avoid relying on default
8331 unobvious parameters.
8332
8333 Example :
8334 # public access (limited to this backend only)
8335 backend public_www
8336 server srv1 192.168.0.1:80
8337 stats enable
8338 stats hide-version
8339 stats scope .
8340 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008341 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008342 stats auth admin1:AdMiN123
8343 stats auth admin2:AdMiN321
8344
8345 # internal monitoring access (unlimited)
8346 backend private_monitoring
8347 stats enable
8348 stats uri /admin?stats
8349 stats refresh 5s
8350
8351 See also : "stats auth", "stats enable", "stats realm", "stats uri"
8352
Willy Tarreaud63335a2010-02-26 12:56:52 +01008353
Willy Tarreauc9705a12010-07-27 20:05:50 +02008354stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +01008355 Enable reporting of a description on the statistics page.
8356 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008357 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008358
Willy Tarreauc9705a12010-07-27 20:05:50 +02008359 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +01008360 description from global section is automatically used instead.
8361
8362 This statement is useful for users that offer shared services to their
8363 customers, where node or description should be different for each customer.
8364
8365 Though this statement alone is enough to enable statistics reporting, it is
8366 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +01008367 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008368
8369 Example :
8370 # internal monitoring access (unlimited)
8371 backend private_monitoring
8372 stats enable
8373 stats show-desc Master node for Europe, Asia, Africa
8374 stats uri /admin?stats
8375 stats refresh 5s
8376
8377 See also: "show-node", "stats enable", "stats uri" and "description" in
8378 global section.
8379
8380
8381stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +02008382 Enable reporting additional information on the statistics page
8383 May be used in sections : defaults | frontend | listen | backend
8384 yes | yes | yes | yes
8385 Arguments : none
8386
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008387 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +01008388 - cap: capabilities (proxy)
8389 - mode: one of tcp, http or health (proxy)
8390 - id: SNMP ID (proxy, socket, server)
8391 - IP (socket, server)
8392 - cookie (backend, server)
8393
8394 Though this statement alone is enough to enable statistics reporting, it is
8395 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +01008396 unobvious parameters. Default behavior is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008397
8398 See also: "stats enable", "stats uri".
8399
8400
8401stats show-node [ <name> ]
8402 Enable reporting of a host name on the statistics page.
8403 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008404 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008405 Arguments:
8406 <name> is an optional name to be reported. If unspecified, the
8407 node name from global section is automatically used instead.
8408
8409 This statement is useful for users that offer shared services to their
8410 customers, where node or description might be different on a stats page
Davor Ocelice9ed2812017-12-25 17:49:28 +01008411 provided for each customer. Default behavior is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008412
8413 Though this statement alone is enough to enable statistics reporting, it is
8414 recommended to set all other settings in order to avoid relying on default
8415 unobvious parameters.
8416
8417 Example:
8418 # internal monitoring access (unlimited)
8419 backend private_monitoring
8420 stats enable
8421 stats show-node Europe-1
8422 stats uri /admin?stats
8423 stats refresh 5s
8424
8425 See also: "show-desc", "stats enable", "stats uri", and "node" in global
8426 section.
8427
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008428
8429stats uri <prefix>
8430 Enable statistics and define the URI prefix to access them
8431 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008432 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008433 Arguments :
8434 <prefix> is the prefix of any URI which will be redirected to stats. This
8435 prefix may contain a question mark ('?') to indicate part of a
8436 query string.
8437
8438 The statistics URI is intercepted on the relayed traffic, so it appears as a
8439 page within the normal application. It is strongly advised to ensure that the
8440 selected URI will never appear in the application, otherwise it will never be
8441 possible to reach it in the application.
8442
8443 The default URI compiled in haproxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008444 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008445 It is generally a good idea to include a question mark in the URI so that
8446 intermediate proxies refrain from caching the results. Also, since any string
8447 beginning with the prefix will be accepted as a stats request, the question
8448 mark helps ensuring that no valid URI will begin with the same words.
8449
8450 It is sometimes very convenient to use "/" as the URI prefix, and put that
8451 statement in a "listen" instance of its own. That makes it easy to dedicate
8452 an address or a port to statistics only.
8453
8454 Though this statement alone is enough to enable statistics reporting, it is
8455 recommended to set all other settings in order to avoid relying on default
8456 unobvious parameters.
8457
8458 Example :
8459 # public access (limited to this backend only)
8460 backend public_www
8461 server srv1 192.168.0.1:80
8462 stats enable
8463 stats hide-version
8464 stats scope .
8465 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008466 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008467 stats auth admin1:AdMiN123
8468 stats auth admin2:AdMiN321
8469
8470 # internal monitoring access (unlimited)
8471 backend private_monitoring
8472 stats enable
8473 stats uri /admin?stats
8474 stats refresh 5s
8475
8476 See also : "stats auth", "stats enable", "stats realm"
8477
8478
Willy Tarreaud63335a2010-02-26 12:56:52 +01008479stick match <pattern> [table <table>] [{if | unless} <cond>]
8480 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008481 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +01008482 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008483
8484 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008485 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008486 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01008487 will be analyzed in the hope to find a matching entry in a
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008488 stickiness table. This rule is mandatory.
8489
8490 <table> is an optional stickiness table name. If unspecified, the same
8491 backend's table is used. A stickiness table is declared using
8492 the "stick-table" statement.
8493
8494 <cond> is an optional matching condition. It makes it possible to match
8495 on a certain criterion only when other conditions are met (or
8496 not met). For instance, it could be used to match on a source IP
8497 address except when a request passes through a known proxy, in
8498 which case we'd match on a header containing that IP address.
8499
8500 Some protocols or applications require complex stickiness rules and cannot
8501 always simply rely on cookies nor hashing. The "stick match" statement
8502 describes a rule to extract the stickiness criterion from an incoming request
8503 or connection. See section 7 for a complete list of possible patterns and
8504 transformation rules.
8505
8506 The table has to be declared using the "stick-table" statement. It must be of
8507 a type compatible with the pattern. By default it is the one which is present
8508 in the same backend. It is possible to share a table with other backends by
8509 referencing it using the "table" keyword. If another table is referenced,
8510 the server's ID inside the backends are used. By default, all server IDs
8511 start at 1 in each backend, so the server ordering is enough. But in case of
8512 doubt, it is highly recommended to force server IDs using their "id" setting.
8513
8514 It is possible to restrict the conditions where a "stick match" statement
8515 will apply, using "if" or "unless" followed by a condition. See section 7 for
8516 ACL based conditions.
8517
8518 There is no limit on the number of "stick match" statements. The first that
8519 applies and matches will cause the request to be directed to the same server
8520 as was used for the request which created the entry. That way, multiple
8521 matches can be used as fallbacks.
8522
8523 The stick rules are checked after the persistence cookies, so they will not
8524 affect stickiness if a cookie has already been used to select a server. That
8525 way, it becomes very easy to insert cookies and match on IP addresses in
8526 order to maintain stickiness between HTTP and HTTPS.
8527
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008528 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8529 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008530 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008531
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008532 Example :
8533 # forward SMTP users to the same server they just used for POP in the
8534 # last 30 minutes
8535 backend pop
8536 mode tcp
8537 balance roundrobin
8538 stick store-request src
8539 stick-table type ip size 200k expire 30m
8540 server s1 192.168.1.1:110
8541 server s2 192.168.1.1:110
8542
8543 backend smtp
8544 mode tcp
8545 balance roundrobin
8546 stick match src table pop
8547 server s1 192.168.1.1:25
8548 server s2 192.168.1.1:25
8549
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008550 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02008551 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008552
8553
8554stick on <pattern> [table <table>] [{if | unless} <condition>]
8555 Define a request pattern to associate a user to a server
8556 May be used in sections : defaults | frontend | listen | backend
8557 no | no | yes | yes
8558
8559 Note : This form is exactly equivalent to "stick match" followed by
8560 "stick store-request", all with the same arguments. Please refer
8561 to both keywords for details. It is only provided as a convenience
8562 for writing more maintainable configurations.
8563
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008564 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8565 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008566 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008567
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008568 Examples :
8569 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +01008570 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008571
8572 # ...is strictly equivalent to this one :
8573 stick match src table pop if !localhost
8574 stick store-request src table pop if !localhost
8575
8576
8577 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
8578 # well as HTTP without cookie. Share the same table between both accesses.
8579 backend http
8580 mode http
8581 balance roundrobin
8582 stick on src table https
8583 cookie SRV insert indirect nocache
8584 server s1 192.168.1.1:80 cookie s1
8585 server s2 192.168.1.1:80 cookie s2
8586
8587 backend https
8588 mode tcp
8589 balance roundrobin
8590 stick-table type ip size 200k expire 30m
8591 stick on src
8592 server s1 192.168.1.1:443
8593 server s2 192.168.1.1:443
8594
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008595 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008596
8597
8598stick store-request <pattern> [table <table>] [{if | unless} <condition>]
8599 Define a request pattern used to create an entry in a stickiness table
8600 May be used in sections : defaults | frontend | listen | backend
8601 no | no | yes | yes
8602
8603 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008604 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008605 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01008606 will be analyzed, extracted and stored in the table once a
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008607 server is selected.
8608
8609 <table> is an optional stickiness table name. If unspecified, the same
8610 backend's table is used. A stickiness table is declared using
8611 the "stick-table" statement.
8612
8613 <cond> is an optional storage condition. It makes it possible to store
8614 certain criteria only when some conditions are met (or not met).
8615 For instance, it could be used to store the source IP address
8616 except when the request passes through a known proxy, in which
8617 case we'd store a converted form of a header containing that IP
8618 address.
8619
8620 Some protocols or applications require complex stickiness rules and cannot
8621 always simply rely on cookies nor hashing. The "stick store-request" statement
8622 describes a rule to decide what to extract from the request and when to do
8623 it, in order to store it into a stickiness table for further requests to
8624 match it using the "stick match" statement. Obviously the extracted part must
8625 make sense and have a chance to be matched in a further request. Storing a
8626 client's IP address for instance often makes sense. Storing an ID found in a
8627 URL parameter also makes sense. Storing a source port will almost never make
8628 any sense because it will be randomly matched. See section 7 for a complete
8629 list of possible patterns and transformation rules.
8630
8631 The table has to be declared using the "stick-table" statement. It must be of
8632 a type compatible with the pattern. By default it is the one which is present
8633 in the same backend. It is possible to share a table with other backends by
8634 referencing it using the "table" keyword. If another table is referenced,
8635 the server's ID inside the backends are used. By default, all server IDs
8636 start at 1 in each backend, so the server ordering is enough. But in case of
8637 doubt, it is highly recommended to force server IDs using their "id" setting.
8638
8639 It is possible to restrict the conditions where a "stick store-request"
8640 statement will apply, using "if" or "unless" followed by a condition. This
8641 condition will be evaluated while parsing the request, so any criteria can be
8642 used. See section 7 for ACL based conditions.
8643
8644 There is no limit on the number of "stick store-request" statements, but
8645 there is a limit of 8 simultaneous stores per request or response. This
8646 makes it possible to store up to 8 criteria, all extracted from either the
8647 request or the response, regardless of the number of rules. Only the 8 first
8648 ones which match will be kept. Using this, it is possible to feed multiple
8649 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01008650 another protocol or access method. Using multiple store-request rules with
8651 the same table is possible and may be used to find the best criterion to rely
8652 on, by arranging the rules by decreasing preference order. Only the first
8653 extracted criterion for a given table will be stored. All subsequent store-
8654 request rules referencing the same table will be skipped and their ACLs will
8655 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008656
8657 The "store-request" rules are evaluated once the server connection has been
8658 established, so that the table will contain the real server that processed
8659 the request.
8660
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008661 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8662 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008663 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008664
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008665 Example :
8666 # forward SMTP users to the same server they just used for POP in the
8667 # last 30 minutes
8668 backend pop
8669 mode tcp
8670 balance roundrobin
8671 stick store-request src
8672 stick-table type ip size 200k expire 30m
8673 server s1 192.168.1.1:110
8674 server s2 192.168.1.1:110
8675
8676 backend smtp
8677 mode tcp
8678 balance roundrobin
8679 stick match src table pop
8680 server s1 192.168.1.1:25
8681 server s2 192.168.1.1:25
8682
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008683 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02008684 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008685
8686
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008687stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Emeric Brunf099e792010-09-27 12:05:28 +02008688 size <size> [expire <expire>] [nopurge] [peers <peersect>]
8689 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +08008690 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008691 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +02008692 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008693
8694 Arguments :
8695 ip a table declared with "type ip" will only store IPv4 addresses.
8696 This form is very compact (about 50 bytes per entry) and allows
8697 very fast entry lookup and stores with almost no overhead. This
8698 is mainly used to store client source IP addresses.
8699
David du Colombier9a6d3c92011-03-17 10:40:24 +01008700 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
8701 This form is very compact (about 60 bytes per entry) and allows
8702 very fast entry lookup and stores with almost no overhead. This
8703 is mainly used to store client source IP addresses.
8704
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008705 integer a table declared with "type integer" will store 32bit integers
8706 which can represent a client identifier found in a request for
8707 instance.
8708
8709 string a table declared with "type string" will store substrings of up
8710 to <len> characters. If the string provided by the pattern
8711 extractor is larger than <len>, it will be truncated before
8712 being stored. During matching, at most <len> characters will be
8713 compared between the string in the table and the extracted
8714 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008715 to 32 characters.
8716
8717 binary a table declared with "type binary" will store binary blocks
8718 of <len> bytes. If the block provided by the pattern
8719 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +02008720 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008721 is shorter than <len>, it will be padded by 0. When not
8722 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008723
8724 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008725 "string" type table (See type "string" above). Or the number
8726 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008727 changing this parameter as memory usage will proportionally
8728 increase.
8729
8730 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +01008731 value directly impacts memory usage. Count approximately
8732 50 bytes per entry, plus the size of a string if any. The size
8733 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008734
8735 [nopurge] indicates that we refuse to purge older entries when the table
8736 is full. When not specified and the table is full when haproxy
8737 wants to store an entry in it, it will flush a few of the oldest
8738 entries in order to release some space for the new ones. This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01008739 most often the desired behavior. In some specific cases, it
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008740 be desirable to refuse new entries instead of purging the older
8741 ones. That may be the case when the amount of data to store is
8742 far above the hardware limits and we prefer not to offer access
8743 to new clients than to reject the ones already connected. When
8744 using this parameter, be sure to properly set the "expire"
8745 parameter (see below).
8746
Emeric Brunf099e792010-09-27 12:05:28 +02008747 <peersect> is the name of the peers section to use for replication. Entries
8748 which associate keys to server IDs are kept synchronized with
8749 the remote peers declared in this section. All entries are also
8750 automatically learned from the local peer (old process) during a
8751 soft restart.
8752
Willy Tarreau1abc6732015-05-01 19:21:02 +02008753 NOTE : each peers section may be referenced only by tables
8754 belonging to the same unique process.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008755
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008756 <expire> defines the maximum duration of an entry in the table since it
8757 was last created, refreshed or matched. The expiration delay is
8758 defined using the standard time format, similarly as the various
8759 timeouts. The maximum duration is slightly above 24 days. See
Jarno Huuskonene0ee0be2017-07-04 10:35:12 +03008760 section 2.4 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +02008761 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008762 be removed once full. Be sure not to use the "nopurge" parameter
8763 if not expiration delay is specified.
8764
Willy Tarreau08d5f982010-06-06 13:34:54 +02008765 <data_type> is used to store additional information in the stick-table. This
8766 may be used by ACLs in order to control various criteria related
8767 to the activity of the client matching the stick-table. For each
8768 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +02008769 that the additional data can fit. Several data types may be
8770 stored with an entry. Multiple data types may be specified after
8771 the "store" keyword, as a comma-separated list. Alternatively,
8772 it is possible to repeat the "store" keyword followed by one or
8773 several data types. Except for the "server_id" type which is
8774 automatically detected and enabled, all data types must be
8775 explicitly declared to be stored. If an ACL references a data
8776 type which is not stored, the ACL will simply not match. Some
8777 data types require an argument which must be passed just after
8778 the type between parenthesis. See below for the supported data
8779 types and their arguments.
8780
8781 The data types that can be stored with an entry are the following :
8782 - server_id : this is an integer which holds the numeric ID of the server a
8783 request was assigned to. It is used by the "stick match", "stick store",
8784 and "stick on" rules. It is automatically enabled when referenced.
8785
8786 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
8787 integer which may be used for anything. Most of the time it will be used
8788 to put a special tag on some entries, for instance to note that a
Davor Ocelice9ed2812017-12-25 17:49:28 +01008789 specific behavior was detected and must be known for future matches.
Willy Tarreauc9705a12010-07-27 20:05:50 +02008790
Willy Tarreauba2ffd12013-05-29 15:54:14 +02008791 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
8792 over a period. It is a positive 32-bit integer integer which may be used
8793 for anything. Just like <gpc0>, it counts events, but instead of keeping
Davor Ocelice9ed2812017-12-25 17:49:28 +01008794 a cumulative number, it maintains the rate at which the counter is
Willy Tarreauba2ffd12013-05-29 15:54:14 +02008795 incremented. Most of the time it will be used to measure the frequency of
Davor Ocelice9ed2812017-12-25 17:49:28 +01008796 occurrence of certain events (e.g. requests to a specific URL).
Willy Tarreauba2ffd12013-05-29 15:54:14 +02008797
Frédéric Lécaille6778b272018-01-29 15:22:53 +01008798 - gpc1 : second General Purpose Counter. It is a positive 32-bit integer
8799 integer which may be used for anything. Most of the time it will be used
8800 to put a special tag on some entries, for instance to note that a
8801 specific behavior was detected and must be known for future matches.
8802
8803 - gpc1_rate(<period>) : increment rate of the second General Purpose Counter
8804 over a period. It is a positive 32-bit integer integer which may be used
8805 for anything. Just like <gpc1>, it counts events, but instead of keeping
8806 a cumulative number, it maintains the rate at which the counter is
8807 incremented. Most of the time it will be used to measure the frequency of
8808 occurrence of certain events (e.g. requests to a specific URL).
8809
Willy Tarreauc9705a12010-07-27 20:05:50 +02008810 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
8811 the absolute number of connections received from clients which matched
8812 this entry. It does not mean the connections were accepted, just that
8813 they were received.
8814
8815 - conn_cur : Current Connections. It is a positive 32-bit integer which
8816 stores the concurrent connection counts for the entry. It is incremented
8817 once an incoming connection matches the entry, and decremented once the
8818 connection leaves. That way it is possible to know at any time the exact
8819 number of concurrent connections for an entry.
8820
8821 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8822 integer parameter <period> which indicates in milliseconds the length
8823 of the period over which the average is measured. It reports the average
8824 incoming connection rate over that period, in connections per period. The
8825 result is an integer which can be matched using ACLs.
8826
8827 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
8828 the absolute number of sessions received from clients which matched this
8829 entry. A session is a connection that was accepted by the layer 4 rules.
8830
8831 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8832 integer parameter <period> which indicates in milliseconds the length
8833 of the period over which the average is measured. It reports the average
8834 incoming session rate over that period, in sessions per period. The
8835 result is an integer which can be matched using ACLs.
8836
8837 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
8838 counts the absolute number of HTTP requests received from clients which
8839 matched this entry. It does not matter whether they are valid requests or
8840 not. Note that this is different from sessions when keep-alive is used on
8841 the client side.
8842
8843 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8844 integer parameter <period> which indicates in milliseconds the length
8845 of the period over which the average is measured. It reports the average
8846 HTTP request rate over that period, in requests per period. The result is
8847 an integer which can be matched using ACLs. It does not matter whether
8848 they are valid requests or not. Note that this is different from sessions
8849 when keep-alive is used on the client side.
8850
8851 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
8852 counts the absolute number of HTTP requests errors induced by clients
8853 which matched this entry. Errors are counted on invalid and truncated
8854 requests, as well as on denied or tarpitted requests, and on failed
8855 authentications. If the server responds with 4xx, then the request is
8856 also counted as an error since it's an error triggered by the client
Davor Ocelice9ed2812017-12-25 17:49:28 +01008857 (e.g. vulnerability scan).
Willy Tarreauc9705a12010-07-27 20:05:50 +02008858
8859 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8860 integer parameter <period> which indicates in milliseconds the length
8861 of the period over which the average is measured. It reports the average
8862 HTTP request error rate over that period, in requests per period (see
8863 http_err_cnt above for what is accounted as an error). The result is an
8864 integer which can be matched using ACLs.
8865
8866 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +01008867 integer which counts the cumulative number of bytes received from clients
Willy Tarreauc9705a12010-07-27 20:05:50 +02008868 which matched this entry. Headers are included in the count. This may be
8869 used to limit abuse of upload features on photo or video servers.
8870
8871 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8872 integer parameter <period> which indicates in milliseconds the length
8873 of the period over which the average is measured. It reports the average
8874 incoming bytes rate over that period, in bytes per period. It may be used
8875 to detect users which upload too much and too fast. Warning: with large
8876 uploads, it is possible that the amount of uploaded data will be counted
8877 once upon termination, thus causing spikes in the average transfer speed
8878 instead of having a smooth one. This may partially be smoothed with
8879 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
8880 recommended for better fairness.
8881
8882 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +01008883 integer which counts the cumulative number of bytes sent to clients which
Willy Tarreauc9705a12010-07-27 20:05:50 +02008884 matched this entry. Headers are included in the count. This may be used
8885 to limit abuse of bots sucking the whole site.
8886
8887 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
8888 an integer parameter <period> which indicates in milliseconds the length
8889 of the period over which the average is measured. It reports the average
8890 outgoing bytes rate over that period, in bytes per period. It may be used
8891 to detect users which download too much and too fast. Warning: with large
8892 transfers, it is possible that the amount of transferred data will be
8893 counted once upon termination, thus causing spikes in the average
8894 transfer speed instead of having a smooth one. This may partially be
8895 smoothed with "option contstats" though this is not perfect yet. Use of
8896 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +02008897
Willy Tarreauc00cdc22010-06-06 16:48:26 +02008898 There is only one stick-table per proxy. At the moment of writing this doc,
8899 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008900 to be required, simply create a dummy backend with a stick-table in it and
8901 reference it.
8902
8903 It is important to understand that stickiness based on learning information
8904 has some limitations, including the fact that all learned associations are
Baptiste Assmann123ff042016-03-06 23:29:28 +01008905 lost upon restart unless peers are properly configured to transfer such
8906 information upon restart (recommended). In general it can be good as a
8907 complement but not always as an exclusive stickiness.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008908
Willy Tarreauc9705a12010-07-27 20:05:50 +02008909 Last, memory requirements may be important when storing many data types.
8910 Indeed, storing all indicators above at once in each entry requires 116 bytes
8911 per entry, or 116 MB for a 1-million entries table. This is definitely not
8912 something that can be ignored.
8913
8914 Example:
8915 # Keep track of counters of up to 1 million IP addresses over 5 minutes
8916 # and store a general purpose counter and the average connection rate
8917 # computed over a sliding window of 30 seconds.
8918 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
8919
Jarno Huuskonene0ee0be2017-07-04 10:35:12 +03008920 See also : "stick match", "stick on", "stick store-request", section 2.4
David du Colombiera13d1b92011-03-17 10:40:22 +01008921 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008922
8923
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008924stick store-response <pattern> [table <table>] [{if | unless} <condition>]
Baptiste Assmann2f2d2ec2016-03-06 23:27:24 +01008925 Define a response pattern used to create an entry in a stickiness table
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008926 May be used in sections : defaults | frontend | listen | backend
8927 no | no | yes | yes
8928
8929 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008930 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008931 describes what elements of the response or connection will
Davor Ocelice9ed2812017-12-25 17:49:28 +01008932 be analyzed, extracted and stored in the table once a
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008933 server is selected.
8934
8935 <table> is an optional stickiness table name. If unspecified, the same
8936 backend's table is used. A stickiness table is declared using
8937 the "stick-table" statement.
8938
8939 <cond> is an optional storage condition. It makes it possible to store
8940 certain criteria only when some conditions are met (or not met).
8941 For instance, it could be used to store the SSL session ID only
8942 when the response is a SSL server hello.
8943
8944 Some protocols or applications require complex stickiness rules and cannot
8945 always simply rely on cookies nor hashing. The "stick store-response"
8946 statement describes a rule to decide what to extract from the response and
8947 when to do it, in order to store it into a stickiness table for further
8948 requests to match it using the "stick match" statement. Obviously the
8949 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008950 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008951 See section 7 for a complete list of possible patterns and transformation
8952 rules.
8953
8954 The table has to be declared using the "stick-table" statement. It must be of
8955 a type compatible with the pattern. By default it is the one which is present
8956 in the same backend. It is possible to share a table with other backends by
8957 referencing it using the "table" keyword. If another table is referenced,
8958 the server's ID inside the backends are used. By default, all server IDs
8959 start at 1 in each backend, so the server ordering is enough. But in case of
8960 doubt, it is highly recommended to force server IDs using their "id" setting.
8961
8962 It is possible to restrict the conditions where a "stick store-response"
8963 statement will apply, using "if" or "unless" followed by a condition. This
8964 condition will be evaluated while parsing the response, so any criteria can
8965 be used. See section 7 for ACL based conditions.
8966
8967 There is no limit on the number of "stick store-response" statements, but
8968 there is a limit of 8 simultaneous stores per request or response. This
8969 makes it possible to store up to 8 criteria, all extracted from either the
8970 request or the response, regardless of the number of rules. Only the 8 first
8971 ones which match will be kept. Using this, it is possible to feed multiple
8972 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01008973 another protocol or access method. Using multiple store-response rules with
8974 the same table is possible and may be used to find the best criterion to rely
8975 on, by arranging the rules by decreasing preference order. Only the first
8976 extracted criterion for a given table will be stored. All subsequent store-
8977 response rules referencing the same table will be skipped and their ACLs will
8978 not be evaluated. However, even if a store-request rule references a table, a
8979 store-response rule may also use the same table. This means that each table
8980 may learn exactly one element from the request and one element from the
8981 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008982
8983 The table will contain the real server that processed the request.
8984
8985 Example :
8986 # Learn SSL session ID from both request and response and create affinity.
8987 backend https
8988 mode tcp
8989 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +02008990 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008991 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008992
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008993 acl clienthello req_ssl_hello_type 1
8994 acl serverhello rep_ssl_hello_type 2
8995
8996 # use tcp content accepts to detects ssl client and server hello.
8997 tcp-request inspect-delay 5s
8998 tcp-request content accept if clienthello
8999
9000 # no timeout on response inspect delay by default.
9001 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009002
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009003 # SSL session ID (SSLID) may be present on a client or server hello.
9004 # Its length is coded on 1 byte at offset 43 and its value starts
9005 # at offset 44.
9006
9007 # Match and learn on request if client hello.
9008 stick on payload_lv(43,1) if clienthello
9009
9010 # Learn on response if server hello.
9011 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +02009012
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009013 server s1 192.168.1.1:443
9014 server s2 192.168.1.1:443
9015
9016 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
9017 extraction.
9018
9019
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009020tcp-check connect [params*]
9021 Opens a new connection
9022 May be used in sections: defaults | frontend | listen | backend
9023 no | no | yes | yes
9024
9025 When an application lies on more than a single TCP port or when HAProxy
9026 load-balance many services in a single backend, it makes sense to probe all
9027 the services individually before considering a server as operational.
9028
9029 When there are no TCP port configured on the server line neither server port
9030 directive, then the 'tcp-check connect port <port>' must be the first step
9031 of the sequence.
9032
9033 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
9034 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
9035 do.
9036
9037 Parameters :
9038 They are optional and can be used to describe how HAProxy should open and
9039 use the TCP connection.
9040
9041 port if not set, check port or server port is used.
9042 It tells HAProxy where to open the connection to.
9043 <port> must be a valid TCP port source integer, from 1 to 65535.
9044
9045 send-proxy send a PROXY protocol string
9046
9047 ssl opens a ciphered connection
9048
9049 Examples:
9050 # check HTTP and HTTPs services on a server.
9051 # first open port 80 thanks to server line port directive, then
9052 # tcp-check opens port 443, ciphered and run a request on it:
9053 option tcp-check
9054 tcp-check connect
9055 tcp-check send GET\ /\ HTTP/1.0\r\n
9056 tcp-check send Host:\ haproxy.1wt.eu\r\n
9057 tcp-check send \r\n
9058 tcp-check expect rstring (2..|3..)
9059 tcp-check connect port 443 ssl
9060 tcp-check send GET\ /\ HTTP/1.0\r\n
9061 tcp-check send Host:\ haproxy.1wt.eu\r\n
9062 tcp-check send \r\n
9063 tcp-check expect rstring (2..|3..)
9064 server www 10.0.0.1 check port 80
9065
9066 # check both POP and IMAP from a single server:
9067 option tcp-check
9068 tcp-check connect port 110
9069 tcp-check expect string +OK\ POP3\ ready
9070 tcp-check connect port 143
9071 tcp-check expect string *\ OK\ IMAP4\ ready
9072 server mail 10.0.0.1 check
9073
9074 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
9075
9076
9077tcp-check expect [!] <match> <pattern>
Davor Ocelice9ed2812017-12-25 17:49:28 +01009078 Specify data to be collected and analyzed during a generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009079 May be used in sections: defaults | frontend | listen | backend
9080 no | no | yes | yes
9081
9082 Arguments :
9083 <match> is a keyword indicating how to look for a specific pattern in the
9084 response. The keyword may be one of "string", "rstring" or
9085 binary.
9086 The keyword may be preceded by an exclamation mark ("!") to negate
9087 the match. Spaces are allowed between the exclamation mark and the
9088 keyword. See below for more details on the supported keywords.
9089
9090 <pattern> is the pattern to look for. It may be a string or a regular
9091 expression. If the pattern contains spaces, they must be escaped
9092 with the usual backslash ('\').
9093 If the match is set to binary, then the pattern must be passed as
Davor Ocelice9ed2812017-12-25 17:49:28 +01009094 a series of hexadecimal digits in an even number. Each sequence of
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009095 two digits will represent a byte. The hexadecimal digits may be
9096 used upper or lower case.
9097
9098
9099 The available matches are intentionally similar to their http-check cousins :
9100
9101 string <string> : test the exact string matches in the response buffer.
9102 A health check response will be considered valid if the
9103 response's buffer contains this exact string. If the
9104 "string" keyword is prefixed with "!", then the response
9105 will be considered invalid if the body contains this
9106 string. This can be used to look for a mandatory pattern
9107 in a protocol response, or to detect a failure when a
9108 specific error appears in a protocol banner.
9109
9110 rstring <regex> : test a regular expression on the response buffer.
9111 A health check response will be considered valid if the
9112 response's buffer matches this expression. If the
9113 "rstring" keyword is prefixed with "!", then the response
9114 will be considered invalid if the body matches the
9115 expression.
9116
9117 binary <hexstring> : test the exact string in its hexadecimal form matches
9118 in the response buffer. A health check response will
9119 be considered valid if the response's buffer contains
9120 this exact hexadecimal string.
9121 Purpose is to match data on binary protocols.
9122
9123 It is important to note that the responses will be limited to a certain size
9124 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
9125 Thus, too large responses may not contain the mandatory pattern when using
9126 "string", "rstring" or binary. If a large response is absolutely required, it
9127 is possible to change the default max size by setting the global variable.
9128 However, it is worth keeping in mind that parsing very large responses can
9129 waste some CPU cycles, especially when regular expressions are used, and that
9130 it is always better to focus the checks on smaller resources. Also, in its
9131 current state, the check will not find any string nor regex past a null
9132 character in the response. Similarly it is not possible to request matching
9133 the null character.
9134
9135 Examples :
9136 # perform a POP check
9137 option tcp-check
9138 tcp-check expect string +OK\ POP3\ ready
9139
9140 # perform an IMAP check
9141 option tcp-check
9142 tcp-check expect string *\ OK\ IMAP4\ ready
9143
9144 # look for the redis master server
9145 option tcp-check
9146 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02009147 tcp-check expect string +PONG
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009148 tcp-check send info\ replication\r\n
9149 tcp-check expect string role:master
9150 tcp-check send QUIT\r\n
9151 tcp-check expect string +OK
9152
9153
9154 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
9155 "tcp-check send-binary", "http-check expect", tune.chksize
9156
9157
9158tcp-check send <data>
9159 Specify a string to be sent as a question during a generic health check
9160 May be used in sections: defaults | frontend | listen | backend
9161 no | no | yes | yes
9162
9163 <data> : the data to be sent as a question during a generic health check
9164 session. For now, <data> must be a string.
9165
9166 Examples :
9167 # look for the redis master server
9168 option tcp-check
9169 tcp-check send info\ replication\r\n
9170 tcp-check expect string role:master
9171
9172 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
9173 "tcp-check send-binary", tune.chksize
9174
9175
Davor Ocelice9ed2812017-12-25 17:49:28 +01009176tcp-check send-binary <hexstring>
9177 Specify a hex digits string to be sent as a binary question during a raw
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009178 tcp health check
9179 May be used in sections: defaults | frontend | listen | backend
9180 no | no | yes | yes
9181
9182 <data> : the data to be sent as a question during a generic health check
9183 session. For now, <data> must be a string.
Davor Ocelice9ed2812017-12-25 17:49:28 +01009184 <hexstring> : test the exact string in its hexadecimal form matches in the
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009185 response buffer. A health check response will be considered
9186 valid if the response's buffer contains this exact
9187 hexadecimal string.
9188 Purpose is to send binary data to ask on binary protocols.
9189
9190 Examples :
9191 # redis check in binary
9192 option tcp-check
9193 tcp-check send-binary 50494e470d0a # PING\r\n
9194 tcp-check expect binary 2b504F4e47 # +PONG
9195
9196
9197 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
9198 "tcp-check send", tune.chksize
9199
9200
Willy Tarreaue9656522010-08-17 15:40:09 +02009201tcp-request connection <action> [{if | unless} <condition>]
9202 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +02009203 May be used in sections : defaults | frontend | listen | backend
9204 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +02009205 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009206 <action> defines the action to perform if the condition applies. See
9207 below.
Willy Tarreau1a687942010-05-23 22:40:30 +02009208
Willy Tarreaue9656522010-08-17 15:40:09 +02009209 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009210
9211 Immediately after acceptance of a new incoming connection, it is possible to
9212 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +02009213 or dropped or have its counters tracked. Those conditions cannot make use of
9214 any data contents because the connection has not been read from yet, and the
9215 buffers are not yet allocated. This is used to selectively and very quickly
9216 accept or drop connections from various sources with a very low overhead. If
9217 some contents need to be inspected in order to take the decision, the
9218 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009219
Willy Tarreaue9656522010-08-17 15:40:09 +02009220 The "tcp-request connection" rules are evaluated in their exact declaration
9221 order. If no rule matches or if there is no rule, the default action is to
9222 accept the incoming connection. There is no specific limit to the number of
9223 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009224
Willy Tarreaua9083d02015-05-08 15:27:59 +02009225 Four types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +02009226 - accept :
9227 accepts the connection if the condition is true (when used with "if")
9228 or false (when used with "unless"). The first such rule executed ends
9229 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009230
Willy Tarreaue9656522010-08-17 15:40:09 +02009231 - reject :
9232 rejects the connection if the condition is true (when used with "if")
9233 or false (when used with "unless"). The first such rule executed ends
9234 the rules evaluation. Rejected connections do not even become a
9235 session, which is why they are accounted separately for in the stats,
9236 as "denied connections". They are not considered for the session
9237 rate-limit and are not logged either. The reason is that these rules
9238 should only be used to filter extremely high connection rates such as
9239 the ones encountered during a massive DDoS attack. Under these extreme
9240 conditions, the simple action of logging each event would make the
9241 system collapse and would considerably lower the filtering capacity. If
9242 logging is absolutely desired, then "tcp-request content" rules should
Willy Tarreau4f614292016-10-21 17:49:36 +02009243 be used instead, as "tcp-request session" rules will not log either.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009244
Willy Tarreau4f0d9192013-06-11 20:40:55 +02009245 - expect-proxy layer4 :
9246 configures the client-facing connection to receive a PROXY protocol
9247 header before any byte is read from the socket. This is equivalent to
9248 having the "accept-proxy" keyword on the "bind" line, except that using
9249 the TCP rule allows the PROXY protocol to be accepted only for certain
9250 IP address ranges using an ACL. This is convenient when multiple layers
9251 of load balancers are passed through by traffic coming from public
9252 hosts.
9253
Bertrand Jacquin90759682016-06-06 15:35:39 +01009254 - expect-netscaler-cip layer4 :
9255 configures the client-facing connection to receive a NetScaler Client
9256 IP insertion protocol header before any byte is read from the socket.
9257 This is equivalent to having the "accept-netscaler-cip" keyword on the
9258 "bind" line, except that using the TCP rule allows the PROXY protocol
9259 to be accepted only for certain IP address ranges using an ACL. This
9260 is convenient when multiple layers of load balancers are passed
9261 through by traffic coming from public hosts.
9262
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009263 - capture <sample> len <length> :
9264 This only applies to "tcp-request content" rules. It captures sample
9265 expression <sample> from the request buffer, and converts it to a
9266 string of at most <len> characters. The resulting string is stored into
9267 the next request "capture" slot, so it will possibly appear next to
9268 some captured HTTP headers. It will then automatically appear in the
9269 logs, and it will be possible to extract it using sample fetch rules to
9270 feed it into headers or anything. The length should be limited given
9271 that this size will be allocated for each capture during the whole
Willy Tarreaua9083d02015-05-08 15:27:59 +02009272 session life. Please check section 7.3 (Fetching samples) and "capture
9273 request header" for more information.
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009274
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009275 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +02009276 enables tracking of sticky counters from current connection. These
Willy Tarreau09448f72014-06-25 18:12:15 +02009277 rules do not stop evaluation and do not change default action. 3 sets
Willy Tarreaue9656522010-08-17 15:40:09 +02009278 of counters may be simultaneously tracked by the same connection. The
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009279 first "track-sc0" rule executed enables tracking of the counters of the
9280 specified table as the first set. The first "track-sc1" rule executed
Willy Tarreaue9656522010-08-17 15:40:09 +02009281 enables tracking of the counters of the specified table as the second
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009282 set. The first "track-sc2" rule executed enables tracking of the
9283 counters of the specified table as the third set. It is a recommended
9284 practice to use the first set of counters for the per-frontend counters
9285 and the second set for the per-backend ones. But this is just a
9286 guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009287
Willy Tarreaue9656522010-08-17 15:40:09 +02009288 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02009289 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +02009290 in section 7.3. It describes what elements of the incoming
Davor Ocelice9ed2812017-12-25 17:49:28 +01009291 request or connection will be analyzed, extracted, combined,
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009292 and used to select which table entry to update the counters.
9293 Note that "tcp-request connection" cannot use content-based
9294 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009295
Willy Tarreaue9656522010-08-17 15:40:09 +02009296 <table> is an optional table to be used instead of the default one,
9297 which is the stick-table declared in the current proxy. All
9298 the counters for the matches and updates for the key will
9299 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009300
Willy Tarreaue9656522010-08-17 15:40:09 +02009301 Once a "track-sc*" rule is executed, the key is looked up in the table
9302 and if it is not found, an entry is allocated for it. Then a pointer to
9303 that entry is kept during all the session's life, and this entry's
9304 counters are updated as often as possible, every time the session's
9305 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009306 Counters are only updated for events that happen after the tracking has
9307 been started. For example, connection counters will not be updated when
9308 tracking layer 7 information, since the connection event happens before
9309 layer7 information is extracted.
9310
Willy Tarreaue9656522010-08-17 15:40:09 +02009311 If the entry tracks concurrent connection counters, one connection is
9312 counted for as long as the entry is tracked, and the entry will not
9313 expire during that time. Tracking counters also provides a performance
9314 advantage over just checking the keys, because only one table lookup is
9315 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009316
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009317 - sc-inc-gpc0(<sc-id>):
9318 The "sc-inc-gpc0" increments the GPC0 counter according to the sticky
9319 counter designated by <sc-id>. If an error occurs, this action silently
9320 fails and the actions evaluation continues.
9321
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009322 - sc-inc-gpc1(<sc-id>):
9323 The "sc-inc-gpc1" increments the GPC1 counter according to the sticky
9324 counter designated by <sc-id>. If an error occurs, this action silently
9325 fails and the actions evaluation continues.
9326
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009327 - sc-set-gpt0(<sc-id>) <int>:
9328 This action sets the GPT0 tag according to the sticky counter designated
9329 by <sc-id> and the value of <int>. The expected result is a boolean. If
9330 an error occurs, this action silently fails and the actions evaluation
9331 continues.
9332
William Lallemand2e785f22016-05-25 01:48:42 +02009333 - set-src <expr> :
9334 Is used to set the source IP address to the value of specified
9335 expression. Useful if you want to mask source IP for privacy.
9336 If you want to provide an IP from a HTTP header use "http-request
9337 set-src"
9338
9339 <expr> Is a standard HAProxy expression formed by a sample-fetch
9340 followed by some converters.
9341
9342 Example:
9343
9344 tcp-request connection set-src src,ipmask(24)
9345
Willy Tarreau0c630532016-10-21 17:52:58 +02009346 When possible, set-src preserves the original source port as long as the
9347 address family allows it, otherwise the source port is set to 0.
William Lallemand2e785f22016-05-25 01:48:42 +02009348
William Lallemand44be6402016-05-25 01:51:35 +02009349 - set-src-port <expr> :
9350 Is used to set the source port address to the value of specified
9351 expression.
9352
9353 <expr> Is a standard HAProxy expression formed by a sample-fetch
9354 followed by some converters.
9355
9356 Example:
9357
9358 tcp-request connection set-src-port int(4000)
9359
Willy Tarreau0c630532016-10-21 17:52:58 +02009360 When possible, set-src-port preserves the original source address as long
9361 as the address family supports a port, otherwise it forces the source
9362 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02009363
William Lallemand13e9b0c2016-05-25 02:34:07 +02009364 - set-dst <expr> :
9365 Is used to set the destination IP address to the value of specified
9366 expression. Useful if you want to mask IP for privacy in log.
9367 If you want to provide an IP from a HTTP header use "http-request
9368 set-dst". If you want to connect to the new address/port, use
9369 '0.0.0.0:0' as a server address in the backend.
9370
9371 <expr> Is a standard HAProxy expression formed by a sample-fetch
9372 followed by some converters.
9373
9374 Example:
9375
9376 tcp-request connection set-dst dst,ipmask(24)
9377 tcp-request connection set-dst ipv4(10.0.0.1)
9378
Willy Tarreau0c630532016-10-21 17:52:58 +02009379 When possible, set-dst preserves the original destination port as long as
9380 the address family allows it, otherwise the destination port is set to 0.
9381
William Lallemand13e9b0c2016-05-25 02:34:07 +02009382 - set-dst-port <expr> :
9383 Is used to set the destination port address to the value of specified
9384 expression. If you want to connect to the new address/port, use
9385 '0.0.0.0:0' as a server address in the backend.
9386
9387
9388 <expr> Is a standard HAProxy expression formed by a sample-fetch
9389 followed by some converters.
9390
9391 Example:
9392
9393 tcp-request connection set-dst-port int(4000)
9394
Willy Tarreau0c630532016-10-21 17:52:58 +02009395 When possible, set-dst-port preserves the original destination address as
9396 long as the address family supports a port, otherwise it forces the
9397 destination address to IPv4 "0.0.0.0" before rewriting the port.
9398
Willy Tarreau2d392c22015-08-24 01:43:45 +02009399 - "silent-drop" :
9400 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +01009401 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +02009402 to prevent the client from being notified. The effect it then that the
9403 client still sees an established connection while there's none on
9404 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
9405 except that it doesn't use any local resource at all on the machine
9406 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +01009407 slow down stronger attackers. It is important to understand the impact
9408 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009409 client and HAProxy (firewalls, proxies, load balancers) will also keep
9410 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +01009411 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009412 TCP_REPAIR socket option is used to block the emission of a TCP
9413 reset. On other systems, the socket's TTL is reduced to 1 so that the
9414 TCP reset doesn't pass the first router, though it's still delivered to
9415 local networks. Do not use it unless you fully understand how it works.
9416
Willy Tarreaue9656522010-08-17 15:40:09 +02009417 Note that the "if/unless" condition is optional. If no condition is set on
9418 the action, it is simply performed unconditionally. That can be useful for
9419 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009420
Willy Tarreaue9656522010-08-17 15:40:09 +02009421 Example: accept all connections from white-listed hosts, reject too fast
9422 connection without counting them, and track accepted connections.
9423 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009424
Willy Tarreaue9656522010-08-17 15:40:09 +02009425 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009426 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009427 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009428
Willy Tarreaue9656522010-08-17 15:40:09 +02009429 Example: accept all connections from white-listed hosts, count all other
9430 connections and reject too fast ones. This results in abusive ones
9431 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009432
Willy Tarreaue9656522010-08-17 15:40:09 +02009433 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009434 tcp-request connection track-sc0 src
9435 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009436
Willy Tarreau4f0d9192013-06-11 20:40:55 +02009437 Example: enable the PROXY protocol for traffic coming from all known proxies.
9438
9439 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
9440
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009441 See section 7 about ACL usage.
9442
Willy Tarreau4f614292016-10-21 17:49:36 +02009443 See also : "tcp-request session", "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009444
9445
Willy Tarreaue9656522010-08-17 15:40:09 +02009446tcp-request content <action> [{if | unless} <condition>]
9447 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +02009448 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02009449 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +02009450 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009451 <action> defines the action to perform if the condition applies. See
9452 below.
Willy Tarreau62644772008-07-16 18:36:06 +02009453
Willy Tarreaue9656522010-08-17 15:40:09 +02009454 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +02009455
Davor Ocelice9ed2812017-12-25 17:49:28 +01009456 A request's contents can be analyzed at an early stage of request processing
Willy Tarreaue9656522010-08-17 15:40:09 +02009457 called "TCP content inspection". During this stage, ACL-based rules are
9458 evaluated every time the request contents are updated, until either an
9459 "accept" or a "reject" rule matches, or the TCP request inspection delay
9460 expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +02009461
Willy Tarreaue9656522010-08-17 15:40:09 +02009462 The first difference between these rules and "tcp-request connection" rules
9463 is that "tcp-request content" rules can make use of contents to take a
9464 decision. Most often, these decisions will consider a protocol recognition or
9465 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +01009466 both frontends and backends. In case of HTTP keep-alive with the client, all
9467 tcp-request content rules are evaluated again, so haproxy keeps a record of
9468 what sticky counters were assigned by a "tcp-request connection" versus a
9469 "tcp-request content" rule, and flushes all the content-related ones after
9470 processing an HTTP request, so that they may be evaluated again by the rules
9471 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009472 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +01009473 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +02009474
Willy Tarreaue9656522010-08-17 15:40:09 +02009475 Content-based rules are evaluated in their exact declaration order. If no
9476 rule matches or if there is no rule, the default action is to accept the
9477 contents. There is no specific limit to the number of rules which may be
9478 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +02009479
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009480 Several types of actions are supported :
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009481 - accept : the request is accepted
9482 - reject : the request is rejected and the connection is closed
9483 - capture : the specified sample expression is captured
Patrick Hemmer268a7072018-05-11 12:52:31 -04009484 - set-priority-class <expr> | set-priority-offset <expr>
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009485 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009486 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009487 - sc-inc-gpc1(<sc-id>)
Thierry Fournierb9125672016-03-29 19:34:37 +02009488 - sc-set-gpt0(<sc-id>) <int>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009489 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +01009490 - unset-var(<var-name>)
Willy Tarreau2d392c22015-08-24 01:43:45 +02009491 - silent-drop
Davor Ocelice9ed2812017-12-25 17:49:28 +01009492 - send-spoe-group <engine-name> <group-name>
Willy Tarreau62644772008-07-16 18:36:06 +02009493
Willy Tarreaue9656522010-08-17 15:40:09 +02009494 They have the same meaning as their counter-parts in "tcp-request connection"
9495 so please refer to that section for a complete description.
Willy Tarreau62644772008-07-16 18:36:06 +02009496
Willy Tarreauf3338342014-01-28 21:40:28 +01009497 While there is nothing mandatory about it, it is recommended to use the
9498 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
9499 content" rules in the frontend, and track-sc2 for "tcp-request content"
9500 rules in the backend, because that makes the configuration more readable
9501 and easier to troubleshoot, but this is just a guideline and all counters
9502 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +02009503
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009504 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +02009505 the action, it is simply performed unconditionally. That can be useful for
9506 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +02009507
Willy Tarreaue9656522010-08-17 15:40:09 +02009508 It is perfectly possible to match layer 7 contents with "tcp-request content"
Willy Tarreauc0239e02012-04-16 14:42:55 +02009509 rules, since HTTP-specific ACL matches are able to preliminarily parse the
9510 contents of a buffer before extracting the required data. If the buffered
9511 contents do not parse as a valid HTTP message, then the ACL does not match.
9512 The parser which is involved there is exactly the same as for all other HTTP
Willy Tarreauf3338342014-01-28 21:40:28 +01009513 processing, so there is no risk of parsing something differently. In an HTTP
9514 backend connected to from an HTTP frontend, it is guaranteed that HTTP
9515 contents will always be immediately present when the rule is evaluated first.
Willy Tarreau62644772008-07-16 18:36:06 +02009516
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009517 Tracking layer7 information is also possible provided that the information
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009518 are present when the rule is processed. The rule processing engine is able to
9519 wait until the inspect delay expires when the data to be tracked is not yet
9520 available.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009521
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009522 The "set-var" is used to set the content of a variable. The variable is
Willy Tarreau4f614292016-10-21 17:49:36 +02009523 declared inline. For "tcp-request session" rules, only session-level
9524 variables can be used, without any layer7 contents.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009525
Daniel Schneller0b547052016-03-21 20:46:57 +01009526 <var-name> The name of the variable starts with an indication about
9527 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01009528 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01009529 "sess" : the variable is shared with the whole session
9530 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009531 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01009532 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009533 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01009534 "res" : the variable is shared only during response
9535 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009536 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01009537 The name may only contain characters 'a-z', 'A-Z', '0-9',
9538 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009539
9540 <expr> Is a standard HAProxy expression formed by a sample-fetch
9541 followed by some converters.
9542
Christopher Faulet85d79c92016-11-09 16:54:56 +01009543 The "unset-var" is used to unset a variable. See above for details about
9544 <var-name>.
9545
Patrick Hemmer268a7072018-05-11 12:52:31 -04009546 The "set-priority-class" is used to set the queue priority class of the
9547 current request. The value must be a sample expression which converts to an
9548 integer in the range -2047..2047. Results outside this range will be
9549 truncated. The priority class determines the order in which queued requests
9550 are processed. Lower values have higher priority.
9551
9552 The "set-priority-offset" is used to set the queue priority timestamp offset
9553 of the current request. The value must be a sample expression which converts
9554 to an integer in the range -524287..524287. Results outside this range will be
9555 truncated. When a request is queued, it is ordered first by the priority
9556 class, then by the current timestamp adjusted by the given offset in
9557 milliseconds. Lower values have higher priority.
9558 Note that the resulting timestamp is is only tracked with enough precision for
9559 524,287ms (8m44s287ms). If the request is queued long enough to where the
9560 adjusted timestamp exceeds this value, it will be misidentified as highest
9561 priority. Thus it is important to set "timeout queue" to a value, where when
9562 combined with the offset, does not exceed this limit.
9563
Christopher Faulet76c09ef2017-09-21 11:03:52 +02009564 The "send-spoe-group" is used to trigger sending of a group of SPOE
9565 messages. To do so, the SPOE engine used to send messages must be defined, as
9566 well as the SPOE group to send. Of course, the SPOE engine must refer to an
9567 existing SPOE filter. If not engine name is provided on the SPOE filter line,
9568 the SPOE agent name must be used.
9569
9570 <engine-name> The SPOE engine name.
9571
9572 <group-name> The SPOE group name as specified in the engine configuration.
9573
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009574 Example:
9575
9576 tcp-request content set-var(sess.my_var) src
Christopher Faulet85d79c92016-11-09 16:54:56 +01009577 tcp-request content unset-var(sess.my_var2)
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009578
Willy Tarreau62644772008-07-16 18:36:06 +02009579 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +02009580 # Accept HTTP requests containing a Host header saying "example.com"
9581 # and reject everything else.
9582 acl is_host_com hdr(Host) -i example.com
9583 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +02009584 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +02009585 tcp-request content reject
9586
9587 Example:
Willy Tarreau62644772008-07-16 18:36:06 +02009588 # reject SMTP connection if client speaks first
9589 tcp-request inspect-delay 30s
9590 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009591 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +02009592
9593 # Forward HTTPS connection only if client speaks
9594 tcp-request inspect-delay 30s
9595 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009596 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +02009597 tcp-request content reject
9598
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009599 Example:
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03009600 # Track the last IP(stick-table type string) from X-Forwarded-For
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009601 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009602 tcp-request content track-sc0 hdr(x-forwarded-for,-1)
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03009603 # Or track the last IP(stick-table type ip|ipv6) from X-Forwarded-For
9604 tcp-request content track-sc0 req.hdr_ip(x-forwarded-for,-1)
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009605
9606 Example:
9607 # track request counts per "base" (concatenation of Host+URL)
9608 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009609 tcp-request content track-sc0 base table req-rate
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009610
Willy Tarreaue9656522010-08-17 15:40:09 +02009611 Example: track per-frontend and per-backend counters, block abusers at the
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03009612 frontend when the backend detects abuse(and marks gpc0).
Willy Tarreaue9656522010-08-17 15:40:09 +02009613
9614 frontend http
Davor Ocelice9ed2812017-12-25 17:49:28 +01009615 # Use General Purpose Counter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +02009616 # protecting all our sites
9617 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009618 tcp-request connection track-sc0 src
9619 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +02009620 ...
9621 use_backend http_dynamic if { path_end .php }
9622
9623 backend http_dynamic
9624 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009625 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +02009626 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009627 acl click_too_fast sc1_http_req_rate gt 10
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03009628 acl mark_as_abuser sc0_inc_gpc0(http) gt 0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009629 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +02009630 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +02009631
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009632 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +02009633
Jarno Huuskonen95b012b2017-04-06 13:59:14 +03009634 See also : "tcp-request connection", "tcp-request session",
9635 "tcp-request inspect-delay", and "http-request".
Willy Tarreau62644772008-07-16 18:36:06 +02009636
9637
9638tcp-request inspect-delay <timeout>
9639 Set the maximum allowed time to wait for data during content inspection
9640 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02009641 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +02009642 Arguments :
9643 <timeout> is the timeout value specified in milliseconds by default, but
9644 can be in any other unit if the number is suffixed by the unit,
9645 as explained at the top of this document.
9646
9647 People using haproxy primarily as a TCP relay are often worried about the
9648 risk of passing any type of protocol to a server without any analysis. In
9649 order to be able to analyze the request contents, we must first withhold
9650 the data then analyze them. This statement simply enables withholding of
9651 data for at most the specified amount of time.
9652
Willy Tarreaufb356202010-08-03 14:02:05 +02009653 TCP content inspection applies very early when a connection reaches a
9654 frontend, then very early when the connection is forwarded to a backend. This
9655 means that a connection may experience a first delay in the frontend and a
9656 second delay in the backend if both have tcp-request rules.
9657
Willy Tarreau62644772008-07-16 18:36:06 +02009658 Note that when performing content inspection, haproxy will evaluate the whole
9659 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009660 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +02009661 a last check is performed upon expiration, this time considering that the
Willy Tarreaud869b242009-03-15 14:43:58 +01009662 contents are definitive. If no delay is set, haproxy will not wait at all
9663 and will immediately apply a verdict based on the available information.
9664 Obviously this is unlikely to be very useful and might even be racy, so such
9665 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +02009666
9667 As soon as a rule matches, the request is released and continues as usual. If
9668 the timeout is reached and no rule matches, the default policy will be to let
9669 it pass through unaffected.
9670
9671 For most protocols, it is enough to set it to a few seconds, as most clients
9672 send the full request immediately upon connection. Add 3 or more seconds to
9673 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +01009674 to use large values, for instance to ensure that the client never talks
Davor Ocelice9ed2812017-12-25 17:49:28 +01009675 before the server (e.g. SMTP), or to wait for a client to talk before passing
9676 data to the server (e.g. SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +02009677 least the inspection delay, otherwise it will expire first. If the client
9678 closes the connection or if the buffer is full, the delay immediately expires
9679 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +02009680
Willy Tarreau55165fe2009-05-10 12:02:55 +02009681 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +02009682 "timeout client".
9683
9684
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009685tcp-response content <action> [{if | unless} <condition>]
9686 Perform an action on a session response depending on a layer 4-7 condition
9687 May be used in sections : defaults | frontend | listen | backend
9688 no | no | yes | yes
9689 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009690 <action> defines the action to perform if the condition applies. See
9691 below.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009692
9693 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
9694
Davor Ocelice9ed2812017-12-25 17:49:28 +01009695 Response contents can be analyzed at an early stage of response processing
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009696 called "TCP content inspection". During this stage, ACL-based rules are
9697 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009698 "accept", "close" or a "reject" rule matches, or a TCP response inspection
9699 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009700
9701 Most often, these decisions will consider a protocol recognition or validity.
9702
9703 Content-based rules are evaluated in their exact declaration order. If no
9704 rule matches or if there is no rule, the default action is to accept the
9705 contents. There is no specific limit to the number of rules which may be
9706 inserted.
9707
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009708 Several types of actions are supported :
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009709 - accept :
9710 accepts the response if the condition is true (when used with "if")
9711 or false (when used with "unless"). The first such rule executed ends
9712 the rules evaluation.
9713
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009714 - close :
9715 immediately closes the connection with the server if the condition is
9716 true (when used with "if"), or false (when used with "unless"). The
9717 first such rule executed ends the rules evaluation. The main purpose of
9718 this action is to force a connection to be finished between a client
9719 and a server after an exchange when the application protocol expects
9720 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009721 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009722 protocols.
9723
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009724 - reject :
9725 rejects the response if the condition is true (when used with "if")
9726 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009727 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009728
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009729 - set-var(<var-name>) <expr>
9730 Sets a variable.
9731
Christopher Faulet85d79c92016-11-09 16:54:56 +01009732 - unset-var(<var-name>)
9733 Unsets a variable.
9734
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009735 - sc-inc-gpc0(<sc-id>):
9736 This action increments the GPC0 counter according to the sticky
9737 counter designated by <sc-id>. If an error occurs, this action fails
9738 silently and the actions evaluation continues.
9739
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009740 - sc-inc-gpc1(<sc-id>):
9741 This action increments the GPC1 counter according to the sticky
9742 counter designated by <sc-id>. If an error occurs, this action fails
9743 silently and the actions evaluation continues.
9744
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009745 - sc-set-gpt0(<sc-id>) <int> :
9746 This action sets the GPT0 tag according to the sticky counter designated
9747 by <sc-id> and the value of <int>. The expected result is a boolean. If
9748 an error occurs, this action silently fails and the actions evaluation
9749 continues.
9750
Willy Tarreau2d392c22015-08-24 01:43:45 +02009751 - "silent-drop" :
9752 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +01009753 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +02009754 to prevent the client from being notified. The effect it then that the
9755 client still sees an established connection while there's none on
9756 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
9757 except that it doesn't use any local resource at all on the machine
9758 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +01009759 slow down stronger attackers. It is important to understand the impact
9760 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009761 client and HAProxy (firewalls, proxies, load balancers) will also keep
9762 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +01009763 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009764 TCP_REPAIR socket option is used to block the emission of a TCP
9765 reset. On other systems, the socket's TTL is reduced to 1 so that the
9766 TCP reset doesn't pass the first router, though it's still delivered to
9767 local networks. Do not use it unless you fully understand how it works.
9768
Christopher Faulet76c09ef2017-09-21 11:03:52 +02009769 - send-spoe-group <engine-name> <group-name>
9770 Send a group of SPOE messages.
9771
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009772 Note that the "if/unless" condition is optional. If no condition is set on
9773 the action, it is simply performed unconditionally. That can be useful for
9774 for changing the default action to a reject.
9775
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009776 It is perfectly possible to match layer 7 contents with "tcp-response
9777 content" rules, but then it is important to ensure that a full response has
9778 been buffered, otherwise no contents will match. In order to achieve this,
9779 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009780 period.
9781
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009782 The "set-var" is used to set the content of a variable. The variable is
9783 declared inline.
9784
Daniel Schneller0b547052016-03-21 20:46:57 +01009785 <var-name> The name of the variable starts with an indication about
9786 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01009787 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01009788 "sess" : the variable is shared with the whole session
9789 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009790 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01009791 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009792 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01009793 "res" : the variable is shared only during response
9794 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009795 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01009796 The name may only contain characters 'a-z', 'A-Z', '0-9',
9797 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009798
9799 <expr> Is a standard HAProxy expression formed by a sample-fetch
9800 followed by some converters.
9801
9802 Example:
9803
9804 tcp-request content set-var(sess.my_var) src
9805
Christopher Faulet85d79c92016-11-09 16:54:56 +01009806 The "unset-var" is used to unset a variable. See above for details about
9807 <var-name>.
9808
9809 Example:
9810
9811 tcp-request content unset-var(sess.my_var)
9812
Christopher Faulet76c09ef2017-09-21 11:03:52 +02009813 The "send-spoe-group" is used to trigger sending of a group of SPOE
9814 messages. To do so, the SPOE engine used to send messages must be defined, as
9815 well as the SPOE group to send. Of course, the SPOE engine must refer to an
9816 existing SPOE filter. If not engine name is provided on the SPOE filter line,
9817 the SPOE agent name must be used.
9818
9819 <engine-name> The SPOE engine name.
9820
9821 <group-name> The SPOE group name as specified in the engine configuration.
9822
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009823 See section 7 about ACL usage.
9824
9825 See also : "tcp-request content", "tcp-response inspect-delay"
9826
9827
Willy Tarreau4f614292016-10-21 17:49:36 +02009828tcp-request session <action> [{if | unless} <condition>]
9829 Perform an action on a validated session depending on a layer 5 condition
9830 May be used in sections : defaults | frontend | listen | backend
9831 no | yes | yes | no
9832 Arguments :
9833 <action> defines the action to perform if the condition applies. See
9834 below.
9835
9836 <condition> is a standard layer5-only ACL-based condition (see section 7).
9837
Davor Ocelice9ed2812017-12-25 17:49:28 +01009838 Once a session is validated, (i.e. after all handshakes have been completed),
Willy Tarreau4f614292016-10-21 17:49:36 +02009839 it is possible to evaluate some conditions to decide whether this session
9840 must be accepted or dropped or have its counters tracked. Those conditions
9841 cannot make use of any data contents because no buffers are allocated yet and
9842 the processing cannot wait at this stage. The main use case it to copy some
9843 early information into variables (since variables are accessible in the
9844 session), or to keep track of some information collected after the handshake,
9845 such as SSL-level elements (SNI, ciphers, client cert's CN) or information
Davor Ocelice9ed2812017-12-25 17:49:28 +01009846 from the PROXY protocol header (e.g. track a source forwarded this way). The
Willy Tarreau4f614292016-10-21 17:49:36 +02009847 extracted information can thus be copied to a variable or tracked using
9848 "track-sc" rules. Of course it is also possible to decide to accept/reject as
9849 with other rulesets. Most operations performed here could also be performed
9850 in "tcp-request content" rules, except that in HTTP these rules are evaluated
9851 for each new request, and that might not always be acceptable. For example a
9852 rule might increment a counter on each evaluation. It would also be possible
9853 that a country is resolved by geolocation from the source IP address,
9854 assigned to a session-wide variable, then the source address rewritten from
9855 an HTTP header for all requests. If some contents need to be inspected in
9856 order to take the decision, the "tcp-request content" statements must be used
9857 instead.
9858
9859 The "tcp-request session" rules are evaluated in their exact declaration
9860 order. If no rule matches or if there is no rule, the default action is to
9861 accept the incoming session. There is no specific limit to the number of
9862 rules which may be inserted.
9863
9864 Several types of actions are supported :
9865 - accept : the request is accepted
9866 - reject : the request is rejected and the connection is closed
9867 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
9868 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009869 - sc-inc-gpc1(<sc-id>)
Willy Tarreau4f614292016-10-21 17:49:36 +02009870 - sc-set-gpt0(<sc-id>) <int>
9871 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +01009872 - unset-var(<var-name>)
Willy Tarreau4f614292016-10-21 17:49:36 +02009873 - silent-drop
9874
9875 These actions have the same meaning as their respective counter-parts in
9876 "tcp-request connection" and "tcp-request content", so please refer to these
9877 sections for a complete description.
9878
9879 Note that the "if/unless" condition is optional. If no condition is set on
9880 the action, it is simply performed unconditionally. That can be useful for
9881 "track-sc*" actions as well as for changing the default action to a reject.
9882
9883 Example: track the original source address by default, or the one advertised
9884 in the PROXY protocol header for connection coming from the local
9885 proxies. The first connection-level rule enables receipt of the
9886 PROXY protocol for these ones, the second rule tracks whatever
9887 address we decide to keep after optional decoding.
9888
9889 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
9890 tcp-request session track-sc0 src
9891
9892 Example: accept all sessions from white-listed hosts, reject too fast
9893 sessions without counting them, and track accepted sessions.
9894 This results in session rate being capped from abusive sources.
9895
9896 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
9897 tcp-request session reject if { src_sess_rate gt 10 }
9898 tcp-request session track-sc0 src
9899
9900 Example: accept all sessions from white-listed hosts, count all other
9901 sessions and reject too fast ones. This results in abusive ones
9902 being blocked as long as they don't slow down.
9903
9904 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
9905 tcp-request session track-sc0 src
9906 tcp-request session reject if { sc0_sess_rate gt 10 }
9907
9908 See section 7 about ACL usage.
9909
9910 See also : "tcp-request connection", "tcp-request content", "stick-table"
9911
9912
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009913tcp-response inspect-delay <timeout>
9914 Set the maximum allowed time to wait for a response during content inspection
9915 May be used in sections : defaults | frontend | listen | backend
9916 no | no | yes | yes
9917 Arguments :
9918 <timeout> is the timeout value specified in milliseconds by default, but
9919 can be in any other unit if the number is suffixed by the unit,
9920 as explained at the top of this document.
9921
9922 See also : "tcp-response content", "tcp-request inspect-delay".
9923
9924
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009925timeout check <timeout>
9926 Set additional check timeout, but only after a connection has been already
9927 established.
9928
9929 May be used in sections: defaults | frontend | listen | backend
9930 yes | no | yes | yes
9931 Arguments:
9932 <timeout> is the timeout value specified in milliseconds by default, but
9933 can be in any other unit if the number is suffixed by the unit,
9934 as explained at the top of this document.
9935
9936 If set, haproxy uses min("timeout connect", "inter") as a connect timeout
9937 for check and "timeout check" as an additional read timeout. The "min" is
Davor Ocelice9ed2812017-12-25 17:49:28 +01009938 used so that people running with *very* long "timeout connect" (e.g. those
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009939 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +01009940 (Please also note that there is no valid reason to have such long connect
9941 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
9942 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009943
9944 If "timeout check" is not set haproxy uses "inter" for complete check
9945 timeout (connect + read) exactly like all <1.3.15 version.
9946
9947 In most cases check request is much simpler and faster to handle than normal
9948 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +01009949 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009950
9951 This parameter is specific to backends, but can be specified once for all in
9952 "defaults" sections. This is in fact one of the easiest solutions not to
9953 forget about it.
9954
Willy Tarreau41a340d2008-01-22 12:25:31 +01009955 See also: "timeout connect", "timeout queue", "timeout server",
9956 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009957
9958
Willy Tarreau0ba27502007-12-24 16:55:16 +01009959timeout client <timeout>
9960timeout clitimeout <timeout> (deprecated)
9961 Set the maximum inactivity time on the client side.
9962 May be used in sections : defaults | frontend | listen | backend
9963 yes | yes | yes | no
9964 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01009965 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +01009966 can be in any other unit if the number is suffixed by the unit,
9967 as explained at the top of this document.
9968
9969 The inactivity timeout applies when the client is expected to acknowledge or
9970 send data. In HTTP mode, this timeout is particularly important to consider
9971 during the first phase, when the client sends the request, and during the
Baptiste Assmann2e1941e2016-03-06 23:24:12 +01009972 response while it is reading data sent by the server. That said, for the
9973 first phase, it is preferable to set the "timeout http-request" to better
9974 protect HAProxy from Slowloris like attacks. The value is specified in
9975 milliseconds by default, but can be in any other unit if the number is
Willy Tarreau0ba27502007-12-24 16:55:16 +01009976 suffixed by the unit, as specified at the top of this document. In TCP mode
9977 (and to a lesser extent, in HTTP mode), it is highly recommended that the
9978 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01009979 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01009980 losses by specifying timeouts that are slightly above multiples of 3 seconds
Davor Ocelice9ed2812017-12-25 17:49:28 +01009981 (e.g. 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
9982 sessions (e.g. WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +02009983 which overrides "timeout client" and "timeout server" for tunnels, as well as
9984 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01009985
9986 This parameter is specific to frontends, but can be specified once for all in
9987 "defaults" sections. This is in fact one of the easiest solutions not to
9988 forget about it. An unspecified timeout results in an infinite timeout, which
9989 is not recommended. Such a usage is accepted and works but reports a warning
9990 during startup because it may results in accumulation of expired sessions in
9991 the system if the system's timeouts are not configured either.
9992
Willy Tarreau95c4e142017-11-26 12:18:55 +01009993 This also applies to HTTP/2 connections, which will be closed with GOAWAY.
Lukas Tribus75df9d72017-11-24 19:05:12 +01009994
Willy Tarreau0ba27502007-12-24 16:55:16 +01009995 This parameter replaces the old, deprecated "clitimeout". It is recommended
9996 to use it to write new configurations. The form "timeout clitimeout" is
9997 provided only by backwards compatibility but its use is strongly discouraged.
9998
Baptiste Assmann2e1941e2016-03-06 23:24:12 +01009999 See also : "clitimeout", "timeout server", "timeout tunnel",
10000 "timeout http-request".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010001
10002
Willy Tarreau05cdd962014-05-10 14:30:07 +020010003timeout client-fin <timeout>
10004 Set the inactivity timeout on the client side for half-closed connections.
10005 May be used in sections : defaults | frontend | listen | backend
10006 yes | yes | yes | no
10007 Arguments :
10008 <timeout> is the timeout value specified in milliseconds by default, but
10009 can be in any other unit if the number is suffixed by the unit,
10010 as explained at the top of this document.
10011
10012 The inactivity timeout applies when the client is expected to acknowledge or
10013 send data while one direction is already shut down. This timeout is different
10014 from "timeout client" in that it only applies to connections which are closed
10015 in one direction. This is particularly useful to avoid keeping connections in
10016 FIN_WAIT state for too long when clients do not disconnect cleanly. This
10017 problem is particularly common long connections such as RDP or WebSocket.
10018 Note that this timeout can override "timeout tunnel" when a connection shuts
Willy Tarreau599391a2017-11-24 10:16:00 +010010019 down in one direction. It is applied to idle HTTP/2 connections once a GOAWAY
10020 frame was sent, often indicating an expectation that the connection quickly
10021 ends.
Willy Tarreau05cdd962014-05-10 14:30:07 +020010022
10023 This parameter is specific to frontends, but can be specified once for all in
10024 "defaults" sections. By default it is not set, so half-closed connections
10025 will use the other timeouts (timeout.client or timeout.tunnel).
10026
10027 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
10028
10029
Willy Tarreau0ba27502007-12-24 16:55:16 +010010030timeout connect <timeout>
10031timeout contimeout <timeout> (deprecated)
10032 Set the maximum time to wait for a connection attempt to a server to succeed.
10033 May be used in sections : defaults | frontend | listen | backend
10034 yes | no | yes | yes
10035 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010010036 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010010037 can be in any other unit if the number is suffixed by the unit,
10038 as explained at the top of this document.
10039
10040 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010010041 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +010010042 cover one or several TCP packet losses by specifying timeouts that are
Davor Ocelice9ed2812017-12-25 17:49:28 +010010043 slightly above multiples of 3 seconds (e.g. 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010044 connect timeout also presets both queue and tarpit timeouts to the same value
10045 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +010010046
10047 This parameter is specific to backends, but can be specified once for all in
10048 "defaults" sections. This is in fact one of the easiest solutions not to
10049 forget about it. An unspecified timeout results in an infinite timeout, which
10050 is not recommended. Such a usage is accepted and works but reports a warning
10051 during startup because it may results in accumulation of failed sessions in
10052 the system if the system's timeouts are not configured either.
10053
10054 This parameter replaces the old, deprecated "contimeout". It is recommended
10055 to use it to write new configurations. The form "timeout contimeout" is
10056 provided only by backwards compatibility but its use is strongly discouraged.
10057
Willy Tarreau41a340d2008-01-22 12:25:31 +010010058 See also: "timeout check", "timeout queue", "timeout server", "contimeout",
10059 "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010060
10061
Willy Tarreaub16a5742010-01-10 14:46:16 +010010062timeout http-keep-alive <timeout>
10063 Set the maximum allowed time to wait for a new HTTP request to appear
10064 May be used in sections : defaults | frontend | listen | backend
10065 yes | yes | yes | yes
10066 Arguments :
10067 <timeout> is the timeout value specified in milliseconds by default, but
10068 can be in any other unit if the number is suffixed by the unit,
10069 as explained at the top of this document.
10070
10071 By default, the time to wait for a new request in case of keep-alive is set
10072 by "timeout http-request". However this is not always convenient because some
10073 people want very short keep-alive timeouts in order to release connections
10074 faster, and others prefer to have larger ones but still have short timeouts
10075 once the request has started to present itself.
10076
10077 The "http-keep-alive" timeout covers these needs. It will define how long to
10078 wait for a new HTTP request to start coming after a response was sent. Once
10079 the first byte of request has been seen, the "http-request" timeout is used
10080 to wait for the complete request to come. Note that empty lines prior to a
10081 new request do not refresh the timeout and are not counted as a new request.
10082
10083 There is also another difference between the two timeouts : when a connection
10084 expires during timeout http-keep-alive, no error is returned, the connection
10085 just closes. If the connection expires in "http-request" while waiting for a
10086 connection to complete, a HTTP 408 error is returned.
10087
10088 In general it is optimal to set this value to a few tens to hundreds of
10089 milliseconds, to allow users to fetch all objects of a page at once but
Davor Ocelice9ed2812017-12-25 17:49:28 +010010090 without waiting for further clicks. Also, if set to a very small value (e.g.
Willy Tarreaub16a5742010-01-10 14:46:16 +010010091 1 millisecond) it will probably only accept pipelined requests but not the
10092 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +020010093 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +010010094
10095 If this parameter is not set, the "http-request" timeout applies, and if both
10096 are not set, "timeout client" still applies at the lower level. It should be
10097 set in the frontend to take effect, unless the frontend is in TCP mode, in
10098 which case the HTTP backend's timeout will be used.
10099
Willy Tarreau95c4e142017-11-26 12:18:55 +010010100 When using HTTP/2 "timeout client" is applied instead. This is so we can keep
10101 using short keep-alive timeouts in HTTP/1.1 while using longer ones in HTTP/2
Lukas Tribus75df9d72017-11-24 19:05:12 +010010102 (where we only have one connection per client and a connection setup).
10103
Willy Tarreaub16a5742010-01-10 14:46:16 +010010104 See also : "timeout http-request", "timeout client".
10105
10106
Willy Tarreau036fae02008-01-06 13:24:40 +010010107timeout http-request <timeout>
10108 Set the maximum allowed time to wait for a complete HTTP request
10109 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +020010110 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +010010111 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010010112 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +010010113 can be in any other unit if the number is suffixed by the unit,
10114 as explained at the top of this document.
10115
10116 In order to offer DoS protection, it may be required to lower the maximum
10117 accepted time to receive a complete HTTP request without affecting the client
10118 timeout. This helps protecting against established connections on which
10119 nothing is sent. The client timeout cannot offer a good protection against
10120 this abuse because it is an inactivity timeout, which means that if the
10121 attacker sends one character every now and then, the timeout will not
10122 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +020010123 types, the request will be aborted if it does not complete in time. When the
10124 timeout expires, an HTTP 408 response is sent to the client to inform it
10125 about the problem, and the connection is closed. The logs will report
10126 termination codes "cR". Some recent browsers are having problems with this
Davor Ocelice9ed2812017-12-25 17:49:28 +010010127 standard, well-documented behavior, so it might be needed to hide the 408
Willy Tarreau0f228a02015-05-01 15:37:53 +020010128 code using "option http-ignore-probes" or "errorfile 408 /dev/null". See
10129 more details in the explanations of the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +010010130
Baptiste Assmanneccdf432015-10-28 13:49:01 +010010131 By default, this timeout only applies to the header part of the request,
10132 and not to any data. As soon as the empty line is received, this timeout is
10133 not used anymore. When combined with "option http-buffer-request", this
10134 timeout also applies to the body of the request..
10135 It is used again on keep-alive connections to wait for a second
Willy Tarreaub16a5742010-01-10 14:46:16 +010010136 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +010010137
10138 Generally it is enough to set it to a few seconds, as most clients send the
10139 full request immediately upon connection. Add 3 or more seconds to cover TCP
Davor Ocelice9ed2812017-12-25 17:49:28 +010010140 retransmits but that's all. Setting it to very low values (e.g. 50 ms) will
Willy Tarreau036fae02008-01-06 13:24:40 +010010141 generally work on local networks as long as there are no packet losses. This
10142 will prevent people from sending bare HTTP requests using telnet.
10143
10144 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +020010145 chunk of the incoming request. It should be set in the frontend to take
10146 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
10147 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +010010148
Willy Tarreau0f228a02015-05-01 15:37:53 +020010149 See also : "errorfile", "http-ignore-probes", "timeout http-keep-alive", and
Baptiste Assmanneccdf432015-10-28 13:49:01 +010010150 "timeout client", "option http-buffer-request".
Willy Tarreau036fae02008-01-06 13:24:40 +010010151
Willy Tarreau844e3c52008-01-11 16:28:18 +010010152
10153timeout queue <timeout>
10154 Set the maximum time to wait in the queue for a connection slot to be free
10155 May be used in sections : defaults | frontend | listen | backend
10156 yes | no | yes | yes
10157 Arguments :
10158 <timeout> is the timeout value specified in milliseconds by default, but
10159 can be in any other unit if the number is suffixed by the unit,
10160 as explained at the top of this document.
10161
10162 When a server's maxconn is reached, connections are left pending in a queue
10163 which may be server-specific or global to the backend. In order not to wait
10164 indefinitely, a timeout is applied to requests pending in the queue. If the
10165 timeout is reached, it is considered that the request will almost never be
10166 served, so it is dropped and a 503 error is returned to the client.
10167
10168 The "timeout queue" statement allows to fix the maximum time for a request to
10169 be left pending in a queue. If unspecified, the same value as the backend's
10170 connection timeout ("timeout connect") is used, for backwards compatibility
10171 with older versions with no "timeout queue" parameter.
10172
10173 See also : "timeout connect", "contimeout".
10174
10175
10176timeout server <timeout>
10177timeout srvtimeout <timeout> (deprecated)
10178 Set the maximum inactivity time on the server side.
10179 May be used in sections : defaults | frontend | listen | backend
10180 yes | no | yes | yes
10181 Arguments :
10182 <timeout> is the timeout value specified in milliseconds by default, but
10183 can be in any other unit if the number is suffixed by the unit,
10184 as explained at the top of this document.
10185
10186 The inactivity timeout applies when the server is expected to acknowledge or
10187 send data. In HTTP mode, this timeout is particularly important to consider
10188 during the first phase of the server's response, when it has to send the
10189 headers, as it directly represents the server's processing time for the
10190 request. To find out what value to put there, it's often good to start with
10191 what would be considered as unacceptable response times, then check the logs
10192 to observe the response time distribution, and adjust the value accordingly.
10193
10194 The value is specified in milliseconds by default, but can be in any other
10195 unit if the number is suffixed by the unit, as specified at the top of this
10196 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
10197 recommended that the client timeout remains equal to the server timeout in
10198 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010010199 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +010010200 packet losses by specifying timeouts that are slightly above multiples of 3
Davor Ocelice9ed2812017-12-25 17:49:28 +010010201 seconds (e.g. 4 or 5 seconds minimum). If some long-lived sessions are mixed
10202 with short-lived sessions (e.g. WebSocket and HTTP), it's worth considering
Willy Tarreauce887fd2012-05-12 12:50:00 +020010203 "timeout tunnel", which overrides "timeout client" and "timeout server" for
10204 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010205
10206 This parameter is specific to backends, but can be specified once for all in
10207 "defaults" sections. This is in fact one of the easiest solutions not to
10208 forget about it. An unspecified timeout results in an infinite timeout, which
10209 is not recommended. Such a usage is accepted and works but reports a warning
10210 during startup because it may results in accumulation of expired sessions in
10211 the system if the system's timeouts are not configured either.
10212
10213 This parameter replaces the old, deprecated "srvtimeout". It is recommended
10214 to use it to write new configurations. The form "timeout srvtimeout" is
10215 provided only by backwards compatibility but its use is strongly discouraged.
10216
Willy Tarreauce887fd2012-05-12 12:50:00 +020010217 See also : "srvtimeout", "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +010010218
Willy Tarreau05cdd962014-05-10 14:30:07 +020010219
10220timeout server-fin <timeout>
10221 Set the inactivity timeout on the server side for half-closed connections.
10222 May be used in sections : defaults | frontend | listen | backend
10223 yes | no | yes | yes
10224 Arguments :
10225 <timeout> is the timeout value specified in milliseconds by default, but
10226 can be in any other unit if the number is suffixed by the unit,
10227 as explained at the top of this document.
10228
10229 The inactivity timeout applies when the server is expected to acknowledge or
10230 send data while one direction is already shut down. This timeout is different
10231 from "timeout server" in that it only applies to connections which are closed
10232 in one direction. This is particularly useful to avoid keeping connections in
10233 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
10234 This problem is particularly common long connections such as RDP or WebSocket.
10235 Note that this timeout can override "timeout tunnel" when a connection shuts
10236 down in one direction. This setting was provided for completeness, but in most
10237 situations, it should not be needed.
10238
10239 This parameter is specific to backends, but can be specified once for all in
10240 "defaults" sections. By default it is not set, so half-closed connections
10241 will use the other timeouts (timeout.server or timeout.tunnel).
10242
10243 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
10244
Willy Tarreau844e3c52008-01-11 16:28:18 +010010245
10246timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +010010247 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +010010248 May be used in sections : defaults | frontend | listen | backend
10249 yes | yes | yes | yes
10250 Arguments :
10251 <timeout> is the tarpit duration specified in milliseconds by default, but
10252 can be in any other unit if the number is suffixed by the unit,
10253 as explained at the top of this document.
10254
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010255 When a connection is tarpitted using "http-request tarpit" or
10256 "reqtarpit", it is maintained open with no activity for a certain
10257 amount of time, then closed. "timeout tarpit" defines how long it will
10258 be maintained open.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010259
10260 The value is specified in milliseconds by default, but can be in any other
10261 unit if the number is suffixed by the unit, as specified at the top of this
10262 document. If unspecified, the same value as the backend's connection timeout
10263 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +010010264 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010265
10266 See also : "timeout connect", "contimeout".
10267
10268
Willy Tarreauce887fd2012-05-12 12:50:00 +020010269timeout tunnel <timeout>
10270 Set the maximum inactivity time on the client and server side for tunnels.
10271 May be used in sections : defaults | frontend | listen | backend
10272 yes | no | yes | yes
10273 Arguments :
10274 <timeout> is the timeout value specified in milliseconds by default, but
10275 can be in any other unit if the number is suffixed by the unit,
10276 as explained at the top of this document.
10277
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010278 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +020010279 between a client and a server, and the connection remains inactive in both
10280 directions. This timeout supersedes both the client and server timeouts once
10281 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
Davor Ocelice9ed2812017-12-25 17:49:28 +010010282 analyzer remains attached to either connection (e.g. tcp content rules are
10283 accepted). In HTTP, this timeout is used when a connection is upgraded (e.g.
Willy Tarreauce887fd2012-05-12 12:50:00 +020010284 when switching to the WebSocket protocol, or forwarding a CONNECT request
10285 to a proxy), or after the first response when no keepalive/close option is
10286 specified.
10287
Willy Tarreau05cdd962014-05-10 14:30:07 +020010288 Since this timeout is usually used in conjunction with long-lived connections,
10289 it usually is a good idea to also set "timeout client-fin" to handle the
10290 situation where a client suddenly disappears from the net and does not
10291 acknowledge a close, or sends a shutdown and does not acknowledge pending
10292 data anymore. This can happen in lossy networks where firewalls are present,
10293 and is detected by the presence of large amounts of sessions in a FIN_WAIT
10294 state.
10295
Willy Tarreauce887fd2012-05-12 12:50:00 +020010296 The value is specified in milliseconds by default, but can be in any other
10297 unit if the number is suffixed by the unit, as specified at the top of this
10298 document. Whatever the expected normal idle time, it is a good practice to
10299 cover at least one or several TCP packet losses by specifying timeouts that
Davor Ocelice9ed2812017-12-25 17:49:28 +010010300 are slightly above multiples of 3 seconds (e.g. 4 or 5 seconds minimum).
Willy Tarreauce887fd2012-05-12 12:50:00 +020010301
10302 This parameter is specific to backends, but can be specified once for all in
10303 "defaults" sections. This is in fact one of the easiest solutions not to
10304 forget about it.
10305
10306 Example :
10307 defaults http
10308 option http-server-close
10309 timeout connect 5s
10310 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +020010311 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +020010312 timeout server 30s
10313 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
10314
Willy Tarreau05cdd962014-05-10 14:30:07 +020010315 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +020010316
10317
Willy Tarreau844e3c52008-01-11 16:28:18 +010010318transparent (deprecated)
10319 Enable client-side transparent proxying
10320 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +010010321 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +010010322 Arguments : none
10323
10324 This keyword was introduced in order to provide layer 7 persistence to layer
10325 3 load balancers. The idea is to use the OS's ability to redirect an incoming
10326 connection for a remote address to a local process (here HAProxy), and let
10327 this process know what address was initially requested. When this option is
10328 used, sessions without cookies will be forwarded to the original destination
10329 IP address of the incoming request (which should match that of another
10330 equipment), while requests with cookies will still be forwarded to the
10331 appropriate server.
10332
10333 The "transparent" keyword is deprecated, use "option transparent" instead.
10334
10335 Note that contrary to a common belief, this option does NOT make HAProxy
10336 present the client's IP to the server when establishing the connection.
10337
Willy Tarreau844e3c52008-01-11 16:28:18 +010010338 See also: "option transparent"
10339
William Lallemanda73203e2012-03-12 12:48:57 +010010340unique-id-format <string>
10341 Generate a unique ID for each request.
10342 May be used in sections : defaults | frontend | listen | backend
10343 yes | yes | yes | no
10344 Arguments :
10345 <string> is a log-format string.
10346
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010347 This keyword creates a ID for each request using the custom log format. A
10348 unique ID is useful to trace a request passing through many components of
10349 a complex infrastructure. The newly created ID may also be logged using the
10350 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +010010351
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010352 The format should be composed from elements that are guaranteed to be
10353 unique when combined together. For instance, if multiple haproxy instances
10354 are involved, it might be important to include the node name. It is often
10355 needed to log the incoming connection's source and destination addresses
10356 and ports. Note that since multiple requests may be performed over the same
10357 connection, including a request counter may help differentiate them.
10358 Similarly, a timestamp may protect against a rollover of the counter.
10359 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +010010360
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010361 It is recommended to use hexadecimal notation for many fields since it
10362 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +010010363
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010364 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010010365
Julien Vehentf21be322014-03-07 08:27:34 -050010366 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010010367
10368 will generate:
10369
10370 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
10371
10372 See also: "unique-id-header"
10373
10374unique-id-header <name>
10375 Add a unique ID header in the HTTP request.
10376 May be used in sections : defaults | frontend | listen | backend
10377 yes | yes | yes | no
10378 Arguments :
10379 <name> is the name of the header.
10380
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010381 Add a unique-id header in the HTTP request sent to the server, using the
10382 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +010010383
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010384 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010010385
Julien Vehentf21be322014-03-07 08:27:34 -050010386 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010010387 unique-id-header X-Unique-ID
10388
10389 will generate:
10390
10391 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
10392
10393 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +010010394
Willy Tarreauf51658d2014-04-23 01:21:56 +020010395use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010396 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010397 May be used in sections : defaults | frontend | listen | backend
10398 no | yes | yes | no
10399 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010400 <backend> is the name of a valid backend or "listen" section, or a
10401 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010402
Willy Tarreauf51658d2014-04-23 01:21:56 +020010403 <condition> is a condition composed of ACLs, as described in section 7. If
10404 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010405
10406 When doing content-switching, connections arrive on a frontend and are then
10407 dispatched to various backends depending on a number of conditions. The
10408 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010409 "use_backend" keyword. While it is normally used with HTTP processing, it can
Davor Ocelice9ed2812017-12-25 17:49:28 +010010410 also be used in pure TCP, either without content using stateless ACLs (e.g.
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010411 source address validation) or combined with a "tcp-request" rule to wait for
10412 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010413
10414 There may be as many "use_backend" rules as desired. All of these rules are
10415 evaluated in their declaration order, and the first one which matches will
10416 assign the backend.
10417
10418 In the first form, the backend will be used if the condition is met. In the
10419 second form, the backend will be used if the condition is not met. If no
10420 condition is valid, the backend defined with "default_backend" will be used.
10421 If no default backend is defined, either the servers in the same section are
10422 used (in case of a "listen" section) or, in case of a frontend, no server is
10423 used and a 503 service unavailable response is returned.
10424
Willy Tarreau51aecc72009-07-12 09:47:04 +020010425 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010426 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +020010427 and backend processing will immediately follow, or the backend will wait for
10428 a complete HTTP request to get in. This feature is useful when a frontend
10429 must decode several protocols on a unique port, one of them being HTTP.
10430
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010431 When <backend> is a simple name, it is resolved at configuration time, and an
10432 error is reported if the specified backend does not exist. If <backend> is
10433 a log-format string instead, no check may be done at configuration time, so
10434 the backend name is resolved dynamically at run time. If the resulting
10435 backend name does not correspond to any valid backend, no other rule is
10436 evaluated, and the default_backend directive is applied instead. Note that
10437 when using dynamic backend names, it is highly recommended to use a prefix
10438 that no other backend uses in order to ensure that an unauthorized backend
10439 cannot be forced from the request.
10440
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010441 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010442 used to detect the association between frontends and backends to compute the
10443 backend's "fullconn" setting. This cannot be done for dynamic names.
10444
10445 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
10446 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +010010447
Willy Tarreau036fae02008-01-06 13:24:40 +010010448
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010449use-server <server> if <condition>
10450use-server <server> unless <condition>
10451 Only use a specific server if/unless an ACL-based condition is matched.
10452 May be used in sections : defaults | frontend | listen | backend
10453 no | no | yes | yes
10454 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010455 <server> is the name of a valid server in the same backend section.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010456
10457 <condition> is a condition composed of ACLs, as described in section 7.
10458
10459 By default, connections which arrive to a backend are load-balanced across
10460 the available servers according to the configured algorithm, unless a
10461 persistence mechanism such as a cookie is used and found in the request.
10462
10463 Sometimes it is desirable to forward a particular request to a specific
10464 server without having to declare a dedicated backend for this server. This
10465 can be achieved using the "use-server" rules. These rules are evaluated after
10466 the "redirect" rules and before evaluating cookies, and they have precedence
10467 on them. There may be as many "use-server" rules as desired. All of these
10468 rules are evaluated in their declaration order, and the first one which
10469 matches will assign the server.
10470
10471 If a rule designates a server which is down, and "option persist" is not used
10472 and no force-persist rule was validated, it is ignored and evaluation goes on
10473 with the next rules until one matches.
10474
10475 In the first form, the server will be used if the condition is met. In the
10476 second form, the server will be used if the condition is not met. If no
10477 condition is valid, the processing continues and the server will be assigned
10478 according to other persistence mechanisms.
10479
10480 Note that even if a rule is matched, cookie processing is still performed but
10481 does not assign the server. This allows prefixed cookies to have their prefix
10482 stripped.
10483
10484 The "use-server" statement works both in HTTP and TCP mode. This makes it
10485 suitable for use with content-based inspection. For instance, a server could
10486 be selected in a farm according to the TLS SNI field. And if these servers
10487 have their weight set to zero, they will not be used for other traffic.
10488
10489 Example :
10490 # intercept incoming TLS requests based on the SNI field
10491 use-server www if { req_ssl_sni -i www.example.com }
10492 server www 192.168.0.1:443 weight 0
10493 use-server mail if { req_ssl_sni -i mail.example.com }
10494 server mail 192.168.0.1:587 weight 0
10495 use-server imap if { req_ssl_sni -i imap.example.com }
Lukas Tribus98a3e3f2017-03-26 12:55:35 +000010496 server imap 192.168.0.1:993 weight 0
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010497 # all the rest is forwarded to this server
10498 server default 192.168.0.2:443 check
10499
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010500 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010501
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010502
Davor Ocelice9ed2812017-12-25 17:49:28 +0100105035. Bind and server options
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010504--------------------------
10505
10506The "bind", "server" and "default-server" keywords support a number of settings
10507depending on some build options and on the system HAProxy was built on. These
10508settings generally each consist in one word sometimes followed by a value,
10509written on the same line as the "bind" or "server" line. All these options are
10510described in this section.
10511
10512
105135.1. Bind options
10514-----------------
10515
10516The "bind" keyword supports a certain number of settings which are all passed
10517as arguments on the same line. The order in which those arguments appear makes
10518no importance, provided that they appear after the bind address. All of these
10519parameters are optional. Some of them consist in a single words (booleans),
10520while other ones expect a value after them. In this case, the value must be
10521provided immediately after the setting name.
10522
10523The currently supported settings are the following ones.
10524
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010525accept-netscaler-cip <magic number>
10526 Enforces the use of the NetScaler Client IP insertion protocol over any
10527 connection accepted by any of the TCP sockets declared on the same line. The
10528 NetScaler Client IP insertion protocol dictates the layer 3/4 addresses of
10529 the incoming connection to be used everywhere an address is used, with the
10530 only exception of "tcp-request connection" rules which will only see the
10531 real connection address. Logs will reflect the addresses indicated in the
10532 protocol, unless it is violated, in which case the real address will still
10533 be used. This keyword combined with support from external components can be
10534 used as an efficient and reliable alternative to the X-Forwarded-For
Bertrand Jacquin90759682016-06-06 15:35:39 +010010535 mechanism which is not always reliable and not even always usable. See also
10536 "tcp-request connection expect-netscaler-cip" for a finer-grained setting of
10537 which client is allowed to use the protocol.
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010538
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010539accept-proxy
10540 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +020010541 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
10542 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010543 3/4 addresses of the incoming connection to be used everywhere an address is
10544 used, with the only exception of "tcp-request connection" rules which will
10545 only see the real connection address. Logs will reflect the addresses
10546 indicated in the protocol, unless it is violated, in which case the real
Davor Ocelice9ed2812017-12-25 17:49:28 +010010547 address will still be used. This keyword combined with support from external
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010548 components can be used as an efficient and reliable alternative to the
10549 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +020010550 usable. See also "tcp-request connection expect-proxy" for a finer-grained
10551 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010552
Olivier Houchardc2aae742017-09-22 18:26:28 +020010553allow-0rtt
Bertrand Jacquina25282b2018-08-14 00:56:13 +010010554 Allow receiving early data when using TLSv1.3. This is disabled by default,
Olivier Houchardc2aae742017-09-22 18:26:28 +020010555 due to security considerations.
10556
Willy Tarreauab861d32013-04-02 02:30:41 +020010557alpn <protocols>
10558 This enables the TLS ALPN extension and advertises the specified protocol
10559 list as supported on top of ALPN. The protocol list consists in a comma-
10560 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
10561 quotes). This requires that the SSL library is build with support for TLS
10562 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
Willy Tarreau95c4e142017-11-26 12:18:55 +010010563 initial NPN extension. ALPN is required to enable HTTP/2 on an HTTP frontend.
10564 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
10565 now obsolete NPN extension. At the time of writing this, most browsers still
10566 support both ALPN and NPN for HTTP/2 so a fallback to NPN may still work for
10567 a while. But ALPN must be used whenever possible. If both HTTP/2 and HTTP/1.1
10568 are expected to be supported, both versions can be advertised, in order of
10569 preference, like below :
10570
10571 bind :443 ssl crt pub.pem alpn h2,http/1.1
Willy Tarreauab861d32013-04-02 02:30:41 +020010572
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010573backlog <backlog>
10574 Sets the socket's backlog to this value. If unspecified, the frontend's
10575 backlog is used instead, which generally defaults to the maxconn value.
10576
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010010577curves <curves>
10578 This setting is only available when support for OpenSSL was built in. It sets
10579 the string describing the list of elliptic curves algorithms ("curve suite")
10580 that are negotiated during the SSL/TLS handshake with ECDHE. The format of the
10581 string is a colon-delimited list of curve name.
10582 Example: "X25519:P-256" (without quote)
10583 When "curves" is set, "ecdhe" parameter is ignored.
10584
Emeric Brun7fb34422012-09-28 15:26:15 +020010585ecdhe <named curve>
10586 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +010010587 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
10588 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +020010589
Emeric Brunfd33a262012-10-11 16:28:27 +020010590ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +020010591 This setting is only available when support for OpenSSL was built in. It
10592 designates a PEM file from which to load CA certificates used to verify
10593 client's certificate.
10594
Emeric Brunb6dc9342012-09-28 17:55:37 +020010595ca-ignore-err [all|<errorID>,...]
10596 This setting is only available when support for OpenSSL was built in.
10597 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
10598 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
10599 error is ignored.
10600
Christopher Faulet31af49d2015-06-09 17:29:50 +020010601ca-sign-file <cafile>
10602 This setting is only available when support for OpenSSL was built in. It
10603 designates a PEM file containing both the CA certificate and the CA private
10604 key used to create and sign server's certificates. This is a mandatory
10605 setting when the dynamic generation of certificates is enabled. See
10606 'generate-certificates' for details.
10607
Bertrand Jacquind4d0a232016-11-13 16:37:12 +000010608ca-sign-pass <passphrase>
Christopher Faulet31af49d2015-06-09 17:29:50 +020010609 This setting is only available when support for OpenSSL was built in. It is
10610 the CA private key passphrase. This setting is optional and used only when
10611 the dynamic generation of certificates is enabled. See
10612 'generate-certificates' for details.
10613
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010614ciphers <ciphers>
10615 This setting is only available when support for OpenSSL was built in. It sets
10616 the string describing the list of cipher algorithms ("cipher suite") that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010617 negotiated during the SSL/TLS handshake. The format of the string is defined
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010618 in "man 1 ciphers" from OpenSSL man pages, and can be for instance a string
10619 such as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes).
Davor Ocelice9ed2812017-12-25 17:49:28 +010010620 Depending on the compatibility and security requirements, the list of suitable
Daniel Schneller87e43022017-09-01 19:29:57 +020010621 ciphers depends on a variety of variables. For background information and
10622 recommendations see e. g. (https://wiki.mozilla.org/Security/Server_Side_TLS)
10623 and (https://mozilla.github.io/server-side-tls/ssl-config-generator/).
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010624
Emeric Brunfd33a262012-10-11 16:28:27 +020010625crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +020010626 This setting is only available when support for OpenSSL was built in. It
10627 designates a PEM file from which to load certificate revocation list used
10628 to verify client's certificate.
10629
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010630crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +000010631 This setting is only available when support for OpenSSL was built in. It
10632 designates a PEM file containing both the required certificates and any
10633 associated private keys. This file can be built by concatenating multiple
10634 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
10635 requires an intermediate certificate, this can also be concatenated into this
10636 file.
10637
10638 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
10639 are loaded.
10640
10641 If a directory name is used instead of a PEM file, then all files found in
Cyril Bonté3180f7b2015-01-25 00:16:08 +010010642 that directory will be loaded in alphabetic order unless their name ends with
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010010643 '.issuer', '.ocsp' or '.sctl' (reserved extensions). This directive may be
10644 specified multiple times in order to load certificates from multiple files or
10645 directories. The certificates will be presented to clients who provide a
10646 valid TLS Server Name Indication field matching one of their CN or alt
Davor Ocelice9ed2812017-12-25 17:49:28 +010010647 subjects. Wildcards are supported, where a wildcard character '*' is used
10648 instead of the first hostname component (e.g. *.example.org matches
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010010649 www.example.org but not www.sub.example.org).
Alex Davies0fbf0162013-03-02 16:04:50 +000010650
10651 If no SNI is provided by the client or if the SSL library does not support
10652 TLS extensions, or if the client provides an SNI hostname which does not
10653 match any certificate, then the first loaded certificate will be presented.
10654 This means that when loading certificates from a directory, it is highly
Cyril Bonté3180f7b2015-01-25 00:16:08 +010010655 recommended to load the default one first as a file or to ensure that it will
10656 always be the first one in the directory.
Alex Davies0fbf0162013-03-02 16:04:50 +000010657
Emeric Brune032bfa2012-09-28 13:01:45 +020010658 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010659
Davor Ocelice9ed2812017-12-25 17:49:28 +010010660 Some CAs (such as GoDaddy) offer a drop down list of server types that do not
Alex Davies0fbf0162013-03-02 16:04:50 +000010661 include HAProxy when obtaining a certificate. If this happens be sure to
Davor Ocelice9ed2812017-12-25 17:49:28 +010010662 choose a web server that the CA believes requires an intermediate CA (for
10663 GoDaddy, selection Apache Tomcat will get the correct bundle, but many
Alex Davies0fbf0162013-03-02 16:04:50 +000010664 others, e.g. nginx, result in a wrong bundle that will not work for some
10665 clients).
10666
Emeric Brun4147b2e2014-06-16 18:36:30 +020010667 For each PEM file, haproxy checks for the presence of file at the same path
10668 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
10669 Status Request extension (also known as "OCSP stapling") is automatically
10670 enabled. The content of this file is optional. If not empty, it must contain
10671 a valid OCSP Response in DER format. In order to be valid an OCSP Response
10672 must comply with the following rules: it has to indicate a good status,
10673 it has to be a single response for the certificate of the PEM file, and it
10674 has to be valid at the moment of addition. If these rules are not respected
10675 the OCSP Response is ignored and a warning is emitted. In order to identify
10676 which certificate an OCSP Response applies to, the issuer's certificate is
10677 necessary. If the issuer's certificate is not found in the PEM file, it will
10678 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
10679 if it exists otherwise it will fail with an error.
10680
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010010681 For each PEM file, haproxy also checks for the presence of file at the same
10682 path suffixed by ".sctl". If such file is found, support for Certificate
10683 Transparency (RFC6962) TLS extension is enabled. The file must contain a
10684 valid Signed Certificate Timestamp List, as described in RFC. File is parsed
10685 to check basic syntax, but no signatures are verified.
10686
yanbzhu6c25e9e2016-01-05 12:52:02 -050010687 There are cases where it is desirable to support multiple key types, e.g. RSA
10688 and ECDSA in the cipher suites offered to the clients. This allows clients
10689 that support EC certificates to be able to use EC ciphers, while
10690 simultaneously supporting older, RSA only clients.
yanbzhud19630c2015-12-14 15:10:25 -050010691
10692 In order to provide this functionality, multiple PEM files, each with a
10693 different key type, are required. To associate these PEM files into a
10694 "cert bundle" that is recognized by haproxy, they must be named in the
10695 following way: All PEM files that are to be bundled must have the same base
10696 name, with a suffix indicating the key type. Currently, three suffixes are
10697 supported: rsa, dsa and ecdsa. For example, if www.example.com has two PEM
10698 files, an RSA file and an ECDSA file, they must be named: "example.pem.rsa"
10699 and "example.pem.ecdsa". The first part of the filename is arbitrary; only the
10700 suffix matters. To load this bundle into haproxy, specify the base name only:
10701
10702 Example : bind :8443 ssl crt example.pem
10703
yanbzhu6c25e9e2016-01-05 12:52:02 -050010704 Note that the suffix is not given to haproxy; this tells haproxy to look for
yanbzhud19630c2015-12-14 15:10:25 -050010705 a cert bundle.
10706
Davor Ocelice9ed2812017-12-25 17:49:28 +010010707 HAProxy will load all PEM files in the bundle at the same time to try to
yanbzhud19630c2015-12-14 15:10:25 -050010708 support multiple key types. PEM files are combined based on Common Name
10709 (CN) and Subject Alternative Name (SAN) to support SNI lookups. This means
10710 that even if you give haproxy a cert bundle, if there are no shared CN/SAN
10711 entries in the certificates in that bundle, haproxy will not be able to
10712 provide multi-cert support.
10713
10714 Assuming bundle in the example above contained the following:
10715
10716 Filename | CN | SAN
10717 -------------------+-----------------+-------------------
10718 example.pem.rsa | www.example.com | rsa.example.com
yanbzhu6c25e9e2016-01-05 12:52:02 -050010719 -------------------+-----------------+-------------------
yanbzhud19630c2015-12-14 15:10:25 -050010720 example.pem.ecdsa | www.example.com | ecdsa.example.com
10721 -------------------+-----------------+-------------------
10722
10723 Users connecting with an SNI of "www.example.com" will be able
10724 to use both RSA and ECDSA cipher suites. Users connecting with an SNI of
10725 "rsa.example.com" will only be able to use RSA cipher suites, and users
10726 connecting with "ecdsa.example.com" will only be able to use ECDSA cipher
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020010727 suites. With BoringSSL and Openssl >= 1.1.1 multi-cert is natively supported,
10728 no need to bundle certificates. ECDSA certificate will be preferred if client
10729 support it.
yanbzhud19630c2015-12-14 15:10:25 -050010730
10731 If a directory name is given as the <cert> argument, haproxy will
10732 automatically search and load bundled files in that directory.
10733
10734 OSCP files (.ocsp) and issuer files (.issuer) are supported with multi-cert
10735 bundling. Each certificate can have its own .ocsp and .issuer file. At this
10736 time, sctl is not supported in multi-certificate bundling.
10737
Emeric Brunb6dc9342012-09-28 17:55:37 +020010738crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +000010739 This setting is only available when support for OpenSSL was built in. Sets a
Davor Ocelice9ed2812017-12-25 17:49:28 +010010740 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010741 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +000010742 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +020010743
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010744crt-list <file>
10745 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010746 designates a list of PEM file with an optional ssl configuration and a SNI
10747 filter per certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010748
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010749 <crtfile> [\[<sslbindconf> ...\]] [[!]<snifilter> ...]
10750
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020010751 sslbindconf support "npn", "alpn", "verify", "ca-file", "no-ca-names",
10752 crl-file", "ecdhe", "curves", "ciphers" configuration. With BoringSSL
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020010753 and Openssl >= 1.1.1 "ssl-min-ver" and "ssl-max-ver" are also supported.
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010754 It override the configuration set in bind line for the certificate.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010755
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020010756 Wildcards are supported in the SNI filter. Negative filter are also supported,
10757 only useful in combination with a wildcard filter to exclude a particular SNI.
10758 The certificates will be presented to clients who provide a valid TLS Server
10759 Name Indication field matching one of the SNI filters. If no SNI filter is
10760 specified, the CN and alt subjects are used. This directive may be specified
10761 multiple times. See the "crt" option for more information. The default
10762 certificate is still needed to meet OpenSSL expectations. If it is not used,
10763 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010764
yanbzhu6c25e9e2016-01-05 12:52:02 -050010765 Multi-cert bundling (see "crt") is supported with crt-list, as long as only
Emmanuel Hocdetd294aea2016-05-13 11:14:06 +020010766 the base name is given in the crt-list. SNI filter will do the same work on
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020010767 all bundled certificates. With BoringSSL and Openssl >= 1.1.1 multi-cert is
10768 natively supported, avoid multi-cert bundling. RSA and ECDSA certificates can
10769 be declared in a row, and set different ssl and filter parameter.
yanbzhud19630c2015-12-14 15:10:25 -050010770
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010771 crt-list file example:
10772 cert1.pem
Emmanuel Hocdet05942112017-02-20 16:11:50 +010010773 cert2.pem [alpn h2,http/1.1]
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010774 certW.pem *.domain.tld !secure.domain.tld
Emmanuel Hocdet05942112017-02-20 16:11:50 +010010775 certS.pem [curves X25519:P-256 ciphers ECDHE-ECDSA-AES256-GCM-SHA384] secure.domain.tld
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010776
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010777defer-accept
10778 Is an optional keyword which is supported only on certain Linux kernels. It
10779 states that a connection will only be accepted once some data arrive on it,
10780 or at worst after the first retransmit. This should be used only on protocols
Davor Ocelice9ed2812017-12-25 17:49:28 +010010781 for which the client talks first (e.g. HTTP). It can slightly improve
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010782 performance by ensuring that most of the request is already available when
10783 the connection is accepted. On the other hand, it will not be able to detect
10784 connections which don't talk. It is important to note that this option is
10785 broken in all kernels up to 2.6.31, as the connection is never accepted until
10786 the client talks. This can cause issues with front firewalls which would see
10787 an established connection while the proxy will only see it in SYN_RECV. This
10788 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
10789
William Lallemandf6975e92017-05-26 17:42:10 +020010790expose-fd listeners
10791 This option is only usable with the stats socket. It gives your stats socket
10792 the capability to pass listeners FD to another HAProxy process.
William Lallemande202b1e2017-06-01 17:38:56 +020010793 During a reload with the master-worker mode, the process is automatically
10794 reexecuted adding -x and one of the stats socket with this option.
Davor Ocelice9ed2812017-12-25 17:49:28 +010010795 See also "-x" in the management guide.
William Lallemandf6975e92017-05-26 17:42:10 +020010796
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010797force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010798 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010799 this listener. SSLv3 is generally less expensive than the TLS counterparts
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010800 for high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010801 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010802
10803force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010804 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010805 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010806 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010807
10808force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010809 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010810 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010811 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010812
10813force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010814 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010815 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010816 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010817
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020010818force-tlsv13
10819 This option enforces use of TLSv1.3 only on SSL connections instantiated from
10820 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010821 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020010822
Christopher Faulet31af49d2015-06-09 17:29:50 +020010823generate-certificates
10824 This setting is only available when support for OpenSSL was built in. It
10825 enables the dynamic SSL certificates generation. A CA certificate and its
10826 private key are necessary (see 'ca-sign-file'). When HAProxy is configured as
10827 a transparent forward proxy, SSL requests generate errors because of a common
10828 name mismatch on the certificate presented to the client. With this option
10829 enabled, HAProxy will try to forge a certificate using the SNI hostname
10830 indicated by the client. This is done only if no certificate matches the SNI
10831 hostname (see 'crt-list'). If an error occurs, the default certificate is
10832 used, else the 'strict-sni' option is set.
10833 It can also be used when HAProxy is configured as a reverse proxy to ease the
10834 deployment of an architecture with many backends.
10835
10836 Creating a SSL certificate is an expensive operation, so a LRU cache is used
10837 to store forged certificates (see 'tune.ssl.ssl-ctx-cache-size'). It
Davor Ocelice9ed2812017-12-25 17:49:28 +010010838 increases the HAProxy's memory footprint to reduce latency when the same
Christopher Faulet31af49d2015-06-09 17:29:50 +020010839 certificate is used many times.
10840
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010841gid <gid>
10842 Sets the group of the UNIX sockets to the designated system gid. It can also
10843 be set by default in the global section's "unix-bind" statement. Note that
10844 some platforms simply ignore this. This setting is equivalent to the "group"
10845 setting except that the group ID is used instead of its name. This setting is
10846 ignored by non UNIX sockets.
10847
10848group <group>
10849 Sets the group of the UNIX sockets to the designated system group. It can
10850 also be set by default in the global section's "unix-bind" statement. Note
10851 that some platforms simply ignore this. This setting is equivalent to the
10852 "gid" setting except that the group name is used instead of its gid. This
10853 setting is ignored by non UNIX sockets.
10854
10855id <id>
10856 Fixes the socket ID. By default, socket IDs are automatically assigned, but
10857 sometimes it is more convenient to fix them to ease monitoring. This value
10858 must be strictly positive and unique within the listener/frontend. This
10859 option can only be used when defining only a single socket.
10860
10861interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +010010862 Restricts the socket to a specific interface. When specified, only packets
10863 received from that particular interface are processed by the socket. This is
10864 currently only supported on Linux. The interface must be a primary system
10865 interface, not an aliased interface. It is also possible to bind multiple
10866 frontends to the same address if they are bound to different interfaces. Note
10867 that binding to a network interface requires root privileges. This parameter
Jérôme Magnin61275192018-02-07 11:39:58 +010010868 is only compatible with TCPv4/TCPv6 sockets. When specified, return traffic
10869 uses the same interface as inbound traffic, and its associated routing table,
10870 even if there are explicit routes through different interfaces configured.
10871 This can prove useful to address asymmetric routing issues when the same
10872 client IP addresses need to be able to reach frontends hosted on different
10873 interfaces.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010874
Willy Tarreauabb175f2012-09-24 12:43:26 +020010875level <level>
10876 This setting is used with the stats sockets only to restrict the nature of
10877 the commands that can be issued on the socket. It is ignored by other
10878 sockets. <level> can be one of :
Davor Ocelice9ed2812017-12-25 17:49:28 +010010879 - "user" is the least privileged level; only non-sensitive stats can be
Willy Tarreauabb175f2012-09-24 12:43:26 +020010880 read, and no change is allowed. It would make sense on systems where it
10881 is not easy to restrict access to the socket.
10882 - "operator" is the default level and fits most common uses. All data can
Davor Ocelice9ed2812017-12-25 17:49:28 +010010883 be read, and only non-sensitive changes are permitted (e.g. clear max
Willy Tarreauabb175f2012-09-24 12:43:26 +020010884 counters).
Davor Ocelice9ed2812017-12-25 17:49:28 +010010885 - "admin" should be used with care, as everything is permitted (e.g. clear
Willy Tarreauabb175f2012-09-24 12:43:26 +020010886 all counters).
10887
Andjelko Iharosc4df59e2017-07-20 11:59:48 +020010888severity-output <format>
10889 This setting is used with the stats sockets only to configure severity
10890 level output prepended to informational feedback messages. Severity
10891 level of messages can range between 0 and 7, conforming to syslog
10892 rfc5424. Valid and successful socket commands requesting data
10893 (i.e. "show map", "get acl foo" etc.) will never have a severity level
10894 prepended. It is ignored by other sockets. <format> can be one of :
10895 - "none" (default) no severity level is prepended to feedback messages.
10896 - "number" severity level is prepended as a number.
10897 - "string" severity level is prepended as a string following the
10898 rfc5424 convention.
10899
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010900maxconn <maxconn>
10901 Limits the sockets to this number of concurrent connections. Extraneous
10902 connections will remain in the system's backlog until a connection is
10903 released. If unspecified, the limit will be the same as the frontend's
10904 maxconn. Note that in case of port ranges or multiple addresses, the same
10905 value will be applied to each socket. This setting enables different
10906 limitations on expensive sockets, for instance SSL entries which may easily
10907 eat all memory.
10908
10909mode <mode>
10910 Sets the octal mode used to define access permissions on the UNIX socket. It
10911 can also be set by default in the global section's "unix-bind" statement.
10912 Note that some platforms simply ignore this. This setting is ignored by non
10913 UNIX sockets.
10914
10915mss <maxseg>
10916 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
10917 connections. This can be used to force a lower MSS for certain specific
10918 ports, for instance for connections passing through a VPN. Note that this
10919 relies on a kernel feature which is theoretically supported under Linux but
10920 was buggy in all versions prior to 2.6.28. It may or may not work on other
10921 operating systems. It may also not change the advertised value but change the
10922 effective size of outgoing segments. The commonly advertised value for TCPv4
10923 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
10924 positive, it will be used as the advertised MSS. If it is negative, it will
10925 indicate by how much to reduce the incoming connection's advertised MSS for
10926 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
10927
10928name <name>
10929 Sets an optional name for these sockets, which will be reported on the stats
10930 page.
10931
Willy Tarreaud72f0f32015-10-13 14:50:22 +020010932namespace <name>
10933 On Linux, it is possible to specify which network namespace a socket will
10934 belong to. This directive makes it possible to explicitly bind a listener to
10935 a namespace different from the default one. Please refer to your operating
10936 system's documentation to find more details about network namespaces.
10937
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010938nice <nice>
10939 Sets the 'niceness' of connections initiated from the socket. Value must be
10940 in the range -1024..1024 inclusive, and defaults to zero. Positive values
10941 means that such connections are more friendly to others and easily offer
10942 their place in the scheduler. On the opposite, negative values mean that
10943 connections want to run with a higher priority than others. The difference
10944 only happens under high loads when the system is close to saturation.
10945 Negative values are appropriate for low-latency or administration services,
10946 and high values are generally recommended for CPU intensive tasks such as SSL
10947 processing or bulk transfers which are less sensible to latency. For example,
10948 it may make sense to use a positive value for an SMTP socket and a negative
10949 one for an RDP socket.
10950
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020010951no-ca-names
10952 This setting is only available when support for OpenSSL was built in. It
10953 prevents from send CA names in server hello message when ca-file is used.
10954
Emeric Brun9b3009b2012-10-05 11:55:06 +020010955no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010956 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010957 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010958 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010959 be enabled using any configuration option. This option is also available on
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010960 global statement "ssl-default-bind-options". Use "ssl-min-ver" and
10961 "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010962
Emeric Brun90ad8722012-10-02 14:00:59 +020010963no-tls-tickets
10964 This setting is only available when support for OpenSSL was built in. It
10965 disables the stateless session resumption (RFC 5077 TLS Ticket
10966 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010967 session resumption is more expensive in CPU usage. This option is also
10968 available on global statement "ssl-default-bind-options".
Emeric Brun90ad8722012-10-02 14:00:59 +020010969
Emeric Brun9b3009b2012-10-05 11:55:06 +020010970no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010971 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010972 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010973 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010974 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010975 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
10976 and "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010977
Emeric Brun9b3009b2012-10-05 11:55:06 +020010978no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +020010979 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010980 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010981 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010982 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010983 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
10984 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020010985
Emeric Brun9b3009b2012-10-05 11:55:06 +020010986no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +020010987 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010988 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010989 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010990 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010991 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
10992 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020010993
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020010994no-tlsv13
10995 This setting is only available when support for OpenSSL was built in. It
10996 disables support for TLSv1.3 on any sockets instantiated from the listener
10997 when SSL is supported. Note that SSLv2 is forced disabled in the code and
10998 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020010999 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
11000 and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011001
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020011002npn <protocols>
11003 This enables the NPN TLS extension and advertises the specified protocol list
11004 as supported on top of NPN. The protocol list consists in a comma-delimited
11005 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
11006 This requires that the SSL library is build with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +020011007 enabled (check with haproxy -vv). Note that the NPN extension has been
Willy Tarreau95c4e142017-11-26 12:18:55 +010011008 replaced with the ALPN extension (see the "alpn" keyword), though this one is
11009 only available starting with OpenSSL 1.0.2. If HTTP/2 is desired on an older
11010 version of OpenSSL, NPN might still be used as most clients still support it
11011 at the time of writing this. It is possible to enable both NPN and ALPN
11012 though it probably doesn't make any sense out of testing.
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020011013
Lukas Tribus53ae85c2017-05-04 15:45:40 +000011014prefer-client-ciphers
11015 Use the client's preference when selecting the cipher suite, by default
11016 the server's preference is enforced. This option is also available on
11017 global statement "ssl-default-bind-options".
Lukas Tribus926594f2018-05-18 17:55:57 +020011018 Note that with OpenSSL >= 1.1.1 ChaCha20-Poly1305 is reprioritized anyway
11019 (without setting this option), if a ChaCha20-Poly1305 cipher is at the top of
11020 the client cipher list.
Lukas Tribus53ae85c2017-05-04 15:45:40 +000011021
Christopher Fauletc644fa92017-11-23 22:44:11 +010011022process <process-set>[/<thread-set>]
11023 This restricts the list of processes and/or threads on which this listener is
11024 allowed to run. It does not enforce any process but eliminates those which do
Davor Ocelice9ed2812017-12-25 17:49:28 +010011025 not match. If the frontend uses a "bind-process" setting, the intersection
Christopher Fauletc644fa92017-11-23 22:44:11 +010011026 between the two is applied. If in the end the listener is not allowed to run
11027 on any remaining process, a warning is emitted, and the listener will either
11028 run on the first process of the listener if a single process was specified,
11029 or on all of its processes if multiple processes were specified. If a thread
Davor Ocelice9ed2812017-12-25 17:49:28 +010011030 set is specified, it limits the threads allowed to process incoming
Christopher Fauletc644fa92017-11-23 22:44:11 +010011031 connections for this listener, for the corresponding process set. For the
11032 unlikely case where several ranges are needed, this directive may be
11033 repeated. <process-set> and <thread-set> must use the format
11034
11035 all | odd | even | number[-[number]]
11036
11037 Ranges can be partially defined. The higher bound can be omitted. In such
11038 case, it is replaced by the corresponding maximum value. The main purpose of
11039 this directive is to be used with the stats sockets and have one different
11040 socket per process. The second purpose is to have multiple bind lines sharing
11041 the same IP:port but not the same process in a listener, so that the system
11042 can distribute the incoming connections into multiple queues and allow a
11043 smoother inter-process load balancing. Currently Linux 3.9 and above is known
11044 for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020011045
Christopher Fauleta717b992018-04-10 14:43:00 +020011046proto <name>
11047 Forces the multiplexer's protocol to use for the incoming connections. It
11048 must be compatible with the mode of the frontend (TCP or HTTP). It must also
11049 be usable on the frontend side. The list of available protocols is reported
11050 in haproxy -vv.
11051 Idea behind this optipon is to bypass the selection of the best multiplexer's
11052 protocol for all connections instantiated from this listening socket. For
11053 instance, it is possible to force the http/2 on clear TCP by specifing "proto
11054 h2" on the bind line.
11055
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011056ssl
11057 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011058 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011059 certificate is necessary (see "crt" above). All contents in the buffers will
11060 appear in clear text, so that ACLs and HTTP processing will only have access
Emmanuel Hocdetbd695fe2017-05-15 15:53:41 +020011061 to deciphered contents. SSLv3 is disabled per default, use "ssl-min-ver SSLv3"
11062 to enable it.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011063
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011064ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11065 This option enforces use of <version> or lower on SSL connections instantiated
11066 from this listener. This option is also available on global statement
11067 "ssl-default-bind-options". See also "ssl-min-ver".
11068
11069ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11070 This option enforces use of <version> or upper on SSL connections instantiated
11071 from this listener. This option is also available on global statement
11072 "ssl-default-bind-options". See also "ssl-max-ver".
11073
Emmanuel Hocdet65623372013-01-24 17:17:15 +010011074strict-sni
11075 This setting is only available when support for OpenSSL was built in. The
11076 SSL/TLS negotiation is allow only if the client provided an SNI which match
11077 a certificate. The default certificate is not used.
11078 See the "crt" option for more information.
11079
Willy Tarreau2af207a2015-02-04 00:45:58 +010011080tcp-ut <delay>
Tim Düsterhus4896c442016-11-29 02:15:19 +010011081 Sets the TCP User Timeout for all incoming connections instantiated from this
Willy Tarreau2af207a2015-02-04 00:45:58 +010011082 listening socket. This option is available on Linux since version 2.6.37. It
11083 allows haproxy to configure a timeout for sockets which contain data not
Davor Ocelice9ed2812017-12-25 17:49:28 +010011084 receiving an acknowledgment for the configured delay. This is especially
Willy Tarreau2af207a2015-02-04 00:45:58 +010011085 useful on long-lived connections experiencing long idle periods such as
11086 remote terminals or database connection pools, where the client and server
11087 timeouts must remain high to allow a long period of idle, but where it is
11088 important to detect that the client has disappeared in order to release all
11089 resources associated with its connection (and the server's session). The
11090 argument is a delay expressed in milliseconds by default. This only works
11091 for regular TCP connections, and is ignored for other protocols.
11092
Willy Tarreau1c862c52012-10-05 16:21:00 +020011093tfo
Lukas Tribus0defb902013-02-13 23:35:39 +010011094 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +020011095 enables TCP Fast Open on the listening socket, which means that clients which
11096 support this feature will be able to send a request and receive a response
11097 during the 3-way handshake starting from second connection, thus saving one
11098 round-trip after the first connection. This only makes sense with protocols
11099 that use high connection rates and where each round trip matters. This can
11100 possibly cause issues with many firewalls which do not accept data on SYN
11101 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +020011102 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
11103 need to build HAProxy with USE_TFO=1 if your libc doesn't define
11104 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +020011105
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010011106tls-ticket-keys <keyfile>
11107 Sets the TLS ticket keys file to load the keys from. The keys need to be 48
11108 bytes long, encoded with base64 (ex. openssl rand -base64 48). Number of keys
11109 is specified by the TLS_TICKETS_NO build option (default 3) and at least as
11110 many keys need to be present in the file. Last TLS_TICKETS_NO keys will be
11111 used for decryption and the penultimate one for encryption. This enables easy
11112 key rotation by just appending new key to the file and reloading the process.
11113 Keys must be periodically rotated (ex. every 12h) or Perfect Forward Secrecy
11114 is compromised. It is also a good idea to keep the keys off any permanent
11115 storage such as hard drives (hint: use tmpfs and don't swap those files).
11116 Lifetime hint can be changed using tune.ssl.timeout.
11117
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011118transparent
11119 Is an optional keyword which is supported only on certain Linux kernels. It
11120 indicates that the addresses will be bound even if they do not belong to the
11121 local machine, and that packets targeting any of these addresses will be
11122 intercepted just as if the addresses were locally configured. This normally
11123 requires that IP forwarding is enabled. Caution! do not use this with the
11124 default address '*', as it would redirect any traffic for the specified port.
11125 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
11126 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
11127 kernel version. Some distribution kernels include backports of the feature,
11128 so check for support with your vendor.
11129
Willy Tarreau77e3af92012-11-24 15:07:23 +010011130v4v6
11131 Is an optional keyword which is supported only on most recent systems
11132 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
11133 and IPv6 when it uses the default address. Doing so is sometimes necessary
11134 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011135 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +010011136
Willy Tarreau9b6700f2012-11-24 11:55:28 +010011137v6only
11138 Is an optional keyword which is supported only on most recent systems
11139 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
11140 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +010011141 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
11142 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +010011143
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011144uid <uid>
11145 Sets the owner of the UNIX sockets to the designated system uid. It can also
11146 be set by default in the global section's "unix-bind" statement. Note that
11147 some platforms simply ignore this. This setting is equivalent to the "user"
11148 setting except that the user numeric ID is used instead of its name. This
11149 setting is ignored by non UNIX sockets.
11150
11151user <user>
11152 Sets the owner of the UNIX sockets to the designated system user. It can also
11153 be set by default in the global section's "unix-bind" statement. Note that
11154 some platforms simply ignore this. This setting is equivalent to the "uid"
11155 setting except that the user name is used instead of its uid. This setting is
11156 ignored by non UNIX sockets.
11157
Emeric Brun1a073b42012-09-28 17:07:34 +020011158verify [none|optional|required]
11159 This setting is only available when support for OpenSSL was built in. If set
11160 to 'none', client certificate is not requested. This is the default. In other
11161 cases, a client certificate is requested. If the client does not provide a
11162 certificate after the request and if 'verify' is set to 'required', then the
11163 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +020011164 certificate provided by the client is always verified using CAs from
11165 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
11166 is aborted, regardless of the 'verify' option, unless the error code exactly
11167 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020011168
Willy Tarreaub6205fd2012-09-24 12:27:33 +0200111695.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +010011170------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020011171
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010011172The "server" and "default-server" keywords support a certain number of settings
11173which are all passed as arguments on the server line. The order in which those
11174arguments appear does not count, and they are all optional. Some of those
11175settings are single words (booleans) while others expect one or several values
11176after them. In this case, the values must immediately follow the setting name.
11177Except default-server, all those settings must be specified after the server's
11178address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +020011179
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011180 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010011181 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +020011182
Frédéric Lécailled2376272017-03-21 18:52:12 +010011183Note that all these settings are supported both by "server" and "default-server"
11184keywords, except "id" which is only supported by "server".
11185
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011186The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +010011187
Willy Tarreauceb4ac92012-04-28 00:41:46 +020011188addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011189 Using the "addr" parameter, it becomes possible to use a different IP address
Baptiste Assmann13f83532016-03-06 23:14:36 +010011190 to send health-checks or to probe the agent-check. On some servers, it may be
11191 desirable to dedicate an IP address to specific component able to perform
11192 complex tests which are more suitable to health-checks than the application.
11193 This parameter is ignored if the "check" parameter is not set. See also the
11194 "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011195
Simon Hormand60d6912013-11-25 10:46:36 +090011196agent-check
11197 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +010011198 health check. An agent health check is performed by making a TCP connection
11199 to the port set by the "agent-port" parameter and reading an ASCII string.
11200 The string is made of a series of words delimited by spaces, tabs or commas
11201 in any order, optionally terminated by '\r' and/or '\n', each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +090011202
Willy Tarreau81f5d942013-12-09 20:51:51 +010011203 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +090011204 Values in this format will set the weight proportional to the initial
Willy Tarreauc5af3a62014-10-07 15:27:33 +020011205 weight of a server as configured when haproxy starts. Note that a zero
11206 weight is reported on the stats page as "DRAIN" since it has the same
11207 effect on the server (it's removed from the LB farm).
Simon Hormand60d6912013-11-25 10:46:36 +090011208
Davor Ocelice9ed2812017-12-25 17:49:28 +010011209 - The string "maxconn:" followed by an integer (no space between). Values
11210 in this format will set the maxconn of a server. The maximum number of
11211 connections advertised needs to be multiplied by the number of load
11212 balancers and different backends that use this health check to get the
11213 total number of connections the server might receive. Example: maxconn:30
Nenad Merdanovic174dd372016-04-24 23:10:06 +020011214
Willy Tarreau81f5d942013-12-09 20:51:51 +010011215 - The word "ready". This will turn the server's administrative state to the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011216 READY mode, thus canceling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +090011217
Willy Tarreau81f5d942013-12-09 20:51:51 +010011218 - The word "drain". This will turn the server's administrative state to the
11219 DRAIN mode, thus it will not accept any new connections other than those
11220 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +090011221
Willy Tarreau81f5d942013-12-09 20:51:51 +010011222 - The word "maint". This will turn the server's administrative state to the
11223 MAINT mode, thus it will not accept any new connections at all, and health
11224 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +090011225
Willy Tarreau81f5d942013-12-09 20:51:51 +010011226 - The words "down", "failed", or "stopped", optionally followed by a
11227 description string after a sharp ('#'). All of these mark the server's
11228 operating state as DOWN, but since the word itself is reported on the stats
11229 page, the difference allows an administrator to know if the situation was
11230 expected or not : the service may intentionally be stopped, may appear up
Davor Ocelice9ed2812017-12-25 17:49:28 +010011231 but fail some validity tests, or may be seen as down (e.g. missing process,
Willy Tarreau81f5d942013-12-09 20:51:51 +010011232 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +090011233
Willy Tarreau81f5d942013-12-09 20:51:51 +010011234 - The word "up" sets back the server's operating state as UP if health checks
11235 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +090011236
Willy Tarreau81f5d942013-12-09 20:51:51 +010011237 Parameters which are not advertised by the agent are not changed. For
11238 example, an agent might be designed to monitor CPU usage and only report a
11239 relative weight and never interact with the operating status. Similarly, an
11240 agent could be designed as an end-user interface with 3 radio buttons
11241 allowing an administrator to change only the administrative state. However,
11242 it is important to consider that only the agent may revert its own actions,
11243 so if a server is set to DRAIN mode or to DOWN state using the agent, the
11244 agent must implement the other equivalent actions to bring the service into
11245 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +090011246
Simon Horman2f1f9552013-11-25 10:46:37 +090011247 Failure to connect to the agent is not considered an error as connectivity
11248 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +010011249 parameter. Warning though, it is not a good idea to stop an agent after it
11250 reports "down", since only an agent reporting "up" will be able to turn the
11251 server up again. Note that the CLI on the Unix stats socket is also able to
Willy Tarreau989222a2016-01-15 10:26:26 +010011252 force an agent's result in order to work around a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +090011253
Willy Tarreau81f5d942013-12-09 20:51:51 +010011254 Requires the "agent-port" parameter to be set. See also the "agent-inter"
Frédéric Lécailled2376272017-03-21 18:52:12 +010011255 and "no-agent-check" parameters.
Simon Hormand60d6912013-11-25 10:46:36 +090011256
James Brown55f9ff12015-10-21 18:19:05 -070011257agent-send <string>
11258 If this option is specified, haproxy will send the given string (verbatim)
11259 to the agent server upon connection. You could, for example, encode
11260 the backend name into this string, which would enable your agent to send
11261 different responses based on the backend. Make sure to include a '\n' if
11262 you want to terminate your request with a newline.
11263
Simon Hormand60d6912013-11-25 10:46:36 +090011264agent-inter <delay>
11265 The "agent-inter" parameter sets the interval between two agent checks
11266 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
11267
11268 Just as with every other time-based parameter, it may be entered in any
11269 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
11270 parameter also serves as a timeout for agent checks "timeout check" is
11271 not set. In order to reduce "resonance" effects when multiple servers are
11272 hosted on the same hardware, the agent and health checks of all servers
11273 are started with a small time offset between them. It is also possible to
11274 add some random noise in the agent and health checks interval using the
11275 global "spread-checks" keyword. This makes sense for instance when a lot
11276 of backends use the same servers.
11277
11278 See also the "agent-check" and "agent-port" parameters.
11279
Misiek768d8602017-01-09 09:52:43 +010011280agent-addr <addr>
11281 The "agent-addr" parameter sets address for agent check.
11282
11283 You can offload agent-check to another target, so you can make single place
11284 managing status and weights of servers defined in haproxy in case you can't
11285 make self-aware and self-managing services. You can specify both IP or
11286 hostname, it will be resolved.
11287
Simon Hormand60d6912013-11-25 10:46:36 +090011288agent-port <port>
11289 The "agent-port" parameter sets the TCP port used for agent checks.
11290
11291 See also the "agent-check" and "agent-inter" parameters.
11292
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011293backup
11294 When "backup" is present on a server line, the server is only used in load
11295 balancing when all other non-backup servers are unavailable. Requests coming
11296 with a persistence cookie referencing the server will always be served
11297 though. By default, only the first operational backup server is used, unless
Frédéric Lécailled2376272017-03-21 18:52:12 +010011298 the "allbackups" option is set in the backend. See also the "no-backup" and
11299 "allbackups" options.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011300
Emeric Brunef42d922012-10-11 16:11:36 +020011301ca-file <cafile>
11302 This setting is only available when support for OpenSSL was built in. It
11303 designates a PEM file from which to load CA certificates used to verify
11304 server's certificate.
11305
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011306check
11307 This option enables health checks on the server. By default, a server is
Patrick Mézardb7aeec62012-01-22 16:01:22 +010011308 always considered available. If "check" is set, the server is available when
11309 accepting periodic TCP connections, to ensure that it is really able to serve
11310 requests. The default address and port to send the tests to are those of the
11311 server, and the default source is the same as the one defined in the
11312 backend. It is possible to change the address using the "addr" parameter, the
11313 port using the "port" parameter, the source address using the "source"
11314 address, and the interval and timers using the "inter", "rise" and "fall"
Simon Hormanafc47ee2013-11-25 10:46:35 +090011315 parameters. The request method is define in the backend using the "httpchk",
11316 "smtpchk", "mysql-check", "pgsql-check" and "ssl-hello-chk" options. Please
Frédéric Lécailled2376272017-03-21 18:52:12 +010011317 refer to those options and parameters for more information. See also
11318 "no-check" option.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011319
Willy Tarreau6c16adc2012-10-05 00:04:16 +020011320check-send-proxy
11321 This option forces emission of a PROXY protocol line with outgoing health
11322 checks, regardless of whether the server uses send-proxy or not for the
11323 normal traffic. By default, the PROXY protocol is enabled for health checks
11324 if it is already enabled for normal traffic and if no "port" nor "addr"
11325 directive is present. However, if such a directive is present, the
11326 "check-send-proxy" option needs to be used to force the use of the
11327 protocol. See also the "send-proxy" option for more information.
11328
Olivier Houchard9130a962017-10-17 17:33:43 +020011329check-sni
11330 This option allows you to specify the SNI to be used when doing health checks
11331 over SSL.
11332
Willy Tarreau763a95b2012-10-04 23:15:39 +020011333check-ssl
11334 This option forces encryption of all health checks over SSL, regardless of
11335 whether the server uses SSL or not for the normal traffic. This is generally
11336 used when an explicit "port" or "addr" directive is specified and SSL health
11337 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011338 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +020011339 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
11340 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
Davor Ocelice9ed2812017-12-25 17:49:28 +010011341 All SSL settings are common to health checks and traffic (e.g. ciphers).
Frédéric Lécailled2376272017-03-21 18:52:12 +010011342 See the "ssl" option for more information and "no-check-ssl" to disable
11343 this option.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011344
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011345ciphers <ciphers>
11346 This option sets the string describing the list of cipher algorithms that is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011347 is negotiated during the SSL/TLS handshake with the server. The format of the
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011348 string is defined in "man 1 ciphers". When SSL is used to communicate with
11349 servers on the local network, it is common to see a weaker set of algorithms
11350 than what is used over the internet. Doing so reduces CPU usage on both the
11351 server and haproxy while still keeping it compatible with deployed software.
11352 Some algorithms such as RC4-SHA1 are reasonably cheap. If no security at all
11353 is needed and just connectivity, using DES can be appropriate.
11354
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011355cookie <value>
11356 The "cookie" parameter sets the cookie value assigned to the server to
11357 <value>. This value will be checked in incoming requests, and the first
11358 operational server possessing the same value will be selected. In return, in
11359 cookie insertion or rewrite modes, this value will be assigned to the cookie
11360 sent to the client. There is nothing wrong in having several servers sharing
11361 the same cookie value, and it is in fact somewhat common between normal and
11362 backup servers. See also the "cookie" keyword in backend section.
11363
Emeric Brunef42d922012-10-11 16:11:36 +020011364crl-file <crlfile>
11365 This setting is only available when support for OpenSSL was built in. It
11366 designates a PEM file from which to load certificate revocation list used
11367 to verify server's certificate.
11368
Emeric Bruna7aa3092012-10-26 12:58:00 +020011369crt <cert>
11370 This setting is only available when support for OpenSSL was built in.
11371 It designates a PEM file from which to load both a certificate and the
11372 associated private key. This file can be built by concatenating both PEM
11373 files into one. This certificate will be sent if the server send a client
11374 certificate request.
11375
Willy Tarreau96839092010-03-29 10:02:24 +020011376disabled
11377 The "disabled" keyword starts the server in the "disabled" state. That means
11378 that it is marked down in maintenance mode, and no connection other than the
11379 ones allowed by persist mode will reach it. It is very well suited to setup
11380 new servers, because normal traffic will never reach them, while it is still
11381 possible to test the service by making use of the force-persist mechanism.
Frédéric Lécailled2376272017-03-21 18:52:12 +010011382 See also "enabled" setting.
Willy Tarreau96839092010-03-29 10:02:24 +020011383
Frédéric Lécailled2376272017-03-21 18:52:12 +010011384enabled
11385 This option may be used as 'server' setting to reset any 'disabled'
11386 setting which would have been inherited from 'default-server' directive as
11387 default value.
11388 It may also be used as 'default-server' setting to reset any previous
11389 'default-server' 'disabled' setting.
Willy Tarreau96839092010-03-29 10:02:24 +020011390
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011391error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +010011392 If health observing is enabled, the "error-limit" parameter specifies the
11393 number of consecutive errors that triggers event selected by the "on-error"
11394 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011395
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011396 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011397
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011398fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011399 The "fall" parameter states that a server will be considered as dead after
11400 <count> consecutive unsuccessful health checks. This value defaults to 3 if
11401 unspecified. See also the "check", "inter" and "rise" parameters.
11402
Emeric Brun8694b9a2012-10-05 14:39:07 +020011403force-sslv3
11404 This option enforces use of SSLv3 only when SSL is used to communicate with
11405 the server. SSLv3 is generally less expensive than the TLS counterparts for
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011406 high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011407 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011408
11409force-tlsv10
11410 This option enforces use of TLSv1.0 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011411 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011412 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011413
11414force-tlsv11
11415 This option enforces use of TLSv1.1 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011416 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011417 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011418
11419force-tlsv12
11420 This option enforces use of TLSv1.2 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011421 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011422 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011423
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011424force-tlsv13
11425 This option enforces use of TLSv1.3 only when SSL is used to communicate with
11426 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011427 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011428
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011429id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +020011430 Set a persistent ID for the server. This ID must be positive and unique for
11431 the proxy. An unused ID will automatically be assigned if unset. The first
11432 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011433
Willy Tarreau6a031d12016-11-07 19:42:35 +010011434init-addr {last | libc | none | <ip>},[...]*
11435 Indicate in what order the server's address should be resolved upon startup
11436 if it uses an FQDN. Attempts are made to resolve the address by applying in
Davor Ocelice9ed2812017-12-25 17:49:28 +010011437 turn each of the methods mentioned in the comma-delimited list. The first
Willy Tarreau6a031d12016-11-07 19:42:35 +010011438 method which succeeds is used. If the end of the list is reached without
11439 finding a working method, an error is thrown. Method "last" suggests to pick
11440 the address which appears in the state file (see "server-state-file"). Method
11441 "libc" uses the libc's internal resolver (gethostbyname() or getaddrinfo()
11442 depending on the operating system and build options). Method "none"
11443 specifically indicates that the server should start without any valid IP
11444 address in a down state. It can be useful to ignore some DNS issues upon
11445 startup, waiting for the situation to get fixed later. Finally, an IP address
11446 (IPv4 or IPv6) may be provided. It can be the currently known address of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011447 server (e.g. filled by a configuration generator), or the address of a dummy
Willy Tarreau6a031d12016-11-07 19:42:35 +010011448 server used to catch old sessions and present them with a decent error
11449 message for example. When the "first" load balancing algorithm is used, this
11450 IP address could point to a fake server used to trigger the creation of new
11451 instances on the fly. This option defaults to "last,libc" indicating that the
11452 previous address found in the state file (if any) is used first, otherwise
11453 the libc's resolver is used. This ensures continued compatibility with the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011454 historic behavior.
Willy Tarreau6a031d12016-11-07 19:42:35 +010011455
11456 Example:
11457 defaults
11458 # never fail on address resolution
11459 default-server init-addr last,libc,none
11460
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011461inter <delay>
11462fastinter <delay>
11463downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011464 The "inter" parameter sets the interval between two consecutive health checks
11465 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
11466 It is also possible to use "fastinter" and "downinter" to optimize delays
11467 between checks depending on the server state :
11468
Pieter Baauw44fc9df2015-09-17 21:30:46 +020011469 Server state | Interval used
11470 ----------------------------------------+----------------------------------
11471 UP 100% (non-transitional) | "inter"
11472 ----------------------------------------+----------------------------------
11473 Transitionally UP (going down "fall"), | "fastinter" if set,
11474 Transitionally DOWN (going up "rise"), | "inter" otherwise.
11475 or yet unchecked. |
11476 ----------------------------------------+----------------------------------
11477 DOWN 100% (non-transitional) | "downinter" if set,
11478 | "inter" otherwise.
11479 ----------------------------------------+----------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +010011480
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011481 Just as with every other time-based parameter, they can be entered in any
11482 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
11483 serves as a timeout for health checks sent to servers if "timeout check" is
11484 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +090011485 hosted on the same hardware, the agent and health checks of all servers
11486 are started with a small time offset between them. It is also possible to
11487 add some random noise in the agent and health checks interval using the
11488 global "spread-checks" keyword. This makes sense for instance when a lot
11489 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011490
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011491maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011492 The "maxconn" parameter specifies the maximal number of concurrent
11493 connections that will be sent to this server. If the number of incoming
11494 concurrent requests goes higher than this value, they will be queued, waiting
11495 for a connection to be released. This parameter is very important as it can
11496 save fragile servers from going down under extreme loads. If a "minconn"
11497 parameter is specified, the limit becomes dynamic. The default value is "0"
11498 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
11499 the backend's "fullconn" keyword.
11500
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011501maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011502 The "maxqueue" parameter specifies the maximal number of connections which
11503 will wait in the queue for this server. If this limit is reached, next
11504 requests will be redispatched to other servers instead of indefinitely
11505 waiting to be served. This will break persistence but may allow people to
11506 quickly re-log in when the server they try to connect to is dying. The
11507 default value is "0" which means the queue is unlimited. See also the
11508 "maxconn" and "minconn" parameters.
11509
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011510minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011511 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
11512 limit following the backend's load. The server will always accept at least
11513 <minconn> connections, never more than <maxconn>, and the limit will be on
11514 the ramp between both values when the backend has less than <fullconn>
11515 concurrent connections. This makes it possible to limit the load on the
11516 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011517 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011518 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011519
Willy Tarreaud72f0f32015-10-13 14:50:22 +020011520namespace <name>
11521 On Linux, it is possible to specify which network namespace a socket will
11522 belong to. This directive makes it possible to explicitly bind a server to
11523 a namespace different from the default one. Please refer to your operating
11524 system's documentation to find more details about network namespaces.
11525
Frédéric Lécailled2376272017-03-21 18:52:12 +010011526no-agent-check
11527 This option may be used as "server" setting to reset any "agent-check"
11528 setting which would have been inherited from "default-server" directive as
11529 default value.
11530 It may also be used as "default-server" setting to reset any previous
11531 "default-server" "agent-check" setting.
11532
11533no-backup
11534 This option may be used as "server" setting to reset any "backup"
11535 setting which would have been inherited from "default-server" directive as
11536 default value.
11537 It may also be used as "default-server" setting to reset any previous
11538 "default-server" "backup" setting.
11539
11540no-check
11541 This option may be used as "server" setting to reset any "check"
11542 setting which would have been inherited from "default-server" directive as
11543 default value.
11544 It may also be used as "default-server" setting to reset any previous
11545 "default-server" "check" setting.
11546
11547no-check-ssl
11548 This option may be used as "server" setting to reset any "check-ssl"
11549 setting which would have been inherited from "default-server" directive as
11550 default value.
11551 It may also be used as "default-server" setting to reset any previous
11552 "default-server" "check-ssl" setting.
11553
Frédéric Lécailled2376272017-03-21 18:52:12 +010011554no-send-proxy
11555 This option may be used as "server" setting to reset any "send-proxy"
11556 setting which would have been inherited from "default-server" directive as
11557 default value.
11558 It may also be used as "default-server" setting to reset any previous
11559 "default-server" "send-proxy" setting.
11560
11561no-send-proxy-v2
11562 This option may be used as "server" setting to reset any "send-proxy-v2"
11563 setting which would have been inherited from "default-server" directive as
11564 default value.
11565 It may also be used as "default-server" setting to reset any previous
11566 "default-server" "send-proxy-v2" setting.
11567
11568no-send-proxy-v2-ssl
11569 This option may be used as "server" setting to reset any "send-proxy-v2-ssl"
11570 setting which would have been inherited from "default-server" directive as
11571 default value.
11572 It may also be used as "default-server" setting to reset any previous
11573 "default-server" "send-proxy-v2-ssl" setting.
11574
11575no-send-proxy-v2-ssl-cn
11576 This option may be used as "server" setting to reset any "send-proxy-v2-ssl-cn"
11577 setting which would have been inherited from "default-server" directive as
11578 default value.
11579 It may also be used as "default-server" setting to reset any previous
11580 "default-server" "send-proxy-v2-ssl-cn" setting.
11581
11582no-ssl
11583 This option may be used as "server" setting to reset any "ssl"
11584 setting which would have been inherited from "default-server" directive as
11585 default value.
11586 It may also be used as "default-server" setting to reset any previous
11587 "default-server" "ssl" setting.
11588
Willy Tarreau2a3fb1c2015-02-05 16:47:07 +010011589no-ssl-reuse
11590 This option disables SSL session reuse when SSL is used to communicate with
11591 the server. It will force the server to perform a full handshake for every
11592 new connection. It's probably only useful for benchmarking, troubleshooting,
11593 and for paranoid users.
11594
Emeric Brun9b3009b2012-10-05 11:55:06 +020011595no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011596 This option disables support for SSLv3 when SSL is used to communicate with
11597 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011598 using any configuration option. Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011599
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020011600 Supported in default-server: No
11601
Emeric Brunf9c5c472012-10-11 15:28:34 +020011602no-tls-tickets
11603 This setting is only available when support for OpenSSL was built in. It
11604 disables the stateless session resumption (RFC 5077 TLS Ticket
11605 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011606 session resumption is more expensive in CPU usage for servers. This option
11607 is also available on global statement "ssl-default-server-options".
Frédéric Lécailled2376272017-03-21 18:52:12 +010011608 See also "tls-tickets".
Emeric Brunf9c5c472012-10-11 15:28:34 +020011609
Emeric Brun9b3009b2012-10-05 11:55:06 +020011610no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +020011611 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020011612 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11613 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011614 often makes sense to disable it when communicating with local servers. This
11615 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011616 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011617
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020011618 Supported in default-server: No
11619
Emeric Brun9b3009b2012-10-05 11:55:06 +020011620no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +020011621 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020011622 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11623 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011624 often makes sense to disable it when communicating with local servers. This
11625 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011626 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011627
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020011628 Supported in default-server: No
11629
Emeric Brun9b3009b2012-10-05 11:55:06 +020011630no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +020011631 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011632 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11633 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011634 often makes sense to disable it when communicating with local servers. This
11635 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011636 Use "ssl-min-ver" and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011637
11638 Supported in default-server: No
11639
11640no-tlsv13
11641 This option disables support for TLSv1.3 when SSL is used to communicate with
11642 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11643 using any configuration option. TLSv1 is more expensive than SSLv3 so it
11644 often makes sense to disable it when communicating with local servers. This
11645 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011646 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011647
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020011648 Supported in default-server: No
11649
Frédéric Lécailled2376272017-03-21 18:52:12 +010011650no-verifyhost
11651 This option may be used as "server" setting to reset any "verifyhost"
11652 setting which would have been inherited from "default-server" directive as
11653 default value.
11654 It may also be used as "default-server" setting to reset any previous
11655 "default-server" "verifyhost" setting.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011656
Simon Hormanfa461682011-06-25 09:39:49 +090011657non-stick
11658 Never add connections allocated to this sever to a stick-table.
11659 This may be used in conjunction with backup to ensure that
11660 stick-table persistence is disabled for backup servers.
11661
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011662observe <mode>
11663 This option enables health adjusting based on observing communication with
11664 the server. By default this functionality is disabled and enabling it also
11665 requires to enable health checks. There are two supported modes: "layer4" and
11666 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
11667 significant. In layer7, which is only allowed for http proxies, responses
11668 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +010011669 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011670
11671 See also the "check", "on-error" and "error-limit".
11672
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011673on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011674 Select what should happen when enough consecutive errors are detected.
11675 Currently, four modes are available:
11676 - fastinter: force fastinter
11677 - fail-check: simulate a failed check, also forces fastinter (default)
11678 - sudden-death: simulate a pre-fatal failed health check, one more failed
11679 check will mark a server down, forces fastinter
11680 - mark-down: mark the server immediately down and force fastinter
11681
11682 See also the "check", "observe" and "error-limit".
11683
Simon Hormane0d1bfb2011-06-21 14:34:58 +090011684on-marked-down <action>
11685 Modify what occurs when a server is marked down.
11686 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070011687 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
11688 all connections to the server are immediately terminated when the server
11689 goes down. It might be used if the health check detects more complex cases
11690 than a simple connection status, and long timeouts would cause the service
11691 to remain unresponsive for too long a time. For instance, a health check
11692 might detect that a database is stuck and that there's no chance to reuse
11693 existing connections anymore. Connections killed this way are logged with
11694 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +090011695
11696 Actions are disabled by default
11697
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070011698on-marked-up <action>
11699 Modify what occurs when a server is marked up.
11700 Currently one action is available:
11701 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
11702 done only if the server is not in backup state and if it is not disabled
11703 (it must have an effective weight > 0). This can be used sometimes to force
11704 an active server to take all the traffic back after recovery when dealing
Davor Ocelice9ed2812017-12-25 17:49:28 +010011705 with long sessions (e.g. LDAP, SQL, ...). Doing this can cause more trouble
11706 than it tries to solve (e.g. incomplete transactions), so use this feature
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070011707 with extreme care. Sessions killed because a server comes up are logged
11708 with an 'U' termination code (for "Up").
11709
11710 Actions are disabled by default
11711
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011712port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011713 Using the "port" parameter, it becomes possible to use a different port to
11714 send health-checks. On some servers, it may be desirable to dedicate a port
11715 to a specific component able to perform complex tests which are more suitable
11716 to health-checks than the application. It is common to run a simple script in
11717 inetd for instance. This parameter is ignored if the "check" parameter is not
11718 set. See also the "addr" parameter.
11719
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020011720proto <name>
11721
11722 Forces the multiplexer's protocol to use for the outgoing connections to this
11723 server. It must be compatible with the mode of the backend (TCP or HTTP). It
11724 must also be usable on the backend side. The list of available protocols is
11725 reported in haproxy -vv.
11726 Idea behind this optipon is to bypass the selection of the best multiplexer's
11727 protocol for all connections established to this server.
11728
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011729redir <prefix>
11730 The "redir" parameter enables the redirection mode for all GET and HEAD
11731 requests addressing this server. This means that instead of having HAProxy
11732 forward the request to the server, it will send an "HTTP 302" response with
11733 the "Location" header composed of this prefix immediately followed by the
11734 requested URI beginning at the leading '/' of the path component. That means
11735 that no trailing slash should be used after <prefix>. All invalid requests
11736 will be rejected, and all non-GET or HEAD requests will be normally served by
11737 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011738 mangling nor cookie insertion is possible in the response. However, cookies in
Davor Ocelice9ed2812017-12-25 17:49:28 +010011739 requests are still analyzed, making this solution completely usable to direct
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011740 users to a remote location in case of local disaster. Main use consists in
11741 increasing bandwidth for static servers by having the clients directly
11742 connect to them. Note: never use a relative location here, it would cause a
11743 loop between the client and HAProxy!
11744
11745 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
11746
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011747rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011748 The "rise" parameter states that a server will be considered as operational
11749 after <count> consecutive successful health checks. This value defaults to 2
11750 if unspecified. See also the "check", "inter" and "fall" parameters.
11751
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020011752resolve-opts <option>,<option>,...
11753 Comma separated list of options to apply to DNS resolution linked to this
11754 server.
11755
11756 Available options:
11757
11758 * allow-dup-ip
11759 By default, HAProxy prevents IP address duplication in a backend when DNS
11760 resolution at runtime is in operation.
11761 That said, for some cases, it makes sense that two servers (in the same
11762 backend, being resolved by the same FQDN) have the same IP address.
11763 For such case, simply enable this option.
11764 This is the opposite of prevent-dup-ip.
11765
11766 * prevent-dup-ip
11767 Ensure HAProxy's default behavior is enforced on a server: prevent re-using
11768 an IP address already set to a server in the same backend and sharing the
11769 same fqdn.
11770 This is the opposite of allow-dup-ip.
11771
11772 Example:
11773 backend b_myapp
11774 default-server init-addr none resolvers dns
11775 server s1 myapp.example.com:80 check resolve-opts allow-dup-ip
11776 server s2 myapp.example.com:81 check resolve-opts allow-dup-ip
11777
11778 With the option allow-dup-ip set:
11779 * if the nameserver returns a single IP address, then both servers will use
11780 it
11781 * If the nameserver returns 2 IP addresses, then each server will pick up a
11782 different address
11783
11784 Default value: not set
11785
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011786resolve-prefer <family>
11787 When DNS resolution is enabled for a server and multiple IP addresses from
11788 different families are returned, HAProxy will prefer using an IP address
11789 from the family mentioned in the "resolve-prefer" parameter.
11790 Available families: "ipv4" and "ipv6"
11791
Baptiste Assmannc4aabae2015-08-04 22:43:06 +020011792 Default value: ipv6
11793
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011794 Example:
11795
11796 server s1 app1.domain.com:80 resolvers mydns resolve-prefer ipv6
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011797
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011798resolve-net <network>[,<network[,...]]
11799 This options prioritize th choice of an ip address matching a network. This is
11800 useful with clouds to prefer a local ip. In some cases, a cloud high
Tim Düsterhus4896c442016-11-29 02:15:19 +010011801 availability service can be announced with many ip addresses on many
Davor Ocelice9ed2812017-12-25 17:49:28 +010011802 different datacenters. The latency between datacenter is not negligible, so
11803 this patch permits to prefer a local datacenter. If no address matches the
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011804 configured network, another address is selected.
11805
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011806 Example:
11807
11808 server s1 app1.domain.com:80 resolvers mydns resolve-net 10.0.0.0/8
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011809
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011810resolvers <id>
11811 Points to an existing "resolvers" section to resolve current server's
11812 hostname.
11813
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011814 Example:
11815
11816 server s1 app1.domain.com:80 check resolvers mydns
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011817
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011818 See also section 5.3
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011819
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010011820send-proxy
11821 The "send-proxy" parameter enforces use of the PROXY protocol over any
11822 connection established to this server. The PROXY protocol informs the other
11823 end about the layer 3/4 addresses of the incoming connection, so that it can
11824 know the client's address or the public address it accessed to, whatever the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010011825 upper layer protocol. For connections accepted by an "accept-proxy" or
11826 "accept-netscaler-cip" listener, the advertised address will be used. Only
11827 TCPv4 and TCPv6 address families are supported. Other families such as
11828 Unix sockets, will report an UNKNOWN family. Servers using this option can
11829 fully be chained to another instance of haproxy listening with an
11830 "accept-proxy" setting. This setting must not be used if the server isn't
11831 aware of the protocol. When health checks are sent to the server, the PROXY
11832 protocol is automatically used when this option is set, unless there is an
11833 explicit "port" or "addr" directive, in which case an explicit
11834 "check-send-proxy" directive would also be needed to use the PROXY protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010011835 See also the "no-send-proxy" option of this section and "accept-proxy" and
11836 "accept-netscaler-cip" option of the "bind" keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010011837
David Safb76832014-05-08 23:42:08 -040011838send-proxy-v2
11839 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
11840 over any connection established to this server. The PROXY protocol informs
11841 the other end about the layer 3/4 addresses of the incoming connection, so
11842 that it can know the client's address or the public address it accessed to,
Emmanuel Hocdet404d9782017-10-24 10:55:14 +020011843 whatever the upper layer protocol. It also send ALPN information if an alpn
11844 have been negotiated. This setting must not be used if the server isn't aware
11845 of this version of the protocol. See also the "no-send-proxy-v2" option of
11846 this section and send-proxy" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040011847
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010011848proxy-v2-options <option>[,<option>]*
11849 The "proxy-v2-options" parameter add option to send in PROXY protocol version
11850 2 when "send-proxy-v2" is used. Options available are "ssl" (see also
Emmanuel Hocdetfa8d0f12018-02-01 15:53:52 +010011851 send-proxy-v2-ssl), "cert-cn" (see also "send-proxy-v2-ssl-cn"), "ssl-cipher":
11852 name of the used cipher, "cert-sig": signature algorithm of the used
Emmanuel Hocdet253c3b72018-02-01 18:29:59 +010011853 certificate, "cert-key": key algorithm of the used certificate), "authority":
11854 host name value passed by the client (only sni from a tls connection is
Emmanuel Hocdet4399c752018-02-05 15:26:43 +010011855 supported), "crc32c": checksum of the proxy protocol v2 header.
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010011856
David Safb76832014-05-08 23:42:08 -040011857send-proxy-v2-ssl
11858 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
11859 2 over any connection established to this server. The PROXY protocol informs
11860 the other end about the layer 3/4 addresses of the incoming connection, so
11861 that it can know the client's address or the public address it accessed to,
11862 whatever the upper layer protocol. In addition, the SSL information extension
11863 of the PROXY protocol is added to the PROXY protocol header. This setting
11864 must not be used if the server isn't aware of this version of the protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010011865 See also the "no-send-proxy-v2-ssl" option of this section and the
11866 "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040011867
11868send-proxy-v2-ssl-cn
11869 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
11870 2 over any connection established to this server. The PROXY protocol informs
11871 the other end about the layer 3/4 addresses of the incoming connection, so
11872 that it can know the client's address or the public address it accessed to,
11873 whatever the upper layer protocol. In addition, the SSL information extension
11874 of the PROXY protocol, along along with the Common Name from the subject of
11875 the client certificate (if any), is added to the PROXY protocol header. This
11876 setting must not be used if the server isn't aware of this version of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011877 protocol. See also the "no-send-proxy-v2-ssl-cn" option of this section and
11878 the "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040011879
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011880slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011881 The "slowstart" parameter for a server accepts a value in milliseconds which
11882 indicates after how long a server which has just come back up will run at
11883 full speed. Just as with every other time-based parameter, it can be entered
11884 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
11885 linearly from 0 to 100% during this time. The limitation applies to two
11886 parameters :
11887
11888 - maxconn: the number of connections accepted by the server will grow from 1
11889 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
11890
11891 - weight: when the backend uses a dynamic weighted algorithm, the weight
11892 grows linearly from 1 to 100%. In this case, the weight is updated at every
11893 health-check. For this reason, it is important that the "inter" parameter
11894 is smaller than the "slowstart", in order to maximize the number of steps.
11895
11896 The slowstart never applies when haproxy starts, otherwise it would cause
11897 trouble to running servers. It only applies when a server has been previously
11898 seen as failed.
11899
Willy Tarreau732eac42015-07-09 11:40:25 +020011900sni <expression>
11901 The "sni" parameter evaluates the sample fetch expression, converts it to a
11902 string and uses the result as the host name sent in the SNI TLS extension to
11903 the server. A typical use case is to send the SNI received from the client in
11904 a bridged HTTPS scenario, using the "ssl_fc_sni" sample fetch for the
Willy Tarreau2ab88672017-07-05 18:23:03 +020011905 expression, though alternatives such as req.hdr(host) can also make sense. If
11906 "verify required" is set (which is the recommended setting), the resulting
Willy Tarreauad92a9a2017-07-28 11:38:41 +020011907 name will also be matched against the server certificate's names. See the
11908 "verify" directive for more details.
Willy Tarreau732eac42015-07-09 11:40:25 +020011909
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011910source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020011911source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011912source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011913 The "source" parameter sets the source address which will be used when
11914 connecting to the server. It follows the exact same parameters and principle
11915 as the backend "source" keyword, except that it only applies to the server
11916 referencing it. Please consult the "source" keyword for details.
11917
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011918 Additionally, the "source" statement on a server line allows one to specify a
11919 source port range by indicating the lower and higher bounds delimited by a
11920 dash ('-'). Some operating systems might require a valid IP address when a
11921 source port range is specified. It is permitted to have the same IP/range for
11922 several servers. Doing so makes it possible to bypass the maximum of 64k
11923 total concurrent connections. The limit will then reach 64k connections per
11924 server.
11925
Lukas Tribus7d56c6d2016-09-13 09:51:15 +000011926 Since Linux 4.2/libc 2.23 IP_BIND_ADDRESS_NO_PORT is set for connections
11927 specifying the source address without port(s).
11928
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011929ssl
Willy Tarreau44f65392013-06-25 07:56:20 +020011930 This option enables SSL ciphering on outgoing connections to the server. It
11931 is critical to verify server certificates using "verify" when using SSL to
11932 connect to servers, otherwise the communication is prone to trivial man in
11933 the-middle attacks rendering SSL useless. When this option is used, health
11934 checks are automatically sent in SSL too unless there is a "port" or an
11935 "addr" directive indicating the check should be sent to a different location.
Frédéric Lécailled2376272017-03-21 18:52:12 +010011936 See the "no-ssl" to disable "ssl" option and "check-ssl" option to force
11937 SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011938
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011939ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11940 This option enforces use of <version> or lower when SSL is used to communicate
11941 with the server. This option is also available on global statement
11942 "ssl-default-server-options". See also "ssl-min-ver".
11943
11944ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11945 This option enforces use of <version> or upper when SSL is used to communicate
11946 with the server. This option is also available on global statement
11947 "ssl-default-server-options". See also "ssl-max-ver".
11948
Frédéric Lécailled2376272017-03-21 18:52:12 +010011949ssl-reuse
11950 This option may be used as "server" setting to reset any "no-ssl-reuse"
11951 setting which would have been inherited from "default-server" directive as
11952 default value.
11953 It may also be used as "default-server" setting to reset any previous
11954 "default-server" "no-ssl-reuse" setting.
11955
11956stick
11957 This option may be used as "server" setting to reset any "non-stick"
11958 setting which would have been inherited from "default-server" directive as
11959 default value.
11960 It may also be used as "default-server" setting to reset any previous
11961 "default-server" "non-stick" setting.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011962
Willy Tarreau163d4622015-10-13 16:16:41 +020011963tcp-ut <delay>
11964 Sets the TCP User Timeout for all outgoing connections to this server. This
11965 option is available on Linux since version 2.6.37. It allows haproxy to
11966 configure a timeout for sockets which contain data not receiving an
Davor Ocelice9ed2812017-12-25 17:49:28 +010011967 acknowledgment for the configured delay. This is especially useful on
Willy Tarreau163d4622015-10-13 16:16:41 +020011968 long-lived connections experiencing long idle periods such as remote
11969 terminals or database connection pools, where the client and server timeouts
11970 must remain high to allow a long period of idle, but where it is important to
11971 detect that the server has disappeared in order to release all resources
11972 associated with its connection (and the client's session). One typical use
11973 case is also to force dead server connections to die when health checks are
11974 too slow or during a soft reload since health checks are then disabled. The
11975 argument is a delay expressed in milliseconds by default. This only works for
11976 regular TCP connections, and is ignored for other protocols.
11977
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011978track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +020011979 This option enables ability to set the current state of the server by tracking
11980 another one. It is possible to track a server which itself tracks another
11981 server, provided that at the end of the chain, a server has health checks
11982 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011983 used, it has to be enabled on both proxies.
11984
Frédéric Lécailled2376272017-03-21 18:52:12 +010011985tls-tickets
11986 This option may be used as "server" setting to reset any "no-tls-tickets"
11987 setting which would have been inherited from "default-server" directive as
11988 default value.
11989 It may also be used as "default-server" setting to reset any previous
11990 "default-server" "no-tlsv-tickets" setting.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011991
Emeric Brunef42d922012-10-11 16:11:36 +020011992verify [none|required]
11993 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +010011994 to 'none', server certificate is not verified. In the other case, The
Willy Tarreauad92a9a2017-07-28 11:38:41 +020011995 certificate provided by the server is verified using CAs from 'ca-file' and
11996 optional CRLs from 'crl-file' after having checked that the names provided in
Davor Ocelice9ed2812017-12-25 17:49:28 +010011997 the certificate's subject and subjectAlternateNames attributes match either
Willy Tarreauad92a9a2017-07-28 11:38:41 +020011998 the name passed using the "sni" directive, or if not provided, the static
11999 host name passed using the "verifyhost" directive. When no name is found, the
12000 certificate's names are ignored. For this reason, without SNI it's important
12001 to use "verifyhost". On verification failure the handshake is aborted. It is
12002 critically important to verify server certificates when using SSL to connect
12003 to servers, otherwise the communication is prone to trivial man-in-the-middle
12004 attacks rendering SSL totally useless. Unless "ssl_server_verify" appears in
12005 the global section, "verify" is set to "required" by default.
Emeric Brunef42d922012-10-11 16:11:36 +020012006
Evan Broderbe554312013-06-27 00:05:25 -070012007verifyhost <hostname>
12008 This setting is only available when support for OpenSSL was built in, and
Willy Tarreauad92a9a2017-07-28 11:38:41 +020012009 only takes effect if 'verify required' is also specified. This directive sets
12010 a default static hostname to check the server's certificate against when no
12011 SNI was used to connect to the server. If SNI is not used, this is the only
12012 way to enable hostname verification. This static hostname, when set, will
12013 also be used for health checks (which cannot provide an SNI value). If none
12014 of the hostnames in the certificate match the specified hostname, the
12015 handshake is aborted. The hostnames in the server-provided certificate may
12016 include wildcards. See also "verify", "sni" and "no-verifyhost" options.
Evan Broderbe554312013-06-27 00:05:25 -070012017
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012018weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012019 The "weight" parameter is used to adjust the server's weight relative to
12020 other servers. All servers will receive a load proportional to their weight
12021 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +020012022 load. The default weight is 1, and the maximal value is 256. A value of 0
12023 means the server will not participate in load-balancing but will still accept
12024 persistent connections. If this parameter is used to distribute the load
12025 according to server's capacity, it is recommended to start with values which
12026 can both grow and shrink, for instance between 10 and 100 to leave enough
12027 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012028
12029
Cyril Bonté46175dd2015-07-02 22:45:32 +0200120305.3. Server IP address resolution using DNS
12031-------------------------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012032
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012033HAProxy allows using a host name on the server line to retrieve its IP address
12034using name servers. By default, HAProxy resolves the name when parsing the
12035configuration file, at startup and cache the result for the process' life.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012036This is not sufficient in some cases, such as in Amazon where a server's IP
12037can change after a reboot or an ELB Virtual IP can change based on current
12038workload.
12039This chapter describes how HAProxy can be configured to process server's name
12040resolution at run time.
12041Whether run time server name resolution has been enable or not, HAProxy will
12042carry on doing the first resolution when parsing the configuration.
12043
12044
Cyril Bonté46175dd2015-07-02 22:45:32 +0200120455.3.1. Global overview
12046----------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012047
12048As we've seen in introduction, name resolution in HAProxy occurs at two
12049different steps of the process life:
12050
12051 1. when starting up, HAProxy parses the server line definition and matches a
12052 host name. It uses libc functions to get the host name resolved. This
12053 resolution relies on /etc/resolv.conf file.
12054
Christopher Faulet67957bd2017-09-27 11:00:59 +020012055 2. at run time, HAProxy performs periodically name resolutions for servers
12056 requiring DNS resolutions.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012057
12058A few other events can trigger a name resolution at run time:
12059 - when a server's health check ends up in a connection timeout: this may be
12060 because the server has a new IP address. So we need to trigger a name
12061 resolution to know this new IP.
12062
Christopher Faulet67957bd2017-09-27 11:00:59 +020012063When using resolvers, the server name can either be a hostname, or a SRV label.
Davor Ocelice9ed2812017-12-25 17:49:28 +010012064HAProxy considers anything that starts with an underscore as a SRV label. If a
Christopher Faulet67957bd2017-09-27 11:00:59 +020012065SRV label is specified, then the corresponding SRV records will be retrieved
12066from the DNS server, and the provided hostnames will be used. The SRV label
12067will be checked periodically, and if any server are added or removed, haproxy
12068will automatically do the same.
Olivier Houchardecfa18d2017-08-07 17:30:03 +020012069
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012070A few things important to notice:
12071 - all the name servers are queried in the mean time. HAProxy will process the
12072 first valid response.
12073
12074 - a resolution is considered as invalid (NX, timeout, refused), when all the
12075 servers return an error.
12076
12077
Cyril Bonté46175dd2015-07-02 22:45:32 +0200120785.3.2. The resolvers section
12079----------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012080
12081This section is dedicated to host information related to name resolution in
Christopher Faulet67957bd2017-09-27 11:00:59 +020012082HAProxy. There can be as many as resolvers section as needed. Each section can
12083contain many name servers.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012084
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012085When multiple name servers are configured in a resolvers section, then HAProxy
12086uses the first valid response. In case of invalid responses, only the last one
12087is treated. Purpose is to give the chance to a slow server to deliver a valid
12088answer after a fast faulty or outdated server.
12089
12090When each server returns a different error type, then only the last error is
Christopher Faulet67957bd2017-09-27 11:00:59 +020012091used by HAProxy. The following processing is applied on this error:
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012092
Christopher Faulet67957bd2017-09-27 11:00:59 +020012093 1. HAProxy retries the same DNS query with a new query type. The A queries are
12094 switch to AAAA or the opposite. SRV queries are not concerned here. Timeout
12095 errors are also excluded.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012096
Christopher Faulet67957bd2017-09-27 11:00:59 +020012097 2. When the fallback on the query type was done (or not applicable), HAProxy
12098 retries the original DNS query, with the preferred query type.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012099
Christopher Faulet67957bd2017-09-27 11:00:59 +020012100 3. HAProxy retries previous steps <resolve_retires> times. If no valid
12101 response is received after that, it stops the DNS resolution and reports
12102 the error.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012103
Christopher Faulet67957bd2017-09-27 11:00:59 +020012104For example, with 2 name servers configured in a resolvers section, the
12105following scenarios are possible:
12106
12107 - First response is valid and is applied directly, second response is
12108 ignored
12109
12110 - First response is invalid and second one is valid, then second response is
12111 applied
12112
12113 - First response is a NX domain and second one a truncated response, then
12114 HAProxy retries the query with a new type
12115
12116 - First response is a NX domain and second one is a timeout, then HAProxy
12117 retries the query with a new type
12118
12119 - Query timed out for both name servers, then HAProxy retries it with the
12120 same query type
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012121
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012122As a DNS server may not answer all the IPs in one DNS request, haproxy keeps
12123a cache of previous answers, an answer will be considered obsolete after
Christopher Faulet67957bd2017-09-27 11:00:59 +020012124<hold obsolete> seconds without the IP returned.
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012125
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012126
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012127resolvers <resolvers id>
Davor Ocelice9ed2812017-12-25 17:49:28 +010012128 Creates a new name server list labeled <resolvers id>
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012129
12130A resolvers section accept the following parameters:
12131
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020012132accepted_payload_size <nb>
Davor Ocelice9ed2812017-12-25 17:49:28 +010012133 Defines the maximum payload size accepted by HAProxy and announced to all the
Christopher Faulet67957bd2017-09-27 11:00:59 +020012134 name servers configured in this resolvers section.
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020012135 <nb> is in bytes. If not set, HAProxy announces 512. (minimal value defined
12136 by RFC 6891)
12137
Baptiste Assmann9d8dbbc2017-08-18 23:35:08 +020012138 Note: the maximum allowed value is 8192.
12139
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012140nameserver <id> <ip>:<port>
12141 DNS server description:
12142 <id> : label of the server, should be unique
12143 <ip> : IP address of the server
12144 <port> : port where the DNS service actually runs
12145
Ben Draut44e609b2018-05-29 15:40:08 -060012146parse-resolv-conf
12147 Adds all nameservers found in /etc/resolv.conf to this resolvers nameservers
12148 list. Ordered as if each nameserver in /etc/resolv.conf was individually
12149 placed in the resolvers section in place of this directive.
12150
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012151hold <status> <period>
12152 Defines <period> during which the last name resolution should be kept based
12153 on last resolution <status>
Baptiste Assmann987e16d2016-11-02 22:23:31 +010012154 <status> : last name resolution status. Acceptable values are "nx",
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012155 "other", "refused", "timeout", "valid", "obsolete".
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012156 <period> : interval between two successive name resolution when the last
12157 answer was in <status>. It follows the HAProxy time format.
12158 <period> is in milliseconds by default.
12159
Baptiste Assmann686408b2017-08-18 10:15:42 +020012160 Default value is 10s for "valid", 0s for "obsolete" and 30s for others.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012161
Christopher Faulet67957bd2017-09-27 11:00:59 +020012162resolution_pool_size <nb> (deprecated)
Baptiste Assmann201c07f2017-05-22 15:17:15 +020012163 Defines the number of resolutions available in the pool for this resolvers.
12164 If not defines, it defaults to 64. If your configuration requires more than
12165 <nb>, then HAProxy will return an error when parsing the configuration.
12166
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012167resolve_retries <nb>
12168 Defines the number <nb> of queries to send to resolve a server name before
12169 giving up.
12170 Default value: 3
12171
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012172 A retry occurs on name server timeout or when the full sequence of DNS query
12173 type failover is over and we need to start up from the default ANY query
12174 type.
12175
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012176timeout <event> <time>
12177 Defines timeouts related to name resolution
12178 <event> : the event on which the <time> timeout period applies to.
12179 events available are:
Christopher Faulet67957bd2017-09-27 11:00:59 +020012180 - resolve : default time to trigger name resolutions when no
12181 other time applied.
12182 Default value: 1s
12183 - retry : time between two DNS queries, when no valid response
12184 have been received.
12185 Default value: 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012186 <time> : time related to the event. It follows the HAProxy time format.
12187 <time> is expressed in milliseconds.
12188
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012189 Example:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012190
12191 resolvers mydns
12192 nameserver dns1 10.0.0.1:53
12193 nameserver dns2 10.0.0.2:53
Ben Draut44e609b2018-05-29 15:40:08 -060012194 parse-resolv-conf
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012195 resolve_retries 3
Christopher Faulet67957bd2017-09-27 11:00:59 +020012196 timeout resolve 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012197 timeout retry 1s
Baptiste Assmann987e16d2016-11-02 22:23:31 +010012198 hold other 30s
12199 hold refused 30s
12200 hold nx 30s
12201 hold timeout 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012202 hold valid 10s
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012203 hold obsolete 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012204
12205
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200122066. HTTP header manipulation
12207---------------------------
12208
12209In HTTP mode, it is possible to rewrite, add or delete some of the request and
12210response headers based on regular expressions. It is also possible to block a
12211request or a response if a particular header matches a regular expression,
12212which is enough to stop most elementary protocol attacks, and to protect
Willy Tarreau70dffda2014-01-30 03:07:23 +010012213against information leak from the internal network.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012214
Willy Tarreau70dffda2014-01-30 03:07:23 +010012215If HAProxy encounters an "Informational Response" (status code 1xx), it is able
12216to process all rsp* rules which can allow, deny, rewrite or delete a header,
12217but it will refuse to add a header to any such messages as this is not
12218HTTP-compliant. The reason for still processing headers in such responses is to
12219stop and/or fix any possible information leak which may happen, for instance
12220because another downstream equipment would unconditionally add a header, or if
12221a server name appears there. When such messages are seen, normal processing
12222still occurs on the next non-informational messages.
Willy Tarreau816b9792009-09-15 21:25:21 +020012223
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012224This section covers common usage of the following keywords, described in detail
12225in section 4.2 :
12226
12227 - reqadd <string>
12228 - reqallow <search>
12229 - reqiallow <search>
12230 - reqdel <search>
12231 - reqidel <search>
12232 - reqdeny <search>
12233 - reqideny <search>
12234 - reqpass <search>
12235 - reqipass <search>
12236 - reqrep <search> <replace>
12237 - reqirep <search> <replace>
12238 - reqtarpit <search>
12239 - reqitarpit <search>
12240 - rspadd <string>
12241 - rspdel <search>
12242 - rspidel <search>
12243 - rspdeny <search>
12244 - rspideny <search>
12245 - rsprep <search> <replace>
12246 - rspirep <search> <replace>
12247
12248With all these keywords, the same conventions are used. The <search> parameter
12249is a POSIX extended regular expression (regex) which supports grouping through
12250parenthesis (without the backslash). Spaces and other delimiters must be
12251prefixed with a backslash ('\') to avoid confusion with a field delimiter.
12252Other characters may be prefixed with a backslash to change their meaning :
12253
12254 \t for a tab
12255 \r for a carriage return (CR)
12256 \n for a new line (LF)
12257 \ to mark a space and differentiate it from a delimiter
12258 \# to mark a sharp and differentiate it from a comment
12259 \\ to use a backslash in a regex
12260 \\\\ to use a backslash in the text (*2 for regex, *2 for haproxy)
12261 \xXX to write the ASCII hex code XX as in the C language
12262
12263The <replace> parameter contains the string to be used to replace the largest
12264portion of text matching the regex. It can make use of the special characters
12265above, and can reference a substring which is delimited by parenthesis in the
12266regex, by writing a backslash ('\') immediately followed by one digit from 0 to
122679 indicating the group position (0 designating the entire line). This practice
12268is very common to users of the "sed" program.
12269
12270The <string> parameter represents the string which will systematically be added
12271after the last header line. It can also use special character sequences above.
12272
12273Notes related to these keywords :
12274---------------------------------
12275 - these keywords are not always convenient to allow/deny based on header
12276 contents. It is strongly recommended to use ACLs with the "block" keyword
12277 instead, resulting in far more flexible and manageable rules.
12278
12279 - lines are always considered as a whole. It is not possible to reference
12280 a header name only or a value only. This is important because of the way
12281 headers are written (notably the number of spaces after the colon).
12282
12283 - the first line is always considered as a header, which makes it possible to
12284 rewrite or filter HTTP requests URIs or response codes, but in turn makes
12285 it harder to distinguish between headers and request line. The regex prefix
12286 ^[^\ \t]*[\ \t] matches any HTTP method followed by a space, and the prefix
12287 ^[^ \t:]*: matches any header name followed by a colon.
12288
12289 - for performances reasons, the number of characters added to a request or to
12290 a response is limited at build time to values between 1 and 4 kB. This
12291 should normally be far more than enough for most usages. If it is too short
12292 on occasional usages, it is possible to gain some space by removing some
12293 useless headers before adding new ones.
12294
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012295 - keywords beginning with "reqi" and "rspi" are the same as their counterpart
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012296 without the 'i' letter except that they ignore case when matching patterns.
12297
12298 - when a request passes through a frontend then a backend, all req* rules
12299 from the frontend will be evaluated, then all req* rules from the backend
12300 will be evaluated. The reverse path is applied to responses.
12301
12302 - req* statements are applied after "block" statements, so that "block" is
12303 always the first one, but before "use_backend" in order to permit rewriting
Willy Tarreaud72758d2010-01-12 10:42:19 +010012304 before switching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012305
12306
Willy Tarreau74ca5042013-06-11 23:12:07 +0200123077. Using ACLs and fetching samples
12308----------------------------------
12309
Davor Ocelice9ed2812017-12-25 17:49:28 +010012310HAProxy is capable of extracting data from request or response streams, from
Willy Tarreau74ca5042013-06-11 23:12:07 +020012311client or server information, from tables, environmental information etc...
12312The action of extracting such data is called fetching a sample. Once retrieved,
12313these samples may be used for various purposes such as a key to a stick-table,
12314but most common usages consist in matching them against predefined constant
12315data called patterns.
12316
12317
123187.1. ACL basics
12319---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012320
12321The use of Access Control Lists (ACL) provides a flexible solution to perform
12322content switching and generally to take decisions based on content extracted
12323from the request, the response or any environmental status. The principle is
12324simple :
12325
Willy Tarreau74ca5042013-06-11 23:12:07 +020012326 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012327 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +020012328 - apply one or multiple pattern matching methods on this sample
12329 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012330
Willy Tarreau74ca5042013-06-11 23:12:07 +020012331The actions generally consist in blocking a request, selecting a backend, or
12332adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012333
12334In order to define a test, the "acl" keyword is used. The syntax is :
12335
Willy Tarreau74ca5042013-06-11 23:12:07 +020012336 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012337
12338This creates a new ACL <aclname> or completes an existing one with new tests.
12339Those tests apply to the portion of request/response specified in <criterion>
12340and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012341an operator which may be specified before the set of values. Optionally some
12342conversion operators may be applied to the sample, and they will be specified
12343as a comma-delimited list of keywords just after the first keyword. The values
12344are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012345
12346ACL names must be formed from upper and lower case letters, digits, '-' (dash),
12347'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
12348which means that "my_acl" and "My_Acl" are two different ACLs.
12349
12350There is no enforced limit to the number of ACLs. The unused ones do not affect
12351performance, they just consume a small amount of memory.
12352
Willy Tarreau74ca5042013-06-11 23:12:07 +020012353The criterion generally is the name of a sample fetch method, or one of its ACL
12354specific declinations. The default test method is implied by the output type of
12355this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012356methods of a same sample fetch method. The sample fetch methods are the only
12357ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012358
12359Sample fetch methods return data which can be of the following types :
12360 - boolean
12361 - integer (signed or unsigned)
12362 - IPv4 or IPv6 address
12363 - string
12364 - data block
12365
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012366Converters transform any of these data into any of these. For example, some
12367converters might convert a string to a lower-case string while other ones
12368would turn a string to an IPv4 address, or apply a netmask to an IP address.
12369The resulting sample is of the type of the last converter applied to the list,
12370which defaults to the type of the sample fetch method.
12371
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012372Each sample or converter returns data of a specific type, specified with its
12373keyword in this documentation. When an ACL is declared using a standard sample
12374fetch method, certain types automatically involved a default matching method
12375which are summarized in the table below :
12376
12377 +---------------------+-----------------+
12378 | Sample or converter | Default |
12379 | output type | matching method |
12380 +---------------------+-----------------+
12381 | boolean | bool |
12382 +---------------------+-----------------+
12383 | integer | int |
12384 +---------------------+-----------------+
12385 | ip | ip |
12386 +---------------------+-----------------+
12387 | string | str |
12388 +---------------------+-----------------+
12389 | binary | none, use "-m" |
12390 +---------------------+-----------------+
12391
12392Note that in order to match a binary samples, it is mandatory to specify a
12393matching method, see below.
12394
Willy Tarreau74ca5042013-06-11 23:12:07 +020012395The ACL engine can match these types against patterns of the following types :
12396 - boolean
12397 - integer or integer range
12398 - IP address / network
12399 - string (exact, substring, suffix, prefix, subdir, domain)
12400 - regular expression
12401 - hex block
12402
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012403The following ACL flags are currently supported :
12404
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012405 -i : ignore case during matching of all subsequent patterns.
12406 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012407 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010012408 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +010012409 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +010012410 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +020012411 -- : force end of flags. Useful when a string looks like one of the flags.
12412
Willy Tarreau74ca5042013-06-11 23:12:07 +020012413The "-f" flag is followed by the name of a file from which all lines will be
12414read as individual values. It is even possible to pass multiple "-f" arguments
12415if the patterns are to be loaded from multiple files. Empty lines as well as
12416lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
12417will be stripped. If it is absolutely necessary to insert a valid pattern
12418beginning with a sharp, just prefix it with a space so that it is not taken for
12419a comment. Depending on the data type and match method, haproxy may load the
12420lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
12421exact string matching. In this case, duplicates will automatically be removed.
12422
Thierry FOURNIER9860c412014-01-29 14:23:29 +010012423The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
12424parsed as two column file. The first column contains the patterns used by the
12425ACL, and the second column contain the samples. The sample can be used later by
12426a map. This can be useful in some rare cases where an ACL would just be used to
12427check for the existence of a pattern in a map before a mapping is applied.
12428
Thierry FOURNIER3534d882014-01-20 17:01:44 +010012429The "-u" flag forces the unique id of the ACL. This unique id is used with the
12430socket interface to identify ACL and dynamically change its values. Note that a
12431file is always identified by its name even if an id is set.
12432
Willy Tarreau74ca5042013-06-11 23:12:07 +020012433Also, note that the "-i" flag applies to subsequent entries and not to entries
12434loaded from files preceding it. For instance :
12435
12436 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
12437
12438In this example, each line of "exact-ua.lst" will be exactly matched against
12439the "user-agent" header of the request. Then each line of "generic-ua" will be
12440case-insensitively matched. Then the word "test" will be insensitively matched
12441as well.
12442
12443The "-m" flag is used to select a specific pattern matching method on the input
12444sample. All ACL-specific criteria imply a pattern matching method and generally
12445do not need this flag. However, this flag is useful with generic sample fetch
12446methods to describe how they're going to be matched against the patterns. This
12447is required for sample fetches which return data type for which there is no
Davor Ocelice9ed2812017-12-25 17:49:28 +010012448obvious matching method (e.g. string or binary). When "-m" is specified and
Willy Tarreau74ca5042013-06-11 23:12:07 +020012449followed by a pattern matching method name, this method is used instead of the
12450default one for the criterion. This makes it possible to match contents in ways
12451that were not initially planned, or with sample fetch methods which return a
12452string. The matching method also affects the way the patterns are parsed.
12453
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010012454The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
12455By default, if the parser cannot parse ip address it considers that the parsed
12456string is maybe a domain name and try dns resolution. The flag "-n" disable this
12457resolution. It is useful for detecting malformed ip lists. Note that if the DNS
12458server is not reachable, the haproxy configuration parsing may last many minutes
12459waiting fir the timeout. During this time no error messages are displayed. The
12460flag "-n" disable this behavior. Note also that during the runtime, this
12461function is disabled for the dynamic acl modifications.
12462
Willy Tarreau74ca5042013-06-11 23:12:07 +020012463There are some restrictions however. Not all methods can be used with all
12464sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
12465be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +020012466
12467 - "found" : only check if the requested sample could be found in the stream,
12468 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +020012469 to pass any pattern to avoid confusion. This matching method is
12470 particularly useful to detect presence of certain contents such
12471 as headers, cookies, etc... even if they are empty and without
12472 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012473
12474 - "bool" : check the value as a boolean. It can only be applied to fetches
12475 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012476 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012477
12478 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +020012479 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012480
12481 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +020012482 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012483
Davor Ocelice9ed2812017-12-25 17:49:28 +010012484 - "bin" : match the contents against a hexadecimal string representing a
Willy Tarreau5adeda12013-03-31 22:13:34 +020012485 binary sequence. This may be used with binary or string samples.
12486
12487 - "len" : match the sample's length as an integer. This may be used with
12488 binary or string samples.
12489
Willy Tarreau74ca5042013-06-11 23:12:07 +020012490 - "str" : exact match : match the contents against a string. This may be
12491 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012492
Willy Tarreau74ca5042013-06-11 23:12:07 +020012493 - "sub" : substring match : check that the contents contain at least one of
12494 the provided string patterns. This may be used with binary or
12495 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012496
Willy Tarreau74ca5042013-06-11 23:12:07 +020012497 - "reg" : regex match : match the contents against a list of regular
12498 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012499
Willy Tarreau74ca5042013-06-11 23:12:07 +020012500 - "beg" : prefix match : check that the contents begin like the provided
12501 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012502
Willy Tarreau74ca5042013-06-11 23:12:07 +020012503 - "end" : suffix match : check that the contents end like the provided
12504 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012505
Willy Tarreau74ca5042013-06-11 23:12:07 +020012506 - "dir" : subdir match : check that a slash-delimited portion of the
12507 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012508 This may be used with binary or string samples.
12509
Willy Tarreau74ca5042013-06-11 23:12:07 +020012510 - "dom" : domain match : check that a dot-delimited portion of the contents
12511 exactly match one of the provided string patterns. This may be
12512 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012513
12514For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
12515request, it is possible to do :
12516
12517 acl jsess_present cook(JSESSIONID) -m found
12518
12519In order to apply a regular expression on the 500 first bytes of data in the
12520buffer, one would use the following acl :
12521
12522 acl script_tag payload(0,500) -m reg -i <script>
12523
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012524On systems where the regex library is much slower when using "-i", it is
12525possible to convert the sample to lowercase before matching, like this :
12526
12527 acl script_tag payload(0,500),lower -m reg <script>
12528
Willy Tarreau74ca5042013-06-11 23:12:07 +020012529All ACL-specific criteria imply a default matching method. Most often, these
12530criteria are composed by concatenating the name of the original sample fetch
12531method and the matching method. For example, "hdr_beg" applies the "beg" match
12532to samples retrieved using the "hdr" fetch method. Since all ACL-specific
12533criteria rely on a sample fetch method, it is always possible instead to use
12534the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012535
Willy Tarreau74ca5042013-06-11 23:12:07 +020012536If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012537the matching method is simply applied to the underlying sample fetch method.
12538For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012539
Willy Tarreau74ca5042013-06-11 23:12:07 +020012540 acl short_form hdr_beg(host) www.
12541 acl alternate1 hdr_beg(host) -m beg www.
12542 acl alternate2 hdr_dom(host) -m beg www.
12543 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012544
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012545
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012546The table below summarizes the compatibility matrix between sample or converter
12547types and the pattern types to fetch against. It indicates for each compatible
12548combination the name of the matching method to be used, surrounded with angle
12549brackets ">" and "<" when the method is the default one and will work by
12550default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +010012551
Willy Tarreau74ca5042013-06-11 23:12:07 +020012552 +-------------------------------------------------+
12553 | Input sample type |
12554 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012555 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012556 +----------------------+---------+---------+---------+---------+---------+
12557 | none (presence only) | found | found | found | found | found |
12558 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012559 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012560 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012561 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012562 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012563 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012564 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012565 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012566 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012567 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012568 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012569 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012570 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012571 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012572 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012573 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012574 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012575 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012576 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012577 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012578 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010012579 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020012580 +----------------------+---------+---------+---------+---------+---------+
12581 | hex block | | | | bin | bin |
12582 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +020012583
12584
Willy Tarreau74ca5042013-06-11 23:12:07 +0200125857.1.1. Matching booleans
12586------------------------
12587
12588In order to match a boolean, no value is needed and all values are ignored.
12589Boolean matching is used by default for all fetch methods of type "boolean".
12590When boolean matching is used, the fetched value is returned as-is, which means
12591that a boolean "true" will always match and a boolean "false" will never match.
12592
12593Boolean matching may also be enforced using "-m bool" on fetch methods which
12594return an integer value. Then, integer value 0 is converted to the boolean
12595"false" and all other values are converted to "true".
12596
Willy Tarreau6a06a402007-07-15 20:15:28 +020012597
Willy Tarreau74ca5042013-06-11 23:12:07 +0200125987.1.2. Matching integers
12599------------------------
12600
12601Integer matching applies by default to integer fetch methods. It can also be
12602enforced on boolean fetches using "-m int". In this case, "false" is converted
12603to the integer 0, and "true" is converted to the integer 1.
12604
12605Integer matching also supports integer ranges and operators. Note that integer
12606matching only applies to positive values. A range is a value expressed with a
12607lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012608
12609For instance, "1024:65535" is a valid range to represent a range of
12610unprivileged ports, and "1024:" would also work. "0:1023" is a valid
12611representation of privileged ports, and ":1023" would also work.
12612
Willy Tarreau62644772008-07-16 18:36:06 +020012613As a special case, some ACL functions support decimal numbers which are in fact
12614two integers separated by a dot. This is used with some version checks for
12615instance. All integer properties apply to those decimal numbers, including
12616ranges and operators.
12617
Willy Tarreau6a06a402007-07-15 20:15:28 +020012618For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +010012619operators with ranges does not make much sense and is strongly discouraged.
12620Similarly, it does not make much sense to perform order comparisons with a set
12621of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012622
Willy Tarreau0ba27502007-12-24 16:55:16 +010012623Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +020012624
12625 eq : true if the tested value equals at least one value
12626 ge : true if the tested value is greater than or equal to at least one value
12627 gt : true if the tested value is greater than at least one value
12628 le : true if the tested value is less than or equal to at least one value
12629 lt : true if the tested value is less than at least one value
12630
Willy Tarreau0ba27502007-12-24 16:55:16 +010012631For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +020012632
12633 acl negative-length hdr_val(content-length) lt 0
12634
Willy Tarreau62644772008-07-16 18:36:06 +020012635This one matches SSL versions between 3.0 and 3.1 (inclusive) :
12636
12637 acl sslv3 req_ssl_ver 3:3.1
12638
Willy Tarreau6a06a402007-07-15 20:15:28 +020012639
Willy Tarreau74ca5042013-06-11 23:12:07 +0200126407.1.3. Matching strings
12641-----------------------
12642
12643String matching applies to string or binary fetch methods, and exists in 6
12644different forms :
12645
12646 - exact match (-m str) : the extracted string must exactly match the
Davor Ocelice9ed2812017-12-25 17:49:28 +010012647 patterns;
Willy Tarreau74ca5042013-06-11 23:12:07 +020012648
12649 - substring match (-m sub) : the patterns are looked up inside the
Davor Ocelice9ed2812017-12-25 17:49:28 +010012650 extracted string, and the ACL matches if any of them is found inside;
Willy Tarreau74ca5042013-06-11 23:12:07 +020012651
12652 - prefix match (-m beg) : the patterns are compared with the beginning of
12653 the extracted string, and the ACL matches if any of them matches.
12654
12655 - suffix match (-m end) : the patterns are compared with the end of the
12656 extracted string, and the ACL matches if any of them matches.
12657
Baptiste Assmann33db6002016-03-06 23:32:10 +010012658 - subdir match (-m dir) : the patterns are looked up inside the extracted
Willy Tarreau74ca5042013-06-11 23:12:07 +020012659 string, delimited with slashes ("/"), and the ACL matches if any of them
12660 matches.
12661
12662 - domain match (-m dom) : the patterns are looked up inside the extracted
12663 string, delimited with dots ("."), and the ACL matches if any of them
12664 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012665
12666String matching applies to verbatim strings as they are passed, with the
12667exception of the backslash ("\") which makes it possible to escape some
12668characters such as the space. If the "-i" flag is passed before the first
12669string, then the matching will be performed ignoring the case. In order
12670to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +010012671before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020012672
12673
Willy Tarreau74ca5042013-06-11 23:12:07 +0200126747.1.4. Matching regular expressions (regexes)
12675---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020012676
12677Just like with string matching, regex matching applies to verbatim strings as
12678they are passed, with the exception of the backslash ("\") which makes it
12679possible to escape some characters such as the space. If the "-i" flag is
12680passed before the first regex, then the matching will be performed ignoring
12681the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +010012682the "--" flag before the first string. Same principle applies of course to
12683match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020012684
12685
Willy Tarreau74ca5042013-06-11 23:12:07 +0200126867.1.5. Matching arbitrary data blocks
12687-------------------------------------
12688
12689It is possible to match some extracted samples against a binary block which may
12690not safely be represented as a string. For this, the patterns must be passed as
12691a series of hexadecimal digits in an even number, when the match method is set
12692to binary. Each sequence of two digits will represent a byte. The hexadecimal
12693digits may be used upper or lower case.
12694
12695Example :
12696 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
12697 acl hello payload(0,6) -m bin 48656c6c6f0a
12698
12699
127007.1.6. Matching IPv4 and IPv6 addresses
12701---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020012702
12703IPv4 addresses values can be specified either as plain addresses or with a
12704netmask appended, in which case the IPv4 address matches whenever it is
12705within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012706host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +010012707difficult to read and debug configurations. If hostnames are used, you should
12708at least ensure that they are present in /etc/hosts so that the configuration
12709does not depend on any random DNS match at the moment the configuration is
12710parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012711
Daniel Schnellereba56342016-04-13 00:26:52 +020012712The dotted IPv4 address notation is supported in both regular as well as the
12713abbreviated form with all-0-octets omitted:
12714
12715 +------------------+------------------+------------------+
12716 | Example 1 | Example 2 | Example 3 |
12717 +------------------+------------------+------------------+
12718 | 192.168.0.1 | 10.0.0.12 | 127.0.0.1 |
12719 | 192.168.1 | 10.12 | 127.1 |
12720 | 192.168.0.1/22 | 10.0.0.12/8 | 127.0.0.1/8 |
12721 | 192.168.1/22 | 10.12/8 | 127.1/8 |
12722 +------------------+------------------+------------------+
12723
12724Notice that this is different from RFC 4632 CIDR address notation in which
12725192.168.42/24 would be equivalent to 192.168.42.0/24.
12726
Willy Tarreauceb4ac92012-04-28 00:41:46 +020012727IPv6 may be entered in their usual form, with or without a netmask appended.
12728Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
12729trouble with randomly resolved IP addresses, host names are never allowed in
12730IPv6 patterns.
12731
12732HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
12733following situations :
12734 - tested address is IPv4, pattern address is IPv4, the match applies
12735 in IPv4 using the supplied mask if any.
12736 - tested address is IPv6, pattern address is IPv6, the match applies
12737 in IPv6 using the supplied mask if any.
12738 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
12739 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
12740 ::IPV4 or ::ffff:IPV4, otherwise it fails.
12741 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
12742 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
12743 applied in IPv6 using the supplied IPv6 mask.
12744
Willy Tarreau74ca5042013-06-11 23:12:07 +020012745
127467.2. Using ACLs to form conditions
12747----------------------------------
12748
12749Some actions are only performed upon a valid condition. A condition is a
12750combination of ACLs with operators. 3 operators are supported :
12751
12752 - AND (implicit)
12753 - OR (explicit with the "or" keyword or the "||" operator)
12754 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +020012755
Willy Tarreau74ca5042013-06-11 23:12:07 +020012756A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +020012757
Willy Tarreau74ca5042013-06-11 23:12:07 +020012758 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +020012759
Willy Tarreau74ca5042013-06-11 23:12:07 +020012760Such conditions are generally used after an "if" or "unless" statement,
12761indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +020012762
Willy Tarreau74ca5042013-06-11 23:12:07 +020012763For instance, to block HTTP requests to the "*" URL with methods other than
12764"OPTIONS", as well as POST requests without content-length, and GET or HEAD
12765requests with a content-length greater than 0, and finally every request which
12766is not either GET/HEAD/POST/OPTIONS !
12767
12768 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030012769 http-request deny if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
12770 http-request deny if METH_GET HTTP_CONTENT
12771 http-request deny unless METH_GET or METH_POST or METH_OPTIONS
Willy Tarreau74ca5042013-06-11 23:12:07 +020012772
12773To select a different backend for requests to static contents on the "www" site
12774and to every request on the "img", "video", "download" and "ftp" hosts :
12775
12776 acl url_static path_beg /static /images /img /css
12777 acl url_static path_end .gif .png .jpg .css .js
12778 acl host_www hdr_beg(host) -i www
12779 acl host_static hdr_beg(host) -i img. video. download. ftp.
12780
Davor Ocelice9ed2812017-12-25 17:49:28 +010012781 # now use backend "static" for all static-only hosts, and for static URLs
Willy Tarreau74ca5042013-06-11 23:12:07 +020012782 # of host "www". Use backend "www" for the rest.
12783 use_backend static if host_static or host_www url_static
12784 use_backend www if host_www
12785
12786It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
12787expressions that are built on the fly without needing to be declared. They must
12788be enclosed between braces, with a space before and after each brace (because
12789the braces must be seen as independent words). Example :
12790
12791 The following rule :
12792
12793 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030012794 http-request deny if METH_POST missing_cl
Willy Tarreau74ca5042013-06-11 23:12:07 +020012795
12796 Can also be written that way :
12797
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030012798 http-request deny if METH_POST { hdr_cnt(Content-length) eq 0 }
Willy Tarreau74ca5042013-06-11 23:12:07 +020012799
12800It is generally not recommended to use this construct because it's a lot easier
12801to leave errors in the configuration when written that way. However, for very
12802simple rules matching only one source IP address for instance, it can make more
12803sense to use them than to declare ACLs with random names. Another example of
12804good use is the following :
12805
12806 With named ACLs :
12807
12808 acl site_dead nbsrv(dynamic) lt 2
12809 acl site_dead nbsrv(static) lt 2
12810 monitor fail if site_dead
12811
12812 With anonymous ACLs :
12813
12814 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
12815
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030012816See section 4.2 for detailed help on the "http-request deny" and "use_backend"
12817keywords.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012818
12819
128207.3. Fetching samples
12821---------------------
12822
12823Historically, sample fetch methods were only used to retrieve data to match
12824against patterns using ACLs. With the arrival of stick-tables, a new class of
12825sample fetch methods was created, most often sharing the same syntax as their
12826ACL counterpart. These sample fetch methods are also known as "fetches". As
12827of now, ACLs and fetches have converged. All ACL fetch methods have been made
12828available as fetch methods, and ACLs may use any sample fetch method as well.
12829
12830This section details all available sample fetch methods and their output type.
12831Some sample fetch methods have deprecated aliases that are used to maintain
12832compatibility with existing configurations. They are then explicitly marked as
12833deprecated and should not be used in new setups.
12834
12835The ACL derivatives are also indicated when available, with their respective
12836matching methods. These ones all have a well defined default pattern matching
12837method, so it is never necessary (though allowed) to pass the "-m" option to
12838indicate how the sample will be matched using ACLs.
12839
12840As indicated in the sample type versus matching compatibility matrix above,
12841when using a generic sample fetch method in an ACL, the "-m" option is
12842mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
12843the same keyword exists as an ACL keyword and as a standard fetch method, the
12844ACL engine will automatically pick the ACL-only one by default.
12845
12846Some of these keywords support one or multiple mandatory arguments, and one or
12847multiple optional arguments. These arguments are strongly typed and are checked
12848when the configuration is parsed so that there is no risk of running with an
Davor Ocelice9ed2812017-12-25 17:49:28 +010012849incorrect argument (e.g. an unresolved backend name). Fetch function arguments
12850are passed between parenthesis and are delimited by commas. When an argument
Willy Tarreau74ca5042013-06-11 23:12:07 +020012851is optional, it will be indicated below between square brackets ('[ ]'). When
12852all arguments are optional, the parenthesis may be omitted.
12853
12854Thus, the syntax of a standard sample fetch method is one of the following :
12855 - name
12856 - name(arg1)
12857 - name(arg1,arg2)
12858
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012859
128607.3.1. Converters
12861-----------------
12862
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012863Sample fetch methods may be combined with transformations to be applied on top
12864of the fetched sample (also called "converters"). These combinations form what
12865is called "sample expressions" and the result is a "sample". Initially this
12866was only supported by "stick on" and "stick store-request" directives but this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012867has now be extended to all places where samples may be used (ACLs, log-format,
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012868unique-id-format, add-header, ...).
12869
12870These transformations are enumerated as a series of specific keywords after the
12871sample fetch method. These keywords may equally be appended immediately after
12872the fetch keyword's argument, delimited by a comma. These keywords can also
Davor Ocelice9ed2812017-12-25 17:49:28 +010012873support some arguments (e.g. a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012874
Willy Tarreau97707872015-01-27 15:12:13 +010012875A certain category of converters are bitwise and arithmetic operators which
12876support performing basic operations on integers. Some bitwise operations are
12877supported (and, or, xor, cpl) and some arithmetic operations are supported
12878(add, sub, mul, div, mod, neg). Some comparators are provided (odd, even, not,
12879bool) which make it possible to report a match without having to write an ACL.
12880
Willy Tarreau74ca5042013-06-11 23:12:07 +020012881The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +010012882
Ben Shillitof25e8e52016-12-02 14:25:37 +00001288351d.single(<prop>[,<prop>*])
12884 Returns values for the properties requested as a string, where values are
12885 separated by the delimiter specified with "51degrees-property-separator".
12886 The device is identified using the User-Agent header passed to the
12887 converter. The function can be passed up to five property names, and if a
12888 property name can't be found, the value "NoData" is returned.
12889
12890 Example :
Davor Ocelice9ed2812017-12-25 17:49:28 +010012891 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request,
12892 # containing values for the three properties requested by using the
Ben Shillitof25e8e52016-12-02 14:25:37 +000012893 # User-Agent passed to the converter.
12894 frontend http-in
12895 bind *:8081
12896 default_backend servers
12897 http-request set-header X-51D-DeviceTypeMobileTablet \
12898 %[req.fhdr(User-Agent),51d.single(DeviceType,IsMobile,IsTablet)]
12899
Willy Tarreau97707872015-01-27 15:12:13 +010012900add(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012901 Adds <value> to the input value of type signed integer, and returns the
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012902 result as a signed integer. <value> can be a numeric value or a variable
Daniel Schneller0b547052016-03-21 20:46:57 +010012903 name. The name of the variable starts with an indication about its scope. The
12904 scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012905 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012906 "sess" : the variable is shared with the whole session
12907 "txn" : the variable is shared with the transaction (request and response)
12908 "req" : the variable is shared only during request processing
12909 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012910 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012911 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012912
12913and(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012914 Performs a bitwise "AND" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012915 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010012916 numeric value or a variable name. The name of the variable starts with an
12917 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012918 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012919 "sess" : the variable is shared with the whole session
12920 "txn" : the variable is shared with the transaction (request and response)
12921 "req" : the variable is shared only during request processing
12922 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012923 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012924 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012925
Holger Just1bfc24b2017-05-06 00:56:53 +020012926b64dec
12927 Converts (decodes) a base64 encoded input string to its binary
12928 representation. It performs the inverse operation of base64().
12929
Emeric Brun53d1a982014-04-30 18:21:37 +020012930base64
12931 Converts a binary input sample to a base64 string. It is used to log or
Davor Ocelice9ed2812017-12-25 17:49:28 +010012932 transfer binary content in a way that can be reliably transferred (e.g.
Emeric Brun53d1a982014-04-30 18:21:37 +020012933 an SSL ID can be copied in a header).
12934
Willy Tarreau97707872015-01-27 15:12:13 +010012935bool
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012936 Returns a boolean TRUE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010012937 non-null, otherwise returns FALSE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010012938 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010012939 presence of a flag).
12940
Emeric Brun54c4ac82014-11-03 15:32:43 +010012941bytes(<offset>[,<length>])
12942 Extracts some bytes from an input binary sample. The result is a binary
12943 sample starting at an offset (in bytes) of the original sample and
Tim Düsterhus4896c442016-11-29 02:15:19 +010012944 optionally truncated at the given length.
Emeric Brun54c4ac82014-11-03 15:32:43 +010012945
Willy Tarreau280f42b2018-02-19 15:34:12 +010012946concat([<start>],[<var>],[<end>])
12947 Concatenates up to 3 fields after the current sample which is then turned to
12948 a string. The first one, <start>, is a constant string, that will be appended
12949 immediately after the existing sample. It may be omitted if not used. The
12950 second one, <var>, is a variable name. The variable will be looked up, its
12951 contents converted to a string, and it will be appended immediately after the
12952 <first> part. If the variable is not found, nothing is appended. It may be
12953 omitted as well. The third field, <end> is a constant string that will be
12954 appended after the variable. It may also be omitted. Together, these elements
12955 allow to concatenate variables with delimiters to an existing set of
12956 variables. This can be used to build new variables made of a succession of
12957 other variables, such as colon-delimited varlues. Note that due to the config
12958 parser, it is not possible to use a comma nor a closing parenthesis as
12959 delimitors.
12960
12961 Example:
12962 tcp-request session set-var(sess.src) src
12963 tcp-request session set-var(sess.dn) ssl_c_s_dn
12964 tcp-request session set-var(txn.sig) str(),concat(<ip=,sess.ip,>),concat(<dn=,sess.dn,>)
12965 http-request set-header x-hap-sig %[var(txn.sig)]
12966
Willy Tarreau97707872015-01-27 15:12:13 +010012967cpl
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012968 Takes the input value of type signed integer, applies a ones-complement
12969 (flips all bits) and returns the result as an signed integer.
Willy Tarreau97707872015-01-27 15:12:13 +010012970
Willy Tarreau80599772015-01-20 19:35:24 +010012971crc32([<avalanche>])
12972 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32
12973 hash function. Optionally, it is possible to apply a full avalanche hash
12974 function to the output if the optional <avalanche> argument equals 1. This
12975 converter uses the same functions as used by the various hash-based load
12976 balancing algorithms, so it will provide exactly the same results. It is
12977 provided for compatibility with other software which want a CRC32 to be
12978 computed on some input keys, so it follows the most common implementation as
12979 found in Ethernet, Gzip, PNG, etc... It is slower than the other algorithms
12980 but may provide a better or at least less predictable distribution. It must
12981 not be used for security purposes as a 32-bit hash is trivial to break. See
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010012982 also "djb2", "sdbm", "wt6", "crc32c" and the "hash-type" directive.
12983
12984crc32c([<avalanche>])
12985 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32C
12986 hash function. Optionally, it is possible to apply a full avalanche hash
12987 function to the output if the optional <avalanche> argument equals 1. This
12988 converter uses the same functions as described in RFC4960, Appendix B [8].
12989 It is provided for compatibility with other software which want a CRC32C to be
12990 computed on some input keys. It is slower than the other algorithms and it must
12991 not be used for security purposes as a 32-bit hash is trivial to break. See
12992 also "djb2", "sdbm", "wt6", "crc32" and the "hash-type" directive.
Willy Tarreau80599772015-01-20 19:35:24 +010012993
David Carlier29b3ca32015-09-25 14:09:21 +010012994da-csv-conv(<prop>[,<prop>*])
David Carlier4542b102015-06-01 13:54:29 +020012995 Asks the DeviceAtlas converter to identify the User Agent string passed on
12996 input, and to emit a string made of the concatenation of the properties
12997 enumerated in argument, delimited by the separator defined by the global
12998 keyword "deviceatlas-property-separator", or by default the pipe character
David Carlier840b0242016-03-16 10:09:55 +000012999 ('|'). There's a limit of 12 different properties imposed by the haproxy
David Carlier4542b102015-06-01 13:54:29 +020013000 configuration language.
13001
13002 Example:
13003 frontend www
Cyril Bonté307ee1e2015-09-28 23:16:06 +020013004 bind *:8881
13005 default_backend servers
David Carlier840b0242016-03-16 10:09:55 +000013006 http-request set-header X-DeviceAtlas-Data %[req.fhdr(User-Agent),da-csv(primaryHardwareType,osName,osVersion,browserName,browserVersion,browserRenderingEngine)]
David Carlier4542b102015-06-01 13:54:29 +020013007
Thierry FOURNIER9687c772015-05-07 15:46:29 +020013008debug
13009 This converter is used as debug tool. It dumps on screen the content and the
13010 type of the input sample. The sample is returned as is on its output. This
13011 converter only exists when haproxy was built with debugging enabled.
13012
Willy Tarreau97707872015-01-27 15:12:13 +010013013div(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013014 Divides the input value of type signed integer by <value>, and returns the
13015 result as an signed integer. If <value> is null, the largest unsigned
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013016 integer is returned (typically 2^63-1). <value> can be a numeric value or a
Daniel Schneller0b547052016-03-21 20:46:57 +010013017 variable name. The name of the variable starts with an indication about its
13018 scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013019 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013020 "sess" : the variable is shared with the whole session
13021 "txn" : the variable is shared with the transaction (request and response)
13022 "req" : the variable is shared only during request processing
13023 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013024 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013025 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013026
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013027djb2([<avalanche>])
13028 Hashes a binary input sample into an unsigned 32-bit quantity using the DJB2
13029 hash function. Optionally, it is possible to apply a full avalanche hash
13030 function to the output if the optional <avalanche> argument equals 1. This
13031 converter uses the same functions as used by the various hash-based load
13032 balancing algorithms, so it will provide exactly the same results. It is
13033 mostly intended for debugging, but can be used as a stick-table entry to
13034 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013035 32-bit hash is trivial to break. See also "crc32", "sdbm", "wt6", "crc32c",
13036 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013037
Willy Tarreau97707872015-01-27 15:12:13 +010013038even
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013039 Returns a boolean TRUE if the input value of type signed integer is even
Willy Tarreau97707872015-01-27 15:12:13 +010013040 otherwise returns FALSE. It is functionally equivalent to "not,and(1),bool".
13041
Marcin Deranek9631a282018-04-16 14:30:46 +020013042field(<index>,<delimiters>[,<count>])
13043 Extracts the substring at the given index counting from the beginning
13044 (positive index) or from the end (negative index) considering given delimiters
13045 from an input string. Indexes start at 1 or -1 and delimiters are a string
13046 formatted list of chars. Optionally you can specify <count> of fields to
13047 extract (default: 1). Value of 0 indicates extraction of all remaining
13048 fields.
13049
13050 Example :
13051 str(f1_f2_f3__f5),field(5,_) # f5
13052 str(f1_f2_f3__f5),field(2,_,0) # f2_f3__f5
13053 str(f1_f2_f3__f5),field(2,_,2) # f2_f3
13054 str(f1_f2_f3__f5),field(-2,_,3) # f2_f3_
13055 str(f1_f2_f3__f5),field(-3,_,0) # f1_f2_f3
Emeric Brunf399b0d2014-11-03 17:07:03 +010013056
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013057hex
Davor Ocelice9ed2812017-12-25 17:49:28 +010013058 Converts a binary input sample to a hex string containing two hex digits per
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013059 input byte. It is used to log or transfer hex dumps of some binary input data
Davor Ocelice9ed2812017-12-25 17:49:28 +010013060 in a way that can be reliably transferred (e.g. an SSL ID can be copied in a
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013061 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +010013062
Dragan Dosen3f957b22017-10-24 09:27:34 +020013063hex2i
13064 Converts a hex string containing two hex digits per input byte to an
13065 integer. If the input value can not be converted, then zero is returned.
13066
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013067http_date([<offset>])
13068 Converts an integer supposed to contain a date since epoch to a string
13069 representing this date in a format suitable for use in HTTP header fields. If
13070 an offset value is specified, then it is a number of seconds that is added to
13071 the date before the conversion is operated. This is particularly useful to
13072 emit Date header fields, Expires values in responses when combined with a
13073 positive offset, or Last-Modified values when the offset is negative.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013074
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013075in_table(<table>)
13076 Uses the string representation of the input sample to perform a look up in
13077 the specified table. If the key is not found in the table, a boolean false
13078 is returned. Otherwise a boolean true is returned. This can be used to verify
Davor Ocelice9ed2812017-12-25 17:49:28 +010013079 the presence of a certain key in a table tracking some elements (e.g. whether
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013080 or not a source IP address or an Authorization header was already seen).
13081
Tim Duesterhus1478aa72018-01-25 16:24:51 +010013082ipmask(<mask4>, [<mask6>])
13083 Apply a mask to an IP address, and use the result for lookups and storage.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013084 This can be used to make all hosts within a certain mask to share the same
Tim Duesterhus1478aa72018-01-25 16:24:51 +010013085 table entries and as such use the same server. The mask4 can be passed in
13086 dotted form (e.g. 255.255.255.0) or in CIDR form (e.g. 24). The mask6 can
13087 be passed in quadruplet form (e.g. ffff:ffff::) or in CIDR form (e.g. 64).
13088 If no mask6 is given IPv6 addresses will fail to convert for backwards
13089 compatibility reasons.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013090
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013091json([<input-code>])
Davor Ocelice9ed2812017-12-25 17:49:28 +010013092 Escapes the input string and produces an ASCII output string ready to use as a
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013093 JSON string. The converter tries to decode the input string according to the
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020013094 <input-code> parameter. It can be "ascii", "utf8", "utf8s", "utf8p" or
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013095 "utf8ps". The "ascii" decoder never fails. The "utf8" decoder detects 3 types
13096 of errors:
13097 - bad UTF-8 sequence (lone continuation byte, bad number of continuation
13098 bytes, ...)
13099 - invalid range (the decoded value is within a UTF-8 prohibited range),
13100 - code overlong (the value is encoded with more bytes than necessary).
13101
13102 The UTF-8 JSON encoding can produce a "too long value" error when the UTF-8
13103 character is greater than 0xffff because the JSON string escape specification
13104 only authorizes 4 hex digits for the value encoding. The UTF-8 decoder exists
13105 in 4 variants designated by a combination of two suffix letters : "p" for
13106 "permissive" and "s" for "silently ignore". The behaviors of the decoders
13107 are :
Davor Ocelice9ed2812017-12-25 17:49:28 +010013108 - "ascii" : never fails;
13109 - "utf8" : fails on any detected errors;
13110 - "utf8s" : never fails, but removes characters corresponding to errors;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013111 - "utf8p" : accepts and fixes the overlong errors, but fails on any other
Davor Ocelice9ed2812017-12-25 17:49:28 +010013112 error;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013113 - "utf8ps" : never fails, accepts and fixes the overlong errors, but removes
13114 characters corresponding to the other errors.
13115
13116 This converter is particularly useful for building properly escaped JSON for
Davor Ocelice9ed2812017-12-25 17:49:28 +010013117 logging to servers which consume JSON-formatted traffic logs.
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013118
13119 Example:
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013120 capture request header Host len 15
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020013121 capture request header user-agent len 150
13122 log-format '{"ip":"%[src]","user-agent":"%[capture.req.hdr(1),json(utf8s)]"}'
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013123
13124 Input request from client 127.0.0.1:
13125 GET / HTTP/1.0
13126 User-Agent: Very "Ugly" UA 1/2
13127
13128 Output log:
13129 {"ip":"127.0.0.1","user-agent":"Very \"Ugly\" UA 1\/2"}
13130
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013131language(<value>[,<default>])
13132 Returns the value with the highest q-factor from a list as extracted from the
13133 "accept-language" header using "req.fhdr". Values with no q-factor have a
13134 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
13135 belong to the list of semi-colon delimited <values> will be considered. The
13136 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
13137 given list and a default value is provided, it is returned. Note that language
13138 names may have a variant after a dash ('-'). If this variant is present in the
13139 list, it will be matched, but if it is not, only the base language is checked.
13140 The match is case-sensitive, and the output string is always one of those
Davor Ocelice9ed2812017-12-25 17:49:28 +010013141 provided in arguments. The ordering of arguments is meaningless, only the
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013142 ordering of the values in the request counts, as the first value among
13143 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +020013144
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013145 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +020013146
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013147 # this configuration switches to the backend matching a
13148 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +020013149
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013150 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
13151 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
13152 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
13153 use_backend spanish if es
13154 use_backend french if fr
13155 use_backend english if en
13156 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +020013157
Willy Tarreau60a2ee72017-12-15 07:13:48 +010013158length
Etienne Carriereed0d24e2017-12-13 13:41:34 +010013159 Get the length of the string. This can only be placed after a string
13160 sample fetch function or after a transformation keyword returning a string
13161 type. The result is of type integer.
13162
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013163lower
13164 Convert a string sample to lower case. This can only be placed after a string
13165 sample fetch function or after a transformation keyword returning a string
13166 type. The result is of type string.
13167
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013168ltime(<format>[,<offset>])
13169 Converts an integer supposed to contain a date since epoch to a string
13170 representing this date in local time using a format defined by the <format>
13171 string using strftime(3). The purpose is to allow any date format to be used
13172 in logs. An optional <offset> in seconds may be applied to the input date
13173 (positive or negative). See the strftime() man page for the format supported
13174 by your operating system. See also the utime converter.
13175
13176 Example :
13177
13178 # Emit two colons, one with the local time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010013179 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013180 log-format %[date,ltime(%Y%m%d%H%M%S)]\ %ci:%cp
13181
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013182map(<map_file>[,<default_value>])
13183map_<match_type>(<map_file>[,<default_value>])
13184map_<match_type>_<output_type>(<map_file>[,<default_value>])
13185 Search the input value from <map_file> using the <match_type> matching method,
13186 and return the associated value converted to the type <output_type>. If the
13187 input value cannot be found in the <map_file>, the converter returns the
13188 <default_value>. If the <default_value> is not set, the converter fails and
13189 acts as if no input value could be fetched. If the <match_type> is not set, it
13190 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
13191 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
13192 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013193
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013194 It is important to avoid overlapping between the keys : IP addresses and
13195 strings are stored in trees, so the first of the finest match will be used.
13196 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013197
Tim Düsterhus4896c442016-11-29 02:15:19 +010013198 The following array contains the list of all map functions available sorted by
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013199 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013200
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013201 input type | match method | output type str | output type int | output type ip
13202 -----------+--------------+-----------------+-----------------+---------------
13203 str | str | map_str | map_str_int | map_str_ip
13204 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +020013205 str | beg | map_beg | map_beg_int | map_end_ip
13206 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013207 str | sub | map_sub | map_sub_int | map_sub_ip
13208 -----------+--------------+-----------------+-----------------+---------------
13209 str | dir | map_dir | map_dir_int | map_dir_ip
13210 -----------+--------------+-----------------+-----------------+---------------
13211 str | dom | map_dom | map_dom_int | map_dom_ip
13212 -----------+--------------+-----------------+-----------------+---------------
13213 str | end | map_end | map_end_int | map_end_ip
13214 -----------+--------------+-----------------+-----------------+---------------
Ruoshan Huang3c5e3742016-12-02 16:25:31 +080013215 str | reg | map_reg | map_reg_int | map_reg_ip
13216 -----------+--------------+-----------------+-----------------+---------------
13217 str | reg | map_regm | map_reg_int | map_reg_ip
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013218 -----------+--------------+-----------------+-----------------+---------------
13219 int | int | map_int | map_int_int | map_int_ip
13220 -----------+--------------+-----------------+-----------------+---------------
13221 ip | ip | map_ip | map_ip_int | map_ip_ip
13222 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013223
Thierry Fournier8feaa662016-02-10 22:55:20 +010013224 The special map called "map_regm" expect matching zone in the regular
13225 expression and modify the output replacing back reference (like "\1") by
13226 the corresponding match text.
13227
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013228 The file contains one key + value per line. Lines which start with '#' are
13229 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
13230 is then the first "word" (series of non-space/tabs characters), and the value
13231 is what follows this series of space/tab till the end of the line excluding
13232 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013233
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013234 Example :
13235
13236 # this is a comment and is ignored
13237 2.22.246.0/23 United Kingdom \n
13238 <-><-----------><--><------------><---->
13239 | | | | `- trailing spaces ignored
13240 | | | `---------- value
13241 | | `-------------------- middle spaces ignored
13242 | `---------------------------- key
13243 `------------------------------------ leading spaces ignored
13244
Willy Tarreau97707872015-01-27 15:12:13 +010013245mod(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013246 Divides the input value of type signed integer by <value>, and returns the
13247 remainder as an signed integer. If <value> is null, then zero is returned.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013248 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013249 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013250 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013251 "sess" : the variable is shared with the whole session
13252 "txn" : the variable is shared with the transaction (request and response)
13253 "req" : the variable is shared only during request processing
13254 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013255 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013256 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013257
13258mul(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013259 Multiplies the input value of type signed integer by <value>, and returns
Thierry FOURNIER00c005c2015-07-08 01:10:21 +020013260 the product as an signed integer. In case of overflow, the largest possible
13261 value for the sign is returned so that the operation doesn't wrap around.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013262 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013263 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013264 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013265 "sess" : the variable is shared with the whole session
13266 "txn" : the variable is shared with the transaction (request and response)
13267 "req" : the variable is shared only during request processing
13268 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013269 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013270 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013271
Nenad Merdanovicb7e7c472017-03-12 21:56:55 +010013272nbsrv
13273 Takes an input value of type string, interprets it as a backend name and
13274 returns the number of usable servers in that backend. Can be used in places
13275 where we want to look up a backend from a dynamic name, like a result of a
13276 map lookup.
13277
Willy Tarreau97707872015-01-27 15:12:13 +010013278neg
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013279 Takes the input value of type signed integer, computes the opposite value,
13280 and returns the remainder as an signed integer. 0 is identity. This operator
13281 is provided for reversed subtracts : in order to subtract the input from a
13282 constant, simply perform a "neg,add(value)".
Willy Tarreau97707872015-01-27 15:12:13 +010013283
13284not
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013285 Returns a boolean FALSE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010013286 non-null, otherwise returns TRUE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010013287 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010013288 absence of a flag).
13289
13290odd
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013291 Returns a boolean TRUE if the input value of type signed integer is odd
Willy Tarreau97707872015-01-27 15:12:13 +010013292 otherwise returns FALSE. It is functionally equivalent to "and(1),bool".
13293
13294or(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013295 Performs a bitwise "OR" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013296 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010013297 numeric value or a variable name. The name of the variable starts with an
13298 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013299 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013300 "sess" : the variable is shared with the whole session
13301 "txn" : the variable is shared with the transaction (request and response)
13302 "req" : the variable is shared only during request processing
13303 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013304 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013305 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013306
Willy Tarreauc4dc3502015-01-23 20:39:28 +010013307regsub(<regex>,<subst>[,<flags>])
Willy Tarreau7eda8492015-01-20 19:47:06 +010013308 Applies a regex-based substitution to the input string. It does the same
13309 operation as the well-known "sed" utility with "s/<regex>/<subst>/". By
13310 default it will replace in the input string the first occurrence of the
13311 largest part matching the regular expression <regex> with the substitution
13312 string <subst>. It is possible to replace all occurrences instead by adding
13313 the flag "g" in the third argument <flags>. It is also possible to make the
13314 regex case insensitive by adding the flag "i" in <flags>. Since <flags> is a
13315 string, it is made up from the concatenation of all desired flags. Thus if
13316 both "i" and "g" are desired, using "gi" or "ig" will have the same effect.
13317 It is important to note that due to the current limitations of the
Baptiste Assmann66025d82016-03-06 23:36:48 +010013318 configuration parser, some characters such as closing parenthesis, closing
13319 square brackets or comma are not possible to use in the arguments. The first
13320 use of this converter is to replace certain characters or sequence of
13321 characters with other ones.
Willy Tarreau7eda8492015-01-20 19:47:06 +010013322
13323 Example :
13324
13325 # de-duplicate "/" in header "x-path".
13326 # input: x-path: /////a///b/c/xzxyz/
13327 # output: x-path: /a/b/c/xzxyz/
13328 http-request set-header x-path %[hdr(x-path),regsub(/+,/,g)]
13329
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013330capture-req(<id>)
13331 Capture the string entry in the request slot <id> and returns the entry as
13332 is. If the slot doesn't exist, the capture fails silently.
13333
13334 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020013335 "http-response capture", "capture.req.hdr" and
13336 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013337
13338capture-res(<id>)
13339 Capture the string entry in the response slot <id> and returns the entry as
13340 is. If the slot doesn't exist, the capture fails silently.
13341
13342 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020013343 "http-response capture", "capture.req.hdr" and
13344 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013345
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013346sdbm([<avalanche>])
13347 Hashes a binary input sample into an unsigned 32-bit quantity using the SDBM
13348 hash function. Optionally, it is possible to apply a full avalanche hash
13349 function to the output if the optional <avalanche> argument equals 1. This
13350 converter uses the same functions as used by the various hash-based load
13351 balancing algorithms, so it will provide exactly the same results. It is
13352 mostly intended for debugging, but can be used as a stick-table entry to
13353 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013354 32-bit hash is trivial to break. See also "crc32", "djb2", "wt6", "crc32c",
13355 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013356
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013357set-var(<var name>)
Davor Ocelice9ed2812017-12-25 17:49:28 +010013358 Sets a variable with the input content and returns the content on the output
13359 as-is. The variable keeps the value and the associated input type. The name of
13360 the variable starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013361 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013362 "sess" : the variable is shared with the whole session
13363 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013364 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013365 "req" : the variable is shared only during request processing,
13366 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013367 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013368 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013369
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020013370sha1
13371 Converts a binary input sample to a SHA1 digest. The result is a binary
13372 sample with length of 20 bytes.
13373
Tim Duesterhusca097c12018-04-27 21:18:45 +020013374strcmp(<var>)
13375 Compares the contents of <var> with the input value of type string. Returns
13376 the result as a signed integer compatible with strcmp(3): 0 if both strings
13377 are identical. A value less than 0 if the left string is lexicographically
13378 smaller than the right string or if the left string is shorter. A value greater
13379 than 0 otherwise (right string greater than left string or the right string is
13380 shorter).
13381
13382 Example :
13383
13384 http-request set-var(txn.host) hdr(host)
13385 # Check whether the client is attempting domain fronting.
13386 acl ssl_sni_http_host_match ssl_fc_sni,strcmp(txn.host) eq 0
13387
13388
Willy Tarreau97707872015-01-27 15:12:13 +010013389sub(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013390 Subtracts <value> from the input value of type signed integer, and returns
13391 the result as an signed integer. Note: in order to subtract the input from
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013392 a constant, simply perform a "neg,add(value)". <value> can be a numeric value
Daniel Schneller0b547052016-03-21 20:46:57 +010013393 or a variable name. The name of the variable starts with an indication about
13394 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013395 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013396 "sess" : the variable is shared with the whole session
13397 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013398 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013399 "req" : the variable is shared only during request processing,
13400 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013401 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013402 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013403
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013404table_bytes_in_rate(<table>)
13405 Uses the string representation of the input sample to perform a look up in
13406 the specified table. If the key is not found in the table, integer value zero
13407 is returned. Otherwise the converter returns the average client-to-server
13408 bytes rate associated with the input sample in the designated table, measured
13409 in amount of bytes over the period configured in the table. See also the
13410 sc_bytes_in_rate sample fetch keyword.
13411
13412
13413table_bytes_out_rate(<table>)
13414 Uses the string representation of the input sample to perform a look up in
13415 the specified table. If the key is not found in the table, integer value zero
13416 is returned. Otherwise the converter returns the average server-to-client
13417 bytes rate associated with the input sample in the designated table, measured
13418 in amount of bytes over the period configured in the table. See also the
13419 sc_bytes_out_rate sample fetch keyword.
13420
13421table_conn_cnt(<table>)
13422 Uses the string representation of the input sample to perform a look up in
13423 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013424 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013425 connections associated with the input sample in the designated table. See
13426 also the sc_conn_cnt sample fetch keyword.
13427
13428table_conn_cur(<table>)
13429 Uses the string representation of the input sample to perform a look up in
13430 the specified table. If the key is not found in the table, integer value zero
13431 is returned. Otherwise the converter returns the current amount of concurrent
13432 tracked connections associated with the input sample in the designated table.
13433 See also the sc_conn_cur sample fetch keyword.
13434
13435table_conn_rate(<table>)
13436 Uses the string representation of the input sample to perform a look up in
13437 the specified table. If the key is not found in the table, integer value zero
13438 is returned. Otherwise the converter returns the average incoming connection
13439 rate associated with the input sample in the designated table. See also the
13440 sc_conn_rate sample fetch keyword.
13441
Thierry FOURNIER236657b2015-08-19 08:25:14 +020013442table_gpt0(<table>)
13443 Uses the string representation of the input sample to perform a look up in
13444 the specified table. If the key is not found in the table, boolean value zero
13445 is returned. Otherwise the converter returns the current value of the first
13446 general purpose tag associated with the input sample in the designated table.
13447 See also the sc_get_gpt0 sample fetch keyword.
13448
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013449table_gpc0(<table>)
13450 Uses the string representation of the input sample to perform a look up in
13451 the specified table. If the key is not found in the table, integer value zero
13452 is returned. Otherwise the converter returns the current value of the first
13453 general purpose counter associated with the input sample in the designated
13454 table. See also the sc_get_gpc0 sample fetch keyword.
13455
13456table_gpc0_rate(<table>)
13457 Uses the string representation of the input sample to perform a look up in
13458 the specified table. If the key is not found in the table, integer value zero
13459 is returned. Otherwise the converter returns the frequency which the gpc0
13460 counter was incremented over the configured period in the table, associated
13461 with the input sample in the designated table. See also the sc_get_gpc0_rate
13462 sample fetch keyword.
13463
Frédéric Lécaille6778b272018-01-29 15:22:53 +010013464table_gpc1(<table>)
13465 Uses the string representation of the input sample to perform a look up in
13466 the specified table. If the key is not found in the table, integer value zero
13467 is returned. Otherwise the converter returns the current value of the second
13468 general purpose counter associated with the input sample in the designated
13469 table. See also the sc_get_gpc1 sample fetch keyword.
13470
13471table_gpc1_rate(<table>)
13472 Uses the string representation of the input sample to perform a look up in
13473 the specified table. If the key is not found in the table, integer value zero
13474 is returned. Otherwise the converter returns the frequency which the gpc1
13475 counter was incremented over the configured period in the table, associated
13476 with the input sample in the designated table. See also the sc_get_gpc1_rate
13477 sample fetch keyword.
13478
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013479table_http_err_cnt(<table>)
13480 Uses the string representation of the input sample to perform a look up in
13481 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013482 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013483 errors associated with the input sample in the designated table. See also the
13484 sc_http_err_cnt sample fetch keyword.
13485
13486table_http_err_rate(<table>)
13487 Uses the string representation of the input sample to perform a look up in
13488 the specified table. If the key is not found in the table, integer value zero
13489 is returned. Otherwise the average rate of HTTP errors associated with the
13490 input sample in the designated table, measured in amount of errors over the
13491 period configured in the table. See also the sc_http_err_rate sample fetch
13492 keyword.
13493
13494table_http_req_cnt(<table>)
13495 Uses the string representation of the input sample to perform a look up in
13496 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013497 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013498 requests associated with the input sample in the designated table. See also
13499 the sc_http_req_cnt sample fetch keyword.
13500
13501table_http_req_rate(<table>)
13502 Uses the string representation of the input sample to perform a look up in
13503 the specified table. If the key is not found in the table, integer value zero
13504 is returned. Otherwise the average rate of HTTP requests associated with the
13505 input sample in the designated table, measured in amount of requests over the
13506 period configured in the table. See also the sc_http_req_rate sample fetch
13507 keyword.
13508
13509table_kbytes_in(<table>)
13510 Uses the string representation of the input sample to perform a look up in
13511 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013512 is returned. Otherwise the converter returns the cumulative number of client-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013513 to-server data associated with the input sample in the designated table,
13514 measured in kilobytes. The test is currently performed on 32-bit integers,
13515 which limits values to 4 terabytes. See also the sc_kbytes_in sample fetch
13516 keyword.
13517
13518table_kbytes_out(<table>)
13519 Uses the string representation of the input sample to perform a look up in
13520 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013521 is returned. Otherwise the converter returns the cumulative number of server-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013522 to-client data associated with the input sample in the designated table,
13523 measured in kilobytes. The test is currently performed on 32-bit integers,
13524 which limits values to 4 terabytes. See also the sc_kbytes_out sample fetch
13525 keyword.
13526
13527table_server_id(<table>)
13528 Uses the string representation of the input sample to perform a look up in
13529 the specified table. If the key is not found in the table, integer value zero
13530 is returned. Otherwise the converter returns the server ID associated with
13531 the input sample in the designated table. A server ID is associated to a
13532 sample by a "stick" rule when a connection to a server succeeds. A server ID
13533 zero means that no server is associated with this key.
13534
13535table_sess_cnt(<table>)
13536 Uses the string representation of the input sample to perform a look up in
13537 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013538 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013539 sessions associated with the input sample in the designated table. Note that
13540 a session here refers to an incoming connection being accepted by the
13541 "tcp-request connection" rulesets. See also the sc_sess_cnt sample fetch
13542 keyword.
13543
13544table_sess_rate(<table>)
13545 Uses the string representation of the input sample to perform a look up in
13546 the specified table. If the key is not found in the table, integer value zero
13547 is returned. Otherwise the converter returns the average incoming session
13548 rate associated with the input sample in the designated table. Note that a
13549 session here refers to an incoming connection being accepted by the
13550 "tcp-request connection" rulesets. See also the sc_sess_rate sample fetch
13551 keyword.
13552
13553table_trackers(<table>)
13554 Uses the string representation of the input sample to perform a look up in
13555 the specified table. If the key is not found in the table, integer value zero
13556 is returned. Otherwise the converter returns the current amount of concurrent
13557 connections tracking the same key as the input sample in the designated
13558 table. It differs from table_conn_cur in that it does not rely on any stored
13559 information but on the table's reference count (the "use" value which is
13560 returned by "show table" on the CLI). This may sometimes be more suited for
13561 layer7 tracking. It can be used to tell a server how many concurrent
13562 connections there are from a given address for example. See also the
13563 sc_trackers sample fetch keyword.
13564
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013565upper
13566 Convert a string sample to upper case. This can only be placed after a string
13567 sample fetch function or after a transformation keyword returning a string
13568 type. The result is of type string.
13569
Thierry FOURNIER82ff3c92015-05-07 15:46:20 +020013570url_dec
13571 Takes an url-encoded string provided as input and returns the decoded
13572 version as output. The input and the output are of type string.
13573
Christopher Faulet85d79c92016-11-09 16:54:56 +010013574unset-var(<var name>)
13575 Unsets a variable if the input content is defined. The name of the variable
13576 starts with an indication about its scope. The scopes allowed are:
13577 "proc" : the variable is shared with the whole process
13578 "sess" : the variable is shared with the whole session
13579 "txn" : the variable is shared with the transaction (request and
13580 response),
13581 "req" : the variable is shared only during request processing,
13582 "res" : the variable is shared only during response processing.
13583 This prefix is followed by a name. The separator is a '.'. The name may only
13584 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
13585
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013586utime(<format>[,<offset>])
13587 Converts an integer supposed to contain a date since epoch to a string
13588 representing this date in UTC time using a format defined by the <format>
13589 string using strftime(3). The purpose is to allow any date format to be used
13590 in logs. An optional <offset> in seconds may be applied to the input date
13591 (positive or negative). See the strftime() man page for the format supported
13592 by your operating system. See also the ltime converter.
13593
13594 Example :
13595
13596 # Emit two colons, one with the UTC time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010013597 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013598 log-format %[date,utime(%Y%m%d%H%M%S)]\ %ci:%cp
13599
Marcin Deranek9631a282018-04-16 14:30:46 +020013600word(<index>,<delimiters>[,<count>])
13601 Extracts the nth word counting from the beginning (positive index) or from
13602 the end (negative index) considering given delimiters from an input string.
13603 Indexes start at 1 or -1 and delimiters are a string formatted list of chars.
13604 Optionally you can specify <count> of words to extract (default: 1).
13605 Value of 0 indicates extraction of all remaining words.
13606
13607 Example :
13608 str(f1_f2_f3__f5),word(4,_) # f5
13609 str(f1_f2_f3__f5),word(2,_,0) # f2_f3__f5
13610 str(f1_f2_f3__f5),word(3,_,2) # f3__f5
13611 str(f1_f2_f3__f5),word(-2,_,3) # f1_f2_f3
13612 str(f1_f2_f3__f5),word(-3,_,0) # f1_f2
Emeric Brunc9a0f6d2014-11-25 14:09:01 +010013613
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013614wt6([<avalanche>])
13615 Hashes a binary input sample into an unsigned 32-bit quantity using the WT6
13616 hash function. Optionally, it is possible to apply a full avalanche hash
13617 function to the output if the optional <avalanche> argument equals 1. This
13618 converter uses the same functions as used by the various hash-based load
13619 balancing algorithms, so it will provide exactly the same results. It is
13620 mostly intended for debugging, but can be used as a stick-table entry to
13621 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013622 32-bit hash is trivial to break. See also "crc32", "djb2", "sdbm", "crc32c",
13623 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013624
Willy Tarreau97707872015-01-27 15:12:13 +010013625xor(<value>)
13626 Performs a bitwise "XOR" (exclusive OR) between <value> and the input value
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013627 of type signed integer, and returns the result as an signed integer.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013628 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013629 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013630 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013631 "sess" : the variable is shared with the whole session
13632 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013633 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013634 "req" : the variable is shared only during request processing,
13635 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013636 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013637 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013638
Thierry FOURNIER01e09742016-12-26 11:46:11 +010013639xxh32([<seed>])
13640 Hashes a binary input sample into an unsigned 32-bit quantity using the 32-bit
13641 variant of the XXHash hash function. This hash supports a seed which defaults
13642 to zero but a different value maybe passed as the <seed> argument. This hash
13643 is known to be very good and very fast so it can be used to hash URLs and/or
13644 URL parameters for use as stick-table keys to collect statistics with a low
13645 collision rate, though care must be taken as the algorithm is not considered
13646 as cryptographically secure.
13647
13648xxh64([<seed>])
13649 Hashes a binary input sample into a signed 64-bit quantity using the 64-bit
13650 variant of the XXHash hash function. This hash supports a seed which defaults
13651 to zero but a different value maybe passed as the <seed> argument. This hash
13652 is known to be very good and very fast so it can be used to hash URLs and/or
13653 URL parameters for use as stick-table keys to collect statistics with a low
13654 collision rate, though care must be taken as the algorithm is not considered
13655 as cryptographically secure.
13656
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013657
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200136587.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020013659--------------------------------------------
13660
13661A first set of sample fetch methods applies to internal information which does
13662not even relate to any client information. These ones are sometimes used with
13663"monitor-fail" directives to report an internal status to external watchers.
13664The sample fetch methods described in this section are usable anywhere.
13665
13666always_false : boolean
13667 Always returns the boolean "false" value. It may be used with ACLs as a
13668 temporary replacement for another one when adjusting configurations.
13669
13670always_true : boolean
13671 Always returns the boolean "true" value. It may be used with ACLs as a
13672 temporary replacement for another one when adjusting configurations.
13673
13674avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010013675 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020013676 divided by the number of active servers. The current backend is used if no
13677 backend is specified. This is very similar to "queue" except that the size of
13678 the farm is considered, in order to give a more accurate measurement of the
13679 time it may take for a new connection to be processed. The main usage is with
13680 ACL to return a sorry page to new users when it becomes certain they will get
13681 a degraded service, or to pass to the backend servers in a header so that
13682 they decide to work in degraded mode or to disable some functions to speed up
13683 the processing a bit. Note that in the event there would not be any active
13684 server anymore, twice the number of queued connections would be considered as
13685 the measured value. This is a fair estimate, as we expect one server to get
13686 back soon anyway, but we still prefer to send new traffic to another backend
13687 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
13688 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010013689
Willy Tarreau74ca5042013-06-11 23:12:07 +020013690be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020013691 Applies to the number of currently established connections on the backend,
13692 possibly including the connection being evaluated. If no backend name is
13693 specified, the current one is used. But it is also possible to check another
13694 backend. It can be used to use a specific farm when the nominal one is full.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040013695 See also the "fe_conn", "queue", "be_conn_free", and "be_sess_rate" criteria.
13696
13697be_conn_free([<backend>]) : integer
13698 Returns an integer value corresponding to the number of available connections
13699 across available servers in the backend. Queue slots are not included. Backup
13700 servers are also not included, unless all other servers are down. If no
13701 backend name is specified, the current one is used. But it is also possible
13702 to check another backend. It can be used to use a specific farm when the
Patrick Hemmer155e93e2018-06-14 18:01:35 -040013703 nominal one is full. See also the "be_conn", "connslots", and "srv_conn_free"
13704 criteria.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040013705
13706 OTHER CAVEATS AND NOTES: if any of the server maxconn, or maxqueue is 0
13707 (meaning unlimited), then this fetch clearly does not make sense, in which
13708 case the value returned will be -1.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013709
Willy Tarreau74ca5042013-06-11 23:12:07 +020013710be_sess_rate([<backend>]) : integer
13711 Returns an integer value corresponding to the sessions creation rate on the
13712 backend, in number of new sessions per second. This is used with ACLs to
13713 switch to an alternate backend when an expensive or fragile one reaches too
Davor Ocelice9ed2812017-12-25 17:49:28 +010013714 high a session rate, or to limit abuse of service (e.g. prevent sucking of an
Willy Tarreau74ca5042013-06-11 23:12:07 +020013715 online dictionary). It can also be useful to add this element to logs using a
13716 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010013717
13718 Example :
13719 # Redirect to an error page if the dictionary is requested too often
13720 backend dynamic
13721 mode http
13722 acl being_scanned be_sess_rate gt 100
13723 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010013724
Davor Ocelice9ed2812017-12-25 17:49:28 +010013725bin(<hex>) : bin
Thierry FOURNIERcc103292015-06-06 19:30:17 +020013726 Returns a binary chain. The input is the hexadecimal representation
13727 of the string.
13728
13729bool(<bool>) : bool
13730 Returns a boolean value. <bool> can be 'true', 'false', '1' or '0'.
13731 'false' and '0' are the same. 'true' and '1' are the same.
13732
Willy Tarreau74ca5042013-06-11 23:12:07 +020013733connslots([<backend>]) : integer
13734 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013735 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020013736 connections on all servers and the maximum queue size. This is probably only
13737 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050013738
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080013739 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020013740 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080013741 usage; see "use_backend" keyword) can be redirected to a different backend.
13742
Willy Tarreau55165fe2009-05-10 12:02:55 +020013743 'connslots' = number of available server connection slots, + number of
13744 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080013745
Willy Tarreaua36af912009-10-10 12:02:45 +020013746 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020013747 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020013748 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020013749 you want to be able to differentiate between different backends, and their
Davor Ocelice9ed2812017-12-25 17:49:28 +010013750 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020013751 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020013752 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080013753
Willy Tarreau55165fe2009-05-10 12:02:55 +020013754 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
13755 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020013756 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020013757 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080013758
Willy Tarreau6236d3a2013-07-25 14:28:25 +020013759date([<offset>]) : integer
13760 Returns the current date as the epoch (number of seconds since 01/01/1970).
13761 If an offset value is specified, then it is a number of seconds that is added
13762 to the current date before returning the value. This is particularly useful
13763 to compute relative dates, as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020013764 It is useful combined with the http_date converter.
13765
13766 Example :
13767
13768 # set an expires header to now+1 hour in every response
13769 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020013770
Etienne Carrierea792a0a2018-01-17 13:43:24 +010013771date_us : integer
13772 Return the microseconds part of the date (the "second" part is returned by
13773 date sample). This sample is coherent with the date sample as it is comes
13774 from the same timeval structure.
13775
Willy Tarreaud716f9b2017-10-13 11:03:15 +020013776distcc_body(<token>[,<occ>]) : binary
13777 Parses a distcc message and returns the body associated to occurrence #<occ>
13778 of the token <token>. Occurrences start at 1, and when unspecified, any may
13779 match though in practice only the first one is checked for now. This can be
13780 used to extract file names or arguments in files built using distcc through
13781 haproxy. Please refer to distcc's protocol documentation for the complete
13782 list of supported tokens.
13783
13784distcc_param(<token>[,<occ>]) : integer
13785 Parses a distcc message and returns the parameter associated to occurrence
13786 #<occ> of the token <token>. Occurrences start at 1, and when unspecified,
13787 any may match though in practice only the first one is checked for now. This
13788 can be used to extract certain information such as the protocol version, the
13789 file size or the argument in files built using distcc through haproxy.
13790 Another use case consists in waiting for the start of the preprocessed file
13791 contents before connecting to the server to avoid keeping idle connections.
13792 Please refer to distcc's protocol documentation for the complete list of
13793 supported tokens.
13794
13795 Example :
13796 # wait up to 20s for the pre-processed file to be uploaded
13797 tcp-request inspect-delay 20s
13798 tcp-request content accept if { distcc_param(DOTI) -m found }
13799 # send large files to the big farm
13800 use_backend big_farm if { distcc_param(DOTI) gt 1000000 }
13801
Willy Tarreau595ec542013-06-12 21:34:28 +020013802env(<name>) : string
13803 Returns a string containing the value of environment variable <name>. As a
13804 reminder, environment variables are per-process and are sampled when the
13805 process starts. This can be useful to pass some information to a next hop
13806 server, or with ACLs to take specific action when the process is started a
13807 certain way.
13808
13809 Examples :
13810 # Pass the Via header to next hop with the local hostname in it
13811 http-request add-header Via 1.1\ %[env(HOSTNAME)]
13812
13813 # reject cookie-less requests when the STOP environment variable is set
13814 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
13815
Willy Tarreau74ca5042013-06-11 23:12:07 +020013816fe_conn([<frontend>]) : integer
13817 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010013818 possibly including the connection being evaluated. If no frontend name is
13819 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020013820 frontend. It can be used to return a sorry page before hard-blocking, or to
13821 use a specific backend to drain new requests when the farm is considered
Davor Ocelice9ed2812017-12-25 17:49:28 +010013822 full. This is mostly used with ACLs but can also be used to pass some
Willy Tarreau74ca5042013-06-11 23:12:07 +020013823 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
13824 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020013825
Nenad Merdanovicad9a7e92016-10-03 04:57:37 +020013826fe_req_rate([<frontend>]) : integer
13827 Returns an integer value corresponding to the number of HTTP requests per
13828 second sent to a frontend. This number can differ from "fe_sess_rate" in
13829 situations where client-side keep-alive is enabled.
13830
Willy Tarreau74ca5042013-06-11 23:12:07 +020013831fe_sess_rate([<frontend>]) : integer
13832 Returns an integer value corresponding to the sessions creation rate on the
13833 frontend, in number of new sessions per second. This is used with ACLs to
13834 limit the incoming session rate to an acceptable range in order to prevent
13835 abuse of service at the earliest moment, for example when combined with other
13836 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
13837 down below the limit. It can also be useful to add this element to logs using
13838 a log-format directive. See also the "rate-limit sessions" directive for use
13839 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010013840
13841 Example :
13842 # This frontend limits incoming mails to 10/s with a max of 100
13843 # concurrent connections. We accept any connection below 10/s, and
13844 # force excess clients to wait for 100 ms. Since clients are limited to
13845 # 100 max, there cannot be more than 10 incoming mails per second.
13846 frontend mail
13847 bind :25
13848 mode tcp
13849 maxconn 100
13850 acl too_fast fe_sess_rate ge 10
13851 tcp-request inspect-delay 100ms
13852 tcp-request content accept if ! too_fast
13853 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010013854
Nenad Merdanovic807a6e72017-03-12 22:00:00 +010013855hostname : string
13856 Returns the system hostname.
13857
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013858int(<integer>) : signed integer
13859 Returns a signed integer.
13860
Thierry FOURNIERcc103292015-06-06 19:30:17 +020013861ipv4(<ipv4>) : ipv4
13862 Returns an ipv4.
13863
13864ipv6(<ipv6>) : ipv6
13865 Returns an ipv6.
13866
13867meth(<method>) : method
13868 Returns a method.
13869
Willy Tarreau0f30d262014-11-24 16:02:05 +010013870nbproc : integer
13871 Returns an integer value corresponding to the number of processes that were
13872 started (it equals the global "nbproc" setting). This is useful for logging
13873 and debugging purposes.
13874
Willy Tarreau74ca5042013-06-11 23:12:07 +020013875nbsrv([<backend>]) : integer
13876 Returns an integer value corresponding to the number of usable servers of
13877 either the current backend or the named backend. This is mostly used with
13878 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010013879 switch to an alternate backend when the number of servers is too low to
13880 to handle some load. It is useful to report a failure when combined with
13881 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010013882
Patrick Hemmerfabb24f2018-08-13 14:07:57 -040013883prio_class : integer
13884 Returns the priority class of the current session for http mode or connection
13885 for tcp mode. The value will be that set by the last call to "http-request
13886 set-priority-class" or "tcp-request content set-priority-class".
13887
13888prio_offset : integer
13889 Returns the priority offset of the current session for http mode or
13890 connection for tcp mode. The value will be that set by the last call to
13891 "http-request set-priority-offset" or "tcp-request content
13892 set-priority-offset".
13893
Willy Tarreau0f30d262014-11-24 16:02:05 +010013894proc : integer
13895 Returns an integer value corresponding to the position of the process calling
13896 the function, between 1 and global.nbproc. This is useful for logging and
13897 debugging purposes.
13898
Willy Tarreau74ca5042013-06-11 23:12:07 +020013899queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010013900 Returns the total number of queued connections of the designated backend,
13901 including all the connections in server queues. If no backend name is
13902 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020013903 one. This is useful with ACLs or to pass statistics to backend servers. This
13904 can be used to take actions when queuing goes above a known level, generally
13905 indicating a surge of traffic or a massive slowdown on the servers. One
13906 possible action could be to reject new users but still accept old ones. See
13907 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
13908
Willy Tarreau84310e22014-02-14 11:59:04 +010013909rand([<range>]) : integer
13910 Returns a random integer value within a range of <range> possible values,
13911 starting at zero. If the range is not specified, it defaults to 2^32, which
13912 gives numbers between 0 and 4294967295. It can be useful to pass some values
13913 needed to take some routing decisions for example, or just for debugging
13914 purposes. This random must not be used for security purposes.
13915
Willy Tarreau74ca5042013-06-11 23:12:07 +020013916srv_conn([<backend>/]<server>) : integer
13917 Returns an integer value corresponding to the number of currently established
13918 connections on the designated server, possibly including the connection being
13919 evaluated. If <backend> is omitted, then the server is looked up in the
13920 current backend. It can be used to use a specific farm when one server is
13921 full, or to inform the server about our view of the number of active
Patrick Hemmer155e93e2018-06-14 18:01:35 -040013922 connections with it. See also the "fe_conn", "be_conn", "queue", and
13923 "srv_conn_free" fetch methods.
13924
13925srv_conn_free([<backend>/]<server>) : integer
13926 Returns an integer value corresponding to the number of available connections
13927 on the designated server, possibly including the connection being evaluated.
13928 The value does not include queue slots. If <backend> is omitted, then the
13929 server is looked up in the current backend. It can be used to use a specific
13930 farm when one server is full, or to inform the server about our view of the
13931 number of active connections with it. See also the "be_conn_free" and
13932 "srv_conn" fetch methods.
13933
13934 OTHER CAVEATS AND NOTES: If the server maxconn is 0, then this fetch clearly
13935 does not make sense, in which case the value returned will be -1.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013936
13937srv_is_up([<backend>/]<server>) : boolean
13938 Returns true when the designated server is UP, and false when it is either
13939 DOWN or in maintenance mode. If <backend> is omitted, then the server is
13940 looked up in the current backend. It is mainly used to take action based on
Davor Ocelice9ed2812017-12-25 17:49:28 +010013941 an external status reported via a health check (e.g. a geographical site's
Willy Tarreau74ca5042013-06-11 23:12:07 +020013942 availability). Another possible use which is more of a hack consists in
13943 using dummy servers as boolean variables that can be enabled or disabled from
13944 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
13945
Willy Tarreauff2b7af2017-10-13 11:46:26 +020013946srv_queue([<backend>/]<server>) : integer
13947 Returns an integer value corresponding to the number of connections currently
13948 pending in the designated server's queue. If <backend> is omitted, then the
13949 server is looked up in the current backend. It can sometimes be used together
13950 with the "use-server" directive to force to use a known faster server when it
13951 is not much loaded. See also the "srv_conn", "avg_queue" and "queue" sample
13952 fetch methods.
13953
Willy Tarreau74ca5042013-06-11 23:12:07 +020013954srv_sess_rate([<backend>/]<server>) : integer
13955 Returns an integer corresponding to the sessions creation rate on the
13956 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013957 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020013958 used with ACLs but can make sense with logs too. This is used to switch to an
13959 alternate backend when an expensive or fragile one reaches too high a session
Davor Ocelice9ed2812017-12-25 17:49:28 +010013960 rate, or to limit abuse of service (e.g. prevent latent requests from
Willy Tarreau74ca5042013-06-11 23:12:07 +020013961 overloading servers).
13962
13963 Example :
13964 # Redirect to a separate back
13965 acl srv1_full srv_sess_rate(be1/srv1) gt 50
13966 acl srv2_full srv_sess_rate(be1/srv2) gt 50
13967 use_backend be2 if srv1_full or srv2_full
13968
Willy Tarreau0f30d262014-11-24 16:02:05 +010013969stopping : boolean
13970 Returns TRUE if the process calling the function is currently stopping. This
13971 can be useful for logging, or for relaxing certain checks or helping close
13972 certain connections upon graceful shutdown.
13973
Thierry FOURNIERcc103292015-06-06 19:30:17 +020013974str(<string>) : string
13975 Returns a string.
13976
Willy Tarreau74ca5042013-06-11 23:12:07 +020013977table_avl([<table>]) : integer
13978 Returns the total number of available entries in the current proxy's
13979 stick-table or in the designated stick-table. See also table_cnt.
13980
13981table_cnt([<table>]) : integer
13982 Returns the total number of entries currently in use in the current proxy's
13983 stick-table or in the designated stick-table. See also src_conn_cnt and
13984 table_avl for other entry counting methods.
13985
Christopher Faulet34adb2a2017-11-21 21:45:38 +010013986thread : integer
13987 Returns an integer value corresponding to the position of the thread calling
13988 the function, between 0 and (global.nbthread-1). This is useful for logging
13989 and debugging purposes.
13990
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013991var(<var-name>) : undefined
13992 Returns a variable with the stored type. If the variable is not set, the
Daniel Schneller0b547052016-03-21 20:46:57 +010013993 sample fetch fails. The name of the variable starts with an indication
13994 about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013995 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013996 "sess" : the variable is shared with the whole session
13997 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013998 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013999 "req" : the variable is shared only during request processing,
14000 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014001 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010014002 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014003
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200140047.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020014005----------------------------------
14006
14007The layer 4 usually describes just the transport layer which in haproxy is
14008closest to the connection, where no content is yet made available. The fetch
14009methods described here are usable as low as the "tcp-request connection" rule
14010sets unless they require some future information. Those generally include
14011TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014012the incoming connection. For retrieving a value from a sticky counters, the
14013counter number can be explicitly set as 0, 1, or 2 using the pre-defined
14014"sc0_", "sc1_", or "sc2_" prefix, or it can be specified as the first integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014015argument when using the "sc_" prefix. An optional table may be specified with
14016the "sc*" form, in which case the currently tracked key will be looked up into
14017this alternate table instead of the table currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014018
14019be_id : integer
14020 Returns an integer containing the current backend's id. It can be used in
14021 frontends with responses to check which backend processed the request.
14022
Marcin Deranekd2471c22016-12-12 14:08:05 +010014023be_name : string
14024 Returns a string containing the current backend's name. It can be used in
14025 frontends with responses to check which backend processed the request.
14026
Willy Tarreau74ca5042013-06-11 23:12:07 +020014027dst : ip
14028 This is the destination IPv4 address of the connection on the client side,
14029 which is the address the client connected to. It can be useful when running
14030 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
14031 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
14032 RFC 4291.
14033
14034dst_conn : integer
14035 Returns an integer value corresponding to the number of currently established
14036 connections on the same socket including the one being evaluated. It is
14037 normally used with ACLs but can as well be used to pass the information to
14038 servers in an HTTP header or in logs. It can be used to either return a sorry
14039 page before hard-blocking, or to use a specific backend to drain new requests
14040 when the socket is considered saturated. This offers the ability to assign
14041 different limits to different listening ports or addresses. See also the
14042 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014043
Willy Tarreau16e01562016-08-09 16:46:18 +020014044dst_is_local : boolean
14045 Returns true if the destination address of the incoming connection is local
14046 to the system, or false if the address doesn't exist on the system, meaning
14047 that it was intercepted in transparent mode. It can be useful to apply
14048 certain rules by default to forwarded traffic and other rules to the traffic
Davor Ocelice9ed2812017-12-25 17:49:28 +010014049 targeting the real address of the machine. For example the stats page could
Willy Tarreau16e01562016-08-09 16:46:18 +020014050 be delivered only on this address, or SSH access could be locally redirected.
14051 Please note that the check involves a few system calls, so it's better to do
14052 it only once per connection.
14053
Willy Tarreau74ca5042013-06-11 23:12:07 +020014054dst_port : integer
14055 Returns an integer value corresponding to the destination TCP port of the
14056 connection on the client side, which is the port the client connected to.
14057 This might be used when running in transparent mode, when assigning dynamic
14058 ports to some clients for a whole application session, to stick all users to
14059 a same server, or to pass the destination port information to a server using
14060 an HTTP header.
14061
Willy Tarreau60ca10a2017-08-18 15:26:54 +020014062fc_http_major : integer
14063 Reports the front connection's HTTP major version encoding, which may be 1
14064 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
14065 encoding and not on the version present in the request header.
14066
Emeric Brun4f603012017-01-05 15:11:44 +010014067fc_rcvd_proxy : boolean
14068 Returns true if the client initiated the connection with a PROXY protocol
14069 header.
14070
Thierry Fournier / OZON.IO6310bef2016-07-24 20:16:50 +020014071fc_rtt(<unit>) : integer
14072 Returns the Round Trip Time (RTT) measured by the kernel for the client
14073 connection. <unit> is facultative, by default the unit is milliseconds. <unit>
14074 can be set to "ms" for milliseconds or "us" for microseconds. If the server
14075 connection is not established, if the connection is not TCP or if the
14076 operating system does not support TCP_INFO, for example Linux kernels before
14077 2.4, the sample fetch fails.
14078
14079fc_rttvar(<unit>) : integer
14080 Returns the Round Trip Time (RTT) variance measured by the kernel for the
14081 client connection. <unit> is facultative, by default the unit is milliseconds.
14082 <unit> can be set to "ms" for milliseconds or "us" for microseconds. If the
14083 server connection is not established, if the connection is not TCP or if the
14084 operating system does not support TCP_INFO, for example Linux kernels before
14085 2.4, the sample fetch fails.
14086
Joe Williams30fcd392016-08-10 07:06:44 -070014087fc_unacked(<unit>) : integer
14088 Returns the unacked counter measured by the kernel for the client connection.
14089 If the server connection is not established, if the connection is not TCP or
14090 if the operating system does not support TCP_INFO, for example Linux kernels
14091 before 2.4, the sample fetch fails.
14092
14093fc_sacked(<unit>) : integer
14094 Returns the sacked counter measured by the kernel for the client connection.
14095 If the server connection is not established, if the connection is not TCP or
14096 if the operating system does not support TCP_INFO, for example Linux kernels
14097 before 2.4, the sample fetch fails.
14098
14099fc_retrans(<unit>) : integer
14100 Returns the retransmits counter measured by the kernel for the client
14101 connection. If the server connection is not established, if the connection is
14102 not TCP or if the operating system does not support TCP_INFO, for example
14103 Linux kernels before 2.4, the sample fetch fails.
14104
14105fc_fackets(<unit>) : integer
14106 Returns the fack counter measured by the kernel for the client
14107 connection. If the server connection is not established, if the connection is
14108 not TCP or if the operating system does not support TCP_INFO, for example
14109 Linux kernels before 2.4, the sample fetch fails.
14110
14111fc_lost(<unit>) : integer
14112 Returns the lost counter measured by the kernel for the client
14113 connection. If the server connection is not established, if the connection is
14114 not TCP or if the operating system does not support TCP_INFO, for example
14115 Linux kernels before 2.4, the sample fetch fails.
14116
14117fc_reordering(<unit>) : integer
14118 Returns the reordering counter measured by the kernel for the client
14119 connection. If the server connection is not established, if the connection is
14120 not TCP or if the operating system does not support TCP_INFO, for example
14121 Linux kernels before 2.4, the sample fetch fails.
14122
Marcin Deranek9a66dfb2018-04-13 14:37:50 +020014123fe_defbe : string
14124 Returns a string containing the frontend's default backend name. It can be
14125 used in frontends to check which backend will handle requests by default.
14126
Willy Tarreau74ca5042013-06-11 23:12:07 +020014127fe_id : integer
14128 Returns an integer containing the current frontend's id. It can be used in
Marcin Deranek6e413ed2016-12-13 12:40:01 +010014129 backends to check from which frontend it was called, or to stick all users
Willy Tarreau74ca5042013-06-11 23:12:07 +020014130 coming via a same frontend to the same server.
14131
Marcin Deranekd2471c22016-12-12 14:08:05 +010014132fe_name : string
14133 Returns a string containing the current frontend's name. It can be used in
14134 backends to check from which frontend it was called, or to stick all users
14135 coming via a same frontend to the same server.
14136
Cyril Bonté62ba8702014-04-22 23:52:25 +020014137sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014138sc0_bytes_in_rate([<table>]) : integer
14139sc1_bytes_in_rate([<table>]) : integer
14140sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014141 Returns the average client-to-server bytes rate from the currently tracked
14142 counters, measured in amount of bytes over the period configured in the
14143 table. See also src_bytes_in_rate.
14144
Cyril Bonté62ba8702014-04-22 23:52:25 +020014145sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014146sc0_bytes_out_rate([<table>]) : integer
14147sc1_bytes_out_rate([<table>]) : integer
14148sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014149 Returns the average server-to-client bytes rate from the currently tracked
14150 counters, measured in amount of bytes over the period configured in the
14151 table. See also src_bytes_out_rate.
14152
Cyril Bonté62ba8702014-04-22 23:52:25 +020014153sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014154sc0_clr_gpc0([<table>]) : integer
14155sc1_clr_gpc0([<table>]) : integer
14156sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020014157 Clears the first General Purpose Counter associated to the currently tracked
14158 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010014159 stored value is zero, so first invocation will always return zero. This is
14160 typically used as a second ACL in an expression in order to mark a connection
14161 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020014162
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014163 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020014164 # block if 5 consecutive requests continue to come faster than 10 sess
14165 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014166 acl abuse sc0_http_req_rate gt 10
14167 acl kill sc0_inc_gpc0 gt 5
14168 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020014169 tcp-request connection accept if !abuse save
14170 tcp-request connection reject if abuse kill
14171
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014172sc_clr_gpc1(<ctr>[,<table>]) : integer
14173sc0_clr_gpc1([<table>]) : integer
14174sc1_clr_gpc1([<table>]) : integer
14175sc2_clr_gpc1([<table>]) : integer
14176 Clears the second General Purpose Counter associated to the currently tracked
14177 counters, and returns its previous value. Before the first invocation, the
14178 stored value is zero, so first invocation will always return zero. This is
14179 typically used as a second ACL in an expression in order to mark a connection
14180 when a first ACL was verified.
14181
Cyril Bonté62ba8702014-04-22 23:52:25 +020014182sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014183sc0_conn_cnt([<table>]) : integer
14184sc1_conn_cnt([<table>]) : integer
14185sc2_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014186 Returns the cumulative number of incoming connections from currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014187 counters. See also src_conn_cnt.
14188
Cyril Bonté62ba8702014-04-22 23:52:25 +020014189sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014190sc0_conn_cur([<table>]) : integer
14191sc1_conn_cur([<table>]) : integer
14192sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014193 Returns the current amount of concurrent connections tracking the same
14194 tracked counters. This number is automatically incremented when tracking
14195 begins and decremented when tracking stops. See also src_conn_cur.
14196
Cyril Bonté62ba8702014-04-22 23:52:25 +020014197sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014198sc0_conn_rate([<table>]) : integer
14199sc1_conn_rate([<table>]) : integer
14200sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014201 Returns the average connection rate from the currently tracked counters,
14202 measured in amount of connections over the period configured in the table.
14203 See also src_conn_rate.
14204
Cyril Bonté62ba8702014-04-22 23:52:25 +020014205sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014206sc0_get_gpc0([<table>]) : integer
14207sc1_get_gpc0([<table>]) : integer
14208sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014209 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014210 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014211
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014212sc_get_gpc1(<ctr>[,<table>]) : integer
14213sc0_get_gpc1([<table>]) : integer
14214sc1_get_gpc1([<table>]) : integer
14215sc2_get_gpc1([<table>]) : integer
14216 Returns the value of the second General Purpose Counter associated to the
14217 currently tracked counters. See also src_get_gpc1 and sc/sc0/sc1/sc2_inc_gpc1.
14218
Thierry FOURNIER236657b2015-08-19 08:25:14 +020014219sc_get_gpt0(<ctr>[,<table>]) : integer
14220sc0_get_gpt0([<table>]) : integer
14221sc1_get_gpt0([<table>]) : integer
14222sc2_get_gpt0([<table>]) : integer
14223 Returns the value of the first General Purpose Tag associated to the
14224 currently tracked counters. See also src_get_gpt0.
14225
Cyril Bonté62ba8702014-04-22 23:52:25 +020014226sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014227sc0_gpc0_rate([<table>]) : integer
14228sc1_gpc0_rate([<table>]) : integer
14229sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014230 Returns the average increment rate of the first General Purpose Counter
14231 associated to the currently tracked counters. It reports the frequency
14232 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014233 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
14234 that the "gpc0_rate" counter must be stored in the stick-table for a value to
14235 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020014236
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014237sc_gpc1_rate(<ctr>[,<table>]) : integer
14238sc0_gpc1_rate([<table>]) : integer
14239sc1_gpc1_rate([<table>]) : integer
14240sc2_gpc1_rate([<table>]) : integer
14241 Returns the average increment rate of the second General Purpose Counter
14242 associated to the currently tracked counters. It reports the frequency
14243 which the gpc1 counter was incremented over the configured period. See also
14244 src_gpcA_rate, sc/sc0/sc1/sc2_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
14245 that the "gpc1_rate" counter must be stored in the stick-table for a value to
14246 be returned, as "gpc1" only holds the event count.
14247
Cyril Bonté62ba8702014-04-22 23:52:25 +020014248sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014249sc0_http_err_cnt([<table>]) : integer
14250sc1_http_err_cnt([<table>]) : integer
14251sc2_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014252 Returns the cumulative number of HTTP errors from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014253 counters. This includes the both request errors and 4xx error responses.
14254 See also src_http_err_cnt.
14255
Cyril Bonté62ba8702014-04-22 23:52:25 +020014256sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014257sc0_http_err_rate([<table>]) : integer
14258sc1_http_err_rate([<table>]) : integer
14259sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014260 Returns the average rate of HTTP errors from the currently tracked counters,
14261 measured in amount of errors over the period configured in the table. This
14262 includes the both request errors and 4xx error responses. See also
14263 src_http_err_rate.
14264
Cyril Bonté62ba8702014-04-22 23:52:25 +020014265sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014266sc0_http_req_cnt([<table>]) : integer
14267sc1_http_req_cnt([<table>]) : integer
14268sc2_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014269 Returns the cumulative number of HTTP requests from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014270 counters. This includes every started request, valid or not. See also
14271 src_http_req_cnt.
14272
Cyril Bonté62ba8702014-04-22 23:52:25 +020014273sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014274sc0_http_req_rate([<table>]) : integer
14275sc1_http_req_rate([<table>]) : integer
14276sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014277 Returns the average rate of HTTP requests from the currently tracked
14278 counters, measured in amount of requests over the period configured in
14279 the table. This includes every started request, valid or not. See also
14280 src_http_req_rate.
14281
Cyril Bonté62ba8702014-04-22 23:52:25 +020014282sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014283sc0_inc_gpc0([<table>]) : integer
14284sc1_inc_gpc0([<table>]) : integer
14285sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014286 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010014287 tracked counters, and returns its new value. Before the first invocation,
14288 the stored value is zero, so first invocation will increase it to 1 and will
14289 return 1. This is typically used as a second ACL in an expression in order
14290 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020014291
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014292 Example:
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014293 acl abuse sc0_http_req_rate gt 10
14294 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020014295 tcp-request connection reject if abuse kill
14296
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014297sc_inc_gpc1(<ctr>[,<table>]) : integer
14298sc0_inc_gpc1([<table>]) : integer
14299sc1_inc_gpc1([<table>]) : integer
14300sc2_inc_gpc1([<table>]) : integer
14301 Increments the second General Purpose Counter associated to the currently
14302 tracked counters, and returns its new value. Before the first invocation,
14303 the stored value is zero, so first invocation will increase it to 1 and will
14304 return 1. This is typically used as a second ACL in an expression in order
14305 to mark a connection when a first ACL was verified.
14306
Cyril Bonté62ba8702014-04-22 23:52:25 +020014307sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014308sc0_kbytes_in([<table>]) : integer
14309sc1_kbytes_in([<table>]) : integer
14310sc2_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020014311 Returns the total amount of client-to-server data from the currently tracked
14312 counters, measured in kilobytes. The test is currently performed on 32-bit
14313 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020014314
Cyril Bonté62ba8702014-04-22 23:52:25 +020014315sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014316sc0_kbytes_out([<table>]) : integer
14317sc1_kbytes_out([<table>]) : integer
14318sc2_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020014319 Returns the total amount of server-to-client data from the currently tracked
14320 counters, measured in kilobytes. The test is currently performed on 32-bit
14321 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020014322
Cyril Bonté62ba8702014-04-22 23:52:25 +020014323sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014324sc0_sess_cnt([<table>]) : integer
14325sc1_sess_cnt([<table>]) : integer
14326sc2_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014327 Returns the cumulative number of incoming connections that were transformed
Willy Tarreaue9656522010-08-17 15:40:09 +020014328 into sessions, which means that they were accepted by a "tcp-request
14329 connection" rule, from the currently tracked counters. A backend may count
14330 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040014331 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020014332 with the client. See also src_sess_cnt.
14333
Cyril Bonté62ba8702014-04-22 23:52:25 +020014334sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014335sc0_sess_rate([<table>]) : integer
14336sc1_sess_rate([<table>]) : integer
14337sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014338 Returns the average session rate from the currently tracked counters,
14339 measured in amount of sessions over the period configured in the table. A
14340 session is a connection that got past the early "tcp-request connection"
14341 rules. A backend may count more sessions than connections because each
14342 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040014343 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020014344
Cyril Bonté62ba8702014-04-22 23:52:25 +020014345sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020014346sc0_tracked([<table>]) : boolean
14347sc1_tracked([<table>]) : boolean
14348sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020014349 Returns true if the designated session counter is currently being tracked by
14350 the current session. This can be useful when deciding whether or not we want
14351 to set some values in a header passed to the server.
14352
Cyril Bonté62ba8702014-04-22 23:52:25 +020014353sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014354sc0_trackers([<table>]) : integer
14355sc1_trackers([<table>]) : integer
14356sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010014357 Returns the current amount of concurrent connections tracking the same
14358 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014359 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010014360 that it does not rely on any stored information but on the table's reference
14361 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020014362 may sometimes be more suited for layer7 tracking. It can be used to tell a
14363 server how many concurrent connections there are from a given address for
14364 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010014365
Willy Tarreau74ca5042013-06-11 23:12:07 +020014366so_id : integer
14367 Returns an integer containing the current listening socket's id. It is useful
14368 in frontends involving many "bind" lines, or to stick all users coming via a
14369 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014370
Willy Tarreau74ca5042013-06-11 23:12:07 +020014371src : ip
Davor Ocelice9ed2812017-12-25 17:49:28 +010014372 This is the source IPv4 address of the client of the session. It is of type
Willy Tarreau74ca5042013-06-11 23:12:07 +020014373 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
14374 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
14375 TCP-level source address which is used, and not the address of a client
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010014376 behind a proxy. However if the "accept-proxy" or "accept-netscaler-cip" bind
14377 directive is used, it can be the address of a client behind another
14378 PROXY-protocol compatible component for all rule sets except
14379 "tcp-request connection" which sees the real address.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014380
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010014381 Example:
14382 # add an HTTP header in requests with the originating address' country
14383 http-request set-header X-Country %[src,map_ip(geoip.lst)]
14384
Willy Tarreau74ca5042013-06-11 23:12:07 +020014385src_bytes_in_rate([<table>]) : integer
14386 Returns the average bytes rate from the incoming connection's source address
14387 in the current proxy's stick-table or in the designated stick-table, measured
14388 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014389 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014390
Willy Tarreau74ca5042013-06-11 23:12:07 +020014391src_bytes_out_rate([<table>]) : integer
14392 Returns the average bytes rate to the incoming connection's source address in
14393 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020014394 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014395 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014396
Willy Tarreau74ca5042013-06-11 23:12:07 +020014397src_clr_gpc0([<table>]) : integer
14398 Clears the first General Purpose Counter associated to the incoming
14399 connection's source address in the current proxy's stick-table or in the
14400 designated stick-table, and returns its previous value. If the address is not
14401 found, an entry is created and 0 is returned. This is typically used as a
14402 second ACL in an expression in order to mark a connection when a first ACL
14403 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020014404
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014405 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020014406 # block if 5 consecutive requests continue to come faster than 10 sess
14407 # per second, and reset the counter as soon as the traffic slows down.
14408 acl abuse src_http_req_rate gt 10
14409 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010014410 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020014411 tcp-request connection accept if !abuse save
14412 tcp-request connection reject if abuse kill
14413
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014414src_clr_gpc1([<table>]) : integer
14415 Clears the second General Purpose Counter associated to the incoming
14416 connection's source address in the current proxy's stick-table or in the
14417 designated stick-table, and returns its previous value. If the address is not
14418 found, an entry is created and 0 is returned. This is typically used as a
14419 second ACL in an expression in order to mark a connection when a first ACL
14420 was verified.
14421
Willy Tarreau74ca5042013-06-11 23:12:07 +020014422src_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014423 Returns the cumulative number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020014424 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020014425 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014426 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014427
Willy Tarreau74ca5042013-06-11 23:12:07 +020014428src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020014429 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020014430 current incoming connection's source address in the current proxy's
14431 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014432 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014433
Willy Tarreau74ca5042013-06-11 23:12:07 +020014434src_conn_rate([<table>]) : integer
14435 Returns the average connection rate from the incoming connection's source
14436 address in the current proxy's stick-table or in the designated stick-table,
14437 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014438 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014439
Willy Tarreau74ca5042013-06-11 23:12:07 +020014440src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020014441 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020014442 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020014443 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014444 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014445
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014446src_get_gpc1([<table>]) : integer
14447 Returns the value of the second General Purpose Counter associated to the
14448 incoming connection's source address in the current proxy's stick-table or in
14449 the designated stick-table. If the address is not found, zero is returned.
14450 See also sc/sc0/sc1/sc2_get_gpc1 and src_inc_gpc1.
14451
Thierry FOURNIER236657b2015-08-19 08:25:14 +020014452src_get_gpt0([<table>]) : integer
14453 Returns the value of the first General Purpose Tag associated to the
14454 incoming connection's source address in the current proxy's stick-table or in
14455 the designated stick-table. If the address is not found, zero is returned.
14456 See also sc/sc0/sc1/sc2_get_gpt0.
14457
Willy Tarreau74ca5042013-06-11 23:12:07 +020014458src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014459 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020014460 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014461 stick-table or in the designated stick-table. It reports the frequency
14462 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014463 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
14464 that the "gpc0_rate" counter must be stored in the stick-table for a value to
14465 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014466
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014467src_gpc1_rate([<table>]) : integer
14468 Returns the average increment rate of the second General Purpose Counter
14469 associated to the incoming connection's source address in the current proxy's
14470 stick-table or in the designated stick-table. It reports the frequency
14471 which the gpc1 counter was incremented over the configured period. See also
14472 sc/sc0/sc1/sc2_gpc1_rate, src_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
14473 that the "gpc1_rate" counter must be stored in the stick-table for a value to
14474 be returned, as "gpc1" only holds the event count.
14475
Willy Tarreau74ca5042013-06-11 23:12:07 +020014476src_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014477 Returns the cumulative number of HTTP errors from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020014478 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020014479 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014480 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020014481 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014482
Willy Tarreau74ca5042013-06-11 23:12:07 +020014483src_http_err_rate([<table>]) : integer
14484 Returns the average rate of HTTP errors from the incoming connection's source
14485 address in the current proxy's stick-table or in the designated stick-table,
14486 measured in amount of errors over the period configured in the table. This
14487 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014488 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014489
Willy Tarreau74ca5042013-06-11 23:12:07 +020014490src_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014491 Returns the cumulative number of HTTP requests from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020014492 source address in the current proxy's stick-table or in the designated stick-
14493 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014494 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014495
Willy Tarreau74ca5042013-06-11 23:12:07 +020014496src_http_req_rate([<table>]) : integer
14497 Returns the average rate of HTTP requests from the incoming connection's
14498 source address in the current proxy's stick-table or in the designated stick-
14499 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020014500 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014501 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014502
Willy Tarreau74ca5042013-06-11 23:12:07 +020014503src_inc_gpc0([<table>]) : integer
14504 Increments the first General Purpose Counter associated to the incoming
14505 connection's source address in the current proxy's stick-table or in the
14506 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014507 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014508 This is typically used as a second ACL in an expression in order to mark a
14509 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020014510
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014511 Example:
Willy Tarreauc9705a12010-07-27 20:05:50 +020014512 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010014513 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020014514 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020014515
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014516src_inc_gpc1([<table>]) : integer
14517 Increments the second General Purpose Counter associated to the incoming
14518 connection's source address in the current proxy's stick-table or in the
14519 designated stick-table, and returns its new value. If the address is not
14520 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc1.
14521 This is typically used as a second ACL in an expression in order to mark a
14522 connection when a first ACL was verified.
14523
Willy Tarreau16e01562016-08-09 16:46:18 +020014524src_is_local : boolean
14525 Returns true if the source address of the incoming connection is local to the
14526 system, or false if the address doesn't exist on the system, meaning that it
14527 comes from a remote machine. Note that UNIX addresses are considered local.
14528 It can be useful to apply certain access restrictions based on where the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014529 client comes from (e.g. require auth or https for remote machines). Please
Willy Tarreau16e01562016-08-09 16:46:18 +020014530 note that the check involves a few system calls, so it's better to do it only
14531 once per connection.
14532
Willy Tarreau74ca5042013-06-11 23:12:07 +020014533src_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020014534 Returns the total amount of data received from the incoming connection's
14535 source address in the current proxy's stick-table or in the designated
14536 stick-table, measured in kilobytes. If the address is not found, zero is
14537 returned. The test is currently performed on 32-bit integers, which limits
14538 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014539
Willy Tarreau74ca5042013-06-11 23:12:07 +020014540src_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020014541 Returns the total amount of data sent to the incoming connection's source
14542 address in the current proxy's stick-table or in the designated stick-table,
14543 measured in kilobytes. If the address is not found, zero is returned. The
14544 test is currently performed on 32-bit integers, which limits values to 4
14545 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020014546
Willy Tarreau74ca5042013-06-11 23:12:07 +020014547src_port : integer
14548 Returns an integer value corresponding to the TCP source port of the
14549 connection on the client side, which is the port the client connected from.
14550 Usage of this function is very limited as modern protocols do not care much
14551 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010014552
Willy Tarreau74ca5042013-06-11 23:12:07 +020014553src_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014554 Returns the cumulative number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020014555 connection's source IPv4 address in the current proxy's stick-table or in the
14556 designated stick-table, that were transformed into sessions, which means that
14557 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014558 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014559
Willy Tarreau74ca5042013-06-11 23:12:07 +020014560src_sess_rate([<table>]) : integer
14561 Returns the average session rate from the incoming connection's source
14562 address in the current proxy's stick-table or in the designated stick-table,
14563 measured in amount of sessions over the period configured in the table. A
14564 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014565 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020014566
Willy Tarreau74ca5042013-06-11 23:12:07 +020014567src_updt_conn_cnt([<table>]) : integer
14568 Creates or updates the entry associated to the incoming connection's source
14569 address in the current proxy's stick-table or in the designated stick-table.
14570 This table must be configured to store the "conn_cnt" data type, otherwise
14571 the match will be ignored. The current count is incremented by one, and the
14572 expiration timer refreshed. The updated count is returned, so this match
14573 can't return zero. This was used to reject service abusers based on their
14574 source address. Note: it is recommended to use the more complete "track-sc*"
14575 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020014576
14577 Example :
14578 # This frontend limits incoming SSH connections to 3 per 10 second for
14579 # each source address, and rejects excess connections until a 10 second
14580 # silence is observed. At most 20 addresses are tracked.
14581 listen ssh
14582 bind :22
14583 mode tcp
14584 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020014585 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020014586 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020014587 server local 127.0.0.1:22
14588
Willy Tarreau74ca5042013-06-11 23:12:07 +020014589srv_id : integer
14590 Returns an integer containing the server's id when processing the response.
14591 While it's almost only used with ACLs, it may be used for logging or
14592 debugging.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020014593
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200145947.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020014595----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020014596
Willy Tarreau74ca5042013-06-11 23:12:07 +020014597The layer 5 usually describes just the session layer which in haproxy is
14598closest to the session once all the connection handshakes are finished, but
14599when no content is yet made available. The fetch methods described here are
14600usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014601future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020014602
Ben Shillitof25e8e52016-12-02 14:25:37 +00001460351d.all(<prop>[,<prop>*]) : string
14604 Returns values for the properties requested as a string, where values are
14605 separated by the delimiter specified with "51degrees-property-separator".
14606 The device is identified using all the important HTTP headers from the
14607 request. The function can be passed up to five property names, and if a
14608 property name can't be found, the value "NoData" is returned.
14609
14610 Example :
14611 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
14612 # containing the three properties requested using all relevant headers from
14613 # the request.
14614 frontend http-in
14615 bind *:8081
14616 default_backend servers
14617 http-request set-header X-51D-DeviceTypeMobileTablet \
14618 %[51d.all(DeviceType,IsMobile,IsTablet)]
14619
Emeric Brun645ae792014-04-30 14:21:06 +020014620ssl_bc : boolean
14621 Returns true when the back connection was made via an SSL/TLS transport
14622 layer and is locally deciphered. This means the outgoing connection was made
14623 other a server with the "ssl" option.
14624
14625ssl_bc_alg_keysize : integer
14626 Returns the symmetric cipher key size supported in bits when the outgoing
14627 connection was made over an SSL/TLS transport layer.
14628
14629ssl_bc_cipher : string
14630 Returns the name of the used cipher when the outgoing connection was made
14631 over an SSL/TLS transport layer.
14632
Emeric Brun74f7ffa2018-02-19 16:14:12 +010014633ssl_bc_is_resumed : boolean
14634 Returns true when the back connection was made over an SSL/TLS transport
14635 layer and the newly created SSL session was resumed using a cached
14636 session or a TLS ticket.
14637
Emeric Brun645ae792014-04-30 14:21:06 +020014638ssl_bc_protocol : string
14639 Returns the name of the used protocol when the outgoing connection was made
14640 over an SSL/TLS transport layer.
14641
Emeric Brunb73a9b02014-04-30 18:49:19 +020014642ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020014643 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020014644 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
14645 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
Emeric Brun645ae792014-04-30 14:21:06 +020014646
14647ssl_bc_session_id : binary
14648 Returns the SSL ID of the back connection when the outgoing connection was
14649 made over an SSL/TLS transport layer. It is useful to log if we want to know
14650 if session was reused or not.
14651
Patrick Hemmere0275472018-04-28 19:15:51 -040014652ssl_bc_session_key : binary
14653 Returns the SSL session master key of the back connection when the outgoing
14654 connection was made over an SSL/TLS transport layer. It is useful to decrypt
14655 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
14656 BoringSSL.
14657
Emeric Brun645ae792014-04-30 14:21:06 +020014658ssl_bc_use_keysize : integer
14659 Returns the symmetric cipher key size used in bits when the outgoing
14660 connection was made over an SSL/TLS transport layer.
14661
Willy Tarreau74ca5042013-06-11 23:12:07 +020014662ssl_c_ca_err : integer
14663 When the incoming connection was made over an SSL/TLS transport layer,
14664 returns the ID of the first error detected during verification of the client
14665 certificate at depth > 0, or 0 if no error was encountered during this
14666 verification process. Please refer to your SSL library's documentation to
14667 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020014668
Willy Tarreau74ca5042013-06-11 23:12:07 +020014669ssl_c_ca_err_depth : integer
14670 When the incoming connection was made over an SSL/TLS transport layer,
14671 returns the depth in the CA chain of the first error detected during the
14672 verification of the client certificate. If no error is encountered, 0 is
14673 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010014674
Emeric Brun43e79582014-10-29 19:03:26 +010014675ssl_c_der : binary
14676 Returns the DER formatted certificate presented by the client when the
14677 incoming connection was made over an SSL/TLS transport layer. When used for
14678 an ACL, the value(s) to match against can be passed in hexadecimal form.
14679
Willy Tarreau74ca5042013-06-11 23:12:07 +020014680ssl_c_err : integer
14681 When the incoming connection was made over an SSL/TLS transport layer,
14682 returns the ID of the first error detected during verification at depth 0, or
14683 0 if no error was encountered during this verification process. Please refer
14684 to your SSL library's documentation to find the exhaustive list of error
14685 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020014686
Willy Tarreau74ca5042013-06-11 23:12:07 +020014687ssl_c_i_dn([<entry>[,<occ>]]) : string
14688 When the incoming connection was made over an SSL/TLS transport layer,
14689 returns the full distinguished name of the issuer of the certificate
14690 presented by the client when no <entry> is specified, or the value of the
14691 first given entry found from the beginning of the DN. If a positive/negative
14692 occurrence number is specified as the optional second argument, it returns
14693 the value of the nth given entry value from the beginning/end of the DN.
14694 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
14695 "ssl_c_i_dn(CN)" retrieves the common name.
Willy Tarreau62644772008-07-16 18:36:06 +020014696
Willy Tarreau74ca5042013-06-11 23:12:07 +020014697ssl_c_key_alg : string
14698 Returns the name of the algorithm used to generate the key of the certificate
14699 presented by the client when the incoming connection was made over an SSL/TLS
14700 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020014701
Willy Tarreau74ca5042013-06-11 23:12:07 +020014702ssl_c_notafter : string
14703 Returns the end date presented by the client as a formatted string
14704 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
14705 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020014706
Willy Tarreau74ca5042013-06-11 23:12:07 +020014707ssl_c_notbefore : string
14708 Returns the start date presented by the client as a formatted string
14709 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
14710 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010014711
Willy Tarreau74ca5042013-06-11 23:12:07 +020014712ssl_c_s_dn([<entry>[,<occ>]]) : string
14713 When the incoming connection was made over an SSL/TLS transport layer,
14714 returns the full distinguished name of the subject of the certificate
14715 presented by the client when no <entry> is specified, or the value of the
14716 first given entry found from the beginning of the DN. If a positive/negative
14717 occurrence number is specified as the optional second argument, it returns
14718 the value of the nth given entry value from the beginning/end of the DN.
14719 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
14720 "ssl_c_s_dn(CN)" retrieves the common name.
Willy Tarreaub6672b52011-12-12 17:23:41 +010014721
Willy Tarreau74ca5042013-06-11 23:12:07 +020014722ssl_c_serial : binary
14723 Returns the serial of the certificate presented by the client when the
14724 incoming connection was made over an SSL/TLS transport layer. When used for
14725 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020014726
Willy Tarreau74ca5042013-06-11 23:12:07 +020014727ssl_c_sha1 : binary
14728 Returns the SHA-1 fingerprint of the certificate presented by the client when
14729 the incoming connection was made over an SSL/TLS transport layer. This can be
14730 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau2d0caa32014-07-02 19:01:22 +020014731 Note that the output is binary, so if you want to pass that signature to the
14732 server, you need to encode it in hex or base64, such as in the example below:
14733
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014734 Example:
Willy Tarreau2d0caa32014-07-02 19:01:22 +020014735 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020014736
Willy Tarreau74ca5042013-06-11 23:12:07 +020014737ssl_c_sig_alg : string
14738 Returns the name of the algorithm used to sign the certificate presented by
14739 the client when the incoming connection was made over an SSL/TLS transport
14740 layer.
Emeric Brun87855892012-10-17 17:39:35 +020014741
Willy Tarreau74ca5042013-06-11 23:12:07 +020014742ssl_c_used : boolean
14743 Returns true if current SSL session uses a client certificate even if current
14744 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020014745
Willy Tarreau74ca5042013-06-11 23:12:07 +020014746ssl_c_verify : integer
14747 Returns the verify result error ID when the incoming connection was made over
14748 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
14749 refer to your SSL library's documentation for an exhaustive list of error
14750 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020014751
Willy Tarreau74ca5042013-06-11 23:12:07 +020014752ssl_c_version : integer
14753 Returns the version of the certificate presented by the client when the
14754 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020014755
Emeric Brun43e79582014-10-29 19:03:26 +010014756ssl_f_der : binary
14757 Returns the DER formatted certificate presented by the frontend when the
14758 incoming connection was made over an SSL/TLS transport layer. When used for
14759 an ACL, the value(s) to match against can be passed in hexadecimal form.
14760
Willy Tarreau74ca5042013-06-11 23:12:07 +020014761ssl_f_i_dn([<entry>[,<occ>]]) : string
14762 When the incoming connection was made over an SSL/TLS transport layer,
14763 returns the full distinguished name of the issuer of the certificate
14764 presented by the frontend when no <entry> is specified, or the value of the
14765 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020014766 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020014767 the value of the nth given entry value from the beginning/end of the DN.
14768 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
14769 "ssl_f_i_dn(CN)" retrieves the common name.
Emeric Brun87855892012-10-17 17:39:35 +020014770
Willy Tarreau74ca5042013-06-11 23:12:07 +020014771ssl_f_key_alg : string
14772 Returns the name of the algorithm used to generate the key of the certificate
14773 presented by the frontend when the incoming connection was made over an
14774 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020014775
Willy Tarreau74ca5042013-06-11 23:12:07 +020014776ssl_f_notafter : string
14777 Returns the end date presented by the frontend as a formatted string
14778 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
14779 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020014780
Willy Tarreau74ca5042013-06-11 23:12:07 +020014781ssl_f_notbefore : string
14782 Returns the start date presented by the frontend as a formatted string
14783 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
14784 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020014785
Willy Tarreau74ca5042013-06-11 23:12:07 +020014786ssl_f_s_dn([<entry>[,<occ>]]) : string
14787 When the incoming connection was made over an SSL/TLS transport layer,
14788 returns the full distinguished name of the subject of the certificate
14789 presented by the frontend when no <entry> is specified, or the value of the
14790 first given entry found from the beginning of the DN. If a positive/negative
14791 occurrence number is specified as the optional second argument, it returns
14792 the value of the nth given entry value from the beginning/end of the DN.
14793 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
14794 "ssl_f_s_dn(CN)" retrieves the common name.
Emeric Brunce5ad802012-10-22 14:11:22 +020014795
Willy Tarreau74ca5042013-06-11 23:12:07 +020014796ssl_f_serial : binary
14797 Returns the serial of the certificate presented by the frontend when the
14798 incoming connection was made over an SSL/TLS transport layer. When used for
14799 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020014800
Emeric Brun55f4fa82014-04-30 17:11:25 +020014801ssl_f_sha1 : binary
14802 Returns the SHA-1 fingerprint of the certificate presented by the frontend
14803 when the incoming connection was made over an SSL/TLS transport layer. This
14804 can be used to know which certificate was chosen using SNI.
14805
Willy Tarreau74ca5042013-06-11 23:12:07 +020014806ssl_f_sig_alg : string
14807 Returns the name of the algorithm used to sign the certificate presented by
14808 the frontend when the incoming connection was made over an SSL/TLS transport
14809 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020014810
Willy Tarreau74ca5042013-06-11 23:12:07 +020014811ssl_f_version : integer
14812 Returns the version of the certificate presented by the frontend when the
14813 incoming connection was made over an SSL/TLS transport layer.
14814
14815ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020014816 Returns true when the front connection was made via an SSL/TLS transport
14817 layer and is locally deciphered. This means it has matched a socket declared
14818 with a "bind" line having the "ssl" option.
14819
Willy Tarreau74ca5042013-06-11 23:12:07 +020014820 Example :
14821 # This passes "X-Proto: https" to servers when client connects over SSL
14822 listen http-https
14823 bind :80
14824 bind :443 ssl crt /etc/haproxy.pem
14825 http-request add-header X-Proto https if { ssl_fc }
14826
14827ssl_fc_alg_keysize : integer
14828 Returns the symmetric cipher key size supported in bits when the incoming
14829 connection was made over an SSL/TLS transport layer.
14830
14831ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014832 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020014833 incoming connection made via a TLS transport layer and locally deciphered by
14834 haproxy. The result is a string containing the protocol name advertised by
14835 the client. The SSL library must have been built with support for TLS
14836 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
14837 not advertised unless the "alpn" keyword on the "bind" line specifies a
14838 protocol list. Also, nothing forces the client to pick a protocol from this
14839 list, any other one may be requested. The TLS ALPN extension is meant to
14840 replace the TLS NPN extension. See also "ssl_fc_npn".
14841
Willy Tarreau74ca5042013-06-11 23:12:07 +020014842ssl_fc_cipher : string
14843 Returns the name of the used cipher when the incoming connection was made
14844 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020014845
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010014846ssl_fc_cipherlist_bin : binary
14847 Returns the binary form of the client hello cipher list. The maximum returned
14848 value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010014849 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010014850
14851ssl_fc_cipherlist_hex : string
14852 Returns the binary form of the client hello cipher list encoded as
14853 hexadecimal. The maximum returned value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010014854 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010014855
14856ssl_fc_cipherlist_str : string
14857 Returns the decoded text form of the client hello cipher list. The maximum
14858 number of ciphers returned is according with the value of
14859 "tune.ssl.capture-cipherlist-size". Note that this sample-fetch is only
Davor Ocelice9ed2812017-12-25 17:49:28 +010014860 available with OpenSSL >= 1.0.2. If the function is not enabled, this
Emmanuel Hocdetddcde192017-09-01 17:32:08 +020014861 sample-fetch returns the hash like "ssl_fc_cipherlist_xxh".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010014862
14863ssl_fc_cipherlist_xxh : integer
14864 Returns a xxh64 of the cipher list. This hash can be return only is the value
14865 "tune.ssl.capture-cipherlist-size" is set greater than 0, however the hash
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010014866 take in account all the data of the cipher list.
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010014867
Willy Tarreau74ca5042013-06-11 23:12:07 +020014868ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020014869 Returns true if a client certificate is present in an incoming connection over
14870 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010014871 Note: on SSL session resumption with Session ID or TLS ticket, client
14872 certificate is not present in the current connection but may be retrieved
14873 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
14874 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020014875
Olivier Houchardccaa7de2017-10-02 11:51:03 +020014876ssl_fc_has_early : boolean
14877 Returns true if early data were sent, and the handshake didn't happen yet. As
14878 it has security implications, it is useful to be able to refuse those, or
14879 wait until the handshake happened.
14880
Willy Tarreau74ca5042013-06-11 23:12:07 +020014881ssl_fc_has_sni : boolean
14882 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020014883 in an incoming connection was made over an SSL/TLS transport layer. Returns
14884 true when the incoming connection presents a TLS SNI field. This requires
14885 that the SSL library is build with support for TLS extensions enabled (check
14886 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020014887
Nenad Merdanovic1516fe32016-05-17 03:31:21 +020014888ssl_fc_is_resumed : boolean
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020014889 Returns true if the SSL/TLS session has been resumed through the use of
Jérôme Magnin4a326cb2018-01-15 14:01:17 +010014890 SSL session cache or TLS tickets on an incoming connection over an SSL/TLS
14891 transport layer.
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020014892
Willy Tarreau74ca5042013-06-11 23:12:07 +020014893ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014894 This extracts the Next Protocol Negotiation field from an incoming connection
Willy Tarreau74ca5042013-06-11 23:12:07 +020014895 made via a TLS transport layer and locally deciphered by haproxy. The result
14896 is a string containing the protocol name advertised by the client. The SSL
14897 library must have been built with support for TLS extensions enabled (check
14898 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
14899 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
14900 forces the client to pick a protocol from this list, any other one may be
14901 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020014902
Willy Tarreau74ca5042013-06-11 23:12:07 +020014903ssl_fc_protocol : string
14904 Returns the name of the used protocol when the incoming connection was made
14905 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020014906
Emeric Brunb73a9b02014-04-30 18:49:19 +020014907ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040014908 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020014909 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
14910 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040014911
Willy Tarreau74ca5042013-06-11 23:12:07 +020014912ssl_fc_session_id : binary
14913 Returns the SSL ID of the front connection when the incoming connection was
14914 made over an SSL/TLS transport layer. It is useful to stick a given client to
14915 a server. It is important to note that some browsers refresh their session ID
14916 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020014917
Patrick Hemmere0275472018-04-28 19:15:51 -040014918ssl_fc_session_key : binary
14919 Returns the SSL session master key of the front connection when the incoming
14920 connection was made over an SSL/TLS transport layer. It is useful to decrypt
14921 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
14922 BoringSSL.
14923
14924
Willy Tarreau74ca5042013-06-11 23:12:07 +020014925ssl_fc_sni : string
14926 This extracts the Server Name Indication TLS extension (SNI) field from an
14927 incoming connection made via an SSL/TLS transport layer and locally
14928 deciphered by haproxy. The result (when present) typically is a string
14929 matching the HTTPS host name (253 chars or less). The SSL library must have
14930 been built with support for TLS extensions enabled (check haproxy -vv).
14931
14932 This fetch is different from "req_ssl_sni" above in that it applies to the
14933 connection being deciphered by haproxy and not to SSL contents being blindly
14934 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020014935 requires that the SSL library is build with support for TLS extensions
14936 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020014937
Willy Tarreau74ca5042013-06-11 23:12:07 +020014938 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020014939 ssl_fc_sni_end : suffix match
14940 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020014941
Willy Tarreau74ca5042013-06-11 23:12:07 +020014942ssl_fc_use_keysize : integer
14943 Returns the symmetric cipher key size used in bits when the incoming
14944 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020014945
Willy Tarreaub6fb4202008-07-20 11:18:28 +020014946
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200149477.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020014948------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020014949
Willy Tarreau74ca5042013-06-11 23:12:07 +020014950Fetching samples from buffer contents is a bit different from the previous
14951sample fetches above because the sampled data are ephemeral. These data can
14952only be used when they're available and will be lost when they're forwarded.
14953For this reason, samples fetched from buffer contents during a request cannot
14954be used in a response for example. Even while the data are being fetched, they
14955can change. Sometimes it is necessary to set some delays or combine multiple
14956sample fetch methods to ensure that the expected data are complete and usable,
14957for example through TCP request content inspection. Please see the "tcp-request
14958content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020014959
Willy Tarreau74ca5042013-06-11 23:12:07 +020014960payload(<offset>,<length>) : binary (deprecated)
Davor Ocelice9ed2812017-12-25 17:49:28 +010014961 This is an alias for "req.payload" when used in the context of a request (e.g.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014962 "stick on", "stick match"), and for "res.payload" when used in the context of
14963 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010014964
Willy Tarreau74ca5042013-06-11 23:12:07 +020014965payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
14966 This is an alias for "req.payload_lv" when used in the context of a request
Davor Ocelice9ed2812017-12-25 17:49:28 +010014967 (e.g. "stick on", "stick match"), and for "res.payload_lv" when used in the
Willy Tarreau74ca5042013-06-11 23:12:07 +020014968 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010014969
Thierry FOURNIERd7d88812017-04-19 15:15:14 +020014970req.hdrs : string
14971 Returns the current request headers as string including the last empty line
14972 separating headers from the request body. The last empty line can be used to
14973 detect a truncated header block. This sample fetch is useful for some SPOE
14974 headers analyzers and for advanced logging.
14975
Thierry FOURNIER5617dce2017-04-09 05:38:19 +020014976req.hdrs_bin : binary
14977 Returns the current request headers contained in preparsed binary form. This
14978 is useful for offloading some processing with SPOE. Each string is described
14979 by a length followed by the number of bytes indicated in the length. The
14980 length is represented using the variable integer encoding detailed in the
14981 SPOE documentation. The end of the list is marked by a couple of empty header
14982 names and values (length of 0 for both).
14983
14984 *(<str:header-name><str:header-value>)<empty string><empty string>
14985
14986 int: refer to the SPOE documentation for the encoding
14987 str: <int:length><bytes>
14988
Willy Tarreau74ca5042013-06-11 23:12:07 +020014989req.len : integer
14990req_len : integer (deprecated)
14991 Returns an integer value corresponding to the number of bytes present in the
14992 request buffer. This is mostly used in ACL. It is important to understand
14993 that this test does not return false as long as the buffer is changing. This
14994 means that a check with equality to zero will almost always immediately match
14995 at the beginning of the session, while a test for more data will wait for
14996 that data to come in and return false only when haproxy is certain that no
14997 more data will come in. This test was designed to be used with TCP request
14998 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020014999
Willy Tarreau74ca5042013-06-11 23:12:07 +020015000req.payload(<offset>,<length>) : binary
15001 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020015002 in the request buffer. As a special case, if the <length> argument is zero,
15003 the the whole buffer from <offset> to the end is extracted. This can be used
15004 with ACLs in order to check for the presence of some content in a buffer at
15005 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015006
Willy Tarreau74ca5042013-06-11 23:12:07 +020015007 ACL alternatives :
15008 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015009
Willy Tarreau74ca5042013-06-11 23:12:07 +020015010req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
15011 This extracts a binary block whose size is specified at <offset1> for <length>
15012 bytes, and which starts at <offset2> if specified or just after the length in
15013 the request buffer. The <offset2> parameter also supports relative offsets if
15014 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015015
Willy Tarreau74ca5042013-06-11 23:12:07 +020015016 ACL alternatives :
15017 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015018
Willy Tarreau74ca5042013-06-11 23:12:07 +020015019 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015020
Willy Tarreau74ca5042013-06-11 23:12:07 +020015021req.proto_http : boolean
15022req_proto_http : boolean (deprecated)
15023 Returns true when data in the request buffer look like HTTP and correctly
15024 parses as such. It is the same parser as the common HTTP request parser which
15025 is used so there should be no surprises. The test does not match until the
15026 request is complete, failed or timed out. This test may be used to report the
15027 protocol in TCP logs, but the biggest use is to block TCP request analysis
15028 until a complete HTTP request is present in the buffer, for example to track
15029 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015030
Willy Tarreau74ca5042013-06-11 23:12:07 +020015031 Example:
15032 # track request counts per "base" (concatenation of Host+URL)
15033 tcp-request inspect-delay 10s
15034 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020015035 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015036
Willy Tarreau74ca5042013-06-11 23:12:07 +020015037req.rdp_cookie([<name>]) : string
15038rdp_cookie([<name>]) : string (deprecated)
15039 When the request buffer looks like the RDP protocol, extracts the RDP cookie
15040 <name>, or any cookie if unspecified. The parser only checks for the first
15041 cookie, as illustrated in the RDP protocol specification. The cookie name is
15042 case insensitive. Generally the "MSTS" cookie name will be used, as it can
15043 contain the user name of the client connecting to the server if properly
15044 configured on the client. The "MSTSHASH" cookie is often used as well for
15045 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015046
Willy Tarreau74ca5042013-06-11 23:12:07 +020015047 This differs from "balance rdp-cookie" in that any balancing algorithm may be
15048 used and thus the distribution of clients to backend servers is not linked to
15049 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
15050 such as "balance roundrobin" or "balance leastconn" will lead to a more even
15051 distribution of clients to backend servers than the hash used by "balance
15052 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015053
Willy Tarreau74ca5042013-06-11 23:12:07 +020015054 ACL derivatives :
15055 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015056
Willy Tarreau74ca5042013-06-11 23:12:07 +020015057 Example :
15058 listen tse-farm
15059 bind 0.0.0.0:3389
15060 # wait up to 5s for an RDP cookie in the request
15061 tcp-request inspect-delay 5s
15062 tcp-request content accept if RDP_COOKIE
15063 # apply RDP cookie persistence
15064 persist rdp-cookie
15065 # Persist based on the mstshash cookie
15066 # This is only useful makes sense if
15067 # balance rdp-cookie is not used
15068 stick-table type string size 204800
15069 stick on req.rdp_cookie(mstshash)
15070 server srv1 1.1.1.1:3389
15071 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015072
Willy Tarreau74ca5042013-06-11 23:12:07 +020015073 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
15074 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015075
Willy Tarreau74ca5042013-06-11 23:12:07 +020015076req.rdp_cookie_cnt([name]) : integer
15077rdp_cookie_cnt([name]) : integer (deprecated)
15078 Tries to parse the request buffer as RDP protocol, then returns an integer
15079 corresponding to the number of RDP cookies found. If an optional cookie name
15080 is passed, only cookies matching this name are considered. This is mostly
15081 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015082
Willy Tarreau74ca5042013-06-11 23:12:07 +020015083 ACL derivatives :
15084 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015085
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020015086req.ssl_ec_ext : boolean
15087 Returns a boolean identifying if client sent the Supported Elliptic Curves
15088 Extension as defined in RFC4492, section 5.1. within the SSL ClientHello
Cyril Bonté307ee1e2015-09-28 23:16:06 +020015089 message. This can be used to present ECC compatible clients with EC
15090 certificate and to use RSA for all others, on the same IP address. Note that
15091 this only applies to raw contents found in the request buffer and not to
15092 contents deciphered via an SSL data layer, so this will not work with "bind"
15093 lines having the "ssl" option.
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020015094
Willy Tarreau74ca5042013-06-11 23:12:07 +020015095req.ssl_hello_type : integer
15096req_ssl_hello_type : integer (deprecated)
15097 Returns an integer value containing the type of the SSL hello message found
15098 in the request buffer if the buffer contains data that parse as a complete
15099 SSL (v3 or superior) client hello message. Note that this only applies to raw
15100 contents found in the request buffer and not to contents deciphered via an
15101 SSL data layer, so this will not work with "bind" lines having the "ssl"
15102 option. This is mostly used in ACL to detect presence of an SSL hello message
15103 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015104
Willy Tarreau74ca5042013-06-11 23:12:07 +020015105req.ssl_sni : string
15106req_ssl_sni : string (deprecated)
15107 Returns a string containing the value of the Server Name TLS extension sent
15108 by a client in a TLS stream passing through the request buffer if the buffer
15109 contains data that parse as a complete SSL (v3 or superior) client hello
15110 message. Note that this only applies to raw contents found in the request
15111 buffer and not to contents deciphered via an SSL data layer, so this will not
15112 work with "bind" lines having the "ssl" option. SNI normally contains the
15113 name of the host the client tries to connect to (for recent browsers). SNI is
15114 useful for allowing or denying access to certain hosts when SSL/TLS is used
15115 by the client. This test was designed to be used with TCP request content
15116 inspection. If content switching is needed, it is recommended to first wait
15117 for a complete client hello (type 1), like in the example below. See also
15118 "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015119
Willy Tarreau74ca5042013-06-11 23:12:07 +020015120 ACL derivatives :
15121 req_ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015122
Willy Tarreau74ca5042013-06-11 23:12:07 +020015123 Examples :
15124 # Wait for a client hello for at most 5 seconds
15125 tcp-request inspect-delay 5s
15126 tcp-request content accept if { req_ssl_hello_type 1 }
15127 use_backend bk_allow if { req_ssl_sni -f allowed_sites }
15128 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015129
Pradeep Jindalbb2acf52015-09-29 10:12:57 +053015130req.ssl_st_ext : integer
15131 Returns 0 if the client didn't send a SessionTicket TLS Extension (RFC5077)
15132 Returns 1 if the client sent SessionTicket TLS Extension
15133 Returns 2 if the client also sent non-zero length TLS SessionTicket
15134 Note that this only applies to raw contents found in the request buffer and
15135 not to contents deciphered via an SSL data layer, so this will not work with
15136 "bind" lines having the "ssl" option. This can for example be used to detect
15137 whether the client sent a SessionTicket or not and stick it accordingly, if
15138 no SessionTicket then stick on SessionID or don't stick as there's no server
15139 side state is there when SessionTickets are in use.
15140
Willy Tarreau74ca5042013-06-11 23:12:07 +020015141req.ssl_ver : integer
15142req_ssl_ver : integer (deprecated)
15143 Returns an integer value containing the version of the SSL/TLS protocol of a
15144 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
15145 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
15146 composed of the major version multiplied by 65536, added to the minor
15147 version. Note that this only applies to raw contents found in the request
15148 buffer and not to contents deciphered via an SSL data layer, so this will not
15149 work with "bind" lines having the "ssl" option. The ACL version of the test
Davor Ocelice9ed2812017-12-25 17:49:28 +010015150 matches against a decimal notation in the form MAJOR.MINOR (e.g. 3.1). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020015151 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015152
Willy Tarreau74ca5042013-06-11 23:12:07 +020015153 ACL derivatives :
15154 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010015155
Willy Tarreau47e8eba2013-09-11 23:28:46 +020015156res.len : integer
15157 Returns an integer value corresponding to the number of bytes present in the
15158 response buffer. This is mostly used in ACL. It is important to understand
15159 that this test does not return false as long as the buffer is changing. This
15160 means that a check with equality to zero will almost always immediately match
15161 at the beginning of the session, while a test for more data will wait for
15162 that data to come in and return false only when haproxy is certain that no
15163 more data will come in. This test was designed to be used with TCP response
15164 content inspection.
15165
Willy Tarreau74ca5042013-06-11 23:12:07 +020015166res.payload(<offset>,<length>) : binary
15167 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020015168 in the response buffer. As a special case, if the <length> argument is zero,
15169 the the whole buffer from <offset> to the end is extracted. This can be used
15170 with ACLs in order to check for the presence of some content in a buffer at
15171 any location.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015172
Willy Tarreau74ca5042013-06-11 23:12:07 +020015173res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
15174 This extracts a binary block whose size is specified at <offset1> for <length>
15175 bytes, and which starts at <offset2> if specified or just after the length in
15176 the response buffer. The <offset2> parameter also supports relative offsets
15177 if prepended with a '+' or '-' sign.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015178
Willy Tarreau74ca5042013-06-11 23:12:07 +020015179 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015180
Willy Tarreau971f7b62015-09-29 14:06:59 +020015181res.ssl_hello_type : integer
15182rep_ssl_hello_type : integer (deprecated)
15183 Returns an integer value containing the type of the SSL hello message found
15184 in the response buffer if the buffer contains data that parses as a complete
15185 SSL (v3 or superior) hello message. Note that this only applies to raw
15186 contents found in the response buffer and not to contents deciphered via an
15187 SSL data layer, so this will not work with "server" lines having the "ssl"
15188 option. This is mostly used in ACL to detect presence of an SSL hello message
15189 that is supposed to contain an SSL session ID usable for stickiness.
15190
Willy Tarreau74ca5042013-06-11 23:12:07 +020015191wait_end : boolean
15192 This fetch either returns true when the inspection period is over, or does
15193 not fetch. It is only used in ACLs, in conjunction with content analysis to
Davor Ocelice9ed2812017-12-25 17:49:28 +010015194 avoid returning a wrong verdict early. It may also be used to delay some
Willy Tarreau74ca5042013-06-11 23:12:07 +020015195 actions, such as a delayed reject for some special addresses. Since it either
15196 stops the rules evaluation or immediately returns true, it is recommended to
Davor Ocelice9ed2812017-12-25 17:49:28 +010015197 use this acl as the last one in a rule. Please note that the default ACL
Willy Tarreau74ca5042013-06-11 23:12:07 +020015198 "WAIT_END" is always usable without prior declaration. This test was designed
15199 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015200
Willy Tarreau74ca5042013-06-11 23:12:07 +020015201 Examples :
15202 # delay every incoming request by 2 seconds
15203 tcp-request inspect-delay 2s
15204 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010015205
Willy Tarreau74ca5042013-06-11 23:12:07 +020015206 # don't immediately tell bad guys they are rejected
15207 tcp-request inspect-delay 10s
15208 acl goodguys src 10.0.0.0/24
15209 acl badguys src 10.0.1.0/24
15210 tcp-request content accept if goodguys
15211 tcp-request content reject if badguys WAIT_END
15212 tcp-request content reject
15213
15214
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200152157.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020015216--------------------------------------
15217
15218It is possible to fetch samples from HTTP contents, requests and responses.
15219This application layer is also called layer 7. It is only possible to fetch the
15220data in this section when a full HTTP request or response has been parsed from
15221its respective request or response buffer. This is always the case with all
15222HTTP specific rules and for sections running with "mode http". When using TCP
15223content inspection, it may be necessary to support an inspection delay in order
15224to let the request or response come in first. These fetches may require a bit
15225more CPU resources than the layer 4 ones, but not much since the request and
15226response are indexed.
15227
15228base : string
15229 This returns the concatenation of the first Host header and the path part of
15230 the request, which starts at the first slash and ends before the question
15231 mark. It can be useful in virtual hosted environments to detect URL abuses as
15232 well as to improve shared caches efficiency. Using this with a limited size
15233 stick table also allows one to collect statistics about most commonly
15234 requested objects by host/path. With ACLs it can allow simple content
15235 switching rules involving the host and the path at the same time, such as
15236 "www.example.com/favicon.ico". See also "path" and "uri".
15237
15238 ACL derivatives :
15239 base : exact string match
15240 base_beg : prefix match
15241 base_dir : subdir match
15242 base_dom : domain match
15243 base_end : suffix match
15244 base_len : length match
15245 base_reg : regex match
15246 base_sub : substring match
15247
15248base32 : integer
15249 This returns a 32-bit hash of the value returned by the "base" fetch method
15250 above. This is useful to track per-URL activity on high traffic sites without
15251 having to store all URLs. Instead a shorter hash is stored, saving a lot of
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020015252 memory. The output type is an unsigned integer. The hash function used is
15253 SDBM with full avalanche on the output. Technically, base32 is exactly equal
15254 to "base,sdbm(1)".
Willy Tarreau74ca5042013-06-11 23:12:07 +020015255
15256base32+src : binary
15257 This returns the concatenation of the base32 fetch above and the src fetch
15258 below. The resulting type is of type binary, with a size of 8 or 20 bytes
15259 depending on the source address family. This can be used to track per-IP,
15260 per-URL counters.
15261
William Lallemand65ad6e12014-01-31 15:08:02 +010015262capture.req.hdr(<idx>) : string
15263 This extracts the content of the header captured by the "capture request
15264 header", idx is the position of the capture keyword in the configuration.
15265 The first entry is an index of 0. See also: "capture request header".
15266
15267capture.req.method : string
15268 This extracts the METHOD of an HTTP request. It can be used in both request
15269 and response. Unlike "method", it can be used in both request and response
15270 because it's allocated.
15271
15272capture.req.uri : string
15273 This extracts the request's URI, which starts at the first slash and ends
15274 before the first space in the request (without the host part). Unlike "path"
15275 and "url", it can be used in both request and response because it's
15276 allocated.
15277
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020015278capture.req.ver : string
15279 This extracts the request's HTTP version and returns either "HTTP/1.0" or
15280 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
15281 logs because it relies on a persistent flag.
15282
William Lallemand65ad6e12014-01-31 15:08:02 +010015283capture.res.hdr(<idx>) : string
15284 This extracts the content of the header captured by the "capture response
15285 header", idx is the position of the capture keyword in the configuration.
15286 The first entry is an index of 0.
15287 See also: "capture response header"
15288
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020015289capture.res.ver : string
15290 This extracts the response's HTTP version and returns either "HTTP/1.0" or
15291 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
15292 persistent flag.
15293
Willy Tarreaua5910cc2015-05-02 00:46:08 +020015294req.body : binary
15295 This returns the HTTP request's available body as a block of data. It
15296 requires that the request body has been buffered made available using
15297 "option http-buffer-request". In case of chunked-encoded body, currently only
15298 the first chunk is analyzed.
15299
Thierry FOURNIER9826c772015-05-20 15:50:54 +020015300req.body_param([<name>) : string
15301 This fetch assumes that the body of the POST request is url-encoded. The user
15302 can check if the "content-type" contains the value
15303 "application/x-www-form-urlencoded". This extracts the first occurrence of the
15304 parameter <name> in the body, which ends before '&'. The parameter name is
15305 case-sensitive. If no name is given, any parameter will match, and the first
15306 one will be returned. The result is a string corresponding to the value of the
15307 parameter <name> as presented in the request body (no URL decoding is
15308 performed). Note that the ACL version of this fetch iterates over multiple
15309 parameters and will iteratively report all parameters values if no name is
15310 given.
15311
Willy Tarreaua5910cc2015-05-02 00:46:08 +020015312req.body_len : integer
15313 This returns the length of the HTTP request's available body in bytes. It may
15314 be lower than the advertised length if the body is larger than the buffer. It
15315 requires that the request body has been buffered made available using
15316 "option http-buffer-request".
15317
15318req.body_size : integer
15319 This returns the advertised length of the HTTP request's body in bytes. It
15320 will represent the advertised Content-Length header, or the size of the first
15321 chunk in case of chunked encoding. In order to parse the chunks, it requires
15322 that the request body has been buffered made available using
15323 "option http-buffer-request".
15324
Willy Tarreau74ca5042013-06-11 23:12:07 +020015325req.cook([<name>]) : string
15326cook([<name>]) : string (deprecated)
15327 This extracts the last occurrence of the cookie name <name> on a "Cookie"
15328 header line from the request, and returns its value as string. If no name is
15329 specified, the first cookie value is returned. When used with ACLs, all
15330 matching cookies are evaluated. Spaces around the name and the value are
15331 ignored as requested by the Cookie header specification (RFC6265). The cookie
15332 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
15333 well return an empty value if it is present. Use the "found" match to detect
15334 presence. Use the res.cook() variant for response cookies sent by the server.
15335
15336 ACL derivatives :
15337 cook([<name>]) : exact string match
15338 cook_beg([<name>]) : prefix match
15339 cook_dir([<name>]) : subdir match
15340 cook_dom([<name>]) : domain match
15341 cook_end([<name>]) : suffix match
15342 cook_len([<name>]) : length match
15343 cook_reg([<name>]) : regex match
15344 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010015345
Willy Tarreau74ca5042013-06-11 23:12:07 +020015346req.cook_cnt([<name>]) : integer
15347cook_cnt([<name>]) : integer (deprecated)
15348 Returns an integer value representing the number of occurrences of the cookie
15349 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015350
Willy Tarreau74ca5042013-06-11 23:12:07 +020015351req.cook_val([<name>]) : integer
15352cook_val([<name>]) : integer (deprecated)
15353 This extracts the last occurrence of the cookie name <name> on a "Cookie"
15354 header line from the request, and converts its value to an integer which is
15355 returned. If no name is specified, the first cookie value is returned. When
15356 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020015357
Willy Tarreau74ca5042013-06-11 23:12:07 +020015358cookie([<name>]) : string (deprecated)
15359 This extracts the last occurrence of the cookie name <name> on a "Cookie"
15360 header line from the request, or a "Set-Cookie" header from the response, and
15361 returns its value as a string. A typical use is to get multiple clients
15362 sharing a same profile use the same server. This can be similar to what
Willy Tarreau294d0f02015-08-10 19:40:12 +020015363 "appsession" did with the "request-learn" statement, but with support for
Willy Tarreau74ca5042013-06-11 23:12:07 +020015364 multi-peer synchronization and state keeping across restarts. If no name is
15365 specified, the first cookie value is returned. This fetch should not be used
15366 anymore and should be replaced by req.cook() or res.cook() instead as it
15367 ambiguously uses the direction based on the context where it is used.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015368
Willy Tarreau74ca5042013-06-11 23:12:07 +020015369hdr([<name>[,<occ>]]) : string
15370 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
15371 used on responses. Please refer to these respective fetches for more details.
15372 In case of doubt about the fetch direction, please use the explicit ones.
15373 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015374 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015375
Willy Tarreau74ca5042013-06-11 23:12:07 +020015376req.fhdr(<name>[,<occ>]) : string
15377 This extracts the last occurrence of header <name> in an HTTP request. When
15378 used from an ACL, all occurrences are iterated over until a match is found.
15379 Optionally, a specific occurrence might be specified as a position number.
15380 Positive values indicate a position from the first occurrence, with 1 being
15381 the first one. Negative values indicate positions relative to the last one,
15382 with -1 being the last one. It differs from req.hdr() in that any commas
15383 present in the value are returned and are not used as delimiters. This is
15384 sometimes useful with headers such as User-Agent.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015385
Willy Tarreau74ca5042013-06-11 23:12:07 +020015386req.fhdr_cnt([<name>]) : integer
15387 Returns an integer value representing the number of occurrences of request
15388 header field name <name>, or the total number of header fields if <name> is
15389 not specified. Contrary to its req.hdr_cnt() cousin, this function returns
15390 the number of full line headers and does not stop on commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015391
Willy Tarreau74ca5042013-06-11 23:12:07 +020015392req.hdr([<name>[,<occ>]]) : string
15393 This extracts the last occurrence of header <name> in an HTTP request. When
15394 used from an ACL, all occurrences are iterated over until a match is found.
15395 Optionally, a specific occurrence might be specified as a position number.
15396 Positive values indicate a position from the first occurrence, with 1 being
15397 the first one. Negative values indicate positions relative to the last one,
15398 with -1 being the last one. A typical use is with the X-Forwarded-For header
15399 once converted to IP, associated with an IP stick-table. The function
15400 considers any comma as a delimiter for distinct values. If full-line headers
Lukas Tribus23953682017-04-28 13:24:30 +000015401 are desired instead, use req.fhdr(). Please carefully check RFC7231 to know
Willy Tarreau74ca5042013-06-11 23:12:07 +020015402 how certain headers are supposed to be parsed. Also, some of them are case
Davor Ocelice9ed2812017-12-25 17:49:28 +010015403 insensitive (e.g. Connection).
Willy Tarreaud63335a2010-02-26 12:56:52 +010015404
Willy Tarreau74ca5042013-06-11 23:12:07 +020015405 ACL derivatives :
15406 hdr([<name>[,<occ>]]) : exact string match
15407 hdr_beg([<name>[,<occ>]]) : prefix match
15408 hdr_dir([<name>[,<occ>]]) : subdir match
15409 hdr_dom([<name>[,<occ>]]) : domain match
15410 hdr_end([<name>[,<occ>]]) : suffix match
15411 hdr_len([<name>[,<occ>]]) : length match
15412 hdr_reg([<name>[,<occ>]]) : regex match
15413 hdr_sub([<name>[,<occ>]]) : substring match
15414
15415req.hdr_cnt([<name>]) : integer
15416hdr_cnt([<header>]) : integer (deprecated)
15417 Returns an integer value representing the number of occurrences of request
15418 header field name <name>, or the total number of header field values if
15419 <name> is not specified. It is important to remember that one header line may
15420 count as several headers if it has several values. The function considers any
15421 comma as a delimiter for distinct values. If full-line headers are desired
15422 instead, req.fhdr_cnt() should be used instead. With ACLs, it can be used to
15423 detect presence, absence or abuse of a specific header, as well as to block
15424 request smuggling attacks by rejecting requests which contain more than one
15425 of certain headers. See "req.hdr" for more information on header matching.
15426
15427req.hdr_ip([<name>[,<occ>]]) : ip
15428hdr_ip([<name>[,<occ>]]) : ip (deprecated)
15429 This extracts the last occurrence of header <name> in an HTTP request,
15430 converts it to an IPv4 or IPv6 address and returns this address. When used
15431 with ACLs, all occurrences are checked, and if <name> is omitted, every value
15432 of every header is checked. Optionally, a specific occurrence might be
15433 specified as a position number. Positive values indicate a position from the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015434 first occurrence, with 1 being the first one. Negative values indicate
Willy Tarreau74ca5042013-06-11 23:12:07 +020015435 positions relative to the last one, with -1 being the last one. A typical use
15436 is with the X-Forwarded-For and X-Client-IP headers.
15437
15438req.hdr_val([<name>[,<occ>]]) : integer
15439hdr_val([<name>[,<occ>]]) : integer (deprecated)
15440 This extracts the last occurrence of header <name> in an HTTP request, and
15441 converts it to an integer value. When used with ACLs, all occurrences are
15442 checked, and if <name> is omitted, every value of every header is checked.
15443 Optionally, a specific occurrence might be specified as a position number.
15444 Positive values indicate a position from the first occurrence, with 1 being
15445 the first one. Negative values indicate positions relative to the last one,
15446 with -1 being the last one. A typical use is with the X-Forwarded-For header.
15447
15448http_auth(<userlist>) : boolean
15449 Returns a boolean indicating whether the authentication data received from
15450 the client match a username & password stored in the specified userlist. This
15451 fetch function is not really useful outside of ACLs. Currently only http
15452 basic auth is supported.
15453
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010015454http_auth_group(<userlist>) : string
15455 Returns a string corresponding to the user name found in the authentication
15456 data received from the client if both the user name and password are valid
15457 according to the specified userlist. The main purpose is to use it in ACLs
15458 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015459 This fetch function is not really useful outside of ACLs. Currently only http
15460 basic auth is supported.
15461
15462 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010015463 http_auth_group(<userlist>) : group ...
15464 Returns true when the user extracted from the request and whose password is
15465 valid according to the specified userlist belongs to at least one of the
15466 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015467
15468http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020015469 Returns true when the request being processed is the first one of the
15470 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020015471 from some requests when a request is not the first one, or to help grouping
15472 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020015473
Willy Tarreau74ca5042013-06-11 23:12:07 +020015474method : integer + string
15475 Returns an integer value corresponding to the method in the HTTP request. For
15476 example, "GET" equals 1 (check sources to establish the matching). Value 9
15477 means "other method" and may be converted to a string extracted from the
15478 stream. This should not be used directly as a sample, this is only meant to
15479 be used from ACLs, which transparently convert methods from patterns to these
15480 integer + string values. Some predefined ACL already check for most common
15481 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015482
Willy Tarreau74ca5042013-06-11 23:12:07 +020015483 ACL derivatives :
15484 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020015485
Willy Tarreau74ca5042013-06-11 23:12:07 +020015486 Example :
15487 # only accept GET and HEAD requests
15488 acl valid_method method GET HEAD
15489 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020015490
Willy Tarreau74ca5042013-06-11 23:12:07 +020015491path : string
15492 This extracts the request's URL path, which starts at the first slash and
15493 ends before the question mark (without the host part). A typical use is with
15494 prefetch-capable caches, and with portals which need to aggregate multiple
15495 information from databases and keep them in caches. Note that with outgoing
15496 caches, it would be wiser to use "url" instead. With ACLs, it's typically
Davor Ocelice9ed2812017-12-25 17:49:28 +010015497 used to match exact file names (e.g. "/login.php"), or directory parts using
Willy Tarreau74ca5042013-06-11 23:12:07 +020015498 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015499
Willy Tarreau74ca5042013-06-11 23:12:07 +020015500 ACL derivatives :
15501 path : exact string match
15502 path_beg : prefix match
15503 path_dir : subdir match
15504 path_dom : domain match
15505 path_end : suffix match
15506 path_len : length match
15507 path_reg : regex match
15508 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020015509
Willy Tarreau49ad95c2015-01-19 15:06:26 +010015510query : string
15511 This extracts the request's query string, which starts after the first
15512 question mark. If no question mark is present, this fetch returns nothing. If
15513 a question mark is present but nothing follows, it returns an empty string.
15514 This means it's possible to easily know whether a query string is present
Tim Düsterhus4896c442016-11-29 02:15:19 +010015515 using the "found" matching method. This fetch is the complement of "path"
Willy Tarreau49ad95c2015-01-19 15:06:26 +010015516 which stops before the question mark.
15517
Willy Tarreaueb27ec72015-02-20 13:55:29 +010015518req.hdr_names([<delim>]) : string
15519 This builds a string made from the concatenation of all header names as they
15520 appear in the request when the rule is evaluated. The default delimiter is
15521 the comma (',') but it may be overridden as an optional argument <delim>. In
15522 this case, only the first character of <delim> is considered.
15523
Willy Tarreau74ca5042013-06-11 23:12:07 +020015524req.ver : string
15525req_ver : string (deprecated)
15526 Returns the version string from the HTTP request, for example "1.1". This can
15527 be useful for logs, but is mostly there for ACL. Some predefined ACL already
15528 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015529
Willy Tarreau74ca5042013-06-11 23:12:07 +020015530 ACL derivatives :
15531 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020015532
Willy Tarreau74ca5042013-06-11 23:12:07 +020015533res.comp : boolean
15534 Returns the boolean "true" value if the response has been compressed by
15535 HAProxy, otherwise returns boolean "false". This may be used to add
15536 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015537
Willy Tarreau74ca5042013-06-11 23:12:07 +020015538res.comp_algo : string
15539 Returns a string containing the name of the algorithm used if the response
15540 was compressed by HAProxy, for example : "deflate". This may be used to add
15541 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015542
Willy Tarreau74ca5042013-06-11 23:12:07 +020015543res.cook([<name>]) : string
15544scook([<name>]) : string (deprecated)
15545 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
15546 header line from the response, and returns its value as string. If no name is
15547 specified, the first cookie value is returned.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020015548
Willy Tarreau74ca5042013-06-11 23:12:07 +020015549 ACL derivatives :
15550 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020015551
Willy Tarreau74ca5042013-06-11 23:12:07 +020015552res.cook_cnt([<name>]) : integer
15553scook_cnt([<name>]) : integer (deprecated)
15554 Returns an integer value representing the number of occurrences of the cookie
15555 <name> in the response, or all cookies if <name> is not specified. This is
15556 mostly useful when combined with ACLs to detect suspicious responses.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015557
Willy Tarreau74ca5042013-06-11 23:12:07 +020015558res.cook_val([<name>]) : integer
15559scook_val([<name>]) : integer (deprecated)
15560 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
15561 header line from the response, and converts its value to an integer which is
15562 returned. If no name is specified, the first cookie value is returned.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015563
Willy Tarreau74ca5042013-06-11 23:12:07 +020015564res.fhdr([<name>[,<occ>]]) : string
15565 This extracts the last occurrence of header <name> in an HTTP response, or of
15566 the last header if no <name> is specified. Optionally, a specific occurrence
15567 might be specified as a position number. Positive values indicate a position
15568 from the first occurrence, with 1 being the first one. Negative values
15569 indicate positions relative to the last one, with -1 being the last one. It
15570 differs from res.hdr() in that any commas present in the value are returned
15571 and are not used as delimiters. If this is not desired, the res.hdr() fetch
15572 should be used instead. This is sometimes useful with headers such as Date or
15573 Expires.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015574
Willy Tarreau74ca5042013-06-11 23:12:07 +020015575res.fhdr_cnt([<name>]) : integer
15576 Returns an integer value representing the number of occurrences of response
15577 header field name <name>, or the total number of header fields if <name> is
15578 not specified. Contrary to its res.hdr_cnt() cousin, this function returns
15579 the number of full line headers and does not stop on commas. If this is not
15580 desired, the res.hdr_cnt() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015581
Willy Tarreau74ca5042013-06-11 23:12:07 +020015582res.hdr([<name>[,<occ>]]) : string
15583shdr([<name>[,<occ>]]) : string (deprecated)
15584 This extracts the last occurrence of header <name> in an HTTP response, or of
15585 the last header if no <name> is specified. Optionally, a specific occurrence
15586 might be specified as a position number. Positive values indicate a position
15587 from the first occurrence, with 1 being the first one. Negative values
15588 indicate positions relative to the last one, with -1 being the last one. This
15589 can be useful to learn some data into a stick-table. The function considers
15590 any comma as a delimiter for distinct values. If this is not desired, the
15591 res.fhdr() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015592
Willy Tarreau74ca5042013-06-11 23:12:07 +020015593 ACL derivatives :
15594 shdr([<name>[,<occ>]]) : exact string match
15595 shdr_beg([<name>[,<occ>]]) : prefix match
15596 shdr_dir([<name>[,<occ>]]) : subdir match
15597 shdr_dom([<name>[,<occ>]]) : domain match
15598 shdr_end([<name>[,<occ>]]) : suffix match
15599 shdr_len([<name>[,<occ>]]) : length match
15600 shdr_reg([<name>[,<occ>]]) : regex match
15601 shdr_sub([<name>[,<occ>]]) : substring match
15602
15603res.hdr_cnt([<name>]) : integer
15604shdr_cnt([<name>]) : integer (deprecated)
15605 Returns an integer value representing the number of occurrences of response
15606 header field name <name>, or the total number of header fields if <name> is
15607 not specified. The function considers any comma as a delimiter for distinct
15608 values. If this is not desired, the res.fhdr_cnt() fetch should be used
15609 instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015610
Willy Tarreau74ca5042013-06-11 23:12:07 +020015611res.hdr_ip([<name>[,<occ>]]) : ip
15612shdr_ip([<name>[,<occ>]]) : ip (deprecated)
15613 This extracts the last occurrence of header <name> in an HTTP response,
15614 convert it to an IPv4 or IPv6 address and returns this address. Optionally, a
15615 specific occurrence might be specified as a position number. Positive values
15616 indicate a position from the first occurrence, with 1 being the first one.
15617 Negative values indicate positions relative to the last one, with -1 being
15618 the last one. This can be useful to learn some data into a stick table.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015619
Willy Tarreaueb27ec72015-02-20 13:55:29 +010015620res.hdr_names([<delim>]) : string
15621 This builds a string made from the concatenation of all header names as they
15622 appear in the response when the rule is evaluated. The default delimiter is
15623 the comma (',') but it may be overridden as an optional argument <delim>. In
15624 this case, only the first character of <delim> is considered.
15625
Willy Tarreau74ca5042013-06-11 23:12:07 +020015626res.hdr_val([<name>[,<occ>]]) : integer
15627shdr_val([<name>[,<occ>]]) : integer (deprecated)
15628 This extracts the last occurrence of header <name> in an HTTP response, and
15629 converts it to an integer value. Optionally, a specific occurrence might be
15630 specified as a position number. Positive values indicate a position from the
15631 first occurrence, with 1 being the first one. Negative values indicate
15632 positions relative to the last one, with -1 being the last one. This can be
15633 useful to learn some data into a stick table.
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010015634
Willy Tarreau74ca5042013-06-11 23:12:07 +020015635res.ver : string
15636resp_ver : string (deprecated)
15637 Returns the version string from the HTTP response, for example "1.1". This
15638 can be useful for logs, but is mostly there for ACL.
Willy Tarreau0e698542011-09-16 08:32:32 +020015639
Willy Tarreau74ca5042013-06-11 23:12:07 +020015640 ACL derivatives :
15641 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010015642
Willy Tarreau74ca5042013-06-11 23:12:07 +020015643set-cookie([<name>]) : string (deprecated)
15644 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
15645 header line from the response and uses the corresponding value to match. This
Willy Tarreau294d0f02015-08-10 19:40:12 +020015646 can be comparable to what "appsession" did with default options, but with
Willy Tarreau74ca5042013-06-11 23:12:07 +020015647 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010015648
Willy Tarreau74ca5042013-06-11 23:12:07 +020015649 This fetch function is deprecated and has been superseded by the "res.cook"
15650 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010015651
Willy Tarreau74ca5042013-06-11 23:12:07 +020015652status : integer
15653 Returns an integer containing the HTTP status code in the HTTP response, for
15654 example, 302. It is mostly used within ACLs and integer ranges, for example,
15655 to remove any Location header if the response is not a 3xx.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015656
Thierry Fournier0e00dca2016-04-07 15:47:40 +020015657unique-id : string
15658 Returns the unique-id attached to the request. The directive
15659 "unique-id-format" must be set. If it is not set, the unique-id sample fetch
15660 fails. Note that the unique-id is usually used with HTTP requests, however this
15661 sample fetch can be used with other protocols. Obviously, if it is used with
15662 other protocols than HTTP, the unique-id-format directive must not contain
15663 HTTP parts. See: unique-id-format and unique-id-header
15664
Willy Tarreau74ca5042013-06-11 23:12:07 +020015665url : string
15666 This extracts the request's URL as presented in the request. A typical use is
15667 with prefetch-capable caches, and with portals which need to aggregate
15668 multiple information from databases and keep them in caches. With ACLs, using
15669 "path" is preferred over using "url", because clients may send a full URL as
15670 is normally done with proxies. The only real use is to match "*" which does
15671 not match in "path", and for which there is already a predefined ACL. See
15672 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015673
Willy Tarreau74ca5042013-06-11 23:12:07 +020015674 ACL derivatives :
15675 url : exact string match
15676 url_beg : prefix match
15677 url_dir : subdir match
15678 url_dom : domain match
15679 url_end : suffix match
15680 url_len : length match
15681 url_reg : regex match
15682 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015683
Willy Tarreau74ca5042013-06-11 23:12:07 +020015684url_ip : ip
15685 This extracts the IP address from the request's URL when the host part is
15686 presented as an IP address. Its use is very limited. For instance, a
15687 monitoring system might use this field as an alternative for the source IP in
15688 order to test what path a given source address would follow, or to force an
15689 entry in a table for a given source address. With ACLs it can be used to
15690 restrict access to certain systems through a proxy, for example when combined
15691 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015692
Willy Tarreau74ca5042013-06-11 23:12:07 +020015693url_port : integer
15694 This extracts the port part from the request's URL. Note that if the port is
15695 not specified in the request, port 80 is assumed. With ACLs it can be used to
15696 restrict access to certain systems through a proxy, for example when combined
15697 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015698
Willy Tarreau1ede1da2015-05-07 16:06:18 +020015699urlp([<name>[,<delim>]]) : string
15700url_param([<name>[,<delim>]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020015701 This extracts the first occurrence of the parameter <name> in the query
15702 string, which begins after either '?' or <delim>, and which ends before '&',
Willy Tarreau1ede1da2015-05-07 16:06:18 +020015703 ';' or <delim>. The parameter name is case-sensitive. If no name is given,
15704 any parameter will match, and the first one will be returned. The result is
15705 a string corresponding to the value of the parameter <name> as presented in
15706 the request (no URL decoding is performed). This can be used for session
Willy Tarreau74ca5042013-06-11 23:12:07 +020015707 stickiness based on a client ID, to extract an application cookie passed as a
15708 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
Willy Tarreau1ede1da2015-05-07 16:06:18 +020015709 this fetch iterates over multiple parameters and will iteratively report all
15710 parameters values if no name is given
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015711
Willy Tarreau74ca5042013-06-11 23:12:07 +020015712 ACL derivatives :
15713 urlp(<name>[,<delim>]) : exact string match
15714 urlp_beg(<name>[,<delim>]) : prefix match
15715 urlp_dir(<name>[,<delim>]) : subdir match
15716 urlp_dom(<name>[,<delim>]) : domain match
15717 urlp_end(<name>[,<delim>]) : suffix match
15718 urlp_len(<name>[,<delim>]) : length match
15719 urlp_reg(<name>[,<delim>]) : regex match
15720 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015721
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015722
Willy Tarreau74ca5042013-06-11 23:12:07 +020015723 Example :
15724 # match http://example.com/foo?PHPSESSIONID=some_id
15725 stick on urlp(PHPSESSIONID)
15726 # match http://example.com/foo;JSESSIONID=some_id
15727 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020015728
Jarno Huuskonen676f6222017-03-30 09:19:45 +030015729urlp_val([<name>[,<delim>]]) : integer
Willy Tarreau74ca5042013-06-11 23:12:07 +020015730 See "urlp" above. This one extracts the URL parameter <name> in the request
15731 and converts it to an integer value. This can be used for session stickiness
15732 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020015733
Dragan Dosen0070cd52016-06-16 12:19:49 +020015734url32 : integer
15735 This returns a 32-bit hash of the value obtained by concatenating the first
15736 Host header and the whole URL including parameters (not only the path part of
15737 the request, as in the "base32" fetch above). This is useful to track per-URL
15738 activity. A shorter hash is stored, saving a lot of memory. The output type
15739 is an unsigned integer.
15740
15741url32+src : binary
15742 This returns the concatenation of the "url32" fetch and the "src" fetch. The
15743 resulting type is of type binary, with a size of 8 or 20 bytes depending on
15744 the source address family. This can be used to track per-IP, per-URL counters.
15745
Willy Tarreau198a7442008-01-17 12:05:32 +010015746
Willy Tarreau74ca5042013-06-11 23:12:07 +0200157477.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015748---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010015749
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015750Some predefined ACLs are hard-coded so that they do not have to be declared in
15751every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020015752order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010015753
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015754ACL name Equivalent to Usage
15755---------------+-----------------------------+---------------------------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015756FALSE always_false never match
Willy Tarreau2492d5b2009-07-11 00:06:00 +020015757HTTP req_proto_http match if protocol is valid HTTP
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015758HTTP_1.0 req_ver 1.0 match HTTP version 1.0
15759HTTP_1.1 req_ver 1.1 match HTTP version 1.1
Willy Tarreaud63335a2010-02-26 12:56:52 +010015760HTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length
15761HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
15762HTTP_URL_SLASH url_beg / match URL beginning with "/"
15763HTTP_URL_STAR url * match URL equal to "*"
15764LOCALHOST src 127.0.0.1/8 match connection from local host
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015765METH_CONNECT method CONNECT match HTTP CONNECT method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020015766METH_DELETE method DELETE match HTTP DELETE method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015767METH_GET method GET HEAD match HTTP GET or HEAD method
15768METH_HEAD method HEAD match HTTP HEAD method
15769METH_OPTIONS method OPTIONS match HTTP OPTIONS method
15770METH_POST method POST match HTTP POST method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020015771METH_PUT method PUT match HTTP PUT method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015772METH_TRACE method TRACE match HTTP TRACE method
Emeric Brunbede3d02009-06-30 17:54:00 +020015773RDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015774REQ_CONTENT req_len gt 0 match data in the request buffer
Willy Tarreaud63335a2010-02-26 12:56:52 +010015775TRUE always_true always match
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015776WAIT_END wait_end wait for end of content analysis
15777---------------+-----------------------------+---------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010015778
Willy Tarreaub937b7e2010-01-12 15:27:54 +010015779
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200157808. Logging
15781----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010015782
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015783One of HAProxy's strong points certainly lies is its precise logs. It probably
15784provides the finest level of information available for such a product, which is
15785very important for troubleshooting complex environments. Standard information
15786provided in logs include client ports, TCP/HTTP state timers, precise session
15787state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010015788to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015789headers.
15790
15791In order to improve administrators reactivity, it offers a great transparency
15792about encountered problems, both internal and external, and it is possible to
15793send logs to different sources at the same time with different level filters :
15794
15795 - global process-level logs (system errors, start/stop, etc..)
15796 - per-instance system and internal errors (lack of resource, bugs, ...)
15797 - per-instance external troubles (servers up/down, max connections)
15798 - per-instance activity (client connections), either at the establishment or
15799 at the termination.
Davor Ocelice9ed2812017-12-25 17:49:28 +010015800 - per-request control of log-level, e.g.
Jim Freeman9e8714b2015-05-26 09:16:34 -060015801 http-request set-log-level silent if sensitive_request
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015802
15803The ability to distribute different levels of logs to different log servers
15804allow several production teams to interact and to fix their problems as soon
15805as possible. For example, the system team might monitor system-wide errors,
15806while the application team might be monitoring the up/down for their servers in
15807real time, and the security team might analyze the activity logs with one hour
15808delay.
15809
15810
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200158118.1. Log levels
15812---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015813
Simon Hormandf791f52011-05-29 15:01:10 +090015814TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015815source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090015816HTTP request, HTTP return code, number of bytes transmitted, conditions
15817in which the session ended, and even exchanged cookies values. For example
15818track a particular user's problems. All messages may be sent to up to two
15819syslog servers. Check the "log" keyword in section 4.2 for more information
15820about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015821
15822
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200158238.2. Log formats
15824----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015825
William Lallemand48940402012-01-30 16:47:22 +010015826HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090015827and will be detailed in the following sections. A few of them may vary
15828slightly with the configuration, due to indicators specific to certain
15829options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015830
15831 - the default format, which is very basic and very rarely used. It only
15832 provides very basic information about the incoming connection at the moment
15833 it is accepted : source IP:port, destination IP:port, and frontend-name.
15834 This mode will eventually disappear so it will not be described to great
15835 extents.
15836
15837 - the TCP format, which is more advanced. This format is enabled when "option
15838 tcplog" is set on the frontend. HAProxy will then usually wait for the
15839 connection to terminate before logging. This format provides much richer
15840 information, such as timers, connection counts, queue size, etc... This
15841 format is recommended for pure TCP proxies.
15842
15843 - the HTTP format, which is the most advanced for HTTP proxying. This format
15844 is enabled when "option httplog" is set on the frontend. It provides the
15845 same information as the TCP format with some HTTP-specific fields such as
15846 the request, the status code, and captures of headers and cookies. This
15847 format is recommended for HTTP proxies.
15848
Emeric Brun3a058f32009-06-30 18:26:00 +020015849 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
15850 fields arranged in the same order as the CLF format. In this mode, all
15851 timers, captures, flags, etc... appear one per field after the end of the
15852 common fields, in the same order they appear in the standard HTTP format.
15853
William Lallemand48940402012-01-30 16:47:22 +010015854 - the custom log format, allows you to make your own log line.
15855
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015856Next sections will go deeper into details for each of these formats. Format
15857specification will be performed on a "field" basis. Unless stated otherwise, a
15858field is a portion of text delimited by any number of spaces. Since syslog
15859servers are susceptible of inserting fields at the beginning of a line, it is
15860always assumed that the first field is the one containing the process name and
15861identifier.
15862
15863Note : Since log lines may be quite long, the log examples in sections below
15864 might be broken into multiple lines. The example log lines will be
15865 prefixed with 3 closing angle brackets ('>>>') and each time a log is
15866 broken into multiple lines, each non-final line will end with a
15867 backslash ('\') and the next line will start indented by two characters.
15868
15869
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200158708.2.1. Default log format
15871-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015872
15873This format is used when no specific option is set. The log is emitted as soon
15874as the connection is accepted. One should note that this currently is the only
15875format which logs the request's destination IP and ports.
15876
15877 Example :
15878 listen www
15879 mode http
15880 log global
15881 server srv1 127.0.0.1:8000
15882
15883 >>> Feb 6 12:12:09 localhost \
15884 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
15885 (www/HTTP)
15886
15887 Field Format Extract from the example above
15888 1 process_name '[' pid ']:' haproxy[14385]:
15889 2 'Connect from' Connect from
15890 3 source_ip ':' source_port 10.0.1.2:33312
15891 4 'to' to
15892 5 destination_ip ':' destination_port 10.0.3.31:8012
15893 6 '(' frontend_name '/' mode ')' (www/HTTP)
15894
15895Detailed fields description :
15896 - "source_ip" is the IP address of the client which initiated the connection.
15897 - "source_port" is the TCP port of the client which initiated the connection.
15898 - "destination_ip" is the IP address the client connected to.
15899 - "destination_port" is the TCP port the client connected to.
15900 - "frontend_name" is the name of the frontend (or listener) which received
15901 and processed the connection.
15902 - "mode is the mode the frontend is operating (TCP or HTTP).
15903
Willy Tarreauceb24bc2010-11-09 12:46:41 +010015904In case of a UNIX socket, the source and destination addresses are marked as
15905"unix:" and the ports reflect the internal ID of the socket which accepted the
15906connection (the same ID as reported in the stats).
15907
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015908It is advised not to use this deprecated format for newer installations as it
15909will eventually disappear.
15910
15911
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200159128.2.2. TCP log format
15913---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015914
15915The TCP format is used when "option tcplog" is specified in the frontend, and
15916is the recommended format for pure TCP proxies. It provides a lot of precious
15917information for troubleshooting. Since this format includes timers and byte
15918counts, the log is normally emitted at the end of the session. It can be
15919emitted earlier if "option logasap" is specified, which makes sense in most
15920environments with long sessions such as remote terminals. Sessions which match
15921the "monitor" rules are never logged. It is also possible not to emit logs for
15922sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020015923specifying "option dontlognull" in the frontend. Successful connections will
15924not be logged if "option dontlog-normal" is specified in the frontend. A few
15925fields may slightly vary depending on some configuration options, those are
15926marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015927
15928 Example :
15929 frontend fnt
15930 mode tcp
15931 option tcplog
15932 log global
15933 default_backend bck
15934
15935 backend bck
15936 server srv1 127.0.0.1:8000
15937
15938 >>> Feb 6 12:12:56 localhost \
15939 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
15940 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
15941
15942 Field Format Extract from the example above
15943 1 process_name '[' pid ']:' haproxy[14387]:
15944 2 client_ip ':' client_port 10.0.1.2:33313
15945 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
15946 4 frontend_name fnt
15947 5 backend_name '/' server_name bck/srv1
15948 6 Tw '/' Tc '/' Tt* 0/0/5007
15949 7 bytes_read* 212
15950 8 termination_state --
15951 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
15952 10 srv_queue '/' backend_queue 0/0
15953
15954Detailed fields description :
15955 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010015956 connection to haproxy. If the connection was accepted on a UNIX socket
15957 instead, the IP address would be replaced with the word "unix". Note that
15958 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010015959 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010015960 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010015961 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015962
15963 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010015964 If the connection was accepted on a UNIX socket instead, the port would be
15965 replaced with the ID of the accepting socket, which is also reported in the
15966 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015967
15968 - "accept_date" is the exact date when the connection was received by haproxy
15969 (which might be very slightly different from the date observed on the
15970 network if there was some queuing in the system's backlog). This is usually
Willy Tarreau590a0512018-09-05 11:56:48 +020015971 the same date which may appear in any upstream firewall's log. When used in
15972 HTTP mode, the accept_date field will be reset to the first moment the
15973 connection is ready to receive a new request (end of previous response for
15974 HTTP/1, immediately after previous request for HTTP/2).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015975
15976 - "frontend_name" is the name of the frontend (or listener) which received
15977 and processed the connection.
15978
15979 - "backend_name" is the name of the backend (or listener) which was selected
15980 to manage the connection to the server. This will be the same as the
15981 frontend if no switching rule has been applied, which is common for TCP
15982 applications.
15983
15984 - "server_name" is the name of the last server to which the connection was
15985 sent, which might differ from the first one if there were connection errors
15986 and a redispatch occurred. Note that this server belongs to the backend
15987 which processed the request. If the connection was aborted before reaching
15988 a server, "<NOSRV>" is indicated instead of a server name.
15989
15990 - "Tw" is the total time in milliseconds spent waiting in the various queues.
15991 It can be "-1" if the connection was aborted before reaching the queue.
15992 See "Timers" below for more details.
15993
15994 - "Tc" is the total time in milliseconds spent waiting for the connection to
15995 establish to the final server, including retries. It can be "-1" if the
15996 connection was aborted before a connection could be established. See
15997 "Timers" below for more details.
15998
15999 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016000 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016001 "option logasap" was specified, then the time counting stops at the moment
16002 the log is emitted. In this case, a '+' sign is prepended before the value,
16003 indicating that the final one will be larger. See "Timers" below for more
16004 details.
16005
16006 - "bytes_read" is the total number of bytes transmitted from the server to
16007 the client when the log is emitted. If "option logasap" is specified, the
16008 this value will be prefixed with a '+' sign indicating that the final one
16009 may be larger. Please note that this value is a 64-bit counter, so log
16010 analysis tools must be able to handle it without overflowing.
16011
16012 - "termination_state" is the condition the session was in when the session
16013 ended. This indicates the session state, which side caused the end of
16014 session to happen, and for what reason (timeout, error, ...). The normal
16015 flags should be "--", indicating the session was closed by either end with
16016 no data remaining in buffers. See below "Session state at disconnection"
16017 for more details.
16018
16019 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040016020 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016021 limits have been reached. For instance, if actconn is close to 512 when
16022 multiple connection errors occur, chances are high that the system limits
16023 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016024 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016025
16026 - "feconn" is the total number of concurrent connections on the frontend when
16027 the session was logged. It is useful to estimate the amount of resource
16028 required to sustain high loads, and to detect when the frontend's "maxconn"
16029 has been reached. Most often when this value increases by huge jumps, it is
16030 because there is congestion on the backend servers, but sometimes it can be
16031 caused by a denial of service attack.
16032
16033 - "beconn" is the total number of concurrent connections handled by the
16034 backend when the session was logged. It includes the total number of
16035 concurrent connections active on servers as well as the number of
16036 connections pending in queues. It is useful to estimate the amount of
16037 additional servers needed to support high loads for a given application.
16038 Most often when this value increases by huge jumps, it is because there is
16039 congestion on the backend servers, but sometimes it can be caused by a
16040 denial of service attack.
16041
16042 - "srv_conn" is the total number of concurrent connections still active on
16043 the server when the session was logged. It can never exceed the server's
16044 configured "maxconn" parameter. If this value is very often close or equal
16045 to the server's "maxconn", it means that traffic regulation is involved a
16046 lot, meaning that either the server's maxconn value is too low, or that
16047 there aren't enough servers to process the load with an optimal response
16048 time. When only one of the server's "srv_conn" is high, it usually means
16049 that this server has some trouble causing the connections to take longer to
16050 be processed than on other servers.
16051
16052 - "retries" is the number of connection retries experienced by this session
16053 when trying to connect to the server. It must normally be zero, unless a
16054 server is being stopped at the same moment the connection was attempted.
16055 Frequent retries generally indicate either a network problem between
16056 haproxy and the server, or a misconfigured system backlog on the server
16057 preventing new connections from being queued. This field may optionally be
16058 prefixed with a '+' sign, indicating that the session has experienced a
16059 redispatch after the maximal retry count has been reached on the initial
16060 server. In this case, the server name appearing in the log is the one the
16061 connection was redispatched to, and not the first one, though both may
16062 sometimes be the same in case of hashing for instance. So as a general rule
16063 of thumb, when a '+' is present in front of the retry count, this count
16064 should not be attributed to the logged server.
16065
16066 - "srv_queue" is the total number of requests which were processed before
16067 this one in the server queue. It is zero when the request has not gone
16068 through the server queue. It makes it possible to estimate the approximate
16069 server's response time by dividing the time spent in queue by the number of
16070 requests in the queue. It is worth noting that if a session experiences a
16071 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016072 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016073 backend queue unless a redispatch occurs.
16074
16075 - "backend_queue" is the total number of requests which were processed before
16076 this one in the backend's global queue. It is zero when the request has not
16077 gone through the global queue. It makes it possible to estimate the average
16078 queue length, which easily translates into a number of missing servers when
16079 divided by a server's "maxconn" parameter. It is worth noting that if a
16080 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010016081 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016082 through both the server queue and the backend queue unless a redispatch
16083 occurs.
16084
16085
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200160868.2.3. HTTP log format
16087----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016088
16089The HTTP format is the most complete and the best suited for HTTP proxies. It
16090is enabled by when "option httplog" is specified in the frontend. It provides
16091the same level of information as the TCP format with additional features which
16092are specific to the HTTP protocol. Just like the TCP format, the log is usually
16093emitted at the end of the session, unless "option logasap" is specified, which
16094generally only makes sense for download sites. A session which matches the
16095"monitor" rules will never logged. It is also possible not to log sessions for
16096which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020016097frontend. Successful connections will not be logged if "option dontlog-normal"
16098is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016099
16100Most fields are shared with the TCP log, some being different. A few fields may
16101slightly vary depending on some configuration options. Those ones are marked
16102with a star ('*') after the field name below.
16103
16104 Example :
16105 frontend http-in
16106 mode http
16107 option httplog
16108 log global
16109 default_backend bck
16110
16111 backend static
16112 server srv1 127.0.0.1:8000
16113
16114 >>> Feb 6 12:14:14 localhost \
16115 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
16116 static/srv1 10/0/30/69/109 200 2750 - - ---- 1/1/1/1/0 0/0 {1wt.eu} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010016117 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016118
16119 Field Format Extract from the example above
16120 1 process_name '[' pid ']:' haproxy[14389]:
16121 2 client_ip ':' client_port 10.0.1.2:33317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016122 3 '[' request_date ']' [06/Feb/2009:12:14:14.655]
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016123 4 frontend_name http-in
16124 5 backend_name '/' server_name static/srv1
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016125 6 TR '/' Tw '/' Tc '/' Tr '/' Ta* 10/0/30/69/109
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016126 7 status_code 200
16127 8 bytes_read* 2750
16128 9 captured_request_cookie -
16129 10 captured_response_cookie -
16130 11 termination_state ----
16131 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
16132 13 srv_queue '/' backend_queue 0/0
16133 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
16134 15 '{' captured_response_headers* '}' {}
16135 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010016136
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016137Detailed fields description :
16138 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016139 connection to haproxy. If the connection was accepted on a UNIX socket
16140 instead, the IP address would be replaced with the word "unix". Note that
16141 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016142 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010016143 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016144 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016145
16146 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016147 If the connection was accepted on a UNIX socket instead, the port would be
16148 replaced with the ID of the accepting socket, which is also reported in the
16149 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016150
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016151 - "request_date" is the exact date when the first byte of the HTTP request
16152 was received by haproxy (log field %tr).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016153
16154 - "frontend_name" is the name of the frontend (or listener) which received
16155 and processed the connection.
16156
16157 - "backend_name" is the name of the backend (or listener) which was selected
16158 to manage the connection to the server. This will be the same as the
16159 frontend if no switching rule has been applied.
16160
16161 - "server_name" is the name of the last server to which the connection was
16162 sent, which might differ from the first one if there were connection errors
16163 and a redispatch occurred. Note that this server belongs to the backend
16164 which processed the request. If the request was aborted before reaching a
16165 server, "<NOSRV>" is indicated instead of a server name. If the request was
16166 intercepted by the stats subsystem, "<STATS>" is indicated instead.
16167
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016168 - "TR" is the total time in milliseconds spent waiting for a full HTTP
16169 request from the client (not counting body) after the first byte was
16170 received. It can be "-1" if the connection was aborted before a complete
16171 request could be received or the a bad request was received. It should
16172 always be very small because a request generally fits in one single packet.
16173 Large times here generally indicate network issues between the client and
Willy Tarreau590a0512018-09-05 11:56:48 +020016174 haproxy or requests being typed by hand. See section 8.4 "Timing Events"
16175 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016176
16177 - "Tw" is the total time in milliseconds spent waiting in the various queues.
16178 It can be "-1" if the connection was aborted before reaching the queue.
Willy Tarreau590a0512018-09-05 11:56:48 +020016179 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016180
16181 - "Tc" is the total time in milliseconds spent waiting for the connection to
16182 establish to the final server, including retries. It can be "-1" if the
Willy Tarreau590a0512018-09-05 11:56:48 +020016183 request was aborted before a connection could be established. See section
16184 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016185
16186 - "Tr" is the total time in milliseconds spent waiting for the server to send
16187 a full HTTP response, not counting data. It can be "-1" if the request was
16188 aborted before a complete response could be received. It generally matches
16189 the server's processing time for the request, though it may be altered by
16190 the amount of data sent by the client to the server. Large times here on
Willy Tarreau590a0512018-09-05 11:56:48 +020016191 "GET" requests generally indicate an overloaded server. See section 8.4
16192 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016193
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016194 - "Ta" is the time the request remained active in haproxy, which is the total
16195 time in milliseconds elapsed between the first byte of the request was
16196 received and the last byte of response was sent. It covers all possible
16197 processing except the handshake (see Th) and idle time (see Ti). There is
16198 one exception, if "option logasap" was specified, then the time counting
16199 stops at the moment the log is emitted. In this case, a '+' sign is
16200 prepended before the value, indicating that the final one will be larger.
Willy Tarreau590a0512018-09-05 11:56:48 +020016201 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016202
16203 - "status_code" is the HTTP status code returned to the client. This status
16204 is generally set by the server, but it might also be set by haproxy when
16205 the server cannot be reached or when its response is blocked by haproxy.
16206
16207 - "bytes_read" is the total number of bytes transmitted to the client when
16208 the log is emitted. This does include HTTP headers. If "option logasap" is
16209 specified, the this value will be prefixed with a '+' sign indicating that
16210 the final one may be larger. Please note that this value is a 64-bit
16211 counter, so log analysis tools must be able to handle it without
16212 overflowing.
16213
16214 - "captured_request_cookie" is an optional "name=value" entry indicating that
16215 the client had this cookie in the request. The cookie name and its maximum
16216 length are defined by the "capture cookie" statement in the frontend
16217 configuration. The field is a single dash ('-') when the option is not
16218 set. Only one cookie may be captured, it is generally used to track session
16219 ID exchanges between a client and a server to detect session crossing
16220 between clients due to application bugs. For more details, please consult
16221 the section "Capturing HTTP headers and cookies" below.
16222
16223 - "captured_response_cookie" is an optional "name=value" entry indicating
16224 that the server has returned a cookie with its response. The cookie name
16225 and its maximum length are defined by the "capture cookie" statement in the
16226 frontend configuration. The field is a single dash ('-') when the option is
16227 not set. Only one cookie may be captured, it is generally used to track
16228 session ID exchanges between a client and a server to detect session
16229 crossing between clients due to application bugs. For more details, please
16230 consult the section "Capturing HTTP headers and cookies" below.
16231
16232 - "termination_state" is the condition the session was in when the session
16233 ended. This indicates the session state, which side caused the end of
16234 session to happen, for what reason (timeout, error, ...), just like in TCP
16235 logs, and information about persistence operations on cookies in the last
16236 two characters. The normal flags should begin with "--", indicating the
16237 session was closed by either end with no data remaining in buffers. See
16238 below "Session state at disconnection" for more details.
16239
16240 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040016241 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016242 limits have been reached. For instance, if actconn is close to 512 or 1024
16243 when multiple connection errors occur, chances are high that the system
16244 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016245 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016246 system.
16247
16248 - "feconn" is the total number of concurrent connections on the frontend when
16249 the session was logged. It is useful to estimate the amount of resource
16250 required to sustain high loads, and to detect when the frontend's "maxconn"
16251 has been reached. Most often when this value increases by huge jumps, it is
16252 because there is congestion on the backend servers, but sometimes it can be
16253 caused by a denial of service attack.
16254
16255 - "beconn" is the total number of concurrent connections handled by the
16256 backend when the session was logged. It includes the total number of
16257 concurrent connections active on servers as well as the number of
16258 connections pending in queues. It is useful to estimate the amount of
16259 additional servers needed to support high loads for a given application.
16260 Most often when this value increases by huge jumps, it is because there is
16261 congestion on the backend servers, but sometimes it can be caused by a
16262 denial of service attack.
16263
16264 - "srv_conn" is the total number of concurrent connections still active on
16265 the server when the session was logged. It can never exceed the server's
16266 configured "maxconn" parameter. If this value is very often close or equal
16267 to the server's "maxconn", it means that traffic regulation is involved a
16268 lot, meaning that either the server's maxconn value is too low, or that
16269 there aren't enough servers to process the load with an optimal response
16270 time. When only one of the server's "srv_conn" is high, it usually means
16271 that this server has some trouble causing the requests to take longer to be
16272 processed than on other servers.
16273
16274 - "retries" is the number of connection retries experienced by this session
16275 when trying to connect to the server. It must normally be zero, unless a
16276 server is being stopped at the same moment the connection was attempted.
16277 Frequent retries generally indicate either a network problem between
16278 haproxy and the server, or a misconfigured system backlog on the server
16279 preventing new connections from being queued. This field may optionally be
16280 prefixed with a '+' sign, indicating that the session has experienced a
16281 redispatch after the maximal retry count has been reached on the initial
16282 server. In this case, the server name appearing in the log is the one the
16283 connection was redispatched to, and not the first one, though both may
16284 sometimes be the same in case of hashing for instance. So as a general rule
16285 of thumb, when a '+' is present in front of the retry count, this count
16286 should not be attributed to the logged server.
16287
16288 - "srv_queue" is the total number of requests which were processed before
16289 this one in the server queue. It is zero when the request has not gone
16290 through the server queue. It makes it possible to estimate the approximate
16291 server's response time by dividing the time spent in queue by the number of
16292 requests in the queue. It is worth noting that if a session experiences a
16293 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016294 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016295 backend queue unless a redispatch occurs.
16296
16297 - "backend_queue" is the total number of requests which were processed before
16298 this one in the backend's global queue. It is zero when the request has not
16299 gone through the global queue. It makes it possible to estimate the average
16300 queue length, which easily translates into a number of missing servers when
16301 divided by a server's "maxconn" parameter. It is worth noting that if a
16302 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010016303 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016304 through both the server queue and the backend queue unless a redispatch
16305 occurs.
16306
16307 - "captured_request_headers" is a list of headers captured in the request due
16308 to the presence of the "capture request header" statement in the frontend.
16309 Multiple headers can be captured, they will be delimited by a vertical bar
16310 ('|'). When no capture is enabled, the braces do not appear, causing a
16311 shift of remaining fields. It is important to note that this field may
16312 contain spaces, and that using it requires a smarter log parser than when
16313 it's not used. Please consult the section "Capturing HTTP headers and
16314 cookies" below for more details.
16315
16316 - "captured_response_headers" is a list of headers captured in the response
16317 due to the presence of the "capture response header" statement in the
16318 frontend. Multiple headers can be captured, they will be delimited by a
16319 vertical bar ('|'). When no capture is enabled, the braces do not appear,
16320 causing a shift of remaining fields. It is important to note that this
16321 field may contain spaces, and that using it requires a smarter log parser
16322 than when it's not used. Please consult the section "Capturing HTTP headers
16323 and cookies" below for more details.
16324
16325 - "http_request" is the complete HTTP request line, including the method,
16326 request and HTTP version string. Non-printable characters are encoded (see
16327 below the section "Non-printable characters"). This is always the last
16328 field, and it is always delimited by quotes and is the only one which can
16329 contain quotes. If new fields are added to the log format, they will be
16330 added before this field. This field might be truncated if the request is
16331 huge and does not fit in the standard syslog buffer (1024 characters). This
16332 is the reason why this field must always remain the last one.
16333
16334
Cyril Bontédc4d9032012-04-08 21:57:39 +0200163358.2.4. Custom log format
16336------------------------
William Lallemand48940402012-01-30 16:47:22 +010016337
Willy Tarreau2beef582012-12-20 17:22:52 +010016338The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010016339mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010016340
Davor Ocelice9ed2812017-12-25 17:49:28 +010016341HAProxy understands some log format variables. % precedes log format variables.
William Lallemand48940402012-01-30 16:47:22 +010016342Variables can take arguments using braces ('{}'), and multiple arguments are
16343separated by commas within the braces. Flags may be added or removed by
16344prefixing them with a '+' or '-' sign.
16345
16346Special variable "%o" may be used to propagate its flags to all other
16347variables on the same format string. This is particularly handy with quoted
Dragan Dosen835b9212016-02-12 13:23:03 +010016348("Q") and escaped ("E") string formats.
William Lallemand48940402012-01-30 16:47:22 +010016349
Willy Tarreauc8368452012-12-21 00:09:23 +010016350If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020016351as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010016352less common information such as the client's SSL certificate's DN, or to log
16353the key that would be used to store an entry into a stick table.
16354
William Lallemand48940402012-01-30 16:47:22 +010016355Note: spaces must be escaped. A space character is considered as a separator.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016356In order to emit a verbatim '%', it must be preceded by another '%' resulting
Willy Tarreau06d97f92013-12-02 17:45:48 +010016357in '%%'. HAProxy will automatically merge consecutive separators.
William Lallemand48940402012-01-30 16:47:22 +010016358
Dragan Dosen835b9212016-02-12 13:23:03 +010016359Note: when using the RFC5424 syslog message format, the characters '"',
16360'\' and ']' inside PARAM-VALUE should be escaped with '\' as prefix (see
16361https://tools.ietf.org/html/rfc5424#section-6.3.3 for more details). In
16362such cases, the use of the flag "E" should be considered.
16363
William Lallemand48940402012-01-30 16:47:22 +010016364Flags are :
16365 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040016366 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
Dragan Dosen835b9212016-02-12 13:23:03 +010016367 * E: escape characters '"', '\' and ']' in a string with '\' as prefix
16368 (intended purpose is for the RFC5424 structured-data log formats)
William Lallemand48940402012-01-30 16:47:22 +010016369
16370 Example:
16371
16372 log-format %T\ %t\ Some\ Text
16373 log-format %{+Q}o\ %t\ %s\ %{-Q}r
16374
Dragan Dosen835b9212016-02-12 13:23:03 +010016375 log-format-sd %{+Q,+E}o\ [exampleSDID@1234\ header=%[capture.req.hdr(0)]]
16376
William Lallemand48940402012-01-30 16:47:22 +010016377At the moment, the default HTTP format is defined this way :
16378
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016379 log-format "%ci:%cp [%tr] %ft %b/%s %TR/%Tw/%Tc/%Tr/%Ta %ST %B %CC \
16380 %CS %tsc %ac/%fc/%bc/%sc/%rc %sq/%bq %hr %hs %{+Q}r"
William Lallemand48940402012-01-30 16:47:22 +010016381
William Lallemandbddd4fd2012-02-27 11:23:10 +010016382the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010016383
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016384 log-format "%{+Q}o %{-Q}ci - - [%trg] %r %ST %B \"\" \"\" %cp \
16385 %ms %ft %b %s %TR %Tw %Tc %Tr %Ta %tsc %ac %fc \
16386 %bc %sc %rc %sq %bq %CC %CS %hrl %hsl"
William Lallemand48940402012-01-30 16:47:22 +010016387
William Lallemandbddd4fd2012-02-27 11:23:10 +010016388and the default TCP format is defined this way :
16389
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016390 log-format "%ci:%cp [%t] %ft %b/%s %Tw/%Tc/%Tt %B %ts \
16391 %ac/%fc/%bc/%sc/%rc %sq/%bq"
William Lallemandbddd4fd2012-02-27 11:23:10 +010016392
William Lallemand48940402012-01-30 16:47:22 +010016393Please refer to the table below for currently defined variables :
16394
William Lallemandbddd4fd2012-02-27 11:23:10 +010016395 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020016396 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016397 +---+------+-----------------------------------------------+-------------+
16398 | | %o | special variable, apply flags on all next var | |
16399 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010016400 | | %B | bytes_read (from server to client) | numeric |
16401 | H | %CC | captured_request_cookie | string |
16402 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020016403 | | %H | hostname | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000016404 | H | %HM | HTTP method (ex: POST) | string |
16405 | H | %HP | HTTP request URI without query string (path) | string |
Andrew Hayworthe63ac872015-07-31 16:14:16 +000016406 | H | %HQ | HTTP request URI query string (ex: ?bar=baz) | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000016407 | H | %HU | HTTP request URI (ex: /foo?bar=baz) | string |
16408 | H | %HV | HTTP version (ex: HTTP/1.0) | string |
William Lallemanda73203e2012-03-12 12:48:57 +010016409 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020016410 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020016411 | | %T | gmt_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016412 | | %Ta | Active time of the request (from TR to end) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016413 | | %Tc | Tc | numeric |
Willy Tarreau27b639d2016-05-17 17:55:27 +020016414 | | %Td | Td = Tt - (Tq + Tw + Tc + Tr) | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080016415 | | %Tl | local_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016416 | | %Th | connection handshake time (SSL, PROXY proto) | numeric |
16417 | H | %Ti | idle time before the HTTP request | numeric |
16418 | H | %Tq | Th + Ti + TR | numeric |
16419 | H | %TR | time to receive the full request from 1st byte| numeric |
16420 | H | %Tr | Tr (response time) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020016421 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016422 | | %Tt | Tt | numeric |
16423 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010016424 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016425 | | %ac | actconn | numeric |
16426 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010016427 | | %bc | beconn (backend concurrent connections) | numeric |
16428 | | %bi | backend_source_ip (connecting address) | IP |
16429 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016430 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010016431 | | %ci | client_ip (accepted address) | IP |
16432 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016433 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010016434 | | %fc | feconn (frontend concurrent connections) | numeric |
16435 | | %fi | frontend_ip (accepting address) | IP |
16436 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020016437 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreau7346acb2014-08-28 15:03:15 +020016438 | | %lc | frontend_log_counter | numeric |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020016439 | | %hr | captured_request_headers default style | string |
16440 | | %hrl | captured_request_headers CLF style | string list |
16441 | | %hs | captured_response_headers default style | string |
16442 | | %hsl | captured_response_headers CLF style | string list |
Willy Tarreau812c88e2015-08-09 10:56:35 +020016443 | | %ms | accept date milliseconds (left-padded with 0) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020016444 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020016445 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016446 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010016447 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016448 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010016449 | | %sc | srv_conn (server concurrent connections) | numeric |
16450 | | %si | server_IP (target address) | IP |
16451 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016452 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020016453 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
16454 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010016455 | | %t | date_time (with millisecond resolution) | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016456 | H | %tr | date_time of HTTP request | date |
16457 | H | %trg | gmt_date_time of start of HTTP request | date |
Jens Bissinger15c64ff2018-08-23 14:11:27 +020016458 | H | %trl | local_date_time of start of HTTP request | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016459 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020016460 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010016461 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010016462
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020016463 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010016464
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010016465
164668.2.5. Error log format
16467-----------------------
16468
16469When an incoming connection fails due to an SSL handshake or an invalid PROXY
16470protocol header, haproxy will log the event using a shorter, fixed line format.
16471By default, logs are emitted at the LOG_INFO level, unless the option
16472"log-separate-errors" is set in the backend, in which case the LOG_ERR level
Davor Ocelice9ed2812017-12-25 17:49:28 +010016473will be used. Connections on which no data are exchanged (e.g. probes) are not
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010016474logged if the "dontlognull" option is set.
16475
16476The format looks like this :
16477
16478 >>> Dec 3 18:27:14 localhost \
16479 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
16480 Connection error during SSL handshake
16481
16482 Field Format Extract from the example above
16483 1 process_name '[' pid ']:' haproxy[6103]:
16484 2 client_ip ':' client_port 127.0.0.1:56059
16485 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
16486 4 frontend_name "/" bind_name ":" frt/f1:
16487 5 message Connection error during SSL handshake
16488
16489These fields just provide minimal information to help debugging connection
16490failures.
16491
16492
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200164938.3. Advanced logging options
16494-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016495
16496Some advanced logging options are often looked for but are not easy to find out
16497just by looking at the various options. Here is an entry point for the few
16498options which can enable better logging. Please refer to the keywords reference
16499for more information about their usage.
16500
16501
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165028.3.1. Disabling logging of external tests
16503------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016504
16505It is quite common to have some monitoring tools perform health checks on
16506haproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
16507commercial load-balancer, and sometimes it will simply be a more complete
16508monitoring system such as Nagios. When the tests are very frequent, users often
16509ask how to disable logging for those checks. There are three possibilities :
16510
16511 - if connections come from everywhere and are just TCP probes, it is often
16512 desired to simply disable logging of connections without data exchange, by
16513 setting "option dontlognull" in the frontend. It also disables logging of
16514 port scans, which may or may not be desired.
16515
16516 - if the connection come from a known source network, use "monitor-net" to
16517 declare this network as monitoring only. Any host in this network will then
16518 only be able to perform health checks, and their requests will not be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016519 logged. This is generally appropriate to designate a list of equipment
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016520 such as other load-balancers.
16521
16522 - if the tests are performed on a known URI, use "monitor-uri" to declare
16523 this URI as dedicated to monitoring. Any host sending this request will
16524 only get the result of a health-check, and the request will not be logged.
16525
16526
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165278.3.2. Logging before waiting for the session to terminate
16528----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016529
16530The problem with logging at end of connection is that you have no clue about
16531what is happening during very long sessions, such as remote terminal sessions
16532or large file downloads. This problem can be worked around by specifying
Davor Ocelice9ed2812017-12-25 17:49:28 +010016533"option logasap" in the frontend. HAProxy will then log as soon as possible,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016534just before data transfer begins. This means that in case of TCP, it will still
16535log the connection status to the server, and in case of HTTP, it will log just
16536after processing the server headers. In this case, the number of bytes reported
16537is the number of header bytes sent to the client. In order to avoid confusion
16538with normal logs, the total time field and the number of bytes are prefixed
16539with a '+' sign which means that real numbers are certainly larger.
16540
16541
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165428.3.3. Raising log level upon errors
16543------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020016544
16545Sometimes it is more convenient to separate normal traffic from errors logs,
16546for instance in order to ease error monitoring from log files. When the option
16547"log-separate-errors" is used, connections which experience errors, timeouts,
16548retries, redispatches or HTTP status codes 5xx will see their syslog level
16549raised from "info" to "err". This will help a syslog daemon store the log in
16550a separate file. It is very important to keep the errors in the normal traffic
16551file too, so that log ordering is not altered. You should also be careful if
16552you already have configured your syslog daemon to store all logs higher than
16553"notice" in an "admin" file, because the "err" level is higher than "notice".
16554
16555
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165568.3.4. Disabling logging of successful connections
16557--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020016558
16559Although this may sound strange at first, some large sites have to deal with
16560multiple thousands of logs per second and are experiencing difficulties keeping
16561them intact for a long time or detecting errors within them. If the option
16562"dontlog-normal" is set on the frontend, all normal connections will not be
16563logged. In this regard, a normal connection is defined as one without any
16564error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
16565and a response with a status 5xx is not considered normal and will be logged
16566too. Of course, doing is is really discouraged as it will remove most of the
16567useful information from the logs. Do this only if you have no other
16568alternative.
16569
16570
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165718.4. Timing events
16572------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016573
16574Timers provide a great help in troubleshooting network problems. All values are
16575reported in milliseconds (ms). These timers should be used in conjunction with
16576the session termination flags. In TCP mode with "option tcplog" set on the
16577frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016578mode, 5 control points are reported under the form "TR/Tw/Tc/Tr/Ta". In
16579addition, three other measures are provided, "Th", "Ti", and "Tq".
16580
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010016581Timings events in HTTP mode:
16582
16583 first request 2nd request
16584 |<-------------------------------->|<-------------- ...
16585 t tr t tr ...
16586 ---|----|----|----|----|----|----|----|----|--
16587 : Th Ti TR Tw Tc Tr Td : Ti ...
16588 :<---- Tq ---->: :
16589 :<-------------- Tt -------------->:
16590 :<--------- Ta --------->:
16591
16592Timings events in TCP mode:
16593
16594 TCP session
16595 |<----------------->|
16596 t t
16597 ---|----|----|----|----|---
16598 | Th Tw Tc Td |
16599 |<------ Tt ------->|
16600
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016601 - Th: total time to accept tcp connection and execute handshakes for low level
Davor Ocelice9ed2812017-12-25 17:49:28 +010016602 protocols. Currently, these protocols are proxy-protocol and SSL. This may
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016603 only happen once during the whole connection's lifetime. A large time here
16604 may indicate that the client only pre-established the connection without
16605 speaking, that it is experiencing network issues preventing it from
Davor Ocelice9ed2812017-12-25 17:49:28 +010016606 completing a handshake in a reasonable time (e.g. MTU issues), or that an
Willy Tarreau590a0512018-09-05 11:56:48 +020016607 SSL handshake was very expensive to compute. Please note that this time is
16608 reported only before the first request, so it is safe to average it over
16609 all request to calculate the amortized value. The second and subsequent
16610 request will always report zero here.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016611
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016612 - Ti: is the idle time before the HTTP request (HTTP mode only). This timer
16613 counts between the end of the handshakes and the first byte of the HTTP
16614 request. When dealing with a second request in keep-alive mode, it starts
Willy Tarreau590a0512018-09-05 11:56:48 +020016615 to count after the end of the transmission the previous response. When a
16616 multiplexed protocol such as HTTP/2 is used, it starts to count immediately
16617 after the previous request. Some browsers pre-establish connections to a
16618 server in order to reduce the latency of a future request, and keep them
16619 pending until they need it. This delay will be reported as the idle time. A
16620 value of -1 indicates that nothing was received on the connection.
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016621
16622 - TR: total time to get the client request (HTTP mode only). It's the time
16623 elapsed between the first bytes received and the moment the proxy received
16624 the empty line marking the end of the HTTP headers. The value "-1"
16625 indicates that the end of headers has never been seen. This happens when
16626 the client closes prematurely or times out. This time is usually very short
16627 since most requests fit in a single packet. A large time may indicate a
16628 request typed by hand during a test.
16629
16630 - Tq: total time to get the client request from the accept date or since the
16631 emission of the last byte of the previous response (HTTP mode only). It's
Davor Ocelice9ed2812017-12-25 17:49:28 +010016632 exactly equal to Th + Ti + TR unless any of them is -1, in which case it
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016633 returns -1 as well. This timer used to be very useful before the arrival of
16634 HTTP keep-alive and browsers' pre-connect feature. It's recommended to drop
16635 it in favor of TR nowadays, as the idle time adds a lot of noise to the
16636 reports.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016637
16638 - Tw: total time spent in the queues waiting for a connection slot. It
16639 accounts for backend queue as well as the server queues, and depends on the
16640 queue size, and the time needed for the server to complete previous
16641 requests. The value "-1" means that the request was killed before reaching
16642 the queue, which is generally what happens with invalid or denied requests.
16643
16644 - Tc: total time to establish the TCP connection to the server. It's the time
16645 elapsed between the moment the proxy sent the connection request, and the
16646 moment it was acknowledged by the server, or between the TCP SYN packet and
16647 the matching SYN/ACK packet in return. The value "-1" means that the
16648 connection never established.
16649
16650 - Tr: server response time (HTTP mode only). It's the time elapsed between
16651 the moment the TCP connection was established to the server and the moment
16652 the server sent its complete response headers. It purely shows its request
16653 processing time, without the network overhead due to the data transmission.
16654 It is worth noting that when the client has data to send to the server, for
16655 instance during a POST request, the time already runs, and this can distort
16656 apparent response time. For this reason, it's generally wise not to trust
16657 too much this field for POST requests initiated from clients behind an
16658 untrusted network. A value of "-1" here means that the last the response
16659 header (empty line) was never seen, most likely because the server timeout
16660 stroke before the server managed to process the request.
16661
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016662 - Ta: total active time for the HTTP request, between the moment the proxy
16663 received the first byte of the request header and the emission of the last
16664 byte of the response body. The exception is when the "logasap" option is
16665 specified. In this case, it only equals (TR+Tw+Tc+Tr), and is prefixed with
16666 a '+' sign. From this field, we can deduce "Td", the data transmission time,
16667 by subtracting other timers when valid :
16668
16669 Td = Ta - (TR + Tw + Tc + Tr)
16670
16671 Timers with "-1" values have to be excluded from this equation. Note that
16672 "Ta" can never be negative.
16673
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016674 - Tt: total session duration time, between the moment the proxy accepted it
16675 and the moment both ends were closed. The exception is when the "logasap"
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016676 option is specified. In this case, it only equals (Th+Ti+TR+Tw+Tc+Tr), and
16677 is prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016678 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016679
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016680 Td = Tt - (Th + Ti + TR + Tw + Tc + Tr)
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016681
16682 Timers with "-1" values have to be excluded from this equation. In TCP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016683 mode, "Ti", "Tq" and "Tr" have to be excluded too. Note that "Tt" can never
16684 be negative and that for HTTP, Tt is simply equal to (Th+Ti+Ta).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016685
16686These timers provide precious indications on trouble causes. Since the TCP
16687protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
16688that timers close to multiples of 3s are nearly always related to lost packets
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016689due to network problems (wires, negotiation, congestion). Moreover, if "Ta" or
16690"Tt" is close to a timeout value specified in the configuration, it often means
16691that a session has been aborted on timeout.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016692
16693Most common cases :
16694
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016695 - If "Th" or "Ti" are close to 3000, a packet has probably been lost between
16696 the client and the proxy. This is very rare on local networks but might
16697 happen when clients are on far remote networks and send large requests. It
16698 may happen that values larger than usual appear here without any network
16699 cause. Sometimes, during an attack or just after a resource starvation has
16700 ended, haproxy may accept thousands of connections in a few milliseconds.
16701 The time spent accepting these connections will inevitably slightly delay
16702 processing of other connections, and it can happen that request times in the
16703 order of a few tens of milliseconds are measured after a few thousands of
16704 new connections have been accepted at once. Using one of the keep-alive
16705 modes may display larger idle times since "Ti" measures the time spent
Patrick Mezard105faca2010-06-12 17:02:46 +020016706 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016707
16708 - If "Tc" is close to 3000, a packet has probably been lost between the
16709 server and the proxy during the server connection phase. This value should
16710 always be very low, such as 1 ms on local networks and less than a few tens
16711 of ms on remote networks.
16712
Willy Tarreau55165fe2009-05-10 12:02:55 +020016713 - If "Tr" is nearly always lower than 3000 except some rare values which seem
16714 to be the average majored by 3000, there are probably some packets lost
16715 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016716
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016717 - If "Ta" is large even for small byte counts, it generally is because
16718 neither the client nor the server decides to close the connection while
16719 haproxy is running in tunnel mode and both have agreed on a keep-alive
16720 connection mode. In order to solve this issue, it will be needed to specify
16721 one of the HTTP options to manipulate keep-alive or close options on either
16722 the frontend or the backend. Having the smallest possible 'Ta' or 'Tt' is
16723 important when connection regulation is used with the "maxconn" option on
16724 the servers, since no new connection will be sent to the server until
16725 another one is released.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016726
16727Other noticeable HTTP log cases ('xx' means any value to be ignored) :
16728
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016729 TR/Tw/Tc/Tr/+Ta The "option logasap" is present on the frontend and the log
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016730 was emitted before the data phase. All the timers are valid
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016731 except "Ta" which is shorter than reality.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016732
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016733 -1/xx/xx/xx/Ta The client was not able to send a complete request in time
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016734 or it aborted too early. Check the session termination flags
16735 then "timeout http-request" and "timeout client" settings.
16736
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016737 TR/-1/xx/xx/Ta It was not possible to process the request, maybe because
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016738 servers were out of order, because the request was invalid
16739 or forbidden by ACL rules. Check the session termination
16740 flags.
16741
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016742 TR/Tw/-1/xx/Ta The connection could not establish on the server. Either it
16743 actively refused it or it timed out after Ta-(TR+Tw) ms.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016744 Check the session termination flags, then check the
16745 "timeout connect" setting. Note that the tarpit action might
16746 return similar-looking patterns, with "Tw" equal to the time
16747 the client connection was maintained open.
16748
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016749 TR/Tw/Tc/-1/Ta The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016750 a complete response in time, or it closed its connection
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016751 unexpectedly after Ta-(TR+Tw+Tc) ms. Check the session
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016752 termination flags, then check the "timeout server" setting.
16753
16754
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200167558.5. Session state at disconnection
16756-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016757
16758TCP and HTTP logs provide a session termination indicator in the
16759"termination_state" field, just before the number of active connections. It is
167602-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
16761each of which has a special meaning :
16762
16763 - On the first character, a code reporting the first event which caused the
16764 session to terminate :
16765
16766 C : the TCP session was unexpectedly aborted by the client.
16767
16768 S : the TCP session was unexpectedly aborted by the server, or the
16769 server explicitly refused it.
16770
16771 P : the session was prematurely aborted by the proxy, because of a
16772 connection limit enforcement, because a DENY filter was matched,
16773 because of a security check which detected and blocked a dangerous
16774 error in server response which might have caused information leak
Davor Ocelice9ed2812017-12-25 17:49:28 +010016775 (e.g. cacheable cookie).
Willy Tarreau570f2212013-06-10 16:42:09 +020016776
16777 L : the session was locally processed by haproxy and was not passed to
16778 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016779
16780 R : a resource on the proxy has been exhausted (memory, sockets, source
16781 ports, ...). Usually, this appears during the connection phase, and
16782 system logs should contain a copy of the precise error. If this
16783 happens, it must be considered as a very serious anomaly which
16784 should be fixed as soon as possible by any means.
16785
16786 I : an internal error was identified by the proxy during a self-check.
16787 This should NEVER happen, and you are encouraged to report any log
16788 containing this, because this would almost certainly be a bug. It
16789 would be wise to preventively restart the process after such an
16790 event too, in case it would be caused by memory corruption.
16791
Simon Horman752dc4a2011-06-21 14:34:59 +090016792 D : the session was killed by haproxy because the server was detected
16793 as down and was configured to kill all connections when going down.
16794
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070016795 U : the session was killed by haproxy on this backup server because an
16796 active server was detected as up and was configured to kill all
16797 backup connections when going up.
16798
Willy Tarreaua2a64e92011-09-07 23:01:56 +020016799 K : the session was actively killed by an admin operating on haproxy.
16800
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016801 c : the client-side timeout expired while waiting for the client to
16802 send or receive data.
16803
16804 s : the server-side timeout expired while waiting for the server to
16805 send or receive data.
16806
16807 - : normal session completion, both the client and the server closed
16808 with nothing left in the buffers.
16809
16810 - on the second character, the TCP or HTTP session state when it was closed :
16811
Willy Tarreauf7b30a92010-12-06 22:59:17 +010016812 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016813 (HTTP mode only). Nothing was sent to any server.
16814
16815 Q : the proxy was waiting in the QUEUE for a connection slot. This can
16816 only happen when servers have a 'maxconn' parameter set. It can
16817 also happen in the global queue after a redispatch consecutive to
16818 a failed attempt to connect to a dying server. If no redispatch is
16819 reported, then no connection attempt was made to any server.
16820
16821 C : the proxy was waiting for the CONNECTION to establish on the
16822 server. The server might at most have noticed a connection attempt.
16823
16824 H : the proxy was waiting for complete, valid response HEADERS from the
16825 server (HTTP only).
16826
16827 D : the session was in the DATA phase.
16828
16829 L : the proxy was still transmitting LAST data to the client while the
16830 server had already finished. This one is very rare as it can only
16831 happen when the client dies while receiving the last packets.
16832
16833 T : the request was tarpitted. It has been held open with the client
16834 during the whole "timeout tarpit" duration or until the client
16835 closed, both of which will be reported in the "Tw" timer.
16836
16837 - : normal session completion after end of data transfer.
16838
16839 - the third character tells whether the persistence cookie was provided by
16840 the client (only in HTTP mode) :
16841
16842 N : the client provided NO cookie. This is usually the case for new
16843 visitors, so counting the number of occurrences of this flag in the
16844 logs generally indicate a valid trend for the site frequentation.
16845
16846 I : the client provided an INVALID cookie matching no known server.
16847 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020016848 cookies between HTTP/HTTPS sites, persistence conditionally
16849 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016850
16851 D : the client provided a cookie designating a server which was DOWN,
16852 so either "option persist" was used and the client was sent to
16853 this server, or it was not set and the client was redispatched to
16854 another server.
16855
Willy Tarreau996a92c2010-10-13 19:30:47 +020016856 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016857 server.
16858
Willy Tarreau996a92c2010-10-13 19:30:47 +020016859 E : the client provided a valid cookie, but with a last date which was
16860 older than what is allowed by the "maxidle" cookie parameter, so
16861 the cookie is consider EXPIRED and is ignored. The request will be
16862 redispatched just as if there was no cookie.
16863
16864 O : the client provided a valid cookie, but with a first date which was
16865 older than what is allowed by the "maxlife" cookie parameter, so
16866 the cookie is consider too OLD and is ignored. The request will be
16867 redispatched just as if there was no cookie.
16868
Willy Tarreauc89ccb62012-04-05 21:18:22 +020016869 U : a cookie was present but was not used to select the server because
16870 some other server selection mechanism was used instead (typically a
16871 "use-server" rule).
16872
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016873 - : does not apply (no cookie set in configuration).
16874
16875 - the last character reports what operations were performed on the persistence
16876 cookie returned by the server (only in HTTP mode) :
16877
16878 N : NO cookie was provided by the server, and none was inserted either.
16879
16880 I : no cookie was provided by the server, and the proxy INSERTED one.
16881 Note that in "cookie insert" mode, if the server provides a cookie,
16882 it will still be overwritten and reported as "I" here.
16883
Willy Tarreau996a92c2010-10-13 19:30:47 +020016884 U : the proxy UPDATED the last date in the cookie that was presented by
16885 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016886 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020016887 date indicated in the cookie. If any other change happens, such as
16888 a redispatch, then the cookie will be marked as inserted instead.
16889
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016890 P : a cookie was PROVIDED by the server and transmitted as-is.
16891
16892 R : the cookie provided by the server was REWRITTEN by the proxy, which
16893 happens in "cookie rewrite" or "cookie prefix" modes.
16894
16895 D : the cookie provided by the server was DELETED by the proxy.
16896
16897 - : does not apply (no cookie set in configuration).
16898
Willy Tarreau996a92c2010-10-13 19:30:47 +020016899The combination of the two first flags gives a lot of information about what
16900was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016901helpful to detect server saturation, network troubles, local system resource
16902starvation, attacks, etc...
16903
16904The most common termination flags combinations are indicated below. They are
16905alphabetically sorted, with the lowercase set just after the upper case for
16906easier finding and understanding.
16907
16908 Flags Reason
16909
16910 -- Normal termination.
16911
16912 CC The client aborted before the connection could be established to the
16913 server. This can happen when haproxy tries to connect to a recently
16914 dead (or unchecked) server, and the client aborts while haproxy is
16915 waiting for the server to respond or for "timeout connect" to expire.
16916
16917 CD The client unexpectedly aborted during data transfer. This can be
16918 caused by a browser crash, by an intermediate equipment between the
16919 client and haproxy which decided to actively break the connection,
16920 by network routing issues between the client and haproxy, or by a
16921 keep-alive session between the server and the client terminated first
16922 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010016923
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016924 cD The client did not send nor acknowledge any data for as long as the
16925 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020016926 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016927
16928 CH The client aborted while waiting for the server to start responding.
16929 It might be the server taking too long to respond or the client
16930 clicking the 'Stop' button too fast.
16931
16932 cH The "timeout client" stroke while waiting for client data during a
16933 POST request. This is sometimes caused by too large TCP MSS values
16934 for PPPoE networks which cannot transport full-sized packets. It can
16935 also happen when client timeout is smaller than server timeout and
16936 the server takes too long to respond.
16937
16938 CQ The client aborted while its session was queued, waiting for a server
16939 with enough empty slots to accept it. It might be that either all the
16940 servers were saturated or that the assigned server was taking too
16941 long a time to respond.
16942
16943 CR The client aborted before sending a full HTTP request. Most likely
16944 the request was typed by hand using a telnet client, and aborted
16945 too early. The HTTP status code is likely a 400 here. Sometimes this
16946 might also be caused by an IDS killing the connection between haproxy
Willy Tarreau0f228a02015-05-01 15:37:53 +020016947 and the client. "option http-ignore-probes" can be used to ignore
16948 connections without any data transfer.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016949
16950 cR The "timeout http-request" stroke before the client sent a full HTTP
16951 request. This is sometimes caused by too large TCP MSS values on the
16952 client side for PPPoE networks which cannot transport full-sized
16953 packets, or by clients sending requests by hand and not typing fast
16954 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020016955 request. The HTTP status code is likely a 408 here. Note: recently,
Willy Tarreau0f228a02015-05-01 15:37:53 +020016956 some browsers started to implement a "pre-connect" feature consisting
16957 in speculatively connecting to some recently visited web sites just
16958 in case the user would like to visit them. This results in many
16959 connections being established to web sites, which end up in 408
16960 Request Timeout if the timeout strikes first, or 400 Bad Request when
16961 the browser decides to close them first. These ones pollute the log
16962 and feed the error counters. Some versions of some browsers have even
16963 been reported to display the error code. It is possible to work
Davor Ocelice9ed2812017-12-25 17:49:28 +010016964 around the undesirable effects of this behavior by adding "option
Willy Tarreau0f228a02015-05-01 15:37:53 +020016965 http-ignore-probes" in the frontend, resulting in connections with
16966 zero data transfer to be totally ignored. This will definitely hide
16967 the errors of people experiencing connectivity issues though.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016968
16969 CT The client aborted while its session was tarpitted. It is important to
16970 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020016971 wrong tarpit rules have been written. If a lot of them happen, it
16972 might make sense to lower the "timeout tarpit" value to something
16973 closer to the average reported "Tw" timer, in order not to consume
16974 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016975
Willy Tarreau570f2212013-06-10 16:42:09 +020016976 LR The request was intercepted and locally handled by haproxy. Generally
16977 it means that this was a redirect or a stats request.
16978
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010016979 SC The server or an equipment between it and haproxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016980 the TCP connection (the proxy received a TCP RST or an ICMP message
16981 in return). Under some circumstances, it can also be the network
Davor Ocelice9ed2812017-12-25 17:49:28 +010016982 stack telling the proxy that the server is unreachable (e.g. no route,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016983 or no ARP response on local network). When this happens in HTTP mode,
16984 the status code is likely a 502 or 503 here.
16985
16986 sC The "timeout connect" stroke before a connection to the server could
16987 complete. When this happens in HTTP mode, the status code is likely a
16988 503 or 504 here.
16989
16990 SD The connection to the server died with an error during the data
16991 transfer. This usually means that haproxy has received an RST from
16992 the server or an ICMP message from an intermediate equipment while
16993 exchanging data with the server. This can be caused by a server crash
16994 or by a network issue on an intermediate equipment.
16995
16996 sD The server did not send nor acknowledge any data for as long as the
16997 "timeout server" setting during the data phase. This is often caused
Davor Ocelice9ed2812017-12-25 17:49:28 +010016998 by too short timeouts on L4 equipment before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016999 load-balancers, ...), as well as keep-alive sessions maintained
17000 between the client and the server expiring first on haproxy.
17001
17002 SH The server aborted before sending its full HTTP response headers, or
17003 it crashed while processing the request. Since a server aborting at
17004 this moment is very rare, it would be wise to inspect its logs to
17005 control whether it crashed and why. The logged request may indicate a
17006 small set of faulty requests, demonstrating bugs in the application.
17007 Sometimes this might also be caused by an IDS killing the connection
17008 between haproxy and the server.
17009
17010 sH The "timeout server" stroke before the server could return its
17011 response headers. This is the most common anomaly, indicating too
17012 long transactions, probably caused by server or database saturation.
17013 The immediate workaround consists in increasing the "timeout server"
17014 setting, but it is important to keep in mind that the user experience
17015 will suffer from these long response times. The only long term
17016 solution is to fix the application.
17017
17018 sQ The session spent too much time in queue and has been expired. See
17019 the "timeout queue" and "timeout connect" settings to find out how to
17020 fix this if it happens too often. If it often happens massively in
17021 short periods, it may indicate general problems on the affected
17022 servers due to I/O or database congestion, or saturation caused by
17023 external attacks.
17024
17025 PC The proxy refused to establish a connection to the server because the
17026 process' socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020017027 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017028 so that it does not happen anymore. This status is very rare and
17029 might happen when the global "ulimit-n" parameter is forced by hand.
17030
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017031 PD The proxy blocked an incorrectly formatted chunked encoded message in
17032 a request or a response, after the server has emitted its headers. In
17033 most cases, this will indicate an invalid message from the server to
Davor Ocelice9ed2812017-12-25 17:49:28 +010017034 the client. HAProxy supports chunk sizes of up to 2GB - 1 (2147483647
Willy Tarreauf3a3e132013-08-31 08:16:26 +020017035 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017036
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017037 PH The proxy blocked the server's response, because it was invalid,
17038 incomplete, dangerous (cache control), or matched a security filter.
17039 In any case, an HTTP 502 error is sent to the client. One possible
17040 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017041 containing unauthorized characters. It is also possible but quite
17042 rare, that the proxy blocked a chunked-encoding request from the
17043 client due to an invalid syntax, before the server responded. In this
17044 case, an HTTP 400 error is sent to the client and reported in the
17045 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017046
17047 PR The proxy blocked the client's HTTP request, either because of an
17048 invalid HTTP syntax, in which case it returned an HTTP 400 error to
17049 the client, or because a deny filter matched, in which case it
17050 returned an HTTP 403 error.
17051
17052 PT The proxy blocked the client's request and has tarpitted its
17053 connection before returning it a 500 server error. Nothing was sent
17054 to the server. The connection was maintained open for as long as
17055 reported by the "Tw" timer field.
17056
17057 RC A local resource has been exhausted (memory, sockets, source ports)
17058 preventing the connection to the server from establishing. The error
17059 logs will tell precisely what was missing. This is very rare and can
17060 only be solved by proper system tuning.
17061
Willy Tarreau996a92c2010-10-13 19:30:47 +020017062The combination of the two last flags gives a lot of information about how
17063persistence was handled by the client, the server and by haproxy. This is very
17064important to troubleshoot disconnections, when users complain they have to
17065re-authenticate. The commonly encountered flags are :
17066
17067 -- Persistence cookie is not enabled.
17068
17069 NN No cookie was provided by the client, none was inserted in the
17070 response. For instance, this can be in insert mode with "postonly"
17071 set on a GET request.
17072
17073 II A cookie designating an invalid server was provided by the client,
17074 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017075 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020017076 value can be presented by a client when no other server knows it.
17077
17078 NI No cookie was provided by the client, one was inserted in the
17079 response. This typically happens for first requests from every user
17080 in "insert" mode, which makes it an easy way to count real users.
17081
17082 VN A cookie was provided by the client, none was inserted in the
17083 response. This happens for most responses for which the client has
17084 already got a cookie.
17085
17086 VU A cookie was provided by the client, with a last visit date which is
17087 not completely up-to-date, so an updated cookie was provided in
17088 response. This can also happen if there was no date at all, or if
17089 there was a date but the "maxidle" parameter was not set, so that the
17090 cookie can be switched to unlimited time.
17091
17092 EI A cookie was provided by the client, with a last visit date which is
17093 too old for the "maxidle" parameter, so the cookie was ignored and a
17094 new cookie was inserted in the response.
17095
17096 OI A cookie was provided by the client, with a first visit date which is
17097 too old for the "maxlife" parameter, so the cookie was ignored and a
17098 new cookie was inserted in the response.
17099
17100 DI The server designated by the cookie was down, a new server was
17101 selected and a new cookie was emitted in the response.
17102
17103 VI The server designated by the cookie was not marked dead but could not
17104 be reached. A redispatch happened and selected another one, which was
17105 then advertised in the response.
17106
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017107
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200171088.6. Non-printable characters
17109-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017110
17111In order not to cause trouble to log analysis tools or terminals during log
17112consulting, non-printable characters are not sent as-is into log files, but are
17113converted to the two-digits hexadecimal representation of their ASCII code,
17114prefixed by the character '#'. The only characters that can be logged without
17115being escaped are comprised between 32 and 126 (inclusive). Obviously, the
17116escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
17117is the same for the character '"' which becomes "#22", as well as '{', '|' and
17118'}' when logging headers.
17119
17120Note that the space character (' ') is not encoded in headers, which can cause
17121issues for tools relying on space count to locate fields. A typical header
17122containing spaces is "User-Agent".
17123
17124Last, it has been observed that some syslog daemons such as syslog-ng escape
17125the quote ('"') with a backslash ('\'). The reverse operation can safely be
17126performed since no quote may appear anywhere else in the logs.
17127
17128
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200171298.7. Capturing HTTP cookies
17130---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017131
17132Cookie capture simplifies the tracking a complete user session. This can be
17133achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017134section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017135cookie will simultaneously be checked in the request ("Cookie:" header) and in
17136the response ("Set-Cookie:" header). The respective values will be reported in
17137the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017138locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017139not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
17140user switches to a new session for example, because the server will reassign it
17141a new cookie. It is also possible to detect if a server unexpectedly sets a
17142wrong cookie to a client, leading to session crossing.
17143
17144 Examples :
17145 # capture the first cookie whose name starts with "ASPSESSION"
17146 capture cookie ASPSESSION len 32
17147
17148 # capture the first cookie whose name is exactly "vgnvisitor"
17149 capture cookie vgnvisitor= len 32
17150
17151
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200171528.8. Capturing HTTP headers
17153---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017154
17155Header captures are useful to track unique request identifiers set by an upper
17156proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
17157the response, one can search for information about the response length, how the
17158server asked the cache to behave, or an object location during a redirection.
17159
17160Header captures are performed using the "capture request header" and "capture
17161response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017162section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017163
17164It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010017165time. Non-existent headers are logged as empty strings, and if one header
17166appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017167are grouped within braces '{' and '}' in the same order as they were declared,
17168and delimited with a vertical bar '|' without any space. Response headers
17169follow the same representation, but are displayed after a space following the
17170request headers block. These blocks are displayed just before the HTTP request
17171in the logs.
17172
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020017173As a special case, it is possible to specify an HTTP header capture in a TCP
17174frontend. The purpose is to enable logging of headers which will be parsed in
17175an HTTP backend if the request is then switched to this HTTP backend.
17176
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017177 Example :
17178 # This instance chains to the outgoing proxy
17179 listen proxy-out
17180 mode http
17181 option httplog
17182 option logasap
17183 log global
17184 server cache1 192.168.1.1:3128
17185
17186 # log the name of the virtual server
17187 capture request header Host len 20
17188
17189 # log the amount of data uploaded during a POST
17190 capture request header Content-Length len 10
17191
17192 # log the beginning of the referrer
17193 capture request header Referer len 20
17194
17195 # server name (useful for outgoing proxies only)
17196 capture response header Server len 20
17197
17198 # logging the content-length is useful with "option logasap"
17199 capture response header Content-Length len 10
17200
Davor Ocelice9ed2812017-12-25 17:49:28 +010017201 # log the expected cache behavior on the response
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017202 capture response header Cache-Control len 8
17203
17204 # the Via header will report the next proxy's name
17205 capture response header Via len 20
17206
17207 # log the URL location during a redirection
17208 capture response header Location len 20
17209
17210 >>> Aug 9 20:26:09 localhost \
17211 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
17212 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
17213 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
17214 "GET http://fr.adserver.yahoo.com/"
17215
17216 >>> Aug 9 20:30:46 localhost \
17217 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
17218 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
17219 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017220 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017221
17222 >>> Aug 9 20:30:46 localhost \
17223 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
17224 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
17225 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
17226 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017227 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017228
17229
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172308.9. Examples of logs
17231---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017232
17233These are real-world examples of logs accompanied with an explanation. Some of
17234them have been made up by hand. The syslog part has been removed for better
17235reading. Their sole purpose is to explain how to decipher them.
17236
17237 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
17238 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
17239 "HEAD / HTTP/1.0"
17240
17241 => long request (6.5s) entered by hand through 'telnet'. The server replied
17242 in 147 ms, and the session ended normally ('----')
17243
17244 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
17245 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
17246 0/9 "HEAD / HTTP/1.0"
17247
17248 => Idem, but the request was queued in the global queue behind 9 other
17249 requests, and waited there for 1230 ms.
17250
17251 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
17252 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
17253 "GET /image.iso HTTP/1.0"
17254
17255 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010017256 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017257 14 ms, 243 bytes of headers were sent to the client, and total time from
17258 accept to first data byte is 30 ms.
17259
17260 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
17261 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
17262 "GET /cgi-bin/bug.cgi? HTTP/1.0"
17263
17264 => the proxy blocked a server response either because of an "rspdeny" or
17265 "rspideny" filter, or because the response was improperly formatted and
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020017266 not HTTP-compliant, or because it blocked sensitive information which
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017267 risked being cached. In this case, the response is replaced with a "502
17268 bad gateway". The flags ("PH--") tell us that it was haproxy who decided
17269 to return the 502 and not the server.
17270
17271 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017272 px-http/<NOSRV> -1/-1/-1/-1/8490 -1 0 - - CR-- 2/2/2/0/0 0/0 ""
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017273
17274 => the client never completed its request and aborted itself ("C---") after
17275 8.5s, while the proxy was waiting for the request headers ("-R--").
17276 Nothing was sent to any server.
17277
17278 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
17279 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
17280
17281 => The client never completed its request, which was aborted by the
17282 time-out ("c---") after 50s, while the proxy was waiting for the request
Davor Ocelice9ed2812017-12-25 17:49:28 +010017283 headers ("-R--"). Nothing was sent to any server, but the proxy could
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017284 send a 408 return code to the client.
17285
17286 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
17287 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
17288
17289 => This log was produced with "option tcplog". The client timed out after
17290 5 seconds ("c----").
17291
17292 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
17293 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017294 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017295
17296 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017297 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017298 (config says 'retries 3'), and no redispatch (otherwise we would have
17299 seen "/+3"). Status code 503 was returned to the client. There were 115
17300 connections on this server, 202 connections on this proxy, and 205 on
17301 the global process. It is possible that the server refused the
17302 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010017303
Willy Tarreau52b2d222011-09-07 23:48:48 +020017304
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200173059. Supported filters
17306--------------------
17307
17308Here are listed officially supported filters with the list of parameters they
17309accept. Depending on compile options, some of these filters might be
17310unavailable. The list of available filters is reported in haproxy -vv.
17311
17312See also : "filter"
17313
173149.1. Trace
17315----------
17316
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010017317filter trace [name <name>] [random-parsing] [random-forwarding] [hexdump]
Christopher Fauletc3fe5332016-04-07 15:30:10 +020017318
17319 Arguments:
17320 <name> is an arbitrary name that will be reported in
17321 messages. If no name is provided, "TRACE" is used.
17322
17323 <random-parsing> enables the random parsing of data exchanged between
17324 the client and the server. By default, this filter
17325 parses all available data. With this parameter, it
17326 only parses a random amount of the available data.
17327
Davor Ocelice9ed2812017-12-25 17:49:28 +010017328 <random-forwarding> enables the random forwarding of parsed data. By
Christopher Fauletc3fe5332016-04-07 15:30:10 +020017329 default, this filter forwards all previously parsed
17330 data. With this parameter, it only forwards a random
17331 amount of the parsed data.
17332
Davor Ocelice9ed2812017-12-25 17:49:28 +010017333 <hexdump> dumps all forwarded data to the server and the client.
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010017334
Christopher Fauletc3fe5332016-04-07 15:30:10 +020017335This filter can be used as a base to develop new filters. It defines all
17336callbacks and print a message on the standard error stream (stderr) with useful
17337information for all of them. It may be useful to debug the activity of other
17338filters or, quite simply, HAProxy's activity.
17339
17340Using <random-parsing> and/or <random-forwarding> parameters is a good way to
17341tests the behavior of a filter that parses data exchanged between a client and
17342a server by adding some latencies in the processing.
17343
17344
173459.2. HTTP compression
17346---------------------
17347
17348filter compression
17349
17350The HTTP compression has been moved in a filter in HAProxy 1.7. "compression"
17351keyword must still be used to enable and configure the HTTP compression. And
17352when no other filter is used, it is enough. But it is mandatory to explicitly
17353use a filter line to enable the HTTP compression when two or more filters are
17354used for the same listener/frontend/backend. This is important to know the
17355filters evaluation order.
17356
17357See also : "compression"
17358
17359
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +0200173609.3. Stream Processing Offload Engine (SPOE)
17361--------------------------------------------
17362
17363filter spoe [engine <name>] config <file>
17364
17365 Arguments :
17366
17367 <name> is the engine name that will be used to find the right scope in
17368 the configuration file. If not provided, all the file will be
17369 parsed.
17370
17371 <file> is the path of the engine configuration file. This file can
17372 contain configuration of several engines. In this case, each
17373 part must be placed in its own scope.
17374
17375The Stream Processing Offload Engine (SPOE) is a filter communicating with
17376external components. It allows the offload of some specifics processing on the
Davor Ocelice9ed2812017-12-25 17:49:28 +010017377streams in tiered applications. These external components and information
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020017378exchanged with them are configured in dedicated files, for the main part. It
17379also requires dedicated backends, defined in HAProxy configuration.
17380
17381SPOE communicates with external components using an in-house binary protocol,
17382the Stream Processing Offload Protocol (SPOP).
17383
Tim Düsterhus4896c442016-11-29 02:15:19 +010017384For all information about the SPOE configuration and the SPOP specification, see
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020017385"doc/SPOE.txt".
17386
17387Important note:
17388 The SPOE filter is highly experimental for now and was not heavily
17389 tested. It is really not production ready. So use it carefully.
17390
William Lallemand86d0df02017-11-24 21:36:45 +01001739110. Cache
17392---------
17393
17394HAProxy provides a cache, which was designed to perform cache on small objects
17395(favicon, css...). This is a minimalist low-maintenance cache which runs in
17396RAM.
17397
17398The cache is based on a memory which is shared between processes and threads,
Cyril Bonté7b888f12017-11-26 22:24:31 +010017399this memory is split in blocks of 1k.
William Lallemand86d0df02017-11-24 21:36:45 +010017400
17401If an object is not used anymore, it can be deleted to store a new object
17402independently of its expiration date. The oldest objects are deleted first
17403when we try to allocate a new one.
17404
Cyril Bonté7b888f12017-11-26 22:24:31 +010017405The cache uses a hash of the host header and the URI as the key.
William Lallemand86d0df02017-11-24 21:36:45 +010017406
17407It's possible to view the status of a cache using the Unix socket command
17408"show cache" consult section 9.3 "Unix Socket commands" of Management Guide
17409for more details.
17410
17411When an object is delivered from the cache, the server name in the log is
17412replaced by "<CACHE>".
17413
Cyril Bonté7b888f12017-11-26 22:24:31 +01001741410.1. Limitation
17415----------------
William Lallemand86d0df02017-11-24 21:36:45 +010017416
17417The cache won't store and won't deliver objects in these cases:
17418
17419- If the response is not a 200
17420- If the response contains a Vary header
17421- If the response does not contain a Content-Length header or if the
17422 Content-Length + the headers size is greater than a buffer size - the
17423 reserve.
17424- If the response is not cacheable
17425
17426- If the request is not a GET
17427- If the HTTP version of the request is smaller than 1.1
William Lallemand8a16fe02018-05-22 11:04:33 +020017428- If the request contains an Authorization header
William Lallemand86d0df02017-11-24 21:36:45 +010017429
17430Caution!: Due to the current limitation of the filters, it is not recommended
17431to use the cache with other filters. Using them can cause undefined behavior
Cyril Bonté7b888f12017-11-26 22:24:31 +010017432if they modify the response (compression for example).
William Lallemand86d0df02017-11-24 21:36:45 +010017433
Cyril Bonté7b888f12017-11-26 22:24:31 +01001743410.2. Setup
17435-----------
William Lallemand86d0df02017-11-24 21:36:45 +010017436
17437To setup a cache, you must define a cache section and use it in a proxy with
17438the corresponding http-request and response actions.
17439
Cyril Bonté7b888f12017-11-26 22:24:31 +01001744010.2.1. Cache section
17441---------------------
William Lallemand86d0df02017-11-24 21:36:45 +010017442
17443cache <name>
17444 Declare a cache section, allocate a shared cache memory named <name>, the
17445 size of cache is mandatory.
17446
17447total-max-size <megabytes>
Davor Ocelice9ed2812017-12-25 17:49:28 +010017448 Define the size in RAM of the cache in megabytes. This size is split in
17449 blocks of 1kB which are used by the cache entries.
William Lallemand86d0df02017-11-24 21:36:45 +010017450
17451max-age <seconds>
17452 Define the maximum expiration duration. The expiration is set has the lowest
17453 value between the s-maxage or max-age (in this order) directive in the
17454 Cache-Control response header and this value. The default value is 60
17455 seconds, which means that you can't cache an object more than 60 seconds by
17456 default.
17457
Cyril Bonté7b888f12017-11-26 22:24:31 +01001745810.2.2. Proxy section
17459---------------------
William Lallemand86d0df02017-11-24 21:36:45 +010017460
17461http-request cache-use <name>
17462 Try to deliver a cached object from the cache <name>. This directive is also
17463 mandatory to store the cache as it calculates the cache hash. If you want to
17464 use a condition for both storage and delivering that's a good idea to put it
17465 after this one.
17466
17467http-response cache-store <name>
17468 Store an http-response within the cache. The storage of the response headers
17469 is done at this step, which means you can use others http-response actions
17470 to modify headers before or after the storage of the response. This action
17471 is responsible for the setup of the cache storage filter.
17472
17473
17474Example:
17475
17476 backend bck1
17477 mode http
17478
17479 http-request cache-use foobar
17480 http-response cache-store foobar
17481 server srv1 127.0.0.1:80
17482
17483 cache foobar
17484 total-max-size 4
17485 max-age 240
17486
Willy Tarreau0ba27502007-12-24 16:55:16 +010017487/*
17488 * Local variables:
17489 * fill-column: 79
17490 * End:
17491 */