blob: 2d715553c9836cde1a4b873346ee79b43c73c2d1 [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 Tarreaua6918e32014-07-12 16:40:52 +02005 version 1.5.2
Willy Tarreau6a06a402007-07-15 20:15:28 +02006 willy tarreau
Willy Tarreaua6918e32014-07-12 16:40:52 +02007 2014/07/12
Willy Tarreau6a06a402007-07-15 20:15:28 +02008
9
10This document covers the configuration language as implemented in the version
11specified above. It does not provide any hint, example or advice. For such
Willy Tarreau0ba27502007-12-24 16:55:16 +010012documentation, please refer to the Reference Manual or the Architecture Manual.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013The summary below is meant to help you search sections by name and navigate
14through 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
22 sometimes useful to prefix all output lines (logs, console outs) with 3
23 closing angle brackets ('>>>') in order to help get the difference between
24 inputs and outputs when it can become ambiguous. If you add sections,
25 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
341.2.1. The Request line
351.2.2. The request headers
361.3. HTTP response
371.3.1. The Response line
381.3.2. The response headers
39
402. Configuring HAProxy
412.1. Configuration file format
422.2. Time format
Patrick Mezard35da19c2010-06-12 17:02:47 +0200432.3. Examples
Willy Tarreauc57f0e22009-05-10 13:12:33 +020044
453. Global parameters
463.1. Process management and security
473.2. Performance tuning
483.3. Debugging
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +0100493.4. Userlists
Cyril Bontédc4d9032012-04-08 21:57:39 +0200503.5. Peers
Willy Tarreauc57f0e22009-05-10 13:12:33 +020051
524. Proxies
534.1. Proxy keywords matrix
544.2. Alphabetically sorted keywords reference
55
Willy Tarreau086fbf52012-09-24 20:34:51 +0200565. Bind and Server options
575.1. Bind options
585.2. Server and default-server options
Willy Tarreauc57f0e22009-05-10 13:12:33 +020059
606. HTTP header manipulation
61
Willy Tarreau74ca5042013-06-11 23:12:07 +0200627. Using ACLs and fetching samples
637.1. ACL basics
647.1.1. Matching booleans
657.1.2. Matching integers
667.1.3. Matching strings
677.1.4. Matching regular expressions (regexes)
687.1.5. Matching arbitrary data blocks
697.1.6. Matching IPv4 and IPv6 addresses
707.2. Using ACLs to form conditions
717.3. Fetching samples
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200727.3.1. Converters
737.3.2. Fetching samples from internal states
747.3.3. Fetching samples at Layer 4
757.3.4. Fetching samples at Layer 5
767.3.5. Fetching samples from buffer contents (Layer 6)
777.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +0200787.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020079
808. Logging
818.1. Log levels
828.2. Log formats
838.2.1. Default log format
848.2.2. TCP log format
858.2.3. HTTP log format
William Lallemand48940402012-01-30 16:47:22 +0100868.2.4. Custom log format
Willy Tarreau5f51e1a2012-12-03 18:40:10 +0100878.2.5. Error log format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200888.3. Advanced logging options
898.3.1. Disabling logging of external tests
908.3.2. Logging before waiting for the session to terminate
918.3.3. Raising log level upon errors
928.3.4. Disabling logging of successful connections
938.4. Timing events
948.5. Session state at disconnection
958.6. Non-printable characters
968.7. Capturing HTTP cookies
978.8. Capturing HTTP headers
988.9. Examples of logs
99
1009. Statistics and monitoring
1019.1. CSV format
1029.2. Unix Socket commands
103
104
1051. Quick reminder about HTTP
106----------------------------
107
108When haproxy is running in HTTP mode, both the request and the response are
109fully analyzed and indexed, thus it becomes possible to build matching criteria
110on almost anything found in the contents.
111
112However, it is important to understand how HTTP requests and responses are
113formed, and how HAProxy decomposes them. It will then become easier to write
114correct rules and to debug existing configurations.
115
116
1171.1. The HTTP transaction model
118-------------------------------
119
120The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100121to one and only one response. Traditionally, a TCP connection is established
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200122from the client to the server, a request is sent by the client on the
123connection, the server responds and the connection is closed. A new request
124will involve a new connection :
125
126 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
127
128In this mode, called the "HTTP close" mode, there are as many connection
129establishments as there are HTTP transactions. Since the connection is closed
130by the server after the response, the client does not need to know the content
131length.
132
133Due to the transactional nature of the protocol, it was possible to improve it
134to avoid closing a connection between two subsequent transactions. In this mode
135however, it is mandatory that the server indicates the content length for each
136response so that the client does not wait indefinitely. For this, a special
137header is used: "Content-length". This mode is called the "keep-alive" mode :
138
139 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
140
141Its advantages are a reduced latency between transactions, and less processing
142power required on the server side. It is generally better than the close mode,
143but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200144a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200145
146A last improvement in the communications is the pipelining mode. It still uses
147keep-alive, but the client does not wait for the first response to send the
148second request. This is useful for fetching large number of images composing a
149page :
150
151 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
152
153This can obviously have a tremendous benefit on performance because the network
154latency is eliminated between subsequent requests. Many HTTP agents do not
155correctly support pipelining since there is no way to associate a response with
156the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100157server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200158
Willy Tarreau70dffda2014-01-30 03:07:23 +0100159By default HAProxy operates in keep-alive mode with regards to persistent
160connections: for each connection it processes each request and response, and
161leaves the connection idle on both sides between the end of a response and the
162start of a new request.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200163
Willy Tarreau70dffda2014-01-30 03:07:23 +0100164HAProxy supports 5 connection modes :
165 - keep alive : all requests and responses are processed (default)
166 - tunnel : only the first request and response are processed,
167 everything else is forwarded with no analysis.
168 - passive close : tunnel with "Connection: close" added in both directions.
169 - server close : the server-facing connection is closed after the response.
170 - forced close : the connection is actively closed after end of response.
171
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172
1731.2. HTTP request
174-----------------
175
176First, let's consider this HTTP request :
177
178 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100179 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200180 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
181 2 Host: www.mydomain.com
182 3 User-agent: my small browser
183 4 Accept: image/jpeg, image/gif
184 5 Accept: image/png
185
186
1871.2.1. The Request line
188-----------------------
189
190Line 1 is the "request line". It is always composed of 3 fields :
191
192 - a METHOD : GET
193 - a URI : /serv/login.php?lang=en&profile=2
194 - a version tag : HTTP/1.1
195
196All of them are delimited by what the standard calls LWS (linear white spaces),
197which are commonly spaces, but can also be tabs or line feeds/carriage returns
198followed by spaces/tabs. The method itself cannot contain any colon (':') and
199is limited to alphabetic letters. All those various combinations make it
200desirable that HAProxy performs the splitting itself rather than leaving it to
201the user to write a complex or inaccurate regular expression.
202
203The URI itself can have several forms :
204
205 - A "relative URI" :
206
207 /serv/login.php?lang=en&profile=2
208
209 It is a complete URL without the host part. This is generally what is
210 received by servers, reverse proxies and transparent proxies.
211
212 - An "absolute URI", also called a "URL" :
213
214 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
215
216 It is composed of a "scheme" (the protocol name followed by '://'), a host
217 name or address, optionally a colon (':') followed by a port number, then
218 a relative URI beginning at the first slash ('/') after the address part.
219 This is generally what proxies receive, but a server supporting HTTP/1.1
220 must accept this form too.
221
222 - a star ('*') : this form is only accepted in association with the OPTIONS
223 method and is not relayable. It is used to inquiry a next hop's
224 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100225
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200226 - an address:port combination : 192.168.0.12:80
227 This is used with the CONNECT method, which is used to establish TCP
228 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
229 other protocols too.
230
231In a relative URI, two sub-parts are identified. The part before the question
232mark is called the "path". It is typically the relative path to static objects
233on the server. The part after the question mark is called the "query string".
234It is mostly used with GET requests sent to dynamic scripts and is very
235specific to the language, framework or application in use.
236
237
2381.2.2. The request headers
239--------------------------
240
241The headers start at the second line. They are composed of a name at the
242beginning of the line, immediately followed by a colon (':'). Traditionally,
243an LWS is added after the colon but that's not required. Then come the values.
244Multiple identical headers may be folded into one single line, delimiting the
245values with commas, provided that their order is respected. This is commonly
246encountered in the "Cookie:" field. A header may span over multiple lines if
247the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
248define a total of 3 values for the "Accept:" header.
249
250Contrary to a common mis-conception, header names are not case-sensitive, and
251their values are not either if they refer to other header names (such as the
252"Connection:" header).
253
254The end of the headers is indicated by the first empty line. People often say
255that it's a double line feed, which is not exact, even if a double line feed
256is one valid form of empty line.
257
258Fortunately, HAProxy takes care of all these complex combinations when indexing
259headers, checking values and counting them, so there is no reason to worry
260about the way they could be written, but it is important not to accuse an
261application of being buggy if it does unusual, valid things.
262
263Important note:
264 As suggested by RFC2616, HAProxy normalizes headers by replacing line breaks
265 in the middle of headers by LWS in order to join multi-line headers. This
266 is necessary for proper analysis and helps less capable HTTP parsers to work
267 correctly and not to be fooled by such complex constructs.
268
269
2701.3. HTTP response
271------------------
272
273An HTTP response looks very much like an HTTP request. Both are called HTTP
274messages. Let's consider this HTTP response :
275
276 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100277 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200278 1 HTTP/1.1 200 OK
279 2 Content-length: 350
280 3 Content-Type: text/html
281
Willy Tarreau816b9792009-09-15 21:25:21 +0200282As a special case, HTTP supports so called "Informational responses" as status
283codes 1xx. These messages are special in that they don't convey any part of the
284response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100285continue to post its request for instance. In the case of a status 100 response
286the requested information will be carried by the next non-100 response message
287following the informational one. This implies that multiple responses may be
288sent to a single request, and that this only works when keep-alive is enabled
289(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
290correctly forward and skip them, and only process the next non-100 response. As
291such, these messages are neither logged nor transformed, unless explicitly
292state otherwise. Status 101 messages indicate that the protocol is changing
293over the same connection and that haproxy must switch to tunnel mode, just as
294if a CONNECT had occurred. Then the Upgrade header would contain additional
295information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200296
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200297
2981.3.1. The Response line
299------------------------
300
301Line 1 is the "response line". It is always composed of 3 fields :
302
303 - a version tag : HTTP/1.1
304 - a status code : 200
305 - a reason : OK
306
307The status code is always 3-digit. The first digit indicates a general status :
Willy Tarreau816b9792009-09-15 21:25:21 +0200308 - 1xx = informational message to be skipped (eg: 100, 101)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200309 - 2xx = OK, content is following (eg: 200, 206)
310 - 3xx = OK, no content following (eg: 302, 304)
311 - 4xx = error caused by the client (eg: 401, 403, 404)
312 - 5xx = error caused by the server (eg: 500, 502, 503)
313
314Please refer to RFC2616 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100315"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200316found there, but it's a common practice to respect the well-established
317messages. It can be composed of one or multiple words, such as "OK", "Found",
318or "Authentication Required".
319
320Haproxy may emit the following status codes by itself :
321
322 Code When / reason
323 200 access to stats page, and when replying to monitoring requests
324 301 when performing a redirection, depending on the configured code
325 302 when performing a redirection, depending on the configured code
326 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100327 307 when performing a redirection, depending on the configured code
328 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200329 400 for an invalid or too large request
330 401 when an authentication is required to perform the action (when
331 accessing the stats page)
332 403 when a request is forbidden by a "block" ACL or "reqdeny" filter
333 408 when the request timeout strikes before the request is complete
334 500 when haproxy encounters an unrecoverable internal error, such as a
335 memory allocation failure, which should never happen
336 502 when the server returns an empty, invalid or incomplete response, or
337 when an "rspdeny" filter blocks the response.
338 503 when no server was available to handle the request, or in response to
339 monitoring requests which match the "monitor fail" condition
340 504 when the response timeout strikes before the server responds
341
342The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3434.2).
344
345
3461.3.2. The response headers
347---------------------------
348
349Response headers work exactly like request headers, and as such, HAProxy uses
350the same parsing function for both. Please refer to paragraph 1.2.2 for more
351details.
352
353
3542. Configuring HAProxy
355----------------------
356
3572.1. Configuration file format
358------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200359
360HAProxy's configuration process involves 3 major sources of parameters :
361
362 - the arguments from the command-line, which always take precedence
363 - the "global" section, which sets process-wide parameters
364 - the proxies sections which can take form of "defaults", "listen",
365 "frontend" and "backend".
366
Willy Tarreau0ba27502007-12-24 16:55:16 +0100367The configuration file syntax consists in lines beginning with a keyword
368referenced in this manual, optionally followed by one or several parameters
369delimited by spaces. If spaces have to be entered in strings, then they must be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100370preceded by a backslash ('\') to be escaped. Backslashes also have to be
Willy Tarreau0ba27502007-12-24 16:55:16 +0100371escaped by doubling them.
372
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200373
3742.2. Time format
375----------------
376
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100377Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100378values are generally expressed in milliseconds (unless explicitly stated
379otherwise) but may be expressed in any other unit by suffixing the unit to the
380numeric value. It is important to consider this because it will not be repeated
381for every keyword. Supported units are :
382
383 - us : microseconds. 1 microsecond = 1/1000000 second
384 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
385 - s : seconds. 1s = 1000ms
386 - m : minutes. 1m = 60s = 60000ms
387 - h : hours. 1h = 60m = 3600s = 3600000ms
388 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
389
390
Patrick Mezard35da19c2010-06-12 17:02:47 +02003912.3. Examples
392-------------
393
394 # Simple configuration for an HTTP proxy listening on port 80 on all
395 # interfaces and forwarding requests to a single backend "servers" with a
396 # single server "server1" listening on 127.0.0.1:8000
397 global
398 daemon
399 maxconn 256
400
401 defaults
402 mode http
403 timeout connect 5000ms
404 timeout client 50000ms
405 timeout server 50000ms
406
407 frontend http-in
408 bind *:80
409 default_backend servers
410
411 backend servers
412 server server1 127.0.0.1:8000 maxconn 32
413
414
415 # The same configuration defined with a single listen block. Shorter but
416 # less expressive, especially in HTTP mode.
417 global
418 daemon
419 maxconn 256
420
421 defaults
422 mode http
423 timeout connect 5000ms
424 timeout client 50000ms
425 timeout server 50000ms
426
427 listen http-in
428 bind *:80
429 server server1 127.0.0.1:8000 maxconn 32
430
431
432Assuming haproxy is in $PATH, test these configurations in a shell with:
433
Willy Tarreauccb289d2010-12-11 20:19:38 +0100434 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200435
436
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004373. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200438--------------------
439
440Parameters in the "global" section are process-wide and often OS-specific. They
441are generally set once for all and do not need being changed once correct. Some
442of them have command-line equivalents.
443
444The following keywords are supported in the "global" section :
445
446 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200447 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200448 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200449 - crt-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200450 - daemon
451 - gid
452 - group
453 - log
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100454 - log-send-hostname
Willy Tarreau6a06a402007-07-15 20:15:28 +0200455 - nbproc
456 - pidfile
457 - uid
458 - ulimit-n
459 - user
Willy Tarreaufbee7132007-10-18 13:53:22 +0200460 - stats
Emeric Brun850efd52014-01-29 12:24:34 +0100461 - ssl-server-verify
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +0200462 - node
463 - description
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100464 - unix-bind
Willy Tarreaud72758d2010-01-12 10:42:19 +0100465
Willy Tarreau6a06a402007-07-15 20:15:28 +0200466 * Performance tuning
Willy Tarreau1746eec2014-04-25 10:46:47 +0200467 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +0200468 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +0200469 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +0100470 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +0100471 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100472 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +0200473 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +0200474 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +0200475 - maxsslrate
Willy Tarreau6a06a402007-07-15 20:15:28 +0200476 - noepoll
477 - nokqueue
478 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100479 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300480 - nogetaddrinfo
Willy Tarreaufe255b72007-10-14 23:09:26 +0200481 - spread-checks
Willy Tarreau27a674e2009-08-17 07:23:33 +0200482 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +0200483 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +0100484 - tune.comp.maxlevel
Willy Tarreau193b8c62012-11-22 00:17:38 +0100485 - tune.http.cookielen
Willy Tarreauac1932d2011-10-24 19:14:41 +0200486 - tune.http.maxhdr
Willy Tarreau7e312732014-02-12 16:35:14 +0100487 - tune.idletimer
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100488 - tune.maxaccept
489 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +0200490 - tune.maxrewrite
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200491 - tune.pipesize
Willy Tarreaue803de22010-01-21 17:43:04 +0100492 - tune.rcvbuf.client
493 - tune.rcvbuf.server
494 - tune.sndbuf.client
495 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +0100496 - tune.ssl.cachesize
Willy Tarreaubfd59462013-02-21 07:46:09 +0100497 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +0200498 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +0100499 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +0200500 - tune.ssl.default-dh-param
William Lallemanda509e4c2012-11-07 16:54:34 +0100501 - tune.zlib.memlevel
502 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +0100503
Willy Tarreau6a06a402007-07-15 20:15:28 +0200504 * Debugging
505 - debug
506 - quiet
Willy Tarreau6a06a402007-07-15 20:15:28 +0200507
508
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005093.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +0200510------------------------------------
511
Emeric Brunc8e8d122012-10-02 18:42:10 +0200512ca-base <dir>
513 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emeric Brunfd33a262012-10-11 16:28:27 +0200514 relative path is used with "ca-file" or "crl-file" directives. Absolute
515 locations specified in "ca-file" and "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +0200516
Willy Tarreau6a06a402007-07-15 20:15:28 +0200517chroot <jail dir>
518 Changes current directory to <jail dir> and performs a chroot() there before
519 dropping privileges. This increases the security level in case an unknown
520 vulnerability would be exploited, since it would make it very hard for the
521 attacker to exploit the system. This only works when the process is started
522 with superuser privileges. It is important to ensure that <jail_dir> is both
523 empty and unwritable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100524
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100525cpu-map <"all"|"odd"|"even"|process_num> <cpu-set>...
526 On Linux 2.6 and above, it is possible to bind a process to a specific CPU
527 set. This means that the process will never run on other CPUs. The "cpu-map"
528 directive specifies CPU sets for process sets. The first argument is the
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100529 process number to bind. This process must have a number between 1 and 32 or
530 64, depending on the machine's word size, and any process IDs above nbproc
531 are ignored. It is possible to specify all processes at once using "all",
532 only odd numbers using "odd" or even numbers using "even", just like with the
533 "bind-process" directive. The second and forthcoming arguments are CPU sets.
534 Each CPU set is either a unique number between 0 and 31 or 63 or a range with
535 two such numbers delimited by a dash ('-'). Multiple CPU numbers or ranges
536 may be specified, and the processes will be allowed to bind to all of them.
537 Obviously, multiple "cpu-map" directives may be specified. Each "cpu-map"
538 directive will replace the previous ones when they overlap.
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100539
Emeric Brunc8e8d122012-10-02 18:42:10 +0200540crt-base <dir>
541 Assigns a default directory to fetch SSL certificates from when a relative
542 path is used with "crtfile" directives. Absolute locations specified after
543 "crtfile" prevail and ignore "crt-base".
544
Willy Tarreau6a06a402007-07-15 20:15:28 +0200545daemon
546 Makes the process fork into background. This is the recommended mode of
547 operation. It is equivalent to the command line "-D" argument. It can be
548 disabled by the command line "-db" argument.
549
550gid <number>
551 Changes the process' group ID to <number>. It is recommended that the group
552 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
553 be started with a user belonging to this group, or with superuser privileges.
Michael Schererab012dd2013-01-12 18:35:19 +0100554 Note that if haproxy is started from a user having supplementary groups, it
555 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200556 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100557
Willy Tarreau6a06a402007-07-15 20:15:28 +0200558group <group name>
559 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
560 See also "gid" and "user".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100561
Willy Tarreaudc2695c2014-06-27 18:10:07 +0200562log <address> [len <length>] <facility> [max level [min level]]
Willy Tarreau6a06a402007-07-15 20:15:28 +0200563 Adds a global syslog server. Up to two global servers can be defined. They
564 will receive logs for startups and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +0100565 configured with "log global".
566
567 <address> can be one of:
568
Willy Tarreau2769aa02007-12-27 18:26:09 +0100569 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +0100570 no port is specified, 514 is used by default (the standard syslog
571 port).
572
David du Colombier24bb5f52011-03-17 10:40:23 +0100573 - An IPv6 address followed by a colon and optionally a UDP port. If
574 no port is specified, 514 is used by default (the standard syslog
575 port).
576
Robert Tsai81ae1952007-12-05 10:47:29 +0100577 - A filesystem path to a UNIX domain socket, keeping in mind
578 considerations for chroot (be sure the path is accessible inside
579 the chroot) and uid/gid (be sure the path is appropriately
580 writeable).
581
Willy Tarreaudad36a32013-03-11 01:20:04 +0100582 Any part of the address string may reference any number of environment
583 variables by preceding their name with a dollar sign ('$') and
584 optionally enclosing them with braces ('{}'), similarly to what is done
585 in Bourne shell.
586
Willy Tarreaudc2695c2014-06-27 18:10:07 +0200587 <length> is an optional maximum line length. Log lines larger than this value
588 will be truncated before being sent. The reason is that syslog
589 servers act differently on log line length. All servers support the
590 default value of 1024, but some servers simply drop larger lines
591 while others do log them. If a server supports long lines, it may
592 make sense to set this value here in order to avoid truncating long
593 lines. Similarly, if a server drops long lines, it is preferable to
594 truncate them before sending them. Accepted values are 80 to 65535
595 inclusive. The default value of 1024 is generally fine for all
596 standard usages. Some specific cases of long captures or
597 JSON-formated logs may require larger values.
598
Robert Tsai81ae1952007-12-05 10:47:29 +0100599 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200600
601 kern user mail daemon auth syslog lpr news
602 uucp cron auth2 ftp ntp audit alert cron2
603 local0 local1 local2 local3 local4 local5 local6 local7
604
605 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +0200606 all messages are sent. If a maximum level is specified, only messages with a
607 severity at least as important as this level will be sent. An optional minimum
608 level can be specified. If it is set, logs emitted with a more severe level
609 than this one will be capped to this level. This is used to avoid sending
610 "emerg" messages on all terminals on some default syslog configurations.
611 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200612
Cyril Bontédc4d9032012-04-08 21:57:39 +0200613 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +0200614
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100615log-send-hostname [<string>]
616 Sets the hostname field in the syslog header. If optional "string" parameter
617 is set the header is set to the string contents, otherwise uses the hostname
618 of the system. Generally used if one is not relaying logs through an
619 intermediate syslog server or for simply customizing the hostname printed in
620 the logs.
621
Kevinm48936af2010-12-22 16:08:21 +0000622log-tag <string>
623 Sets the tag field in the syslog header to this string. It defaults to the
624 program name as launched from the command line, which usually is "haproxy".
625 Sometimes it can be useful to differentiate between multiple processes
626 running on the same host.
627
Willy Tarreau6a06a402007-07-15 20:15:28 +0200628nbproc <number>
629 Creates <number> processes when going daemon. This requires the "daemon"
630 mode. By default, only one process is created, which is the recommended mode
631 of operation. For systems limited to small sets of file descriptors per
632 process, it may be needed to fork multiple daemons. USING MULTIPLE PROCESSES
633 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. See also "daemon".
634
635pidfile <pidfile>
636 Writes pids of all daemons into file <pidfile>. This option is equivalent to
637 the "-p" command line argument. The file must be accessible to the user
638 starting the process. See also "daemon".
639
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100640stats bind-process [ all | odd | even | <number 1-64>[-<number 1-64>] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +0200641 Limits the stats socket to a certain set of processes numbers. By default the
642 stats socket is bound to all processes, causing a warning to be emitted when
643 nbproc is greater than 1 because there is no way to select the target process
644 when connecting. However, by using this setting, it becomes possible to pin
645 the stats socket to a specific set of processes, typically the first one. The
646 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100647 the number of processes used. The maximum process ID depends on the machine's
Willy Tarreauae302532014-05-07 19:22:24 +0200648 word size (32 or 64). A better option consists in using the "process" setting
649 of the "stats socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +0200650
Willy Tarreau610f04b2014-02-13 11:36:41 +0100651ssl-default-bind-ciphers <ciphers>
652 This setting is only available when support for OpenSSL was built in. It sets
653 the default string describing the list of cipher algorithms ("cipher suite")
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300654 that are negotiated during the SSL/TLS handshake for all "bind" lines which
Willy Tarreau610f04b2014-02-13 11:36:41 +0100655 do not explicitly define theirs. The format of the string is defined in
656 "man 1 ciphers" from OpenSSL man pages, and can be for instance a string such
657 as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes). Please check the
658 "bind" keyword for more information.
659
660ssl-default-server-ciphers <ciphers>
661 This setting is only available when support for OpenSSL was built in. It
662 sets the default string describing the list of cipher algorithms that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300663 negotiated during the SSL/TLS handshake with the server, for all "server"
Willy Tarreau610f04b2014-02-13 11:36:41 +0100664 lines which do not explicitly define theirs. The format of the string is
665 defined in "man 1 ciphers". Please check the "server" keyword for more
666 information.
667
Emeric Brun850efd52014-01-29 12:24:34 +0100668ssl-server-verify [none|required]
669 The default behavior for SSL verify on servers side. If specified to 'none',
670 servers certificates are not verified. The default is 'required' except if
671 forced using cmdline option '-dV'.
672
Willy Tarreauabb175f2012-09-24 12:43:26 +0200673stats socket [<address:port>|<path>] [param*]
674 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
675 Connections to this socket will return various statistics outputs and even
676 allow some commands to be issued to change some runtime settings. Please
677 consult section 9.2 "Unix Socket commands" for more details.
Willy Tarreau6162db22009-10-10 17:13:00 +0200678
Willy Tarreauabb175f2012-09-24 12:43:26 +0200679 All parameters supported by "bind" lines are supported, for instance to
680 restrict access to some users or their access rights. Please consult
681 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +0200682
683stats timeout <timeout, in milliseconds>
684 The default timeout on the stats socket is set to 10 seconds. It is possible
685 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +0100686 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +0200687
688stats maxconn <connections>
689 By default, the stats socket is limited to 10 concurrent connections. It is
690 possible to change this value with "stats maxconn".
691
Willy Tarreau6a06a402007-07-15 20:15:28 +0200692uid <number>
693 Changes the process' user ID to <number>. It is recommended that the user ID
694 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
695 be started with superuser privileges in order to be able to switch to another
696 one. See also "gid" and "user".
697
698ulimit-n <number>
699 Sets the maximum number of per-process file-descriptors to <number>. By
700 default, it is automatically computed, so it is recommended not to use this
701 option.
702
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100703unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
704 [ group <group> ] [ gid <gid> ]
705
706 Fixes common settings to UNIX listening sockets declared in "bind" statements.
707 This is mainly used to simplify declaration of those UNIX sockets and reduce
708 the risk of errors, since those settings are most commonly required but are
709 also process-specific. The <prefix> setting can be used to force all socket
710 path to be relative to that directory. This might be needed to access another
711 component's chroot. Note that those paths are resolved before haproxy chroots
712 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
713 all have the same meaning as their homonyms used by the "bind" statement. If
714 both are specified, the "bind" statement has priority, meaning that the
715 "unix-bind" settings may be seen as process-wide default settings.
716
Willy Tarreau6a06a402007-07-15 20:15:28 +0200717user <user name>
718 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
719 See also "uid" and "group".
720
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +0200721node <name>
722 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
723
724 This statement is useful in HA configurations where two or more processes or
725 servers share the same IP address. By setting a different node-name on all
726 nodes, it becomes easy to immediately spot what server is handling the
727 traffic.
728
729description <text>
730 Add a text that describes the instance.
731
732 Please note that it is required to escape certain characters (# for example)
733 and this text is inserted into a html page so you should avoid using
734 "<" and ">" characters.
735
Willy Tarreau6a06a402007-07-15 20:15:28 +0200736
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007373.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +0200738-----------------------
739
Willy Tarreau1746eec2014-04-25 10:46:47 +0200740max-spread-checks <delay in milliseconds>
741 By default, haproxy tries to spread the start of health checks across the
742 smallest health check interval of all the servers in a farm. The principle is
743 to avoid hammering services running on the same server. But when using large
744 check intervals (10 seconds or more), the last servers in the farm take some
745 time before starting to be tested, which can be a problem. This parameter is
746 used to enforce an upper bound on delay between the first and the last check,
747 even if the servers' check intervals are larger. When servers run with
748 shorter intervals, their intervals will be respected though.
749
Willy Tarreau6a06a402007-07-15 20:15:28 +0200750maxconn <number>
751 Sets the maximum per-process number of concurrent connections to <number>. It
752 is equivalent to the command-line argument "-n". Proxies will stop accepting
753 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +0200754 automatically adjusted according to this value. See also "ulimit-n". Note:
755 the "select" poller cannot reliably use more than 1024 file descriptors on
756 some platforms. If your platform only supports select and reports "select
757 FAILED" on startup, you need to reduce maxconn until it works (slightly
758 below 500 in general).
Willy Tarreau6a06a402007-07-15 20:15:28 +0200759
Willy Tarreau81c25d02011-09-07 15:17:21 +0200760maxconnrate <number>
761 Sets the maximum per-process number of connections per second to <number>.
762 Proxies will stop accepting connections when this limit is reached. It can be
763 used to limit the global capacity regardless of each frontend capacity. It is
764 important to note that this can only be used as a service protection measure,
765 as there will not necessarily be a fair share between frontends when the
766 limit is reached, so it's a good idea to also limit each frontend to some
767 value close to its expected share. Also, lowering tune.maxaccept can improve
768 fairness.
769
William Lallemandd85f9172012-11-09 17:05:39 +0100770maxcomprate <number>
771 Sets the maximum per-process input compression rate to <number> kilobytes
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300772 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +0100773 level will be decreased during the session. If the maximum is reached at the
774 beginning of a session, the session will not compress at all. If the maximum
775 is not reached, the compression level will be increased up to
776 tune.comp.maxlevel. A value of zero means there is no limit, this is the
777 default value.
778
William Lallemand072a2bf2012-11-20 17:01:01 +0100779maxcompcpuusage <number>
780 Sets the maximum CPU usage HAProxy can reach before stopping the compression
781 for new requests or decreasing the compression level of current requests.
782 It works like 'maxcomprate' but measures CPU usage instead of incoming data
783 bandwidth. The value is expressed in percent of the CPU used by haproxy. In
784 case of multiple processes (nbproc > 1), each process manages its individual
785 usage. A value of 100 disable the limit. The default value is 100. Setting
786 a lower value will prevent the compression work from slowing the whole
787 process down and from introducing high latencies.
788
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100789maxpipes <number>
790 Sets the maximum per-process number of pipes to <number>. Currently, pipes
791 are only used by kernel-based tcp splicing. Since a pipe contains two file
792 descriptors, the "ulimit-n" value will be increased accordingly. The default
793 value is maxconn/4, which seems to be more than enough for most heavy usages.
794 The splice code dynamically allocates and releases pipes, and can fall back
795 to standard copy, so setting this value too low may only impact performance.
796
Willy Tarreau93e7c002013-10-07 18:51:07 +0200797maxsessrate <number>
798 Sets the maximum per-process number of sessions per second to <number>.
799 Proxies will stop accepting connections when this limit is reached. It can be
800 used to limit the global capacity regardless of each frontend capacity. It is
801 important to note that this can only be used as a service protection measure,
802 as there will not necessarily be a fair share between frontends when the
803 limit is reached, so it's a good idea to also limit each frontend to some
804 value close to its expected share. Also, lowering tune.maxaccept can improve
805 fairness.
806
Willy Tarreau403edff2012-09-06 11:58:37 +0200807maxsslconn <number>
808 Sets the maximum per-process number of concurrent SSL connections to
809 <number>. By default there is no SSL-specific limit, which means that the
810 global maxconn setting will apply to all connections. Setting this limit
811 avoids having openssl use too much memory and crash when malloc returns NULL
812 (since it unfortunately does not reliably check for such conditions). Note
813 that the limit applies both to incoming and outgoing connections, so one
814 connection which is deciphered then ciphered accounts for 2 SSL connections.
815
Willy Tarreaue43d5322013-10-07 20:01:52 +0200816maxsslrate <number>
817 Sets the maximum per-process number of SSL sessions per second to <number>.
818 SSL listeners will stop accepting connections when this limit is reached. It
819 can be used to limit the global SSL CPU usage regardless of each frontend
820 capacity. It is important to note that this can only be used as a service
821 protection measure, as there will not necessarily be a fair share between
822 frontends when the limit is reached, so it's a good idea to also limit each
823 frontend to some value close to its expected share. It is also important to
824 note that the sessions are accounted before they enter the SSL stack and not
825 after, which also protects the stack against bad handshakes. Also, lowering
826 tune.maxaccept can improve fairness.
827
William Lallemand9d5f5482012-11-07 16:12:57 +0100828maxzlibmem <number>
829 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
830 When the maximum amount is reached, future sessions will not compress as long
831 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +0100832 The default value is 0. The value is available in bytes on the UNIX socket
833 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
834 "ZlibMemUsage" in bytes.
835
Willy Tarreau6a06a402007-07-15 20:15:28 +0200836noepoll
837 Disables the use of the "epoll" event polling system on Linux. It is
838 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +0100839 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +0200840
841nokqueue
842 Disables the use of the "kqueue" event polling system on BSD. It is
843 equivalent to the command-line argument "-dk". The next polling system
844 used will generally be "poll". See also "nopoll".
845
846nopoll
847 Disables the use of the "poll" event polling system. It is equivalent to the
848 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +0100849 It should never be needed to disable "poll" since it's available on all
Willy Tarreaue9f49e72012-11-11 17:42:00 +0100850 platforms supported by HAProxy. See also "nokqueue" and "noepoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +0200851
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100852nosplice
853 Disables the use of kernel tcp splicing between sockets on Linux. It is
854 equivalent to the command line argument "-dS". Data will then be copied
855 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100856 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100857 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
858 be used. This option makes it easier to globally disable kernel splicing in
859 case of doubt. See also "option splice-auto", "option splice-request" and
860 "option splice-response".
861
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300862nogetaddrinfo
863 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
864 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
865
Willy Tarreaufe255b72007-10-14 23:09:26 +0200866spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +0900867 Sometimes it is desirable to avoid sending agent and health checks to
868 servers at exact intervals, for instance when many logical servers are
869 located on the same physical server. With the help of this parameter, it
870 becomes possible to add some randomness in the check interval between 0
871 and +/- 50%. A value between 2 and 5 seems to show good results. The
872 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +0200873
Willy Tarreau27a674e2009-08-17 07:23:33 +0200874tune.bufsize <number>
875 Sets the buffer size to this size (in bytes). Lower values allow more
876 sessions to coexist in the same amount of RAM, and higher values allow some
877 applications with very large cookies to work. The default value is 16384 and
878 can be changed at build time. It is strongly recommended not to change this
879 from the default value, as very low values will break some services such as
880 statistics, and values larger than default size will increase memory usage,
881 possibly causing the system to run out of memory. At least the global maxconn
882 parameter should be decreased by the same factor as this one is increased.
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +0400883 If HTTP request is larger than (tune.bufsize - tune.maxrewrite), haproxy will
884 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
885 than this size, haproxy will return HTTP 502 (Bad Gateway).
Willy Tarreau27a674e2009-08-17 07:23:33 +0200886
Willy Tarreau43961d52010-10-04 20:39:20 +0200887tune.chksize <number>
888 Sets the check buffer size to this size (in bytes). Higher values may help
889 find string or regex patterns in very large pages, though doing so may imply
890 more memory and CPU usage. The default value is 16384 and can be changed at
891 build time. It is not recommended to change this value, but to use better
892 checks whenever possible.
893
William Lallemandf3747832012-11-09 12:33:10 +0100894tune.comp.maxlevel <number>
895 Sets the maximum compression level. The compression level affects CPU
896 usage during compression. This value affects CPU usage during compression.
897 Each session using compression initializes the compression algorithm with
898 this value. The default value is 1.
899
Willy Tarreau193b8c62012-11-22 00:17:38 +0100900tune.http.cookielen <number>
901 Sets the maximum length of captured cookies. This is the maximum value that
902 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
903 will automatically be truncated to this one. It is important not to set too
904 high a value because all cookie captures still allocate this size whatever
905 their configured value (they share a same pool). This value is per request
906 per response, so the memory allocated is twice this value per connection.
907 When not specified, the limit is set to 63 characters. It is recommended not
908 to change this value.
909
Willy Tarreauac1932d2011-10-24 19:14:41 +0200910tune.http.maxhdr <number>
911 Sets the maximum number of headers in a request. When a request comes with a
912 number of headers greater than this value (including the first line), it is
913 rejected with a "400 Bad Request" status code. Similarly, too large responses
914 are blocked with "502 Bad Gateway". The default value is 101, which is enough
915 for all usages, considering that the widely deployed Apache server uses the
916 same limit. It can be useful to push this limit further to temporarily allow
917 a buggy application to work by the time it gets fixed. Keep in mind that each
918 new header consumes 32bits of memory for each session, so don't push this
919 limit too high.
920
Willy Tarreau7e312732014-02-12 16:35:14 +0100921tune.idletimer <timeout>
922 Sets the duration after which haproxy will consider that an empty buffer is
923 probably associated with an idle stream. This is used to optimally adjust
924 some packet sizes while forwarding large and small data alternatively. The
925 decision to use splice() or to send large buffers in SSL is modulated by this
926 parameter. The value is in milliseconds between 0 and 65535. A value of zero
927 means that haproxy will not try to detect idle streams. The default is 1000,
928 which seems to correctly detect end user pauses (eg: read a page before
929 clicking). There should be not reason for changing this value. Please check
930 tune.ssl.maxrecord below.
931
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100932tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +0100933 Sets the maximum number of consecutive connections a process may accept in a
934 row before switching to other work. In single process mode, higher numbers
935 give better performance at high connection rates. However in multi-process
936 modes, keeping a bit of fairness between processes generally is better to
937 increase performance. This value applies individually to each listener, so
938 that the number of processes a listener is bound to is taken into account.
939 This value defaults to 64. In multi-process mode, it is divided by twice
940 the number of processes the listener is bound to. Setting this value to -1
941 completely disables the limitation. It should normally not be needed to tweak
942 this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100943
944tune.maxpollevents <number>
945 Sets the maximum amount of events that can be processed at once in a call to
946 the polling system. The default value is adapted to the operating system. It
947 has been noticed that reducing it below 200 tends to slightly decrease
948 latency at the expense of network bandwidth, and increasing it above 200
949 tends to trade latency for slightly increased bandwidth.
950
Willy Tarreau27a674e2009-08-17 07:23:33 +0200951tune.maxrewrite <number>
952 Sets the reserved buffer space to this size in bytes. The reserved space is
953 used for header rewriting or appending. The first reads on sockets will never
954 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
955 bufsize, though that does not make much sense since there are rarely large
956 numbers of headers to add. Setting it too high prevents processing of large
957 requests or responses. Setting it too low prevents addition of new headers
958 to already large requests or to POST requests. It is generally wise to set it
959 to about 1024. It is automatically readjusted to half of bufsize if it is
960 larger than that. This means you don't have to worry about it when changing
961 bufsize.
962
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200963tune.pipesize <number>
964 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
965 are the default size for the system. But sometimes when using TCP splicing,
966 it can improve performance to increase pipe sizes, especially if it is
967 suspected that pipes are not filled and that many calls to splice() are
968 performed. This has an impact on the kernel's memory footprint, so this must
969 not be changed if impacts are not understood.
970
Willy Tarreaue803de22010-01-21 17:43:04 +0100971tune.rcvbuf.client <number>
972tune.rcvbuf.server <number>
973 Forces the kernel socket receive buffer size on the client or the server side
974 to the specified value in bytes. This value applies to all TCP/HTTP frontends
975 and backends. It should normally never be set, and the default size (0) lets
976 the kernel autotune this value depending on the amount of available memory.
977 However it can sometimes help to set it to very low values (eg: 4096) in
978 order to save kernel memory by preventing it from buffering too large amounts
979 of received data. Lower values will significantly increase CPU usage though.
980
981tune.sndbuf.client <number>
982tune.sndbuf.server <number>
983 Forces the kernel socket send buffer size on the client or the server side to
984 the specified value in bytes. This value applies to all TCP/HTTP frontends
985 and backends. It should normally never be set, and the default size (0) lets
986 the kernel autotune this value depending on the amount of available memory.
987 However it can sometimes help to set it to very low values (eg: 4096) in
988 order to save kernel memory by preventing it from buffering too large amounts
989 of received data. Lower values will significantly increase CPU usage though.
990 Another use case is to prevent write timeouts with extremely slow clients due
991 to the kernel waiting for a large part of the buffer to be read before
992 notifying haproxy again.
993
Willy Tarreau6ec58db2012-11-16 16:32:15 +0100994tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +0100995 Sets the size of the global SSL session cache, in a number of blocks. A block
996 is large enough to contain an encoded session without peer certificate.
997 An encoded session with peer certificate is stored in multiple blocks
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300998 depending on the size of the peer certificate. A block uses approximately
Emeric Brunaf9619d2012-11-28 18:47:52 +0100999 200 bytes of memory. The default value may be forced at build time, otherwise
1000 defaults to 20000. When the cache is full, the most idle entries are purged
1001 and reassigned. Higher values reduce the occurrence of such a purge, hence
1002 the number of CPU-intensive SSL handshakes by ensuring that all users keep
1003 their session as long as possible. All entries are pre-allocated upon startup
Emeric Brun22890a12012-12-28 14:41:32 +01001004 and are shared between all processes if "nbproc" is greater than 1. Setting
1005 this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001006
Emeric Brun8dc60392014-05-09 13:52:00 +02001007tune.ssl.force-private-cache
1008 This boolean disables SSL session cache sharing between all processes. It
1009 should normally not be used since it will force many renegotiations due to
1010 clients hitting a random process. But it may be required on some operating
1011 systems where none of the SSL cache synchronization method may be used. In
1012 this case, adding a first layer of hash-based load balancing before the SSL
1013 layer might limit the impact of the lack of session sharing.
1014
Emeric Brun4f65bff2012-11-16 15:11:00 +01001015tune.ssl.lifetime <timeout>
1016 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001017 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01001018 does not guarantee that sessions will last that long, because if the cache is
1019 full, the longest idle sessions will be purged despite their configured
1020 lifetime. The real usefulness of this setting is to prevent sessions from
1021 being used for too long.
1022
Willy Tarreaubfd59462013-02-21 07:46:09 +01001023tune.ssl.maxrecord <number>
1024 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
1025 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
1026 data only once it has received a full record. With large records, it means
1027 that clients might have to download up to 16kB of data before starting to
1028 process them. Limiting the value can improve page load times on browsers
1029 located over high latency or low bandwidth networks. It is suggested to find
1030 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
1031 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
1032 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
1033 2859 gave good results during tests. Use "strace -e trace=write" to find the
Willy Tarreau7e312732014-02-12 16:35:14 +01001034 best value. Haproxy will automatically switch to this setting after an idle
1035 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01001036
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001037tune.ssl.default-dh-param <number>
1038 Sets the maximum size of the Diffie-Hellman parameters used for generating
1039 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
1040 final size will try to match the size of the server's RSA (or DSA) key (e.g,
1041 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
1042 this maximum value. Default value if 1024. Only 1024 or higher values are
1043 allowed. Higher values will increase the CPU load, and values greater than
1044 1024 bits are not supported by Java 7 and earlier clients. This value is not
1045 used if static Diffie-Hellman parameters are supplied via the certificate file.
1046
William Lallemanda509e4c2012-11-07 16:54:34 +01001047tune.zlib.memlevel <number>
1048 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001049 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01001050 state. A value of 1 uses minimum memory but is slow and reduces compression
1051 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
1052 between 1 and 9. The default value is 8.
1053
1054tune.zlib.windowsize <number>
1055 Sets the window size (the size of the history buffer) as a parameter of the
1056 zlib initialization for each session. Larger values of this parameter result
1057 in better compression at the expense of memory usage. Can be a value between
1058 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001059
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010603.3. Debugging
1061--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02001062
1063debug
1064 Enables debug mode which dumps to stdout all exchanges, and disables forking
1065 into background. It is the equivalent of the command-line argument "-d". It
1066 should never be used in a production configuration since it may prevent full
1067 system startup.
1068
1069quiet
1070 Do not display any message during startup. It is equivalent to the command-
1071 line argument "-q".
1072
Emeric Brunf099e792010-09-27 12:05:28 +02001073
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010010743.4. Userlists
1075--------------
1076It is possible to control access to frontend/backend/listen sections or to
1077http stats by allowing only authenticated and authorized users. To do this,
1078it is required to create at least one userlist and to define users.
1079
1080userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01001081 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001082 used to store authentication & authorization data for independent customers.
1083
1084group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01001085 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001086 attach users to this group by using a comma separated list of names
1087 proceeded by "users" keyword.
1088
Cyril Bontéf0c60612010-02-06 14:44:47 +01001089user <username> [password|insecure-password <password>]
1090 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001091 Adds user <username> to the current userlist. Both secure (encrypted) and
1092 insecure (unencrypted) passwords can be used. Encrypted passwords are
Cyril Bonté78caf842010-03-10 22:41:43 +01001093 evaluated using the crypt(3) function so depending of the system's
1094 capabilities, different algorithms are supported. For example modern Glibc
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001095 based Linux system supports MD5, SHA-256, SHA-512 and of course classic,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001096 DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001097
1098
1099 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01001100 userlist L1
1101 group G1 users tiger,scott
1102 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001103
Cyril Bontéf0c60612010-02-06 14:44:47 +01001104 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
1105 user scott insecure-password elgato
1106 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001107
Cyril Bontéf0c60612010-02-06 14:44:47 +01001108 userlist L2
1109 group G1
1110 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001111
Cyril Bontéf0c60612010-02-06 14:44:47 +01001112 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
1113 user scott insecure-password elgato groups G1,G2
1114 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001115
1116 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001117
Emeric Brunf099e792010-09-27 12:05:28 +02001118
11193.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02001120----------
Emeric Brunf099e792010-09-27 12:05:28 +02001121It is possible to synchronize server entries in stick tables between several
1122haproxy instances over TCP connections in a multi-master fashion. Each instance
1123pushes its local updates and insertions to remote peers. Server IDs are used to
1124identify servers remotely, so it is important that configurations look similar
1125or at least that the same IDs are forced on each server on all participants.
1126Interrupted exchanges are automatically detected and recovered from the last
1127known point. In addition, during a soft restart, the old process connects to
1128the new one using such a TCP connection to push all its entries before the new
1129process tries to connect to other peers. That ensures very fast replication
1130during a reload, it typically takes a fraction of a second even for large
1131tables.
1132
1133peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001134 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02001135 which is referenced by one or more stick-tables.
1136
1137peer <peername> <ip>:<port>
1138 Defines a peer inside a peers section.
1139 If <peername> is set to the local peer name (by default hostname, or forced
1140 using "-L" command line option), haproxy will listen for incoming remote peer
1141 connection on <ip>:<port>. Otherwise, <ip>:<port> defines where to connect to
1142 to join the remote peer, and <peername> is used at the protocol level to
1143 identify and validate the remote peer on the server side.
1144
1145 During a soft restart, local peer <ip>:<port> is used by the old instance to
1146 connect the new one and initiate a complete replication (teaching process).
1147
1148 It is strongly recommended to have the exact same peers declaration on all
1149 peers and to only rely on the "-L" command line argument to change the local
1150 peer name. This makes it easier to maintain coherent configuration files
1151 across all peers.
1152
Willy Tarreaudad36a32013-03-11 01:20:04 +01001153 Any part of the address string may reference any number of environment
1154 variables by preceding their name with a dollar sign ('$') and optionally
1155 enclosing them with braces ('{}'), similarly to what is done in Bourne shell.
1156
Cyril Bontédc4d9032012-04-08 21:57:39 +02001157 Example:
Emeric Brunf099e792010-09-27 12:05:28 +02001158 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001159 peer haproxy1 192.168.0.1:1024
1160 peer haproxy2 192.168.0.2:1024
1161 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02001162
1163 backend mybackend
1164 mode tcp
1165 balance roundrobin
1166 stick-table type ip size 20k peers mypeers
1167 stick on src
1168
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001169 server srv1 192.168.0.30:80
1170 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02001171
1172
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011734. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02001174----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001175
Willy Tarreau6a06a402007-07-15 20:15:28 +02001176Proxy configuration can be located in a set of sections :
1177 - defaults <name>
1178 - frontend <name>
1179 - backend <name>
1180 - listen <name>
1181
1182A "defaults" section sets default parameters for all other sections following
1183its declaration. Those default parameters are reset by the next "defaults"
1184section. See below for the list of parameters which can be set in a "defaults"
Willy Tarreau0ba27502007-12-24 16:55:16 +01001185section. The name is optional but its use is encouraged for better readability.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001186
1187A "frontend" section describes a set of listening sockets accepting client
1188connections.
1189
1190A "backend" section describes a set of servers to which the proxy will connect
1191to forward incoming connections.
1192
1193A "listen" section defines a complete proxy with its frontend and backend
1194parts combined in one section. It is generally useful for TCP-only traffic.
1195
Willy Tarreau0ba27502007-12-24 16:55:16 +01001196All proxy names must be formed from upper and lower case letters, digits,
1197'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
1198case-sensitive, which means that "www" and "WWW" are two different proxies.
1199
1200Historically, all proxy names could overlap, it just caused troubles in the
1201logs. Since the introduction of content switching, it is mandatory that two
1202proxies with overlapping capabilities (frontend/backend) have different names.
1203However, it is still permitted that a frontend and a backend share the same
1204name, as this configuration seems to be commonly encountered.
1205
1206Right now, two major proxy modes are supported : "tcp", also known as layer 4,
1207and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001208bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01001209protocol, and can interact with it by allowing, blocking, switching, adding,
1210modifying, or removing arbitrary contents in requests or responses, based on
1211arbitrary criteria.
1212
Willy Tarreau70dffda2014-01-30 03:07:23 +01001213In HTTP mode, the processing applied to requests and responses flowing over
1214a connection depends in the combination of the frontend's HTTP options and
1215the backend's. HAProxy supports 5 connection modes :
1216
1217 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
1218 requests and responses are processed, and connections remain open but idle
1219 between responses and new requests.
1220
1221 - TUN: tunnel ("option http-tunnel") : this was the default mode for versions
1222 1.0 to 1.5-dev21 : only the first request and response are processed, and
1223 everything else is forwarded with no analysis at all. This mode should not
1224 be used as it creates lots of trouble with logging and HTTP processing.
1225
1226 - PCL: passive close ("option httpclose") : exactly the same as tunnel mode,
1227 but with "Connection: close" appended in both directions to try to make
1228 both ends close after the first request/response exchange.
1229
1230 - SCL: server close ("option http-server-close") : the server-facing
1231 connection is closed after the end of the response is received, but the
1232 client-facing connection remains open.
1233
1234 - FCL: forced close ("option forceclose") : the connection is actively closed
1235 after the end of the response.
1236
1237The effective mode that will be applied to a connection passing through a
1238frontend and a backend can be determined by both proxy modes according to the
1239following matrix, but in short, the modes are symmetric, keep-alive is the
1240weakest option and force close is the strongest.
1241
1242 Backend mode
1243
1244 | KAL | TUN | PCL | SCL | FCL
1245 ----+-----+-----+-----+-----+----
1246 KAL | KAL | TUN | PCL | SCL | FCL
1247 ----+-----+-----+-----+-----+----
1248 TUN | TUN | TUN | PCL | SCL | FCL
1249 Frontend ----+-----+-----+-----+-----+----
1250 mode PCL | PCL | PCL | PCL | FCL | FCL
1251 ----+-----+-----+-----+-----+----
1252 SCL | SCL | SCL | FCL | SCL | FCL
1253 ----+-----+-----+-----+-----+----
1254 FCL | FCL | FCL | FCL | FCL | FCL
1255
Willy Tarreau0ba27502007-12-24 16:55:16 +01001256
Willy Tarreau70dffda2014-01-30 03:07:23 +01001257
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012584.1. Proxy keywords matrix
1259--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001260
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001261The following list of keywords is supported. Most of them may only be used in a
1262limited set of section types. Some of them are marked as "deprecated" because
1263they are inherited from an old syntax which may be confusing or functionally
1264limited, and there are new recommended keywords to replace them. Keywords
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001265marked with "(*)" can be optionally inverted using the "no" prefix, eg. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001266option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02001267and must be disabled for a specific instance. Such options may also be prefixed
1268with "default" in order to restore default settings regardless of what has been
1269specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001270
Willy Tarreau6a06a402007-07-15 20:15:28 +02001271
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001272 keyword defaults frontend listen backend
1273------------------------------------+----------+----------+---------+---------
1274acl - X X X
1275appsession - - X X
1276backlog X X X -
1277balance X - X X
1278bind - X X -
1279bind-process X X X X
1280block - X X X
1281capture cookie - X X -
1282capture request header - X X -
1283capture response header - X X -
1284clitimeout (deprecated) X X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02001285compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001286contimeout (deprecated) X - X X
1287cookie X - X X
1288default-server X - X X
1289default_backend X X X -
1290description - X X X
1291disabled X X X X
1292dispatch - - X X
1293enabled X X X X
1294errorfile X X X X
1295errorloc X X X X
1296errorloc302 X X X X
1297-- keyword -------------------------- defaults - frontend - listen -- backend -
1298errorloc303 X X X X
Cyril Bonté0d4bf012010-04-25 23:21:46 +02001299force-persist - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001300fullconn X - X X
1301grace X X X X
1302hash-type X - X X
1303http-check disable-on-404 X - X X
Willy Tarreaubd741542010-03-16 18:46:54 +01001304http-check expect - - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02001305http-check send-state X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001306http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02001307http-response - X X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02001308http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001309id - X X X
Cyril Bonté0d4bf012010-04-25 23:21:46 +02001310ignore-persist - X X X
William Lallemand0f99e342011-10-12 17:50:54 +02001311log (*) X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02001312max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001313maxconn X X X -
1314mode X X X X
1315monitor fail - X X -
1316monitor-net X X X -
1317monitor-uri X X X -
1318option abortonclose (*) X - X X
1319option accept-invalid-http-request (*) X X X -
1320option accept-invalid-http-response (*) X - X X
1321option allbackups (*) X - X X
1322option checkcache (*) X - X X
1323option clitcpka (*) X X X -
1324option contstats (*) X X X -
1325option dontlog-normal (*) X X X -
1326option dontlognull (*) X X X -
1327option forceclose (*) X X X X
1328-- keyword -------------------------- defaults - frontend - listen -- backend -
1329option forwardfor X X X X
Willy Tarreau16bfb022010-01-16 19:48:41 +01001330option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02001331option http-no-delay (*) X X X X
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02001332option http-pretend-keepalive (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001333option http-server-close (*) X X X X
Willy Tarreau02bce8b2014-01-30 00:15:28 +01001334option http-tunnel (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001335option http-use-proxy-header (*) X X X -
1336option httpchk X - X X
1337option httpclose (*) X X X X
1338option httplog X X X X
1339option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001340option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02001341option ldap-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001342option log-health-checks (*) X - X X
1343option log-separate-errors (*) X X X -
1344option logasap (*) X X X -
1345option mysql-check X - X X
Rauf Kuliyev38b41562011-01-04 15:14:13 +01001346option pgsql-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001347option nolinger (*) X X X X
1348option originalto X X X X
1349option persist (*) X - X X
1350option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02001351option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001352option smtpchk X - X X
1353option socket-stats (*) X X X -
1354option splice-auto (*) X X X X
1355option splice-request (*) X X X X
1356option splice-response (*) X X X X
1357option srvtcpka (*) X - X X
1358option ssl-hello-chk X - X X
1359-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01001360option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001361option tcp-smart-accept (*) X X X -
1362option tcp-smart-connect (*) X - X X
1363option tcpka X X X X
1364option tcplog X X X X
1365option transparent (*) X - X X
1366persist rdp-cookie X - X X
1367rate-limit sessions X X X -
1368redirect - X X X
1369redisp (deprecated) X - X X
1370redispatch (deprecated) X - X X
1371reqadd - X X X
1372reqallow - X X X
1373reqdel - X X X
1374reqdeny - X X X
1375reqiallow - X X X
1376reqidel - X X X
1377reqideny - X X X
1378reqipass - X X X
1379reqirep - X X X
1380reqisetbe - X X X
1381reqitarpit - X X X
1382reqpass - X X X
1383reqrep - X X X
1384-- keyword -------------------------- defaults - frontend - listen -- backend -
1385reqsetbe - X X X
1386reqtarpit - X X X
1387retries X - X X
1388rspadd - X X X
1389rspdel - X X X
1390rspdeny - X X X
1391rspidel - X X X
1392rspideny - X X X
1393rspirep - X X X
1394rsprep - X X X
1395server - - X X
1396source X - X X
1397srvtimeout (deprecated) X - X X
Cyril Bonté66c327d2010-10-12 00:14:37 +02001398stats admin - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001399stats auth X - X X
1400stats enable X - X X
1401stats hide-version X - X X
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02001402stats http-request - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001403stats realm X - X X
1404stats refresh X - X X
1405stats scope X - X X
1406stats show-desc X - X X
1407stats show-legends X - X X
1408stats show-node X - X X
1409stats uri X - X X
1410-- keyword -------------------------- defaults - frontend - listen -- backend -
1411stick match - - X X
1412stick on - - X X
1413stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02001414stick store-response - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001415stick-table - - X X
Willy Tarreau938c7fe2014-04-25 14:21:39 +02001416tcp-check connect - - X X
1417tcp-check expect - - X X
1418tcp-check send - - X X
1419tcp-check send-binary - - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02001420tcp-request connection - X X -
1421tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02001422tcp-request inspect-delay - X X X
Emeric Brun0a3b67f2010-09-24 15:34:53 +02001423tcp-response content - - X X
1424tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001425timeout check X - X X
1426timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02001427timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001428timeout clitimeout (deprecated) X X X -
1429timeout connect X - X X
1430timeout contimeout (deprecated) X - X X
1431timeout http-keep-alive X X X X
1432timeout http-request X X X X
1433timeout queue X - X X
1434timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02001435timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001436timeout srvtimeout (deprecated) X - X X
1437timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02001438timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001439transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01001440unique-id-format X X X -
1441unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001442use_backend - X X -
Willy Tarreau4a5cade2012-04-05 21:09:48 +02001443use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001444------------------------------------+----------+----------+---------+---------
1445 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02001446
Willy Tarreau0ba27502007-12-24 16:55:16 +01001447
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014484.2. Alphabetically sorted keywords reference
1449---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001450
1451This section provides a description of each keyword and its usage.
1452
1453
1454acl <aclname> <criterion> [flags] [operator] <value> ...
1455 Declare or complete an access list.
1456 May be used in sections : defaults | frontend | listen | backend
1457 no | yes | yes | yes
1458 Example:
1459 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
1460 acl invalid_src src_port 0:1023
1461 acl local_dst hdr(host) -i localhost
1462
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001463 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001464
1465
Cyril Bontéb21570a2009-11-29 20:04:48 +01001466appsession <cookie> len <length> timeout <holdtime>
1467 [request-learn] [prefix] [mode <path-parameters|query-string>]
Willy Tarreau0ba27502007-12-24 16:55:16 +01001468 Define session stickiness on an existing application cookie.
1469 May be used in sections : defaults | frontend | listen | backend
1470 no | no | yes | yes
1471 Arguments :
1472 <cookie> this is the name of the cookie used by the application and which
1473 HAProxy will have to learn for each new session.
1474
Cyril Bontéb21570a2009-11-29 20:04:48 +01001475 <length> this is the max number of characters that will be memorized and
Willy Tarreau0ba27502007-12-24 16:55:16 +01001476 checked in each cookie value.
1477
1478 <holdtime> this is the time after which the cookie will be removed from
1479 memory if unused. If no unit is specified, this time is in
1480 milliseconds.
1481
Cyril Bontébf47aeb2009-10-15 00:15:40 +02001482 request-learn
1483 If this option is specified, then haproxy will be able to learn
1484 the cookie found in the request in case the server does not
1485 specify any in response. This is typically what happens with
1486 PHPSESSID cookies, or when haproxy's session expires before
1487 the application's session and the correct server is selected.
1488 It is recommended to specify this option to improve reliability.
1489
Cyril Bontéb21570a2009-11-29 20:04:48 +01001490 prefix When this option is specified, haproxy will match on the cookie
1491 prefix (or URL parameter prefix). The appsession value is the
1492 data following this prefix.
1493
1494 Example :
1495 appsession ASPSESSIONID len 64 timeout 3h prefix
1496
1497 This will match the cookie ASPSESSIONIDXXXX=XXXXX,
1498 the appsession value will be XXXX=XXXXX.
1499
1500 mode This option allows to change the URL parser mode.
1501 2 modes are currently supported :
1502 - path-parameters :
1503 The parser looks for the appsession in the path parameters
1504 part (each parameter is separated by a semi-colon), which is
1505 convenient for JSESSIONID for example.
1506 This is the default mode if the option is not set.
1507 - query-string :
1508 In this mode, the parser will look for the appsession in the
1509 query string.
1510
Willy Tarreau0ba27502007-12-24 16:55:16 +01001511 When an application cookie is defined in a backend, HAProxy will check when
1512 the server sets such a cookie, and will store its value in a table, and
1513 associate it with the server's identifier. Up to <length> characters from
1514 the value will be retained. On each connection, haproxy will look for this
Cyril Bontéb21570a2009-11-29 20:04:48 +01001515 cookie both in the "Cookie:" headers, and as a URL parameter (depending on
1516 the mode used). If a known value is found, the client will be directed to the
1517 server associated with this value. Otherwise, the load balancing algorithm is
Willy Tarreau0ba27502007-12-24 16:55:16 +01001518 applied. Cookies are automatically removed from memory when they have been
1519 unused for a duration longer than <holdtime>.
1520
1521 The definition of an application cookie is limited to one per backend.
1522
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01001523 Note : Consider not using this feature in multi-process mode (nbproc > 1)
1524 unless you know what you do : memory is not shared between the
1525 processes, which can result in random behaviours.
1526
Willy Tarreau0ba27502007-12-24 16:55:16 +01001527 Example :
1528 appsession JSESSIONID len 52 timeout 3h
1529
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01001530 See also : "cookie", "capture cookie", "balance", "stick", "stick-table",
1531 "ignore-persist", "nbproc" and "bind-process".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001532
1533
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01001534backlog <conns>
1535 Give hints to the system about the approximate listen backlog desired size
1536 May be used in sections : defaults | frontend | listen | backend
1537 yes | yes | yes | no
1538 Arguments :
1539 <conns> is the number of pending connections. Depending on the operating
1540 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02001541 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01001542
1543 In order to protect against SYN flood attacks, one solution is to increase
1544 the system's SYN backlog size. Depending on the system, sometimes it is just
1545 tunable via a system parameter, sometimes it is not adjustable at all, and
1546 sometimes the system relies on hints given by the application at the time of
1547 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
1548 to the listen() syscall. On systems which can make use of this value, it can
1549 sometimes be useful to be able to specify a different value, hence this
1550 backlog parameter.
1551
1552 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
1553 used as a hint and the system accepts up to the smallest greater power of
1554 two, and never more than some limits (usually 32768).
1555
1556 See also : "maxconn" and the target operating system's tuning guide.
1557
1558
Willy Tarreau0ba27502007-12-24 16:55:16 +01001559balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02001560balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01001561 Define the load balancing algorithm to be used in a backend.
1562 May be used in sections : defaults | frontend | listen | backend
1563 yes | no | yes | yes
1564 Arguments :
1565 <algorithm> is the algorithm used to select a server when doing load
1566 balancing. This only applies when no persistence information
1567 is available, or when a connection is redispatched to another
1568 server. <algorithm> may be one of the following :
1569
1570 roundrobin Each server is used in turns, according to their weights.
1571 This is the smoothest and fairest algorithm when the server's
1572 processing time remains equally distributed. This algorithm
1573 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02001574 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08001575 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02001576 large farms, when a server becomes up after having been down
1577 for a very short time, it may sometimes take a few hundreds
1578 requests for it to be re-integrated into the farm and start
1579 receiving traffic. This is normal, though very rare. It is
1580 indicated here in case you would have the chance to observe
1581 it, so that you don't worry.
1582
1583 static-rr Each server is used in turns, according to their weights.
1584 This algorithm is as similar to roundrobin except that it is
1585 static, which means that changing a server's weight on the
1586 fly will have no effect. On the other hand, it has no design
1587 limitation on the number of servers, and when a server goes
1588 up, it is always immediately reintroduced into the farm, once
1589 the full map is recomputed. It also uses slightly less CPU to
1590 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01001591
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01001592 leastconn The server with the lowest number of connections receives the
1593 connection. Round-robin is performed within groups of servers
1594 of the same load to ensure that all servers will be used. Use
1595 of this algorithm is recommended where very long sessions are
1596 expected, such as LDAP, SQL, TSE, etc... but is not very well
1597 suited for protocols using short sessions such as HTTP. This
1598 algorithm is dynamic, which means that server weights may be
1599 adjusted on the fly for slow starts for instance.
1600
Willy Tarreauf09c6602012-02-13 17:12:08 +01001601 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001602 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01001603 identifier to the highest (see server parameter "id"), which
1604 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02001605 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01001606 not make sense to use this algorithm without setting maxconn.
1607 The purpose of this algorithm is to always use the smallest
1608 number of servers so that extra servers can be powered off
1609 during non-intensive hours. This algorithm ignores the server
1610 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02001611 or IMAP than HTTP, though it can be useful there too. In
1612 order to use this algorithm efficiently, it is recommended
1613 that a cloud controller regularly checks server usage to turn
1614 them off when unused, and regularly checks backend queue to
1615 turn new servers on when the queue inflates. Alternatively,
1616 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01001617
Willy Tarreau0ba27502007-12-24 16:55:16 +01001618 source The source IP address is hashed and divided by the total
1619 weight of the running servers to designate which server will
1620 receive the request. This ensures that the same client IP
1621 address will always reach the same server as long as no
1622 server goes down or up. If the hash result changes due to the
1623 number of running servers changing, many clients will be
1624 directed to a different server. This algorithm is generally
1625 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001626 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01001627 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02001628 static by default, which means that changing a server's
1629 weight on the fly will have no effect, but this can be
1630 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001631
Oskar Stolc8dc41842012-05-19 10:19:54 +01001632 uri This algorithm hashes either the left part of the URI (before
1633 the question mark) or the whole URI (if the "whole" parameter
1634 is present) and divides the hash value by the total weight of
1635 the running servers. The result designates which server will
1636 receive the request. This ensures that the same URI will
1637 always be directed to the same server as long as no server
1638 goes up or down. This is used with proxy caches and
1639 anti-virus proxies in order to maximize the cache hit rate.
1640 Note that this algorithm may only be used in an HTTP backend.
1641 This algorithm is static by default, which means that
1642 changing a server's weight on the fly will have no effect,
1643 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001644
Oskar Stolc8dc41842012-05-19 10:19:54 +01001645 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02001646 "depth", both followed by a positive integer number. These
1647 options may be helpful when it is needed to balance servers
1648 based on the beginning of the URI only. The "len" parameter
1649 indicates that the algorithm should only consider that many
1650 characters at the beginning of the URI to compute the hash.
1651 Note that having "len" set to 1 rarely makes sense since most
1652 URIs start with a leading "/".
1653
1654 The "depth" parameter indicates the maximum directory depth
1655 to be used to compute the hash. One level is counted for each
1656 slash in the request. If both parameters are specified, the
1657 evaluation stops when either is reached.
1658
Willy Tarreau0ba27502007-12-24 16:55:16 +01001659 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001660 the query string of each HTTP GET request.
1661
1662 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02001663 request entity will be searched for the parameter argument,
1664 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02001665 ('?') in the URL. The message body will only start to be
1666 analyzed once either the advertised amount of data has been
1667 received or the request buffer is full. In the unlikely event
1668 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02001669 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02001670 be randomly balanced if at all. This keyword used to support
1671 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001672
1673 If the parameter is found followed by an equal sign ('=') and
1674 a value, then the value is hashed and divided by the total
1675 weight of the running servers. The result designates which
1676 server will receive the request.
1677
1678 This is used to track user identifiers in requests and ensure
1679 that a same user ID will always be sent to the same server as
1680 long as no server goes up or down. If no value is found or if
1681 the parameter is not found, then a round robin algorithm is
1682 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02001683 backend. This algorithm is static by default, which means
1684 that changing a server's weight on the fly will have no
1685 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001686
Cyril Bontédc4d9032012-04-08 21:57:39 +02001687 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
1688 request. Just as with the equivalent ACL 'hdr()' function,
1689 the header name in parenthesis is not case sensitive. If the
1690 header is absent or if it does not contain any value, the
1691 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01001692
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001693 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01001694 reducing the hash algorithm to the main domain part with some
1695 specific headers such as 'Host'. For instance, in the Host
1696 value "haproxy.1wt.eu", only "1wt" will be considered.
1697
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02001698 This algorithm is static by default, which means that
1699 changing a server's weight on the fly will have no effect,
1700 but this can be changed using "hash-type".
1701
Emeric Brun736aa232009-06-30 17:56:00 +02001702 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02001703 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02001704 The RDP cookie <name> (or "mstshash" if omitted) will be
1705 looked up and hashed for each incoming TCP request. Just as
1706 with the equivalent ACL 'req_rdp_cookie()' function, the name
1707 is not case-sensitive. This mechanism is useful as a degraded
1708 persistence mode, as it makes it possible to always send the
1709 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001710 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02001711 used instead.
1712
1713 Note that for this to work, the frontend must ensure that an
1714 RDP cookie is already present in the request buffer. For this
1715 you must use 'tcp-request content accept' rule combined with
1716 a 'req_rdp_cookie_cnt' ACL.
1717
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02001718 This algorithm is static by default, which means that
1719 changing a server's weight on the fly will have no effect,
1720 but this can be changed using "hash-type".
1721
Cyril Bontédc4d9032012-04-08 21:57:39 +02001722 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09001723
Willy Tarreau0ba27502007-12-24 16:55:16 +01001724 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02001725 algorithms. Right now, only "url_param" and "uri" support an
1726 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001727
Willy Tarreau3cd9af22009-03-15 14:06:41 +01001728 The load balancing algorithm of a backend is set to roundrobin when no other
1729 algorithm, mode nor option have been set. The algorithm may only be set once
1730 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001731
1732 Examples :
1733 balance roundrobin
1734 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001735 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01001736 balance hdr(User-Agent)
1737 balance hdr(host)
1738 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001739
1740 Note: the following caveats and limitations on using the "check_post"
1741 extension with "url_param" must be considered :
1742
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001743 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001744 to determine if the parameters will be found in the body or entity which
1745 may contain binary data. Therefore another method may be required to
1746 restrict consideration of POST requests that have no URL parameters in
1747 the body. (see acl reqideny http_end)
1748
1749 - using a <max_wait> value larger than the request buffer size does not
1750 make sense and is useless. The buffer size is set at build time, and
1751 defaults to 16 kB.
1752
1753 - Content-Encoding is not supported, the parameter search will probably
1754 fail; and load balancing will fall back to Round Robin.
1755
1756 - Expect: 100-continue is not supported, load balancing will fall back to
1757 Round Robin.
1758
1759 - Transfer-Encoding (RFC2616 3.6.1) is only supported in the first chunk.
1760 If the entire parameter value is not present in the first chunk, the
1761 selection of server is undefined (actually, defined by how little
1762 actually appeared in the first chunk).
1763
1764 - This feature does not support generation of a 100, 411 or 501 response.
1765
1766 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001767 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02001768 white space or control characters are found, indicating the end of what
1769 might be a URL parameter list. This is probably not a concern with SGML
1770 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001771
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02001772 See also : "dispatch", "cookie", "appsession", "transparent", "hash-type" and
1773 "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001774
1775
Willy Tarreaub6205fd2012-09-24 12:27:33 +02001776bind [<address>]:<port_range> [, ...] [param*]
1777bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01001778 Define one or several listening addresses and/or ports in a frontend.
1779 May be used in sections : defaults | frontend | listen | backend
1780 no | yes | yes | no
1781 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01001782 <address> is optional and can be a host name, an IPv4 address, an IPv6
1783 address, or '*'. It designates the address the frontend will
1784 listen on. If unset, all IPv4 addresses of the system will be
1785 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01001786 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01001787 Optionally, an address family prefix may be used before the
1788 address to force the family regardless of the address format,
1789 which can be useful to specify a path to a unix socket with
1790 no slash ('/'). Currently supported prefixes are :
1791 - 'ipv4@' -> address is always IPv4
1792 - 'ipv6@' -> address is always IPv6
1793 - 'unix@' -> address is a path to a local unix socket
Willy Tarreaub4fca5d2014-07-08 00:37:50 +02001794 - 'abns@' -> address is in abstract namespace (Linux only).
1795 Note: since abstract sockets are not "rebindable", they
1796 do not cope well with multi-process mode during
1797 soft-restart, so it is better to avoid them if
1798 nbproc is greater than 1. The effect is that if the
1799 new process fails to start, only one of the old ones
1800 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01001801 - 'fd@<n>' -> use file descriptor <n> inherited from the
1802 parent. The fd must be bound and may or may not already
1803 be listening.
Willy Tarreaudad36a32013-03-11 01:20:04 +01001804 Any part of the address string may reference any number of
1805 environment variables by preceding their name with a dollar
1806 sign ('$') and optionally enclosing them with braces ('{}'),
1807 similarly to what is done in Bourne shell.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01001808
Willy Tarreauc5011ca2010-03-22 11:53:56 +01001809 <port_range> is either a unique TCP port, or a port range for which the
1810 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001811 above. The port is mandatory for TCP listeners. Note that in
1812 the case of an IPv6 address, the port is always the number
1813 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01001814 - a numerical port (ex: '80')
1815 - a dash-delimited ports range explicitly stating the lower
1816 and upper bounds (ex: '2000-2100') which are included in
1817 the range.
1818
1819 Particular care must be taken against port ranges, because
1820 every <address:port> couple consumes one socket (= a file
1821 descriptor), so it's easy to consume lots of descriptors
1822 with a simple range, and to run out of sockets. Also, each
1823 <address:port> couple must be used only once among all
1824 instances running on a same system. Please note that binding
1825 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001826 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01001827 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001828
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001829 <path> is a UNIX socket path beginning with a slash ('/'). This is
1830 alternative to the TCP listening port. Haproxy will then
1831 receive UNIX connections on the socket located at this place.
1832 The path must begin with a slash and by default is absolute.
1833 It can be relative to the prefix defined by "unix-bind" in
1834 the global section. Note that the total length of the prefix
1835 followed by the socket path cannot exceed some system limits
1836 for UNIX sockets, which commonly are set to 107 characters.
1837
Willy Tarreaub6205fd2012-09-24 12:27:33 +02001838 <param*> is a list of parameters common to all sockets declared on the
1839 same line. These numerous parameters depend on OS and build
1840 options and have a complete section dedicated to them. Please
1841 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02001842
Willy Tarreau0ba27502007-12-24 16:55:16 +01001843 It is possible to specify a list of address:port combinations delimited by
1844 commas. The frontend will then listen on all of these addresses. There is no
1845 fixed limit to the number of addresses and ports which can be listened on in
1846 a frontend, as well as there is no limit to the number of "bind" statements
1847 in a frontend.
1848
1849 Example :
1850 listen http_proxy
1851 bind :80,:443
1852 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001853 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01001854
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02001855 listen http_https_proxy
1856 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02001857 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02001858
Willy Tarreau24709282013-03-10 21:32:12 +01001859 listen http_https_proxy_explicit
1860 bind ipv6@:80
1861 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
1862 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
1863
Willy Tarreaudad36a32013-03-11 01:20:04 +01001864 listen external_bind_app1
1865 bind fd@${FD_APP1}
1866
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001867 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02001868 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001869
1870
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001871bind-process [ all | odd | even | <number 1-64>[-<number 1-64>] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001872 Limit visibility of an instance to a certain set of processes numbers.
1873 May be used in sections : defaults | frontend | listen | backend
1874 yes | yes | yes | yes
1875 Arguments :
1876 all All process will see this instance. This is the default. It
1877 may be used to override a default value.
1878
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001879 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001880 option may be combined with other numbers.
1881
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001882 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001883 option may be combined with other numbers. Do not use it
1884 with less than 2 processes otherwise some instances might be
1885 missing from all processes.
1886
Willy Tarreau110ecc12012-11-15 17:50:01 +01001887 number The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001888 whose values must all be between 1 and 32 or 64 depending on
Willy Tarreau102df612014-05-07 23:56:38 +02001889 the machine's word size. If a proxy is bound to process
1890 numbers greater than the configured global.nbproc, it will
1891 either be forced to process #1 if a single process was
1892 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001893
1894 This keyword limits binding of certain instances to certain processes. This
1895 is useful in order not to have too many processes listening to the same
1896 ports. For instance, on a dual-core machine, it might make sense to set
1897 'nbproc 2' in the global section, then distributes the listeners among 'odd'
1898 and 'even' instances.
1899
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001900 At the moment, it is not possible to reference more than 32 or 64 processes
1901 using this keyword, but this should be more than enough for most setups.
1902 Please note that 'all' really means all processes regardless of the machine's
1903 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001904
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02001905 Each "bind" line may further be limited to a subset of the proxy's processes,
1906 please consult the "process" bind keyword in section 5.1.
1907
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001908 If some backends are referenced by frontends bound to other processes, the
1909 backend automatically inherits the frontend's processes.
1910
1911 Example :
1912 listen app_ip1
1913 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02001914 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001915
1916 listen app_ip2
1917 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02001918 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001919
1920 listen management
1921 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02001922 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001923
Willy Tarreau110ecc12012-11-15 17:50:01 +01001924 listen management
1925 bind 10.0.0.4:80
1926 bind-process 1-4
1927
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02001928 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01001929
1930
Willy Tarreau0ba27502007-12-24 16:55:16 +01001931block { if | unless } <condition>
1932 Block a layer 7 request if/unless a condition is matched
1933 May be used in sections : defaults | frontend | listen | backend
1934 no | yes | yes | yes
1935
1936 The HTTP request will be blocked very early in the layer 7 processing
1937 if/unless <condition> is matched. A 403 error will be returned if the request
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001938 is blocked. The condition has to reference ACLs (see section 7). This is
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02001939 typically used to deny access to certain sensitive resources if some
Willy Tarreau0ba27502007-12-24 16:55:16 +01001940 conditions are met or not met. There is no fixed limit to the number of
1941 "block" statements per instance.
1942
1943 Example:
1944 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
1945 acl invalid_src src_port 0:1023
1946 acl local_dst hdr(host) -i localhost
1947 block if invalid_src || local_dst
1948
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001949 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001950
1951
1952capture cookie <name> len <length>
1953 Capture and log a cookie in the request and in the response.
1954 May be used in sections : defaults | frontend | listen | backend
1955 no | yes | yes | no
1956 Arguments :
1957 <name> is the beginning of the name of the cookie to capture. In order
1958 to match the exact name, simply suffix the name with an equal
1959 sign ('='). The full name will appear in the logs, which is
1960 useful with application servers which adjust both the cookie name
1961 and value (eg: ASPSESSIONXXXXX).
1962
1963 <length> is the maximum number of characters to report in the logs, which
1964 include the cookie name, the equal sign and the value, all in the
1965 standard "name=value" form. The string will be truncated on the
1966 right if it exceeds <length>.
1967
1968 Only the first cookie is captured. Both the "cookie" request headers and the
1969 "set-cookie" response headers are monitored. This is particularly useful to
1970 check for application bugs causing session crossing or stealing between
1971 users, because generally the user's cookies can only change on a login page.
1972
1973 When the cookie was not presented by the client, the associated log column
1974 will report "-". When a request does not cause a cookie to be assigned by the
1975 server, a "-" is reported in the response column.
1976
1977 The capture is performed in the frontend only because it is necessary that
1978 the log format does not change for a given frontend depending on the
1979 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01001980 "capture cookie" statement in a frontend. The maximum capture length is set
1981 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
1982 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001983
1984 Example:
1985 capture cookie ASPSESSION len 32
1986
1987 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001988 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001989
1990
1991capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01001992 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001993 May be used in sections : defaults | frontend | listen | backend
1994 no | yes | yes | no
1995 Arguments :
1996 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01001997 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01001998 appear in the requests, with the first letter of each word in
1999 upper case. The header name will not appear in the logs, only the
2000 value is reported, but the position in the logs is respected.
2001
2002 <length> is the maximum number of characters to extract from the value and
2003 report in the logs. The string will be truncated on the right if
2004 it exceeds <length>.
2005
Willy Tarreau4460d032012-11-21 23:37:37 +01002006 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002007 value will be added to the logs between braces ('{}'). If multiple headers
2008 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002009 in the same order they were declared in the configuration. Non-existent
2010 headers will be logged just as an empty string. Common uses for request
2011 header captures include the "Host" field in virtual hosting environments, the
2012 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002013 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002014 environments to find where the request came from.
2015
2016 Note that when capturing headers such as "User-agent", some spaces may be
2017 logged, making the log analysis more difficult. Thus be careful about what
2018 you log if you know your log parser is not smart enough to rely on the
2019 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002020
Willy Tarreau0900abb2012-11-22 00:21:46 +01002021 There is no limit to the number of captured request headers nor to their
2022 length, though it is wise to keep them low to limit memory usage per session.
2023 In order to keep log format consistent for a same frontend, header captures
2024 can only be declared in a frontend. It is not possible to specify a capture
2025 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002026
2027 Example:
2028 capture request header Host len 15
2029 capture request header X-Forwarded-For len 15
2030 capture request header Referrer len 15
2031
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002032 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002033 about logging.
2034
2035
2036capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01002037 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002038 May be used in sections : defaults | frontend | listen | backend
2039 no | yes | yes | no
2040 Arguments :
2041 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002042 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01002043 appear in the response, with the first letter of each word in
2044 upper case. The header name will not appear in the logs, only the
2045 value is reported, but the position in the logs is respected.
2046
2047 <length> is the maximum number of characters to extract from the value and
2048 report in the logs. The string will be truncated on the right if
2049 it exceeds <length>.
2050
Willy Tarreau4460d032012-11-21 23:37:37 +01002051 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002052 result will be added to the logs between braces ('{}') after the captured
2053 request headers. If multiple headers are captured, they will be delimited by
2054 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002055 the configuration. Non-existent headers will be logged just as an empty
2056 string. Common uses for response header captures include the "Content-length"
2057 header which indicates how many bytes are expected to be returned, the
2058 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002059
Willy Tarreau0900abb2012-11-22 00:21:46 +01002060 There is no limit to the number of captured response headers nor to their
2061 length, though it is wise to keep them low to limit memory usage per session.
2062 In order to keep log format consistent for a same frontend, header captures
2063 can only be declared in a frontend. It is not possible to specify a capture
2064 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002065
2066 Example:
2067 capture response header Content-length len 9
2068 capture response header Location len 15
2069
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002070 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002071 about logging.
2072
2073
Cyril Bontéf0c60612010-02-06 14:44:47 +01002074clitimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002075 Set the maximum inactivity time on the client side.
2076 May be used in sections : defaults | frontend | listen | backend
2077 yes | yes | yes | no
2078 Arguments :
2079 <timeout> is the timeout value is specified in milliseconds by default, but
2080 can be in any other unit if the number is suffixed by the unit,
2081 as explained at the top of this document.
2082
2083 The inactivity timeout applies when the client is expected to acknowledge or
2084 send data. In HTTP mode, this timeout is particularly important to consider
2085 during the first phase, when the client sends the request, and during the
2086 response while it is reading data sent by the server. The value is specified
2087 in milliseconds by default, but can be in any other unit if the number is
2088 suffixed by the unit, as specified at the top of this document. In TCP mode
2089 (and to a lesser extent, in HTTP mode), it is highly recommended that the
2090 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002091 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01002092 losses by specifying timeouts that are slightly above multiples of 3 seconds
2093 (eg: 4 or 5 seconds).
2094
2095 This parameter is specific to frontends, but can be specified once for all in
2096 "defaults" sections. This is in fact one of the easiest solutions not to
2097 forget about it. An unspecified timeout results in an infinite timeout, which
2098 is not recommended. Such a usage is accepted and works but reports a warning
2099 during startup because it may results in accumulation of expired sessions in
2100 the system if the system's timeouts are not configured either.
2101
2102 This parameter is provided for compatibility but is currently deprecated.
2103 Please use "timeout client" instead.
2104
Willy Tarreau036fae02008-01-06 13:24:40 +01002105 See also : "timeout client", "timeout http-request", "timeout server", and
2106 "srvtimeout".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002107
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002108compression algo <algorithm> ...
2109compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02002110compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02002111 Enable HTTP compression.
2112 May be used in sections : defaults | frontend | listen | backend
2113 yes | yes | yes | yes
2114 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002115 algo is followed by the list of supported compression algorithms.
2116 type is followed by the list of MIME types that will be compressed.
2117 offload makes haproxy work as a compression offloader only (see notes).
2118
2119 The currently supported algorithms are :
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04002120 identity this is mostly for debugging, and it was useful for developing
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002121 the compression feature. Identity does not apply any change on
2122 data.
2123
2124 gzip applies gzip compression. This setting is only available when
2125 support for zlib was built in.
2126
2127 deflate same as gzip, but with deflate algorithm and zlib format.
2128 Note that this algorithm has ambiguous support on many browsers
2129 and no support at all from recent ones. It is strongly
2130 recommended not to use it for anything else than experimentation.
2131 This setting is only available when support for zlib was built
2132 in.
2133
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04002134 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002135 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002136 If backend servers support HTTP compression, these directives
2137 will be no-op: haproxy will see the compressed response and will not
2138 compress again. If backend servers do not support HTTP compression and
2139 there is Accept-Encoding header in request, haproxy will compress the
2140 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02002141
2142 The "offload" setting makes haproxy remove the Accept-Encoding header to
2143 prevent backend servers from compressing responses. It is strongly
2144 recommended not to do this because this means that all the compression work
2145 will be done on the single point where haproxy is located. However in some
2146 deployment scenarios, haproxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002147 with broken HTTP compression implementation which can't be turned off.
2148 In that case haproxy can be used to prevent that gateway from emitting
2149 invalid payloads. In this case, simply removing the header in the
2150 configuration does not work because it applies before the header is parsed,
2151 so that prevents haproxy from compressing. The "offload" setting should
Willy Tarreau4cf57af2014-07-12 16:37:02 +02002152 then be used for such scenarios. Note: for now, the "offload" setting is
2153 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02002154
William Lallemand05097442012-11-20 12:14:28 +01002155 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002156 * the request does not advertise a supported compression algorithm in the
2157 "Accept-Encoding" header
2158 * the response message is not HTTP/1.1
William Lallemandd3002612012-11-26 14:34:47 +01002159 * HTTP status code is not 200
William Lallemand8bb4e342013-12-10 17:28:48 +01002160 * response header "Transfer-Encoding" contains "chunked" (Temporary
2161 Workaround)
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002162 * response contain neither a "Content-Length" header nor a
2163 "Transfer-Encoding" whose last value is "chunked"
2164 * response contains a "Content-Type" header whose first value starts with
2165 "multipart"
2166 * the response contains the "no-transform" value in the "Cache-control"
2167 header
2168 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
2169 and later
2170 * The response contains a "Content-Encoding" header, indicating that the
2171 response is already compressed (see compression offload)
William Lallemand05097442012-11-20 12:14:28 +01002172
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002173 Note: The compression does not rewrite Etag headers, and does not emit the
2174 Warning header.
William Lallemand05097442012-11-20 12:14:28 +01002175
William Lallemand82fe75c2012-10-23 10:25:10 +02002176 Examples :
2177 compression algo gzip
2178 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01002179
Cyril Bontéf0c60612010-02-06 14:44:47 +01002180contimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002181 Set the maximum time to wait for a connection attempt to a server to succeed.
2182 May be used in sections : defaults | frontend | listen | backend
2183 yes | no | yes | yes
2184 Arguments :
2185 <timeout> is the timeout value is specified in milliseconds by default, but
2186 can be in any other unit if the number is suffixed by the unit,
2187 as explained at the top of this document.
2188
2189 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002190 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +01002191 cover one or several TCP packet losses by specifying timeouts that are
Willy Tarreau0ba27502007-12-24 16:55:16 +01002192 slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the
2193 connect timeout also presets the queue timeout to the same value if this one
2194 has not been specified. Historically, the contimeout was also used to set the
2195 tarpit timeout in a listen section, which is not possible in a pure frontend.
2196
2197 This parameter is specific to backends, but can be specified once for all in
2198 "defaults" sections. This is in fact one of the easiest solutions not to
2199 forget about it. An unspecified timeout results in an infinite timeout, which
2200 is not recommended. Such a usage is accepted and works but reports a warning
2201 during startup because it may results in accumulation of failed sessions in
2202 the system if the system's timeouts are not configured either.
2203
2204 This parameter is provided for backwards compatibility but is currently
2205 deprecated. Please use "timeout connect", "timeout queue" or "timeout tarpit"
2206 instead.
2207
2208 See also : "timeout connect", "timeout queue", "timeout tarpit",
2209 "timeout server", "contimeout".
2210
2211
Willy Tarreau55165fe2009-05-10 12:02:55 +02002212cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02002213 [ postonly ] [ preserve ] [ httponly ] [ secure ]
2214 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002215 Enable cookie-based persistence in a backend.
2216 May be used in sections : defaults | frontend | listen | backend
2217 yes | no | yes | yes
2218 Arguments :
2219 <name> is the name of the cookie which will be monitored, modified or
2220 inserted in order to bring persistence. This cookie is sent to
2221 the client via a "Set-Cookie" header in the response, and is
2222 brought back by the client in a "Cookie" header in all requests.
2223 Special care should be taken to choose a name which does not
2224 conflict with any likely application cookie. Also, if the same
2225 backends are subject to be used by the same clients (eg:
2226 HTTP/HTTPS), care should be taken to use different cookie names
2227 between all backends if persistence between them is not desired.
2228
2229 rewrite This keyword indicates that the cookie will be provided by the
2230 server and that haproxy will have to modify its value to set the
2231 server's identifier in it. This mode is handy when the management
2232 of complex combinations of "Set-cookie" and "Cache-control"
2233 headers is left to the application. The application can then
2234 decide whether or not it is appropriate to emit a persistence
2235 cookie. Since all responses should be monitored, this mode only
2236 works in HTTP close mode. Unless the application behaviour is
2237 very complex and/or broken, it is advised not to start with this
2238 mode for new deployments. This keyword is incompatible with
2239 "insert" and "prefix".
2240
2241 insert This keyword indicates that the persistence cookie will have to
Willy Tarreaua79094d2010-08-31 22:54:15 +02002242 be inserted by haproxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002243
Willy Tarreaua79094d2010-08-31 22:54:15 +02002244 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002245 server. When used without the "preserve" option, if the server
2246 emits a cookie with the same name, it will be remove before
2247 processing. For this reason, this mode can be used to upgrade
2248 existing configurations running in the "rewrite" mode. The cookie
2249 will only be a session cookie and will not be stored on the
2250 client's disk. By default, unless the "indirect" option is added,
2251 the server will see the cookies emitted by the client. Due to
2252 caching effects, it is generally wise to add the "nocache" or
2253 "postonly" keywords (see below). The "insert" keyword is not
2254 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002255
2256 prefix This keyword indicates that instead of relying on a dedicated
2257 cookie for the persistence, an existing one will be completed.
2258 This may be needed in some specific environments where the client
2259 does not support more than one single cookie and the application
2260 already needs it. In this case, whenever the server sets a cookie
2261 named <name>, it will be prefixed with the server's identifier
2262 and a delimiter. The prefix will be removed from all client
2263 requests so that the server still finds the cookie it emitted.
2264 Since all requests and responses are subject to being modified,
2265 this mode requires the HTTP close mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02002266 not compatible with "rewrite" and "insert". Note: it is highly
2267 recommended not to use "indirect" with "prefix", otherwise server
2268 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002269
Willy Tarreaua79094d2010-08-31 22:54:15 +02002270 indirect When this option is specified, no cookie will be emitted to a
2271 client which already has a valid one for the server which has
2272 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002273 it will be removed, unless the "preserve" option is also set. In
2274 "insert" mode, this will additionally remove cookies from the
2275 requests transmitted to the server, making the persistence
2276 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02002277 Note: it is highly recommended not to use "indirect" with
2278 "prefix", otherwise server cookie updates would not be sent to
2279 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002280
2281 nocache This option is recommended in conjunction with the insert mode
2282 when there is a cache between the client and HAProxy, as it
2283 ensures that a cacheable response will be tagged non-cacheable if
2284 a cookie needs to be inserted. This is important because if all
2285 persistence cookies are added on a cacheable home page for
2286 instance, then all customers will then fetch the page from an
2287 outer cache and will all share the same persistence cookie,
2288 leading to one server receiving much more traffic than others.
2289 See also the "insert" and "postonly" options.
2290
2291 postonly This option ensures that cookie insertion will only be performed
2292 on responses to POST requests. It is an alternative to the
2293 "nocache" option, because POST responses are not cacheable, so
2294 this ensures that the persistence cookie will never get cached.
2295 Since most sites do not need any sort of persistence before the
2296 first POST which generally is a login request, this is a very
2297 efficient method to optimize caching without risking to find a
2298 persistence cookie in the cache.
2299 See also the "insert" and "nocache" options.
2300
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002301 preserve This option may only be used with "insert" and/or "indirect". It
2302 allows the server to emit the persistence cookie itself. In this
2303 case, if a cookie is found in the response, haproxy will leave it
2304 untouched. This is useful in order to end persistence after a
2305 logout request for instance. For this, the server just has to
2306 emit a cookie with an invalid value (eg: empty) or with a date in
2307 the past. By combining this mechanism with the "disable-on-404"
2308 check option, it is possible to perform a completely graceful
2309 shutdown because users will definitely leave the server after
2310 they logout.
2311
Willy Tarreau4992dd22012-05-31 21:02:17 +02002312 httponly This option tells haproxy to add an "HttpOnly" cookie attribute
2313 when a cookie is inserted. This attribute is used so that a
2314 user agent doesn't share the cookie with non-HTTP components.
2315 Please check RFC6265 for more information on this attribute.
2316
2317 secure This option tells haproxy to add a "Secure" cookie attribute when
2318 a cookie is inserted. This attribute is used so that a user agent
2319 never emits this cookie over non-secure channels, which means
2320 that a cookie learned with this flag will be presented only over
2321 SSL/TLS connections. Please check RFC6265 for more information on
2322 this attribute.
2323
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02002324 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002325 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01002326 name. If the domain begins with a dot, the browser is allowed to
2327 use it for any host ending with that name. It is also possible to
2328 specify several domain names by invoking this option multiple
2329 times. Some browsers might have small limits on the number of
2330 domains, so be careful when doing that. For the record, sending
2331 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02002332
Willy Tarreau996a92c2010-10-13 19:30:47 +02002333 maxidle This option allows inserted cookies to be ignored after some idle
2334 time. It only works with insert-mode cookies. When a cookie is
2335 sent to the client, the date this cookie was emitted is sent too.
2336 Upon further presentations of this cookie, if the date is older
2337 than the delay indicated by the parameter (in seconds), it will
2338 be ignored. Otherwise, it will be refreshed if needed when the
2339 response is sent to the client. This is particularly useful to
2340 prevent users who never close their browsers from remaining for
2341 too long on the same server (eg: after a farm size change). When
2342 this option is set and a cookie has no date, it is always
2343 accepted, but gets refreshed in the response. This maintains the
2344 ability for admins to access their sites. Cookies that have a
2345 date in the future further than 24 hours are ignored. Doing so
2346 lets admins fix timezone issues without risking kicking users off
2347 the site.
2348
2349 maxlife This option allows inserted cookies to be ignored after some life
2350 time, whether they're in use or not. It only works with insert
2351 mode cookies. When a cookie is first sent to the client, the date
2352 this cookie was emitted is sent too. Upon further presentations
2353 of this cookie, if the date is older than the delay indicated by
2354 the parameter (in seconds), it will be ignored. If the cookie in
2355 the request has no date, it is accepted and a date will be set.
2356 Cookies that have a date in the future further than 24 hours are
2357 ignored. Doing so lets admins fix timezone issues without risking
2358 kicking users off the site. Contrary to maxidle, this value is
2359 not refreshed, only the first visit date counts. Both maxidle and
2360 maxlife may be used at the time. This is particularly useful to
2361 prevent users who never close their browsers from remaining for
2362 too long on the same server (eg: after a farm size change). This
2363 is stronger than the maxidle method in that it forces a
2364 redispatch after some absolute delay.
2365
Willy Tarreau0ba27502007-12-24 16:55:16 +01002366 There can be only one persistence cookie per HTTP backend, and it can be
2367 declared in a defaults section. The value of the cookie will be the value
2368 indicated after the "cookie" keyword in a "server" statement. If no cookie
2369 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002370
Willy Tarreau0ba27502007-12-24 16:55:16 +01002371 Examples :
2372 cookie JSESSIONID prefix
2373 cookie SRV insert indirect nocache
2374 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02002375 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01002376
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02002377 See also : "appsession", "balance source", "capture cookie", "server"
Cyril Bonté0d4bf012010-04-25 23:21:46 +02002378 and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002379
Willy Tarreau983e01e2010-01-11 18:42:06 +01002380
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002381default-server [param*]
2382 Change default options for a server in a backend
2383 May be used in sections : defaults | frontend | listen | backend
2384 yes | no | yes | yes
2385 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01002386 <param*> is a list of parameters for this server. The "default-server"
2387 keyword accepts an important number of options and has a complete
2388 section dedicated to it. Please refer to section 5 for more
2389 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002390
Willy Tarreau983e01e2010-01-11 18:42:06 +01002391 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002392 default-server inter 1000 weight 13
2393
2394 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01002395
Willy Tarreau983e01e2010-01-11 18:42:06 +01002396
Willy Tarreau0ba27502007-12-24 16:55:16 +01002397default_backend <backend>
2398 Specify the backend to use when no "use_backend" rule has been matched.
2399 May be used in sections : defaults | frontend | listen | backend
2400 yes | yes | yes | no
2401 Arguments :
2402 <backend> is the name of the backend to use.
2403
2404 When doing content-switching between frontend and backends using the
2405 "use_backend" keyword, it is often useful to indicate which backend will be
2406 used when no rule has matched. It generally is the dynamic backend which
2407 will catch all undetermined requests.
2408
Willy Tarreau0ba27502007-12-24 16:55:16 +01002409 Example :
2410
2411 use_backend dynamic if url_dyn
2412 use_backend static if url_css url_img extension_img
2413 default_backend dynamic
2414
Willy Tarreau2769aa02007-12-27 18:26:09 +01002415 See also : "use_backend", "reqsetbe", "reqisetbe"
2416
Willy Tarreau0ba27502007-12-24 16:55:16 +01002417
Baptiste Assmann27f51342013-10-09 06:51:49 +02002418description <string>
2419 Describe a listen, frontend or backend.
2420 May be used in sections : defaults | frontend | listen | backend
2421 no | yes | yes | yes
2422 Arguments : string
2423
2424 Allows to add a sentence to describe the related object in the HAProxy HTML
2425 stats page. The description will be printed on the right of the object name
2426 it describes.
2427 No need to backslash spaces in the <string> arguments.
2428
2429
Willy Tarreau0ba27502007-12-24 16:55:16 +01002430disabled
2431 Disable a proxy, frontend or backend.
2432 May be used in sections : defaults | frontend | listen | backend
2433 yes | yes | yes | yes
2434 Arguments : none
2435
2436 The "disabled" keyword is used to disable an instance, mainly in order to
2437 liberate a listening port or to temporarily disable a service. The instance
2438 will still be created and its configuration will be checked, but it will be
2439 created in the "stopped" state and will appear as such in the statistics. It
2440 will not receive any traffic nor will it send any health-checks or logs. It
2441 is possible to disable many instances at once by adding the "disabled"
2442 keyword in a "defaults" section.
2443
2444 See also : "enabled"
2445
2446
Willy Tarreau5ce94572010-06-07 14:35:41 +02002447dispatch <address>:<port>
2448 Set a default server address
2449 May be used in sections : defaults | frontend | listen | backend
2450 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02002451 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02002452
2453 <address> is the IPv4 address of the default server. Alternatively, a
2454 resolvable hostname is supported, but this name will be resolved
2455 during start-up.
2456
2457 <ports> is a mandatory port specification. All connections will be sent
2458 to this port, and it is not permitted to use port offsets as is
2459 possible with normal servers.
2460
Willy Tarreau787aed52011-04-15 06:45:37 +02002461 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02002462 server can take the connection. In the past it was used to forward non
2463 persistent connections to an auxiliary load balancer. Due to its simple
2464 syntax, it has also been used for simple TCP relays. It is recommended not to
2465 use it for more clarity, and to use the "server" directive instead.
2466
2467 See also : "server"
2468
2469
Willy Tarreau0ba27502007-12-24 16:55:16 +01002470enabled
2471 Enable a proxy, frontend or backend.
2472 May be used in sections : defaults | frontend | listen | backend
2473 yes | yes | yes | yes
2474 Arguments : none
2475
2476 The "enabled" keyword is used to explicitly enable an instance, when the
2477 defaults has been set to "disabled". This is very rarely used.
2478
2479 See also : "disabled"
2480
2481
2482errorfile <code> <file>
2483 Return a file contents instead of errors generated by HAProxy
2484 May be used in sections : defaults | frontend | listen | backend
2485 yes | yes | yes | yes
2486 Arguments :
2487 <code> is the HTTP status code. Currently, HAProxy is capable of
Willy Tarreauae94d4d2011-05-11 16:28:49 +02002488 generating codes 200, 400, 403, 408, 500, 502, 503, and 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002489
2490 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002491 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01002492 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01002493 error pages, and to use absolute paths, since files are read
2494 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002495
2496 It is important to understand that this keyword is not meant to rewrite
2497 errors returned by the server, but errors detected and returned by HAProxy.
2498 This is why the list of supported errors is limited to a small set.
2499
Willy Tarreauae94d4d2011-05-11 16:28:49 +02002500 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
2501
Willy Tarreau0ba27502007-12-24 16:55:16 +01002502 The files are returned verbatim on the TCP socket. This allows any trick such
2503 as redirections to another URL or site, as well as tricks to clean cookies,
2504 force enable or disable caching, etc... The package provides default error
2505 files returning the same contents as default errors.
2506
Willy Tarreau59140a22009-02-22 12:02:30 +01002507 The files should not exceed the configured buffer size (BUFSIZE), which
2508 generally is 8 or 16 kB, otherwise they will be truncated. It is also wise
2509 not to put any reference to local contents (eg: images) in order to avoid
2510 loops between the client and HAProxy when all servers are down, causing an
2511 error to be returned instead of an image. For better HTTP compliance, it is
2512 recommended that all header lines end with CR-LF and not LF alone.
2513
Willy Tarreau0ba27502007-12-24 16:55:16 +01002514 The files are read at the same time as the configuration and kept in memory.
2515 For this reason, the errors continue to be returned even when the process is
2516 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01002517 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01002518 403 status code and interrogating a blocked URL.
2519
2520 See also : "errorloc", "errorloc302", "errorloc303"
2521
Willy Tarreau59140a22009-02-22 12:02:30 +01002522 Example :
2523 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau2705a612014-05-23 17:38:34 +02002524 errorfile 408 /dev/null # workaround Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01002525 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
2526 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
2527
Willy Tarreau2769aa02007-12-27 18:26:09 +01002528
2529errorloc <code> <url>
2530errorloc302 <code> <url>
2531 Return an HTTP redirection to a URL instead of errors generated by HAProxy
2532 May be used in sections : defaults | frontend | listen | backend
2533 yes | yes | yes | yes
2534 Arguments :
2535 <code> is the HTTP status code. Currently, HAProxy is capable of
Willy Tarreauae94d4d2011-05-11 16:28:49 +02002536 generating codes 200, 400, 403, 408, 500, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01002537
2538 <url> it is the exact contents of the "Location" header. It may contain
2539 either a relative URI to an error page hosted on the same site,
2540 or an absolute URI designating an error page on another site.
2541 Special care should be given to relative URIs to avoid redirect
2542 loops if the URI itself may generate the same error (eg: 500).
2543
2544 It is important to understand that this keyword is not meant to rewrite
2545 errors returned by the server, but errors detected and returned by HAProxy.
2546 This is why the list of supported errors is limited to a small set.
2547
Willy Tarreauae94d4d2011-05-11 16:28:49 +02002548 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
2549
Willy Tarreau2769aa02007-12-27 18:26:09 +01002550 Note that both keyword return the HTTP 302 status code, which tells the
2551 client to fetch the designated URL using the same HTTP method. This can be
2552 quite problematic in case of non-GET methods such as POST, because the URL
2553 sent to the client might not be allowed for something other than GET. To
2554 workaround this problem, please use "errorloc303" which send the HTTP 303
2555 status code, indicating to the client that the URL must be fetched with a GET
2556 request.
2557
2558 See also : "errorfile", "errorloc303"
2559
2560
2561errorloc303 <code> <url>
2562 Return an HTTP redirection to a URL instead of errors generated by HAProxy
2563 May be used in sections : defaults | frontend | listen | backend
2564 yes | yes | yes | yes
2565 Arguments :
2566 <code> is the HTTP status code. Currently, HAProxy is capable of
2567 generating codes 400, 403, 408, 500, 502, 503, and 504.
2568
2569 <url> it is the exact contents of the "Location" header. It may contain
2570 either a relative URI to an error page hosted on the same site,
2571 or an absolute URI designating an error page on another site.
2572 Special care should be given to relative URIs to avoid redirect
2573 loops if the URI itself may generate the same error (eg: 500).
2574
2575 It is important to understand that this keyword is not meant to rewrite
2576 errors returned by the server, but errors detected and returned by HAProxy.
2577 This is why the list of supported errors is limited to a small set.
2578
Willy Tarreauae94d4d2011-05-11 16:28:49 +02002579 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
2580
Willy Tarreau2769aa02007-12-27 18:26:09 +01002581 Note that both keyword return the HTTP 303 status code, which tells the
2582 client to fetch the designated URL using the same HTTP GET method. This
2583 solves the usual problems associated with "errorloc" and the 302 code. It is
2584 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002585 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01002586
2587 See also : "errorfile", "errorloc", "errorloc302"
2588
2589
Willy Tarreau4de91492010-01-22 19:10:05 +01002590force-persist { if | unless } <condition>
2591 Declare a condition to force persistence on down servers
2592 May be used in sections: defaults | frontend | listen | backend
2593 no | yes | yes | yes
2594
2595 By default, requests are not dispatched to down servers. It is possible to
2596 force this using "option persist", but it is unconditional and redispatches
2597 to a valid server if "option redispatch" is set. That leaves with very little
2598 possibilities to force some requests to reach a server which is artificially
2599 marked down for maintenance operations.
2600
2601 The "force-persist" statement allows one to declare various ACL-based
2602 conditions which, when met, will cause a request to ignore the down status of
2603 a server and still try to connect to it. That makes it possible to start a
2604 server, still replying an error to the health checks, and run a specially
2605 configured browser to test the service. Among the handy methods, one could
2606 use a specific source IP address, or a specific cookie. The cookie also has
2607 the advantage that it can easily be added/removed on the browser from a test
2608 page. Once the service is validated, it is then possible to open the service
2609 to the world by returning a valid response to health checks.
2610
2611 The forced persistence is enabled when an "if" condition is met, or unless an
2612 "unless" condition is met. The final redispatch is always disabled when this
2613 is used.
2614
Cyril Bonté0d4bf012010-04-25 23:21:46 +02002615 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02002616 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01002617
2618
Willy Tarreau2769aa02007-12-27 18:26:09 +01002619fullconn <conns>
2620 Specify at what backend load the servers will reach their maxconn
2621 May be used in sections : defaults | frontend | listen | backend
2622 yes | no | yes | yes
2623 Arguments :
2624 <conns> is the number of connections on the backend which will make the
2625 servers use the maximal number of connections.
2626
Willy Tarreau198a7442008-01-17 12:05:32 +01002627 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01002628 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01002629 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01002630 load. The server will then always accept at least <minconn> connections,
2631 never more than <maxconn>, and the limit will be on the ramp between both
2632 values when the backend has less than <conns> concurrent connections. This
2633 makes it possible to limit the load on the servers during normal loads, but
2634 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002635 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01002636
Willy Tarreaufbb78422011-06-05 15:38:35 +02002637 Since it's hard to get this value right, haproxy automatically sets it to
2638 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01002639 backend (based on "use_backend" and "default_backend" rules). That way it's
2640 safe to leave it unset. However, "use_backend" involving dynamic names are
2641 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02002642
Willy Tarreau2769aa02007-12-27 18:26:09 +01002643 Example :
2644 # The servers will accept between 100 and 1000 concurrent connections each
2645 # and the maximum of 1000 will be reached when the backend reaches 10000
2646 # connections.
2647 backend dynamic
2648 fullconn 10000
2649 server srv1 dyn1:80 minconn 100 maxconn 1000
2650 server srv2 dyn2:80 minconn 100 maxconn 1000
2651
2652 See also : "maxconn", "server"
2653
2654
2655grace <time>
2656 Maintain a proxy operational for some time after a soft stop
2657 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01002658 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01002659 Arguments :
2660 <time> is the time (by default in milliseconds) for which the instance
2661 will remain operational with the frontend sockets still listening
2662 when a soft-stop is received via the SIGUSR1 signal.
2663
2664 This may be used to ensure that the services disappear in a certain order.
2665 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002666 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01002667 needed by the equipment to detect the failure.
2668
2669 Note that currently, there is very little benefit in using this parameter,
2670 and it may in fact complicate the soft-reconfiguration process more than
2671 simplify it.
2672
Willy Tarreau0ba27502007-12-24 16:55:16 +01002673
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05002674hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002675 Specify a method to use for mapping hashes to servers
2676 May be used in sections : defaults | frontend | listen | backend
2677 yes | no | yes | yes
2678 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04002679 <method> is the method used to select a server from the hash computed by
2680 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002681
Bhaskar98634f02013-10-29 23:30:51 -04002682 map-based the hash table is a static array containing all alive servers.
2683 The hashes will be very smooth, will consider weights, but
2684 will be static in that weight changes while a server is up
2685 will be ignored. This means that there will be no slow start.
2686 Also, since a server is selected by its position in the array,
2687 most mappings are changed when the server count changes. This
2688 means that when a server goes up or down, or when a server is
2689 added to a farm, most connections will be redistributed to
2690 different servers. This can be inconvenient with caches for
2691 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01002692
Bhaskar98634f02013-10-29 23:30:51 -04002693 consistent the hash table is a tree filled with many occurrences of each
2694 server. The hash key is looked up in the tree and the closest
2695 server is chosen. This hash is dynamic, it supports changing
2696 weights while the servers are up, so it is compatible with the
2697 slow start feature. It has the advantage that when a server
2698 goes up or down, only its associations are moved. When a
2699 server is added to the farm, only a few part of the mappings
2700 are redistributed, making it an ideal method for caches.
2701 However, due to its principle, the distribution will never be
2702 very smooth and it may sometimes be necessary to adjust a
2703 server's weight or its ID to get a more balanced distribution.
2704 In order to get the same distribution on multiple load
2705 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05002706 same IDs. Note: consistent hash uses sdbm and avalanche if no
2707 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04002708
2709 <function> is the hash function to be used :
2710
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002711 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04002712 reimplementation of ndbm) database library. It was found to do
2713 well in scrambling bits, causing better distribution of the keys
2714 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05002715 function with good distribution, unless the total server weight
2716 is a multiple of 64, in which case applying the avalanche
2717 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04002718
2719 djb2 this function was first proposed by Dan Bernstein many years ago
2720 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05002721 function provides a better distribution than sdbm. It generally
2722 works well with text-based inputs though it can perform extremely
2723 poorly with numeric-only input or when the total server weight is
2724 a multiple of 33, unless the avalanche modifier is also used.
2725
Willy Tarreaua0f42712013-11-14 14:30:35 +01002726 wt6 this function was designed for haproxy while testing other
2727 functions in the past. It is not as smooth as the other ones, but
2728 is much less sensible to the input data set or to the number of
2729 servers. It can make sense as an alternative to sdbm+avalanche or
2730 djb2+avalanche for consistent hashing or when hashing on numeric
2731 data such as a source IP address or a visitor identifier in a URL
2732 parameter.
2733
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05002734 <modifier> indicates an optional method applied after hashing the key :
2735
2736 avalanche This directive indicates that the result from the hash
2737 function above should not be used in its raw form but that
2738 a 4-byte full avalanche hash must be applied first. The
2739 purpose of this step is to mix the resulting bits from the
2740 previous hash in order to avoid any undesired effect when
2741 the input contains some limited values or when the number of
2742 servers is a multiple of one of the hash's components (64
2743 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
2744 result less predictable, but it's also not as smooth as when
2745 using the original function. Some testing might be needed
2746 with some workloads. This hash is one of the many proposed
2747 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002748
Bhaskar98634f02013-10-29 23:30:51 -04002749 The default hash type is "map-based" and is recommended for most usages. The
2750 default function is "sdbm", the selection of a function should be based on
2751 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002752
2753 See also : "balance", "server"
2754
2755
Willy Tarreau0ba27502007-12-24 16:55:16 +01002756http-check disable-on-404
2757 Enable a maintenance mode upon HTTP/404 response to health-checks
2758 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01002759 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01002760 Arguments : none
2761
2762 When this option is set, a server which returns an HTTP code 404 will be
2763 excluded from further load-balancing, but will still receive persistent
2764 connections. This provides a very convenient method for Web administrators
2765 to perform a graceful shutdown of their servers. It is also important to note
2766 that a server which is detected as failed while it was in this mode will not
2767 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
2768 will immediately be reinserted into the farm. The status on the stats page
2769 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01002770 option only works in conjunction with the "httpchk" option. If this option
2771 is used with "http-check expect", then it has precedence over it so that 404
2772 responses will still be considered as soft-stop.
2773
2774 See also : "option httpchk", "http-check expect"
2775
2776
2777http-check expect [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002778 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01002779 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02002780 yes | no | yes | yes
Willy Tarreaubd741542010-03-16 18:46:54 +01002781 Arguments :
2782 <match> is a keyword indicating how to look for a specific pattern in the
2783 response. The keyword may be one of "status", "rstatus",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002784 "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01002785 exclamation mark ("!") to negate the match. Spaces are allowed
2786 between the exclamation mark and the keyword. See below for more
2787 details on the supported keywords.
2788
2789 <pattern> is the pattern to look for. It may be a string or a regular
2790 expression. If the pattern contains spaces, they must be escaped
2791 with the usual backslash ('\').
2792
2793 By default, "option httpchk" considers that response statuses 2xx and 3xx
2794 are valid, and that others are invalid. When "http-check expect" is used,
2795 it defines what is considered valid or invalid. Only one "http-check"
2796 statement is supported in a backend. If a server fails to respond or times
2797 out, the check obviously fails. The available matches are :
2798
2799 status <string> : test the exact string match for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002800 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01002801 response's status code is exactly this string. If the
2802 "status" keyword is prefixed with "!", then the response
2803 will be considered invalid if the status code matches.
2804
2805 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002806 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01002807 response's status code matches the expression. If the
2808 "rstatus" keyword is prefixed with "!", then the response
2809 will be considered invalid if the status code matches.
2810 This is mostly used to check for multiple codes.
2811
2812 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002813 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01002814 response's body contains this exact string. If the
2815 "string" keyword is prefixed with "!", then the response
2816 will be considered invalid if the body contains this
2817 string. This can be used to look for a mandatory word at
2818 the end of a dynamic page, or to detect a failure when a
2819 specific error appears on the check page (eg: a stack
2820 trace).
2821
2822 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04002823 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01002824 response's body matches this expression. If the "rstring"
2825 keyword is prefixed with "!", then the response will be
2826 considered invalid if the body matches the expression.
2827 This can be used to look for a mandatory word at the end
2828 of a dynamic page, or to detect a failure when a specific
2829 error appears on the check page (eg: a stack trace).
2830
2831 It is important to note that the responses will be limited to a certain size
2832 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
2833 Thus, too large responses may not contain the mandatory pattern when using
2834 "string" or "rstring". If a large response is absolutely required, it is
2835 possible to change the default max size by setting the global variable.
2836 However, it is worth keeping in mind that parsing very large responses can
2837 waste some CPU cycles, especially when regular expressions are used, and that
2838 it is always better to focus the checks on smaller resources.
2839
2840 Last, if "http-check expect" is combined with "http-check disable-on-404",
2841 then this last one has precedence when the server responds with 404.
2842
2843 Examples :
2844 # only accept status 200 as valid
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01002845 http-check expect status 200
Willy Tarreaubd741542010-03-16 18:46:54 +01002846
2847 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01002848 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01002849
2850 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01002851 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01002852
2853 # check that we have a correct hexadecimal tag before /html
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01002854 http-check expect rstring <!--tag:[0-9a-f]*</html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01002855
Willy Tarreaubd741542010-03-16 18:46:54 +01002856 See also : "option httpchk", "http-check disable-on-404"
Willy Tarreau2769aa02007-12-27 18:26:09 +01002857
2858
Willy Tarreauef781042010-01-27 11:53:01 +01002859http-check send-state
2860 Enable emission of a state header with HTTP health checks
2861 May be used in sections : defaults | frontend | listen | backend
2862 yes | no | yes | yes
2863 Arguments : none
2864
2865 When this option is set, haproxy will systematically send a special header
2866 "X-Haproxy-Server-State" with a list of parameters indicating to each server
2867 how they are seen by haproxy. This can be used for instance when a server is
2868 manipulated without access to haproxy and the operator needs to know whether
2869 haproxy still sees it up or not, or if the server is the last one in a farm.
2870
2871 The header is composed of fields delimited by semi-colons, the first of which
2872 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
2873 checks on the total number before transition, just as appears in the stats
2874 interface. Next headers are in the form "<variable>=<value>", indicating in
2875 no specific order some values available in the stats interface :
2876 - a variable "name", containing the name of the backend followed by a slash
2877 ("/") then the name of the server. This can be used when a server is
2878 checked in multiple backends.
2879
2880 - a variable "node" containing the name of the haproxy node, as set in the
2881 global "node" variable, otherwise the system's hostname if unspecified.
2882
2883 - a variable "weight" indicating the weight of the server, a slash ("/")
2884 and the total weight of the farm (just counting usable servers). This
2885 helps to know if other servers are available to handle the load when this
2886 one fails.
2887
2888 - a variable "scur" indicating the current number of concurrent connections
2889 on the server, followed by a slash ("/") then the total number of
2890 connections on all servers of the same backend.
2891
2892 - a variable "qcur" indicating the current number of requests in the
2893 server's queue.
2894
2895 Example of a header received by the application server :
2896 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
2897 scur=13/22; qcur=0
2898
2899 See also : "option httpchk", "http-check disable-on-404"
2900
Willy Tarreauccbcc372012-12-27 12:37:57 +01002901http-request { allow | deny | tarpit | auth [realm <realm>] | redirect <rule> |
Willy Tarreauf4c43c12013-06-11 17:01:13 +02002902 add-header <name> <fmt> | set-header <name> <fmt> |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02002903 del-header <name> | set-nice <nice> | set-log-level <level> |
Sasha Pachev218f0642014-06-16 12:05:59 -06002904 replace-header <name> <match-regex> <replace-fmt> |
2905 replace-value <name> <match-regex> <replace-fmt> |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02002906 set-tos <tos> | set-mark <mark> |
2907 add-acl(<file name>) <key fmt> |
2908 del-acl(<file name>) <key fmt> |
2909 del-map(<file name>) <key fmt> |
2910 set-map(<file name>) <key fmt> <value fmt>
2911 }
Cyril Bontéf0c60612010-02-06 14:44:47 +01002912 [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002913 Access control for Layer 7 requests
2914
2915 May be used in sections: defaults | frontend | listen | backend
2916 no | yes | yes | yes
2917
Willy Tarreau20b0de52012-12-24 15:45:22 +01002918 The http-request statement defines a set of rules which apply to layer 7
2919 processing. The rules are evaluated in their declaration order when they are
2920 met in a frontend, listen or backend section. Any rule may optionally be
2921 followed by an ACL-based condition, in which case it will only be evaluated
2922 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002923
Willy Tarreau20b0de52012-12-24 15:45:22 +01002924 The first keyword is the rule's action. Currently supported actions include :
2925 - "allow" : this stops the evaluation of the rules and lets the request
2926 pass the check. No further "http-request" rules are evaluated.
2927
2928 - "deny" : this stops the evaluation of the rules and immediately rejects
2929 the request and emits an HTTP 403 error. No further "http-request" rules
2930 are evaluated.
2931
Willy Tarreauccbcc372012-12-27 12:37:57 +01002932 - "tarpit" : this stops the evaluation of the rules and immediately blocks
2933 the request without responding for a delay specified by "timeout tarpit"
2934 or "timeout connect" if the former is not set. After that delay, if the
2935 client is still connected, an HTTP error 500 is returned so that the
2936 client does not suspect it has been tarpitted. Logs will report the flags
2937 "PT". The goal of the tarpit rule is to slow down robots during an attack
2938 when they're limited on the number of concurrent requests. It can be very
2939 efficient against very dumb robots, and will significantly reduce the
2940 load on firewalls compared to a "deny" rule. But when facing "correctly"
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002941 developed robots, it can make things worse by forcing haproxy and the
Willy Tarreauccbcc372012-12-27 12:37:57 +01002942 front firewall to support insane number of concurrent connections.
2943
Willy Tarreau20b0de52012-12-24 15:45:22 +01002944 - "auth" : this stops the evaluation of the rules and immediately responds
2945 with an HTTP 401 or 407 error code to invite the user to present a valid
2946 user name and password. No further "http-request" rules are evaluated. An
2947 optional "realm" parameter is supported, it sets the authentication realm
2948 that is returned with the response (typically the application's name).
2949
Willy Tarreau81499eb2012-12-27 12:19:02 +01002950 - "redirect" : this performs an HTTP redirection based on a redirect rule.
2951 This is exactly the same as the "redirect" statement except that it
2952 inserts a redirect rule which can be processed in the middle of other
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01002953 "http-request" rules and that these rules use the "log-format" strings.
2954 See the "redirect" keyword for the rule's syntax.
Willy Tarreau81499eb2012-12-27 12:19:02 +01002955
Willy Tarreau20b0de52012-12-24 15:45:22 +01002956 - "add-header" appends an HTTP header field whose name is specified in
2957 <name> and whose value is defined by <fmt> which follows the log-format
2958 rules (see Custom Log Format in section 8.2.4). This is particularly
2959 useful to pass connection-specific information to the server (eg: the
2960 client's SSL certificate), or to combine several headers into one. This
2961 rule is not final, so it is possible to add other similar rules. Note
2962 that header addition is performed immediately, so one rule might reuse
2963 the resulting header from a previous rule.
2964
2965 - "set-header" does the same as "add-header" except that the header name
2966 is first removed if it existed. This is useful when passing security
2967 information to the server, where the header must not be manipulated by
2968 external users.
2969
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02002970 - "del-header" removes all HTTP header fields whose name is specified in
2971 <name>.
2972
Sasha Pachev218f0642014-06-16 12:05:59 -06002973 - "replace-header" matches the regular expression in all occurrences of
2974 header field <name> according to <match-regex>, and replaces them with
2975 the <replace-fmt> argument. Format characters are allowed in replace-fmt
2976 and work like in <fmt> arguments in "add-header". The match is only
2977 case-sensitive. It is important to understand that this action only
2978 considers whole header lines, regardless of the number of values they
2979 may contain. This usage is suited to headers naturally containing commas
2980 in their value, such as If-Modified-Since and so on.
2981
2982 Example:
2983
2984 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
2985
2986 applied to:
2987
2988 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
2989
2990 outputs:
2991
2992 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
2993
2994 assuming the backend IP is 192.168.1.20
2995
2996 - "replace-value" works like "replace-header" except that it matches the
2997 regex against every comma-delimited value of the header field <name>
2998 instead of the entire header. This is suited for all headers which are
2999 allowed to carry more than one value. An example could be the Accept
3000 header.
3001
3002 Example:
3003
3004 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
3005
3006 applied to:
3007
3008 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
3009
3010 outputs:
3011
3012 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
3013
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003014 - "set-nice" sets the "nice" factor of the current request being processed.
3015 It only has effect against the other requests being processed at the same
3016 time. The default value is 0, unless altered by the "nice" setting on the
3017 "bind" line. The accepted range is -1024..1024. The higher the value, the
3018 nicest the request will be. Lower values will make the request more
3019 important than other ones. This can be useful to improve the speed of
3020 some requests, or lower the priority of non-important requests. Using
3021 this setting without prior experimentation can cause some major slowdown.
3022
Willy Tarreau9a355ec2013-06-11 17:45:46 +02003023 - "set-log-level" is used to change the log level of the current request
3024 when a certain condition is met. Valid levels are the 8 syslog levels
3025 (see the "log" keyword) plus the special level "silent" which disables
3026 logging for this request. This rule is not final so the last matching
3027 rule wins. This rule can be useful to disable health checks coming from
3028 another equipment.
3029
Willy Tarreau42cf39e2013-06-11 18:51:32 +02003030 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
3031 the client to the value passed in <tos> on platforms which support this.
3032 This value represents the whole 8 bits of the IP TOS field, and can be
3033 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
3034 that only the 6 higher bits are used in DSCP or TOS, and the two lower
3035 bits are always 0. This can be used to adjust some routing behaviour on
3036 border routers based on some information from the request. See RFC 2474,
3037 2597, 3260 and 4594 for more information.
3038
Willy Tarreau51347ed2013-06-11 19:34:13 +02003039 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
3040 client to the value passed in <mark> on platforms which support it. This
3041 value is an unsigned 32 bit value which can be matched by netfilter and
3042 by the routing table. It can be expressed both in decimal or hexadecimal
3043 format (prefixed by "0x"). This can be useful to force certain packets to
3044 take a different route (for example a cheaper network path for bulk
3045 downloads). This works on Linux kernels 2.6.32 and above and requires
3046 admin privileges.
3047
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003048 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
3049 from a file (even a dummy empty file). The file name of the ACL to be
3050 updated is passed between parentheses. It takes one argument: <key fmt>,
3051 which follows log-format rules, to collect content of the new entry. It
3052 performs a lookup in the ACL before insertion, to avoid duplicated (or
3053 more) values. This lookup is done by a linear search and can be expensive
3054 with large lists! It is the equivalent of the "add acl" command from the
3055 stats socket, but can be triggered by an HTTP request.
3056
3057 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
3058 from a file (even a dummy empty file). The file name of the ACL to be
3059 updated is passed between parentheses. It takes one argument: <key fmt>,
3060 which follows log-format rules, to collect content of the entry to delete.
3061 It is the equivalent of the "del acl" command from the stats socket, but
3062 can be triggered by an HTTP request.
3063
3064 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
3065 from a file (even a dummy empty file). The file name of the MAP to be
3066 updated is passed between parentheses. It takes one argument: <key fmt>,
3067 which follows log-format rules, to collect content of the entry to delete.
3068 It takes one argument: "file name" It is the equivalent of the "del map"
3069 command from the stats socket, but can be triggered by an HTTP request.
3070
3071 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
3072 from a file (even a dummy empty file). The file name of the MAP to be
3073 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
3074 which follows log-format rules, used to collect MAP key, and <value fmt>,
3075 which follows log-format rules, used to collect content for the new entry.
3076 It performs a lookup in the MAP before insertion, to avoid duplicated (or
3077 more) values. This lookup is done by a linear search and can be expensive
3078 with large lists! It is the equivalent of the "set map" command from the
3079 stats socket, but can be triggered by an HTTP request.
3080
Willy Tarreau20b0de52012-12-24 15:45:22 +01003081 There is no limit to the number of http-request statements per instance.
3082
3083 It is important to know that http-request rules are processed very early in
3084 the HTTP processing, just after "block" rules and before "reqdel" or "reqrep"
3085 rules. That way, headers added by "add-header"/"set-header" are visible by
3086 almost all further ACL rules.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003087
3088 Example:
Cyril Bonté78caf842010-03-10 22:41:43 +01003089 acl nagios src 192.168.129.3
3090 acl local_net src 192.168.0.0/16
3091 acl auth_ok http_auth(L1)
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003092
Cyril Bonté78caf842010-03-10 22:41:43 +01003093 http-request allow if nagios
3094 http-request allow if local_net auth_ok
3095 http-request auth realm Gimme if local_net auth_ok
3096 http-request deny
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003097
Cyril Bonté78caf842010-03-10 22:41:43 +01003098 Example:
3099 acl auth_ok http_auth_group(L1) G1
Cyril Bonté78caf842010-03-10 22:41:43 +01003100 http-request auth unless auth_ok
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003101
Willy Tarreau20b0de52012-12-24 15:45:22 +01003102 Example:
3103 http-request set-header X-Haproxy-Current-Date %T
3104 http-request set-header X-SSL %[ssl_fc]
3105 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id]
3106 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
3107 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
3108 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
3109 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
3110 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
3111 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
3112
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003113 Example:
3114 acl key req.hdr(X-Add-Acl-Key) -m found
3115 acl add path /addacl
3116 acl del path /delacl
3117
3118 acl myhost hdr(Host) -f myhost.lst
3119
3120 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
3121 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
3122
3123 Example:
3124 acl value req.hdr(X-Value) -m found
3125 acl setmap path /setmap
3126 acl delmap path /delmap
3127
3128 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
3129
3130 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
3131 http-request del-map(map.lst) %[src] if delmap
3132
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02003133 See also : "stats http-request", section 3.4 about userlists and section 7
3134 about ACL usage.
Willy Tarreauef781042010-01-27 11:53:01 +01003135
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003136http-response { allow | deny | add-header <name> <fmt> | set-nice <nice> |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02003137 set-header <name> <fmt> | del-header <name> |
Sasha Pachev218f0642014-06-16 12:05:59 -06003138 replace-header <name> <regex-match> <replace-fmt> |
3139 replace-value <name> <regex-match> <replace-fmt> |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003140 set-log-level <level> | set-mark <mark> | set-tos <tos> |
3141 add-acl(<file name>) <key fmt> |
3142 del-acl(<file name>) <key fmt> |
3143 del-map(<file name>) <key fmt> |
3144 set-map(<file name>) <key fmt> <value fmt>
3145 }
Lukas Tribus2dd1d1a2013-06-19 23:34:41 +02003146 [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003147 Access control for Layer 7 responses
3148
3149 May be used in sections: defaults | frontend | listen | backend
3150 no | yes | yes | yes
3151
3152 The http-response statement defines a set of rules which apply to layer 7
3153 processing. The rules are evaluated in their declaration order when they are
3154 met in a frontend, listen or backend section. Any rule may optionally be
3155 followed by an ACL-based condition, in which case it will only be evaluated
3156 if the condition is true. Since these rules apply on responses, the backend
3157 rules are applied first, followed by the frontend's rules.
3158
3159 The first keyword is the rule's action. Currently supported actions include :
3160 - "allow" : this stops the evaluation of the rules and lets the response
3161 pass the check. No further "http-response" rules are evaluated for the
3162 current section.
3163
3164 - "deny" : this stops the evaluation of the rules and immediately rejects
3165 the response and emits an HTTP 502 error. No further "http-response"
3166 rules are evaluated.
3167
3168 - "add-header" appends an HTTP header field whose name is specified in
3169 <name> and whose value is defined by <fmt> which follows the log-format
3170 rules (see Custom Log Format in section 8.2.4). This may be used to send
3171 a cookie to a client for example, or to pass some internal information.
3172 This rule is not final, so it is possible to add other similar rules.
3173 Note that header addition is performed immediately, so one rule might
3174 reuse the resulting header from a previous rule.
3175
3176 - "set-header" does the same as "add-header" except that the header name
3177 is first removed if it existed. This is useful when passing security
3178 information to the server, where the header must not be manipulated by
3179 external users.
3180
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02003181 - "del-header" removes all HTTP header fields whose name is specified in
3182 <name>.
3183
Sasha Pachev218f0642014-06-16 12:05:59 -06003184 - "replace-header" matches the regular expression in all occurrences of
3185 header field <name> according to <match-regex>, and replaces them with
3186 the <replace-fmt> argument. Format characters are allowed in replace-fmt
3187 and work like in <fmt> arguments in "add-header". The match is only
3188 case-sensitive. It is important to understand that this action only
3189 considers whole header lines, regardless of the number of values they
3190 may contain. This usage is suited to headers naturally containing commas
3191 in their value, such as Set-Cookie, Expires and so on.
3192
3193 Example:
3194
3195 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
3196
3197 applied to:
3198
3199 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
3200
3201 outputs:
3202
3203 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
3204
3205 assuming the backend IP is 192.168.1.20.
3206
3207 - "replace-value" works like "replace-header" except that it matches the
3208 regex against every comma-delimited value of the header field <name>
3209 instead of the entire header. This is suited for all headers which are
3210 allowed to carry more than one value. An example could be the Accept
3211 header.
3212
3213 Example:
3214
3215 http-response replace-value Cache-control ^public$ private
3216
3217 applied to:
3218
3219 Cache-Control: max-age=3600, public
3220
3221 outputs:
3222
3223 Cache-Control: max-age=3600, private
3224
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003225 - "set-nice" sets the "nice" factor of the current request being processed.
3226 It only has effect against the other requests being processed at the same
3227 time. The default value is 0, unless altered by the "nice" setting on the
3228 "bind" line. The accepted range is -1024..1024. The higher the value, the
3229 nicest the request will be. Lower values will make the request more
3230 important than other ones. This can be useful to improve the speed of
3231 some requests, or lower the priority of non-important requests. Using
3232 this setting without prior experimentation can cause some major slowdown.
3233
Willy Tarreau9a355ec2013-06-11 17:45:46 +02003234 - "set-log-level" is used to change the log level of the current request
3235 when a certain condition is met. Valid levels are the 8 syslog levels
3236 (see the "log" keyword) plus the special level "silent" which disables
3237 logging for this request. This rule is not final so the last matching
3238 rule wins. This rule can be useful to disable health checks coming from
3239 another equipment.
3240
Willy Tarreau42cf39e2013-06-11 18:51:32 +02003241 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
3242 the client to the value passed in <tos> on platforms which support this.
3243 This value represents the whole 8 bits of the IP TOS field, and can be
3244 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
3245 that only the 6 higher bits are used in DSCP or TOS, and the two lower
3246 bits are always 0. This can be used to adjust some routing behaviour on
3247 border routers based on some information from the request. See RFC 2474,
3248 2597, 3260 and 4594 for more information.
3249
Willy Tarreau51347ed2013-06-11 19:34:13 +02003250 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
3251 client to the value passed in <mark> on platforms which support it. This
3252 value is an unsigned 32 bit value which can be matched by netfilter and
3253 by the routing table. It can be expressed both in decimal or hexadecimal
3254 format (prefixed by "0x"). This can be useful to force certain packets to
3255 take a different route (for example a cheaper network path for bulk
3256 downloads). This works on Linux kernels 2.6.32 and above and requires
3257 admin privileges.
3258
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003259 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
3260 from a file (even a dummy empty file). The file name of the ACL to be
3261 updated is passed between parentheses. It takes one argument: <key fmt>,
3262 which follows log-format rules, to collect content of the new entry. It
3263 performs a lookup in the ACL before insertion, to avoid duplicated (or
3264 more) values. This lookup is done by a linear search and can be expensive
3265 with large lists! It is the equivalent of the "add acl" command from the
3266 stats socket, but can be triggered by an HTTP response.
3267
3268 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
3269 from a file (even a dummy empty file). The file name of the ACL to be
3270 updated is passed between parentheses. It takes one argument: <key fmt>,
3271 which follows log-format rules, to collect content of the entry to delete.
3272 It is the equivalent of the "del acl" command from the stats socket, but
3273 can be triggered by an HTTP response.
3274
3275 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
3276 from a file (even a dummy empty file). The file name of the MAP to be
3277 updated is passed between parentheses. It takes one argument: <key fmt>,
3278 which follows log-format rules, to collect content of the entry to delete.
3279 It takes one argument: "file name" It is the equivalent of the "del map"
3280 command from the stats socket, but can be triggered by an HTTP response.
3281
3282 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
3283 from a file (even a dummy empty file). The file name of the MAP to be
3284 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
3285 which follows log-format rules, used to collect MAP key, and <value fmt>,
3286 which follows log-format rules, used to collect content for the new entry.
3287 It performs a lookup in the MAP before insertion, to avoid duplicated (or
3288 more) values. This lookup is done by a linear search and can be expensive
3289 with large lists! It is the equivalent of the "set map" command from the
3290 stats socket, but can be triggered by an HTTP response.
3291
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003292 There is no limit to the number of http-response statements per instance.
3293
Godbach09250262013-07-02 01:19:15 +08003294 It is important to know that http-response rules are processed very early in
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003295 the HTTP processing, before "reqdel" or "reqrep" rules. That way, headers
3296 added by "add-header"/"set-header" are visible by almost all further ACL
3297 rules.
3298
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003299 Example:
3300 acl key_acl res.hdr(X-Acl-Key) -m found
3301
3302 acl myhost hdr(Host) -f myhost.lst
3303
3304 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
3305 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
3306
3307 Example:
3308 acl value res.hdr(X-Value) -m found
3309
3310 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
3311
3312 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
3313 http-response del-map(map.lst) %[src] if ! value
3314
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003315 See also : "http-request", section 3.4 about userlists and section 7 about
3316 ACL usage.
3317
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02003318
Mark Lamourinec2247f02012-01-04 13:02:01 -05003319http-send-name-header [<header>]
3320 Add the server name to a request. Use the header string given by <header>
3321
3322 May be used in sections: defaults | frontend | listen | backend
3323 yes | no | yes | yes
3324
3325 Arguments :
3326
3327 <header> The header string to use to send the server name
3328
3329 The "http-send-name-header" statement causes the name of the target
3330 server to be added to the headers of an HTTP request. The name
3331 is added with the header string proved.
3332
3333 See also : "server"
3334
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01003335id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02003336 Set a persistent ID to a proxy.
3337 May be used in sections : defaults | frontend | listen | backend
3338 no | yes | yes | yes
3339 Arguments : none
3340
3341 Set a persistent ID for the proxy. This ID must be unique and positive.
3342 An unused ID will automatically be assigned if unset. The first assigned
3343 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01003344
3345
Cyril Bonté0d4bf012010-04-25 23:21:46 +02003346ignore-persist { if | unless } <condition>
3347 Declare a condition to ignore persistence
3348 May be used in sections: defaults | frontend | listen | backend
3349 no | yes | yes | yes
3350
3351 By default, when cookie persistence is enabled, every requests containing
3352 the cookie are unconditionally persistent (assuming the target server is up
3353 and running).
3354
3355 The "ignore-persist" statement allows one to declare various ACL-based
3356 conditions which, when met, will cause a request to ignore persistence.
3357 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003358 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02003359 persistence for a specific User-Agent (for example, some web crawler bots).
3360
3361 Combined with "appsession", it can also help reduce HAProxy memory usage, as
3362 the appsession table won't grow if persistence is ignored.
3363
3364 The persistence is ignored when an "if" condition is met, or unless an
3365 "unless" condition is met.
3366
3367 See also : "force-persist", "cookie", and section 7 about ACL usage.
3368
3369
Willy Tarreau2769aa02007-12-27 18:26:09 +01003370log global
Willy Tarreaudc2695c2014-06-27 18:10:07 +02003371log <address> [len <length>] <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02003372no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01003373 Enable per-instance logging of events and traffic.
3374 May be used in sections : defaults | frontend | listen | backend
3375 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02003376
3377 Prefix :
3378 no should be used when the logger list must be flushed. For example,
3379 if you don't want to inherit from the default logger list. This
3380 prefix does not allow arguments.
3381
Willy Tarreau2769aa02007-12-27 18:26:09 +01003382 Arguments :
3383 global should be used when the instance's logging parameters are the
3384 same as the global ones. This is the most common usage. "global"
3385 replaces <address>, <facility> and <level> with those of the log
3386 entries found in the "global" section. Only one "log global"
3387 statement may be used per instance, and this form takes no other
3388 parameter.
3389
3390 <address> indicates where to send the logs. It takes the same format as
3391 for the "global" section's logs, and can be one of :
3392
3393 - An IPv4 address optionally followed by a colon (':') and a UDP
3394 port. If no port is specified, 514 is used by default (the
3395 standard syslog port).
3396
David du Colombier24bb5f52011-03-17 10:40:23 +01003397 - An IPv6 address followed by a colon (':') and optionally a UDP
3398 port. If no port is specified, 514 is used by default (the
3399 standard syslog port).
3400
Willy Tarreau2769aa02007-12-27 18:26:09 +01003401 - A filesystem path to a UNIX domain socket, keeping in mind
3402 considerations for chroot (be sure the path is accessible
3403 inside the chroot) and uid/gid (be sure the path is
3404 appropriately writeable).
3405
Willy Tarreaudad36a32013-03-11 01:20:04 +01003406 Any part of the address string may reference any number of
3407 environment variables by preceding their name with a dollar
3408 sign ('$') and optionally enclosing them with braces ('{}'),
3409 similarly to what is done in Bourne shell.
3410
Willy Tarreaudc2695c2014-06-27 18:10:07 +02003411 <length> is an optional maximum line length. Log lines larger than this
3412 value will be truncated before being sent. The reason is that
3413 syslog servers act differently on log line length. All servers
3414 support the default value of 1024, but some servers simply drop
3415 larger lines while others do log them. If a server supports long
3416 lines, it may make sense to set this value here in order to avoid
3417 truncating long lines. Similarly, if a server drops long lines,
3418 it is preferable to truncate them before sending them. Accepted
3419 values are 80 to 65535 inclusive. The default value of 1024 is
3420 generally fine for all standard usages. Some specific cases of
3421 long captures or JSON-formated logs may require larger values.
3422
Willy Tarreau2769aa02007-12-27 18:26:09 +01003423 <facility> must be one of the 24 standard syslog facilities :
3424
3425 kern user mail daemon auth syslog lpr news
3426 uucp cron auth2 ftp ntp audit alert cron2
3427 local0 local1 local2 local3 local4 local5 local6 local7
3428
3429 <level> is optional and can be specified to filter outgoing messages. By
3430 default, all messages are sent. If a level is specified, only
3431 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02003432 will be sent. An optional minimum level can be specified. If it
3433 is set, logs emitted with a more severe level than this one will
3434 be capped to this level. This is used to avoid sending "emerg"
3435 messages on all terminals on some default syslog configurations.
3436 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01003437
3438 emerg alert crit err warning notice info debug
3439
William Lallemand0f99e342011-10-12 17:50:54 +02003440 It is important to keep in mind that it is the frontend which decides what to
3441 log from a connection, and that in case of content switching, the log entries
3442 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01003443
3444 However, backend log declaration define how and where servers status changes
3445 will be logged. Level "notice" will be used to indicate a server going up,
3446 "warning" will be used for termination signals and definitive service
3447 termination, and "alert" will be used for when a server goes down.
3448
3449 Note : According to RFC3164, messages are truncated to 1024 bytes before
3450 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003451
3452 Example :
3453 log global
Willy Tarreauf7edefa2009-05-10 17:20:05 +02003454 log 127.0.0.1:514 local0 notice # only send important events
3455 log 127.0.0.1:514 local0 notice notice # same but limit output level
Willy Tarreaudad36a32013-03-11 01:20:04 +01003456 log ${LOCAL_SYSLOG}:514 local0 notice # send to local server
3457
Willy Tarreau2769aa02007-12-27 18:26:09 +01003458
William Lallemand48940402012-01-30 16:47:22 +01003459log-format <string>
3460 Allows you to custom a log line.
3461
3462 See also : Custom Log Format (8.2.4)
3463
Willy Tarreau2769aa02007-12-27 18:26:09 +01003464
Willy Tarreauc35362a2014-04-25 13:58:37 +02003465max-keep-alive-queue <value>
3466 Set the maximum server queue size for maintaining keep-alive connections
3467 May be used in sections: defaults | frontend | listen | backend
3468 yes | no | yes | yes
3469
3470 HTTP keep-alive tries to reuse the same server connection whenever possible,
3471 but sometimes it can be counter-productive, for example if a server has a lot
3472 of connections while other ones are idle. This is especially true for static
3473 servers.
3474
3475 The purpose of this setting is to set a threshold on the number of queued
3476 connections at which haproxy stops trying to reuse the same server and prefers
3477 to find another one. The default value, -1, means there is no limit. A value
3478 of zero means that keep-alive requests will never be queued. For very close
3479 servers which can be reached with a low latency and which are not sensible to
3480 breaking keep-alive, a low value is recommended (eg: local static server can
3481 use a value of 10 or less). For remote servers suffering from a high latency,
3482 higher values might be needed to cover for the latency and/or the cost of
3483 picking a different server.
3484
3485 Note that this has no impact on responses which are maintained to the same
3486 server consecutively to a 401 response. They will still go to the same server
3487 even if they have to be queued.
3488
3489 See also : "option http-server-close", "option prefer-last-server", server
3490 "maxconn" and cookie persistence.
3491
3492
Willy Tarreau2769aa02007-12-27 18:26:09 +01003493maxconn <conns>
3494 Fix the maximum number of concurrent connections on a frontend
3495 May be used in sections : defaults | frontend | listen | backend
3496 yes | yes | yes | no
3497 Arguments :
3498 <conns> is the maximum number of concurrent connections the frontend will
3499 accept to serve. Excess connections will be queued by the system
3500 in the socket's listen queue and will be served once a connection
3501 closes.
3502
3503 If the system supports it, it can be useful on big sites to raise this limit
3504 very high so that haproxy manages connection queues, instead of leaving the
3505 clients with unanswered connection attempts. This value should not exceed the
3506 global maxconn. Also, keep in mind that a connection contains two buffers
3507 of 8kB each, as well as some other data resulting in about 17 kB of RAM being
3508 consumed per established connection. That means that a medium system equipped
3509 with 1GB of RAM can withstand around 40000-50000 concurrent connections if
3510 properly tuned.
3511
3512 Also, when <conns> is set to large values, it is possible that the servers
3513 are not sized to accept such loads, and for this reason it is generally wise
3514 to assign them some reasonable connection limits.
3515
Vincent Bernat6341be52012-06-27 17:18:30 +02003516 By default, this value is set to 2000.
3517
Willy Tarreau2769aa02007-12-27 18:26:09 +01003518 See also : "server", global section's "maxconn", "fullconn"
3519
3520
3521mode { tcp|http|health }
3522 Set the running mode or protocol of the instance
3523 May be used in sections : defaults | frontend | listen | backend
3524 yes | yes | yes | yes
3525 Arguments :
3526 tcp The instance will work in pure TCP mode. A full-duplex connection
3527 will be established between clients and servers, and no layer 7
3528 examination will be performed. This is the default mode. It
3529 should be used for SSL, SSH, SMTP, ...
3530
3531 http The instance will work in HTTP mode. The client request will be
3532 analyzed in depth before connecting to any server. Any request
3533 which is not RFC-compliant will be rejected. Layer 7 filtering,
3534 processing and switching will be possible. This is the mode which
3535 brings HAProxy most of its value.
3536
3537 health The instance will work in "health" mode. It will just reply "OK"
Willy Tarreau82569f92012-09-27 23:48:56 +02003538 to incoming connections and close the connection. Alternatively,
3539 If the "httpchk" option is set, "HTTP/1.0 200 OK" will be sent
3540 instead. Nothing will be logged in either case. This mode is used
3541 to reply to external components health checks. This mode is
3542 deprecated and should not be used anymore as it is possible to do
3543 the same and even better by combining TCP or HTTP modes with the
3544 "monitor" keyword.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003545
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003546 When doing content switching, it is mandatory that the frontend and the
3547 backend are in the same mode (generally HTTP), otherwise the configuration
3548 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003549
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003550 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01003551 defaults http_instances
3552 mode http
3553
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003554 See also : "monitor", "monitor-net"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003555
Willy Tarreau0ba27502007-12-24 16:55:16 +01003556
Cyril Bontéf0c60612010-02-06 14:44:47 +01003557monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01003558 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003559 May be used in sections : defaults | frontend | listen | backend
3560 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01003561 Arguments :
3562 if <cond> the monitor request will fail if the condition is satisfied,
3563 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003564 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01003565 are met, for instance a low number of servers both in a
3566 backend and its backup.
3567
3568 unless <cond> the monitor request will succeed only if the condition is
3569 satisfied, and will fail otherwise. Such a condition may be
3570 based on a test on the presence of a minimum number of active
3571 servers in a list of backends.
3572
3573 This statement adds a condition which can force the response to a monitor
3574 request to report a failure. By default, when an external component queries
3575 the URI dedicated to monitoring, a 200 response is returned. When one of the
3576 conditions above is met, haproxy will return 503 instead of 200. This is
3577 very useful to report a site failure to an external component which may base
3578 routing advertisements between multiple sites on the availability reported by
3579 haproxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003580 criterion. Note that "monitor fail" only works in HTTP mode. Both status
3581 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003582
3583 Example:
3584 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01003585 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01003586 acl site_dead nbsrv(dynamic) lt 2
3587 acl site_dead nbsrv(static) lt 2
3588 monitor-uri /site_alive
3589 monitor fail if site_dead
3590
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003591 See also : "monitor-net", "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003592
3593
3594monitor-net <source>
3595 Declare a source network which is limited to monitor requests
3596 May be used in sections : defaults | frontend | listen | backend
3597 yes | yes | yes | no
3598 Arguments :
3599 <source> is the source IPv4 address or network which will only be able to
3600 get monitor responses to any request. It can be either an IPv4
3601 address, a host name, or an address followed by a slash ('/')
3602 followed by a mask.
3603
3604 In TCP mode, any connection coming from a source matching <source> will cause
3605 the connection to be immediately closed without any log. This allows another
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003606 equipment to probe the port and verify that it is still listening, without
Willy Tarreau2769aa02007-12-27 18:26:09 +01003607 forwarding the connection to a remote server.
3608
3609 In HTTP mode, a connection coming from a source matching <source> will be
3610 accepted, the following response will be sent without waiting for a request,
3611 then the connection will be closed : "HTTP/1.0 200 OK". This is normally
3612 enough for any front-end HTTP probe to detect that the service is UP and
Willy Tarreau82569f92012-09-27 23:48:56 +02003613 running without forwarding the request to a backend server. Note that this
3614 response is sent in raw format, without any transformation. This is important
3615 as it means that it will not be SSL-encrypted on SSL listeners.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003616
Willy Tarreau82569f92012-09-27 23:48:56 +02003617 Monitor requests are processed very early, just after tcp-request connection
3618 ACLs which are the only ones able to block them. These connections are short
3619 lived and never wait for any data from the client. They cannot be logged, and
3620 it is the intended purpose. They are only used to report HAProxy's health to
3621 an upper component, nothing more. Please note that "monitor fail" rules do
3622 not apply to connections intercepted by "monitor-net".
Willy Tarreau2769aa02007-12-27 18:26:09 +01003623
Willy Tarreau95cd2832010-03-04 23:36:33 +01003624 Last, please note that only one "monitor-net" statement can be specified in
3625 a frontend. If more than one is found, only the last one will be considered.
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003626
Willy Tarreau2769aa02007-12-27 18:26:09 +01003627 Example :
3628 # addresses .252 and .253 are just probing us.
3629 frontend www
3630 monitor-net 192.168.0.252/31
3631
3632 See also : "monitor fail", "monitor-uri"
3633
3634
3635monitor-uri <uri>
3636 Intercept a URI used by external components' monitor requests
3637 May be used in sections : defaults | frontend | listen | backend
3638 yes | yes | yes | no
3639 Arguments :
3640 <uri> is the exact URI which we want to intercept to return HAProxy's
3641 health status instead of forwarding the request.
3642
3643 When an HTTP request referencing <uri> will be received on a frontend,
3644 HAProxy will not forward it nor log it, but instead will return either
3645 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
3646 conditions defined with "monitor fail". This is normally enough for any
3647 front-end HTTP probe to detect that the service is UP and running without
3648 forwarding the request to a backend server. Note that the HTTP method, the
3649 version and all headers are ignored, but the request must at least be valid
3650 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
3651
3652 Monitor requests are processed very early. It is not possible to block nor
3653 divert them using ACLs. They cannot be logged either, and it is the intended
3654 purpose. They are only used to report HAProxy's health to an upper component,
3655 nothing more. However, it is possible to add any number of conditions using
3656 "monitor fail" and ACLs so that the result can be adjusted to whatever check
3657 can be imagined (most often the number of available servers in a backend).
3658
3659 Example :
3660 # Use /haproxy_test to report haproxy's status
3661 frontend www
3662 mode http
3663 monitor-uri /haproxy_test
3664
3665 See also : "monitor fail", "monitor-net"
3666
Willy Tarreau0ba27502007-12-24 16:55:16 +01003667
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003668option abortonclose
3669no option abortonclose
3670 Enable or disable early dropping of aborted requests pending in queues.
3671 May be used in sections : defaults | frontend | listen | backend
3672 yes | no | yes | yes
3673 Arguments : none
3674
3675 In presence of very high loads, the servers will take some time to respond.
3676 The per-instance connection queue will inflate, and the response time will
3677 increase respective to the size of the queue times the average per-session
3678 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01003679 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003680 the queue, and slowing down other users, and the servers as well, because the
3681 request will eventually be served, then aborted at the first error
3682 encountered while delivering the response.
3683
3684 As there is no way to distinguish between a full STOP and a simple output
3685 close on the client side, HTTP agents should be conservative and consider
3686 that the client might only have closed its output channel while waiting for
3687 the response. However, this introduces risks of congestion when lots of users
3688 do the same, and is completely useless nowadays because probably no client at
3689 all will close the session while waiting for the response. Some HTTP agents
Willy Tarreaud72758d2010-01-12 10:42:19 +01003690 support this behaviour (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003691 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01003692 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003693 of being the single component to break rare but valid traffic is extremely
3694 low, which adds to the temptation to be able to abort a session early while
3695 still not served and not pollute the servers.
3696
3697 In HAProxy, the user can choose the desired behaviour using the option
3698 "abortonclose". By default (without the option) the behaviour is HTTP
3699 compliant and aborted requests will be served. But when the option is
3700 specified, a session with an incoming channel closed will be aborted while
3701 it is still possible, either pending in the queue for a connection slot, or
3702 during the connection establishment if the server has not yet acknowledged
3703 the connection request. This considerably reduces the queue size and the load
3704 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01003705 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003706
3707 If this option has been enabled in a "defaults" section, it can be disabled
3708 in a specific instance by prepending the "no" keyword before it.
3709
3710 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
3711
3712
Willy Tarreau4076a152009-04-02 15:18:36 +02003713option accept-invalid-http-request
3714no option accept-invalid-http-request
3715 Enable or disable relaxing of HTTP request parsing
3716 May be used in sections : defaults | frontend | listen | backend
3717 yes | yes | yes | no
3718 Arguments : none
3719
3720 By default, HAProxy complies with RFC2616 in terms of message parsing. This
3721 means that invalid characters in header names are not permitted and cause an
3722 error to be returned to the client. This is the desired behaviour as such
3723 forbidden characters are essentially used to build attacks exploiting server
3724 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
3725 server will emit invalid header names for whatever reason (configuration,
3726 implementation) and the issue will not be immediately fixed. In such a case,
3727 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01003728 even if that does not make sense, by specifying this option. Similarly, the
3729 list of characters allowed to appear in a URI is well defined by RFC3986, and
3730 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
3731 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
3732 not allowed at all. Haproxy always blocks a number of them (0..32, 127). The
3733 remaining ones are blocked by default unless this option is enabled.
Willy Tarreau4076a152009-04-02 15:18:36 +02003734
3735 This option should never be enabled by default as it hides application bugs
3736 and open security breaches. It should only be deployed after a problem has
3737 been confirmed.
3738
3739 When this option is enabled, erroneous header names will still be accepted in
3740 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01003741 analysis using the "show errors" request on the UNIX stats socket. Similarly,
3742 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02003743 also helps confirming that the issue has been solved.
3744
3745 If this option has been enabled in a "defaults" section, it can be disabled
3746 in a specific instance by prepending the "no" keyword before it.
3747
3748 See also : "option accept-invalid-http-response" and "show errors" on the
3749 stats socket.
3750
3751
3752option accept-invalid-http-response
3753no option accept-invalid-http-response
3754 Enable or disable relaxing of HTTP response parsing
3755 May be used in sections : defaults | frontend | listen | backend
3756 yes | no | yes | yes
3757 Arguments : none
3758
3759 By default, HAProxy complies with RFC2616 in terms of message parsing. This
3760 means that invalid characters in header names are not permitted and cause an
3761 error to be returned to the client. This is the desired behaviour as such
3762 forbidden characters are essentially used to build attacks exploiting server
3763 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
3764 server will emit invalid header names for whatever reason (configuration,
3765 implementation) and the issue will not be immediately fixed. In such a case,
3766 it is possible to relax HAProxy's header name parser to accept any character
3767 even if that does not make sense, by specifying this option.
3768
3769 This option should never be enabled by default as it hides application bugs
3770 and open security breaches. It should only be deployed after a problem has
3771 been confirmed.
3772
3773 When this option is enabled, erroneous header names will still be accepted in
3774 responses, but the complete response will be captured in order to permit
3775 later analysis using the "show errors" request on the UNIX stats socket.
3776 Doing this also helps confirming that the issue has been solved.
3777
3778 If this option has been enabled in a "defaults" section, it can be disabled
3779 in a specific instance by prepending the "no" keyword before it.
3780
3781 See also : "option accept-invalid-http-request" and "show errors" on the
3782 stats socket.
3783
3784
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003785option allbackups
3786no option allbackups
3787 Use either all backup servers at a time or only the first one
3788 May be used in sections : defaults | frontend | listen | backend
3789 yes | no | yes | yes
3790 Arguments : none
3791
3792 By default, the first operational backup server gets all traffic when normal
3793 servers are all down. Sometimes, it may be preferred to use multiple backups
3794 at once, because one will not be enough. When "option allbackups" is enabled,
3795 the load balancing will be performed among all backup servers when all normal
3796 ones are unavailable. The same load balancing algorithm will be used and the
3797 servers' weights will be respected. Thus, there will not be any priority
3798 order between the backup servers anymore.
3799
3800 This option is mostly used with static server farms dedicated to return a
3801 "sorry" page when an application is completely offline.
3802
3803 If this option has been enabled in a "defaults" section, it can be disabled
3804 in a specific instance by prepending the "no" keyword before it.
3805
3806
3807option checkcache
3808no option checkcache
Godbach7056a352013-12-11 20:01:07 +08003809 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003810 May be used in sections : defaults | frontend | listen | backend
3811 yes | no | yes | yes
3812 Arguments : none
3813
3814 Some high-level frameworks set application cookies everywhere and do not
3815 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003816 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003817 high risk of session crossing or stealing between users traversing the same
3818 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02003819 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003820
3821 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003822 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01003823 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003824 response to check if there's a risk of caching a cookie on a client-side
3825 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01003826 to the client are :
3827 - all those without "Set-Cookie" header ;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003828 - all those with a return code other than 200, 203, 206, 300, 301, 410,
Willy Tarreau198a7442008-01-17 12:05:32 +01003829 provided that the server has not set a "Cache-control: public" header ;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003830 - all those that come from a POST request, provided that the server has not
3831 set a 'Cache-Control: public' header ;
3832 - those with a 'Pragma: no-cache' header
3833 - those with a 'Cache-control: private' header
3834 - those with a 'Cache-control: no-store' header
3835 - those with a 'Cache-control: max-age=0' header
3836 - those with a 'Cache-control: s-maxage=0' header
3837 - those with a 'Cache-control: no-cache' header
3838 - those with a 'Cache-control: no-cache="set-cookie"' header
3839 - those with a 'Cache-control: no-cache="set-cookie,' header
3840 (allowing other fields after set-cookie)
3841
3842 If a response doesn't respect these requirements, then it will be blocked
Willy Tarreau198a7442008-01-17 12:05:32 +01003843 just as if it was from an "rspdeny" filter, with an "HTTP 502 bad gateway".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003844 The session state shows "PH--" meaning that the proxy blocked the response
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003845 during headers processing. Additionally, an alert will be sent in the logs so
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003846 that admins are informed that there's something to be fixed.
3847
3848 Due to the high impact on the application, the application should be tested
3849 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003850 good practice to always activate it during tests, even if it is not used in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003851 production, as it will report potentially dangerous application behaviours.
3852
3853 If this option has been enabled in a "defaults" section, it can be disabled
3854 in a specific instance by prepending the "no" keyword before it.
3855
3856
3857option clitcpka
3858no option clitcpka
3859 Enable or disable the sending of TCP keepalive packets on the client side
3860 May be used in sections : defaults | frontend | listen | backend
3861 yes | yes | yes | no
3862 Arguments : none
3863
3864 When there is a firewall or any session-aware component between a client and
3865 a server, and when the protocol involves very long sessions with long idle
3866 periods (eg: remote desktops), there is a risk that one of the intermediate
3867 components decides to expire a session which has remained idle for too long.
3868
3869 Enabling socket-level TCP keep-alives makes the system regularly send packets
3870 to the other end of the connection, leaving it active. The delay between
3871 keep-alive probes is controlled by the system only and depends both on the
3872 operating system and its tuning parameters.
3873
3874 It is important to understand that keep-alive packets are neither emitted nor
3875 received at the application level. It is only the network stacks which sees
3876 them. For this reason, even if one side of the proxy already uses keep-alives
3877 to maintain its connection alive, those keep-alive packets will not be
3878 forwarded to the other side of the proxy.
3879
3880 Please note that this has nothing to do with HTTP keep-alive.
3881
3882 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
3883 client side of a connection, which should help when session expirations are
3884 noticed between HAProxy and a client.
3885
3886 If this option has been enabled in a "defaults" section, it can be disabled
3887 in a specific instance by prepending the "no" keyword before it.
3888
3889 See also : "option srvtcpka", "option tcpka"
3890
3891
Willy Tarreau0ba27502007-12-24 16:55:16 +01003892option contstats
3893 Enable continuous traffic statistics updates
3894 May be used in sections : defaults | frontend | listen | backend
3895 yes | yes | yes | no
3896 Arguments : none
3897
3898 By default, counters used for statistics calculation are incremented
3899 only when a session finishes. It works quite well when serving small
3900 objects, but with big ones (for example large images or archives) or
3901 with A/V streaming, a graph generated from haproxy counters looks like
3902 a hedgehog. With this option enabled counters get incremented continuously,
3903 during a whole session. Recounting touches a hotpath directly so
3904 it is not enabled by default, as it has small performance impact (~0.5%).
3905
3906
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02003907option dontlog-normal
3908no option dontlog-normal
3909 Enable or disable logging of normal, successful connections
3910 May be used in sections : defaults | frontend | listen | backend
3911 yes | yes | yes | no
3912 Arguments : none
3913
3914 There are large sites dealing with several thousand connections per second
3915 and for which logging is a major pain. Some of them are even forced to turn
3916 logs off and cannot debug production issues. Setting this option ensures that
3917 normal connections, those which experience no error, no timeout, no retry nor
3918 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
3919 mode, the response status code is checked and return codes 5xx will still be
3920 logged.
3921
3922 It is strongly discouraged to use this option as most of the time, the key to
3923 complex issues is in the normal logs which will not be logged here. If you
3924 need to separate logs, see the "log-separate-errors" option instead.
3925
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003926 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02003927 logging.
3928
3929
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003930option dontlognull
3931no option dontlognull
3932 Enable or disable logging of null connections
3933 May be used in sections : defaults | frontend | listen | backend
3934 yes | yes | yes | no
3935 Arguments : none
3936
3937 In certain environments, there are components which will regularly connect to
3938 various systems to ensure that they are still alive. It can be the case from
3939 another load balancer as well as from monitoring systems. By default, even a
3940 simple port probe or scan will produce a log. If those connections pollute
3941 the logs too much, it is possible to enable option "dontlognull" to indicate
3942 that a connection on which no data has been transferred will not be logged,
3943 which typically corresponds to those probes.
3944
3945 It is generally recommended not to use this option in uncontrolled
3946 environments (eg: internet), otherwise scans and other malicious activities
3947 would not be logged.
3948
3949 If this option has been enabled in a "defaults" section, it can be disabled
3950 in a specific instance by prepending the "no" keyword before it.
3951
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003952 See also : "log", "monitor-net", "monitor-uri" and section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003953
3954
3955option forceclose
3956no option forceclose
3957 Enable or disable active connection closing after response is transferred.
3958 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaua31e5df2009-12-30 01:10:35 +01003959 yes | yes | yes | yes
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003960 Arguments : none
3961
3962 Some HTTP servers do not necessarily close the connections when they receive
3963 the "Connection: close" set by "option httpclose", and if the client does not
3964 close either, then the connection remains open till the timeout expires. This
3965 causes high number of simultaneous connections on the servers and shows high
3966 global session times in the logs.
3967
3968 When this happens, it is possible to use "option forceclose". It will
Willy Tarreau82eeaf22009-12-29 12:09:05 +01003969 actively close the outgoing server channel as soon as the server has finished
Cyril Bonté653dcd62014-02-20 00:13:15 +01003970 to respond and release some resources earlier than with "option httpclose".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003971
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02003972 This option may also be combined with "option http-pretend-keepalive", which
3973 will disable sending of the "Connection: close" header, but will still cause
3974 the connection to be closed once the whole response is received.
3975
Cyril Bonté653dcd62014-02-20 00:13:15 +01003976 This option disables and replaces any previous "option httpclose", "option
3977 http-server-close", "option http-keep-alive", or "option http-tunnel".
Willy Tarreau02bce8b2014-01-30 00:15:28 +01003978
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003979 If this option has been enabled in a "defaults" section, it can be disabled
3980 in a specific instance by prepending the "no" keyword before it.
3981
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02003982 See also : "option httpclose" and "option http-pretend-keepalive"
Willy Tarreaubf1f8162007-12-28 17:42:56 +01003983
3984
Willy Tarreau87cf5142011-08-19 22:57:24 +02003985option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01003986 Enable insertion of the X-Forwarded-For header to requests sent to servers
3987 May be used in sections : defaults | frontend | listen | backend
3988 yes | yes | yes | yes
3989 Arguments :
3990 <network> is an optional argument used to disable this option for sources
3991 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02003992 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01003993 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01003994
3995 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
3996 their client address. This is sometimes annoying when the client's IP address
3997 is expected in server logs. To solve this problem, the well-known HTTP header
3998 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
3999 This header contains a value representing the client's IP address. Since this
4000 header is always appended at the end of the existing header list, the server
4001 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02004002 the server's manual to find how to enable use of this standard header. Note
4003 that only the last occurrence of the header must be used, since it is really
4004 possible that the client has already brought one.
4005
Willy Tarreaud72758d2010-01-12 10:42:19 +01004006 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02004007 the default "X-Forwarded-For". This can be useful where you might already
Willy Tarreaud72758d2010-01-12 10:42:19 +01004008 have a "X-Forwarded-For" header from a different application (eg: stunnel),
4009 and you need preserve it. Also if your backend server doesn't use the
Ross Westaf72a1d2008-08-03 10:51:45 +02004010 "X-Forwarded-For" header and requires different one (eg: Zeus Web Servers
4011 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01004012
4013 Sometimes, a same HAProxy instance may be shared between a direct client
4014 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
4015 used to decrypt HTTPS traffic). It is possible to disable the addition of the
4016 header for a known source address or network by adding the "except" keyword
4017 followed by the network address. In this case, any source IP matching the
4018 network will not cause an addition of this header. Most common uses are with
4019 private networks or 127.0.0.1.
4020
Willy Tarreau87cf5142011-08-19 22:57:24 +02004021 Alternatively, the keyword "if-none" states that the header will only be
4022 added if it is not present. This should only be used in perfectly trusted
4023 environment, as this might cause a security issue if headers reaching haproxy
4024 are under the control of the end-user.
4025
Willy Tarreauc27debf2008-01-06 08:57:02 +01004026 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02004027 least one of them uses it, the header will be added. Note that the backend's
4028 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02004029 both are defined. In the case of the "if-none" argument, if at least one of
4030 the frontend or the backend does not specify it, it wants the addition to be
4031 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004032
Ross Westaf72a1d2008-08-03 10:51:45 +02004033 Examples :
Willy Tarreauc27debf2008-01-06 08:57:02 +01004034 # Public HTTP address also used by stunnel on the same machine
4035 frontend www
4036 mode http
4037 option forwardfor except 127.0.0.1 # stunnel already adds the header
4038
Ross Westaf72a1d2008-08-03 10:51:45 +02004039 # Those servers want the IP Address in X-Client
4040 backend www
4041 mode http
4042 option forwardfor header X-Client
4043
Willy Tarreau87cf5142011-08-19 22:57:24 +02004044 See also : "option httpclose", "option http-server-close",
Willy Tarreau70dffda2014-01-30 03:07:23 +01004045 "option forceclose", "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01004046
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02004047
Willy Tarreau16bfb022010-01-16 19:48:41 +01004048option http-keep-alive
4049no option http-keep-alive
4050 Enable or disable HTTP keep-alive from client to server
4051 May be used in sections : defaults | frontend | listen | backend
4052 yes | yes | yes | yes
4053 Arguments : none
4054
Willy Tarreau70dffda2014-01-30 03:07:23 +01004055 By default HAProxy operates in keep-alive mode with regards to persistent
4056 connections: for each connection it processes each request and response, and
4057 leaves the connection idle on both sides between the end of a response and the
4058 start of a new request. This mode may be changed by several options such as
4059 "option http-server-close", "option forceclose", "option httpclose" or
4060 "option http-tunnel". This option allows to set back the keep-alive mode,
4061 which can be useful when another mode was used in a defaults section.
4062
4063 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
4064 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01004065 network) and the fastest session reuse on the server side at the expense
4066 of maintaining idle connections to the servers. In general, it is possible
4067 with this option to achieve approximately twice the request rate that the
4068 "http-server-close" option achieves on small objects. There are mainly two
4069 situations where this option may be useful :
4070
4071 - when the server is non-HTTP compliant and authenticates the connection
4072 instead of requests (eg: NTLM authentication)
4073
4074 - when the cost of establishing the connection to the server is significant
4075 compared to the cost of retrieving the associated object from the server.
4076
4077 This last case can happen when the server is a fast static server of cache.
4078 In this case, the server will need to be properly tuned to support high enough
4079 connection counts because connections will last until the client sends another
4080 request.
4081
4082 If the client request has to go to another backend or another server due to
4083 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01004084 immediately be closed and a new one re-opened. Option "prefer-last-server" is
4085 available to try optimize server selection so that if the server currently
4086 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01004087
4088 In general it is preferred to use "option http-server-close" with application
4089 servers, and some static servers might benefit from "option http-keep-alive".
4090
4091 At the moment, logs will not indicate whether requests came from the same
4092 session or not. The accept date reported in the logs corresponds to the end
4093 of the previous request, and the request time corresponds to the time spent
4094 waiting for a new request. The keep-alive request time is still bound to the
4095 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
4096 not set.
4097
Cyril Bonté653dcd62014-02-20 00:13:15 +01004098 This option disables and replaces any previous "option httpclose", "option
4099 http-server-close", "option forceclose" or "option http-tunnel". When backend
Willy Tarreau70dffda2014-01-30 03:07:23 +01004100 and frontend options differ, all of these 4 options have precedence over
Cyril Bonté653dcd62014-02-20 00:13:15 +01004101 "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01004102
4103 See also : "option forceclose", "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01004104 "option prefer-last-server", "option http-pretend-keepalive",
4105 "option httpclose", and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01004106
4107
Willy Tarreau96e31212011-05-30 18:10:30 +02004108option http-no-delay
4109no option http-no-delay
4110 Instruct the system to favor low interactive delays over performance in HTTP
4111 May be used in sections : defaults | frontend | listen | backend
4112 yes | yes | yes | yes
4113 Arguments : none
4114
4115 In HTTP, each payload is unidirectional and has no notion of interactivity.
4116 Any agent is expected to queue data somewhat for a reasonably low delay.
4117 There are some very rare server-to-server applications that abuse the HTTP
4118 protocol and expect the payload phase to be highly interactive, with many
4119 interleaved data chunks in both directions within a single request. This is
4120 absolutely not supported by the HTTP specification and will not work across
4121 most proxies or servers. When such applications attempt to do this through
4122 haproxy, it works but they will experience high delays due to the network
4123 optimizations which favor performance by instructing the system to wait for
4124 enough data to be available in order to only send full packets. Typical
4125 delays are around 200 ms per round trip. Note that this only happens with
4126 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
4127 affected.
4128
4129 When "option http-no-delay" is present in either the frontend or the backend
4130 used by a connection, all such optimizations will be disabled in order to
4131 make the exchanges as fast as possible. Of course this offers no guarantee on
4132 the functionality, as it may break at any other place. But if it works via
4133 HAProxy, it will work as fast as possible. This option should never be used
4134 by default, and should never be used at all unless such a buggy application
4135 is discovered. The impact of using this option is an increase of bandwidth
4136 usage and CPU usage, which may significantly lower performance in high
4137 latency environments.
4138
4139
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02004140option http-pretend-keepalive
4141no option http-pretend-keepalive
4142 Define whether haproxy will announce keepalive to the server or not
4143 May be used in sections : defaults | frontend | listen | backend
4144 yes | yes | yes | yes
4145 Arguments : none
4146
4147 When running with "option http-server-close" or "option forceclose", haproxy
4148 adds a "Connection: close" header to the request forwarded to the server.
4149 Unfortunately, when some servers see this header, they automatically refrain
4150 from using the chunked encoding for responses of unknown length, while this
4151 is totally unrelated. The immediate effect is that this prevents haproxy from
4152 maintaining the client connection alive. A second effect is that a client or
4153 a cache could receive an incomplete response without being aware of it, and
4154 consider the response complete.
4155
4156 By setting "option http-pretend-keepalive", haproxy will make the server
4157 believe it will keep the connection alive. The server will then not fall back
4158 to the abnormal undesired above. When haproxy gets the whole response, it
4159 will close the connection with the server just as it would do with the
4160 "forceclose" option. That way the client gets a normal response and the
4161 connection is correctly closed on the server side.
4162
4163 It is recommended not to enable this option by default, because most servers
4164 will more efficiently close the connection themselves after the last packet,
4165 and release its buffers slightly earlier. Also, the added packet on the
4166 network could slightly reduce the overall peak performance. However it is
4167 worth noting that when this option is enabled, haproxy will have slightly
4168 less work to do. So if haproxy is the bottleneck on the whole architecture,
4169 enabling this option might save a few CPU cycles.
4170
4171 This option may be set both in a frontend and in a backend. It is enabled if
4172 at least one of the frontend or backend holding a connection has it enabled.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004173 This option may be combined with "option httpclose", which will cause
Willy Tarreau22a95342010-09-29 14:31:41 +02004174 keepalive to be announced to the server and close to be announced to the
4175 client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02004176
4177 If this option has been enabled in a "defaults" section, it can be disabled
4178 in a specific instance by prepending the "no" keyword before it.
4179
Willy Tarreau16bfb022010-01-16 19:48:41 +01004180 See also : "option forceclose", "option http-server-close", and
4181 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02004182
Willy Tarreauc27debf2008-01-06 08:57:02 +01004183
Willy Tarreaub608feb2010-01-02 22:47:18 +01004184option http-server-close
4185no option http-server-close
4186 Enable or disable HTTP connection closing on the server side
4187 May be used in sections : defaults | frontend | listen | backend
4188 yes | yes | yes | yes
4189 Arguments : none
4190
Willy Tarreau70dffda2014-01-30 03:07:23 +01004191 By default HAProxy operates in keep-alive mode with regards to persistent
4192 connections: for each connection it processes each request and response, and
4193 leaves the connection idle on both sides between the end of a response and
4194 the start of a new request. This mode may be changed by several options such
4195 as "option http-server-close", "option forceclose", "option httpclose" or
4196 "option http-tunnel". Setting "option http-server-close" enables HTTP
4197 connection-close mode on the server side while keeping the ability to support
4198 HTTP keep-alive and pipelining on the client side. This provides the lowest
4199 latency on the client side (slow network) and the fastest session reuse on
4200 the server side to save server resources, similarly to "option forceclose".
4201 It also permits non-keepalive capable servers to be served in keep-alive mode
4202 to the clients if they conform to the requirements of RFC2616. Please note
4203 that some servers do not always conform to those requirements when they see
4204 "Connection: close" in the request. The effect will be that keep-alive will
4205 never be used. A workaround consists in enabling "option
4206 http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01004207
4208 At the moment, logs will not indicate whether requests came from the same
4209 session or not. The accept date reported in the logs corresponds to the end
4210 of the previous request, and the request time corresponds to the time spent
4211 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01004212 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
4213 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01004214
4215 This option may be set both in a frontend and in a backend. It is enabled if
4216 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01004217 It disables and replaces any previous "option httpclose", "option forceclose",
4218 "option http-tunnel" or "option http-keep-alive". Please check section 4
Willy Tarreau70dffda2014-01-30 03:07:23 +01004219 ("Proxies") to see how this option combines with others when frontend and
4220 backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01004221
4222 If this option has been enabled in a "defaults" section, it can be disabled
4223 in a specific instance by prepending the "no" keyword before it.
4224
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02004225 See also : "option forceclose", "option http-pretend-keepalive",
Willy Tarreau16bfb022010-01-16 19:48:41 +01004226 "option httpclose", "option http-keep-alive", and
4227 "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01004228
4229
Willy Tarreau02bce8b2014-01-30 00:15:28 +01004230option http-tunnel
4231no option http-tunnel
4232 Disable or enable HTTP connection processing after first transaction
4233 May be used in sections : defaults | frontend | listen | backend
4234 yes | yes | yes | yes
4235 Arguments : none
4236
Willy Tarreau70dffda2014-01-30 03:07:23 +01004237 By default HAProxy operates in keep-alive mode with regards to persistent
4238 connections: for each connection it processes each request and response, and
4239 leaves the connection idle on both sides between the end of a response and
4240 the start of a new request. This mode may be changed by several options such
4241 as "option http-server-close", "option forceclose", "option httpclose" or
4242 "option http-tunnel".
4243
4244 Option "http-tunnel" disables any HTTP processing past the first request and
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03004245 the first response. This is the mode which was used by default in versions
Willy Tarreau70dffda2014-01-30 03:07:23 +01004246 1.0 to 1.5-dev21. It is the mode with the lowest processing overhead, which
4247 is normally not needed anymore unless in very specific cases such as when
4248 using an in-house protocol that looks like HTTP but is not compatible, or
4249 just to log one request per client in order to reduce log size. Note that
4250 everything which works at the HTTP level, including header parsing/addition,
4251 cookie processing or content switching will only work for the first request
4252 and will be ignored after the first response.
Willy Tarreau02bce8b2014-01-30 00:15:28 +01004253
4254 If this option has been enabled in a "defaults" section, it can be disabled
4255 in a specific instance by prepending the "no" keyword before it.
4256
4257 See also : "option forceclose", "option http-server-close",
4258 "option httpclose", "option http-keep-alive", and
4259 "1.1. The HTTP transaction model".
4260
4261
Willy Tarreau88d349d2010-01-25 12:15:43 +01004262option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01004263no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01004264 Make use of non-standard Proxy-Connection header instead of Connection
4265 May be used in sections : defaults | frontend | listen | backend
4266 yes | yes | yes | no
4267 Arguments : none
4268
4269 While RFC2616 explicitly states that HTTP/1.1 agents must use the
4270 Connection header to indicate their wish of persistent or non-persistent
4271 connections, both browsers and proxies ignore this header for proxied
4272 connections and make use of the undocumented, non-standard Proxy-Connection
4273 header instead. The issue begins when trying to put a load balancer between
4274 browsers and such proxies, because there will be a difference between what
4275 haproxy understands and what the client and the proxy agree on.
4276
4277 By setting this option in a frontend, haproxy can automatically switch to use
4278 that non-standard header if it sees proxied requests. A proxied request is
4279 defined here as one where the URI begins with neither a '/' nor a '*'. The
4280 choice of header only affects requests passing through proxies making use of
4281 one of the "httpclose", "forceclose" and "http-server-close" options. Note
4282 that this option can only be specified in a frontend and will affect the
4283 request along its whole life.
4284
Willy Tarreau844a7e72010-01-31 21:46:18 +01004285 Also, when this option is set, a request which requires authentication will
4286 automatically switch to use proxy authentication headers if it is itself a
4287 proxied request. That makes it possible to check or enforce authentication in
4288 front of an existing proxy.
4289
Willy Tarreau88d349d2010-01-25 12:15:43 +01004290 This option should normally never be used, except in front of a proxy.
4291
4292 See also : "option httpclose", "option forceclose" and "option
4293 http-server-close".
4294
4295
Willy Tarreaud63335a2010-02-26 12:56:52 +01004296option httpchk
4297option httpchk <uri>
4298option httpchk <method> <uri>
4299option httpchk <method> <uri> <version>
4300 Enable HTTP protocol to check on the servers health
4301 May be used in sections : defaults | frontend | listen | backend
4302 yes | no | yes | yes
4303 Arguments :
4304 <method> is the optional HTTP method used with the requests. When not set,
4305 the "OPTIONS" method is used, as it generally requires low server
4306 processing and is easy to filter out from the logs. Any method
4307 may be used, though it is not recommended to invent non-standard
4308 ones.
4309
4310 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
4311 which is accessible by default on almost any server, but may be
4312 changed to any other URI. Query strings are permitted.
4313
4314 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
4315 but some servers might behave incorrectly in HTTP 1.0, so turning
4316 it to HTTP/1.1 may sometimes help. Note that the Host field is
4317 mandatory in HTTP/1.1, and as a trick, it is possible to pass it
4318 after "\r\n" following the version string.
4319
4320 By default, server health checks only consist in trying to establish a TCP
4321 connection. When "option httpchk" is specified, a complete HTTP request is
4322 sent once the TCP connection is established, and responses 2xx and 3xx are
4323 considered valid, while all other ones indicate a server failure, including
4324 the lack of any response.
4325
4326 The port and interval are specified in the server configuration.
4327
4328 This option does not necessarily require an HTTP backend, it also works with
4329 plain TCP backends. This is particularly useful to check simple scripts bound
4330 to some dedicated ports using the inetd daemon.
4331
4332 Examples :
4333 # Relay HTTPS traffic to Apache instance and check service availability
4334 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
4335 backend https_relay
4336 mode tcp
4337 option httpchk OPTIONS * HTTP/1.1\r\nHost:\ www
4338 server apache1 192.168.1.1:443 check port 80
4339
Simon Hormanafc47ee2013-11-25 10:46:35 +09004340 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
4341 "option pgsql-check", "http-check" and the "check", "port" and
4342 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01004343
4344
Willy Tarreauc27debf2008-01-06 08:57:02 +01004345option httpclose
4346no option httpclose
4347 Enable or disable passive HTTP connection closing
4348 May be used in sections : defaults | frontend | listen | backend
4349 yes | yes | yes | yes
4350 Arguments : none
4351
Willy Tarreau70dffda2014-01-30 03:07:23 +01004352 By default HAProxy operates in keep-alive mode with regards to persistent
4353 connections: for each connection it processes each request and response, and
4354 leaves the connection idle on both sides between the end of a response and
4355 the start of a new request. This mode may be changed by several options such
Cyril Bonté653dcd62014-02-20 00:13:15 +01004356 as "option http-server-close", "option forceclose", "option httpclose" or
Willy Tarreau70dffda2014-01-30 03:07:23 +01004357 "option http-tunnel".
4358
4359 If "option httpclose" is set, HAProxy will work in HTTP tunnel mode and check
4360 if a "Connection: close" header is already set in each direction, and will
4361 add one if missing. Each end should react to this by actively closing the TCP
4362 connection after each transfer, thus resulting in a switch to the HTTP close
4363 mode. Any "Connection" header different from "close" will also be removed.
4364 Note that this option is deprecated since what it does is very cheap but not
4365 reliable. Using "option http-server-close" or "option forceclose" is strongly
4366 recommended instead.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004367
4368 It seldom happens that some servers incorrectly ignore this header and do not
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004369 close the connection even though they reply "Connection: close". For this
Willy Tarreau0dfdf192010-01-05 11:33:11 +01004370 reason, they are not compatible with older HTTP 1.0 browsers. If this happens
4371 it is possible to use the "option forceclose" which actively closes the
4372 request connection once the server responds. Option "forceclose" also
4373 releases the server connection earlier because it does not have to wait for
4374 the client to acknowledge it.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004375
4376 This option may be set both in a frontend and in a backend. It is enabled if
4377 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01004378 It disables and replaces any previous "option http-server-close",
4379 "option forceclose", "option http-keep-alive" or "option http-tunnel". Please
Willy Tarreau70dffda2014-01-30 03:07:23 +01004380 check section 4 ("Proxies") to see how this option combines with others when
4381 frontend and backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004382
4383 If this option has been enabled in a "defaults" section, it can be disabled
4384 in a specific instance by prepending the "no" keyword before it.
4385
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02004386 See also : "option forceclose", "option http-server-close" and
4387 "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01004388
4389
Emeric Brun3a058f32009-06-30 18:26:00 +02004390option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01004391 Enable logging of HTTP request, session state and timers
4392 May be used in sections : defaults | frontend | listen | backend
4393 yes | yes | yes | yes
Emeric Brun3a058f32009-06-30 18:26:00 +02004394 Arguments :
4395 clf if the "clf" argument is added, then the output format will be
4396 the CLF format instead of HAProxy's default HTTP format. You can
4397 use this when you need to feed HAProxy's logs through a specific
4398 log analyser which only support the CLF format and which is not
4399 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004400
4401 By default, the log output format is very poor, as it only contains the
4402 source and destination addresses, and the instance name. By specifying
4403 "option httplog", each log line turns into a much richer format including,
4404 but not limited to, the HTTP request, the connection timers, the session
4405 status, the connections numbers, the captured headers and cookies, the
4406 frontend, backend and server name, and of course the source address and
4407 ports.
4408
4409 This option may be set either in the frontend or the backend.
4410
Emeric Brun3a058f32009-06-30 18:26:00 +02004411 If this option has been enabled in a "defaults" section, it can be disabled
4412 in a specific instance by prepending the "no" keyword before it. Specifying
4413 only "option httplog" will automatically clear the 'clf' mode if it was set
4414 by default.
4415
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004416 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01004417
Willy Tarreau55165fe2009-05-10 12:02:55 +02004418
4419option http_proxy
4420no option http_proxy
4421 Enable or disable plain HTTP proxy mode
4422 May be used in sections : defaults | frontend | listen | backend
4423 yes | yes | yes | yes
4424 Arguments : none
4425
4426 It sometimes happens that people need a pure HTTP proxy which understands
4427 basic proxy requests without caching nor any fancy feature. In this case,
4428 it may be worth setting up an HAProxy instance with the "option http_proxy"
4429 set. In this mode, no server is declared, and the connection is forwarded to
4430 the IP address and port found in the URL after the "http://" scheme.
4431
4432 No host address resolution is performed, so this only works when pure IP
4433 addresses are passed. Since this option's usage perimeter is rather limited,
4434 it will probably be used only by experts who know they need exactly it. Last,
4435 if the clients are susceptible of sending keep-alive requests, it will be
Cyril Bonté2409e682010-12-14 22:47:51 +01004436 needed to add "option httpclose" to ensure that all requests will correctly
Willy Tarreau55165fe2009-05-10 12:02:55 +02004437 be analyzed.
4438
4439 If this option has been enabled in a "defaults" section, it can be disabled
4440 in a specific instance by prepending the "no" keyword before it.
4441
4442 Example :
4443 # this backend understands HTTP proxy requests and forwards them directly.
4444 backend direct_forward
4445 option httpclose
4446 option http_proxy
4447
4448 See also : "option httpclose"
4449
Willy Tarreau211ad242009-10-03 21:45:07 +02004450
Jamie Gloudon801a0a32012-08-25 00:18:33 -04004451option independent-streams
4452no option independent-streams
4453 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02004454 May be used in sections : defaults | frontend | listen | backend
4455 yes | yes | yes | yes
4456 Arguments : none
4457
4458 By default, when data is sent over a socket, both the write timeout and the
4459 read timeout for that socket are refreshed, because we consider that there is
4460 activity on that socket, and we have no other means of guessing if we should
4461 receive data or not.
4462
4463 While this default behaviour is desirable for almost all applications, there
4464 exists a situation where it is desirable to disable it, and only refresh the
4465 read timeout if there are incoming data. This happens on sessions with large
4466 timeouts and low amounts of exchanged data such as telnet session. If the
4467 server suddenly disappears, the output data accumulates in the system's
4468 socket buffers, both timeouts are correctly refreshed, and there is no way
4469 to know the server does not receive them, so we don't timeout. However, when
4470 the underlying protocol always echoes sent data, it would be enough by itself
4471 to detect the issue using the read timeout. Note that this problem does not
4472 happen with more verbose protocols because data won't accumulate long in the
4473 socket buffers.
4474
4475 When this option is set on the frontend, it will disable read timeout updates
4476 on data sent to the client. There probably is little use of this case. When
4477 the option is set on the backend, it will disable read timeout updates on
4478 data sent to the server. Doing so will typically break large HTTP posts from
4479 slow lines, so use it with caution.
4480
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03004481 Note: older versions used to call this setting "option independent-streams"
Jamie Gloudon801a0a32012-08-25 00:18:33 -04004482 with a spelling mistake. This spelling is still supported but
4483 deprecated.
4484
Willy Tarreauce887fd2012-05-12 12:50:00 +02004485 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02004486
4487
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02004488option ldap-check
4489 Use LDAPv3 health checks for server testing
4490 May be used in sections : defaults | frontend | listen | backend
4491 yes | no | yes | yes
4492 Arguments : none
4493
4494 It is possible to test that the server correctly talks LDAPv3 instead of just
4495 testing that it accepts the TCP connection. When this option is set, an
4496 LDAPv3 anonymous simple bind message is sent to the server, and the response
4497 is analyzed to find an LDAPv3 bind response message.
4498
4499 The server is considered valid only when the LDAP response contains success
4500 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
4501
4502 Logging of bind requests is server dependent see your documentation how to
4503 configure it.
4504
4505 Example :
4506 option ldap-check
4507
4508 See also : "option httpchk"
4509
4510
Willy Tarreau211ad242009-10-03 21:45:07 +02004511option log-health-checks
4512no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02004513 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02004514 May be used in sections : defaults | frontend | listen | backend
4515 yes | no | yes | yes
4516 Arguments : none
4517
Willy Tarreaubef1b322014-05-13 21:01:39 +02004518 By default, failed health check are logged if server is UP and successful
4519 health checks are logged if server is DOWN, so the amount of additional
4520 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02004521
Willy Tarreaubef1b322014-05-13 21:01:39 +02004522 When this option is enabled, any change of the health check status or to
4523 the server's health will be logged, so that it becomes possible to know
4524 that a server was failing occasional checks before crashing, or exactly when
4525 it failed to respond a valid HTTP status, then when the port started to
4526 reject connections, then when the server stopped responding at all.
4527
4528 Note that status changes not caused by health checks (eg: enable/disable on
4529 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02004530
Willy Tarreaubef1b322014-05-13 21:01:39 +02004531 See also: "option httpchk", "option ldap-check", "option mysql-check",
4532 "option pgsql-check", "option redis-check", "option smtpchk",
4533 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02004534
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02004535
4536option log-separate-errors
4537no option log-separate-errors
4538 Change log level for non-completely successful connections
4539 May be used in sections : defaults | frontend | listen | backend
4540 yes | yes | yes | no
4541 Arguments : none
4542
4543 Sometimes looking for errors in logs is not easy. This option makes haproxy
4544 raise the level of logs containing potentially interesting information such
4545 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
4546 level changes from "info" to "err". This makes it possible to log them
4547 separately to a different file with most syslog daemons. Be careful not to
4548 remove them from the original file, otherwise you would lose ordering which
4549 provides very important information.
4550
4551 Using this option, large sites dealing with several thousand connections per
4552 second may log normal traffic to a rotating buffer and only archive smaller
4553 error logs.
4554
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004555 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02004556 logging.
4557
Willy Tarreauc27debf2008-01-06 08:57:02 +01004558
4559option logasap
4560no option logasap
4561 Enable or disable early logging of HTTP requests
4562 May be used in sections : defaults | frontend | listen | backend
4563 yes | yes | yes | no
4564 Arguments : none
4565
4566 By default, HTTP requests are logged upon termination so that the total
4567 transfer time and the number of bytes appear in the logs. When large objects
4568 are being transferred, it may take a while before the request appears in the
4569 logs. Using "option logasap", the request gets logged as soon as the server
4570 sends the complete headers. The only missing information in the logs will be
4571 the total number of bytes which will indicate everything except the amount
4572 of data transferred, and the total time which will not take the transfer
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004573 time into account. In such a situation, it's a good practice to capture the
Willy Tarreauc27debf2008-01-06 08:57:02 +01004574 "Content-Length" response header so that the logs at least indicate how many
4575 bytes are expected to be transferred.
4576
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004577 Examples :
4578 listen http_proxy 0.0.0.0:80
4579 mode http
4580 option httplog
4581 option logasap
4582 log 192.168.2.200 local3
4583
4584 >>> Feb 6 12:14:14 localhost \
4585 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
4586 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
4587 "GET /image.iso HTTP/1.0"
4588
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004589 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01004590 logging.
4591
4592
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02004593option mysql-check [ user <username> [ post-41 ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02004594 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01004595 May be used in sections : defaults | frontend | listen | backend
4596 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02004597 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004598 <username> This is the username which will be used when connecting to MySQL
4599 server.
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02004600 post-41 Send post v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02004601
4602 If you specify a username, the check consists of sending two MySQL packet,
4603 one Client Authentication packet, and one QUIT packet, to correctly close
4604 MySQL session. We then parse the MySQL Handshake Initialisation packet and/or
4605 Error packet. It is a basic but useful test which does not produce error nor
4606 aborted connect on the server. However, it requires adding an authorization
4607 in the MySQL table, like this :
4608
4609 USE mysql;
4610 INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');
4611 FLUSH PRIVILEGES;
4612
4613 If you don't specify a username (it is deprecated and not recommended), the
4614 check only consists in parsing the Mysql Handshake Initialisation packet or
4615 Error packet, we don't send anything in this mode. It was reported that it
4616 can generate lockout if check is too frequent and/or if there is not enough
4617 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
4618 value as if a connection is established successfully within fewer than MySQL
4619 "max_connect_errors" attempts after a previous connection was interrupted,
4620 the error count for the host is cleared to zero. If HAProxy's server get
4621 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
4622
4623 Remember that this does not check database presence nor database consistency.
4624 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01004625
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02004626 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01004627
4628 Most often, an incoming MySQL server needs to see the client's IP address for
4629 various purposes, including IP privilege matching and connection logging.
4630 When possible, it is often wise to masquerade the client's IP address when
4631 connecting to the server using the "usesrc" argument of the "source" keyword,
4632 which requires the cttproxy feature to be compiled in, and the MySQL server
4633 to route the client via the machine hosting haproxy.
4634
4635 See also: "option httpchk"
4636
4637
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004638option nolinger
4639no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004640 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004641 May be used in sections: defaults | frontend | listen | backend
4642 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01004643 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004644
4645 When clients or servers abort connections in a dirty way (eg: they are
4646 physically disconnected), the session timeouts triggers and the session is
4647 closed. But it will remain in FIN_WAIT1 state for some time in the system,
4648 using some resources and possibly limiting the ability to establish newer
4649 connections.
4650
4651 When this happens, it is possible to activate "option nolinger" which forces
4652 the system to immediately remove any socket's pending data on close. Thus,
4653 the session is instantly purged from the system's tables. This usually has
4654 side effects such as increased number of TCP resets due to old retransmits
4655 getting immediately rejected. Some firewalls may sometimes complain about
4656 this too.
4657
4658 For this reason, it is not recommended to use this option when not absolutely
4659 needed. You know that you need it when you have thousands of FIN_WAIT1
4660 sessions on your system (TIME_WAIT ones do not count).
4661
4662 This option may be used both on frontends and backends, depending on the side
4663 where it is required. Use it on the frontend for clients, and on the backend
4664 for servers.
4665
4666 If this option has been enabled in a "defaults" section, it can be disabled
4667 in a specific instance by prepending the "no" keyword before it.
4668
4669
Willy Tarreau55165fe2009-05-10 12:02:55 +02004670option originalto [ except <network> ] [ header <name> ]
4671 Enable insertion of the X-Original-To header to requests sent to servers
4672 May be used in sections : defaults | frontend | listen | backend
4673 yes | yes | yes | yes
4674 Arguments :
4675 <network> is an optional argument used to disable this option for sources
4676 matching <network>
4677 <name> an optional argument to specify a different "X-Original-To"
4678 header name.
4679
4680 Since HAProxy can work in transparent mode, every request from a client can
4681 be redirected to the proxy and HAProxy itself can proxy every request to a
4682 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
4683 be lost. This is annoying when you want access rules based on destination ip
4684 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
4685 added by HAProxy to all requests sent to the server. This header contains a
4686 value representing the original destination IP address. Since this must be
4687 configured to always use the last occurrence of this header only. Note that
4688 only the last occurrence of the header must be used, since it is really
4689 possible that the client has already brought one.
4690
4691 The keyword "header" may be used to supply a different header name to replace
4692 the default "X-Original-To". This can be useful where you might already
4693 have a "X-Original-To" header from a different application, and you need
4694 preserve it. Also if your backend server doesn't use the "X-Original-To"
4695 header and requires different one.
4696
4697 Sometimes, a same HAProxy instance may be shared between a direct client
4698 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
4699 used to decrypt HTTPS traffic). It is possible to disable the addition of the
4700 header for a known source address or network by adding the "except" keyword
4701 followed by the network address. In this case, any source IP matching the
4702 network will not cause an addition of this header. Most common uses are with
4703 private networks or 127.0.0.1.
4704
4705 This option may be specified either in the frontend or in the backend. If at
4706 least one of them uses it, the header will be added. Note that the backend's
4707 setting of the header subargument takes precedence over the frontend's if
4708 both are defined.
4709
Willy Tarreau55165fe2009-05-10 12:02:55 +02004710 Examples :
4711 # Original Destination address
4712 frontend www
4713 mode http
4714 option originalto except 127.0.0.1
4715
4716 # Those servers want the IP Address in X-Client-Dst
4717 backend www
4718 mode http
4719 option originalto header X-Client-Dst
4720
Willy Tarreau87cf5142011-08-19 22:57:24 +02004721 See also : "option httpclose", "option http-server-close",
4722 "option forceclose"
Willy Tarreau55165fe2009-05-10 12:02:55 +02004723
4724
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004725option persist
4726no option persist
4727 Enable or disable forced persistence on down servers
4728 May be used in sections: defaults | frontend | listen | backend
4729 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01004730 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004731
4732 When an HTTP request reaches a backend with a cookie which references a dead
4733 server, by default it is redispatched to another server. It is possible to
4734 force the request to be sent to the dead server first using "option persist"
4735 if absolutely needed. A common use case is when servers are under extreme
4736 load and spend their time flapping. In this case, the users would still be
4737 directed to the server they opened the session on, in the hope they would be
4738 correctly served. It is recommended to use "option redispatch" in conjunction
4739 with this option so that in the event it would not be possible to connect to
4740 the server at all (server definitely dead), the client would finally be
4741 redirected to another valid server.
4742
4743 If this option has been enabled in a "defaults" section, it can be disabled
4744 in a specific instance by prepending the "no" keyword before it.
4745
Willy Tarreau4de91492010-01-22 19:10:05 +01004746 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004747
4748
Willy Tarreau0c122822013-12-15 18:49:01 +01004749option pgsql-check [ user <username> ]
4750 Use PostgreSQL health checks for server testing
4751 May be used in sections : defaults | frontend | listen | backend
4752 yes | no | yes | yes
4753 Arguments :
4754 <username> This is the username which will be used when connecting to
4755 PostgreSQL server.
4756
4757 The check sends a PostgreSQL StartupMessage and waits for either
4758 Authentication request or ErrorResponse message. It is a basic but useful
4759 test which does not produce error nor aborted connect on the server.
4760 This check is identical with the "mysql-check".
4761
4762 See also: "option httpchk"
4763
4764
Willy Tarreau9420b122013-12-15 18:58:25 +01004765option prefer-last-server
4766no option prefer-last-server
4767 Allow multiple load balanced requests to remain on the same server
4768 May be used in sections: defaults | frontend | listen | backend
4769 yes | no | yes | yes
4770 Arguments : none
4771
4772 When the load balancing algorithm in use is not deterministic, and a previous
4773 request was sent to a server to which haproxy still holds a connection, it is
4774 sometimes desirable that subsequent requests on a same session go to the same
4775 server as much as possible. Note that this is different from persistence, as
4776 we only indicate a preference which haproxy tries to apply without any form
4777 of warranty. The real use is for keep-alive connections sent to servers. When
4778 this option is used, haproxy will try to reuse the same connection that is
4779 attached to the server instead of rebalancing to another server, causing a
4780 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01004781 not make much sense to use this in combination with hashing algorithms. Note,
4782 haproxy already automatically tries to stick to a server which sends a 401 or
4783 to a proxy which sends a 407 (authentication required). This is mandatory for
4784 use with the broken NTLM authentication challenge, and significantly helps in
4785 troubleshooting some faulty applications. Option prefer-last-server might be
4786 desirable in these environments as well, to avoid redistributing the traffic
4787 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01004788
4789 If this option has been enabled in a "defaults" section, it can be disabled
4790 in a specific instance by prepending the "no" keyword before it.
4791
4792 See also: "option http-keep-alive"
4793
4794
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01004795option redispatch
4796no option redispatch
4797 Enable or disable session redistribution in case of connection failure
4798 May be used in sections: defaults | frontend | listen | backend
4799 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01004800 Arguments : none
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01004801
4802 In HTTP mode, if a server designated by a cookie is down, clients may
4803 definitely stick to it because they cannot flush the cookie, so they will not
4804 be able to access the service anymore.
4805
4806 Specifying "option redispatch" will allow the proxy to break their
4807 persistence and redistribute them to a working server.
4808
4809 It also allows to retry last connection to another server in case of multiple
4810 connection failures. Of course, it requires having "retries" set to a nonzero
4811 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01004812
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01004813 This form is the preferred form, which replaces both the "redispatch" and
4814 "redisp" keywords.
4815
4816 If this option has been enabled in a "defaults" section, it can be disabled
4817 in a specific instance by prepending the "no" keyword before it.
4818
Willy Tarreau4de91492010-01-22 19:10:05 +01004819 See also : "redispatch", "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01004820
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004821
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02004822option redis-check
4823 Use redis health checks for server testing
4824 May be used in sections : defaults | frontend | listen | backend
4825 yes | no | yes | yes
4826 Arguments : none
4827
4828 It is possible to test that the server correctly talks REDIS protocol instead
4829 of just testing that it accepts the TCP connection. When this option is set,
4830 a PING redis command is sent to the server, and the response is analyzed to
4831 find the "+PONG" response message.
4832
4833 Example :
4834 option redis-check
4835
4836 See also : "option httpchk"
4837
4838
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004839option smtpchk
4840option smtpchk <hello> <domain>
4841 Use SMTP health checks for server testing
4842 May be used in sections : defaults | frontend | listen | backend
4843 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01004844 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01004845 <hello> is an optional argument. It is the "hello" command to use. It can
4846 be either "HELO" (for SMTP) or "EHLO" (for ESTMP). All other
4847 values will be turned into the default command ("HELO").
4848
4849 <domain> is the domain name to present to the server. It may only be
4850 specified (and is mandatory) if the hello command has been
4851 specified. By default, "localhost" is used.
4852
4853 When "option smtpchk" is set, the health checks will consist in TCP
4854 connections followed by an SMTP command. By default, this command is
4855 "HELO localhost". The server's return code is analyzed and only return codes
4856 starting with a "2" will be considered as valid. All other responses,
4857 including a lack of response will constitute an error and will indicate a
4858 dead server.
4859
4860 This test is meant to be used with SMTP servers or relays. Depending on the
4861 request, it is possible that some servers do not log each connection attempt,
4862 so you may want to experiment to improve the behaviour. Using telnet on port
4863 25 is often easier than adjusting the configuration.
4864
4865 Most often, an incoming SMTP server needs to see the client's IP address for
4866 various purposes, including spam filtering, anti-spoofing and logging. When
4867 possible, it is often wise to masquerade the client's IP address when
4868 connecting to the server using the "usesrc" argument of the "source" keyword,
4869 which requires the cttproxy feature to be compiled in.
4870
4871 Example :
4872 option smtpchk HELO mydomain.org
4873
4874 See also : "option httpchk", "source"
4875
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01004876
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02004877option socket-stats
4878no option socket-stats
4879
4880 Enable or disable collecting & providing separate statistics for each socket.
4881 May be used in sections : defaults | frontend | listen | backend
4882 yes | yes | yes | no
4883
4884 Arguments : none
4885
4886
Willy Tarreauff4f82d2009-02-06 11:28:13 +01004887option splice-auto
4888no option splice-auto
4889 Enable or disable automatic kernel acceleration on sockets in both directions
4890 May be used in sections : defaults | frontend | listen | backend
4891 yes | yes | yes | yes
4892 Arguments : none
4893
4894 When this option is enabled either on a frontend or on a backend, haproxy
4895 will automatically evaluate the opportunity to use kernel tcp splicing to
4896 forward data between the client and the server, in either direction. Haproxy
4897 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004898 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01004899 are not much aggressive in order to limit excessive use of splicing. This
4900 option requires splicing to be enabled at compile time, and may be globally
4901 disabled with the global option "nosplice". Since splice uses pipes, using it
4902 requires that there are enough spare pipes.
4903
4904 Important note: kernel-based TCP splicing is a Linux-specific feature which
4905 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
4906 transfer data between sockets without copying these data to user-space, thus
4907 providing noticeable performance gains and CPU cycles savings. Since many
4908 early implementations are buggy, corrupt data and/or are inefficient, this
4909 feature is not enabled by default, and it should be used with extreme care.
4910 While it is not possible to detect the correctness of an implementation,
4911 2.6.29 is the first version offering a properly working implementation. In
4912 case of doubt, splicing may be globally disabled using the global "nosplice"
4913 keyword.
4914
4915 Example :
4916 option splice-auto
4917
4918 If this option has been enabled in a "defaults" section, it can be disabled
4919 in a specific instance by prepending the "no" keyword before it.
4920
4921 See also : "option splice-request", "option splice-response", and global
4922 options "nosplice" and "maxpipes"
4923
4924
4925option splice-request
4926no option splice-request
4927 Enable or disable automatic kernel acceleration on sockets for requests
4928 May be used in sections : defaults | frontend | listen | backend
4929 yes | yes | yes | yes
4930 Arguments : none
4931
4932 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004933 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01004934 the client to the server. It might still use the recv/send scheme if there
4935 are no spare pipes left. This option requires splicing to be enabled at
4936 compile time, and may be globally disabled with the global option "nosplice".
4937 Since splice uses pipes, using it requires that there are enough spare pipes.
4938
4939 Important note: see "option splice-auto" for usage limitations.
4940
4941 Example :
4942 option splice-request
4943
4944 If this option has been enabled in a "defaults" section, it can be disabled
4945 in a specific instance by prepending the "no" keyword before it.
4946
4947 See also : "option splice-auto", "option splice-response", and global options
4948 "nosplice" and "maxpipes"
4949
4950
4951option splice-response
4952no option splice-response
4953 Enable or disable automatic kernel acceleration on sockets for responses
4954 May be used in sections : defaults | frontend | listen | backend
4955 yes | yes | yes | yes
4956 Arguments : none
4957
4958 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004959 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01004960 the server to the client. It might still use the recv/send scheme if there
4961 are no spare pipes left. This option requires splicing to be enabled at
4962 compile time, and may be globally disabled with the global option "nosplice".
4963 Since splice uses pipes, using it requires that there are enough spare pipes.
4964
4965 Important note: see "option splice-auto" for usage limitations.
4966
4967 Example :
4968 option splice-response
4969
4970 If this option has been enabled in a "defaults" section, it can be disabled
4971 in a specific instance by prepending the "no" keyword before it.
4972
4973 See also : "option splice-auto", "option splice-request", and global options
4974 "nosplice" and "maxpipes"
4975
4976
Willy Tarreaubf1f8162007-12-28 17:42:56 +01004977option srvtcpka
4978no option srvtcpka
4979 Enable or disable the sending of TCP keepalive packets on the server side
4980 May be used in sections : defaults | frontend | listen | backend
4981 yes | no | yes | yes
4982 Arguments : none
4983
4984 When there is a firewall or any session-aware component between a client and
4985 a server, and when the protocol involves very long sessions with long idle
4986 periods (eg: remote desktops), there is a risk that one of the intermediate
4987 components decides to expire a session which has remained idle for too long.
4988
4989 Enabling socket-level TCP keep-alives makes the system regularly send packets
4990 to the other end of the connection, leaving it active. The delay between
4991 keep-alive probes is controlled by the system only and depends both on the
4992 operating system and its tuning parameters.
4993
4994 It is important to understand that keep-alive packets are neither emitted nor
4995 received at the application level. It is only the network stacks which sees
4996 them. For this reason, even if one side of the proxy already uses keep-alives
4997 to maintain its connection alive, those keep-alive packets will not be
4998 forwarded to the other side of the proxy.
4999
5000 Please note that this has nothing to do with HTTP keep-alive.
5001
5002 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
5003 server side of a connection, which should help when session expirations are
5004 noticed between HAProxy and a server.
5005
5006 If this option has been enabled in a "defaults" section, it can be disabled
5007 in a specific instance by prepending the "no" keyword before it.
5008
5009 See also : "option clitcpka", "option tcpka"
5010
5011
Willy Tarreaua453bdd2008-01-08 19:50:52 +01005012option ssl-hello-chk
5013 Use SSLv3 client hello health checks for server testing
5014 May be used in sections : defaults | frontend | listen | backend
5015 yes | no | yes | yes
5016 Arguments : none
5017
5018 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
5019 possible to test that the server correctly talks SSL instead of just testing
5020 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
5021 SSLv3 client hello messages are sent once the connection is established to
5022 the server, and the response is analyzed to find an SSL server hello message.
5023 The server is considered valid only when the response contains this server
5024 hello message.
5025
5026 All servers tested till there correctly reply to SSLv3 client hello messages,
5027 and most servers tested do not even log the requests containing only hello
5028 messages, which is appreciable.
5029
Willy Tarreau763a95b2012-10-04 23:15:39 +02005030 Note that this check works even when SSL support was not built into haproxy
5031 because it forges the SSL message. When SSL support is available, it is best
5032 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01005033
Willy Tarreau763a95b2012-10-04 23:15:39 +02005034 See also: "option httpchk", "check-ssl"
5035
Willy Tarreaua453bdd2008-01-08 19:50:52 +01005036
Willy Tarreaued179852013-12-16 01:07:00 +01005037option tcp-check
5038 Perform health checks using tcp-check send/expect sequences
5039 May be used in sections: defaults | frontend | listen | backend
5040 yes | no | yes | yes
5041
5042 This health check method is intended to be combined with "tcp-check" command
5043 lists in order to support send/expect types of health check sequences.
5044
5045 TCP checks currently support 4 modes of operations :
5046 - no "tcp-check" directive : the health check only consists in a connection
5047 attempt, which remains the default mode.
5048
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005049 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01005050 used to send a string along with a connection opening. With some
5051 protocols, it helps sending a "QUIT" message for example that prevents
5052 the server from logging a connection error for each health check. The
5053 check result will still be based on the ability to open the connection
5054 only.
5055
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005056 - "tcp-check expect" only is mentioned : this is used to test a banner.
Willy Tarreaued179852013-12-16 01:07:00 +01005057 The connection is opened and haproxy waits for the server to present some
5058 contents which must validate some rules. The check result will be based
5059 on the matching between the contents and the rules. This is suited for
5060 POP, IMAP, SMTP, FTP, SSH, TELNET.
5061
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005062 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01005063 used to test a hello-type protocol. Haproxy sends a message, the server
5064 responds and its response is analysed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005065 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01005066 suited for protocols which require a binding or a request/response model.
5067 LDAP, MySQL, Redis and SSL are example of such protocols, though they
5068 already all have their dedicated checks with a deeper understanding of
5069 the respective protocols.
5070 In this mode, many questions may be sent and many answers may be
5071 analysed.
5072
5073 Examples :
5074 # perform a POP check (analyse only server's banner)
5075 option tcp-check
5076 tcp-check expect string +OK\ POP3\ ready
5077
5078 # perform an IMAP check (analyse only server's banner)
5079 option tcp-check
5080 tcp-check expect string *\ OK\ IMAP4\ ready
5081
5082 # look for the redis master server after ensuring it speaks well
5083 # redis protocol, then it exits properly.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005084 # (send a command then analyse the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01005085 option tcp-check
5086 tcp-check send PING\r\n
5087 tcp-check expect +PONG
5088 tcp-check send info\ replication\r\n
5089 tcp-check expect string role:master
5090 tcp-check send QUIT\r\n
5091 tcp-check expect string +OK
5092
5093 forge a HTTP request, then analyse the response
5094 (send many headers before analyzing)
5095 option tcp-check
5096 tcp-check send HEAD\ /\ HTTP/1.1\r\n
5097 tcp-check send Host:\ www.mydomain.com\r\n
5098 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
5099 tcp-check send \r\n
5100 tcp-check expect rstring HTTP/1\..\ (2..|3..)
5101
5102
5103 See also : "tcp-check expect", "tcp-check send"
5104
5105
Willy Tarreau9ea05a72009-06-14 12:07:01 +02005106option tcp-smart-accept
5107no option tcp-smart-accept
5108 Enable or disable the saving of one ACK packet during the accept sequence
5109 May be used in sections : defaults | frontend | listen | backend
5110 yes | yes | yes | no
5111 Arguments : none
5112
5113 When an HTTP connection request comes in, the system acknowledges it on
5114 behalf of HAProxy, then the client immediately sends its request, and the
5115 system acknowledges it too while it is notifying HAProxy about the new
5116 connection. HAProxy then reads the request and responds. This means that we
5117 have one TCP ACK sent by the system for nothing, because the request could
5118 very well be acknowledged by HAProxy when it sends its response.
5119
5120 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
5121 sending this useless ACK on platforms which support it (currently at least
5122 Linux). It must not cause any problem, because the system will send it anyway
5123 after 40 ms if the response takes more time than expected to come.
5124
5125 During complex network debugging sessions, it may be desirable to disable
5126 this optimization because delayed ACKs can make troubleshooting more complex
5127 when trying to identify where packets are delayed. It is then possible to
5128 fall back to normal behaviour by specifying "no option tcp-smart-accept".
5129
5130 It is also possible to force it for non-HTTP proxies by simply specifying
5131 "option tcp-smart-accept". For instance, it can make sense with some services
5132 such as SMTP where the server speaks first.
5133
5134 It is recommended to avoid forcing this option in a defaults section. In case
5135 of doubt, consider setting it back to automatic values by prepending the
5136 "default" keyword before it, or disabling it using the "no" keyword.
5137
Willy Tarreaud88edf22009-06-14 15:48:17 +02005138 See also : "option tcp-smart-connect"
5139
5140
5141option tcp-smart-connect
5142no option tcp-smart-connect
5143 Enable or disable the saving of one ACK packet during the connect sequence
5144 May be used in sections : defaults | frontend | listen | backend
5145 yes | no | yes | yes
5146 Arguments : none
5147
5148 On certain systems (at least Linux), HAProxy can ask the kernel not to
5149 immediately send an empty ACK upon a connection request, but to directly
5150 send the buffer request instead. This saves one packet on the network and
5151 thus boosts performance. It can also be useful for some servers, because they
5152 immediately get the request along with the incoming connection.
5153
5154 This feature is enabled when "option tcp-smart-connect" is set in a backend.
5155 It is not enabled by default because it makes network troubleshooting more
5156 complex.
5157
5158 It only makes sense to enable it with protocols where the client speaks first
5159 such as HTTP. In other situations, if there is no data to send in place of
5160 the ACK, a normal ACK is sent.
5161
5162 If this option has been enabled in a "defaults" section, it can be disabled
5163 in a specific instance by prepending the "no" keyword before it.
5164
5165 See also : "option tcp-smart-accept"
5166
Willy Tarreau9ea05a72009-06-14 12:07:01 +02005167
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005168option tcpka
5169 Enable or disable the sending of TCP keepalive packets on both sides
5170 May be used in sections : defaults | frontend | listen | backend
5171 yes | yes | yes | yes
5172 Arguments : none
5173
5174 When there is a firewall or any session-aware component between a client and
5175 a server, and when the protocol involves very long sessions with long idle
5176 periods (eg: remote desktops), there is a risk that one of the intermediate
5177 components decides to expire a session which has remained idle for too long.
5178
5179 Enabling socket-level TCP keep-alives makes the system regularly send packets
5180 to the other end of the connection, leaving it active. The delay between
5181 keep-alive probes is controlled by the system only and depends both on the
5182 operating system and its tuning parameters.
5183
5184 It is important to understand that keep-alive packets are neither emitted nor
5185 received at the application level. It is only the network stacks which sees
5186 them. For this reason, even if one side of the proxy already uses keep-alives
5187 to maintain its connection alive, those keep-alive packets will not be
5188 forwarded to the other side of the proxy.
5189
5190 Please note that this has nothing to do with HTTP keep-alive.
5191
5192 Using option "tcpka" enables the emission of TCP keep-alive probes on both
5193 the client and server sides of a connection. Note that this is meaningful
5194 only in "defaults" or "listen" sections. If this option is used in a
5195 frontend, only the client side will get keep-alives, and if this option is
5196 used in a backend, only the server side will get keep-alives. For this
5197 reason, it is strongly recommended to explicitly use "option clitcpka" and
5198 "option srvtcpka" when the configuration is split between frontends and
5199 backends.
5200
5201 See also : "option clitcpka", "option srvtcpka"
5202
Willy Tarreau844e3c52008-01-11 16:28:18 +01005203
5204option tcplog
5205 Enable advanced logging of TCP connections with session state and timers
5206 May be used in sections : defaults | frontend | listen | backend
5207 yes | yes | yes | yes
5208 Arguments : none
5209
5210 By default, the log output format is very poor, as it only contains the
5211 source and destination addresses, and the instance name. By specifying
5212 "option tcplog", each log line turns into a much richer format including, but
5213 not limited to, the connection timers, the session status, the connections
5214 numbers, the frontend, backend and server name, and of course the source
5215 address and ports. This option is useful for pure TCP proxies in order to
5216 find which of the client or server disconnects or times out. For normal HTTP
5217 proxies, it's better to use "option httplog" which is even more complete.
5218
5219 This option may be set either in the frontend or the backend.
5220
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005221 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01005222
5223
Willy Tarreau844e3c52008-01-11 16:28:18 +01005224option transparent
5225no option transparent
5226 Enable client-side transparent proxying
5227 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01005228 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01005229 Arguments : none
5230
5231 This option was introduced in order to provide layer 7 persistence to layer 3
5232 load balancers. The idea is to use the OS's ability to redirect an incoming
5233 connection for a remote address to a local process (here HAProxy), and let
5234 this process know what address was initially requested. When this option is
5235 used, sessions without cookies will be forwarded to the original destination
5236 IP address of the incoming request (which should match that of another
5237 equipment), while requests with cookies will still be forwarded to the
5238 appropriate server.
5239
5240 Note that contrary to a common belief, this option does NOT make HAProxy
5241 present the client's IP to the server when establishing the connection.
5242
Willy Tarreaua1146052011-03-01 09:51:54 +01005243 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005244 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01005245
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005246
Emeric Brun647caf12009-06-30 17:57:00 +02005247persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02005248persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02005249 Enable RDP cookie-based persistence
5250 May be used in sections : defaults | frontend | listen | backend
5251 yes | no | yes | yes
5252 Arguments :
5253 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02005254 default cookie name "msts" will be used. There currently is no
5255 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02005256
5257 This statement enables persistence based on an RDP cookie. The RDP cookie
5258 contains all information required to find the server in the list of known
5259 servers. So when this option is set in the backend, the request is analysed
5260 and if an RDP cookie is found, it is decoded. If it matches a known server
5261 which is still UP (or if "option persist" is set), then the connection is
5262 forwarded to this server.
5263
5264 Note that this only makes sense in a TCP backend, but for this to work, the
5265 frontend must have waited long enough to ensure that an RDP cookie is present
5266 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005267 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02005268 a single "listen" section.
5269
Willy Tarreau61e28f22010-05-16 22:31:05 +02005270 Also, it is important to understand that the terminal server will emit this
5271 RDP cookie only if it is configured for "token redirection mode", which means
5272 that the "IP address redirection" option is disabled.
5273
Emeric Brun647caf12009-06-30 17:57:00 +02005274 Example :
5275 listen tse-farm
5276 bind :3389
5277 # wait up to 5s for an RDP cookie in the request
5278 tcp-request inspect-delay 5s
5279 tcp-request content accept if RDP_COOKIE
5280 # apply RDP cookie persistence
5281 persist rdp-cookie
5282 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02005283 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02005284 balance rdp-cookie
5285 server srv1 1.1.1.1:3389
5286 server srv2 1.1.1.2:3389
5287
Simon Hormanab814e02011-06-24 14:50:20 +09005288 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
5289 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02005290
5291
Willy Tarreau3a7d2072009-03-05 23:48:25 +01005292rate-limit sessions <rate>
5293 Set a limit on the number of new sessions accepted per second on a frontend
5294 May be used in sections : defaults | frontend | listen | backend
5295 yes | yes | yes | no
5296 Arguments :
5297 <rate> The <rate> parameter is an integer designating the maximum number
5298 of new sessions per second to accept on the frontend.
5299
5300 When the frontend reaches the specified number of new sessions per second, it
5301 stops accepting new connections until the rate drops below the limit again.
5302 During this time, the pending sessions will be kept in the socket's backlog
5303 (in system buffers) and haproxy will not even be aware that sessions are
5304 pending. When applying very low limit on a highly loaded service, it may make
5305 sense to increase the socket's backlog using the "backlog" keyword.
5306
5307 This feature is particularly efficient at blocking connection-based attacks
5308 or service abuse on fragile servers. Since the session rate is measured every
5309 millisecond, it is extremely accurate. Also, the limit applies immediately,
5310 no delay is needed at all to detect the threshold.
5311
5312 Example : limit the connection rate on SMTP to 10 per second max
5313 listen smtp
5314 mode tcp
5315 bind :25
5316 rate-limit sessions 10
5317 server 127.0.0.1:1025
5318
Willy Tarreaua17c2d92011-07-25 08:16:20 +02005319 Note : when the maximum rate is reached, the frontend's status is not changed
5320 but its sockets appear as "WAITING" in the statistics if the
5321 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01005322
5323 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
5324
5325
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005326redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
5327redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
5328redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02005329 Return an HTTP redirection if/unless a condition is matched
5330 May be used in sections : defaults | frontend | listen | backend
5331 no | yes | yes | yes
5332
5333 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +01005334 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02005335
Willy Tarreau0140f252008-11-19 21:07:09 +01005336 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005337 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01005338 the HTTP "Location" header. When used in an "http-request" rule,
5339 <loc> value follows the log-format rules and can include some
5340 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005341
5342 <pfx> With "redirect prefix", the "Location" header is built from the
5343 concatenation of <pfx> and the complete URI path, including the
5344 query string, unless the "drop-query" option is specified (see
5345 below). As a special case, if <pfx> equals exactly "/", then
5346 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01005347 redirect to the same URL (for instance, to insert a cookie). When
5348 used in an "http-request" rule, <pfx> value follows the log-format
5349 rules and can include some dynamic values (see Custom Log Format
5350 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005351
5352 <sch> With "redirect scheme", then the "Location" header is built by
5353 concatenating <sch> with "://" then the first occurrence of the
5354 "Host" header, and then the URI path, including the query string
5355 unless the "drop-query" option is specified (see below). If no
5356 path is found or if the path is "*", then "/" is used instead. If
5357 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005358 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005359 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01005360 HTTPS. When used in an "http-request" rule, <sch> value follows
5361 the log-format rules and can include some dynamic values (see
5362 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +01005363
5364 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +01005365 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
5366 with 302 used by default if no code is specified. 301 means
5367 "Moved permanently", and a browser may cache the Location. 302
5368 means "Moved permanently" and means that the browser should not
5369 cache the redirection. 303 is equivalent to 302 except that the
5370 browser will fetch the location with a GET method. 307 is just
5371 like 302 but makes it clear that the same method must be reused.
5372 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +01005373
5374 <option> There are several options which can be specified to adjust the
5375 expected behaviour of a redirection :
5376
5377 - "drop-query"
5378 When this keyword is used in a prefix-based redirection, then the
5379 location will be set without any possible query-string, which is useful
5380 for directing users to a non-secure page for instance. It has no effect
5381 with a location-type redirect.
5382
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01005383 - "append-slash"
5384 This keyword may be used in conjunction with "drop-query" to redirect
5385 users who use a URL not ending with a '/' to the same one with the '/'.
5386 It can be useful to ensure that search engines will only see one URL.
5387 For this, a return code 301 is preferred.
5388
Willy Tarreau0140f252008-11-19 21:07:09 +01005389 - "set-cookie NAME[=value]"
5390 A "Set-Cookie" header will be added with NAME (and optionally "=value")
5391 to the response. This is sometimes used to indicate that a user has
5392 been seen, for instance to protect against some types of DoS. No other
5393 cookie option is added, so the cookie will be a session cookie. Note
5394 that for a browser, a sole cookie name without an equal sign is
5395 different from a cookie with an equal sign.
5396
5397 - "clear-cookie NAME[=]"
5398 A "Set-Cookie" header will be added with NAME (and optionally "="), but
5399 with the "Max-Age" attribute set to zero. This will tell the browser to
5400 delete this cookie. It is useful for instance on logout pages. It is
5401 important to note that clearing the cookie "NAME" will not remove a
5402 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
5403 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02005404
5405 Example: move the login URL only to HTTPS.
5406 acl clear dst_port 80
5407 acl secure dst_port 8080
5408 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +01005409 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +01005410 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +01005411 acl cookie_set hdr_sub(cookie) SEEN=1
5412
5413 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +01005414 redirect prefix https://mysite.com if login_page !secure
5415 redirect prefix http://mysite.com drop-query if login_page !uid_given
5416 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +01005417 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +02005418
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01005419 Example: send redirects for request for articles without a '/'.
5420 acl missing_slash path_reg ^/article/[^/]*$
5421 redirect code 301 prefix / drop-query append-slash if missing_slash
5422
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005423 Example: redirect all HTTP traffic to HTTPS when SSL is handled by haproxy.
David BERARDe7153042012-11-03 00:11:31 +01005424 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +02005425
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01005426 Example: append 'www.' prefix in front of all hosts not having it
5427 http-request redirect code 301 location www.%[hdr(host)]%[req.uri] \
5428 unless { hdr_beg(host) -i www }
5429
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005430 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02005431
5432
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01005433redisp (deprecated)
5434redispatch (deprecated)
5435 Enable or disable session redistribution in case of connection failure
5436 May be used in sections: defaults | frontend | listen | backend
5437 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005438 Arguments : none
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01005439
5440 In HTTP mode, if a server designated by a cookie is down, clients may
5441 definitely stick to it because they cannot flush the cookie, so they will not
5442 be able to access the service anymore.
5443
5444 Specifying "redispatch" will allow the proxy to break their persistence and
5445 redistribute them to a working server.
5446
5447 It also allows to retry last connection to another server in case of multiple
5448 connection failures. Of course, it requires having "retries" set to a nonzero
5449 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01005450
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01005451 This form is deprecated, do not use it in any new configuration, use the new
5452 "option redispatch" instead.
5453
5454 See also : "option redispatch"
5455
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005456
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01005457reqadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01005458 Add a header at the end of the HTTP request
5459 May be used in sections : defaults | frontend | listen | backend
5460 no | yes | yes | yes
5461 Arguments :
5462 <string> is the complete line to be added. Any space or known delimiter
5463 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005464 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01005465
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01005466 <cond> is an optional matching condition built from ACLs. It makes it
5467 possible to ignore this rule when other conditions are not met.
5468
Willy Tarreau303c0352008-01-17 19:01:39 +01005469 A new line consisting in <string> followed by a line feed will be added after
5470 the last header of an HTTP request.
5471
5472 Header transformations only apply to traffic which passes through HAProxy,
5473 and not to traffic generated by HAProxy, such as health-checks or error
5474 responses.
5475
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01005476 Example : add "X-Proto: SSL" to requests coming via port 81
5477 acl is-ssl dst_port 81
5478 reqadd X-Proto:\ SSL if is-ssl
5479
5480 See also: "rspadd", section 6 about HTTP header manipulation, and section 7
5481 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005482
5483
Willy Tarreau5321c422010-01-28 20:35:13 +01005484reqallow <search> [{if | unless} <cond>]
5485reqiallow <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005486 Definitely allow an HTTP request if a line matches a regular expression
5487 May be used in sections : defaults | frontend | listen | backend
5488 no | yes | yes | yes
5489 Arguments :
5490 <search> is the regular expression applied to HTTP headers and to the
5491 request line. This is an extended regular expression. Parenthesis
5492 grouping is supported and no preliminary backslash is required.
5493 Any space or known delimiter must be escaped using a backslash
5494 ('\'). The pattern applies to a full line at a time. The
5495 "reqallow" keyword strictly matches case while "reqiallow"
5496 ignores case.
5497
Willy Tarreau5321c422010-01-28 20:35:13 +01005498 <cond> is an optional matching condition built from ACLs. It makes it
5499 possible to ignore this rule when other conditions are not met.
5500
Willy Tarreau303c0352008-01-17 19:01:39 +01005501 A request containing any line which matches extended regular expression
5502 <search> will mark the request as allowed, even if any later test would
5503 result in a deny. The test applies both to the request line and to request
5504 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01005505 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01005506
5507 It is easier, faster and more powerful to use ACLs to write access policies.
5508 Reqdeny, reqallow and reqpass should be avoided in new designs.
5509
5510 Example :
5511 # allow www.* but refuse *.local
5512 reqiallow ^Host:\ www\.
5513 reqideny ^Host:\ .*\.local
5514
Willy Tarreau5321c422010-01-28 20:35:13 +01005515 See also: "reqdeny", "block", section 6 about HTTP header manipulation, and
5516 section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005517
5518
Willy Tarreau5321c422010-01-28 20:35:13 +01005519reqdel <search> [{if | unless} <cond>]
5520reqidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005521 Delete all headers matching a regular expression in an HTTP request
5522 May be used in sections : defaults | frontend | listen | backend
5523 no | yes | yes | yes
5524 Arguments :
5525 <search> is the regular expression applied to HTTP headers and to the
5526 request line. This is an extended regular expression. Parenthesis
5527 grouping is supported and no preliminary backslash is required.
5528 Any space or known delimiter must be escaped using a backslash
5529 ('\'). The pattern applies to a full line at a time. The "reqdel"
5530 keyword strictly matches case while "reqidel" ignores case.
5531
Willy Tarreau5321c422010-01-28 20:35:13 +01005532 <cond> is an optional matching condition built from ACLs. It makes it
5533 possible to ignore this rule when other conditions are not met.
5534
Willy Tarreau303c0352008-01-17 19:01:39 +01005535 Any header line matching extended regular expression <search> in the request
5536 will be completely deleted. Most common use of this is to remove unwanted
5537 and/or dangerous headers or cookies from a request before passing it to the
5538 next servers.
5539
5540 Header transformations only apply to traffic which passes through HAProxy,
5541 and not to traffic generated by HAProxy, such as health-checks or error
5542 responses. Keep in mind that header names are not case-sensitive.
5543
5544 Example :
5545 # remove X-Forwarded-For header and SERVER cookie
5546 reqidel ^X-Forwarded-For:.*
5547 reqidel ^Cookie:.*SERVER=
5548
Willy Tarreau5321c422010-01-28 20:35:13 +01005549 See also: "reqadd", "reqrep", "rspdel", section 6 about HTTP header
5550 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005551
5552
Willy Tarreau5321c422010-01-28 20:35:13 +01005553reqdeny <search> [{if | unless} <cond>]
5554reqideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005555 Deny an HTTP request if a line matches a regular expression
5556 May be used in sections : defaults | frontend | listen | backend
5557 no | yes | yes | yes
5558 Arguments :
5559 <search> is the regular expression applied to HTTP headers and to the
5560 request line. This is an extended regular expression. Parenthesis
5561 grouping is supported and no preliminary backslash is required.
5562 Any space or known delimiter must be escaped using a backslash
5563 ('\'). The pattern applies to a full line at a time. The
5564 "reqdeny" keyword strictly matches case while "reqideny" ignores
5565 case.
5566
Willy Tarreau5321c422010-01-28 20:35:13 +01005567 <cond> is an optional matching condition built from ACLs. It makes it
5568 possible to ignore this rule when other conditions are not met.
5569
Willy Tarreau303c0352008-01-17 19:01:39 +01005570 A request containing any line which matches extended regular expression
5571 <search> will mark the request as denied, even if any later test would
5572 result in an allow. The test applies both to the request line and to request
5573 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01005574 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01005575
Willy Tarreauced27012008-01-17 20:35:34 +01005576 A denied request will generate an "HTTP 403 forbidden" response once the
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01005577 complete request has been parsed. This is consistent with what is practiced
Willy Tarreaud72758d2010-01-12 10:42:19 +01005578 using ACLs.
Willy Tarreauced27012008-01-17 20:35:34 +01005579
Willy Tarreau303c0352008-01-17 19:01:39 +01005580 It is easier, faster and more powerful to use ACLs to write access policies.
5581 Reqdeny, reqallow and reqpass should be avoided in new designs.
5582
5583 Example :
5584 # refuse *.local, then allow www.*
5585 reqideny ^Host:\ .*\.local
5586 reqiallow ^Host:\ www\.
5587
Willy Tarreau5321c422010-01-28 20:35:13 +01005588 See also: "reqallow", "rspdeny", "block", section 6 about HTTP header
5589 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005590
5591
Willy Tarreau5321c422010-01-28 20:35:13 +01005592reqpass <search> [{if | unless} <cond>]
5593reqipass <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005594 Ignore any HTTP request line matching a regular expression in next rules
5595 May be used in sections : defaults | frontend | listen | backend
5596 no | yes | yes | yes
5597 Arguments :
5598 <search> is the regular expression applied to HTTP headers and to the
5599 request line. This is an extended regular expression. Parenthesis
5600 grouping is supported and no preliminary backslash is required.
5601 Any space or known delimiter must be escaped using a backslash
5602 ('\'). The pattern applies to a full line at a time. The
5603 "reqpass" keyword strictly matches case while "reqipass" ignores
5604 case.
5605
Willy Tarreau5321c422010-01-28 20:35:13 +01005606 <cond> is an optional matching condition built from ACLs. It makes it
5607 possible to ignore this rule when other conditions are not met.
5608
Willy Tarreau303c0352008-01-17 19:01:39 +01005609 A request containing any line which matches extended regular expression
5610 <search> will skip next rules, without assigning any deny or allow verdict.
5611 The test applies both to the request line and to request headers. Keep in
5612 mind that URLs in request line are case-sensitive while header names are not.
5613
5614 It is easier, faster and more powerful to use ACLs to write access policies.
5615 Reqdeny, reqallow and reqpass should be avoided in new designs.
5616
5617 Example :
5618 # refuse *.local, then allow www.*, but ignore "www.private.local"
5619 reqipass ^Host:\ www.private\.local
5620 reqideny ^Host:\ .*\.local
5621 reqiallow ^Host:\ www\.
5622
Willy Tarreau5321c422010-01-28 20:35:13 +01005623 See also: "reqallow", "reqdeny", "block", section 6 about HTTP header
5624 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005625
5626
Willy Tarreau5321c422010-01-28 20:35:13 +01005627reqrep <search> <string> [{if | unless} <cond>]
5628reqirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005629 Replace a regular expression with a string in an HTTP request line
5630 May be used in sections : defaults | frontend | listen | backend
5631 no | yes | yes | yes
5632 Arguments :
5633 <search> is the regular expression applied to HTTP headers and to the
5634 request line. This is an extended regular expression. Parenthesis
5635 grouping is supported and no preliminary backslash is required.
5636 Any space or known delimiter must be escaped using a backslash
5637 ('\'). The pattern applies to a full line at a time. The "reqrep"
5638 keyword strictly matches case while "reqirep" ignores case.
5639
5640 <string> is the complete line to be added. Any space or known delimiter
5641 must be escaped using a backslash ('\'). References to matched
5642 pattern groups are possible using the common \N form, with N
5643 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005644 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01005645
Willy Tarreau5321c422010-01-28 20:35:13 +01005646 <cond> is an optional matching condition built from ACLs. It makes it
5647 possible to ignore this rule when other conditions are not met.
5648
Willy Tarreau303c0352008-01-17 19:01:39 +01005649 Any line matching extended regular expression <search> in the request (both
5650 the request line and header lines) will be completely replaced with <string>.
5651 Most common use of this is to rewrite URLs or domain names in "Host" headers.
5652
5653 Header transformations only apply to traffic which passes through HAProxy,
5654 and not to traffic generated by HAProxy, such as health-checks or error
5655 responses. Note that for increased readability, it is suggested to add enough
5656 spaces between the request and the response. Keep in mind that URLs in
5657 request line are case-sensitive while header names are not.
5658
5659 Example :
5660 # replace "/static/" with "/" at the beginning of any request path.
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04005661 reqrep ^([^\ :]*)\ /static/(.*) \1\ /\2
Willy Tarreau303c0352008-01-17 19:01:39 +01005662 # replace "www.mydomain.com" with "www" in the host name.
5663 reqirep ^Host:\ www.mydomain.com Host:\ www
5664
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04005665 See also: "reqadd", "reqdel", "rsprep", "tune.bufsize", section 6 about
5666 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005667
5668
Willy Tarreau5321c422010-01-28 20:35:13 +01005669reqtarpit <search> [{if | unless} <cond>]
5670reqitarpit <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005671 Tarpit an HTTP request containing a line matching a regular expression
5672 May be used in sections : defaults | frontend | listen | backend
5673 no | yes | yes | yes
5674 Arguments :
5675 <search> is the regular expression applied to HTTP headers and to the
5676 request line. This is an extended regular expression. Parenthesis
5677 grouping is supported and no preliminary backslash is required.
5678 Any space or known delimiter must be escaped using a backslash
5679 ('\'). The pattern applies to a full line at a time. The
5680 "reqtarpit" keyword strictly matches case while "reqitarpit"
5681 ignores case.
5682
Willy Tarreau5321c422010-01-28 20:35:13 +01005683 <cond> is an optional matching condition built from ACLs. It makes it
5684 possible to ignore this rule when other conditions are not met.
5685
Willy Tarreau303c0352008-01-17 19:01:39 +01005686 A request containing any line which matches extended regular expression
5687 <search> will be tarpitted, which means that it will connect to nowhere, will
Willy Tarreauced27012008-01-17 20:35:34 +01005688 be kept open for a pre-defined time, then will return an HTTP error 500 so
5689 that the attacker does not suspect it has been tarpitted. The status 500 will
5690 be reported in the logs, but the completion flags will indicate "PT". The
Willy Tarreau303c0352008-01-17 19:01:39 +01005691 delay is defined by "timeout tarpit", or "timeout connect" if the former is
5692 not set.
5693
5694 The goal of the tarpit is to slow down robots attacking servers with
5695 identifiable requests. Many robots limit their outgoing number of connections
5696 and stay connected waiting for a reply which can take several minutes to
5697 come. Depending on the environment and attack, it may be particularly
5698 efficient at reducing the load on the network and firewalls.
5699
Willy Tarreau5321c422010-01-28 20:35:13 +01005700 Examples :
Willy Tarreau303c0352008-01-17 19:01:39 +01005701 # ignore user-agents reporting any flavour of "Mozilla" or "MSIE", but
5702 # block all others.
5703 reqipass ^User-Agent:\.*(Mozilla|MSIE)
5704 reqitarpit ^User-Agent:
5705
Willy Tarreau5321c422010-01-28 20:35:13 +01005706 # block bad guys
5707 acl badguys src 10.1.0.3 172.16.13.20/28
5708 reqitarpit . if badguys
5709
5710 See also: "reqallow", "reqdeny", "reqpass", section 6 about HTTP header
5711 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005712
5713
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02005714retries <value>
5715 Set the number of retries to perform on a server after a connection failure
5716 May be used in sections: defaults | frontend | listen | backend
5717 yes | no | yes | yes
5718 Arguments :
5719 <value> is the number of times a connection attempt should be retried on
5720 a server when a connection either is refused or times out. The
5721 default value is 3.
5722
5723 It is important to understand that this value applies to the number of
5724 connection attempts, not full requests. When a connection has effectively
5725 been established to a server, there will be no more retry.
5726
5727 In order to avoid immediate reconnections to a server which is restarting,
5728 a turn-around timer of 1 second is applied before a retry occurs.
5729
5730 When "option redispatch" is set, the last retry may be performed on another
5731 server even if a cookie references a different server.
5732
5733 See also : "option redispatch"
5734
5735
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005736rspadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01005737 Add a header at the end of the HTTP response
5738 May be used in sections : defaults | frontend | listen | backend
5739 no | yes | yes | yes
5740 Arguments :
5741 <string> is the complete line to be added. Any space or known delimiter
5742 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005743 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01005744
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005745 <cond> is an optional matching condition built from ACLs. It makes it
5746 possible to ignore this rule when other conditions are not met.
5747
Willy Tarreau303c0352008-01-17 19:01:39 +01005748 A new line consisting in <string> followed by a line feed will be added after
5749 the last header of an HTTP response.
5750
5751 Header transformations only apply to traffic which passes through HAProxy,
5752 and not to traffic generated by HAProxy, such as health-checks or error
5753 responses.
5754
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005755 See also: "reqadd", section 6 about HTTP header manipulation, and section 7
5756 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005757
5758
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005759rspdel <search> [{if | unless} <cond>]
5760rspidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005761 Delete all headers matching a regular expression in an HTTP response
5762 May be used in sections : defaults | frontend | listen | backend
5763 no | yes | yes | yes
5764 Arguments :
5765 <search> is the regular expression applied to HTTP headers and to the
5766 response line. This is an extended regular expression, so
5767 parenthesis grouping is supported and no preliminary backslash
5768 is required. Any space or known delimiter must be escaped using
5769 a backslash ('\'). The pattern applies to a full line at a time.
5770 The "rspdel" keyword strictly matches case while "rspidel"
5771 ignores case.
5772
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005773 <cond> is an optional matching condition built from ACLs. It makes it
5774 possible to ignore this rule when other conditions are not met.
5775
Willy Tarreau303c0352008-01-17 19:01:39 +01005776 Any header line matching extended regular expression <search> in the response
5777 will be completely deleted. Most common use of this is to remove unwanted
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02005778 and/or sensitive headers or cookies from a response before passing it to the
Willy Tarreau303c0352008-01-17 19:01:39 +01005779 client.
5780
5781 Header transformations only apply to traffic which passes through HAProxy,
5782 and not to traffic generated by HAProxy, such as health-checks or error
5783 responses. Keep in mind that header names are not case-sensitive.
5784
5785 Example :
5786 # remove the Server header from responses
Willy Tarreau5e80e022013-05-25 08:31:25 +02005787 rspidel ^Server:.*
Willy Tarreau303c0352008-01-17 19:01:39 +01005788
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005789 See also: "rspadd", "rsprep", "reqdel", section 6 about HTTP header
5790 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005791
5792
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005793rspdeny <search> [{if | unless} <cond>]
5794rspideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005795 Block an HTTP response if a line matches a regular expression
5796 May be used in sections : defaults | frontend | listen | backend
5797 no | yes | yes | yes
5798 Arguments :
5799 <search> is the regular expression applied to HTTP headers and to the
5800 response line. This is an extended regular expression, so
5801 parenthesis grouping is supported and no preliminary backslash
5802 is required. Any space or known delimiter must be escaped using
5803 a backslash ('\'). The pattern applies to a full line at a time.
5804 The "rspdeny" keyword strictly matches case while "rspideny"
5805 ignores case.
5806
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005807 <cond> is an optional matching condition built from ACLs. It makes it
5808 possible to ignore this rule when other conditions are not met.
5809
Willy Tarreau303c0352008-01-17 19:01:39 +01005810 A response containing any line which matches extended regular expression
5811 <search> will mark the request as denied. The test applies both to the
5812 response line and to response headers. Keep in mind that header names are not
5813 case-sensitive.
5814
5815 Main use of this keyword is to prevent sensitive information leak and to
Willy Tarreauced27012008-01-17 20:35:34 +01005816 block the response before it reaches the client. If a response is denied, it
5817 will be replaced with an HTTP 502 error so that the client never retrieves
5818 any sensitive data.
Willy Tarreau303c0352008-01-17 19:01:39 +01005819
5820 It is easier, faster and more powerful to use ACLs to write access policies.
5821 Rspdeny should be avoided in new designs.
5822
5823 Example :
5824 # Ensure that no content type matching ms-word will leak
5825 rspideny ^Content-type:\.*/ms-word
5826
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005827 See also: "reqdeny", "acl", "block", section 6 about HTTP header manipulation
5828 and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005829
5830
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005831rsprep <search> <string> [{if | unless} <cond>]
5832rspirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01005833 Replace a regular expression with a string in an HTTP response line
5834 May be used in sections : defaults | frontend | listen | backend
5835 no | yes | yes | yes
5836 Arguments :
5837 <search> is the regular expression applied to HTTP headers and to the
5838 response line. This is an extended regular expression, so
5839 parenthesis grouping is supported and no preliminary backslash
5840 is required. Any space or known delimiter must be escaped using
5841 a backslash ('\'). The pattern applies to a full line at a time.
5842 The "rsprep" keyword strictly matches case while "rspirep"
5843 ignores case.
5844
5845 <string> is the complete line to be added. Any space or known delimiter
5846 must be escaped using a backslash ('\'). References to matched
5847 pattern groups are possible using the common \N form, with N
5848 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005849 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01005850
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005851 <cond> is an optional matching condition built from ACLs. It makes it
5852 possible to ignore this rule when other conditions are not met.
5853
Willy Tarreau303c0352008-01-17 19:01:39 +01005854 Any line matching extended regular expression <search> in the response (both
5855 the response line and header lines) will be completely replaced with
5856 <string>. Most common use of this is to rewrite Location headers.
5857
5858 Header transformations only apply to traffic which passes through HAProxy,
5859 and not to traffic generated by HAProxy, such as health-checks or error
5860 responses. Note that for increased readability, it is suggested to add enough
5861 spaces between the request and the response. Keep in mind that header names
5862 are not case-sensitive.
5863
5864 Example :
5865 # replace "Location: 127.0.0.1:8080" with "Location: www.mydomain.com"
5866 rspirep ^Location:\ 127.0.0.1:8080 Location:\ www.mydomain.com
5867
Willy Tarreaufdb563c2010-01-31 15:43:27 +01005868 See also: "rspadd", "rspdel", "reqrep", section 6 about HTTP header
5869 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01005870
5871
David du Colombier486df472011-03-17 10:40:26 +01005872server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005873 Declare a server in a backend
5874 May be used in sections : defaults | frontend | listen | backend
5875 no | no | yes | yes
5876 Arguments :
5877 <name> is the internal name assigned to this server. This name will
Cyril Bonté941a0c62012-10-15 19:44:24 +02005878 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -05005879 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005880
David du Colombier486df472011-03-17 10:40:26 +01005881 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
5882 resolvable hostname is supported, but this name will be resolved
5883 during start-up. Address "0.0.0.0" or "*" has a special meaning.
5884 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +02005885 address as the one from the client connection. This is useful in
5886 transparent proxy architectures where the client's connection is
5887 intercepted and haproxy must forward to the original destination
5888 address. This is more or less what the "transparent" keyword does
5889 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +01005890 to report statistics. Optionally, an address family prefix may be
5891 used before the address to force the family regardless of the
5892 address format, which can be useful to specify a path to a unix
5893 socket with no slash ('/'). Currently supported prefixes are :
5894 - 'ipv4@' -> address is always IPv4
5895 - 'ipv6@' -> address is always IPv6
5896 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02005897 - 'abns@' -> address is in abstract namespace (Linux only)
Willy Tarreaudad36a32013-03-11 01:20:04 +01005898 Any part of the address string may reference any number of
5899 environment variables by preceding their name with a dollar
5900 sign ('$') and optionally enclosing them with braces ('{}'),
5901 similarly to what is done in Bourne shell.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005902
Willy Tarreaub6205fd2012-09-24 12:27:33 +02005903 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005904 be sent to this port. If unset, the same port the client
5905 connected to will be used. The port may also be prefixed by a "+"
5906 or a "-". In this case, the server's port will be determined by
5907 adding this value to the client's port.
5908
5909 <param*> is a list of parameters for this server. The "server" keywords
5910 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005911 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005912
5913 Examples :
5914 server first 10.1.1.1:1080 cookie first check inter 1000
5915 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +01005916 server transp ipv4@
Willy Tarreaudad36a32013-03-11 01:20:04 +01005917 server backup ${SRV_BACKUP}:1080 backup
5918 server www1_dc1 ${LAN_DC1}.101:80
5919 server www1_dc2 ${LAN_DC2}.101:80
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005920
Mark Lamourinec2247f02012-01-04 13:02:01 -05005921 See also: "default-server", "http-send-name-header" and section 5 about
5922 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005923
5924
5925source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02005926source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +01005927source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005928 Set the source address for outgoing connections
5929 May be used in sections : defaults | frontend | listen | backend
5930 yes | no | yes | yes
5931 Arguments :
5932 <addr> is the IPv4 address HAProxy will bind to before connecting to a
5933 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +01005934
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005935 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +01005936 the most appropriate address to reach its destination. Optionally
5937 an address family prefix may be used before the address to force
5938 the family regardless of the address format, which can be useful
5939 to specify a path to a unix socket with no slash ('/'). Currently
5940 supported prefixes are :
5941 - 'ipv4@' -> address is always IPv4
5942 - 'ipv6@' -> address is always IPv6
5943 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02005944 - 'abns@' -> address is in abstract namespace (Linux only)
Willy Tarreaudad36a32013-03-11 01:20:04 +01005945 Any part of the address string may reference any number of
5946 environment variables by preceding their name with a dollar
5947 sign ('$') and optionally enclosing them with braces ('{}'),
5948 similarly to what is done in Bourne shell.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005949
5950 <port> is an optional port. It is normally not needed but may be useful
5951 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02005952 the system will select a free port. Note that port ranges are not
5953 supported in the backend. If you want to force port ranges, you
5954 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005955
5956 <addr2> is the IP address to present to the server when connections are
5957 forwarded in full transparent proxy mode. This is currently only
5958 supported on some patched Linux kernels. When this address is
5959 specified, clients connecting to the server will be presented
5960 with this address, while health checks will still use the address
5961 <addr>.
5962
5963 <port2> is the optional port to present to the server when connections
5964 are forwarded in full transparent proxy mode (see <addr2> above).
5965 The default value of zero means the system will select a free
5966 port.
5967
Willy Tarreaubce70882009-09-07 11:51:47 +02005968 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
5969 This is the name of a comma-separated header list which can
5970 contain multiple IP addresses. By default, the last occurrence is
5971 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +01005972 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +02005973 by previous proxy, typically Stunnel. In order to use another
5974 occurrence from the last one, please see the <occ> parameter
5975 below. When the header (or occurrence) is not found, no binding
5976 is performed so that the proxy's default IP address is used. Also
5977 keep in mind that the header name is case insensitive, as for any
5978 HTTP header.
5979
5980 <occ> is the occurrence number of a value to be used in a multi-value
5981 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005982 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +02005983 address. Positive values indicate a position from the first
5984 occurrence, 1 being the first one. Negative values indicate
5985 positions relative to the last one, -1 being the last one. This
5986 is helpful for situations where an X-Forwarded-For header is set
5987 at the entry point of an infrastructure and must be used several
5988 proxy layers away. When this value is not specified, -1 is
5989 assumed. Passing a zero here disables the feature.
5990
Willy Tarreaud53f96b2009-02-04 18:46:54 +01005991 <name> is an optional interface name to which to bind to for outgoing
5992 traffic. On systems supporting this features (currently, only
5993 Linux), this allows one to bind all traffic to the server to
5994 this interface even if it is not the one the system would select
5995 based on routing tables. This should be used with extreme care.
5996 Note that using this option requires root privileges.
5997
Willy Tarreaueabeafa2008-01-16 16:17:06 +01005998 The "source" keyword is useful in complex environments where a specific
5999 address only is allowed to connect to the servers. It may be needed when a
6000 private address must be used through a public gateway for instance, and it is
6001 known that the system cannot determine the adequate source address by itself.
6002
6003 An extension which is available on certain patched Linux kernels may be used
6004 through the "usesrc" optional keyword. It makes it possible to connect to the
6005 servers with an IP address which does not belong to the system itself. This
6006 is called "full transparent proxy mode". For this to work, the destination
6007 servers have to route their traffic back to this address through the machine
6008 running HAProxy, and IP forwarding must generally be enabled on this machine.
6009
6010 In this "full transparent proxy" mode, it is possible to force a specific IP
6011 address to be presented to the servers. This is not much used in fact. A more
6012 common use is to tell HAProxy to present the client's IP address. For this,
6013 there are two methods :
6014
6015 - present the client's IP and port addresses. This is the most transparent
6016 mode, but it can cause problems when IP connection tracking is enabled on
6017 the machine, because a same connection may be seen twice with different
6018 states. However, this solution presents the huge advantage of not
6019 limiting the system to the 64k outgoing address+port couples, because all
6020 of the client ranges may be used.
6021
6022 - present only the client's IP address and select a spare port. This
6023 solution is still quite elegant but slightly less transparent (downstream
6024 firewalls logs will not match upstream's). It also presents the downside
6025 of limiting the number of concurrent connections to the usual 64k ports.
6026 However, since the upstream and downstream ports are different, local IP
6027 connection tracking on the machine will not be upset by the reuse of the
6028 same session.
6029
6030 Note that depending on the transparent proxy technology used, it may be
6031 required to force the source address. In fact, cttproxy version 2 requires an
6032 IP address in <addr> above, and does not support setting of "0.0.0.0" as the
6033 IP address because it creates NAT entries which much match the exact outgoing
6034 address. Tproxy version 4 and some other kernel patches which work in pure
6035 forwarding mode generally will not have this limitation.
6036
6037 This option sets the default source for all servers in the backend. It may
6038 also be specified in a "defaults" section. Finer source address specification
6039 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006040 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006041
6042 Examples :
6043 backend private
6044 # Connect to the servers using our 192.168.1.200 source address
6045 source 192.168.1.200
6046
6047 backend transparent_ssl1
6048 # Connect to the SSL farm from the client's source address
6049 source 192.168.1.200 usesrc clientip
6050
6051 backend transparent_ssl2
6052 # Connect to the SSL farm from the client's source address and port
6053 # not recommended if IP conntrack is present on the local machine.
6054 source 192.168.1.200 usesrc client
6055
6056 backend transparent_ssl3
6057 # Connect to the SSL farm from the client's source address. It
6058 # is more conntrack-friendly.
6059 source 192.168.1.200 usesrc clientip
6060
6061 backend transparent_smtp
6062 # Connect to the SMTP farm from the client's source address/port
6063 # with Tproxy version 4.
6064 source 0.0.0.0 usesrc clientip
6065
Willy Tarreaubce70882009-09-07 11:51:47 +02006066 backend transparent_http
6067 # Connect to the servers using the client's IP as seen by previous
6068 # proxy.
6069 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
6070
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006071 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006072 the Linux kernel on www.balabit.com, the "bind" keyword.
6073
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006074
Willy Tarreau844e3c52008-01-11 16:28:18 +01006075srvtimeout <timeout> (deprecated)
6076 Set the maximum inactivity time on the server side.
6077 May be used in sections : defaults | frontend | listen | backend
6078 yes | no | yes | yes
6079 Arguments :
6080 <timeout> is the timeout value specified in milliseconds by default, but
6081 can be in any other unit if the number is suffixed by the unit,
6082 as explained at the top of this document.
6083
6084 The inactivity timeout applies when the server is expected to acknowledge or
6085 send data. In HTTP mode, this timeout is particularly important to consider
6086 during the first phase of the server's response, when it has to send the
6087 headers, as it directly represents the server's processing time for the
6088 request. To find out what value to put there, it's often good to start with
6089 what would be considered as unacceptable response times, then check the logs
6090 to observe the response time distribution, and adjust the value accordingly.
6091
6092 The value is specified in milliseconds by default, but can be in any other
6093 unit if the number is suffixed by the unit, as specified at the top of this
6094 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
6095 recommended that the client timeout remains equal to the server timeout in
6096 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01006097 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +01006098 packet losses by specifying timeouts that are slightly above multiples of 3
Willy Tarreaud72758d2010-01-12 10:42:19 +01006099 seconds (eg: 4 or 5 seconds minimum).
Willy Tarreau844e3c52008-01-11 16:28:18 +01006100
6101 This parameter is specific to backends, but can be specified once for all in
6102 "defaults" sections. This is in fact one of the easiest solutions not to
6103 forget about it. An unspecified timeout results in an infinite timeout, which
6104 is not recommended. Such a usage is accepted and works but reports a warning
6105 during startup because it may results in accumulation of expired sessions in
6106 the system if the system's timeouts are not configured either.
6107
6108 This parameter is provided for compatibility but is currently deprecated.
6109 Please use "timeout server" instead.
6110
Willy Tarreauce887fd2012-05-12 12:50:00 +02006111 See also : "timeout server", "timeout tunnel", "timeout client" and
6112 "clitimeout".
Willy Tarreau844e3c52008-01-11 16:28:18 +01006113
6114
Cyril Bonté66c327d2010-10-12 00:14:37 +02006115stats admin { if | unless } <cond>
6116 Enable statistics admin level if/unless a condition is matched
6117 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006118 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +02006119
6120 This statement enables the statistics admin level if/unless a condition is
6121 matched.
6122
6123 The admin level allows to enable/disable servers from the web interface. By
6124 default, statistics page is read-only for security reasons.
6125
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006126 Note : Consider not using this feature in multi-process mode (nbproc > 1)
6127 unless you know what you do : memory is not shared between the
6128 processes, which can result in random behaviours.
6129
Cyril Bonté23b39d92011-02-10 22:54:44 +01006130 Currently, the POST request is limited to the buffer size minus the reserved
6131 buffer space, which means that if the list of servers is too long, the
6132 request won't be processed. It is recommended to alter few servers at a
6133 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +02006134
6135 Example :
6136 # statistics admin level only for localhost
6137 backend stats_localhost
6138 stats enable
6139 stats admin if LOCALHOST
6140
6141 Example :
6142 # statistics admin level always enabled because of the authentication
6143 backend stats_auth
6144 stats enable
6145 stats auth admin:AdMiN123
6146 stats admin if TRUE
6147
6148 Example :
6149 # statistics admin level depends on the authenticated user
6150 userlist stats-auth
6151 group admin users admin
6152 user admin insecure-password AdMiN123
6153 group readonly users haproxy
6154 user haproxy insecure-password haproxy
6155
6156 backend stats_auth
6157 stats enable
6158 acl AUTH http_auth(stats-auth)
6159 acl AUTH_ADMIN http_auth_group(stats-auth) admin
6160 stats http-request auth unless AUTH
6161 stats admin if AUTH_ADMIN
6162
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006163 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
6164 "bind-process", section 3.4 about userlists and section 7 about
6165 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +02006166
6167
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006168stats auth <user>:<passwd>
6169 Enable statistics with authentication and grant access to an account
6170 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006171 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006172 Arguments :
6173 <user> is a user name to grant access to
6174
6175 <passwd> is the cleartext password associated to this user
6176
6177 This statement enables statistics with default settings, and restricts access
6178 to declared users only. It may be repeated as many times as necessary to
6179 allow as many users as desired. When a user tries to access the statistics
6180 without a valid account, a "401 Forbidden" response will be returned so that
6181 the browser asks the user to provide a valid user and password. The real
6182 which will be returned to the browser is configurable using "stats realm".
6183
6184 Since the authentication method is HTTP Basic Authentication, the passwords
6185 circulate in cleartext on the network. Thus, it was decided that the
6186 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02006187 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006188
6189 It is also possible to reduce the scope of the proxies which appear in the
6190 report using "stats scope".
6191
6192 Though this statement alone is enough to enable statistics reporting, it is
6193 recommended to set all other settings in order to avoid relying on default
6194 unobvious parameters.
6195
6196 Example :
6197 # public access (limited to this backend only)
6198 backend public_www
6199 server srv1 192.168.0.1:80
6200 stats enable
6201 stats hide-version
6202 stats scope .
6203 stats uri /admin?stats
6204 stats realm Haproxy\ Statistics
6205 stats auth admin1:AdMiN123
6206 stats auth admin2:AdMiN321
6207
6208 # internal monitoring access (unlimited)
6209 backend private_monitoring
6210 stats enable
6211 stats uri /admin?stats
6212 stats refresh 5s
6213
6214 See also : "stats enable", "stats realm", "stats scope", "stats uri"
6215
6216
6217stats enable
6218 Enable statistics reporting with default settings
6219 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006220 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006221 Arguments : none
6222
6223 This statement enables statistics reporting with default settings defined
6224 at build time. Unless stated otherwise, these settings are used :
6225 - stats uri : /haproxy?stats
6226 - stats realm : "HAProxy Statistics"
6227 - stats auth : no authentication
6228 - stats scope : no restriction
6229
6230 Though this statement alone is enough to enable statistics reporting, it is
6231 recommended to set all other settings in order to avoid relying on default
6232 unobvious parameters.
6233
6234 Example :
6235 # public access (limited to this backend only)
6236 backend public_www
6237 server srv1 192.168.0.1:80
6238 stats enable
6239 stats hide-version
6240 stats scope .
6241 stats uri /admin?stats
6242 stats realm Haproxy\ Statistics
6243 stats auth admin1:AdMiN123
6244 stats auth admin2:AdMiN321
6245
6246 # internal monitoring access (unlimited)
6247 backend private_monitoring
6248 stats enable
6249 stats uri /admin?stats
6250 stats refresh 5s
6251
6252 See also : "stats auth", "stats realm", "stats uri"
6253
6254
Willy Tarreaud63335a2010-02-26 12:56:52 +01006255stats hide-version
6256 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006257 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006258 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01006259 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006260
Willy Tarreaud63335a2010-02-26 12:56:52 +01006261 By default, the stats page reports some useful status information along with
6262 the statistics. Among them is HAProxy's version. However, it is generally
6263 considered dangerous to report precise version to anyone, as it can help them
6264 target known weaknesses with specific attacks. The "stats hide-version"
6265 statement removes the version from the statistics report. This is recommended
6266 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006267
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02006268 Though this statement alone is enough to enable statistics reporting, it is
6269 recommended to set all other settings in order to avoid relying on default
6270 unobvious parameters.
6271
Willy Tarreaud63335a2010-02-26 12:56:52 +01006272 Example :
6273 # public access (limited to this backend only)
6274 backend public_www
6275 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02006276 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01006277 stats hide-version
6278 stats scope .
6279 stats uri /admin?stats
6280 stats realm Haproxy\ Statistics
6281 stats auth admin1:AdMiN123
6282 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006283
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006284 # internal monitoring access (unlimited)
6285 backend private_monitoring
6286 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01006287 stats uri /admin?stats
6288 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +01006289
Willy Tarreaud63335a2010-02-26 12:56:52 +01006290 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02006291
Willy Tarreau983e01e2010-01-11 18:42:06 +01006292
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02006293stats http-request { allow | deny | auth [realm <realm>] }
6294 [ { if | unless } <condition> ]
6295 Access control for statistics
6296
6297 May be used in sections: defaults | frontend | listen | backend
6298 no | no | yes | yes
6299
6300 As "http-request", these set of options allow to fine control access to
6301 statistics. Each option may be followed by if/unless and acl.
6302 First option with matched condition (or option without condition) is final.
6303 For "deny" a 403 error will be returned, for "allow" normal processing is
6304 performed, for "auth" a 401/407 error code is returned so the client
6305 should be asked to enter a username and password.
6306
6307 There is no fixed limit to the number of http-request statements per
6308 instance.
6309
6310 See also : "http-request", section 3.4 about userlists and section 7
6311 about ACL usage.
6312
6313
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006314stats realm <realm>
6315 Enable statistics and set authentication realm
6316 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006317 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006318 Arguments :
6319 <realm> is the name of the HTTP Basic Authentication realm reported to
6320 the browser. The browser uses it to display it in the pop-up
6321 inviting the user to enter a valid username and password.
6322
6323 The realm is read as a single word, so any spaces in it should be escaped
6324 using a backslash ('\').
6325
6326 This statement is useful only in conjunction with "stats auth" since it is
6327 only related to authentication.
6328
6329 Though this statement alone is enough to enable statistics reporting, it is
6330 recommended to set all other settings in order to avoid relying on default
6331 unobvious parameters.
6332
6333 Example :
6334 # public access (limited to this backend only)
6335 backend public_www
6336 server srv1 192.168.0.1:80
6337 stats enable
6338 stats hide-version
6339 stats scope .
6340 stats uri /admin?stats
6341 stats realm Haproxy\ Statistics
6342 stats auth admin1:AdMiN123
6343 stats auth admin2:AdMiN321
6344
6345 # internal monitoring access (unlimited)
6346 backend private_monitoring
6347 stats enable
6348 stats uri /admin?stats
6349 stats refresh 5s
6350
6351 See also : "stats auth", "stats enable", "stats uri"
6352
6353
6354stats refresh <delay>
6355 Enable statistics with automatic refresh
6356 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006357 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006358 Arguments :
6359 <delay> is the suggested refresh delay, specified in seconds, which will
6360 be returned to the browser consulting the report page. While the
6361 browser is free to apply any delay, it will generally respect it
6362 and refresh the page this every seconds. The refresh interval may
6363 be specified in any other non-default time unit, by suffixing the
6364 unit after the value, as explained at the top of this document.
6365
6366 This statement is useful on monitoring displays with a permanent page
6367 reporting the load balancer's activity. When set, the HTML report page will
6368 include a link "refresh"/"stop refresh" so that the user can select whether
6369 he wants automatic refresh of the page or not.
6370
6371 Though this statement alone is enough to enable statistics reporting, it is
6372 recommended to set all other settings in order to avoid relying on default
6373 unobvious parameters.
6374
6375 Example :
6376 # public access (limited to this backend only)
6377 backend public_www
6378 server srv1 192.168.0.1:80
6379 stats enable
6380 stats hide-version
6381 stats scope .
6382 stats uri /admin?stats
6383 stats realm Haproxy\ Statistics
6384 stats auth admin1:AdMiN123
6385 stats auth admin2:AdMiN321
6386
6387 # internal monitoring access (unlimited)
6388 backend private_monitoring
6389 stats enable
6390 stats uri /admin?stats
6391 stats refresh 5s
6392
6393 See also : "stats auth", "stats enable", "stats realm", "stats uri"
6394
6395
6396stats scope { <name> | "." }
6397 Enable statistics and limit access scope
6398 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006399 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006400 Arguments :
6401 <name> is the name of a listen, frontend or backend section to be
6402 reported. The special name "." (a single dot) designates the
6403 section in which the statement appears.
6404
6405 When this statement is specified, only the sections enumerated with this
6406 statement will appear in the report. All other ones will be hidden. This
6407 statement may appear as many times as needed if multiple sections need to be
6408 reported. Please note that the name checking is performed as simple string
6409 comparisons, and that it is never checked that a give section name really
6410 exists.
6411
6412 Though this statement alone is enough to enable statistics reporting, it is
6413 recommended to set all other settings in order to avoid relying on default
6414 unobvious parameters.
6415
6416 Example :
6417 # public access (limited to this backend only)
6418 backend public_www
6419 server srv1 192.168.0.1:80
6420 stats enable
6421 stats hide-version
6422 stats scope .
6423 stats uri /admin?stats
6424 stats realm Haproxy\ Statistics
6425 stats auth admin1:AdMiN123
6426 stats auth admin2:AdMiN321
6427
6428 # internal monitoring access (unlimited)
6429 backend private_monitoring
6430 stats enable
6431 stats uri /admin?stats
6432 stats refresh 5s
6433
6434 See also : "stats auth", "stats enable", "stats realm", "stats uri"
6435
Willy Tarreaud63335a2010-02-26 12:56:52 +01006436
Willy Tarreauc9705a12010-07-27 20:05:50 +02006437stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +01006438 Enable reporting of a description on the statistics page.
6439 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006440 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01006441
Willy Tarreauc9705a12010-07-27 20:05:50 +02006442 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +01006443 description from global section is automatically used instead.
6444
6445 This statement is useful for users that offer shared services to their
6446 customers, where node or description should be different for each customer.
6447
6448 Though this statement alone is enough to enable statistics reporting, it is
6449 recommended to set all other settings in order to avoid relying on default
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04006450 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +01006451
6452 Example :
6453 # internal monitoring access (unlimited)
6454 backend private_monitoring
6455 stats enable
6456 stats show-desc Master node for Europe, Asia, Africa
6457 stats uri /admin?stats
6458 stats refresh 5s
6459
6460 See also: "show-node", "stats enable", "stats uri" and "description" in
6461 global section.
6462
6463
6464stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +02006465 Enable reporting additional information on the statistics page
6466 May be used in sections : defaults | frontend | listen | backend
6467 yes | yes | yes | yes
6468 Arguments : none
6469
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006470 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +01006471 - cap: capabilities (proxy)
6472 - mode: one of tcp, http or health (proxy)
6473 - id: SNMP ID (proxy, socket, server)
6474 - IP (socket, server)
6475 - cookie (backend, server)
6476
6477 Though this statement alone is enough to enable statistics reporting, it is
6478 recommended to set all other settings in order to avoid relying on default
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04006479 unobvious parameters. Default behaviour is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +01006480
6481 See also: "stats enable", "stats uri".
6482
6483
6484stats show-node [ <name> ]
6485 Enable reporting of a host name on the statistics page.
6486 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006487 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01006488 Arguments:
6489 <name> is an optional name to be reported. If unspecified, the
6490 node name from global section is automatically used instead.
6491
6492 This statement is useful for users that offer shared services to their
6493 customers, where node or description might be different on a stats page
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04006494 provided for each customer. Default behaviour is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +01006495
6496 Though this statement alone is enough to enable statistics reporting, it is
6497 recommended to set all other settings in order to avoid relying on default
6498 unobvious parameters.
6499
6500 Example:
6501 # internal monitoring access (unlimited)
6502 backend private_monitoring
6503 stats enable
6504 stats show-node Europe-1
6505 stats uri /admin?stats
6506 stats refresh 5s
6507
6508 See also: "show-desc", "stats enable", "stats uri", and "node" in global
6509 section.
6510
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006511
6512stats uri <prefix>
6513 Enable statistics and define the URI prefix to access them
6514 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02006515 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006516 Arguments :
6517 <prefix> is the prefix of any URI which will be redirected to stats. This
6518 prefix may contain a question mark ('?') to indicate part of a
6519 query string.
6520
6521 The statistics URI is intercepted on the relayed traffic, so it appears as a
6522 page within the normal application. It is strongly advised to ensure that the
6523 selected URI will never appear in the application, otherwise it will never be
6524 possible to reach it in the application.
6525
6526 The default URI compiled in haproxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006527 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006528 It is generally a good idea to include a question mark in the URI so that
6529 intermediate proxies refrain from caching the results. Also, since any string
6530 beginning with the prefix will be accepted as a stats request, the question
6531 mark helps ensuring that no valid URI will begin with the same words.
6532
6533 It is sometimes very convenient to use "/" as the URI prefix, and put that
6534 statement in a "listen" instance of its own. That makes it easy to dedicate
6535 an address or a port to statistics only.
6536
6537 Though this statement alone is enough to enable statistics reporting, it is
6538 recommended to set all other settings in order to avoid relying on default
6539 unobvious parameters.
6540
6541 Example :
6542 # public access (limited to this backend only)
6543 backend public_www
6544 server srv1 192.168.0.1:80
6545 stats enable
6546 stats hide-version
6547 stats scope .
6548 stats uri /admin?stats
6549 stats realm Haproxy\ Statistics
6550 stats auth admin1:AdMiN123
6551 stats auth admin2:AdMiN321
6552
6553 # internal monitoring access (unlimited)
6554 backend private_monitoring
6555 stats enable
6556 stats uri /admin?stats
6557 stats refresh 5s
6558
6559 See also : "stats auth", "stats enable", "stats realm"
6560
6561
Willy Tarreaud63335a2010-02-26 12:56:52 +01006562stick match <pattern> [table <table>] [{if | unless} <cond>]
6563 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006564 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +01006565 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006566
6567 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02006568 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006569 describes what elements of the incoming request or connection
6570 will be analysed in the hope to find a matching entry in a
6571 stickiness table. This rule is mandatory.
6572
6573 <table> is an optional stickiness table name. If unspecified, the same
6574 backend's table is used. A stickiness table is declared using
6575 the "stick-table" statement.
6576
6577 <cond> is an optional matching condition. It makes it possible to match
6578 on a certain criterion only when other conditions are met (or
6579 not met). For instance, it could be used to match on a source IP
6580 address except when a request passes through a known proxy, in
6581 which case we'd match on a header containing that IP address.
6582
6583 Some protocols or applications require complex stickiness rules and cannot
6584 always simply rely on cookies nor hashing. The "stick match" statement
6585 describes a rule to extract the stickiness criterion from an incoming request
6586 or connection. See section 7 for a complete list of possible patterns and
6587 transformation rules.
6588
6589 The table has to be declared using the "stick-table" statement. It must be of
6590 a type compatible with the pattern. By default it is the one which is present
6591 in the same backend. It is possible to share a table with other backends by
6592 referencing it using the "table" keyword. If another table is referenced,
6593 the server's ID inside the backends are used. By default, all server IDs
6594 start at 1 in each backend, so the server ordering is enough. But in case of
6595 doubt, it is highly recommended to force server IDs using their "id" setting.
6596
6597 It is possible to restrict the conditions where a "stick match" statement
6598 will apply, using "if" or "unless" followed by a condition. See section 7 for
6599 ACL based conditions.
6600
6601 There is no limit on the number of "stick match" statements. The first that
6602 applies and matches will cause the request to be directed to the same server
6603 as was used for the request which created the entry. That way, multiple
6604 matches can be used as fallbacks.
6605
6606 The stick rules are checked after the persistence cookies, so they will not
6607 affect stickiness if a cookie has already been used to select a server. That
6608 way, it becomes very easy to insert cookies and match on IP addresses in
6609 order to maintain stickiness between HTTP and HTTPS.
6610
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006611 Note : Consider not using this feature in multi-process mode (nbproc > 1)
6612 unless you know what you do : memory is not shared between the
6613 processes, which can result in random behaviours.
6614
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006615 Example :
6616 # forward SMTP users to the same server they just used for POP in the
6617 # last 30 minutes
6618 backend pop
6619 mode tcp
6620 balance roundrobin
6621 stick store-request src
6622 stick-table type ip size 200k expire 30m
6623 server s1 192.168.1.1:110
6624 server s2 192.168.1.1:110
6625
6626 backend smtp
6627 mode tcp
6628 balance roundrobin
6629 stick match src table pop
6630 server s1 192.168.1.1:25
6631 server s2 192.168.1.1:25
6632
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006633 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02006634 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006635
6636
6637stick on <pattern> [table <table>] [{if | unless} <condition>]
6638 Define a request pattern to associate a user to a server
6639 May be used in sections : defaults | frontend | listen | backend
6640 no | no | yes | yes
6641
6642 Note : This form is exactly equivalent to "stick match" followed by
6643 "stick store-request", all with the same arguments. Please refer
6644 to both keywords for details. It is only provided as a convenience
6645 for writing more maintainable configurations.
6646
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006647 Note : Consider not using this feature in multi-process mode (nbproc > 1)
6648 unless you know what you do : memory is not shared between the
6649 processes, which can result in random behaviours.
6650
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006651 Examples :
6652 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +01006653 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006654
6655 # ...is strictly equivalent to this one :
6656 stick match src table pop if !localhost
6657 stick store-request src table pop if !localhost
6658
6659
6660 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
6661 # well as HTTP without cookie. Share the same table between both accesses.
6662 backend http
6663 mode http
6664 balance roundrobin
6665 stick on src table https
6666 cookie SRV insert indirect nocache
6667 server s1 192.168.1.1:80 cookie s1
6668 server s2 192.168.1.1:80 cookie s2
6669
6670 backend https
6671 mode tcp
6672 balance roundrobin
6673 stick-table type ip size 200k expire 30m
6674 stick on src
6675 server s1 192.168.1.1:443
6676 server s2 192.168.1.1:443
6677
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006678 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006679
6680
6681stick store-request <pattern> [table <table>] [{if | unless} <condition>]
6682 Define a request pattern used to create an entry in a stickiness table
6683 May be used in sections : defaults | frontend | listen | backend
6684 no | no | yes | yes
6685
6686 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02006687 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006688 describes what elements of the incoming request or connection
6689 will be analysed, extracted and stored in the table once a
6690 server is selected.
6691
6692 <table> is an optional stickiness table name. If unspecified, the same
6693 backend's table is used. A stickiness table is declared using
6694 the "stick-table" statement.
6695
6696 <cond> is an optional storage condition. It makes it possible to store
6697 certain criteria only when some conditions are met (or not met).
6698 For instance, it could be used to store the source IP address
6699 except when the request passes through a known proxy, in which
6700 case we'd store a converted form of a header containing that IP
6701 address.
6702
6703 Some protocols or applications require complex stickiness rules and cannot
6704 always simply rely on cookies nor hashing. The "stick store-request" statement
6705 describes a rule to decide what to extract from the request and when to do
6706 it, in order to store it into a stickiness table for further requests to
6707 match it using the "stick match" statement. Obviously the extracted part must
6708 make sense and have a chance to be matched in a further request. Storing a
6709 client's IP address for instance often makes sense. Storing an ID found in a
6710 URL parameter also makes sense. Storing a source port will almost never make
6711 any sense because it will be randomly matched. See section 7 for a complete
6712 list of possible patterns and transformation rules.
6713
6714 The table has to be declared using the "stick-table" statement. It must be of
6715 a type compatible with the pattern. By default it is the one which is present
6716 in the same backend. It is possible to share a table with other backends by
6717 referencing it using the "table" keyword. If another table is referenced,
6718 the server's ID inside the backends are used. By default, all server IDs
6719 start at 1 in each backend, so the server ordering is enough. But in case of
6720 doubt, it is highly recommended to force server IDs using their "id" setting.
6721
6722 It is possible to restrict the conditions where a "stick store-request"
6723 statement will apply, using "if" or "unless" followed by a condition. This
6724 condition will be evaluated while parsing the request, so any criteria can be
6725 used. See section 7 for ACL based conditions.
6726
6727 There is no limit on the number of "stick store-request" statements, but
6728 there is a limit of 8 simultaneous stores per request or response. This
6729 makes it possible to store up to 8 criteria, all extracted from either the
6730 request or the response, regardless of the number of rules. Only the 8 first
6731 ones which match will be kept. Using this, it is possible to feed multiple
6732 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01006733 another protocol or access method. Using multiple store-request rules with
6734 the same table is possible and may be used to find the best criterion to rely
6735 on, by arranging the rules by decreasing preference order. Only the first
6736 extracted criterion for a given table will be stored. All subsequent store-
6737 request rules referencing the same table will be skipped and their ACLs will
6738 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006739
6740 The "store-request" rules are evaluated once the server connection has been
6741 established, so that the table will contain the real server that processed
6742 the request.
6743
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006744 Note : Consider not using this feature in multi-process mode (nbproc > 1)
6745 unless you know what you do : memory is not shared between the
6746 processes, which can result in random behaviours.
6747
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006748 Example :
6749 # forward SMTP users to the same server they just used for POP in the
6750 # last 30 minutes
6751 backend pop
6752 mode tcp
6753 balance roundrobin
6754 stick store-request src
6755 stick-table type ip size 200k expire 30m
6756 server s1 192.168.1.1:110
6757 server s2 192.168.1.1:110
6758
6759 backend smtp
6760 mode tcp
6761 balance roundrobin
6762 stick match src table pop
6763 server s1 192.168.1.1:25
6764 server s2 192.168.1.1:25
6765
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006766 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02006767 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006768
6769
Emeric Brun7c6b82e2010-09-24 16:34:28 +02006770stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Emeric Brunf099e792010-09-27 12:05:28 +02006771 size <size> [expire <expire>] [nopurge] [peers <peersect>]
6772 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +08006773 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006774 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +02006775 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006776
6777 Arguments :
6778 ip a table declared with "type ip" will only store IPv4 addresses.
6779 This form is very compact (about 50 bytes per entry) and allows
6780 very fast entry lookup and stores with almost no overhead. This
6781 is mainly used to store client source IP addresses.
6782
David du Colombier9a6d3c92011-03-17 10:40:24 +01006783 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
6784 This form is very compact (about 60 bytes per entry) and allows
6785 very fast entry lookup and stores with almost no overhead. This
6786 is mainly used to store client source IP addresses.
6787
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006788 integer a table declared with "type integer" will store 32bit integers
6789 which can represent a client identifier found in a request for
6790 instance.
6791
6792 string a table declared with "type string" will store substrings of up
6793 to <len> characters. If the string provided by the pattern
6794 extractor is larger than <len>, it will be truncated before
6795 being stored. During matching, at most <len> characters will be
6796 compared between the string in the table and the extracted
6797 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +02006798 to 32 characters.
6799
6800 binary a table declared with "type binary" will store binary blocks
6801 of <len> bytes. If the block provided by the pattern
6802 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +02006803 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +02006804 is shorter than <len>, it will be padded by 0. When not
6805 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006806
6807 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +02006808 "string" type table (See type "string" above). Or the number
6809 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006810 changing this parameter as memory usage will proportionally
6811 increase.
6812
6813 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +01006814 value directly impacts memory usage. Count approximately
6815 50 bytes per entry, plus the size of a string if any. The size
6816 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006817
6818 [nopurge] indicates that we refuse to purge older entries when the table
6819 is full. When not specified and the table is full when haproxy
6820 wants to store an entry in it, it will flush a few of the oldest
6821 entries in order to release some space for the new ones. This is
6822 most often the desired behaviour. In some specific cases, it
6823 be desirable to refuse new entries instead of purging the older
6824 ones. That may be the case when the amount of data to store is
6825 far above the hardware limits and we prefer not to offer access
6826 to new clients than to reject the ones already connected. When
6827 using this parameter, be sure to properly set the "expire"
6828 parameter (see below).
6829
Emeric Brunf099e792010-09-27 12:05:28 +02006830 <peersect> is the name of the peers section to use for replication. Entries
6831 which associate keys to server IDs are kept synchronized with
6832 the remote peers declared in this section. All entries are also
6833 automatically learned from the local peer (old process) during a
6834 soft restart.
6835
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01006836 NOTE : peers can't be used in multi-process mode.
6837
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006838 <expire> defines the maximum duration of an entry in the table since it
6839 was last created, refreshed or matched. The expiration delay is
6840 defined using the standard time format, similarly as the various
6841 timeouts. The maximum duration is slightly above 24 days. See
6842 section 2.2 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +02006843 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006844 be removed once full. Be sure not to use the "nopurge" parameter
6845 if not expiration delay is specified.
6846
Willy Tarreau08d5f982010-06-06 13:34:54 +02006847 <data_type> is used to store additional information in the stick-table. This
6848 may be used by ACLs in order to control various criteria related
6849 to the activity of the client matching the stick-table. For each
6850 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +02006851 that the additional data can fit. Several data types may be
6852 stored with an entry. Multiple data types may be specified after
6853 the "store" keyword, as a comma-separated list. Alternatively,
6854 it is possible to repeat the "store" keyword followed by one or
6855 several data types. Except for the "server_id" type which is
6856 automatically detected and enabled, all data types must be
6857 explicitly declared to be stored. If an ACL references a data
6858 type which is not stored, the ACL will simply not match. Some
6859 data types require an argument which must be passed just after
6860 the type between parenthesis. See below for the supported data
6861 types and their arguments.
6862
6863 The data types that can be stored with an entry are the following :
6864 - server_id : this is an integer which holds the numeric ID of the server a
6865 request was assigned to. It is used by the "stick match", "stick store",
6866 and "stick on" rules. It is automatically enabled when referenced.
6867
6868 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
6869 integer which may be used for anything. Most of the time it will be used
6870 to put a special tag on some entries, for instance to note that a
6871 specific behaviour was detected and must be known for future matches.
6872
Willy Tarreauba2ffd12013-05-29 15:54:14 +02006873 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
6874 over a period. It is a positive 32-bit integer integer which may be used
6875 for anything. Just like <gpc0>, it counts events, but instead of keeping
6876 a cumulative count, it maintains the rate at which the counter is
6877 incremented. Most of the time it will be used to measure the frequency of
6878 occurrence of certain events (eg: requests to a specific URL).
6879
Willy Tarreauc9705a12010-07-27 20:05:50 +02006880 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
6881 the absolute number of connections received from clients which matched
6882 this entry. It does not mean the connections were accepted, just that
6883 they were received.
6884
6885 - conn_cur : Current Connections. It is a positive 32-bit integer which
6886 stores the concurrent connection counts for the entry. It is incremented
6887 once an incoming connection matches the entry, and decremented once the
6888 connection leaves. That way it is possible to know at any time the exact
6889 number of concurrent connections for an entry.
6890
6891 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
6892 integer parameter <period> which indicates in milliseconds the length
6893 of the period over which the average is measured. It reports the average
6894 incoming connection rate over that period, in connections per period. The
6895 result is an integer which can be matched using ACLs.
6896
6897 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
6898 the absolute number of sessions received from clients which matched this
6899 entry. A session is a connection that was accepted by the layer 4 rules.
6900
6901 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
6902 integer parameter <period> which indicates in milliseconds the length
6903 of the period over which the average is measured. It reports the average
6904 incoming session rate over that period, in sessions per period. The
6905 result is an integer which can be matched using ACLs.
6906
6907 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
6908 counts the absolute number of HTTP requests received from clients which
6909 matched this entry. It does not matter whether they are valid requests or
6910 not. Note that this is different from sessions when keep-alive is used on
6911 the client side.
6912
6913 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
6914 integer parameter <period> which indicates in milliseconds the length
6915 of the period over which the average is measured. It reports the average
6916 HTTP request rate over that period, in requests per period. The result is
6917 an integer which can be matched using ACLs. It does not matter whether
6918 they are valid requests or not. Note that this is different from sessions
6919 when keep-alive is used on the client side.
6920
6921 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
6922 counts the absolute number of HTTP requests errors induced by clients
6923 which matched this entry. Errors are counted on invalid and truncated
6924 requests, as well as on denied or tarpitted requests, and on failed
6925 authentications. If the server responds with 4xx, then the request is
6926 also counted as an error since it's an error triggered by the client
6927 (eg: vulnerability scan).
6928
6929 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
6930 integer parameter <period> which indicates in milliseconds the length
6931 of the period over which the average is measured. It reports the average
6932 HTTP request error rate over that period, in requests per period (see
6933 http_err_cnt above for what is accounted as an error). The result is an
6934 integer which can be matched using ACLs.
6935
6936 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
6937 integer which counts the cumulated amount of bytes received from clients
6938 which matched this entry. Headers are included in the count. This may be
6939 used to limit abuse of upload features on photo or video servers.
6940
6941 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
6942 integer parameter <period> which indicates in milliseconds the length
6943 of the period over which the average is measured. It reports the average
6944 incoming bytes rate over that period, in bytes per period. It may be used
6945 to detect users which upload too much and too fast. Warning: with large
6946 uploads, it is possible that the amount of uploaded data will be counted
6947 once upon termination, thus causing spikes in the average transfer speed
6948 instead of having a smooth one. This may partially be smoothed with
6949 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
6950 recommended for better fairness.
6951
6952 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
6953 integer which counts the cumulated amount of bytes sent to clients which
6954 matched this entry. Headers are included in the count. This may be used
6955 to limit abuse of bots sucking the whole site.
6956
6957 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
6958 an integer parameter <period> which indicates in milliseconds the length
6959 of the period over which the average is measured. It reports the average
6960 outgoing bytes rate over that period, in bytes per period. It may be used
6961 to detect users which download too much and too fast. Warning: with large
6962 transfers, it is possible that the amount of transferred data will be
6963 counted once upon termination, thus causing spikes in the average
6964 transfer speed instead of having a smooth one. This may partially be
6965 smoothed with "option contstats" though this is not perfect yet. Use of
6966 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +02006967
Willy Tarreauc00cdc22010-06-06 16:48:26 +02006968 There is only one stick-table per proxy. At the moment of writing this doc,
6969 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006970 to be required, simply create a dummy backend with a stick-table in it and
6971 reference it.
6972
6973 It is important to understand that stickiness based on learning information
6974 has some limitations, including the fact that all learned associations are
6975 lost upon restart. In general it can be good as a complement but not always
6976 as an exclusive stickiness.
6977
Willy Tarreauc9705a12010-07-27 20:05:50 +02006978 Last, memory requirements may be important when storing many data types.
6979 Indeed, storing all indicators above at once in each entry requires 116 bytes
6980 per entry, or 116 MB for a 1-million entries table. This is definitely not
6981 something that can be ignored.
6982
6983 Example:
6984 # Keep track of counters of up to 1 million IP addresses over 5 minutes
6985 # and store a general purpose counter and the average connection rate
6986 # computed over a sliding window of 30 seconds.
6987 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
6988
6989 See also : "stick match", "stick on", "stick store-request", section 2.2
David du Colombiera13d1b92011-03-17 10:40:22 +01006990 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01006991
6992
Emeric Brun6a1cefa2010-09-24 18:15:17 +02006993stick store-response <pattern> [table <table>] [{if | unless} <condition>]
6994 Define a request pattern used to create an entry in a stickiness table
6995 May be used in sections : defaults | frontend | listen | backend
6996 no | no | yes | yes
6997
6998 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02006999 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007000 describes what elements of the response or connection will
7001 be analysed, extracted and stored in the table once a
7002 server is selected.
7003
7004 <table> is an optional stickiness table name. If unspecified, the same
7005 backend's table is used. A stickiness table is declared using
7006 the "stick-table" statement.
7007
7008 <cond> is an optional storage condition. It makes it possible to store
7009 certain criteria only when some conditions are met (or not met).
7010 For instance, it could be used to store the SSL session ID only
7011 when the response is a SSL server hello.
7012
7013 Some protocols or applications require complex stickiness rules and cannot
7014 always simply rely on cookies nor hashing. The "stick store-response"
7015 statement describes a rule to decide what to extract from the response and
7016 when to do it, in order to store it into a stickiness table for further
7017 requests to match it using the "stick match" statement. Obviously the
7018 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007019 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007020 See section 7 for a complete list of possible patterns and transformation
7021 rules.
7022
7023 The table has to be declared using the "stick-table" statement. It must be of
7024 a type compatible with the pattern. By default it is the one which is present
7025 in the same backend. It is possible to share a table with other backends by
7026 referencing it using the "table" keyword. If another table is referenced,
7027 the server's ID inside the backends are used. By default, all server IDs
7028 start at 1 in each backend, so the server ordering is enough. But in case of
7029 doubt, it is highly recommended to force server IDs using their "id" setting.
7030
7031 It is possible to restrict the conditions where a "stick store-response"
7032 statement will apply, using "if" or "unless" followed by a condition. This
7033 condition will be evaluated while parsing the response, so any criteria can
7034 be used. See section 7 for ACL based conditions.
7035
7036 There is no limit on the number of "stick store-response" statements, but
7037 there is a limit of 8 simultaneous stores per request or response. This
7038 makes it possible to store up to 8 criteria, all extracted from either the
7039 request or the response, regardless of the number of rules. Only the 8 first
7040 ones which match will be kept. Using this, it is possible to feed multiple
7041 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01007042 another protocol or access method. Using multiple store-response rules with
7043 the same table is possible and may be used to find the best criterion to rely
7044 on, by arranging the rules by decreasing preference order. Only the first
7045 extracted criterion for a given table will be stored. All subsequent store-
7046 response rules referencing the same table will be skipped and their ACLs will
7047 not be evaluated. However, even if a store-request rule references a table, a
7048 store-response rule may also use the same table. This means that each table
7049 may learn exactly one element from the request and one element from the
7050 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007051
7052 The table will contain the real server that processed the request.
7053
7054 Example :
7055 # Learn SSL session ID from both request and response and create affinity.
7056 backend https
7057 mode tcp
7058 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +02007059 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007060 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007061
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007062 acl clienthello req_ssl_hello_type 1
7063 acl serverhello rep_ssl_hello_type 2
7064
7065 # use tcp content accepts to detects ssl client and server hello.
7066 tcp-request inspect-delay 5s
7067 tcp-request content accept if clienthello
7068
7069 # no timeout on response inspect delay by default.
7070 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007071
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007072 # SSL session ID (SSLID) may be present on a client or server hello.
7073 # Its length is coded on 1 byte at offset 43 and its value starts
7074 # at offset 44.
7075
7076 # Match and learn on request if client hello.
7077 stick on payload_lv(43,1) if clienthello
7078
7079 # Learn on response if server hello.
7080 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +02007081
Emeric Brun6a1cefa2010-09-24 18:15:17 +02007082 server s1 192.168.1.1:443
7083 server s2 192.168.1.1:443
7084
7085 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
7086 extraction.
7087
7088
Willy Tarreau938c7fe2014-04-25 14:21:39 +02007089tcp-check connect [params*]
7090 Opens a new connection
7091 May be used in sections: defaults | frontend | listen | backend
7092 no | no | yes | yes
7093
7094 When an application lies on more than a single TCP port or when HAProxy
7095 load-balance many services in a single backend, it makes sense to probe all
7096 the services individually before considering a server as operational.
7097
7098 When there are no TCP port configured on the server line neither server port
7099 directive, then the 'tcp-check connect port <port>' must be the first step
7100 of the sequence.
7101
7102 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
7103 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
7104 do.
7105
7106 Parameters :
7107 They are optional and can be used to describe how HAProxy should open and
7108 use the TCP connection.
7109
7110 port if not set, check port or server port is used.
7111 It tells HAProxy where to open the connection to.
7112 <port> must be a valid TCP port source integer, from 1 to 65535.
7113
7114 send-proxy send a PROXY protocol string
7115
7116 ssl opens a ciphered connection
7117
7118 Examples:
7119 # check HTTP and HTTPs services on a server.
7120 # first open port 80 thanks to server line port directive, then
7121 # tcp-check opens port 443, ciphered and run a request on it:
7122 option tcp-check
7123 tcp-check connect
7124 tcp-check send GET\ /\ HTTP/1.0\r\n
7125 tcp-check send Host:\ haproxy.1wt.eu\r\n
7126 tcp-check send \r\n
7127 tcp-check expect rstring (2..|3..)
7128 tcp-check connect port 443 ssl
7129 tcp-check send GET\ /\ HTTP/1.0\r\n
7130 tcp-check send Host:\ haproxy.1wt.eu\r\n
7131 tcp-check send \r\n
7132 tcp-check expect rstring (2..|3..)
7133 server www 10.0.0.1 check port 80
7134
7135 # check both POP and IMAP from a single server:
7136 option tcp-check
7137 tcp-check connect port 110
7138 tcp-check expect string +OK\ POP3\ ready
7139 tcp-check connect port 143
7140 tcp-check expect string *\ OK\ IMAP4\ ready
7141 server mail 10.0.0.1 check
7142
7143 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
7144
7145
7146tcp-check expect [!] <match> <pattern>
7147 Specify data to be collected and analysed during a generic health check
7148 May be used in sections: defaults | frontend | listen | backend
7149 no | no | yes | yes
7150
7151 Arguments :
7152 <match> is a keyword indicating how to look for a specific pattern in the
7153 response. The keyword may be one of "string", "rstring" or
7154 binary.
7155 The keyword may be preceded by an exclamation mark ("!") to negate
7156 the match. Spaces are allowed between the exclamation mark and the
7157 keyword. See below for more details on the supported keywords.
7158
7159 <pattern> is the pattern to look for. It may be a string or a regular
7160 expression. If the pattern contains spaces, they must be escaped
7161 with the usual backslash ('\').
7162 If the match is set to binary, then the pattern must be passed as
7163 a serie of hexadecimal digits in an even number. Each sequence of
7164 two digits will represent a byte. The hexadecimal digits may be
7165 used upper or lower case.
7166
7167
7168 The available matches are intentionally similar to their http-check cousins :
7169
7170 string <string> : test the exact string matches in the response buffer.
7171 A health check response will be considered valid if the
7172 response's buffer contains this exact string. If the
7173 "string" keyword is prefixed with "!", then the response
7174 will be considered invalid if the body contains this
7175 string. This can be used to look for a mandatory pattern
7176 in a protocol response, or to detect a failure when a
7177 specific error appears in a protocol banner.
7178
7179 rstring <regex> : test a regular expression on the response buffer.
7180 A health check response will be considered valid if the
7181 response's buffer matches this expression. If the
7182 "rstring" keyword is prefixed with "!", then the response
7183 will be considered invalid if the body matches the
7184 expression.
7185
7186 binary <hexstring> : test the exact string in its hexadecimal form matches
7187 in the response buffer. A health check response will
7188 be considered valid if the response's buffer contains
7189 this exact hexadecimal string.
7190 Purpose is to match data on binary protocols.
7191
7192 It is important to note that the responses will be limited to a certain size
7193 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
7194 Thus, too large responses may not contain the mandatory pattern when using
7195 "string", "rstring" or binary. If a large response is absolutely required, it
7196 is possible to change the default max size by setting the global variable.
7197 However, it is worth keeping in mind that parsing very large responses can
7198 waste some CPU cycles, especially when regular expressions are used, and that
7199 it is always better to focus the checks on smaller resources. Also, in its
7200 current state, the check will not find any string nor regex past a null
7201 character in the response. Similarly it is not possible to request matching
7202 the null character.
7203
7204 Examples :
7205 # perform a POP check
7206 option tcp-check
7207 tcp-check expect string +OK\ POP3\ ready
7208
7209 # perform an IMAP check
7210 option tcp-check
7211 tcp-check expect string *\ OK\ IMAP4\ ready
7212
7213 # look for the redis master server
7214 option tcp-check
7215 tcp-check send PING\r\n
7216 tcp-check expect +PONG
7217 tcp-check send info\ replication\r\n
7218 tcp-check expect string role:master
7219 tcp-check send QUIT\r\n
7220 tcp-check expect string +OK
7221
7222
7223 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
7224 "tcp-check send-binary", "http-check expect", tune.chksize
7225
7226
7227tcp-check send <data>
7228 Specify a string to be sent as a question during a generic health check
7229 May be used in sections: defaults | frontend | listen | backend
7230 no | no | yes | yes
7231
7232 <data> : the data to be sent as a question during a generic health check
7233 session. For now, <data> must be a string.
7234
7235 Examples :
7236 # look for the redis master server
7237 option tcp-check
7238 tcp-check send info\ replication\r\n
7239 tcp-check expect string role:master
7240
7241 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
7242 "tcp-check send-binary", tune.chksize
7243
7244
7245tcp-check send-binary <hexastring>
7246 Specify an hexa digits string to be sent as a binary question during a raw
7247 tcp health check
7248 May be used in sections: defaults | frontend | listen | backend
7249 no | no | yes | yes
7250
7251 <data> : the data to be sent as a question during a generic health check
7252 session. For now, <data> must be a string.
7253 <hexastring> : test the exact string in its hexadecimal form matches in the
7254 response buffer. A health check response will be considered
7255 valid if the response's buffer contains this exact
7256 hexadecimal string.
7257 Purpose is to send binary data to ask on binary protocols.
7258
7259 Examples :
7260 # redis check in binary
7261 option tcp-check
7262 tcp-check send-binary 50494e470d0a # PING\r\n
7263 tcp-check expect binary 2b504F4e47 # +PONG
7264
7265
7266 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
7267 "tcp-check send", tune.chksize
7268
7269
Willy Tarreaue9656522010-08-17 15:40:09 +02007270tcp-request connection <action> [{if | unless} <condition>]
7271 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +02007272 May be used in sections : defaults | frontend | listen | backend
7273 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +02007274 Arguments :
7275 <action> defines the action to perform if the condition applies. Valid
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007276 actions include : "accept", "reject", "track-sc0", "track-sc1",
7277 "track-sc2", and "expect-proxy". See below for more details.
Willy Tarreau1a687942010-05-23 22:40:30 +02007278
Willy Tarreaue9656522010-08-17 15:40:09 +02007279 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007280
7281 Immediately after acceptance of a new incoming connection, it is possible to
7282 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +02007283 or dropped or have its counters tracked. Those conditions cannot make use of
7284 any data contents because the connection has not been read from yet, and the
7285 buffers are not yet allocated. This is used to selectively and very quickly
7286 accept or drop connections from various sources with a very low overhead. If
7287 some contents need to be inspected in order to take the decision, the
7288 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007289
Willy Tarreaue9656522010-08-17 15:40:09 +02007290 The "tcp-request connection" rules are evaluated in their exact declaration
7291 order. If no rule matches or if there is no rule, the default action is to
7292 accept the incoming connection. There is no specific limit to the number of
7293 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007294
Willy Tarreau18bf01e2014-06-13 16:18:52 +02007295 Five types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +02007296 - accept :
7297 accepts the connection if the condition is true (when used with "if")
7298 or false (when used with "unless"). The first such rule executed ends
7299 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007300
Willy Tarreaue9656522010-08-17 15:40:09 +02007301 - reject :
7302 rejects the connection if the condition is true (when used with "if")
7303 or false (when used with "unless"). The first such rule executed ends
7304 the rules evaluation. Rejected connections do not even become a
7305 session, which is why they are accounted separately for in the stats,
7306 as "denied connections". They are not considered for the session
7307 rate-limit and are not logged either. The reason is that these rules
7308 should only be used to filter extremely high connection rates such as
7309 the ones encountered during a massive DDoS attack. Under these extreme
7310 conditions, the simple action of logging each event would make the
7311 system collapse and would considerably lower the filtering capacity. If
7312 logging is absolutely desired, then "tcp-request content" rules should
7313 be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007314
Willy Tarreau4f0d9192013-06-11 20:40:55 +02007315 - expect-proxy layer4 :
7316 configures the client-facing connection to receive a PROXY protocol
7317 header before any byte is read from the socket. This is equivalent to
7318 having the "accept-proxy" keyword on the "bind" line, except that using
7319 the TCP rule allows the PROXY protocol to be accepted only for certain
7320 IP address ranges using an ACL. This is convenient when multiple layers
7321 of load balancers are passed through by traffic coming from public
7322 hosts.
7323
Willy Tarreau18bf01e2014-06-13 16:18:52 +02007324 - capture <sample> len <length> :
7325 This only applies to "tcp-request content" rules. It captures sample
7326 expression <sample> from the request buffer, and converts it to a
7327 string of at most <len> characters. The resulting string is stored into
7328 the next request "capture" slot, so it will possibly appear next to
7329 some captured HTTP headers. It will then automatically appear in the
7330 logs, and it will be possible to extract it using sample fetch rules to
7331 feed it into headers or anything. The length should be limited given
7332 that this size will be allocated for each capture during the whole
7333 session life. Since it applies to Please check section 7.3 (Fetching
7334 samples) and "capture request header" for more information.
7335
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007336 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +02007337 enables tracking of sticky counters from current connection. These
7338 rules do not stop evaluation and do not change default action. Two sets
7339 of counters may be simultaneously tracked by the same connection. The
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007340 first "track-sc0" rule executed enables tracking of the counters of the
7341 specified table as the first set. The first "track-sc1" rule executed
Willy Tarreaue9656522010-08-17 15:40:09 +02007342 enables tracking of the counters of the specified table as the second
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007343 set. The first "track-sc2" rule executed enables tracking of the
7344 counters of the specified table as the third set. It is a recommended
7345 practice to use the first set of counters for the per-frontend counters
7346 and the second set for the per-backend ones. But this is just a
7347 guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007348
Willy Tarreaue9656522010-08-17 15:40:09 +02007349 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02007350 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +02007351 in section 7.3. It describes what elements of the incoming
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007352 request or connection will be analysed, extracted, combined,
7353 and used to select which table entry to update the counters.
7354 Note that "tcp-request connection" cannot use content-based
7355 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007356
Willy Tarreaue9656522010-08-17 15:40:09 +02007357 <table> is an optional table to be used instead of the default one,
7358 which is the stick-table declared in the current proxy. All
7359 the counters for the matches and updates for the key will
7360 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007361
Willy Tarreaue9656522010-08-17 15:40:09 +02007362 Once a "track-sc*" rule is executed, the key is looked up in the table
7363 and if it is not found, an entry is allocated for it. Then a pointer to
7364 that entry is kept during all the session's life, and this entry's
7365 counters are updated as often as possible, every time the session's
7366 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007367 Counters are only updated for events that happen after the tracking has
7368 been started. For example, connection counters will not be updated when
7369 tracking layer 7 information, since the connection event happens before
7370 layer7 information is extracted.
7371
Willy Tarreaue9656522010-08-17 15:40:09 +02007372 If the entry tracks concurrent connection counters, one connection is
7373 counted for as long as the entry is tracked, and the entry will not
7374 expire during that time. Tracking counters also provides a performance
7375 advantage over just checking the keys, because only one table lookup is
7376 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007377
Willy Tarreaue9656522010-08-17 15:40:09 +02007378 Note that the "if/unless" condition is optional. If no condition is set on
7379 the action, it is simply performed unconditionally. That can be useful for
7380 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007381
Willy Tarreaue9656522010-08-17 15:40:09 +02007382 Example: accept all connections from white-listed hosts, reject too fast
7383 connection without counting them, and track accepted connections.
7384 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007385
Willy Tarreaue9656522010-08-17 15:40:09 +02007386 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007387 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007388 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007389
Willy Tarreaue9656522010-08-17 15:40:09 +02007390 Example: accept all connections from white-listed hosts, count all other
7391 connections and reject too fast ones. This results in abusive ones
7392 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007393
Willy Tarreaue9656522010-08-17 15:40:09 +02007394 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007395 tcp-request connection track-sc0 src
7396 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007397
Willy Tarreau4f0d9192013-06-11 20:40:55 +02007398 Example: enable the PROXY protocol for traffic coming from all known proxies.
7399
7400 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
7401
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007402 See section 7 about ACL usage.
7403
Willy Tarreaue9656522010-08-17 15:40:09 +02007404 See also : "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007405
7406
Willy Tarreaue9656522010-08-17 15:40:09 +02007407tcp-request content <action> [{if | unless} <condition>]
7408 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +02007409 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02007410 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +02007411 Arguments :
7412 <action> defines the action to perform if the condition applies. Valid
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007413 actions include : "accept", "reject", "track-sc0", "track-sc1",
Willy Tarreau18bf01e2014-06-13 16:18:52 +02007414 "track-sc2" and "capture". See "tcp-request connection" above
7415 for their signification.
Willy Tarreau62644772008-07-16 18:36:06 +02007416
Willy Tarreaue9656522010-08-17 15:40:09 +02007417 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +02007418
Willy Tarreaue9656522010-08-17 15:40:09 +02007419 A request's contents can be analysed at an early stage of request processing
7420 called "TCP content inspection". During this stage, ACL-based rules are
7421 evaluated every time the request contents are updated, until either an
7422 "accept" or a "reject" rule matches, or the TCP request inspection delay
7423 expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +02007424
Willy Tarreaue9656522010-08-17 15:40:09 +02007425 The first difference between these rules and "tcp-request connection" rules
7426 is that "tcp-request content" rules can make use of contents to take a
7427 decision. Most often, these decisions will consider a protocol recognition or
7428 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +01007429 both frontends and backends. In case of HTTP keep-alive with the client, all
7430 tcp-request content rules are evaluated again, so haproxy keeps a record of
7431 what sticky counters were assigned by a "tcp-request connection" versus a
7432 "tcp-request content" rule, and flushes all the content-related ones after
7433 processing an HTTP request, so that they may be evaluated again by the rules
7434 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007435 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +01007436 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +02007437
Willy Tarreaue9656522010-08-17 15:40:09 +02007438 Content-based rules are evaluated in their exact declaration order. If no
7439 rule matches or if there is no rule, the default action is to accept the
7440 contents. There is no specific limit to the number of rules which may be
7441 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +02007442
Willy Tarreau18bf01e2014-06-13 16:18:52 +02007443 Four types of actions are supported :
7444 - accept : the request is accepted
7445 - reject : the request is rejected and the connection is closed
7446 - capture : the specified sample expression is captured
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007447 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Willy Tarreau62644772008-07-16 18:36:06 +02007448
Willy Tarreaue9656522010-08-17 15:40:09 +02007449 They have the same meaning as their counter-parts in "tcp-request connection"
7450 so please refer to that section for a complete description.
Willy Tarreau62644772008-07-16 18:36:06 +02007451
Willy Tarreauf3338342014-01-28 21:40:28 +01007452 While there is nothing mandatory about it, it is recommended to use the
7453 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
7454 content" rules in the frontend, and track-sc2 for "tcp-request content"
7455 rules in the backend, because that makes the configuration more readable
7456 and easier to troubleshoot, but this is just a guideline and all counters
7457 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +02007458
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007459 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +02007460 the action, it is simply performed unconditionally. That can be useful for
7461 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +02007462
Willy Tarreaue9656522010-08-17 15:40:09 +02007463 It is perfectly possible to match layer 7 contents with "tcp-request content"
Willy Tarreauc0239e02012-04-16 14:42:55 +02007464 rules, since HTTP-specific ACL matches are able to preliminarily parse the
7465 contents of a buffer before extracting the required data. If the buffered
7466 contents do not parse as a valid HTTP message, then the ACL does not match.
7467 The parser which is involved there is exactly the same as for all other HTTP
Willy Tarreauf3338342014-01-28 21:40:28 +01007468 processing, so there is no risk of parsing something differently. In an HTTP
7469 backend connected to from an HTTP frontend, it is guaranteed that HTTP
7470 contents will always be immediately present when the rule is evaluated first.
Willy Tarreau62644772008-07-16 18:36:06 +02007471
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007472 Tracking layer7 information is also possible provided that the information
7473 are present when the rule is processed. The current solution for making the
7474 rule engine wait for such information is to set an inspect delay and to
7475 condition its execution with an ACL relying on such information.
7476
Willy Tarreau62644772008-07-16 18:36:06 +02007477 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +02007478 # Accept HTTP requests containing a Host header saying "example.com"
7479 # and reject everything else.
7480 acl is_host_com hdr(Host) -i example.com
7481 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +02007482 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +02007483 tcp-request content reject
7484
7485 Example:
Willy Tarreau62644772008-07-16 18:36:06 +02007486 # reject SMTP connection if client speaks first
7487 tcp-request inspect-delay 30s
7488 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007489 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +02007490
7491 # Forward HTTPS connection only if client speaks
7492 tcp-request inspect-delay 30s
7493 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02007494 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +02007495 tcp-request content reject
7496
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007497 Example:
7498 # Track the last IP from X-Forwarded-For
7499 tcp-request inspect-delay 10s
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007500 tcp-request content track-sc0 hdr(x-forwarded-for,-1) if HTTP
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007501
7502 Example:
7503 # track request counts per "base" (concatenation of Host+URL)
7504 tcp-request inspect-delay 10s
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007505 tcp-request content track-sc0 base table req-rate if HTTP
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01007506
Willy Tarreaue9656522010-08-17 15:40:09 +02007507 Example: track per-frontend and per-backend counters, block abusers at the
7508 frontend when the backend detects abuse.
7509
7510 frontend http
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007511 # Use General Purpose Couter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +02007512 # protecting all our sites
7513 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007514 tcp-request connection track-sc0 src
7515 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +02007516 ...
7517 use_backend http_dynamic if { path_end .php }
7518
7519 backend http_dynamic
7520 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007521 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +02007522 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02007523 acl click_too_fast sc1_http_req_rate gt 10
7524 acl mark_as_abuser sc0_inc_gpc0 gt 0
7525 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +02007526 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +02007527
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007528 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +02007529
Willy Tarreaue9656522010-08-17 15:40:09 +02007530 See also : "tcp-request connection", "tcp-request inspect-delay"
Willy Tarreau62644772008-07-16 18:36:06 +02007531
7532
7533tcp-request inspect-delay <timeout>
7534 Set the maximum allowed time to wait for data during content inspection
7535 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02007536 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +02007537 Arguments :
7538 <timeout> is the timeout value specified in milliseconds by default, but
7539 can be in any other unit if the number is suffixed by the unit,
7540 as explained at the top of this document.
7541
7542 People using haproxy primarily as a TCP relay are often worried about the
7543 risk of passing any type of protocol to a server without any analysis. In
7544 order to be able to analyze the request contents, we must first withhold
7545 the data then analyze them. This statement simply enables withholding of
7546 data for at most the specified amount of time.
7547
Willy Tarreaufb356202010-08-03 14:02:05 +02007548 TCP content inspection applies very early when a connection reaches a
7549 frontend, then very early when the connection is forwarded to a backend. This
7550 means that a connection may experience a first delay in the frontend and a
7551 second delay in the backend if both have tcp-request rules.
7552
Willy Tarreau62644772008-07-16 18:36:06 +02007553 Note that when performing content inspection, haproxy will evaluate the whole
7554 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007555 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +02007556 a last check is performed upon expiration, this time considering that the
Willy Tarreaud869b242009-03-15 14:43:58 +01007557 contents are definitive. If no delay is set, haproxy will not wait at all
7558 and will immediately apply a verdict based on the available information.
7559 Obviously this is unlikely to be very useful and might even be racy, so such
7560 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +02007561
7562 As soon as a rule matches, the request is released and continues as usual. If
7563 the timeout is reached and no rule matches, the default policy will be to let
7564 it pass through unaffected.
7565
7566 For most protocols, it is enough to set it to a few seconds, as most clients
7567 send the full request immediately upon connection. Add 3 or more seconds to
7568 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +01007569 to use large values, for instance to ensure that the client never talks
Willy Tarreau62644772008-07-16 18:36:06 +02007570 before the server (eg: SMTP), or to wait for a client to talk before passing
7571 data to the server (eg: SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +02007572 least the inspection delay, otherwise it will expire first. If the client
7573 closes the connection or if the buffer is full, the delay immediately expires
7574 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +02007575
Willy Tarreau55165fe2009-05-10 12:02:55 +02007576 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +02007577 "timeout client".
7578
7579
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007580tcp-response content <action> [{if | unless} <condition>]
7581 Perform an action on a session response depending on a layer 4-7 condition
7582 May be used in sections : defaults | frontend | listen | backend
7583 no | no | yes | yes
7584 Arguments :
7585 <action> defines the action to perform if the condition applies. Valid
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02007586 actions include : "accept", "close", "reject".
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007587
7588 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
7589
7590 Response contents can be analysed at an early stage of response processing
7591 called "TCP content inspection". During this stage, ACL-based rules are
7592 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02007593 "accept", "close" or a "reject" rule matches, or a TCP response inspection
7594 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007595
7596 Most often, these decisions will consider a protocol recognition or validity.
7597
7598 Content-based rules are evaluated in their exact declaration order. If no
7599 rule matches or if there is no rule, the default action is to accept the
7600 contents. There is no specific limit to the number of rules which may be
7601 inserted.
7602
7603 Two types of actions are supported :
7604 - accept :
7605 accepts the response if the condition is true (when used with "if")
7606 or false (when used with "unless"). The first such rule executed ends
7607 the rules evaluation.
7608
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02007609 - close :
7610 immediately closes the connection with the server if the condition is
7611 true (when used with "if"), or false (when used with "unless"). The
7612 first such rule executed ends the rules evaluation. The main purpose of
7613 this action is to force a connection to be finished between a client
7614 and a server after an exchange when the application protocol expects
7615 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007616 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02007617 protocols.
7618
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007619 - reject :
7620 rejects the response if the condition is true (when used with "if")
7621 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007622 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007623
7624 Note that the "if/unless" condition is optional. If no condition is set on
7625 the action, it is simply performed unconditionally. That can be useful for
7626 for changing the default action to a reject.
7627
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007628 It is perfectly possible to match layer 7 contents with "tcp-response
7629 content" rules, but then it is important to ensure that a full response has
7630 been buffered, otherwise no contents will match. In order to achieve this,
7631 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +02007632 period.
7633
7634 See section 7 about ACL usage.
7635
7636 See also : "tcp-request content", "tcp-response inspect-delay"
7637
7638
7639tcp-response inspect-delay <timeout>
7640 Set the maximum allowed time to wait for a response during content inspection
7641 May be used in sections : defaults | frontend | listen | backend
7642 no | no | yes | yes
7643 Arguments :
7644 <timeout> is the timeout value specified in milliseconds by default, but
7645 can be in any other unit if the number is suffixed by the unit,
7646 as explained at the top of this document.
7647
7648 See also : "tcp-response content", "tcp-request inspect-delay".
7649
7650
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01007651timeout check <timeout>
7652 Set additional check timeout, but only after a connection has been already
7653 established.
7654
7655 May be used in sections: defaults | frontend | listen | backend
7656 yes | no | yes | yes
7657 Arguments:
7658 <timeout> is the timeout value specified in milliseconds by default, but
7659 can be in any other unit if the number is suffixed by the unit,
7660 as explained at the top of this document.
7661
7662 If set, haproxy uses min("timeout connect", "inter") as a connect timeout
7663 for check and "timeout check" as an additional read timeout. The "min" is
7664 used so that people running with *very* long "timeout connect" (eg. those
7665 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +01007666 (Please also note that there is no valid reason to have such long connect
7667 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
7668 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01007669
7670 If "timeout check" is not set haproxy uses "inter" for complete check
7671 timeout (connect + read) exactly like all <1.3.15 version.
7672
7673 In most cases check request is much simpler and faster to handle than normal
7674 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +01007675 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01007676
7677 This parameter is specific to backends, but can be specified once for all in
7678 "defaults" sections. This is in fact one of the easiest solutions not to
7679 forget about it.
7680
Willy Tarreau41a340d2008-01-22 12:25:31 +01007681 See also: "timeout connect", "timeout queue", "timeout server",
7682 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01007683
7684
Willy Tarreau0ba27502007-12-24 16:55:16 +01007685timeout client <timeout>
7686timeout clitimeout <timeout> (deprecated)
7687 Set the maximum inactivity time on the client side.
7688 May be used in sections : defaults | frontend | listen | backend
7689 yes | yes | yes | no
7690 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01007691 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +01007692 can be in any other unit if the number is suffixed by the unit,
7693 as explained at the top of this document.
7694
7695 The inactivity timeout applies when the client is expected to acknowledge or
7696 send data. In HTTP mode, this timeout is particularly important to consider
7697 during the first phase, when the client sends the request, and during the
7698 response while it is reading data sent by the server. The value is specified
7699 in milliseconds by default, but can be in any other unit if the number is
7700 suffixed by the unit, as specified at the top of this document. In TCP mode
7701 (and to a lesser extent, in HTTP mode), it is highly recommended that the
7702 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007703 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01007704 losses by specifying timeouts that are slightly above multiples of 3 seconds
Willy Tarreauce887fd2012-05-12 12:50:00 +02007705 (eg: 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
7706 sessions (eg: WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +02007707 which overrides "timeout client" and "timeout server" for tunnels, as well as
7708 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007709
7710 This parameter is specific to frontends, but can be specified once for all in
7711 "defaults" sections. This is in fact one of the easiest solutions not to
7712 forget about it. An unspecified timeout results in an infinite timeout, which
7713 is not recommended. Such a usage is accepted and works but reports a warning
7714 during startup because it may results in accumulation of expired sessions in
7715 the system if the system's timeouts are not configured either.
7716
7717 This parameter replaces the old, deprecated "clitimeout". It is recommended
7718 to use it to write new configurations. The form "timeout clitimeout" is
7719 provided only by backwards compatibility but its use is strongly discouraged.
7720
Willy Tarreauce887fd2012-05-12 12:50:00 +02007721 See also : "clitimeout", "timeout server", "timeout tunnel".
Willy Tarreau0ba27502007-12-24 16:55:16 +01007722
7723
Willy Tarreau05cdd962014-05-10 14:30:07 +02007724timeout client-fin <timeout>
7725 Set the inactivity timeout on the client side for half-closed connections.
7726 May be used in sections : defaults | frontend | listen | backend
7727 yes | yes | yes | no
7728 Arguments :
7729 <timeout> is the timeout value specified in milliseconds by default, but
7730 can be in any other unit if the number is suffixed by the unit,
7731 as explained at the top of this document.
7732
7733 The inactivity timeout applies when the client is expected to acknowledge or
7734 send data while one direction is already shut down. This timeout is different
7735 from "timeout client" in that it only applies to connections which are closed
7736 in one direction. This is particularly useful to avoid keeping connections in
7737 FIN_WAIT state for too long when clients do not disconnect cleanly. This
7738 problem is particularly common long connections such as RDP or WebSocket.
7739 Note that this timeout can override "timeout tunnel" when a connection shuts
7740 down in one direction.
7741
7742 This parameter is specific to frontends, but can be specified once for all in
7743 "defaults" sections. By default it is not set, so half-closed connections
7744 will use the other timeouts (timeout.client or timeout.tunnel).
7745
7746 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
7747
7748
Willy Tarreau0ba27502007-12-24 16:55:16 +01007749timeout connect <timeout>
7750timeout contimeout <timeout> (deprecated)
7751 Set the maximum time to wait for a connection attempt to a server to succeed.
7752 May be used in sections : defaults | frontend | listen | backend
7753 yes | no | yes | yes
7754 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01007755 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +01007756 can be in any other unit if the number is suffixed by the unit,
7757 as explained at the top of this document.
7758
7759 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007760 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +01007761 cover one or several TCP packet losses by specifying timeouts that are
Willy Tarreau0ba27502007-12-24 16:55:16 +01007762 slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01007763 connect timeout also presets both queue and tarpit timeouts to the same value
7764 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007765
7766 This parameter is specific to backends, but can be specified once for all in
7767 "defaults" sections. This is in fact one of the easiest solutions not to
7768 forget about it. An unspecified timeout results in an infinite timeout, which
7769 is not recommended. Such a usage is accepted and works but reports a warning
7770 during startup because it may results in accumulation of failed sessions in
7771 the system if the system's timeouts are not configured either.
7772
7773 This parameter replaces the old, deprecated "contimeout". It is recommended
7774 to use it to write new configurations. The form "timeout contimeout" is
7775 provided only by backwards compatibility but its use is strongly discouraged.
7776
Willy Tarreau41a340d2008-01-22 12:25:31 +01007777 See also: "timeout check", "timeout queue", "timeout server", "contimeout",
7778 "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +01007779
7780
Willy Tarreaub16a5742010-01-10 14:46:16 +01007781timeout http-keep-alive <timeout>
7782 Set the maximum allowed time to wait for a new HTTP request to appear
7783 May be used in sections : defaults | frontend | listen | backend
7784 yes | yes | yes | yes
7785 Arguments :
7786 <timeout> is the timeout value specified in milliseconds by default, but
7787 can be in any other unit if the number is suffixed by the unit,
7788 as explained at the top of this document.
7789
7790 By default, the time to wait for a new request in case of keep-alive is set
7791 by "timeout http-request". However this is not always convenient because some
7792 people want very short keep-alive timeouts in order to release connections
7793 faster, and others prefer to have larger ones but still have short timeouts
7794 once the request has started to present itself.
7795
7796 The "http-keep-alive" timeout covers these needs. It will define how long to
7797 wait for a new HTTP request to start coming after a response was sent. Once
7798 the first byte of request has been seen, the "http-request" timeout is used
7799 to wait for the complete request to come. Note that empty lines prior to a
7800 new request do not refresh the timeout and are not counted as a new request.
7801
7802 There is also another difference between the two timeouts : when a connection
7803 expires during timeout http-keep-alive, no error is returned, the connection
7804 just closes. If the connection expires in "http-request" while waiting for a
7805 connection to complete, a HTTP 408 error is returned.
7806
7807 In general it is optimal to set this value to a few tens to hundreds of
7808 milliseconds, to allow users to fetch all objects of a page at once but
7809 without waiting for further clicks. Also, if set to a very small value (eg:
7810 1 millisecond) it will probably only accept pipelined requests but not the
7811 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +02007812 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +01007813
7814 If this parameter is not set, the "http-request" timeout applies, and if both
7815 are not set, "timeout client" still applies at the lower level. It should be
7816 set in the frontend to take effect, unless the frontend is in TCP mode, in
7817 which case the HTTP backend's timeout will be used.
7818
7819 See also : "timeout http-request", "timeout client".
7820
7821
Willy Tarreau036fae02008-01-06 13:24:40 +01007822timeout http-request <timeout>
7823 Set the maximum allowed time to wait for a complete HTTP request
7824 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +02007825 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +01007826 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01007827 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +01007828 can be in any other unit if the number is suffixed by the unit,
7829 as explained at the top of this document.
7830
7831 In order to offer DoS protection, it may be required to lower the maximum
7832 accepted time to receive a complete HTTP request without affecting the client
7833 timeout. This helps protecting against established connections on which
7834 nothing is sent. The client timeout cannot offer a good protection against
7835 this abuse because it is an inactivity timeout, which means that if the
7836 attacker sends one character every now and then, the timeout will not
7837 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +02007838 types, the request will be aborted if it does not complete in time. When the
7839 timeout expires, an HTTP 408 response is sent to the client to inform it
7840 about the problem, and the connection is closed. The logs will report
7841 termination codes "cR". Some recent browsers are having problems with this
7842 standard, well-documented behaviour, so it might be needed to hide the 408
7843 code using "errorfile 408 /dev/null". See more details in the explanations of
7844 the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +01007845
7846 Note that this timeout only applies to the header part of the request, and
7847 not to any data. As soon as the empty line is received, this timeout is not
Willy Tarreaub16a5742010-01-10 14:46:16 +01007848 used anymore. It is used again on keep-alive connections to wait for a second
7849 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +01007850
7851 Generally it is enough to set it to a few seconds, as most clients send the
7852 full request immediately upon connection. Add 3 or more seconds to cover TCP
7853 retransmits but that's all. Setting it to very low values (eg: 50 ms) will
7854 generally work on local networks as long as there are no packet losses. This
7855 will prevent people from sending bare HTTP requests using telnet.
7856
7857 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +02007858 chunk of the incoming request. It should be set in the frontend to take
7859 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
7860 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +01007861
Willy Tarreau2705a612014-05-23 17:38:34 +02007862 See also : "errorfile", "timeout http-keep-alive", "timeout client".
Willy Tarreau036fae02008-01-06 13:24:40 +01007863
Willy Tarreau844e3c52008-01-11 16:28:18 +01007864
7865timeout queue <timeout>
7866 Set the maximum time to wait in the queue for a connection slot to be free
7867 May be used in sections : defaults | frontend | listen | backend
7868 yes | no | yes | yes
7869 Arguments :
7870 <timeout> is the timeout value specified in milliseconds by default, but
7871 can be in any other unit if the number is suffixed by the unit,
7872 as explained at the top of this document.
7873
7874 When a server's maxconn is reached, connections are left pending in a queue
7875 which may be server-specific or global to the backend. In order not to wait
7876 indefinitely, a timeout is applied to requests pending in the queue. If the
7877 timeout is reached, it is considered that the request will almost never be
7878 served, so it is dropped and a 503 error is returned to the client.
7879
7880 The "timeout queue" statement allows to fix the maximum time for a request to
7881 be left pending in a queue. If unspecified, the same value as the backend's
7882 connection timeout ("timeout connect") is used, for backwards compatibility
7883 with older versions with no "timeout queue" parameter.
7884
7885 See also : "timeout connect", "contimeout".
7886
7887
7888timeout server <timeout>
7889timeout srvtimeout <timeout> (deprecated)
7890 Set the maximum inactivity time on the server side.
7891 May be used in sections : defaults | frontend | listen | backend
7892 yes | no | yes | yes
7893 Arguments :
7894 <timeout> is the timeout value specified in milliseconds by default, but
7895 can be in any other unit if the number is suffixed by the unit,
7896 as explained at the top of this document.
7897
7898 The inactivity timeout applies when the server is expected to acknowledge or
7899 send data. In HTTP mode, this timeout is particularly important to consider
7900 during the first phase of the server's response, when it has to send the
7901 headers, as it directly represents the server's processing time for the
7902 request. To find out what value to put there, it's often good to start with
7903 what would be considered as unacceptable response times, then check the logs
7904 to observe the response time distribution, and adjust the value accordingly.
7905
7906 The value is specified in milliseconds by default, but can be in any other
7907 unit if the number is suffixed by the unit, as specified at the top of this
7908 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
7909 recommended that the client timeout remains equal to the server timeout in
7910 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007911 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +01007912 packet losses by specifying timeouts that are slightly above multiples of 3
Willy Tarreauce887fd2012-05-12 12:50:00 +02007913 seconds (eg: 4 or 5 seconds minimum). If some long-lived sessions are mixed
7914 with short-lived sessions (eg: WebSocket and HTTP), it's worth considering
7915 "timeout tunnel", which overrides "timeout client" and "timeout server" for
7916 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +01007917
7918 This parameter is specific to backends, but can be specified once for all in
7919 "defaults" sections. This is in fact one of the easiest solutions not to
7920 forget about it. An unspecified timeout results in an infinite timeout, which
7921 is not recommended. Such a usage is accepted and works but reports a warning
7922 during startup because it may results in accumulation of expired sessions in
7923 the system if the system's timeouts are not configured either.
7924
7925 This parameter replaces the old, deprecated "srvtimeout". It is recommended
7926 to use it to write new configurations. The form "timeout srvtimeout" is
7927 provided only by backwards compatibility but its use is strongly discouraged.
7928
Willy Tarreauce887fd2012-05-12 12:50:00 +02007929 See also : "srvtimeout", "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +01007930
Willy Tarreau05cdd962014-05-10 14:30:07 +02007931
7932timeout server-fin <timeout>
7933 Set the inactivity timeout on the server side for half-closed connections.
7934 May be used in sections : defaults | frontend | listen | backend
7935 yes | no | yes | yes
7936 Arguments :
7937 <timeout> is the timeout value specified in milliseconds by default, but
7938 can be in any other unit if the number is suffixed by the unit,
7939 as explained at the top of this document.
7940
7941 The inactivity timeout applies when the server is expected to acknowledge or
7942 send data while one direction is already shut down. This timeout is different
7943 from "timeout server" in that it only applies to connections which are closed
7944 in one direction. This is particularly useful to avoid keeping connections in
7945 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
7946 This problem is particularly common long connections such as RDP or WebSocket.
7947 Note that this timeout can override "timeout tunnel" when a connection shuts
7948 down in one direction. This setting was provided for completeness, but in most
7949 situations, it should not be needed.
7950
7951 This parameter is specific to backends, but can be specified once for all in
7952 "defaults" sections. By default it is not set, so half-closed connections
7953 will use the other timeouts (timeout.server or timeout.tunnel).
7954
7955 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
7956
Willy Tarreau844e3c52008-01-11 16:28:18 +01007957
7958timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +01007959 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +01007960 May be used in sections : defaults | frontend | listen | backend
7961 yes | yes | yes | yes
7962 Arguments :
7963 <timeout> is the tarpit duration specified in milliseconds by default, but
7964 can be in any other unit if the number is suffixed by the unit,
7965 as explained at the top of this document.
7966
7967 When a connection is tarpitted using "reqtarpit", it is maintained open with
7968 no activity for a certain amount of time, then closed. "timeout tarpit"
7969 defines how long it will be maintained open.
7970
7971 The value is specified in milliseconds by default, but can be in any other
7972 unit if the number is suffixed by the unit, as specified at the top of this
7973 document. If unspecified, the same value as the backend's connection timeout
7974 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +01007975 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +01007976
7977 See also : "timeout connect", "contimeout".
7978
7979
Willy Tarreauce887fd2012-05-12 12:50:00 +02007980timeout tunnel <timeout>
7981 Set the maximum inactivity time on the client and server side for tunnels.
7982 May be used in sections : defaults | frontend | listen | backend
7983 yes | no | yes | yes
7984 Arguments :
7985 <timeout> is the timeout value specified in milliseconds by default, but
7986 can be in any other unit if the number is suffixed by the unit,
7987 as explained at the top of this document.
7988
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007989 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +02007990 between a client and a server, and the connection remains inactive in both
7991 directions. This timeout supersedes both the client and server timeouts once
7992 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
7993 analyser remains attached to either connection (eg: tcp content rules are
7994 accepted). In HTTP, this timeout is used when a connection is upgraded (eg:
7995 when switching to the WebSocket protocol, or forwarding a CONNECT request
7996 to a proxy), or after the first response when no keepalive/close option is
7997 specified.
7998
Willy Tarreau05cdd962014-05-10 14:30:07 +02007999 Since this timeout is usually used in conjunction with long-lived connections,
8000 it usually is a good idea to also set "timeout client-fin" to handle the
8001 situation where a client suddenly disappears from the net and does not
8002 acknowledge a close, or sends a shutdown and does not acknowledge pending
8003 data anymore. This can happen in lossy networks where firewalls are present,
8004 and is detected by the presence of large amounts of sessions in a FIN_WAIT
8005 state.
8006
Willy Tarreauce887fd2012-05-12 12:50:00 +02008007 The value is specified in milliseconds by default, but can be in any other
8008 unit if the number is suffixed by the unit, as specified at the top of this
8009 document. Whatever the expected normal idle time, it is a good practice to
8010 cover at least one or several TCP packet losses by specifying timeouts that
8011 are slightly above multiples of 3 seconds (eg: 4 or 5 seconds minimum).
8012
8013 This parameter is specific to backends, but can be specified once for all in
8014 "defaults" sections. This is in fact one of the easiest solutions not to
8015 forget about it.
8016
8017 Example :
8018 defaults http
8019 option http-server-close
8020 timeout connect 5s
8021 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +02008022 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +02008023 timeout server 30s
8024 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
8025
Willy Tarreau05cdd962014-05-10 14:30:07 +02008026 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +02008027
8028
Willy Tarreau844e3c52008-01-11 16:28:18 +01008029transparent (deprecated)
8030 Enable client-side transparent proxying
8031 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01008032 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01008033 Arguments : none
8034
8035 This keyword was introduced in order to provide layer 7 persistence to layer
8036 3 load balancers. The idea is to use the OS's ability to redirect an incoming
8037 connection for a remote address to a local process (here HAProxy), and let
8038 this process know what address was initially requested. When this option is
8039 used, sessions without cookies will be forwarded to the original destination
8040 IP address of the incoming request (which should match that of another
8041 equipment), while requests with cookies will still be forwarded to the
8042 appropriate server.
8043
8044 The "transparent" keyword is deprecated, use "option transparent" instead.
8045
8046 Note that contrary to a common belief, this option does NOT make HAProxy
8047 present the client's IP to the server when establishing the connection.
8048
Willy Tarreau844e3c52008-01-11 16:28:18 +01008049 See also: "option transparent"
8050
William Lallemanda73203e2012-03-12 12:48:57 +01008051unique-id-format <string>
8052 Generate a unique ID for each request.
8053 May be used in sections : defaults | frontend | listen | backend
8054 yes | yes | yes | no
8055 Arguments :
8056 <string> is a log-format string.
8057
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008058 This keyword creates a ID for each request using the custom log format. A
8059 unique ID is useful to trace a request passing through many components of
8060 a complex infrastructure. The newly created ID may also be logged using the
8061 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +01008062
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008063 The format should be composed from elements that are guaranteed to be
8064 unique when combined together. For instance, if multiple haproxy instances
8065 are involved, it might be important to include the node name. It is often
8066 needed to log the incoming connection's source and destination addresses
8067 and ports. Note that since multiple requests may be performed over the same
8068 connection, including a request counter may help differentiate them.
8069 Similarly, a timestamp may protect against a rollover of the counter.
8070 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +01008071
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008072 It is recommended to use hexadecimal notation for many fields since it
8073 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +01008074
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008075 Example:
William Lallemanda73203e2012-03-12 12:48:57 +01008076
Julien Vehentf21be322014-03-07 08:27:34 -05008077 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +01008078
8079 will generate:
8080
8081 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
8082
8083 See also: "unique-id-header"
8084
8085unique-id-header <name>
8086 Add a unique ID header in the HTTP request.
8087 May be used in sections : defaults | frontend | listen | backend
8088 yes | yes | yes | no
8089 Arguments :
8090 <name> is the name of the header.
8091
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008092 Add a unique-id header in the HTTP request sent to the server, using the
8093 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +01008094
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008095 Example:
William Lallemanda73203e2012-03-12 12:48:57 +01008096
Julien Vehentf21be322014-03-07 08:27:34 -05008097 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +01008098 unique-id-header X-Unique-ID
8099
8100 will generate:
8101
8102 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
8103
8104 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +01008105
Willy Tarreauf51658d2014-04-23 01:21:56 +02008106use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +02008107 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +01008108 May be used in sections : defaults | frontend | listen | backend
8109 no | yes | yes | no
8110 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01008111 <backend> is the name of a valid backend or "listen" section, or a
8112 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +01008113
Willy Tarreauf51658d2014-04-23 01:21:56 +02008114 <condition> is a condition composed of ACLs, as described in section 7. If
8115 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +01008116
8117 When doing content-switching, connections arrive on a frontend and are then
8118 dispatched to various backends depending on a number of conditions. The
8119 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +02008120 "use_backend" keyword. While it is normally used with HTTP processing, it can
8121 also be used in pure TCP, either without content using stateless ACLs (eg:
8122 source address validation) or combined with a "tcp-request" rule to wait for
8123 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +01008124
8125 There may be as many "use_backend" rules as desired. All of these rules are
8126 evaluated in their declaration order, and the first one which matches will
8127 assign the backend.
8128
8129 In the first form, the backend will be used if the condition is met. In the
8130 second form, the backend will be used if the condition is not met. If no
8131 condition is valid, the backend defined with "default_backend" will be used.
8132 If no default backend is defined, either the servers in the same section are
8133 used (in case of a "listen" section) or, in case of a frontend, no server is
8134 used and a 503 service unavailable response is returned.
8135
Willy Tarreau51aecc72009-07-12 09:47:04 +02008136 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008137 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +02008138 and backend processing will immediately follow, or the backend will wait for
8139 a complete HTTP request to get in. This feature is useful when a frontend
8140 must decode several protocols on a unique port, one of them being HTTP.
8141
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01008142 When <backend> is a simple name, it is resolved at configuration time, and an
8143 error is reported if the specified backend does not exist. If <backend> is
8144 a log-format string instead, no check may be done at configuration time, so
8145 the backend name is resolved dynamically at run time. If the resulting
8146 backend name does not correspond to any valid backend, no other rule is
8147 evaluated, and the default_backend directive is applied instead. Note that
8148 when using dynamic backend names, it is highly recommended to use a prefix
8149 that no other backend uses in order to ensure that an unauthorized backend
8150 cannot be forced from the request.
8151
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008152 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01008153 used to detect the association between frontends and backends to compute the
8154 backend's "fullconn" setting. This cannot be done for dynamic names.
8155
8156 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
8157 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +01008158
Willy Tarreau036fae02008-01-06 13:24:40 +01008159
Willy Tarreau4a5cade2012-04-05 21:09:48 +02008160use-server <server> if <condition>
8161use-server <server> unless <condition>
8162 Only use a specific server if/unless an ACL-based condition is matched.
8163 May be used in sections : defaults | frontend | listen | backend
8164 no | no | yes | yes
8165 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008166 <server> is the name of a valid server in the same backend section.
Willy Tarreau4a5cade2012-04-05 21:09:48 +02008167
8168 <condition> is a condition composed of ACLs, as described in section 7.
8169
8170 By default, connections which arrive to a backend are load-balanced across
8171 the available servers according to the configured algorithm, unless a
8172 persistence mechanism such as a cookie is used and found in the request.
8173
8174 Sometimes it is desirable to forward a particular request to a specific
8175 server without having to declare a dedicated backend for this server. This
8176 can be achieved using the "use-server" rules. These rules are evaluated after
8177 the "redirect" rules and before evaluating cookies, and they have precedence
8178 on them. There may be as many "use-server" rules as desired. All of these
8179 rules are evaluated in their declaration order, and the first one which
8180 matches will assign the server.
8181
8182 If a rule designates a server which is down, and "option persist" is not used
8183 and no force-persist rule was validated, it is ignored and evaluation goes on
8184 with the next rules until one matches.
8185
8186 In the first form, the server will be used if the condition is met. In the
8187 second form, the server will be used if the condition is not met. If no
8188 condition is valid, the processing continues and the server will be assigned
8189 according to other persistence mechanisms.
8190
8191 Note that even if a rule is matched, cookie processing is still performed but
8192 does not assign the server. This allows prefixed cookies to have their prefix
8193 stripped.
8194
8195 The "use-server" statement works both in HTTP and TCP mode. This makes it
8196 suitable for use with content-based inspection. For instance, a server could
8197 be selected in a farm according to the TLS SNI field. And if these servers
8198 have their weight set to zero, they will not be used for other traffic.
8199
8200 Example :
8201 # intercept incoming TLS requests based on the SNI field
8202 use-server www if { req_ssl_sni -i www.example.com }
8203 server www 192.168.0.1:443 weight 0
8204 use-server mail if { req_ssl_sni -i mail.example.com }
8205 server mail 192.168.0.1:587 weight 0
8206 use-server imap if { req_ssl_sni -i imap.example.com }
8207 server mail 192.168.0.1:993 weight 0
8208 # all the rest is forwarded to this server
8209 server default 192.168.0.2:443 check
8210
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008211 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +02008212
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008213
82145. Bind and Server options
8215--------------------------
8216
8217The "bind", "server" and "default-server" keywords support a number of settings
8218depending on some build options and on the system HAProxy was built on. These
8219settings generally each consist in one word sometimes followed by a value,
8220written on the same line as the "bind" or "server" line. All these options are
8221described in this section.
8222
8223
82245.1. Bind options
8225-----------------
8226
8227The "bind" keyword supports a certain number of settings which are all passed
8228as arguments on the same line. The order in which those arguments appear makes
8229no importance, provided that they appear after the bind address. All of these
8230parameters are optional. Some of them consist in a single words (booleans),
8231while other ones expect a value after them. In this case, the value must be
8232provided immediately after the setting name.
8233
8234The currently supported settings are the following ones.
8235
8236accept-proxy
8237 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +02008238 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
8239 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008240 3/4 addresses of the incoming connection to be used everywhere an address is
8241 used, with the only exception of "tcp-request connection" rules which will
8242 only see the real connection address. Logs will reflect the addresses
8243 indicated in the protocol, unless it is violated, in which case the real
8244 address will still be used. This keyword combined with support from external
8245 components can be used as an efficient and reliable alternative to the
8246 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +02008247 usable. See also "tcp-request connection expect-proxy" for a finer-grained
8248 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008249
Willy Tarreauab861d32013-04-02 02:30:41 +02008250alpn <protocols>
8251 This enables the TLS ALPN extension and advertises the specified protocol
8252 list as supported on top of ALPN. The protocol list consists in a comma-
8253 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
8254 quotes). This requires that the SSL library is build with support for TLS
8255 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
8256 initial NPN extension.
8257
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008258backlog <backlog>
8259 Sets the socket's backlog to this value. If unspecified, the frontend's
8260 backlog is used instead, which generally defaults to the maxconn value.
8261
Emeric Brun7fb34422012-09-28 15:26:15 +02008262ecdhe <named curve>
8263 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +01008264 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
8265 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +02008266
Emeric Brunfd33a262012-10-11 16:28:27 +02008267ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +02008268 This setting is only available when support for OpenSSL was built in. It
8269 designates a PEM file from which to load CA certificates used to verify
8270 client's certificate.
8271
Emeric Brunb6dc9342012-09-28 17:55:37 +02008272ca-ignore-err [all|<errorID>,...]
8273 This setting is only available when support for OpenSSL was built in.
8274 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
8275 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
8276 error is ignored.
8277
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008278ciphers <ciphers>
8279 This setting is only available when support for OpenSSL was built in. It sets
8280 the string describing the list of cipher algorithms ("cipher suite") that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008281 negotiated during the SSL/TLS handshake. The format of the string is defined
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008282 in "man 1 ciphers" from OpenSSL man pages, and can be for instance a string
8283 such as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes).
8284
Emeric Brunfd33a262012-10-11 16:28:27 +02008285crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +02008286 This setting is only available when support for OpenSSL was built in. It
8287 designates a PEM file from which to load certificate revocation list used
8288 to verify client's certificate.
8289
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008290crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +00008291 This setting is only available when support for OpenSSL was built in. It
8292 designates a PEM file containing both the required certificates and any
8293 associated private keys. This file can be built by concatenating multiple
8294 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
8295 requires an intermediate certificate, this can also be concatenated into this
8296 file.
8297
8298 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
8299 are loaded.
8300
8301 If a directory name is used instead of a PEM file, then all files found in
Emeric Brun4147b2e2014-06-16 18:36:30 +02008302 that directory will be loaded unless their name ends with '.issuer' or
8303 '.ocsp' (reserved extensions). This directive may be specified multiple times
Alex Davies0fbf0162013-03-02 16:04:50 +00008304 in order to load certificates from multiple files or directories. The
8305 certificates will be presented to clients who provide a valid TLS Server Name
8306 Indication field matching one of their CN or alt subjects. Wildcards are
8307 supported, where a wildcard character '*' is used instead of the first
8308 hostname component (eg: *.example.org matches www.example.org but not
8309 www.sub.example.org).
8310
8311 If no SNI is provided by the client or if the SSL library does not support
8312 TLS extensions, or if the client provides an SNI hostname which does not
8313 match any certificate, then the first loaded certificate will be presented.
8314 This means that when loading certificates from a directory, it is highly
8315 recommended to load the default one first as a file.
8316
Emeric Brune032bfa2012-09-28 13:01:45 +02008317 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008318
Alex Davies0fbf0162013-03-02 16:04:50 +00008319 Some CAs (such as Godaddy) offer a drop down list of server types that do not
8320 include HAProxy when obtaining a certificate. If this happens be sure to
Godbach8bf60a12014-04-21 21:42:41 +08008321 choose a webserver that the CA believes requires an intermediate CA (for
Alex Davies0fbf0162013-03-02 16:04:50 +00008322 Godaddy, selection Apache Tomcat will get the correct bundle, but many
8323 others, e.g. nginx, result in a wrong bundle that will not work for some
8324 clients).
8325
Emeric Brun4147b2e2014-06-16 18:36:30 +02008326 For each PEM file, haproxy checks for the presence of file at the same path
8327 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
8328 Status Request extension (also known as "OCSP stapling") is automatically
8329 enabled. The content of this file is optional. If not empty, it must contain
8330 a valid OCSP Response in DER format. In order to be valid an OCSP Response
8331 must comply with the following rules: it has to indicate a good status,
8332 it has to be a single response for the certificate of the PEM file, and it
8333 has to be valid at the moment of addition. If these rules are not respected
8334 the OCSP Response is ignored and a warning is emitted. In order to identify
8335 which certificate an OCSP Response applies to, the issuer's certificate is
8336 necessary. If the issuer's certificate is not found in the PEM file, it will
8337 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
8338 if it exists otherwise it will fail with an error.
8339
Emeric Brunb6dc9342012-09-28 17:55:37 +02008340crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +00008341 This setting is only available when support for OpenSSL was built in. Sets a
8342 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008343 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +00008344 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +02008345
Emmanuel Hocdetfe616562013-01-22 15:31:15 +01008346crt-list <file>
8347 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +02008348 designates a list of PEM file with an optional list of SNI filter per
8349 certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +01008350
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +02008351 <crtfile> [[!]<snifilter> ...]
Emmanuel Hocdetfe616562013-01-22 15:31:15 +01008352
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +02008353 Wildcards are supported in the SNI filter. Negative filter are also supported,
8354 only useful in combination with a wildcard filter to exclude a particular SNI.
8355 The certificates will be presented to clients who provide a valid TLS Server
8356 Name Indication field matching one of the SNI filters. If no SNI filter is
8357 specified, the CN and alt subjects are used. This directive may be specified
8358 multiple times. See the "crt" option for more information. The default
8359 certificate is still needed to meet OpenSSL expectations. If it is not used,
8360 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +01008361
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008362defer-accept
8363 Is an optional keyword which is supported only on certain Linux kernels. It
8364 states that a connection will only be accepted once some data arrive on it,
8365 or at worst after the first retransmit. This should be used only on protocols
8366 for which the client talks first (eg: HTTP). It can slightly improve
8367 performance by ensuring that most of the request is already available when
8368 the connection is accepted. On the other hand, it will not be able to detect
8369 connections which don't talk. It is important to note that this option is
8370 broken in all kernels up to 2.6.31, as the connection is never accepted until
8371 the client talks. This can cause issues with front firewalls which would see
8372 an established connection while the proxy will only see it in SYN_RECV. This
8373 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
8374
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008375force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008376 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008377 this listener. SSLv3 is generally less expensive than the TLS counterparts
8378 for high connection rates. See also "force-tls*", "no-sslv3", and "no-tls*".
8379
8380force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008381 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008382 this listener. See also "force-tls*", "no-sslv3", and "no-tls*".
8383
8384force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008385 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008386 this listener. See also "force-tls*", "no-sslv3", and "no-tls*".
8387
8388force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008389 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008390 this listener. See also "force-tls*", "no-sslv3", and "no-tls*".
8391
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008392gid <gid>
8393 Sets the group of the UNIX sockets to the designated system gid. It can also
8394 be set by default in the global section's "unix-bind" statement. Note that
8395 some platforms simply ignore this. This setting is equivalent to the "group"
8396 setting except that the group ID is used instead of its name. This setting is
8397 ignored by non UNIX sockets.
8398
8399group <group>
8400 Sets the group of the UNIX sockets to the designated system group. It can
8401 also be set by default in the global section's "unix-bind" statement. Note
8402 that some platforms simply ignore this. This setting is equivalent to the
8403 "gid" setting except that the group name is used instead of its gid. This
8404 setting is ignored by non UNIX sockets.
8405
8406id <id>
8407 Fixes the socket ID. By default, socket IDs are automatically assigned, but
8408 sometimes it is more convenient to fix them to ease monitoring. This value
8409 must be strictly positive and unique within the listener/frontend. This
8410 option can only be used when defining only a single socket.
8411
8412interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +01008413 Restricts the socket to a specific interface. When specified, only packets
8414 received from that particular interface are processed by the socket. This is
8415 currently only supported on Linux. The interface must be a primary system
8416 interface, not an aliased interface. It is also possible to bind multiple
8417 frontends to the same address if they are bound to different interfaces. Note
8418 that binding to a network interface requires root privileges. This parameter
8419 is only compatible with TCPv4/TCPv6 sockets.
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008420
Willy Tarreauabb175f2012-09-24 12:43:26 +02008421level <level>
8422 This setting is used with the stats sockets only to restrict the nature of
8423 the commands that can be issued on the socket. It is ignored by other
8424 sockets. <level> can be one of :
8425 - "user" is the least privileged level ; only non-sensitive stats can be
8426 read, and no change is allowed. It would make sense on systems where it
8427 is not easy to restrict access to the socket.
8428 - "operator" is the default level and fits most common uses. All data can
8429 be read, and only non-sensitive changes are permitted (eg: clear max
8430 counters).
8431 - "admin" should be used with care, as everything is permitted (eg: clear
8432 all counters).
8433
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008434maxconn <maxconn>
8435 Limits the sockets to this number of concurrent connections. Extraneous
8436 connections will remain in the system's backlog until a connection is
8437 released. If unspecified, the limit will be the same as the frontend's
8438 maxconn. Note that in case of port ranges or multiple addresses, the same
8439 value will be applied to each socket. This setting enables different
8440 limitations on expensive sockets, for instance SSL entries which may easily
8441 eat all memory.
8442
8443mode <mode>
8444 Sets the octal mode used to define access permissions on the UNIX socket. It
8445 can also be set by default in the global section's "unix-bind" statement.
8446 Note that some platforms simply ignore this. This setting is ignored by non
8447 UNIX sockets.
8448
8449mss <maxseg>
8450 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
8451 connections. This can be used to force a lower MSS for certain specific
8452 ports, for instance for connections passing through a VPN. Note that this
8453 relies on a kernel feature which is theoretically supported under Linux but
8454 was buggy in all versions prior to 2.6.28. It may or may not work on other
8455 operating systems. It may also not change the advertised value but change the
8456 effective size of outgoing segments. The commonly advertised value for TCPv4
8457 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
8458 positive, it will be used as the advertised MSS. If it is negative, it will
8459 indicate by how much to reduce the incoming connection's advertised MSS for
8460 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
8461
8462name <name>
8463 Sets an optional name for these sockets, which will be reported on the stats
8464 page.
8465
8466nice <nice>
8467 Sets the 'niceness' of connections initiated from the socket. Value must be
8468 in the range -1024..1024 inclusive, and defaults to zero. Positive values
8469 means that such connections are more friendly to others and easily offer
8470 their place in the scheduler. On the opposite, negative values mean that
8471 connections want to run with a higher priority than others. The difference
8472 only happens under high loads when the system is close to saturation.
8473 Negative values are appropriate for low-latency or administration services,
8474 and high values are generally recommended for CPU intensive tasks such as SSL
8475 processing or bulk transfers which are less sensible to latency. For example,
8476 it may make sense to use a positive value for an SMTP socket and a negative
8477 one for an RDP socket.
8478
Emeric Brun9b3009b2012-10-05 11:55:06 +02008479no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008480 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008481 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008482 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008483 be enabled using any configuration option. See also "force-tls*",
8484 and "force-sslv3".
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008485
Emeric Brun90ad8722012-10-02 14:00:59 +02008486no-tls-tickets
8487 This setting is only available when support for OpenSSL was built in. It
8488 disables the stateless session resumption (RFC 5077 TLS Ticket
8489 extension) and force to use stateful session resumption. Stateless
8490 session resumption is more expensive in CPU usage.
8491
Emeric Brun9b3009b2012-10-05 11:55:06 +02008492no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008493 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008494 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008495 when SSL is supported. Note that SSLv2 is forced disabled in the code and
8496 cannot be enabled using any configuration option. See also "force-tls*",
8497 and "force-sslv3".
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008498
Emeric Brun9b3009b2012-10-05 11:55:06 +02008499no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +02008500 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008501 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008502 when SSL is supported. Note that SSLv2 is forced disabled in the code and
8503 cannot be enabled using any configuration option. See also "force-tls*",
8504 and "force-sslv3".
Emeric Brunf5da4932012-09-28 19:42:54 +02008505
Emeric Brun9b3009b2012-10-05 11:55:06 +02008506no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +02008507 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008508 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +02008509 when SSL is supported. Note that SSLv2 is forced disabled in the code and
8510 cannot be enabled using any configuration option. See also "force-tls*",
8511 and "force-sslv3".
Emeric Brunf5da4932012-09-28 19:42:54 +02008512
Willy Tarreau6c9a3d52012-10-18 18:57:14 +02008513npn <protocols>
8514 This enables the NPN TLS extension and advertises the specified protocol list
8515 as supported on top of NPN. The protocol list consists in a comma-delimited
8516 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
8517 This requires that the SSL library is build with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +02008518 enabled (check with haproxy -vv). Note that the NPN extension has been
8519 replaced with the ALPN extension (see the "alpn" keyword).
Willy Tarreau6c9a3d52012-10-18 18:57:14 +02008520
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02008521process [ all | odd | even | <number 1-64>[-<number 1-64>] ]
8522 This restricts the list of processes on which this listener is allowed to
8523 run. It does not enforce any process but eliminates those which do not match.
8524 If the frontend uses a "bind-process" setting, the intersection between the
8525 two is applied. If in the end the listener is not allowed to run on any
8526 remaining process, a warning is emitted, and the listener will either run on
8527 the first process of the listener if a single process was specified, or on
8528 all of its processes if multiple processes were specified. For the unlikely
Willy Tarreauae302532014-05-07 19:22:24 +02008529 case where several ranges are needed, this directive may be repeated. The
8530 main purpose of this directive is to be used with the stats sockets and have
8531 one different socket per process. The second purpose is to have multiple bind
8532 lines sharing the same IP:port but not the same process in a listener, so
8533 that the system can distribute the incoming connections into multiple queues
8534 and allow a smoother inter-process load balancing. Currently Linux 3.9 and
8535 above is known for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02008536
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008537ssl
8538 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008539 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008540 certificate is necessary (see "crt" above). All contents in the buffers will
8541 appear in clear text, so that ACLs and HTTP processing will only have access
8542 to deciphered contents.
8543
Emmanuel Hocdet65623372013-01-24 17:17:15 +01008544strict-sni
8545 This setting is only available when support for OpenSSL was built in. The
8546 SSL/TLS negotiation is allow only if the client provided an SNI which match
8547 a certificate. The default certificate is not used.
8548 See the "crt" option for more information.
8549
Willy Tarreau1c862c52012-10-05 16:21:00 +02008550tfo
Lukas Tribus0defb902013-02-13 23:35:39 +01008551 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +02008552 enables TCP Fast Open on the listening socket, which means that clients which
8553 support this feature will be able to send a request and receive a response
8554 during the 3-way handshake starting from second connection, thus saving one
8555 round-trip after the first connection. This only makes sense with protocols
8556 that use high connection rates and where each round trip matters. This can
8557 possibly cause issues with many firewalls which do not accept data on SYN
8558 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +02008559 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
8560 need to build HAProxy with USE_TFO=1 if your libc doesn't define
8561 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +02008562
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008563transparent
8564 Is an optional keyword which is supported only on certain Linux kernels. It
8565 indicates that the addresses will be bound even if they do not belong to the
8566 local machine, and that packets targeting any of these addresses will be
8567 intercepted just as if the addresses were locally configured. This normally
8568 requires that IP forwarding is enabled. Caution! do not use this with the
8569 default address '*', as it would redirect any traffic for the specified port.
8570 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
8571 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
8572 kernel version. Some distribution kernels include backports of the feature,
8573 so check for support with your vendor.
8574
Willy Tarreau77e3af92012-11-24 15:07:23 +01008575v4v6
8576 Is an optional keyword which is supported only on most recent systems
8577 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
8578 and IPv6 when it uses the default address. Doing so is sometimes necessary
8579 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008580 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +01008581
Willy Tarreau9b6700f2012-11-24 11:55:28 +01008582v6only
8583 Is an optional keyword which is supported only on most recent systems
8584 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
8585 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +01008586 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
8587 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +01008588
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008589uid <uid>
8590 Sets the owner of the UNIX sockets to the designated system uid. It can also
8591 be set by default in the global section's "unix-bind" statement. Note that
8592 some platforms simply ignore this. This setting is equivalent to the "user"
8593 setting except that the user numeric ID is used instead of its name. This
8594 setting is ignored by non UNIX sockets.
8595
8596user <user>
8597 Sets the owner of the UNIX sockets to the designated system user. It can also
8598 be set by default in the global section's "unix-bind" statement. Note that
8599 some platforms simply ignore this. This setting is equivalent to the "uid"
8600 setting except that the user name is used instead of its uid. This setting is
8601 ignored by non UNIX sockets.
8602
Emeric Brun1a073b42012-09-28 17:07:34 +02008603verify [none|optional|required]
8604 This setting is only available when support for OpenSSL was built in. If set
8605 to 'none', client certificate is not requested. This is the default. In other
8606 cases, a client certificate is requested. If the client does not provide a
8607 certificate after the request and if 'verify' is set to 'required', then the
8608 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +02008609 certificate provided by the client is always verified using CAs from
8610 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
8611 is aborted, regardless of the 'verify' option, unless the error code exactly
8612 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +02008613
Willy Tarreaub6205fd2012-09-24 12:27:33 +020086145.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +01008615------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02008616
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01008617The "server" and "default-server" keywords support a certain number of settings
8618which are all passed as arguments on the server line. The order in which those
8619arguments appear does not count, and they are all optional. Some of those
8620settings are single words (booleans) while others expect one or several values
8621after them. In this case, the values must immediately follow the setting name.
8622Except default-server, all those settings must be specified after the server's
8623address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +02008624
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008625 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01008626 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +02008627
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008628The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +01008629
Willy Tarreauceb4ac92012-04-28 00:41:46 +02008630addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008631 Using the "addr" parameter, it becomes possible to use a different IP address
8632 to send health-checks. On some servers, it may be desirable to dedicate an IP
8633 address to specific component able to perform complex tests which are more
8634 suitable to health-checks than the application. This parameter is ignored if
8635 the "check" parameter is not set. See also the "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +02008636
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008637 Supported in default-server: No
8638
Simon Hormand60d6912013-11-25 10:46:36 +09008639agent-check
8640 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +01008641 health check. An agent health check is performed by making a TCP connection
8642 to the port set by the "agent-port" parameter and reading an ASCII string.
8643 The string is made of a series of words delimited by spaces, tabs or commas
8644 in any order, optionally terminated by '\r' and/or '\n', each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +09008645
Willy Tarreau81f5d942013-12-09 20:51:51 +01008646 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +09008647 Values in this format will set the weight proportional to the initial
8648 weight of a server as configured when haproxy starts.
8649
Willy Tarreau81f5d942013-12-09 20:51:51 +01008650 - The word "ready". This will turn the server's administrative state to the
8651 READY mode, thus cancelling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +09008652
Willy Tarreau81f5d942013-12-09 20:51:51 +01008653 - The word "drain". This will turn the server's administrative state to the
8654 DRAIN mode, thus it will not accept any new connections other than those
8655 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +09008656
Willy Tarreau81f5d942013-12-09 20:51:51 +01008657 - The word "maint". This will turn the server's administrative state to the
8658 MAINT mode, thus it will not accept any new connections at all, and health
8659 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +09008660
Willy Tarreau81f5d942013-12-09 20:51:51 +01008661 - The words "down", "failed", or "stopped", optionally followed by a
8662 description string after a sharp ('#'). All of these mark the server's
8663 operating state as DOWN, but since the word itself is reported on the stats
8664 page, the difference allows an administrator to know if the situation was
8665 expected or not : the service may intentionally be stopped, may appear up
8666 but fail some validity tests, or may be seen as down (eg: missing process,
8667 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +09008668
Willy Tarreau81f5d942013-12-09 20:51:51 +01008669 - The word "up" sets back the server's operating state as UP if health checks
8670 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +09008671
Willy Tarreau81f5d942013-12-09 20:51:51 +01008672 Parameters which are not advertised by the agent are not changed. For
8673 example, an agent might be designed to monitor CPU usage and only report a
8674 relative weight and never interact with the operating status. Similarly, an
8675 agent could be designed as an end-user interface with 3 radio buttons
8676 allowing an administrator to change only the administrative state. However,
8677 it is important to consider that only the agent may revert its own actions,
8678 so if a server is set to DRAIN mode or to DOWN state using the agent, the
8679 agent must implement the other equivalent actions to bring the service into
8680 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +09008681
Simon Horman2f1f9552013-11-25 10:46:37 +09008682 Failure to connect to the agent is not considered an error as connectivity
8683 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +01008684 parameter. Warning though, it is not a good idea to stop an agent after it
8685 reports "down", since only an agent reporting "up" will be able to turn the
8686 server up again. Note that the CLI on the Unix stats socket is also able to
8687 force an agent's result in order to workaround a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +09008688
Willy Tarreau81f5d942013-12-09 20:51:51 +01008689 Requires the "agent-port" parameter to be set. See also the "agent-inter"
8690 parameter.
Simon Hormand60d6912013-11-25 10:46:36 +09008691
8692 Supported in default-server: No
8693
8694agent-inter <delay>
8695 The "agent-inter" parameter sets the interval between two agent checks
8696 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
8697
8698 Just as with every other time-based parameter, it may be entered in any
8699 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
8700 parameter also serves as a timeout for agent checks "timeout check" is
8701 not set. In order to reduce "resonance" effects when multiple servers are
8702 hosted on the same hardware, the agent and health checks of all servers
8703 are started with a small time offset between them. It is also possible to
8704 add some random noise in the agent and health checks interval using the
8705 global "spread-checks" keyword. This makes sense for instance when a lot
8706 of backends use the same servers.
8707
8708 See also the "agent-check" and "agent-port" parameters.
8709
8710 Supported in default-server: Yes
8711
8712agent-port <port>
8713 The "agent-port" parameter sets the TCP port used for agent checks.
8714
8715 See also the "agent-check" and "agent-inter" parameters.
8716
8717 Supported in default-server: Yes
8718
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008719backup
8720 When "backup" is present on a server line, the server is only used in load
8721 balancing when all other non-backup servers are unavailable. Requests coming
8722 with a persistence cookie referencing the server will always be served
8723 though. By default, only the first operational backup server is used, unless
8724 the "allbackups" option is set in the backend. See also the "allbackups"
8725 option.
Willy Tarreau6a06a402007-07-15 20:15:28 +02008726
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008727 Supported in default-server: No
8728
Emeric Brunef42d922012-10-11 16:11:36 +02008729ca-file <cafile>
8730 This setting is only available when support for OpenSSL was built in. It
8731 designates a PEM file from which to load CA certificates used to verify
8732 server's certificate.
8733
8734 Supported in default-server: No
8735
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008736check
8737 This option enables health checks on the server. By default, a server is
Patrick Mézardb7aeec62012-01-22 16:01:22 +01008738 always considered available. If "check" is set, the server is available when
8739 accepting periodic TCP connections, to ensure that it is really able to serve
8740 requests. The default address and port to send the tests to are those of the
8741 server, and the default source is the same as the one defined in the
8742 backend. It is possible to change the address using the "addr" parameter, the
8743 port using the "port" parameter, the source address using the "source"
8744 address, and the interval and timers using the "inter", "rise" and "fall"
Simon Hormanafc47ee2013-11-25 10:46:35 +09008745 parameters. The request method is define in the backend using the "httpchk",
8746 "smtpchk", "mysql-check", "pgsql-check" and "ssl-hello-chk" options. Please
8747 refer to those options and parameters for more information.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008748
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008749 Supported in default-server: No
8750
Willy Tarreau6c16adc2012-10-05 00:04:16 +02008751check-send-proxy
8752 This option forces emission of a PROXY protocol line with outgoing health
8753 checks, regardless of whether the server uses send-proxy or not for the
8754 normal traffic. By default, the PROXY protocol is enabled for health checks
8755 if it is already enabled for normal traffic and if no "port" nor "addr"
8756 directive is present. However, if such a directive is present, the
8757 "check-send-proxy" option needs to be used to force the use of the
8758 protocol. See also the "send-proxy" option for more information.
8759
8760 Supported in default-server: No
8761
Willy Tarreau763a95b2012-10-04 23:15:39 +02008762check-ssl
8763 This option forces encryption of all health checks over SSL, regardless of
8764 whether the server uses SSL or not for the normal traffic. This is generally
8765 used when an explicit "port" or "addr" directive is specified and SSL health
8766 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008767 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +02008768 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
8769 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
8770 All SSL settings are common to health checks and traffic (eg: ciphers).
8771 See the "ssl" option for more information.
8772
8773 Supported in default-server: No
8774
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008775ciphers <ciphers>
8776 This option sets the string describing the list of cipher algorithms that is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008777 is negotiated during the SSL/TLS handshake with the server. The format of the
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008778 string is defined in "man 1 ciphers". When SSL is used to communicate with
8779 servers on the local network, it is common to see a weaker set of algorithms
8780 than what is used over the internet. Doing so reduces CPU usage on both the
8781 server and haproxy while still keeping it compatible with deployed software.
8782 Some algorithms such as RC4-SHA1 are reasonably cheap. If no security at all
8783 is needed and just connectivity, using DES can be appropriate.
8784
Willy Tarreau763a95b2012-10-04 23:15:39 +02008785 Supported in default-server: No
8786
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008787cookie <value>
8788 The "cookie" parameter sets the cookie value assigned to the server to
8789 <value>. This value will be checked in incoming requests, and the first
8790 operational server possessing the same value will be selected. In return, in
8791 cookie insertion or rewrite modes, this value will be assigned to the cookie
8792 sent to the client. There is nothing wrong in having several servers sharing
8793 the same cookie value, and it is in fact somewhat common between normal and
8794 backup servers. See also the "cookie" keyword in backend section.
8795
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008796 Supported in default-server: No
8797
Emeric Brunef42d922012-10-11 16:11:36 +02008798crl-file <crlfile>
8799 This setting is only available when support for OpenSSL was built in. It
8800 designates a PEM file from which to load certificate revocation list used
8801 to verify server's certificate.
8802
8803 Supported in default-server: No
8804
Emeric Bruna7aa3092012-10-26 12:58:00 +02008805crt <cert>
8806 This setting is only available when support for OpenSSL was built in.
8807 It designates a PEM file from which to load both a certificate and the
8808 associated private key. This file can be built by concatenating both PEM
8809 files into one. This certificate will be sent if the server send a client
8810 certificate request.
8811
8812 Supported in default-server: No
8813
Willy Tarreau96839092010-03-29 10:02:24 +02008814disabled
8815 The "disabled" keyword starts the server in the "disabled" state. That means
8816 that it is marked down in maintenance mode, and no connection other than the
8817 ones allowed by persist mode will reach it. It is very well suited to setup
8818 new servers, because normal traffic will never reach them, while it is still
8819 possible to test the service by making use of the force-persist mechanism.
8820
8821 Supported in default-server: No
8822
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008823error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +01008824 If health observing is enabled, the "error-limit" parameter specifies the
8825 number of consecutive errors that triggers event selected by the "on-error"
8826 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008827
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008828 Supported in default-server: Yes
8829
8830 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008831
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008832fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008833 The "fall" parameter states that a server will be considered as dead after
8834 <count> consecutive unsuccessful health checks. This value defaults to 3 if
8835 unspecified. See also the "check", "inter" and "rise" parameters.
8836
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008837 Supported in default-server: Yes
8838
Emeric Brun8694b9a2012-10-05 14:39:07 +02008839force-sslv3
8840 This option enforces use of SSLv3 only when SSL is used to communicate with
8841 the server. SSLv3 is generally less expensive than the TLS counterparts for
8842 high connection rates. See also "no-tlsv*", "no-sslv3".
8843
8844 Supported in default-server: No
8845
8846force-tlsv10
8847 This option enforces use of TLSv1.0 only when SSL is used to communicate with
8848 the server. See also "no-tlsv*", "no-sslv3".
8849
8850 Supported in default-server: No
8851
8852force-tlsv11
8853 This option enforces use of TLSv1.1 only when SSL is used to communicate with
8854 the server. See also "no-tlsv*", "no-sslv3".
8855
8856 Supported in default-server: No
8857
8858force-tlsv12
8859 This option enforces use of TLSv1.2 only when SSL is used to communicate with
8860 the server. See also "no-tlsv*", "no-sslv3".
8861
8862 Supported in default-server: No
8863
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008864id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02008865 Set a persistent ID for the server. This ID must be positive and unique for
8866 the proxy. An unused ID will automatically be assigned if unset. The first
8867 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008868
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008869 Supported in default-server: No
8870
8871inter <delay>
8872fastinter <delay>
8873downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008874 The "inter" parameter sets the interval between two consecutive health checks
8875 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
8876 It is also possible to use "fastinter" and "downinter" to optimize delays
8877 between checks depending on the server state :
8878
8879 Server state | Interval used
8880 ---------------------------------+-----------------------------------------
8881 UP 100% (non-transitional) | "inter"
8882 ---------------------------------+-----------------------------------------
8883 Transitionally UP (going down), |
8884 Transitionally DOWN (going up), | "fastinter" if set, "inter" otherwise.
8885 or yet unchecked. |
8886 ---------------------------------+-----------------------------------------
8887 DOWN 100% (non-transitional) | "downinter" if set, "inter" otherwise.
8888 ---------------------------------+-----------------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +01008889
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008890 Just as with every other time-based parameter, they can be entered in any
8891 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
8892 serves as a timeout for health checks sent to servers if "timeout check" is
8893 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +09008894 hosted on the same hardware, the agent and health checks of all servers
8895 are started with a small time offset between them. It is also possible to
8896 add some random noise in the agent and health checks interval using the
8897 global "spread-checks" keyword. This makes sense for instance when a lot
8898 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008899
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008900 Supported in default-server: Yes
8901
8902maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008903 The "maxconn" parameter specifies the maximal number of concurrent
8904 connections that will be sent to this server. If the number of incoming
8905 concurrent requests goes higher than this value, they will be queued, waiting
8906 for a connection to be released. This parameter is very important as it can
8907 save fragile servers from going down under extreme loads. If a "minconn"
8908 parameter is specified, the limit becomes dynamic. The default value is "0"
8909 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
8910 the backend's "fullconn" keyword.
8911
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008912 Supported in default-server: Yes
8913
8914maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008915 The "maxqueue" parameter specifies the maximal number of connections which
8916 will wait in the queue for this server. If this limit is reached, next
8917 requests will be redispatched to other servers instead of indefinitely
8918 waiting to be served. This will break persistence but may allow people to
8919 quickly re-log in when the server they try to connect to is dying. The
8920 default value is "0" which means the queue is unlimited. See also the
8921 "maxconn" and "minconn" parameters.
8922
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008923 Supported in default-server: Yes
8924
8925minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008926 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
8927 limit following the backend's load. The server will always accept at least
8928 <minconn> connections, never more than <maxconn>, and the limit will be on
8929 the ramp between both values when the backend has less than <fullconn>
8930 concurrent connections. This makes it possible to limit the load on the
8931 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008932 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008933 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008934
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008935 Supported in default-server: Yes
8936
Emeric Brun9b3009b2012-10-05 11:55:06 +02008937no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008938 This option disables support for SSLv3 when SSL is used to communicate with
8939 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emeric Brun8694b9a2012-10-05 14:39:07 +02008940 using any configuration option. See also "force-sslv3", "force-tlsv*".
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008941
Willy Tarreau763a95b2012-10-04 23:15:39 +02008942 Supported in default-server: No
8943
Emeric Brunf9c5c472012-10-11 15:28:34 +02008944no-tls-tickets
8945 This setting is only available when support for OpenSSL was built in. It
8946 disables the stateless session resumption (RFC 5077 TLS Ticket
8947 extension) and force to use stateful session resumption. Stateless
8948 session resumption is more expensive in CPU usage for servers.
8949
8950 Supported in default-server: No
8951
Emeric Brun9b3009b2012-10-05 11:55:06 +02008952no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +02008953 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +02008954 the server. Note that SSLv2 is disabled in the code and cannot be enabled
8955 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun8694b9a2012-10-05 14:39:07 +02008956 often makes sense to disable it when communicating with local servers. See
8957 also "force-sslv3", "force-tlsv*".
Emeric Brunf5da4932012-09-28 19:42:54 +02008958
Willy Tarreau763a95b2012-10-04 23:15:39 +02008959 Supported in default-server: No
8960
Emeric Brun9b3009b2012-10-05 11:55:06 +02008961no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +02008962 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +02008963 the server. Note that SSLv2 is disabled in the code and cannot be enabled
8964 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun8694b9a2012-10-05 14:39:07 +02008965 often makes sense to disable it when communicating with local servers. See
8966 also "force-sslv3", "force-tlsv*".
Emeric Brunf5da4932012-09-28 19:42:54 +02008967
Willy Tarreau763a95b2012-10-04 23:15:39 +02008968 Supported in default-server: No
8969
Emeric Brun9b3009b2012-10-05 11:55:06 +02008970no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +02008971 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008972 the server. Note that SSLv2 is disabled in the code and cannot be enabled
8973 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun8694b9a2012-10-05 14:39:07 +02008974 often makes sense to disable it when communicating with local servers. See
8975 also "force-sslv3", "force-tlsv*".
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02008976
Willy Tarreau763a95b2012-10-04 23:15:39 +02008977 Supported in default-server: No
8978
Simon Hormanfa461682011-06-25 09:39:49 +09008979non-stick
8980 Never add connections allocated to this sever to a stick-table.
8981 This may be used in conjunction with backup to ensure that
8982 stick-table persistence is disabled for backup servers.
8983
Willy Tarreau763a95b2012-10-04 23:15:39 +02008984 Supported in default-server: No
8985
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008986observe <mode>
8987 This option enables health adjusting based on observing communication with
8988 the server. By default this functionality is disabled and enabling it also
8989 requires to enable health checks. There are two supported modes: "layer4" and
8990 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
8991 significant. In layer7, which is only allowed for http proxies, responses
8992 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +01008993 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008994
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008995 Supported in default-server: No
8996
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01008997 See also the "check", "on-error" and "error-limit".
8998
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01008999on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01009000 Select what should happen when enough consecutive errors are detected.
9001 Currently, four modes are available:
9002 - fastinter: force fastinter
9003 - fail-check: simulate a failed check, also forces fastinter (default)
9004 - sudden-death: simulate a pre-fatal failed health check, one more failed
9005 check will mark a server down, forces fastinter
9006 - mark-down: mark the server immediately down and force fastinter
9007
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009008 Supported in default-server: Yes
9009
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +01009010 See also the "check", "observe" and "error-limit".
9011
Simon Hormane0d1bfb2011-06-21 14:34:58 +09009012on-marked-down <action>
9013 Modify what occurs when a server is marked down.
9014 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -07009015 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
9016 all connections to the server are immediately terminated when the server
9017 goes down. It might be used if the health check detects more complex cases
9018 than a simple connection status, and long timeouts would cause the service
9019 to remain unresponsive for too long a time. For instance, a health check
9020 might detect that a database is stuck and that there's no chance to reuse
9021 existing connections anymore. Connections killed this way are logged with
9022 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +09009023
9024 Actions are disabled by default
9025
9026 Supported in default-server: Yes
9027
Justin Karnegeseb2c24a2012-05-24 15:28:52 -07009028on-marked-up <action>
9029 Modify what occurs when a server is marked up.
9030 Currently one action is available:
9031 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
9032 done only if the server is not in backup state and if it is not disabled
9033 (it must have an effective weight > 0). This can be used sometimes to force
9034 an active server to take all the traffic back after recovery when dealing
9035 with long sessions (eg: LDAP, SQL, ...). Doing this can cause more trouble
9036 than it tries to solve (eg: incomplete transactions), so use this feature
9037 with extreme care. Sessions killed because a server comes up are logged
9038 with an 'U' termination code (for "Up").
9039
9040 Actions are disabled by default
9041
9042 Supported in default-server: Yes
9043
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009044port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009045 Using the "port" parameter, it becomes possible to use a different port to
9046 send health-checks. On some servers, it may be desirable to dedicate a port
9047 to a specific component able to perform complex tests which are more suitable
9048 to health-checks than the application. It is common to run a simple script in
9049 inetd for instance. This parameter is ignored if the "check" parameter is not
9050 set. See also the "addr" parameter.
9051
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009052 Supported in default-server: Yes
9053
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009054redir <prefix>
9055 The "redir" parameter enables the redirection mode for all GET and HEAD
9056 requests addressing this server. This means that instead of having HAProxy
9057 forward the request to the server, it will send an "HTTP 302" response with
9058 the "Location" header composed of this prefix immediately followed by the
9059 requested URI beginning at the leading '/' of the path component. That means
9060 that no trailing slash should be used after <prefix>. All invalid requests
9061 will be rejected, and all non-GET or HEAD requests will be normally served by
9062 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009063 mangling nor cookie insertion is possible in the response. However, cookies in
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009064 requests are still analysed, making this solution completely usable to direct
9065 users to a remote location in case of local disaster. Main use consists in
9066 increasing bandwidth for static servers by having the clients directly
9067 connect to them. Note: never use a relative location here, it would cause a
9068 loop between the client and HAProxy!
9069
9070 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
9071
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009072 Supported in default-server: No
9073
9074rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009075 The "rise" parameter states that a server will be considered as operational
9076 after <count> consecutive successful health checks. This value defaults to 2
9077 if unspecified. See also the "check", "inter" and "fall" parameters.
9078
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009079 Supported in default-server: Yes
9080
Willy Tarreau5ab04ec2011-03-20 10:32:26 +01009081send-proxy
9082 The "send-proxy" parameter enforces use of the PROXY protocol over any
9083 connection established to this server. The PROXY protocol informs the other
9084 end about the layer 3/4 addresses of the incoming connection, so that it can
9085 know the client's address or the public address it accessed to, whatever the
9086 upper layer protocol. For connections accepted by an "accept-proxy" listener,
9087 the advertised address will be used. Only TCPv4 and TCPv6 address families
9088 are supported. Other families such as Unix sockets, will report an UNKNOWN
9089 family. Servers using this option can fully be chained to another instance of
9090 haproxy listening with an "accept-proxy" setting. This setting must not be
Willy Tarreau6c16adc2012-10-05 00:04:16 +02009091 used if the server isn't aware of the protocol. When health checks are sent
9092 to the server, the PROXY protocol is automatically used when this option is
9093 set, unless there is an explicit "port" or "addr" directive, in which case an
9094 explicit "check-send-proxy" directive would also be needed to use the PROXY
9095 protocol. See also the "accept-proxy" option of the "bind" keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +01009096
9097 Supported in default-server: No
9098
David Safb76832014-05-08 23:42:08 -04009099send-proxy-v2
9100 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
9101 over any connection established to this server. The PROXY protocol informs
9102 the other end about the layer 3/4 addresses of the incoming connection, so
9103 that it can know the client's address or the public address it accessed to,
9104 whatever the upper layer protocol. This setting must not be used if the
9105 server isn't aware of this version of the protocol. See also the "send-proxy"
9106 option of the "bind" keyword.
9107
9108 Supported in default-server: No
9109
9110send-proxy-v2-ssl
9111 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
9112 2 over any connection established to this server. The PROXY protocol informs
9113 the other end about the layer 3/4 addresses of the incoming connection, so
9114 that it can know the client's address or the public address it accessed to,
9115 whatever the upper layer protocol. In addition, the SSL information extension
9116 of the PROXY protocol is added to the PROXY protocol header. This setting
9117 must not be used if the server isn't aware of this version of the protocol.
9118 See also the "send-proxy-v2" option of the "bind" keyword.
9119
9120 Supported in default-server: No
9121
9122send-proxy-v2-ssl-cn
9123 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
9124 2 over any connection established to this server. The PROXY protocol informs
9125 the other end about the layer 3/4 addresses of the incoming connection, so
9126 that it can know the client's address or the public address it accessed to,
9127 whatever the upper layer protocol. In addition, the SSL information extension
9128 of the PROXY protocol, along along with the Common Name from the subject of
9129 the client certificate (if any), is added to the PROXY protocol header. This
9130 setting must not be used if the server isn't aware of this version of the
9131 protocol. See also the "send-proxy-v2" option of the "bind" keyword.
9132
9133 Supported in default-server: No
9134
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009135slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009136 The "slowstart" parameter for a server accepts a value in milliseconds which
9137 indicates after how long a server which has just come back up will run at
9138 full speed. Just as with every other time-based parameter, it can be entered
9139 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
9140 linearly from 0 to 100% during this time. The limitation applies to two
9141 parameters :
9142
9143 - maxconn: the number of connections accepted by the server will grow from 1
9144 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
9145
9146 - weight: when the backend uses a dynamic weighted algorithm, the weight
9147 grows linearly from 1 to 100%. In this case, the weight is updated at every
9148 health-check. For this reason, it is important that the "inter" parameter
9149 is smaller than the "slowstart", in order to maximize the number of steps.
9150
9151 The slowstart never applies when haproxy starts, otherwise it would cause
9152 trouble to running servers. It only applies when a server has been previously
9153 seen as failed.
9154
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009155 Supported in default-server: Yes
9156
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02009157source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02009158source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02009159source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009160 The "source" parameter sets the source address which will be used when
9161 connecting to the server. It follows the exact same parameters and principle
9162 as the backend "source" keyword, except that it only applies to the server
9163 referencing it. Please consult the "source" keyword for details.
9164
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02009165 Additionally, the "source" statement on a server line allows one to specify a
9166 source port range by indicating the lower and higher bounds delimited by a
9167 dash ('-'). Some operating systems might require a valid IP address when a
9168 source port range is specified. It is permitted to have the same IP/range for
9169 several servers. Doing so makes it possible to bypass the maximum of 64k
9170 total concurrent connections. The limit will then reach 64k connections per
9171 server.
9172
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009173 Supported in default-server: No
9174
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02009175ssl
Willy Tarreau44f65392013-06-25 07:56:20 +02009176 This option enables SSL ciphering on outgoing connections to the server. It
9177 is critical to verify server certificates using "verify" when using SSL to
9178 connect to servers, otherwise the communication is prone to trivial man in
9179 the-middle attacks rendering SSL useless. When this option is used, health
9180 checks are automatically sent in SSL too unless there is a "port" or an
9181 "addr" directive indicating the check should be sent to a different location.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009182 See the "check-ssl" option to force SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +02009183
9184 Supported in default-server: No
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02009185
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009186track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +02009187 This option enables ability to set the current state of the server by tracking
9188 another one. It is possible to track a server which itself tracks another
9189 server, provided that at the end of the chain, a server has health checks
9190 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009191 used, it has to be enabled on both proxies.
9192
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009193 Supported in default-server: No
9194
Emeric Brunef42d922012-10-11 16:11:36 +02009195verify [none|required]
9196 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +01009197 to 'none', server certificate is not verified. In the other case, The
9198 certificate provided by the server is verified using CAs from 'ca-file'
9199 and optional CRLs from 'crl-file'. If 'ssl_server_verify' is not specified
9200 in global section, this is the default. On verify failure the handshake
Willy Tarreau44f65392013-06-25 07:56:20 +02009201 is aborted. It is critically important to verify server certificates when
9202 using SSL to connect to servers, otherwise the communication is prone to
9203 trivial man-in-the-middle attacks rendering SSL totally useless.
Emeric Brunef42d922012-10-11 16:11:36 +02009204
9205 Supported in default-server: No
9206
Evan Broderbe554312013-06-27 00:05:25 -07009207verifyhost <hostname>
9208 This setting is only available when support for OpenSSL was built in, and
9209 only takes effect if 'verify required' is also specified. When set, the
9210 hostnames in the subject and subjectAlternateNames of the certificate
9211 provided by the server are checked. If none of the hostnames in the
9212 certificate match the specified hostname, the handshake is aborted. The
9213 hostnames in the server-provided certificate may include wildcards.
9214
9215 Supported in default-server: No
9216
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009217weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009218 The "weight" parameter is used to adjust the server's weight relative to
9219 other servers. All servers will receive a load proportional to their weight
9220 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +02009221 load. The default weight is 1, and the maximal value is 256. A value of 0
9222 means the server will not participate in load-balancing but will still accept
9223 persistent connections. If this parameter is used to distribute the load
9224 according to server's capacity, it is recommended to start with values which
9225 can both grow and shrink, for instance between 10 and 100 to leave enough
9226 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009227
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +01009228 Supported in default-server: Yes
9229
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009230
92316. HTTP header manipulation
9232---------------------------
9233
9234In HTTP mode, it is possible to rewrite, add or delete some of the request and
9235response headers based on regular expressions. It is also possible to block a
9236request or a response if a particular header matches a regular expression,
9237which is enough to stop most elementary protocol attacks, and to protect
Willy Tarreau70dffda2014-01-30 03:07:23 +01009238against information leak from the internal network.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009239
Willy Tarreau70dffda2014-01-30 03:07:23 +01009240If HAProxy encounters an "Informational Response" (status code 1xx), it is able
9241to process all rsp* rules which can allow, deny, rewrite or delete a header,
9242but it will refuse to add a header to any such messages as this is not
9243HTTP-compliant. The reason for still processing headers in such responses is to
9244stop and/or fix any possible information leak which may happen, for instance
9245because another downstream equipment would unconditionally add a header, or if
9246a server name appears there. When such messages are seen, normal processing
9247still occurs on the next non-informational messages.
Willy Tarreau816b9792009-09-15 21:25:21 +02009248
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009249This section covers common usage of the following keywords, described in detail
9250in section 4.2 :
9251
9252 - reqadd <string>
9253 - reqallow <search>
9254 - reqiallow <search>
9255 - reqdel <search>
9256 - reqidel <search>
9257 - reqdeny <search>
9258 - reqideny <search>
9259 - reqpass <search>
9260 - reqipass <search>
9261 - reqrep <search> <replace>
9262 - reqirep <search> <replace>
9263 - reqtarpit <search>
9264 - reqitarpit <search>
9265 - rspadd <string>
9266 - rspdel <search>
9267 - rspidel <search>
9268 - rspdeny <search>
9269 - rspideny <search>
9270 - rsprep <search> <replace>
9271 - rspirep <search> <replace>
9272
9273With all these keywords, the same conventions are used. The <search> parameter
9274is a POSIX extended regular expression (regex) which supports grouping through
9275parenthesis (without the backslash). Spaces and other delimiters must be
9276prefixed with a backslash ('\') to avoid confusion with a field delimiter.
9277Other characters may be prefixed with a backslash to change their meaning :
9278
9279 \t for a tab
9280 \r for a carriage return (CR)
9281 \n for a new line (LF)
9282 \ to mark a space and differentiate it from a delimiter
9283 \# to mark a sharp and differentiate it from a comment
9284 \\ to use a backslash in a regex
9285 \\\\ to use a backslash in the text (*2 for regex, *2 for haproxy)
9286 \xXX to write the ASCII hex code XX as in the C language
9287
9288The <replace> parameter contains the string to be used to replace the largest
9289portion of text matching the regex. It can make use of the special characters
9290above, and can reference a substring which is delimited by parenthesis in the
9291regex, by writing a backslash ('\') immediately followed by one digit from 0 to
92929 indicating the group position (0 designating the entire line). This practice
9293is very common to users of the "sed" program.
9294
9295The <string> parameter represents the string which will systematically be added
9296after the last header line. It can also use special character sequences above.
9297
9298Notes related to these keywords :
9299---------------------------------
9300 - these keywords are not always convenient to allow/deny based on header
9301 contents. It is strongly recommended to use ACLs with the "block" keyword
9302 instead, resulting in far more flexible and manageable rules.
9303
9304 - lines are always considered as a whole. It is not possible to reference
9305 a header name only or a value only. This is important because of the way
9306 headers are written (notably the number of spaces after the colon).
9307
9308 - the first line is always considered as a header, which makes it possible to
9309 rewrite or filter HTTP requests URIs or response codes, but in turn makes
9310 it harder to distinguish between headers and request line. The regex prefix
9311 ^[^\ \t]*[\ \t] matches any HTTP method followed by a space, and the prefix
9312 ^[^ \t:]*: matches any header name followed by a colon.
9313
9314 - for performances reasons, the number of characters added to a request or to
9315 a response is limited at build time to values between 1 and 4 kB. This
9316 should normally be far more than enough for most usages. If it is too short
9317 on occasional usages, it is possible to gain some space by removing some
9318 useless headers before adding new ones.
9319
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009320 - keywords beginning with "reqi" and "rspi" are the same as their counterpart
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009321 without the 'i' letter except that they ignore case when matching patterns.
9322
9323 - when a request passes through a frontend then a backend, all req* rules
9324 from the frontend will be evaluated, then all req* rules from the backend
9325 will be evaluated. The reverse path is applied to responses.
9326
9327 - req* statements are applied after "block" statements, so that "block" is
9328 always the first one, but before "use_backend" in order to permit rewriting
Willy Tarreaud72758d2010-01-12 10:42:19 +01009329 before switching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009330
9331
Willy Tarreau74ca5042013-06-11 23:12:07 +020093327. Using ACLs and fetching samples
9333----------------------------------
9334
9335Haproxy is capable of extracting data from request or response streams, from
9336client or server information, from tables, environmental information etc...
9337The action of extracting such data is called fetching a sample. Once retrieved,
9338these samples may be used for various purposes such as a key to a stick-table,
9339but most common usages consist in matching them against predefined constant
9340data called patterns.
9341
9342
93437.1. ACL basics
9344---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009345
9346The use of Access Control Lists (ACL) provides a flexible solution to perform
9347content switching and generally to take decisions based on content extracted
9348from the request, the response or any environmental status. The principle is
9349simple :
9350
Willy Tarreau74ca5042013-06-11 23:12:07 +02009351 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009352 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +02009353 - apply one or multiple pattern matching methods on this sample
9354 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009355
Willy Tarreau74ca5042013-06-11 23:12:07 +02009356The actions generally consist in blocking a request, selecting a backend, or
9357adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009358
9359In order to define a test, the "acl" keyword is used. The syntax is :
9360
Willy Tarreau74ca5042013-06-11 23:12:07 +02009361 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009362
9363This creates a new ACL <aclname> or completes an existing one with new tests.
9364Those tests apply to the portion of request/response specified in <criterion>
9365and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009366an operator which may be specified before the set of values. Optionally some
9367conversion operators may be applied to the sample, and they will be specified
9368as a comma-delimited list of keywords just after the first keyword. The values
9369are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009370
9371ACL names must be formed from upper and lower case letters, digits, '-' (dash),
9372'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
9373which means that "my_acl" and "My_Acl" are two different ACLs.
9374
9375There is no enforced limit to the number of ACLs. The unused ones do not affect
9376performance, they just consume a small amount of memory.
9377
Willy Tarreau74ca5042013-06-11 23:12:07 +02009378The criterion generally is the name of a sample fetch method, or one of its ACL
9379specific declinations. The default test method is implied by the output type of
9380this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009381methods of a same sample fetch method. The sample fetch methods are the only
9382ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +02009383
9384Sample fetch methods return data which can be of the following types :
9385 - boolean
9386 - integer (signed or unsigned)
9387 - IPv4 or IPv6 address
9388 - string
9389 - data block
9390
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009391Converters transform any of these data into any of these. For example, some
9392converters might convert a string to a lower-case string while other ones
9393would turn a string to an IPv4 address, or apply a netmask to an IP address.
9394The resulting sample is of the type of the last converter applied to the list,
9395which defaults to the type of the sample fetch method.
9396
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009397Each sample or converter returns data of a specific type, specified with its
9398keyword in this documentation. When an ACL is declared using a standard sample
9399fetch method, certain types automatically involved a default matching method
9400which are summarized in the table below :
9401
9402 +---------------------+-----------------+
9403 | Sample or converter | Default |
9404 | output type | matching method |
9405 +---------------------+-----------------+
9406 | boolean | bool |
9407 +---------------------+-----------------+
9408 | integer | int |
9409 +---------------------+-----------------+
9410 | ip | ip |
9411 +---------------------+-----------------+
9412 | string | str |
9413 +---------------------+-----------------+
9414 | binary | none, use "-m" |
9415 +---------------------+-----------------+
9416
9417Note that in order to match a binary samples, it is mandatory to specify a
9418matching method, see below.
9419
Willy Tarreau74ca5042013-06-11 23:12:07 +02009420The ACL engine can match these types against patterns of the following types :
9421 - boolean
9422 - integer or integer range
9423 - IP address / network
9424 - string (exact, substring, suffix, prefix, subdir, domain)
9425 - regular expression
9426 - hex block
9427
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009428The following ACL flags are currently supported :
9429
Willy Tarreau2b5285d2010-05-09 23:45:24 +02009430 -i : ignore case during matching of all subsequent patterns.
9431 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +02009432 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +01009433 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +01009434 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +01009435 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +02009436 -- : force end of flags. Useful when a string looks like one of the flags.
9437
Willy Tarreau74ca5042013-06-11 23:12:07 +02009438The "-f" flag is followed by the name of a file from which all lines will be
9439read as individual values. It is even possible to pass multiple "-f" arguments
9440if the patterns are to be loaded from multiple files. Empty lines as well as
9441lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
9442will be stripped. If it is absolutely necessary to insert a valid pattern
9443beginning with a sharp, just prefix it with a space so that it is not taken for
9444a comment. Depending on the data type and match method, haproxy may load the
9445lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
9446exact string matching. In this case, duplicates will automatically be removed.
9447
Thierry FOURNIER9860c412014-01-29 14:23:29 +01009448The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
9449parsed as two column file. The first column contains the patterns used by the
9450ACL, and the second column contain the samples. The sample can be used later by
9451a map. This can be useful in some rare cases where an ACL would just be used to
9452check for the existence of a pattern in a map before a mapping is applied.
9453
Thierry FOURNIER3534d882014-01-20 17:01:44 +01009454The "-u" flag forces the unique id of the ACL. This unique id is used with the
9455socket interface to identify ACL and dynamically change its values. Note that a
9456file is always identified by its name even if an id is set.
9457
Willy Tarreau74ca5042013-06-11 23:12:07 +02009458Also, note that the "-i" flag applies to subsequent entries and not to entries
9459loaded from files preceding it. For instance :
9460
9461 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
9462
9463In this example, each line of "exact-ua.lst" will be exactly matched against
9464the "user-agent" header of the request. Then each line of "generic-ua" will be
9465case-insensitively matched. Then the word "test" will be insensitively matched
9466as well.
9467
9468The "-m" flag is used to select a specific pattern matching method on the input
9469sample. All ACL-specific criteria imply a pattern matching method and generally
9470do not need this flag. However, this flag is useful with generic sample fetch
9471methods to describe how they're going to be matched against the patterns. This
9472is required for sample fetches which return data type for which there is no
9473obvious matching method (eg: string or binary). When "-m" is specified and
9474followed by a pattern matching method name, this method is used instead of the
9475default one for the criterion. This makes it possible to match contents in ways
9476that were not initially planned, or with sample fetch methods which return a
9477string. The matching method also affects the way the patterns are parsed.
9478
Thierry FOURNIERb7729c92014-02-11 16:24:41 +01009479The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
9480By default, if the parser cannot parse ip address it considers that the parsed
9481string is maybe a domain name and try dns resolution. The flag "-n" disable this
9482resolution. It is useful for detecting malformed ip lists. Note that if the DNS
9483server is not reachable, the haproxy configuration parsing may last many minutes
9484waiting fir the timeout. During this time no error messages are displayed. The
9485flag "-n" disable this behavior. Note also that during the runtime, this
9486function is disabled for the dynamic acl modifications.
9487
Willy Tarreau74ca5042013-06-11 23:12:07 +02009488There are some restrictions however. Not all methods can be used with all
9489sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
9490be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +02009491
9492 - "found" : only check if the requested sample could be found in the stream,
9493 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +02009494 to pass any pattern to avoid confusion. This matching method is
9495 particularly useful to detect presence of certain contents such
9496 as headers, cookies, etc... even if they are empty and without
9497 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009498
9499 - "bool" : check the value as a boolean. It can only be applied to fetches
9500 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +02009501 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009502
9503 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +02009504 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009505
9506 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +02009507 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009508
9509 - "bin" : match the contents against an hexadecimal string representing a
9510 binary sequence. This may be used with binary or string samples.
9511
9512 - "len" : match the sample's length as an integer. This may be used with
9513 binary or string samples.
9514
Willy Tarreau74ca5042013-06-11 23:12:07 +02009515 - "str" : exact match : match the contents against a string. This may be
9516 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009517
Willy Tarreau74ca5042013-06-11 23:12:07 +02009518 - "sub" : substring match : check that the contents contain at least one of
9519 the provided string patterns. This may be used with binary or
9520 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009521
Willy Tarreau74ca5042013-06-11 23:12:07 +02009522 - "reg" : regex match : match the contents against a list of regular
9523 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009524
Willy Tarreau74ca5042013-06-11 23:12:07 +02009525 - "beg" : prefix match : check that the contents begin like the provided
9526 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009527
Willy Tarreau74ca5042013-06-11 23:12:07 +02009528 - "end" : suffix match : check that the contents end like the provided
9529 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009530
Willy Tarreau74ca5042013-06-11 23:12:07 +02009531 - "dir" : subdir match : check that a slash-delimited portion of the
9532 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009533 This may be used with binary or string samples.
9534
Willy Tarreau74ca5042013-06-11 23:12:07 +02009535 - "dom" : domain match : check that a dot-delimited portion of the contents
9536 exactly match one of the provided string patterns. This may be
9537 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +02009538
9539For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
9540request, it is possible to do :
9541
9542 acl jsess_present cook(JSESSIONID) -m found
9543
9544In order to apply a regular expression on the 500 first bytes of data in the
9545buffer, one would use the following acl :
9546
9547 acl script_tag payload(0,500) -m reg -i <script>
9548
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009549On systems where the regex library is much slower when using "-i", it is
9550possible to convert the sample to lowercase before matching, like this :
9551
9552 acl script_tag payload(0,500),lower -m reg <script>
9553
Willy Tarreau74ca5042013-06-11 23:12:07 +02009554All ACL-specific criteria imply a default matching method. Most often, these
9555criteria are composed by concatenating the name of the original sample fetch
9556method and the matching method. For example, "hdr_beg" applies the "beg" match
9557to samples retrieved using the "hdr" fetch method. Since all ACL-specific
9558criteria rely on a sample fetch method, it is always possible instead to use
9559the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +02009560
Willy Tarreau74ca5042013-06-11 23:12:07 +02009561If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009562the matching method is simply applied to the underlying sample fetch method.
9563For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +02009564
Willy Tarreau74ca5042013-06-11 23:12:07 +02009565 acl short_form hdr_beg(host) www.
9566 acl alternate1 hdr_beg(host) -m beg www.
9567 acl alternate2 hdr_dom(host) -m beg www.
9568 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +02009569
Willy Tarreau2b5285d2010-05-09 23:45:24 +02009570
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009571The table below summarizes the compatibility matrix between sample or converter
9572types and the pattern types to fetch against. It indicates for each compatible
9573combination the name of the matching method to be used, surrounded with angle
9574brackets ">" and "<" when the method is the default one and will work by
9575default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +01009576
Willy Tarreau74ca5042013-06-11 23:12:07 +02009577 +-------------------------------------------------+
9578 | Input sample type |
9579 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009580 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009581 +----------------------+---------+---------+---------+---------+---------+
9582 | none (presence only) | found | found | found | found | found |
9583 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009584 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009585 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009586 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009587 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009588 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009589 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009590 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009591 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +02009592 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009593 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009594 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009595 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009596 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009597 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009598 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009599 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009600 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009601 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009602 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009603 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +01009604 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +02009605 +----------------------+---------+---------+---------+---------+---------+
9606 | hex block | | | | bin | bin |
9607 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +02009608
9609
Willy Tarreau74ca5042013-06-11 23:12:07 +020096107.1.1. Matching booleans
9611------------------------
9612
9613In order to match a boolean, no value is needed and all values are ignored.
9614Boolean matching is used by default for all fetch methods of type "boolean".
9615When boolean matching is used, the fetched value is returned as-is, which means
9616that a boolean "true" will always match and a boolean "false" will never match.
9617
9618Boolean matching may also be enforced using "-m bool" on fetch methods which
9619return an integer value. Then, integer value 0 is converted to the boolean
9620"false" and all other values are converted to "true".
9621
Willy Tarreau6a06a402007-07-15 20:15:28 +02009622
Willy Tarreau74ca5042013-06-11 23:12:07 +020096237.1.2. Matching integers
9624------------------------
9625
9626Integer matching applies by default to integer fetch methods. It can also be
9627enforced on boolean fetches using "-m int". In this case, "false" is converted
9628to the integer 0, and "true" is converted to the integer 1.
9629
9630Integer matching also supports integer ranges and operators. Note that integer
9631matching only applies to positive values. A range is a value expressed with a
9632lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +02009633
9634For instance, "1024:65535" is a valid range to represent a range of
9635unprivileged ports, and "1024:" would also work. "0:1023" is a valid
9636representation of privileged ports, and ":1023" would also work.
9637
Willy Tarreau62644772008-07-16 18:36:06 +02009638As a special case, some ACL functions support decimal numbers which are in fact
9639two integers separated by a dot. This is used with some version checks for
9640instance. All integer properties apply to those decimal numbers, including
9641ranges and operators.
9642
Willy Tarreau6a06a402007-07-15 20:15:28 +02009643For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +01009644operators with ranges does not make much sense and is strongly discouraged.
9645Similarly, it does not make much sense to perform order comparisons with a set
9646of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +02009647
Willy Tarreau0ba27502007-12-24 16:55:16 +01009648Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +02009649
9650 eq : true if the tested value equals at least one value
9651 ge : true if the tested value is greater than or equal to at least one value
9652 gt : true if the tested value is greater than at least one value
9653 le : true if the tested value is less than or equal to at least one value
9654 lt : true if the tested value is less than at least one value
9655
Willy Tarreau0ba27502007-12-24 16:55:16 +01009656For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +02009657
9658 acl negative-length hdr_val(content-length) lt 0
9659
Willy Tarreau62644772008-07-16 18:36:06 +02009660This one matches SSL versions between 3.0 and 3.1 (inclusive) :
9661
9662 acl sslv3 req_ssl_ver 3:3.1
9663
Willy Tarreau6a06a402007-07-15 20:15:28 +02009664
Willy Tarreau74ca5042013-06-11 23:12:07 +020096657.1.3. Matching strings
9666-----------------------
9667
9668String matching applies to string or binary fetch methods, and exists in 6
9669different forms :
9670
9671 - exact match (-m str) : the extracted string must exactly match the
9672 patterns ;
9673
9674 - substring match (-m sub) : the patterns are looked up inside the
9675 extracted string, and the ACL matches if any of them is found inside ;
9676
9677 - prefix match (-m beg) : the patterns are compared with the beginning of
9678 the extracted string, and the ACL matches if any of them matches.
9679
9680 - suffix match (-m end) : the patterns are compared with the end of the
9681 extracted string, and the ACL matches if any of them matches.
9682
9683 - subdir match (-m sub) : the patterns are looked up inside the extracted
9684 string, delimited with slashes ("/"), and the ACL matches if any of them
9685 matches.
9686
9687 - domain match (-m dom) : the patterns are looked up inside the extracted
9688 string, delimited with dots ("."), and the ACL matches if any of them
9689 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +02009690
9691String matching applies to verbatim strings as they are passed, with the
9692exception of the backslash ("\") which makes it possible to escape some
9693characters such as the space. If the "-i" flag is passed before the first
9694string, then the matching will be performed ignoring the case. In order
9695to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +01009696before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +02009697
9698
Willy Tarreau74ca5042013-06-11 23:12:07 +020096997.1.4. Matching regular expressions (regexes)
9700---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02009701
9702Just like with string matching, regex matching applies to verbatim strings as
9703they are passed, with the exception of the backslash ("\") which makes it
9704possible to escape some characters such as the space. If the "-i" flag is
9705passed before the first regex, then the matching will be performed ignoring
9706the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +01009707the "--" flag before the first string. Same principle applies of course to
9708match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +02009709
9710
Willy Tarreau74ca5042013-06-11 23:12:07 +020097117.1.5. Matching arbitrary data blocks
9712-------------------------------------
9713
9714It is possible to match some extracted samples against a binary block which may
9715not safely be represented as a string. For this, the patterns must be passed as
9716a series of hexadecimal digits in an even number, when the match method is set
9717to binary. Each sequence of two digits will represent a byte. The hexadecimal
9718digits may be used upper or lower case.
9719
9720Example :
9721 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
9722 acl hello payload(0,6) -m bin 48656c6c6f0a
9723
9724
97257.1.6. Matching IPv4 and IPv6 addresses
9726---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02009727
9728IPv4 addresses values can be specified either as plain addresses or with a
9729netmask appended, in which case the IPv4 address matches whenever it is
9730within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01009731host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +01009732difficult to read and debug configurations. If hostnames are used, you should
9733at least ensure that they are present in /etc/hosts so that the configuration
9734does not depend on any random DNS match at the moment the configuration is
9735parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +02009736
Willy Tarreauceb4ac92012-04-28 00:41:46 +02009737IPv6 may be entered in their usual form, with or without a netmask appended.
9738Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
9739trouble with randomly resolved IP addresses, host names are never allowed in
9740IPv6 patterns.
9741
9742HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
9743following situations :
9744 - tested address is IPv4, pattern address is IPv4, the match applies
9745 in IPv4 using the supplied mask if any.
9746 - tested address is IPv6, pattern address is IPv6, the match applies
9747 in IPv6 using the supplied mask if any.
9748 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
9749 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
9750 ::IPV4 or ::ffff:IPV4, otherwise it fails.
9751 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
9752 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
9753 applied in IPv6 using the supplied IPv6 mask.
9754
Willy Tarreau74ca5042013-06-11 23:12:07 +02009755
97567.2. Using ACLs to form conditions
9757----------------------------------
9758
9759Some actions are only performed upon a valid condition. A condition is a
9760combination of ACLs with operators. 3 operators are supported :
9761
9762 - AND (implicit)
9763 - OR (explicit with the "or" keyword or the "||" operator)
9764 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +02009765
Willy Tarreau74ca5042013-06-11 23:12:07 +02009766A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +02009767
Willy Tarreau74ca5042013-06-11 23:12:07 +02009768 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +02009769
Willy Tarreau74ca5042013-06-11 23:12:07 +02009770Such conditions are generally used after an "if" or "unless" statement,
9771indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +02009772
Willy Tarreau74ca5042013-06-11 23:12:07 +02009773For instance, to block HTTP requests to the "*" URL with methods other than
9774"OPTIONS", as well as POST requests without content-length, and GET or HEAD
9775requests with a content-length greater than 0, and finally every request which
9776is not either GET/HEAD/POST/OPTIONS !
9777
9778 acl missing_cl hdr_cnt(Content-length) eq 0
9779 block if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
9780 block if METH_GET HTTP_CONTENT
9781 block unless METH_GET or METH_POST or METH_OPTIONS
9782
9783To select a different backend for requests to static contents on the "www" site
9784and to every request on the "img", "video", "download" and "ftp" hosts :
9785
9786 acl url_static path_beg /static /images /img /css
9787 acl url_static path_end .gif .png .jpg .css .js
9788 acl host_www hdr_beg(host) -i www
9789 acl host_static hdr_beg(host) -i img. video. download. ftp.
9790
9791 # now use backend "static" for all static-only hosts, and for static urls
9792 # of host "www". Use backend "www" for the rest.
9793 use_backend static if host_static or host_www url_static
9794 use_backend www if host_www
9795
9796It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
9797expressions that are built on the fly without needing to be declared. They must
9798be enclosed between braces, with a space before and after each brace (because
9799the braces must be seen as independent words). Example :
9800
9801 The following rule :
9802
9803 acl missing_cl hdr_cnt(Content-length) eq 0
9804 block if METH_POST missing_cl
9805
9806 Can also be written that way :
9807
9808 block if METH_POST { hdr_cnt(Content-length) eq 0 }
9809
9810It is generally not recommended to use this construct because it's a lot easier
9811to leave errors in the configuration when written that way. However, for very
9812simple rules matching only one source IP address for instance, it can make more
9813sense to use them than to declare ACLs with random names. Another example of
9814good use is the following :
9815
9816 With named ACLs :
9817
9818 acl site_dead nbsrv(dynamic) lt 2
9819 acl site_dead nbsrv(static) lt 2
9820 monitor fail if site_dead
9821
9822 With anonymous ACLs :
9823
9824 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
9825
9826See section 4.2 for detailed help on the "block" and "use_backend" keywords.
9827
9828
98297.3. Fetching samples
9830---------------------
9831
9832Historically, sample fetch methods were only used to retrieve data to match
9833against patterns using ACLs. With the arrival of stick-tables, a new class of
9834sample fetch methods was created, most often sharing the same syntax as their
9835ACL counterpart. These sample fetch methods are also known as "fetches". As
9836of now, ACLs and fetches have converged. All ACL fetch methods have been made
9837available as fetch methods, and ACLs may use any sample fetch method as well.
9838
9839This section details all available sample fetch methods and their output type.
9840Some sample fetch methods have deprecated aliases that are used to maintain
9841compatibility with existing configurations. They are then explicitly marked as
9842deprecated and should not be used in new setups.
9843
9844The ACL derivatives are also indicated when available, with their respective
9845matching methods. These ones all have a well defined default pattern matching
9846method, so it is never necessary (though allowed) to pass the "-m" option to
9847indicate how the sample will be matched using ACLs.
9848
9849As indicated in the sample type versus matching compatibility matrix above,
9850when using a generic sample fetch method in an ACL, the "-m" option is
9851mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
9852the same keyword exists as an ACL keyword and as a standard fetch method, the
9853ACL engine will automatically pick the ACL-only one by default.
9854
9855Some of these keywords support one or multiple mandatory arguments, and one or
9856multiple optional arguments. These arguments are strongly typed and are checked
9857when the configuration is parsed so that there is no risk of running with an
9858incorrect argument (eg: an unresolved backend name). Fetch function arguments
9859are passed between parenthesis and are delimited by commas. When an argument
9860is optional, it will be indicated below between square brackets ('[ ]'). When
9861all arguments are optional, the parenthesis may be omitted.
9862
9863Thus, the syntax of a standard sample fetch method is one of the following :
9864 - name
9865 - name(arg1)
9866 - name(arg1,arg2)
9867
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009868
98697.3.1. Converters
9870-----------------
9871
Willy Tarreaue6b11e42013-11-26 19:02:32 +01009872Sample fetch methods may be combined with transformations to be applied on top
9873of the fetched sample (also called "converters"). These combinations form what
9874is called "sample expressions" and the result is a "sample". Initially this
9875was only supported by "stick on" and "stick store-request" directives but this
9876has now be extended to all places where samples may be used (acls, log-format,
9877unique-id-format, add-header, ...).
9878
9879These transformations are enumerated as a series of specific keywords after the
9880sample fetch method. These keywords may equally be appended immediately after
9881the fetch keyword's argument, delimited by a comma. These keywords can also
9882support some arguments (eg: a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +01009883
Willy Tarreau74ca5042013-06-11 23:12:07 +02009884The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +01009885
Emeric Brun53d1a982014-04-30 18:21:37 +02009886base64
9887 Converts a binary input sample to a base64 string. It is used to log or
9888 transfer binary content in a way that can be reliably transferred (eg:
9889 an SSL ID can be copied in a header).
9890
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009891hex
9892 Converts a binary input sample to an hex string containing two hex digits per
9893 input byte. It is used to log or transfer hex dumps of some binary input data
9894 in a way that can be reliably transferred (eg: an SSL ID can be copied in a
9895 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +01009896
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009897http_date([<offset>])
9898 Converts an integer supposed to contain a date since epoch to a string
9899 representing this date in a format suitable for use in HTTP header fields. If
9900 an offset value is specified, then it is a number of seconds that is added to
9901 the date before the conversion is operated. This is particularly useful to
9902 emit Date header fields, Expires values in responses when combined with a
9903 positive offset, or Last-Modified values when the offset is negative.
Willy Tarreau74ca5042013-06-11 23:12:07 +02009904
Willy Tarreau644d9ef2014-07-10 16:29:08 +02009905ipmask(<mask>)
9906 Apply a mask to an IPv4 address, and use the result for lookups and storage.
9907 This can be used to make all hosts within a certain mask to share the same
9908 table entries and as such use the same server. The mask can be passed in
9909 dotted form (eg: 255.255.255.0) or in CIDR form (eg: 24).
9910
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009911language(<value>[,<default>])
9912 Returns the value with the highest q-factor from a list as extracted from the
9913 "accept-language" header using "req.fhdr". Values with no q-factor have a
9914 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
9915 belong to the list of semi-colon delimited <values> will be considered. The
9916 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
9917 given list and a default value is provided, it is returned. Note that language
9918 names may have a variant after a dash ('-'). If this variant is present in the
9919 list, it will be matched, but if it is not, only the base language is checked.
9920 The match is case-sensitive, and the output string is always one of those
9921 provided in arguments. The ordering of arguments is meaningless, only the
9922 ordering of the values in the request counts, as the first value among
9923 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +02009924
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009925 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +02009926
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009927 # this configuration switches to the backend matching a
9928 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +02009929
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009930 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
9931 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
9932 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
9933 use_backend spanish if es
9934 use_backend french if fr
9935 use_backend english if en
9936 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +02009937
Willy Tarreau644d9ef2014-07-10 16:29:08 +02009938lower
9939 Convert a string sample to lower case. This can only be placed after a string
9940 sample fetch function or after a transformation keyword returning a string
9941 type. The result is of type string.
9942
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009943map(<map_file>[,<default_value>])
9944map_<match_type>(<map_file>[,<default_value>])
9945map_<match_type>_<output_type>(<map_file>[,<default_value>])
9946 Search the input value from <map_file> using the <match_type> matching method,
9947 and return the associated value converted to the type <output_type>. If the
9948 input value cannot be found in the <map_file>, the converter returns the
9949 <default_value>. If the <default_value> is not set, the converter fails and
9950 acts as if no input value could be fetched. If the <match_type> is not set, it
9951 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
9952 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
9953 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +01009954
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009955 It is important to avoid overlapping between the keys : IP addresses and
9956 strings are stored in trees, so the first of the finest match will be used.
9957 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +01009958
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009959 The following array contains the list of all map functions avalaible sorted by
9960 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +01009961
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009962 input type | match method | output type str | output type int | output type ip
9963 -----------+--------------+-----------------+-----------------+---------------
9964 str | str | map_str | map_str_int | map_str_ip
9965 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +02009966 str | beg | map_beg | map_beg_int | map_end_ip
9967 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009968 str | sub | map_sub | map_sub_int | map_sub_ip
9969 -----------+--------------+-----------------+-----------------+---------------
9970 str | dir | map_dir | map_dir_int | map_dir_ip
9971 -----------+--------------+-----------------+-----------------+---------------
9972 str | dom | map_dom | map_dom_int | map_dom_ip
9973 -----------+--------------+-----------------+-----------------+---------------
9974 str | end | map_end | map_end_int | map_end_ip
9975 -----------+--------------+-----------------+-----------------+---------------
9976 str | reg | map_reg | map_reg_int | map_reg_ip
9977 -----------+--------------+-----------------+-----------------+---------------
9978 int | int | map_int | map_int_int | map_int_ip
9979 -----------+--------------+-----------------+-----------------+---------------
9980 ip | ip | map_ip | map_ip_int | map_ip_ip
9981 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +01009982
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009983 The file contains one key + value per line. Lines which start with '#' are
9984 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
9985 is then the first "word" (series of non-space/tabs characters), and the value
9986 is what follows this series of space/tab till the end of the line excluding
9987 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +01009988
Thierry FOURNIER060762e2014-04-23 13:29:15 +02009989 Example :
9990
9991 # this is a comment and is ignored
9992 2.22.246.0/23 United Kingdom \n
9993 <-><-----------><--><------------><---->
9994 | | | | `- trailing spaces ignored
9995 | | | `---------- value
9996 | | `-------------------- middle spaces ignored
9997 | `---------------------------- key
9998 `------------------------------------ leading spaces ignored
9999
Willy Tarreau644d9ef2014-07-10 16:29:08 +020010000upper
10001 Convert a string sample to upper case. This can only be placed after a string
10002 sample fetch function or after a transformation keyword returning a string
10003 type. The result is of type string.
10004
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010010005
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200100067.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020010007--------------------------------------------
10008
10009A first set of sample fetch methods applies to internal information which does
10010not even relate to any client information. These ones are sometimes used with
10011"monitor-fail" directives to report an internal status to external watchers.
10012The sample fetch methods described in this section are usable anywhere.
10013
10014always_false : boolean
10015 Always returns the boolean "false" value. It may be used with ACLs as a
10016 temporary replacement for another one when adjusting configurations.
10017
10018always_true : boolean
10019 Always returns the boolean "true" value. It may be used with ACLs as a
10020 temporary replacement for another one when adjusting configurations.
10021
10022avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010010023 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020010024 divided by the number of active servers. The current backend is used if no
10025 backend is specified. This is very similar to "queue" except that the size of
10026 the farm is considered, in order to give a more accurate measurement of the
10027 time it may take for a new connection to be processed. The main usage is with
10028 ACL to return a sorry page to new users when it becomes certain they will get
10029 a degraded service, or to pass to the backend servers in a header so that
10030 they decide to work in degraded mode or to disable some functions to speed up
10031 the processing a bit. Note that in the event there would not be any active
10032 server anymore, twice the number of queued connections would be considered as
10033 the measured value. This is a fair estimate, as we expect one server to get
10034 back soon anyway, but we still prefer to send new traffic to another backend
10035 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
10036 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010010037
Willy Tarreau74ca5042013-06-11 23:12:07 +020010038be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020010039 Applies to the number of currently established connections on the backend,
10040 possibly including the connection being evaluated. If no backend name is
10041 specified, the current one is used. But it is also possible to check another
10042 backend. It can be used to use a specific farm when the nominal one is full.
10043 See also the "fe_conn", "queue" and "be_sess_rate" criteria.
Willy Tarreau6a06a402007-07-15 20:15:28 +020010044
Willy Tarreau74ca5042013-06-11 23:12:07 +020010045be_sess_rate([<backend>]) : integer
10046 Returns an integer value corresponding to the sessions creation rate on the
10047 backend, in number of new sessions per second. This is used with ACLs to
10048 switch to an alternate backend when an expensive or fragile one reaches too
10049 high a session rate, or to limit abuse of service (eg. prevent sucking of an
10050 online dictionary). It can also be useful to add this element to logs using a
10051 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010052
10053 Example :
10054 # Redirect to an error page if the dictionary is requested too often
10055 backend dynamic
10056 mode http
10057 acl being_scanned be_sess_rate gt 100
10058 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010010059
Willy Tarreau74ca5042013-06-11 23:12:07 +020010060connslots([<backend>]) : integer
10061 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010062 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020010063 connections on all servers and the maximum queue size. This is probably only
10064 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050010065
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080010066 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020010067 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080010068 usage; see "use_backend" keyword) can be redirected to a different backend.
10069
Willy Tarreau55165fe2009-05-10 12:02:55 +020010070 'connslots' = number of available server connection slots, + number of
10071 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080010072
Willy Tarreaua36af912009-10-10 12:02:45 +020010073 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020010074 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020010075 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020010076 you want to be able to differentiate between different backends, and their
10077 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020010078 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020010079 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080010080
Willy Tarreau55165fe2009-05-10 12:02:55 +020010081 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
10082 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020010083 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020010084 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080010085
Willy Tarreau6236d3a2013-07-25 14:28:25 +020010086date([<offset>]) : integer
10087 Returns the current date as the epoch (number of seconds since 01/01/1970).
10088 If an offset value is specified, then it is a number of seconds that is added
10089 to the current date before returning the value. This is particularly useful
10090 to compute relative dates, as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020010091 It is useful combined with the http_date converter.
10092
10093 Example :
10094
10095 # set an expires header to now+1 hour in every response
10096 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020010097
Willy Tarreau595ec542013-06-12 21:34:28 +020010098env(<name>) : string
10099 Returns a string containing the value of environment variable <name>. As a
10100 reminder, environment variables are per-process and are sampled when the
10101 process starts. This can be useful to pass some information to a next hop
10102 server, or with ACLs to take specific action when the process is started a
10103 certain way.
10104
10105 Examples :
10106 # Pass the Via header to next hop with the local hostname in it
10107 http-request add-header Via 1.1\ %[env(HOSTNAME)]
10108
10109 # reject cookie-less requests when the STOP environment variable is set
10110 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
10111
Willy Tarreau74ca5042013-06-11 23:12:07 +020010112fe_conn([<frontend>]) : integer
10113 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010010114 possibly including the connection being evaluated. If no frontend name is
10115 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020010116 frontend. It can be used to return a sorry page before hard-blocking, or to
10117 use a specific backend to drain new requests when the farm is considered
10118 full. This is mostly used with ACLs but can also be used to pass some
10119 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
10120 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020010121
Willy Tarreau74ca5042013-06-11 23:12:07 +020010122fe_sess_rate([<frontend>]) : integer
10123 Returns an integer value corresponding to the sessions creation rate on the
10124 frontend, in number of new sessions per second. This is used with ACLs to
10125 limit the incoming session rate to an acceptable range in order to prevent
10126 abuse of service at the earliest moment, for example when combined with other
10127 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
10128 down below the limit. It can also be useful to add this element to logs using
10129 a log-format directive. See also the "rate-limit sessions" directive for use
10130 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010010131
10132 Example :
10133 # This frontend limits incoming mails to 10/s with a max of 100
10134 # concurrent connections. We accept any connection below 10/s, and
10135 # force excess clients to wait for 100 ms. Since clients are limited to
10136 # 100 max, there cannot be more than 10 incoming mails per second.
10137 frontend mail
10138 bind :25
10139 mode tcp
10140 maxconn 100
10141 acl too_fast fe_sess_rate ge 10
10142 tcp-request inspect-delay 100ms
10143 tcp-request content accept if ! too_fast
10144 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010010145
Willy Tarreau74ca5042013-06-11 23:12:07 +020010146nbsrv([<backend>]) : integer
10147 Returns an integer value corresponding to the number of usable servers of
10148 either the current backend or the named backend. This is mostly used with
10149 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010010150 switch to an alternate backend when the number of servers is too low to
10151 to handle some load. It is useful to report a failure when combined with
10152 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010010153
Willy Tarreau74ca5042013-06-11 23:12:07 +020010154queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010010155 Returns the total number of queued connections of the designated backend,
10156 including all the connections in server queues. If no backend name is
10157 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020010158 one. This is useful with ACLs or to pass statistics to backend servers. This
10159 can be used to take actions when queuing goes above a known level, generally
10160 indicating a surge of traffic or a massive slowdown on the servers. One
10161 possible action could be to reject new users but still accept old ones. See
10162 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
10163
Willy Tarreau84310e22014-02-14 11:59:04 +010010164rand([<range>]) : integer
10165 Returns a random integer value within a range of <range> possible values,
10166 starting at zero. If the range is not specified, it defaults to 2^32, which
10167 gives numbers between 0 and 4294967295. It can be useful to pass some values
10168 needed to take some routing decisions for example, or just for debugging
10169 purposes. This random must not be used for security purposes.
10170
Willy Tarreau74ca5042013-06-11 23:12:07 +020010171srv_conn([<backend>/]<server>) : integer
10172 Returns an integer value corresponding to the number of currently established
10173 connections on the designated server, possibly including the connection being
10174 evaluated. If <backend> is omitted, then the server is looked up in the
10175 current backend. It can be used to use a specific farm when one server is
10176 full, or to inform the server about our view of the number of active
10177 connections with it. See also the "fe_conn", "be_conn" and "queue" fetch
10178 methods.
10179
10180srv_is_up([<backend>/]<server>) : boolean
10181 Returns true when the designated server is UP, and false when it is either
10182 DOWN or in maintenance mode. If <backend> is omitted, then the server is
10183 looked up in the current backend. It is mainly used to take action based on
10184 an external status reported via a health check (eg: a geographical site's
10185 availability). Another possible use which is more of a hack consists in
10186 using dummy servers as boolean variables that can be enabled or disabled from
10187 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
10188
10189srv_sess_rate([<backend>/]<server>) : integer
10190 Returns an integer corresponding to the sessions creation rate on the
10191 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010192 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020010193 used with ACLs but can make sense with logs too. This is used to switch to an
10194 alternate backend when an expensive or fragile one reaches too high a session
10195 rate, or to limit abuse of service (eg. prevent latent requests from
10196 overloading servers).
10197
10198 Example :
10199 # Redirect to a separate back
10200 acl srv1_full srv_sess_rate(be1/srv1) gt 50
10201 acl srv2_full srv_sess_rate(be1/srv2) gt 50
10202 use_backend be2 if srv1_full or srv2_full
10203
10204table_avl([<table>]) : integer
10205 Returns the total number of available entries in the current proxy's
10206 stick-table or in the designated stick-table. See also table_cnt.
10207
10208table_cnt([<table>]) : integer
10209 Returns the total number of entries currently in use in the current proxy's
10210 stick-table or in the designated stick-table. See also src_conn_cnt and
10211 table_avl for other entry counting methods.
10212
10213
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200102147.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020010215----------------------------------
10216
10217The layer 4 usually describes just the transport layer which in haproxy is
10218closest to the connection, where no content is yet made available. The fetch
10219methods described here are usable as low as the "tcp-request connection" rule
10220sets unless they require some future information. Those generally include
10221TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010222the incoming connection. For retrieving a value from a sticky counters, the
10223counter number can be explicitly set as 0, 1, or 2 using the pre-defined
10224"sc0_", "sc1_", or "sc2_" prefix, or it can be specified as the first integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010225argument when using the "sc_" prefix. An optional table may be specified with
10226the "sc*" form, in which case the currently tracked key will be looked up into
10227this alternate table instead of the table currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020010228
10229be_id : integer
10230 Returns an integer containing the current backend's id. It can be used in
10231 frontends with responses to check which backend processed the request.
10232
10233dst : ip
10234 This is the destination IPv4 address of the connection on the client side,
10235 which is the address the client connected to. It can be useful when running
10236 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
10237 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
10238 RFC 4291.
10239
10240dst_conn : integer
10241 Returns an integer value corresponding to the number of currently established
10242 connections on the same socket including the one being evaluated. It is
10243 normally used with ACLs but can as well be used to pass the information to
10244 servers in an HTTP header or in logs. It can be used to either return a sorry
10245 page before hard-blocking, or to use a specific backend to drain new requests
10246 when the socket is considered saturated. This offers the ability to assign
10247 different limits to different listening ports or addresses. See also the
10248 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010249
Willy Tarreau74ca5042013-06-11 23:12:07 +020010250dst_port : integer
10251 Returns an integer value corresponding to the destination TCP port of the
10252 connection on the client side, which is the port the client connected to.
10253 This might be used when running in transparent mode, when assigning dynamic
10254 ports to some clients for a whole application session, to stick all users to
10255 a same server, or to pass the destination port information to a server using
10256 an HTTP header.
10257
10258fe_id : integer
10259 Returns an integer containing the current frontend's id. It can be used in
10260 backends to check from which backend it was called, or to stick all users
10261 coming via a same frontend to the same server.
10262
Cyril Bonté62ba8702014-04-22 23:52:25 +020010263sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010264sc0_bytes_in_rate([<table>]) : integer
10265sc1_bytes_in_rate([<table>]) : integer
10266sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010267 Returns the average client-to-server bytes rate from the currently tracked
10268 counters, measured in amount of bytes over the period configured in the
10269 table. See also src_bytes_in_rate.
10270
Cyril Bonté62ba8702014-04-22 23:52:25 +020010271sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010272sc0_bytes_out_rate([<table>]) : integer
10273sc1_bytes_out_rate([<table>]) : integer
10274sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010275 Returns the average server-to-client bytes rate from the currently tracked
10276 counters, measured in amount of bytes over the period configured in the
10277 table. See also src_bytes_out_rate.
10278
Cyril Bonté62ba8702014-04-22 23:52:25 +020010279sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010280sc0_clr_gpc0([<table>]) : integer
10281sc1_clr_gpc0([<table>]) : integer
10282sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020010283 Clears the first General Purpose Counter associated to the currently tracked
10284 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010010285 stored value is zero, so first invocation will always return zero. This is
10286 typically used as a second ACL in an expression in order to mark a connection
10287 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020010288
10289 # block if 5 consecutive requests continue to come faster than 10 sess
10290 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010291 acl abuse sc0_http_req_rate gt 10
10292 acl kill sc0_inc_gpc0 gt 5
10293 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020010294 tcp-request connection accept if !abuse save
10295 tcp-request connection reject if abuse kill
10296
Cyril Bonté62ba8702014-04-22 23:52:25 +020010297sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010298sc0_conn_cnt([<table>]) : integer
10299sc1_conn_cnt([<table>]) : integer
10300sc2_conn_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010301 Returns the cumulated number of incoming connections from currently tracked
10302 counters. See also src_conn_cnt.
10303
Cyril Bonté62ba8702014-04-22 23:52:25 +020010304sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010305sc0_conn_cur([<table>]) : integer
10306sc1_conn_cur([<table>]) : integer
10307sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010308 Returns the current amount of concurrent connections tracking the same
10309 tracked counters. This number is automatically incremented when tracking
10310 begins and decremented when tracking stops. See also src_conn_cur.
10311
Cyril Bonté62ba8702014-04-22 23:52:25 +020010312sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010313sc0_conn_rate([<table>]) : integer
10314sc1_conn_rate([<table>]) : integer
10315sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010316 Returns the average connection rate from the currently tracked counters,
10317 measured in amount of connections over the period configured in the table.
10318 See also src_conn_rate.
10319
Cyril Bonté62ba8702014-04-22 23:52:25 +020010320sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010321sc0_get_gpc0([<table>]) : integer
10322sc1_get_gpc0([<table>]) : integer
10323sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010324 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010325 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010326
Cyril Bonté62ba8702014-04-22 23:52:25 +020010327sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010328sc0_gpc0_rate([<table>]) : integer
10329sc1_gpc0_rate([<table>]) : integer
10330sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010331 Returns the average increment rate of the first General Purpose Counter
10332 associated to the currently tracked counters. It reports the frequency
10333 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010334 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
10335 that the "gpc0_rate" counter must be stored in the stick-table for a value to
10336 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020010337
Cyril Bonté62ba8702014-04-22 23:52:25 +020010338sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010339sc0_http_err_cnt([<table>]) : integer
10340sc1_http_err_cnt([<table>]) : integer
10341sc2_http_err_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010342 Returns the cumulated number of HTTP errors from the currently tracked
10343 counters. This includes the both request errors and 4xx error responses.
10344 See also src_http_err_cnt.
10345
Cyril Bonté62ba8702014-04-22 23:52:25 +020010346sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010347sc0_http_err_rate([<table>]) : integer
10348sc1_http_err_rate([<table>]) : integer
10349sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010350 Returns the average rate of HTTP errors from the currently tracked counters,
10351 measured in amount of errors over the period configured in the table. This
10352 includes the both request errors and 4xx error responses. See also
10353 src_http_err_rate.
10354
Cyril Bonté62ba8702014-04-22 23:52:25 +020010355sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010356sc0_http_req_cnt([<table>]) : integer
10357sc1_http_req_cnt([<table>]) : integer
10358sc2_http_req_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010359 Returns the cumulated number of HTTP requests from the currently tracked
10360 counters. This includes every started request, valid or not. See also
10361 src_http_req_cnt.
10362
Cyril Bonté62ba8702014-04-22 23:52:25 +020010363sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010364sc0_http_req_rate([<table>]) : integer
10365sc1_http_req_rate([<table>]) : integer
10366sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010367 Returns the average rate of HTTP requests from the currently tracked
10368 counters, measured in amount of requests over the period configured in
10369 the table. This includes every started request, valid or not. See also
10370 src_http_req_rate.
10371
Cyril Bonté62ba8702014-04-22 23:52:25 +020010372sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010373sc0_inc_gpc0([<table>]) : integer
10374sc1_inc_gpc0([<table>]) : integer
10375sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010376 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010010377 tracked counters, and returns its new value. Before the first invocation,
10378 the stored value is zero, so first invocation will increase it to 1 and will
10379 return 1. This is typically used as a second ACL in an expression in order
10380 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020010381
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010382 acl abuse sc0_http_req_rate gt 10
10383 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020010384 tcp-request connection reject if abuse kill
10385
Cyril Bonté62ba8702014-04-22 23:52:25 +020010386sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010387sc0_kbytes_in([<table>]) : integer
10388sc1_kbytes_in([<table>]) : integer
10389sc2_kbytes_in([<table>]) : integer
Willy Tarreau911780c2014-07-10 15:29:24 +020010390 Returns the total amount of client-to-server data from the currently tracked
10391 counters, measured in kilobytes. The test is currently performed on 32-bit
10392 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020010393
Cyril Bonté62ba8702014-04-22 23:52:25 +020010394sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010395sc0_kbytes_out([<table>]) : integer
10396sc1_kbytes_out([<table>]) : integer
10397sc2_kbytes_out([<table>]) : integer
Willy Tarreau911780c2014-07-10 15:29:24 +020010398 Returns the total amount of server-to-client data from the currently tracked
10399 counters, measured in kilobytes. The test is currently performed on 32-bit
10400 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020010401
Cyril Bonté62ba8702014-04-22 23:52:25 +020010402sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010403sc0_sess_cnt([<table>]) : integer
10404sc1_sess_cnt([<table>]) : integer
10405sc2_sess_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010406 Returns the cumulated number of incoming connections that were transformed
10407 into sessions, which means that they were accepted by a "tcp-request
10408 connection" rule, from the currently tracked counters. A backend may count
10409 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010410 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020010411 with the client. See also src_sess_cnt.
10412
Cyril Bonté62ba8702014-04-22 23:52:25 +020010413sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010414sc0_sess_rate([<table>]) : integer
10415sc1_sess_rate([<table>]) : integer
10416sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020010417 Returns the average session rate from the currently tracked counters,
10418 measured in amount of sessions over the period configured in the table. A
10419 session is a connection that got past the early "tcp-request connection"
10420 rules. A backend may count more sessions than connections because each
10421 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010422 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020010423
Cyril Bonté62ba8702014-04-22 23:52:25 +020010424sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020010425sc0_tracked([<table>]) : boolean
10426sc1_tracked([<table>]) : boolean
10427sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020010428 Returns true if the designated session counter is currently being tracked by
10429 the current session. This can be useful when deciding whether or not we want
10430 to set some values in a header passed to the server.
10431
Cyril Bonté62ba8702014-04-22 23:52:25 +020010432sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020010433sc0_trackers([<table>]) : integer
10434sc1_trackers([<table>]) : integer
10435sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010010436 Returns the current amount of concurrent connections tracking the same
10437 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010438 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010010439 that it does not rely on any stored information but on the table's reference
10440 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020010441 may sometimes be more suited for layer7 tracking. It can be used to tell a
10442 server how many concurrent connections there are from a given address for
10443 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010010444
Willy Tarreau74ca5042013-06-11 23:12:07 +020010445so_id : integer
10446 Returns an integer containing the current listening socket's id. It is useful
10447 in frontends involving many "bind" lines, or to stick all users coming via a
10448 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010449
Willy Tarreau74ca5042013-06-11 23:12:07 +020010450src : ip
10451 This is the source IPv4 address of the client of the session. It is of type
10452 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
10453 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
10454 TCP-level source address which is used, and not the address of a client
10455 behind a proxy. However if the "accept-proxy" bind directive is used, it can
10456 be the address of a client behind another PROXY-protocol compatible component
10457 for all rule sets except "tcp-request connection" which sees the real address.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010458
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010010459 Example:
10460 # add an HTTP header in requests with the originating address' country
10461 http-request set-header X-Country %[src,map_ip(geoip.lst)]
10462
Willy Tarreau74ca5042013-06-11 23:12:07 +020010463src_bytes_in_rate([<table>]) : integer
10464 Returns the average bytes rate from the incoming connection's source address
10465 in the current proxy's stick-table or in the designated stick-table, measured
10466 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010467 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010468
Willy Tarreau74ca5042013-06-11 23:12:07 +020010469src_bytes_out_rate([<table>]) : integer
10470 Returns the average bytes rate to the incoming connection's source address in
10471 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020010472 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010473 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010474
Willy Tarreau74ca5042013-06-11 23:12:07 +020010475src_clr_gpc0([<table>]) : integer
10476 Clears the first General Purpose Counter associated to the incoming
10477 connection's source address in the current proxy's stick-table or in the
10478 designated stick-table, and returns its previous value. If the address is not
10479 found, an entry is created and 0 is returned. This is typically used as a
10480 second ACL in an expression in order to mark a connection when a first ACL
10481 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020010482
10483 # block if 5 consecutive requests continue to come faster than 10 sess
10484 # per second, and reset the counter as soon as the traffic slows down.
10485 acl abuse src_http_req_rate gt 10
10486 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010010487 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020010488 tcp-request connection accept if !abuse save
10489 tcp-request connection reject if abuse kill
10490
Willy Tarreau74ca5042013-06-11 23:12:07 +020010491src_conn_cnt([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020010492 Returns the cumulated number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020010493 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020010494 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010495 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010496
Willy Tarreau74ca5042013-06-11 23:12:07 +020010497src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020010498 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020010499 current incoming connection's source address in the current proxy's
10500 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010501 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010502
Willy Tarreau74ca5042013-06-11 23:12:07 +020010503src_conn_rate([<table>]) : integer
10504 Returns the average connection rate from the incoming connection's source
10505 address in the current proxy's stick-table or in the designated stick-table,
10506 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010507 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010508
Willy Tarreau74ca5042013-06-11 23:12:07 +020010509src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020010510 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020010511 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020010512 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010513 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010514
Willy Tarreau74ca5042013-06-11 23:12:07 +020010515src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010516 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020010517 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010518 stick-table or in the designated stick-table. It reports the frequency
10519 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010520 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
10521 that the "gpc0_rate" counter must be stored in the stick-table for a value to
10522 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010523
Willy Tarreau74ca5042013-06-11 23:12:07 +020010524src_http_err_cnt([<table>]) : integer
10525 Returns the cumulated number of HTTP errors from the incoming connection's
10526 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020010527 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010528 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020010529 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010530
Willy Tarreau74ca5042013-06-11 23:12:07 +020010531src_http_err_rate([<table>]) : integer
10532 Returns the average rate of HTTP errors from the incoming connection's source
10533 address in the current proxy's stick-table or in the designated stick-table,
10534 measured in amount of errors over the period configured in the table. This
10535 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010536 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010537
Willy Tarreau74ca5042013-06-11 23:12:07 +020010538src_http_req_cnt([<table>]) : integer
10539 Returns the cumulated number of HTTP requests from the incoming connection's
10540 source address in the current proxy's stick-table or in the designated stick-
10541 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010542 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010543
Willy Tarreau74ca5042013-06-11 23:12:07 +020010544src_http_req_rate([<table>]) : integer
10545 Returns the average rate of HTTP requests from the incoming connection's
10546 source address in the current proxy's stick-table or in the designated stick-
10547 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020010548 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010549 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010550
Willy Tarreau74ca5042013-06-11 23:12:07 +020010551src_inc_gpc0([<table>]) : integer
10552 Increments the first General Purpose Counter associated to the incoming
10553 connection's source address in the current proxy's stick-table or in the
10554 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010555 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020010556 This is typically used as a second ACL in an expression in order to mark a
10557 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020010558
10559 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010010560 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020010561 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020010562
Willy Tarreau74ca5042013-06-11 23:12:07 +020010563src_kbytes_in([<table>]) : integer
Willy Tarreau911780c2014-07-10 15:29:24 +020010564 Returns the total amount of data received from the incoming connection's
10565 source address in the current proxy's stick-table or in the designated
10566 stick-table, measured in kilobytes. If the address is not found, zero is
10567 returned. The test is currently performed on 32-bit integers, which limits
10568 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010569
Willy Tarreau74ca5042013-06-11 23:12:07 +020010570src_kbytes_out([<table>]) : integer
Willy Tarreau911780c2014-07-10 15:29:24 +020010571 Returns the total amount of data sent to the incoming connection's source
10572 address in the current proxy's stick-table or in the designated stick-table,
10573 measured in kilobytes. If the address is not found, zero is returned. The
10574 test is currently performed on 32-bit integers, which limits values to 4
10575 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020010576
Willy Tarreau74ca5042013-06-11 23:12:07 +020010577src_port : integer
10578 Returns an integer value corresponding to the TCP source port of the
10579 connection on the client side, which is the port the client connected from.
10580 Usage of this function is very limited as modern protocols do not care much
10581 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010010582
Willy Tarreau74ca5042013-06-11 23:12:07 +020010583src_sess_cnt([<table>]) : integer
10584 Returns the cumulated number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020010585 connection's source IPv4 address in the current proxy's stick-table or in the
10586 designated stick-table, that were transformed into sessions, which means that
10587 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010588 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010589
Willy Tarreau74ca5042013-06-11 23:12:07 +020010590src_sess_rate([<table>]) : integer
10591 Returns the average session rate from the incoming connection's source
10592 address in the current proxy's stick-table or in the designated stick-table,
10593 measured in amount of sessions over the period configured in the table. A
10594 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020010595 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010596
Willy Tarreau74ca5042013-06-11 23:12:07 +020010597src_updt_conn_cnt([<table>]) : integer
10598 Creates or updates the entry associated to the incoming connection's source
10599 address in the current proxy's stick-table or in the designated stick-table.
10600 This table must be configured to store the "conn_cnt" data type, otherwise
10601 the match will be ignored. The current count is incremented by one, and the
10602 expiration timer refreshed. The updated count is returned, so this match
10603 can't return zero. This was used to reject service abusers based on their
10604 source address. Note: it is recommended to use the more complete "track-sc*"
10605 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020010606
10607 Example :
10608 # This frontend limits incoming SSH connections to 3 per 10 second for
10609 # each source address, and rejects excess connections until a 10 second
10610 # silence is observed. At most 20 addresses are tracked.
10611 listen ssh
10612 bind :22
10613 mode tcp
10614 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020010615 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020010616 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020010617 server local 127.0.0.1:22
10618
Willy Tarreau74ca5042013-06-11 23:12:07 +020010619srv_id : integer
10620 Returns an integer containing the server's id when processing the response.
10621 While it's almost only used with ACLs, it may be used for logging or
10622 debugging.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020010623
Hervé COMMOWICK35ed8012010-12-15 14:04:51 +010010624
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200106257.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020010626----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020010627
Willy Tarreau74ca5042013-06-11 23:12:07 +020010628The layer 5 usually describes just the session layer which in haproxy is
10629closest to the session once all the connection handshakes are finished, but
10630when no content is yet made available. The fetch methods described here are
10631usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010632future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020010633
Emeric Brun645ae792014-04-30 14:21:06 +020010634ssl_bc : boolean
10635 Returns true when the back connection was made via an SSL/TLS transport
10636 layer and is locally deciphered. This means the outgoing connection was made
10637 other a server with the "ssl" option.
10638
10639ssl_bc_alg_keysize : integer
10640 Returns the symmetric cipher key size supported in bits when the outgoing
10641 connection was made over an SSL/TLS transport layer.
10642
10643ssl_bc_cipher : string
10644 Returns the name of the used cipher when the outgoing connection was made
10645 over an SSL/TLS transport layer.
10646
10647ssl_bc_protocol : string
10648 Returns the name of the used protocol when the outgoing connection was made
10649 over an SSL/TLS transport layer.
10650
Emeric Brunb73a9b02014-04-30 18:49:19 +020010651ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020010652 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020010653 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
10654 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
Emeric Brun645ae792014-04-30 14:21:06 +020010655
10656ssl_bc_session_id : binary
10657 Returns the SSL ID of the back connection when the outgoing connection was
10658 made over an SSL/TLS transport layer. It is useful to log if we want to know
10659 if session was reused or not.
10660
10661ssl_bc_use_keysize : integer
10662 Returns the symmetric cipher key size used in bits when the outgoing
10663 connection was made over an SSL/TLS transport layer.
10664
Willy Tarreau74ca5042013-06-11 23:12:07 +020010665ssl_c_ca_err : integer
10666 When the incoming connection was made over an SSL/TLS transport layer,
10667 returns the ID of the first error detected during verification of the client
10668 certificate at depth > 0, or 0 if no error was encountered during this
10669 verification process. Please refer to your SSL library's documentation to
10670 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020010671
Willy Tarreau74ca5042013-06-11 23:12:07 +020010672ssl_c_ca_err_depth : integer
10673 When the incoming connection was made over an SSL/TLS transport layer,
10674 returns the depth in the CA chain of the first error detected during the
10675 verification of the client certificate. If no error is encountered, 0 is
10676 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010010677
Willy Tarreau74ca5042013-06-11 23:12:07 +020010678ssl_c_err : integer
10679 When the incoming connection was made over an SSL/TLS transport layer,
10680 returns the ID of the first error detected during verification at depth 0, or
10681 0 if no error was encountered during this verification process. Please refer
10682 to your SSL library's documentation to find the exhaustive list of error
10683 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020010684
Willy Tarreau74ca5042013-06-11 23:12:07 +020010685ssl_c_i_dn([<entry>[,<occ>]]) : string
10686 When the incoming connection was made over an SSL/TLS transport layer,
10687 returns the full distinguished name of the issuer of the certificate
10688 presented by the client when no <entry> is specified, or the value of the
10689 first given entry found from the beginning of the DN. If a positive/negative
10690 occurrence number is specified as the optional second argument, it returns
10691 the value of the nth given entry value from the beginning/end of the DN.
10692 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
10693 "ssl_c_i_dn(CN)" retrieves the common name.
Willy Tarreau62644772008-07-16 18:36:06 +020010694
Willy Tarreau74ca5042013-06-11 23:12:07 +020010695ssl_c_key_alg : string
10696 Returns the name of the algorithm used to generate the key of the certificate
10697 presented by the client when the incoming connection was made over an SSL/TLS
10698 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020010699
Willy Tarreau74ca5042013-06-11 23:12:07 +020010700ssl_c_notafter : string
10701 Returns the end date presented by the client as a formatted string
10702 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
10703 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020010704
Willy Tarreau74ca5042013-06-11 23:12:07 +020010705ssl_c_notbefore : string
10706 Returns the start date presented by the client as a formatted string
10707 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
10708 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010010709
Willy Tarreau74ca5042013-06-11 23:12:07 +020010710ssl_c_s_dn([<entry>[,<occ>]]) : string
10711 When the incoming connection was made over an SSL/TLS transport layer,
10712 returns the full distinguished name of the subject of the certificate
10713 presented by the client when no <entry> is specified, or the value of the
10714 first given entry found from the beginning of the DN. If a positive/negative
10715 occurrence number is specified as the optional second argument, it returns
10716 the value of the nth given entry value from the beginning/end of the DN.
10717 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
10718 "ssl_c_s_dn(CN)" retrieves the common name.
Willy Tarreaub6672b52011-12-12 17:23:41 +010010719
Willy Tarreau74ca5042013-06-11 23:12:07 +020010720ssl_c_serial : binary
10721 Returns the serial of the certificate presented by the client when the
10722 incoming connection was made over an SSL/TLS transport layer. When used for
10723 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020010724
Willy Tarreau74ca5042013-06-11 23:12:07 +020010725ssl_c_sha1 : binary
10726 Returns the SHA-1 fingerprint of the certificate presented by the client when
10727 the incoming connection was made over an SSL/TLS transport layer. This can be
10728 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau9fe4cb62014-07-02 19:01:22 +020010729 Note that the output is binary, so if you want to pass that signature to the
10730 server, you need to encode it in hex or base64, such as in the example below:
10731
10732 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020010733
Willy Tarreau74ca5042013-06-11 23:12:07 +020010734ssl_c_sig_alg : string
10735 Returns the name of the algorithm used to sign the certificate presented by
10736 the client when the incoming connection was made over an SSL/TLS transport
10737 layer.
Emeric Brun87855892012-10-17 17:39:35 +020010738
Willy Tarreau74ca5042013-06-11 23:12:07 +020010739ssl_c_used : boolean
10740 Returns true if current SSL session uses a client certificate even if current
10741 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020010742
Willy Tarreau74ca5042013-06-11 23:12:07 +020010743ssl_c_verify : integer
10744 Returns the verify result error ID when the incoming connection was made over
10745 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
10746 refer to your SSL library's documentation for an exhaustive list of error
10747 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020010748
Willy Tarreau74ca5042013-06-11 23:12:07 +020010749ssl_c_version : integer
10750 Returns the version of the certificate presented by the client when the
10751 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020010752
Willy Tarreau74ca5042013-06-11 23:12:07 +020010753ssl_f_i_dn([<entry>[,<occ>]]) : string
10754 When the incoming connection was made over an SSL/TLS transport layer,
10755 returns the full distinguished name of the issuer of the certificate
10756 presented by the frontend when no <entry> is specified, or the value of the
10757 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020010758 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020010759 the value of the nth given entry value from the beginning/end of the DN.
10760 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
10761 "ssl_f_i_dn(CN)" retrieves the common name.
Emeric Brun87855892012-10-17 17:39:35 +020010762
Willy Tarreau74ca5042013-06-11 23:12:07 +020010763ssl_f_key_alg : string
10764 Returns the name of the algorithm used to generate the key of the certificate
10765 presented by the frontend when the incoming connection was made over an
10766 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020010767
Willy Tarreau74ca5042013-06-11 23:12:07 +020010768ssl_f_notafter : string
10769 Returns the end date presented by the frontend as a formatted string
10770 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
10771 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020010772
Willy Tarreau74ca5042013-06-11 23:12:07 +020010773ssl_f_notbefore : string
10774 Returns the start date presented by the frontend as a formatted string
10775 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
10776 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020010777
Willy Tarreau74ca5042013-06-11 23:12:07 +020010778ssl_f_s_dn([<entry>[,<occ>]]) : string
10779 When the incoming connection was made over an SSL/TLS transport layer,
10780 returns the full distinguished name of the subject of the certificate
10781 presented by the frontend when no <entry> is specified, or the value of the
10782 first given entry found from the beginning of the DN. If a positive/negative
10783 occurrence number is specified as the optional second argument, it returns
10784 the value of the nth given entry value from the beginning/end of the DN.
10785 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
10786 "ssl_f_s_dn(CN)" retrieves the common name.
Emeric Brunce5ad802012-10-22 14:11:22 +020010787
Willy Tarreau74ca5042013-06-11 23:12:07 +020010788ssl_f_serial : binary
10789 Returns the serial of the certificate presented by the frontend when the
10790 incoming connection was made over an SSL/TLS transport layer. When used for
10791 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020010792
Emeric Brun55f4fa82014-04-30 17:11:25 +020010793ssl_f_sha1 : binary
10794 Returns the SHA-1 fingerprint of the certificate presented by the frontend
10795 when the incoming connection was made over an SSL/TLS transport layer. This
10796 can be used to know which certificate was chosen using SNI.
10797
Willy Tarreau74ca5042013-06-11 23:12:07 +020010798ssl_f_sig_alg : string
10799 Returns the name of the algorithm used to sign the certificate presented by
10800 the frontend when the incoming connection was made over an SSL/TLS transport
10801 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020010802
Willy Tarreau74ca5042013-06-11 23:12:07 +020010803ssl_f_version : integer
10804 Returns the version of the certificate presented by the frontend when the
10805 incoming connection was made over an SSL/TLS transport layer.
10806
10807ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020010808 Returns true when the front connection was made via an SSL/TLS transport
10809 layer and is locally deciphered. This means it has matched a socket declared
10810 with a "bind" line having the "ssl" option.
10811
Willy Tarreau74ca5042013-06-11 23:12:07 +020010812 Example :
10813 # This passes "X-Proto: https" to servers when client connects over SSL
10814 listen http-https
10815 bind :80
10816 bind :443 ssl crt /etc/haproxy.pem
10817 http-request add-header X-Proto https if { ssl_fc }
10818
10819ssl_fc_alg_keysize : integer
10820 Returns the symmetric cipher key size supported in bits when the incoming
10821 connection was made over an SSL/TLS transport layer.
10822
10823ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010824 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020010825 incoming connection made via a TLS transport layer and locally deciphered by
10826 haproxy. The result is a string containing the protocol name advertised by
10827 the client. The SSL library must have been built with support for TLS
10828 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
10829 not advertised unless the "alpn" keyword on the "bind" line specifies a
10830 protocol list. Also, nothing forces the client to pick a protocol from this
10831 list, any other one may be requested. The TLS ALPN extension is meant to
10832 replace the TLS NPN extension. See also "ssl_fc_npn".
10833
Willy Tarreau74ca5042013-06-11 23:12:07 +020010834ssl_fc_cipher : string
10835 Returns the name of the used cipher when the incoming connection was made
10836 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020010837
Willy Tarreau74ca5042013-06-11 23:12:07 +020010838ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020010839 Returns true if a client certificate is present in an incoming connection over
10840 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010010841 Note: on SSL session resumption with Session ID or TLS ticket, client
10842 certificate is not present in the current connection but may be retrieved
10843 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
10844 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020010845
Willy Tarreau74ca5042013-06-11 23:12:07 +020010846ssl_fc_has_sni : boolean
10847 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020010848 in an incoming connection was made over an SSL/TLS transport layer. Returns
10849 true when the incoming connection presents a TLS SNI field. This requires
10850 that the SSL library is build with support for TLS extensions enabled (check
10851 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020010852
Willy Tarreau74ca5042013-06-11 23:12:07 +020010853ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010854 This extracts the Next Protocol Negotiation field from an incoming connection
Willy Tarreau74ca5042013-06-11 23:12:07 +020010855 made via a TLS transport layer and locally deciphered by haproxy. The result
10856 is a string containing the protocol name advertised by the client. The SSL
10857 library must have been built with support for TLS extensions enabled (check
10858 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
10859 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
10860 forces the client to pick a protocol from this list, any other one may be
10861 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020010862
Willy Tarreau74ca5042013-06-11 23:12:07 +020010863ssl_fc_protocol : string
10864 Returns the name of the used protocol when the incoming connection was made
10865 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020010866
Emeric Brunb73a9b02014-04-30 18:49:19 +020010867ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040010868 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020010869 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
10870 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040010871
Willy Tarreau74ca5042013-06-11 23:12:07 +020010872ssl_fc_session_id : binary
10873 Returns the SSL ID of the front connection when the incoming connection was
10874 made over an SSL/TLS transport layer. It is useful to stick a given client to
10875 a server. It is important to note that some browsers refresh their session ID
10876 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020010877
Willy Tarreau74ca5042013-06-11 23:12:07 +020010878ssl_fc_sni : string
10879 This extracts the Server Name Indication TLS extension (SNI) field from an
10880 incoming connection made via an SSL/TLS transport layer and locally
10881 deciphered by haproxy. The result (when present) typically is a string
10882 matching the HTTPS host name (253 chars or less). The SSL library must have
10883 been built with support for TLS extensions enabled (check haproxy -vv).
10884
10885 This fetch is different from "req_ssl_sni" above in that it applies to the
10886 connection being deciphered by haproxy and not to SSL contents being blindly
10887 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020010888 requires that the SSL library is build with support for TLS extensions
10889 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020010890
Willy Tarreau74ca5042013-06-11 23:12:07 +020010891 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020010892 ssl_fc_sni_end : suffix match
10893 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020010894
Willy Tarreau74ca5042013-06-11 23:12:07 +020010895ssl_fc_use_keysize : integer
10896 Returns the symmetric cipher key size used in bits when the incoming
10897 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020010898
Willy Tarreaub6fb4202008-07-20 11:18:28 +020010899
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200109007.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020010901------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020010902
Willy Tarreau74ca5042013-06-11 23:12:07 +020010903Fetching samples from buffer contents is a bit different from the previous
10904sample fetches above because the sampled data are ephemeral. These data can
10905only be used when they're available and will be lost when they're forwarded.
10906For this reason, samples fetched from buffer contents during a request cannot
10907be used in a response for example. Even while the data are being fetched, they
10908can change. Sometimes it is necessary to set some delays or combine multiple
10909sample fetch methods to ensure that the expected data are complete and usable,
10910for example through TCP request content inspection. Please see the "tcp-request
10911content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020010912
Willy Tarreau74ca5042013-06-11 23:12:07 +020010913payload(<offset>,<length>) : binary (deprecated)
10914 This is an alias for "req.payload" when used in the context of a request (eg:
10915 "stick on", "stick match"), and for "res.payload" when used in the context of
10916 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010917
Willy Tarreau74ca5042013-06-11 23:12:07 +020010918payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
10919 This is an alias for "req.payload_lv" when used in the context of a request
10920 (eg: "stick on", "stick match"), and for "res.payload_lv" when used in the
10921 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010922
Willy Tarreau74ca5042013-06-11 23:12:07 +020010923req.len : integer
10924req_len : integer (deprecated)
10925 Returns an integer value corresponding to the number of bytes present in the
10926 request buffer. This is mostly used in ACL. It is important to understand
10927 that this test does not return false as long as the buffer is changing. This
10928 means that a check with equality to zero will almost always immediately match
10929 at the beginning of the session, while a test for more data will wait for
10930 that data to come in and return false only when haproxy is certain that no
10931 more data will come in. This test was designed to be used with TCP request
10932 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010933
Willy Tarreau74ca5042013-06-11 23:12:07 +020010934req.payload(<offset>,<length>) : binary
10935 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020010936 in the request buffer. As a special case, if the <length> argument is zero,
10937 the the whole buffer from <offset> to the end is extracted. This can be used
10938 with ACLs in order to check for the presence of some content in a buffer at
10939 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010940
Willy Tarreau74ca5042013-06-11 23:12:07 +020010941 ACL alternatives :
10942 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010943
Willy Tarreau74ca5042013-06-11 23:12:07 +020010944req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
10945 This extracts a binary block whose size is specified at <offset1> for <length>
10946 bytes, and which starts at <offset2> if specified or just after the length in
10947 the request buffer. The <offset2> parameter also supports relative offsets if
10948 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010949
Willy Tarreau74ca5042013-06-11 23:12:07 +020010950 ACL alternatives :
10951 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010952
Willy Tarreau74ca5042013-06-11 23:12:07 +020010953 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010954
Willy Tarreau74ca5042013-06-11 23:12:07 +020010955req.proto_http : boolean
10956req_proto_http : boolean (deprecated)
10957 Returns true when data in the request buffer look like HTTP and correctly
10958 parses as such. It is the same parser as the common HTTP request parser which
10959 is used so there should be no surprises. The test does not match until the
10960 request is complete, failed or timed out. This test may be used to report the
10961 protocol in TCP logs, but the biggest use is to block TCP request analysis
10962 until a complete HTTP request is present in the buffer, for example to track
10963 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010964
Willy Tarreau74ca5042013-06-11 23:12:07 +020010965 Example:
10966 # track request counts per "base" (concatenation of Host+URL)
10967 tcp-request inspect-delay 10s
10968 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010969 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020010970
Willy Tarreau74ca5042013-06-11 23:12:07 +020010971req.rdp_cookie([<name>]) : string
10972rdp_cookie([<name>]) : string (deprecated)
10973 When the request buffer looks like the RDP protocol, extracts the RDP cookie
10974 <name>, or any cookie if unspecified. The parser only checks for the first
10975 cookie, as illustrated in the RDP protocol specification. The cookie name is
10976 case insensitive. Generally the "MSTS" cookie name will be used, as it can
10977 contain the user name of the client connecting to the server if properly
10978 configured on the client. The "MSTSHASH" cookie is often used as well for
10979 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020010980
Willy Tarreau74ca5042013-06-11 23:12:07 +020010981 This differs from "balance rdp-cookie" in that any balancing algorithm may be
10982 used and thus the distribution of clients to backend servers is not linked to
10983 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
10984 such as "balance roundrobin" or "balance leastconn" will lead to a more even
10985 distribution of clients to backend servers than the hash used by "balance
10986 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020010987
Willy Tarreau74ca5042013-06-11 23:12:07 +020010988 ACL derivatives :
10989 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020010990
Willy Tarreau74ca5042013-06-11 23:12:07 +020010991 Example :
10992 listen tse-farm
10993 bind 0.0.0.0:3389
10994 # wait up to 5s for an RDP cookie in the request
10995 tcp-request inspect-delay 5s
10996 tcp-request content accept if RDP_COOKIE
10997 # apply RDP cookie persistence
10998 persist rdp-cookie
10999 # Persist based on the mstshash cookie
11000 # This is only useful makes sense if
11001 # balance rdp-cookie is not used
11002 stick-table type string size 204800
11003 stick on req.rdp_cookie(mstshash)
11004 server srv1 1.1.1.1:3389
11005 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011006
Willy Tarreau74ca5042013-06-11 23:12:07 +020011007 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
11008 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011009
Willy Tarreau74ca5042013-06-11 23:12:07 +020011010req.rdp_cookie_cnt([name]) : integer
11011rdp_cookie_cnt([name]) : integer (deprecated)
11012 Tries to parse the request buffer as RDP protocol, then returns an integer
11013 corresponding to the number of RDP cookies found. If an optional cookie name
11014 is passed, only cookies matching this name are considered. This is mostly
11015 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011016
Willy Tarreau74ca5042013-06-11 23:12:07 +020011017 ACL derivatives :
11018 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011019
Willy Tarreau74ca5042013-06-11 23:12:07 +020011020req.ssl_hello_type : integer
11021req_ssl_hello_type : integer (deprecated)
11022 Returns an integer value containing the type of the SSL hello message found
11023 in the request buffer if the buffer contains data that parse as a complete
11024 SSL (v3 or superior) client hello message. Note that this only applies to raw
11025 contents found in the request buffer and not to contents deciphered via an
11026 SSL data layer, so this will not work with "bind" lines having the "ssl"
11027 option. This is mostly used in ACL to detect presence of an SSL hello message
11028 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011029
Willy Tarreau74ca5042013-06-11 23:12:07 +020011030req.ssl_sni : string
11031req_ssl_sni : string (deprecated)
11032 Returns a string containing the value of the Server Name TLS extension sent
11033 by a client in a TLS stream passing through the request buffer if the buffer
11034 contains data that parse as a complete SSL (v3 or superior) client hello
11035 message. Note that this only applies to raw contents found in the request
11036 buffer and not to contents deciphered via an SSL data layer, so this will not
11037 work with "bind" lines having the "ssl" option. SNI normally contains the
11038 name of the host the client tries to connect to (for recent browsers). SNI is
11039 useful for allowing or denying access to certain hosts when SSL/TLS is used
11040 by the client. This test was designed to be used with TCP request content
11041 inspection. If content switching is needed, it is recommended to first wait
11042 for a complete client hello (type 1), like in the example below. See also
11043 "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011044
Willy Tarreau74ca5042013-06-11 23:12:07 +020011045 ACL derivatives :
11046 req_ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011047
Willy Tarreau74ca5042013-06-11 23:12:07 +020011048 Examples :
11049 # Wait for a client hello for at most 5 seconds
11050 tcp-request inspect-delay 5s
11051 tcp-request content accept if { req_ssl_hello_type 1 }
11052 use_backend bk_allow if { req_ssl_sni -f allowed_sites }
11053 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020011054
Willy Tarreau74ca5042013-06-11 23:12:07 +020011055res.ssl_hello_type : integer
11056rep_ssl_hello_type : integer (deprecated)
11057 Returns an integer value containing the type of the SSL hello message found
11058 in the response buffer if the buffer contains data that parses as a complete
11059 SSL (v3 or superior) hello message. Note that this only applies to raw
11060 contents found in the response buffer and not to contents deciphered via an
11061 SSL data layer, so this will not work with "server" lines having the "ssl"
11062 option. This is mostly used in ACL to detect presence of an SSL hello message
11063 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau51539362012-05-08 12:46:28 +020011064
Willy Tarreau74ca5042013-06-11 23:12:07 +020011065req.ssl_ver : integer
11066req_ssl_ver : integer (deprecated)
11067 Returns an integer value containing the version of the SSL/TLS protocol of a
11068 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
11069 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
11070 composed of the major version multiplied by 65536, added to the minor
11071 version. Note that this only applies to raw contents found in the request
11072 buffer and not to contents deciphered via an SSL data layer, so this will not
11073 work with "bind" lines having the "ssl" option. The ACL version of the test
11074 matches against a decimal notation in the form MAJOR.MINOR (eg: 3.1). This
11075 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011076
Willy Tarreau74ca5042013-06-11 23:12:07 +020011077 ACL derivatives :
11078 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010011079
Willy Tarreau47e8eba2013-09-11 23:28:46 +020011080res.len : integer
11081 Returns an integer value corresponding to the number of bytes present in the
11082 response buffer. This is mostly used in ACL. It is important to understand
11083 that this test does not return false as long as the buffer is changing. This
11084 means that a check with equality to zero will almost always immediately match
11085 at the beginning of the session, while a test for more data will wait for
11086 that data to come in and return false only when haproxy is certain that no
11087 more data will come in. This test was designed to be used with TCP response
11088 content inspection.
11089
Willy Tarreau74ca5042013-06-11 23:12:07 +020011090res.payload(<offset>,<length>) : binary
11091 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020011092 in the response buffer. As a special case, if the <length> argument is zero,
11093 the the whole buffer from <offset> to the end is extracted. This can be used
11094 with ACLs in order to check for the presence of some content in a buffer at
11095 any location.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011096
Willy Tarreau74ca5042013-06-11 23:12:07 +020011097res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
11098 This extracts a binary block whose size is specified at <offset1> for <length>
11099 bytes, and which starts at <offset2> if specified or just after the length in
11100 the response buffer. The <offset2> parameter also supports relative offsets
11101 if prepended with a '+' or '-' sign.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011102
Willy Tarreau74ca5042013-06-11 23:12:07 +020011103 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011104
Willy Tarreau74ca5042013-06-11 23:12:07 +020011105wait_end : boolean
11106 This fetch either returns true when the inspection period is over, or does
11107 not fetch. It is only used in ACLs, in conjunction with content analysis to
11108 avoid returning a wrong verdict early. It may also be used to delay some
11109 actions, such as a delayed reject for some special addresses. Since it either
11110 stops the rules evaluation or immediately returns true, it is recommended to
11111 use this acl as the last one in a rule. Please note that the default ACL
11112 "WAIT_END" is always usable without prior declaration. This test was designed
11113 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011114
Willy Tarreau74ca5042013-06-11 23:12:07 +020011115 Examples :
11116 # delay every incoming request by 2 seconds
11117 tcp-request inspect-delay 2s
11118 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010011119
Willy Tarreau74ca5042013-06-11 23:12:07 +020011120 # don't immediately tell bad guys they are rejected
11121 tcp-request inspect-delay 10s
11122 acl goodguys src 10.0.0.0/24
11123 acl badguys src 10.0.1.0/24
11124 tcp-request content accept if goodguys
11125 tcp-request content reject if badguys WAIT_END
11126 tcp-request content reject
11127
11128
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200111297.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020011130--------------------------------------
11131
11132It is possible to fetch samples from HTTP contents, requests and responses.
11133This application layer is also called layer 7. It is only possible to fetch the
11134data in this section when a full HTTP request or response has been parsed from
11135its respective request or response buffer. This is always the case with all
11136HTTP specific rules and for sections running with "mode http". When using TCP
11137content inspection, it may be necessary to support an inspection delay in order
11138to let the request or response come in first. These fetches may require a bit
11139more CPU resources than the layer 4 ones, but not much since the request and
11140response are indexed.
11141
11142base : string
11143 This returns the concatenation of the first Host header and the path part of
11144 the request, which starts at the first slash and ends before the question
11145 mark. It can be useful in virtual hosted environments to detect URL abuses as
11146 well as to improve shared caches efficiency. Using this with a limited size
11147 stick table also allows one to collect statistics about most commonly
11148 requested objects by host/path. With ACLs it can allow simple content
11149 switching rules involving the host and the path at the same time, such as
11150 "www.example.com/favicon.ico". See also "path" and "uri".
11151
11152 ACL derivatives :
11153 base : exact string match
11154 base_beg : prefix match
11155 base_dir : subdir match
11156 base_dom : domain match
11157 base_end : suffix match
11158 base_len : length match
11159 base_reg : regex match
11160 base_sub : substring match
11161
11162base32 : integer
11163 This returns a 32-bit hash of the value returned by the "base" fetch method
11164 above. This is useful to track per-URL activity on high traffic sites without
11165 having to store all URLs. Instead a shorter hash is stored, saving a lot of
11166 memory. The output type is an unsigned integer.
11167
11168base32+src : binary
11169 This returns the concatenation of the base32 fetch above and the src fetch
11170 below. The resulting type is of type binary, with a size of 8 or 20 bytes
11171 depending on the source address family. This can be used to track per-IP,
11172 per-URL counters.
11173
William Lallemand65ad6e12014-01-31 15:08:02 +010011174capture.req.hdr(<idx>) : string
11175 This extracts the content of the header captured by the "capture request
11176 header", idx is the position of the capture keyword in the configuration.
11177 The first entry is an index of 0. See also: "capture request header".
11178
11179capture.req.method : string
11180 This extracts the METHOD of an HTTP request. It can be used in both request
11181 and response. Unlike "method", it can be used in both request and response
11182 because it's allocated.
11183
11184capture.req.uri : string
11185 This extracts the request's URI, which starts at the first slash and ends
11186 before the first space in the request (without the host part). Unlike "path"
11187 and "url", it can be used in both request and response because it's
11188 allocated.
11189
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020011190capture.req.ver : string
11191 This extracts the request's HTTP version and returns either "HTTP/1.0" or
11192 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
11193 logs because it relies on a persistent flag.
11194
William Lallemand65ad6e12014-01-31 15:08:02 +010011195capture.res.hdr(<idx>) : string
11196 This extracts the content of the header captured by the "capture response
11197 header", idx is the position of the capture keyword in the configuration.
11198 The first entry is an index of 0.
11199 See also: "capture response header"
11200
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020011201capture.res.ver : string
11202 This extracts the response's HTTP version and returns either "HTTP/1.0" or
11203 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
11204 persistent flag.
11205
Willy Tarreau74ca5042013-06-11 23:12:07 +020011206req.cook([<name>]) : string
11207cook([<name>]) : string (deprecated)
11208 This extracts the last occurrence of the cookie name <name> on a "Cookie"
11209 header line from the request, and returns its value as string. If no name is
11210 specified, the first cookie value is returned. When used with ACLs, all
11211 matching cookies are evaluated. Spaces around the name and the value are
11212 ignored as requested by the Cookie header specification (RFC6265). The cookie
11213 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
11214 well return an empty value if it is present. Use the "found" match to detect
11215 presence. Use the res.cook() variant for response cookies sent by the server.
11216
11217 ACL derivatives :
11218 cook([<name>]) : exact string match
11219 cook_beg([<name>]) : prefix match
11220 cook_dir([<name>]) : subdir match
11221 cook_dom([<name>]) : domain match
11222 cook_end([<name>]) : suffix match
11223 cook_len([<name>]) : length match
11224 cook_reg([<name>]) : regex match
11225 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010011226
Willy Tarreau74ca5042013-06-11 23:12:07 +020011227req.cook_cnt([<name>]) : integer
11228cook_cnt([<name>]) : integer (deprecated)
11229 Returns an integer value representing the number of occurrences of the cookie
11230 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011231
Willy Tarreau74ca5042013-06-11 23:12:07 +020011232req.cook_val([<name>]) : integer
11233cook_val([<name>]) : integer (deprecated)
11234 This extracts the last occurrence of the cookie name <name> on a "Cookie"
11235 header line from the request, and converts its value to an integer which is
11236 returned. If no name is specified, the first cookie value is returned. When
11237 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020011238
Willy Tarreau74ca5042013-06-11 23:12:07 +020011239cookie([<name>]) : string (deprecated)
11240 This extracts the last occurrence of the cookie name <name> on a "Cookie"
11241 header line from the request, or a "Set-Cookie" header from the response, and
11242 returns its value as a string. A typical use is to get multiple clients
11243 sharing a same profile use the same server. This can be similar to what
11244 "appsession" does with the "request-learn" statement, but with support for
11245 multi-peer synchronization and state keeping across restarts. If no name is
11246 specified, the first cookie value is returned. This fetch should not be used
11247 anymore and should be replaced by req.cook() or res.cook() instead as it
11248 ambiguously uses the direction based on the context where it is used.
11249 See also : "appsession".
Willy Tarreaud63335a2010-02-26 12:56:52 +010011250
Willy Tarreau74ca5042013-06-11 23:12:07 +020011251hdr([<name>[,<occ>]]) : string
11252 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
11253 used on responses. Please refer to these respective fetches for more details.
11254 In case of doubt about the fetch direction, please use the explicit ones.
11255 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011256 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011257
Willy Tarreau74ca5042013-06-11 23:12:07 +020011258req.fhdr(<name>[,<occ>]) : string
11259 This extracts the last occurrence of header <name> in an HTTP request. When
11260 used from an ACL, all occurrences are iterated over until a match is found.
11261 Optionally, a specific occurrence might be specified as a position number.
11262 Positive values indicate a position from the first occurrence, with 1 being
11263 the first one. Negative values indicate positions relative to the last one,
11264 with -1 being the last one. It differs from req.hdr() in that any commas
11265 present in the value are returned and are not used as delimiters. This is
11266 sometimes useful with headers such as User-Agent.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011267
Willy Tarreau74ca5042013-06-11 23:12:07 +020011268req.fhdr_cnt([<name>]) : integer
11269 Returns an integer value representing the number of occurrences of request
11270 header field name <name>, or the total number of header fields if <name> is
11271 not specified. Contrary to its req.hdr_cnt() cousin, this function returns
11272 the number of full line headers and does not stop on commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011273
Willy Tarreau74ca5042013-06-11 23:12:07 +020011274req.hdr([<name>[,<occ>]]) : string
11275 This extracts the last occurrence of header <name> in an HTTP request. When
11276 used from an ACL, all occurrences are iterated over until a match is found.
11277 Optionally, a specific occurrence might be specified as a position number.
11278 Positive values indicate a position from the first occurrence, with 1 being
11279 the first one. Negative values indicate positions relative to the last one,
11280 with -1 being the last one. A typical use is with the X-Forwarded-For header
11281 once converted to IP, associated with an IP stick-table. The function
11282 considers any comma as a delimiter for distinct values. If full-line headers
11283 are desired instead, use req.fhdr(). Please carefully check RFC2616 to know
11284 how certain headers are supposed to be parsed. Also, some of them are case
11285 insensitive (eg: Connection).
Willy Tarreaud63335a2010-02-26 12:56:52 +010011286
Willy Tarreau74ca5042013-06-11 23:12:07 +020011287 ACL derivatives :
11288 hdr([<name>[,<occ>]]) : exact string match
11289 hdr_beg([<name>[,<occ>]]) : prefix match
11290 hdr_dir([<name>[,<occ>]]) : subdir match
11291 hdr_dom([<name>[,<occ>]]) : domain match
11292 hdr_end([<name>[,<occ>]]) : suffix match
11293 hdr_len([<name>[,<occ>]]) : length match
11294 hdr_reg([<name>[,<occ>]]) : regex match
11295 hdr_sub([<name>[,<occ>]]) : substring match
11296
11297req.hdr_cnt([<name>]) : integer
11298hdr_cnt([<header>]) : integer (deprecated)
11299 Returns an integer value representing the number of occurrences of request
11300 header field name <name>, or the total number of header field values if
11301 <name> is not specified. It is important to remember that one header line may
11302 count as several headers if it has several values. The function considers any
11303 comma as a delimiter for distinct values. If full-line headers are desired
11304 instead, req.fhdr_cnt() should be used instead. With ACLs, it can be used to
11305 detect presence, absence or abuse of a specific header, as well as to block
11306 request smuggling attacks by rejecting requests which contain more than one
11307 of certain headers. See "req.hdr" for more information on header matching.
11308
11309req.hdr_ip([<name>[,<occ>]]) : ip
11310hdr_ip([<name>[,<occ>]]) : ip (deprecated)
11311 This extracts the last occurrence of header <name> in an HTTP request,
11312 converts it to an IPv4 or IPv6 address and returns this address. When used
11313 with ACLs, all occurrences are checked, and if <name> is omitted, every value
11314 of every header is checked. Optionally, a specific occurrence might be
11315 specified as a position number. Positive values indicate a position from the
11316 first occurrence, with 1 being the first one. Negative values indicate
11317 positions relative to the last one, with -1 being the last one. A typical use
11318 is with the X-Forwarded-For and X-Client-IP headers.
11319
11320req.hdr_val([<name>[,<occ>]]) : integer
11321hdr_val([<name>[,<occ>]]) : integer (deprecated)
11322 This extracts the last occurrence of header <name> in an HTTP request, and
11323 converts it to an integer value. When used with ACLs, all occurrences are
11324 checked, and if <name> is omitted, every value of every header is checked.
11325 Optionally, a specific occurrence might be specified as a position number.
11326 Positive values indicate a position from the first occurrence, with 1 being
11327 the first one. Negative values indicate positions relative to the last one,
11328 with -1 being the last one. A typical use is with the X-Forwarded-For header.
11329
11330http_auth(<userlist>) : boolean
11331 Returns a boolean indicating whether the authentication data received from
11332 the client match a username & password stored in the specified userlist. This
11333 fetch function is not really useful outside of ACLs. Currently only http
11334 basic auth is supported.
11335
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010011336http_auth_group(<userlist>) : string
11337 Returns a string corresponding to the user name found in the authentication
11338 data received from the client if both the user name and password are valid
11339 according to the specified userlist. The main purpose is to use it in ACLs
11340 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020011341 This fetch function is not really useful outside of ACLs. Currently only http
11342 basic auth is supported.
11343
11344 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010011345 http_auth_group(<userlist>) : group ...
11346 Returns true when the user extracted from the request and whose password is
11347 valid according to the specified userlist belongs to at least one of the
11348 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020011349
11350http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020011351 Returns true when the request being processed is the first one of the
11352 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020011353 from some requests when a request is not the first one, or to help grouping
11354 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020011355
Willy Tarreau74ca5042013-06-11 23:12:07 +020011356method : integer + string
11357 Returns an integer value corresponding to the method in the HTTP request. For
11358 example, "GET" equals 1 (check sources to establish the matching). Value 9
11359 means "other method" and may be converted to a string extracted from the
11360 stream. This should not be used directly as a sample, this is only meant to
11361 be used from ACLs, which transparently convert methods from patterns to these
11362 integer + string values. Some predefined ACL already check for most common
11363 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011364
Willy Tarreau74ca5042013-06-11 23:12:07 +020011365 ACL derivatives :
11366 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020011367
Willy Tarreau74ca5042013-06-11 23:12:07 +020011368 Example :
11369 # only accept GET and HEAD requests
11370 acl valid_method method GET HEAD
11371 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020011372
Willy Tarreau74ca5042013-06-11 23:12:07 +020011373path : string
11374 This extracts the request's URL path, which starts at the first slash and
11375 ends before the question mark (without the host part). A typical use is with
11376 prefetch-capable caches, and with portals which need to aggregate multiple
11377 information from databases and keep them in caches. Note that with outgoing
11378 caches, it would be wiser to use "url" instead. With ACLs, it's typically
11379 used to match exact file names (eg: "/login.php"), or directory parts using
11380 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011381
Willy Tarreau74ca5042013-06-11 23:12:07 +020011382 ACL derivatives :
11383 path : exact string match
11384 path_beg : prefix match
11385 path_dir : subdir match
11386 path_dom : domain match
11387 path_end : suffix match
11388 path_len : length match
11389 path_reg : regex match
11390 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020011391
Willy Tarreau74ca5042013-06-11 23:12:07 +020011392req.ver : string
11393req_ver : string (deprecated)
11394 Returns the version string from the HTTP request, for example "1.1". This can
11395 be useful for logs, but is mostly there for ACL. Some predefined ACL already
11396 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011397
Willy Tarreau74ca5042013-06-11 23:12:07 +020011398 ACL derivatives :
11399 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020011400
Willy Tarreau74ca5042013-06-11 23:12:07 +020011401res.comp : boolean
11402 Returns the boolean "true" value if the response has been compressed by
11403 HAProxy, otherwise returns boolean "false". This may be used to add
11404 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011405
Willy Tarreau74ca5042013-06-11 23:12:07 +020011406res.comp_algo : string
11407 Returns a string containing the name of the algorithm used if the response
11408 was compressed by HAProxy, for example : "deflate". This may be used to add
11409 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011410
Willy Tarreau74ca5042013-06-11 23:12:07 +020011411res.cook([<name>]) : string
11412scook([<name>]) : string (deprecated)
11413 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
11414 header line from the response, and returns its value as string. If no name is
11415 specified, the first cookie value is returned.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020011416
Willy Tarreau74ca5042013-06-11 23:12:07 +020011417 ACL derivatives :
11418 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020011419
Willy Tarreau74ca5042013-06-11 23:12:07 +020011420res.cook_cnt([<name>]) : integer
11421scook_cnt([<name>]) : integer (deprecated)
11422 Returns an integer value representing the number of occurrences of the cookie
11423 <name> in the response, or all cookies if <name> is not specified. This is
11424 mostly useful when combined with ACLs to detect suspicious responses.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011425
Willy Tarreau74ca5042013-06-11 23:12:07 +020011426res.cook_val([<name>]) : integer
11427scook_val([<name>]) : integer (deprecated)
11428 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
11429 header line from the response, and converts its value to an integer which is
11430 returned. If no name is specified, the first cookie value is returned.
Willy Tarreaud63335a2010-02-26 12:56:52 +010011431
Willy Tarreau74ca5042013-06-11 23:12:07 +020011432res.fhdr([<name>[,<occ>]]) : string
11433 This extracts the last occurrence of header <name> in an HTTP response, or of
11434 the last header if no <name> is specified. Optionally, a specific occurrence
11435 might be specified as a position number. Positive values indicate a position
11436 from the first occurrence, with 1 being the first one. Negative values
11437 indicate positions relative to the last one, with -1 being the last one. It
11438 differs from res.hdr() in that any commas present in the value are returned
11439 and are not used as delimiters. If this is not desired, the res.hdr() fetch
11440 should be used instead. This is sometimes useful with headers such as Date or
11441 Expires.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011442
Willy Tarreau74ca5042013-06-11 23:12:07 +020011443res.fhdr_cnt([<name>]) : integer
11444 Returns an integer value representing the number of occurrences of response
11445 header field name <name>, or the total number of header fields if <name> is
11446 not specified. Contrary to its res.hdr_cnt() cousin, this function returns
11447 the number of full line headers and does not stop on commas. If this is not
11448 desired, the res.hdr_cnt() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011449
Willy Tarreau74ca5042013-06-11 23:12:07 +020011450res.hdr([<name>[,<occ>]]) : string
11451shdr([<name>[,<occ>]]) : string (deprecated)
11452 This extracts the last occurrence of header <name> in an HTTP response, or of
11453 the last header if no <name> is specified. Optionally, a specific occurrence
11454 might be specified as a position number. Positive values indicate a position
11455 from the first occurrence, with 1 being the first one. Negative values
11456 indicate positions relative to the last one, with -1 being the last one. This
11457 can be useful to learn some data into a stick-table. The function considers
11458 any comma as a delimiter for distinct values. If this is not desired, the
11459 res.fhdr() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011460
Willy Tarreau74ca5042013-06-11 23:12:07 +020011461 ACL derivatives :
11462 shdr([<name>[,<occ>]]) : exact string match
11463 shdr_beg([<name>[,<occ>]]) : prefix match
11464 shdr_dir([<name>[,<occ>]]) : subdir match
11465 shdr_dom([<name>[,<occ>]]) : domain match
11466 shdr_end([<name>[,<occ>]]) : suffix match
11467 shdr_len([<name>[,<occ>]]) : length match
11468 shdr_reg([<name>[,<occ>]]) : regex match
11469 shdr_sub([<name>[,<occ>]]) : substring match
11470
11471res.hdr_cnt([<name>]) : integer
11472shdr_cnt([<name>]) : integer (deprecated)
11473 Returns an integer value representing the number of occurrences of response
11474 header field name <name>, or the total number of header fields if <name> is
11475 not specified. The function considers any comma as a delimiter for distinct
11476 values. If this is not desired, the res.fhdr_cnt() fetch should be used
11477 instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011478
Willy Tarreau74ca5042013-06-11 23:12:07 +020011479res.hdr_ip([<name>[,<occ>]]) : ip
11480shdr_ip([<name>[,<occ>]]) : ip (deprecated)
11481 This extracts the last occurrence of header <name> in an HTTP response,
11482 convert it to an IPv4 or IPv6 address and returns this address. Optionally, a
11483 specific occurrence might be specified as a position number. Positive values
11484 indicate a position from the first occurrence, with 1 being the first one.
11485 Negative values indicate positions relative to the last one, with -1 being
11486 the last one. This can be useful to learn some data into a stick table.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011487
Willy Tarreau74ca5042013-06-11 23:12:07 +020011488res.hdr_val([<name>[,<occ>]]) : integer
11489shdr_val([<name>[,<occ>]]) : integer (deprecated)
11490 This extracts the last occurrence of header <name> in an HTTP response, and
11491 converts it to an integer value. Optionally, a specific occurrence might be
11492 specified as a position number. Positive values indicate a position from the
11493 first occurrence, with 1 being the first one. Negative values indicate
11494 positions relative to the last one, with -1 being the last one. This can be
11495 useful to learn some data into a stick table.
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010011496
Willy Tarreau74ca5042013-06-11 23:12:07 +020011497res.ver : string
11498resp_ver : string (deprecated)
11499 Returns the version string from the HTTP response, for example "1.1". This
11500 can be useful for logs, but is mostly there for ACL.
Willy Tarreau0e698542011-09-16 08:32:32 +020011501
Willy Tarreau74ca5042013-06-11 23:12:07 +020011502 ACL derivatives :
11503 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010011504
Willy Tarreau74ca5042013-06-11 23:12:07 +020011505set-cookie([<name>]) : string (deprecated)
11506 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
11507 header line from the response and uses the corresponding value to match. This
11508 can be comparable to what "appsession" does with default options, but with
11509 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010011510
Willy Tarreau74ca5042013-06-11 23:12:07 +020011511 This fetch function is deprecated and has been superseded by the "res.cook"
11512 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010011513
Willy Tarreau74ca5042013-06-11 23:12:07 +020011514 See also : "appsession"
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011515
Willy Tarreau74ca5042013-06-11 23:12:07 +020011516status : integer
11517 Returns an integer containing the HTTP status code in the HTTP response, for
11518 example, 302. It is mostly used within ACLs and integer ranges, for example,
11519 to remove any Location header if the response is not a 3xx.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011520
Willy Tarreau74ca5042013-06-11 23:12:07 +020011521url : string
11522 This extracts the request's URL as presented in the request. A typical use is
11523 with prefetch-capable caches, and with portals which need to aggregate
11524 multiple information from databases and keep them in caches. With ACLs, using
11525 "path" is preferred over using "url", because clients may send a full URL as
11526 is normally done with proxies. The only real use is to match "*" which does
11527 not match in "path", and for which there is already a predefined ACL. See
11528 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011529
Willy Tarreau74ca5042013-06-11 23:12:07 +020011530 ACL derivatives :
11531 url : exact string match
11532 url_beg : prefix match
11533 url_dir : subdir match
11534 url_dom : domain match
11535 url_end : suffix match
11536 url_len : length match
11537 url_reg : regex match
11538 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011539
Willy Tarreau74ca5042013-06-11 23:12:07 +020011540url_ip : ip
11541 This extracts the IP address from the request's URL when the host part is
11542 presented as an IP address. Its use is very limited. For instance, a
11543 monitoring system might use this field as an alternative for the source IP in
11544 order to test what path a given source address would follow, or to force an
11545 entry in a table for a given source address. With ACLs it can be used to
11546 restrict access to certain systems through a proxy, for example when combined
11547 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011548
Willy Tarreau74ca5042013-06-11 23:12:07 +020011549url_port : integer
11550 This extracts the port part from the request's URL. Note that if the port is
11551 not specified in the request, port 80 is assumed. With ACLs it can be used to
11552 restrict access to certain systems through a proxy, for example when combined
11553 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011554
Willy Tarreau74ca5042013-06-11 23:12:07 +020011555urlp(<name>[,<delim>]) : string
11556url_param(<name>[,<delim>]) : string
11557 This extracts the first occurrence of the parameter <name> in the query
11558 string, which begins after either '?' or <delim>, and which ends before '&',
11559 ';' or <delim>. The parameter name is case-sensitive. The result is a string
11560 corresponding to the value of the parameter <name> as presented in the
11561 request (no URL decoding is performed). This can be used for session
11562 stickiness based on a client ID, to extract an application cookie passed as a
11563 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
11564 this fetch do not iterate over multiple parameters and stop at the first one
11565 as well.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011566
Willy Tarreau74ca5042013-06-11 23:12:07 +020011567 ACL derivatives :
11568 urlp(<name>[,<delim>]) : exact string match
11569 urlp_beg(<name>[,<delim>]) : prefix match
11570 urlp_dir(<name>[,<delim>]) : subdir match
11571 urlp_dom(<name>[,<delim>]) : domain match
11572 urlp_end(<name>[,<delim>]) : suffix match
11573 urlp_len(<name>[,<delim>]) : length match
11574 urlp_reg(<name>[,<delim>]) : regex match
11575 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011576
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011577
Willy Tarreau74ca5042013-06-11 23:12:07 +020011578 Example :
11579 # match http://example.com/foo?PHPSESSIONID=some_id
11580 stick on urlp(PHPSESSIONID)
11581 # match http://example.com/foo;JSESSIONID=some_id
11582 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020011583
Willy Tarreau74ca5042013-06-11 23:12:07 +020011584urlp_val(<name>[,<delim>]) : integer
11585 See "urlp" above. This one extracts the URL parameter <name> in the request
11586 and converts it to an integer value. This can be used for session stickiness
11587 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020011588
Willy Tarreau198a7442008-01-17 12:05:32 +010011589
Willy Tarreau74ca5042013-06-11 23:12:07 +0200115907.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011591---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010011592
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011593Some predefined ACLs are hard-coded so that they do not have to be declared in
11594every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020011595order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010011596
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011597ACL name Equivalent to Usage
11598---------------+-----------------------------+---------------------------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011599FALSE always_false never match
Willy Tarreau2492d5b2009-07-11 00:06:00 +020011600HTTP req_proto_http match if protocol is valid HTTP
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011601HTTP_1.0 req_ver 1.0 match HTTP version 1.0
11602HTTP_1.1 req_ver 1.1 match HTTP version 1.1
Willy Tarreaud63335a2010-02-26 12:56:52 +010011603HTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length
11604HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
11605HTTP_URL_SLASH url_beg / match URL beginning with "/"
11606HTTP_URL_STAR url * match URL equal to "*"
11607LOCALHOST src 127.0.0.1/8 match connection from local host
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011608METH_CONNECT method CONNECT match HTTP CONNECT method
11609METH_GET method GET HEAD match HTTP GET or HEAD method
11610METH_HEAD method HEAD match HTTP HEAD method
11611METH_OPTIONS method OPTIONS match HTTP OPTIONS method
11612METH_POST method POST match HTTP POST method
11613METH_TRACE method TRACE match HTTP TRACE method
Emeric Brunbede3d02009-06-30 17:54:00 +020011614RDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011615REQ_CONTENT req_len gt 0 match data in the request buffer
Willy Tarreaud63335a2010-02-26 12:56:52 +010011616TRUE always_true always match
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011617WAIT_END wait_end wait for end of content analysis
11618---------------+-----------------------------+---------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010011619
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011620
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200116218. Logging
11622----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010011623
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011624One of HAProxy's strong points certainly lies is its precise logs. It probably
11625provides the finest level of information available for such a product, which is
11626very important for troubleshooting complex environments. Standard information
11627provided in logs include client ports, TCP/HTTP state timers, precise session
11628state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011629to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011630headers.
11631
11632In order to improve administrators reactivity, it offers a great transparency
11633about encountered problems, both internal and external, and it is possible to
11634send logs to different sources at the same time with different level filters :
11635
11636 - global process-level logs (system errors, start/stop, etc..)
11637 - per-instance system and internal errors (lack of resource, bugs, ...)
11638 - per-instance external troubles (servers up/down, max connections)
11639 - per-instance activity (client connections), either at the establishment or
11640 at the termination.
11641
11642The ability to distribute different levels of logs to different log servers
11643allow several production teams to interact and to fix their problems as soon
11644as possible. For example, the system team might monitor system-wide errors,
11645while the application team might be monitoring the up/down for their servers in
11646real time, and the security team might analyze the activity logs with one hour
11647delay.
11648
11649
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200116508.1. Log levels
11651---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011652
Simon Hormandf791f52011-05-29 15:01:10 +090011653TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011654source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090011655HTTP request, HTTP return code, number of bytes transmitted, conditions
11656in which the session ended, and even exchanged cookies values. For example
11657track a particular user's problems. All messages may be sent to up to two
11658syslog servers. Check the "log" keyword in section 4.2 for more information
11659about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011660
11661
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200116628.2. Log formats
11663----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011664
William Lallemand48940402012-01-30 16:47:22 +010011665HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090011666and will be detailed in the following sections. A few of them may vary
11667slightly with the configuration, due to indicators specific to certain
11668options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011669
11670 - the default format, which is very basic and very rarely used. It only
11671 provides very basic information about the incoming connection at the moment
11672 it is accepted : source IP:port, destination IP:port, and frontend-name.
11673 This mode will eventually disappear so it will not be described to great
11674 extents.
11675
11676 - the TCP format, which is more advanced. This format is enabled when "option
11677 tcplog" is set on the frontend. HAProxy will then usually wait for the
11678 connection to terminate before logging. This format provides much richer
11679 information, such as timers, connection counts, queue size, etc... This
11680 format is recommended for pure TCP proxies.
11681
11682 - the HTTP format, which is the most advanced for HTTP proxying. This format
11683 is enabled when "option httplog" is set on the frontend. It provides the
11684 same information as the TCP format with some HTTP-specific fields such as
11685 the request, the status code, and captures of headers and cookies. This
11686 format is recommended for HTTP proxies.
11687
Emeric Brun3a058f32009-06-30 18:26:00 +020011688 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
11689 fields arranged in the same order as the CLF format. In this mode, all
11690 timers, captures, flags, etc... appear one per field after the end of the
11691 common fields, in the same order they appear in the standard HTTP format.
11692
William Lallemand48940402012-01-30 16:47:22 +010011693 - the custom log format, allows you to make your own log line.
11694
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011695Next sections will go deeper into details for each of these formats. Format
11696specification will be performed on a "field" basis. Unless stated otherwise, a
11697field is a portion of text delimited by any number of spaces. Since syslog
11698servers are susceptible of inserting fields at the beginning of a line, it is
11699always assumed that the first field is the one containing the process name and
11700identifier.
11701
11702Note : Since log lines may be quite long, the log examples in sections below
11703 might be broken into multiple lines. The example log lines will be
11704 prefixed with 3 closing angle brackets ('>>>') and each time a log is
11705 broken into multiple lines, each non-final line will end with a
11706 backslash ('\') and the next line will start indented by two characters.
11707
11708
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200117098.2.1. Default log format
11710-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011711
11712This format is used when no specific option is set. The log is emitted as soon
11713as the connection is accepted. One should note that this currently is the only
11714format which logs the request's destination IP and ports.
11715
11716 Example :
11717 listen www
11718 mode http
11719 log global
11720 server srv1 127.0.0.1:8000
11721
11722 >>> Feb 6 12:12:09 localhost \
11723 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
11724 (www/HTTP)
11725
11726 Field Format Extract from the example above
11727 1 process_name '[' pid ']:' haproxy[14385]:
11728 2 'Connect from' Connect from
11729 3 source_ip ':' source_port 10.0.1.2:33312
11730 4 'to' to
11731 5 destination_ip ':' destination_port 10.0.3.31:8012
11732 6 '(' frontend_name '/' mode ')' (www/HTTP)
11733
11734Detailed fields description :
11735 - "source_ip" is the IP address of the client which initiated the connection.
11736 - "source_port" is the TCP port of the client which initiated the connection.
11737 - "destination_ip" is the IP address the client connected to.
11738 - "destination_port" is the TCP port the client connected to.
11739 - "frontend_name" is the name of the frontend (or listener) which received
11740 and processed the connection.
11741 - "mode is the mode the frontend is operating (TCP or HTTP).
11742
Willy Tarreauceb24bc2010-11-09 12:46:41 +010011743In case of a UNIX socket, the source and destination addresses are marked as
11744"unix:" and the ports reflect the internal ID of the socket which accepted the
11745connection (the same ID as reported in the stats).
11746
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011747It is advised not to use this deprecated format for newer installations as it
11748will eventually disappear.
11749
11750
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200117518.2.2. TCP log format
11752---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011753
11754The TCP format is used when "option tcplog" is specified in the frontend, and
11755is the recommended format for pure TCP proxies. It provides a lot of precious
11756information for troubleshooting. Since this format includes timers and byte
11757counts, the log is normally emitted at the end of the session. It can be
11758emitted earlier if "option logasap" is specified, which makes sense in most
11759environments with long sessions such as remote terminals. Sessions which match
11760the "monitor" rules are never logged. It is also possible not to emit logs for
11761sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020011762specifying "option dontlognull" in the frontend. Successful connections will
11763not be logged if "option dontlog-normal" is specified in the frontend. A few
11764fields may slightly vary depending on some configuration options, those are
11765marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011766
11767 Example :
11768 frontend fnt
11769 mode tcp
11770 option tcplog
11771 log global
11772 default_backend bck
11773
11774 backend bck
11775 server srv1 127.0.0.1:8000
11776
11777 >>> Feb 6 12:12:56 localhost \
11778 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
11779 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
11780
11781 Field Format Extract from the example above
11782 1 process_name '[' pid ']:' haproxy[14387]:
11783 2 client_ip ':' client_port 10.0.1.2:33313
11784 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
11785 4 frontend_name fnt
11786 5 backend_name '/' server_name bck/srv1
11787 6 Tw '/' Tc '/' Tt* 0/0/5007
11788 7 bytes_read* 212
11789 8 termination_state --
11790 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
11791 10 srv_queue '/' backend_queue 0/0
11792
11793Detailed fields description :
11794 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010011795 connection to haproxy. If the connection was accepted on a UNIX socket
11796 instead, the IP address would be replaced with the word "unix". Note that
11797 when the connection is accepted on a socket configured with "accept-proxy"
11798 and the PROXY protocol is correctly used, then the logs will reflect the
11799 forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011800
11801 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010011802 If the connection was accepted on a UNIX socket instead, the port would be
11803 replaced with the ID of the accepting socket, which is also reported in the
11804 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011805
11806 - "accept_date" is the exact date when the connection was received by haproxy
11807 (which might be very slightly different from the date observed on the
11808 network if there was some queuing in the system's backlog). This is usually
11809 the same date which may appear in any upstream firewall's log.
11810
11811 - "frontend_name" is the name of the frontend (or listener) which received
11812 and processed the connection.
11813
11814 - "backend_name" is the name of the backend (or listener) which was selected
11815 to manage the connection to the server. This will be the same as the
11816 frontend if no switching rule has been applied, which is common for TCP
11817 applications.
11818
11819 - "server_name" is the name of the last server to which the connection was
11820 sent, which might differ from the first one if there were connection errors
11821 and a redispatch occurred. Note that this server belongs to the backend
11822 which processed the request. If the connection was aborted before reaching
11823 a server, "<NOSRV>" is indicated instead of a server name.
11824
11825 - "Tw" is the total time in milliseconds spent waiting in the various queues.
11826 It can be "-1" if the connection was aborted before reaching the queue.
11827 See "Timers" below for more details.
11828
11829 - "Tc" is the total time in milliseconds spent waiting for the connection to
11830 establish to the final server, including retries. It can be "-1" if the
11831 connection was aborted before a connection could be established. See
11832 "Timers" below for more details.
11833
11834 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011835 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011836 "option logasap" was specified, then the time counting stops at the moment
11837 the log is emitted. In this case, a '+' sign is prepended before the value,
11838 indicating that the final one will be larger. See "Timers" below for more
11839 details.
11840
11841 - "bytes_read" is the total number of bytes transmitted from the server to
11842 the client when the log is emitted. If "option logasap" is specified, the
11843 this value will be prefixed with a '+' sign indicating that the final one
11844 may be larger. Please note that this value is a 64-bit counter, so log
11845 analysis tools must be able to handle it without overflowing.
11846
11847 - "termination_state" is the condition the session was in when the session
11848 ended. This indicates the session state, which side caused the end of
11849 session to happen, and for what reason (timeout, error, ...). The normal
11850 flags should be "--", indicating the session was closed by either end with
11851 no data remaining in buffers. See below "Session state at disconnection"
11852 for more details.
11853
11854 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040011855 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011856 limits have been reached. For instance, if actconn is close to 512 when
11857 multiple connection errors occur, chances are high that the system limits
11858 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011859 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011860
11861 - "feconn" is the total number of concurrent connections on the frontend when
11862 the session was logged. It is useful to estimate the amount of resource
11863 required to sustain high loads, and to detect when the frontend's "maxconn"
11864 has been reached. Most often when this value increases by huge jumps, it is
11865 because there is congestion on the backend servers, but sometimes it can be
11866 caused by a denial of service attack.
11867
11868 - "beconn" is the total number of concurrent connections handled by the
11869 backend when the session was logged. It includes the total number of
11870 concurrent connections active on servers as well as the number of
11871 connections pending in queues. It is useful to estimate the amount of
11872 additional servers needed to support high loads for a given application.
11873 Most often when this value increases by huge jumps, it is because there is
11874 congestion on the backend servers, but sometimes it can be caused by a
11875 denial of service attack.
11876
11877 - "srv_conn" is the total number of concurrent connections still active on
11878 the server when the session was logged. It can never exceed the server's
11879 configured "maxconn" parameter. If this value is very often close or equal
11880 to the server's "maxconn", it means that traffic regulation is involved a
11881 lot, meaning that either the server's maxconn value is too low, or that
11882 there aren't enough servers to process the load with an optimal response
11883 time. When only one of the server's "srv_conn" is high, it usually means
11884 that this server has some trouble causing the connections to take longer to
11885 be processed than on other servers.
11886
11887 - "retries" is the number of connection retries experienced by this session
11888 when trying to connect to the server. It must normally be zero, unless a
11889 server is being stopped at the same moment the connection was attempted.
11890 Frequent retries generally indicate either a network problem between
11891 haproxy and the server, or a misconfigured system backlog on the server
11892 preventing new connections from being queued. This field may optionally be
11893 prefixed with a '+' sign, indicating that the session has experienced a
11894 redispatch after the maximal retry count has been reached on the initial
11895 server. In this case, the server name appearing in the log is the one the
11896 connection was redispatched to, and not the first one, though both may
11897 sometimes be the same in case of hashing for instance. So as a general rule
11898 of thumb, when a '+' is present in front of the retry count, this count
11899 should not be attributed to the logged server.
11900
11901 - "srv_queue" is the total number of requests which were processed before
11902 this one in the server queue. It is zero when the request has not gone
11903 through the server queue. It makes it possible to estimate the approximate
11904 server's response time by dividing the time spent in queue by the number of
11905 requests in the queue. It is worth noting that if a session experiences a
11906 redispatch and passes through two server queues, their positions will be
11907 cumulated. A request should not pass through both the server queue and the
11908 backend queue unless a redispatch occurs.
11909
11910 - "backend_queue" is the total number of requests which were processed before
11911 this one in the backend's global queue. It is zero when the request has not
11912 gone through the global queue. It makes it possible to estimate the average
11913 queue length, which easily translates into a number of missing servers when
11914 divided by a server's "maxconn" parameter. It is worth noting that if a
11915 session experiences a redispatch, it may pass twice in the backend's queue,
11916 and then both positions will be cumulated. A request should not pass
11917 through both the server queue and the backend queue unless a redispatch
11918 occurs.
11919
11920
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200119218.2.3. HTTP log format
11922----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011923
11924The HTTP format is the most complete and the best suited for HTTP proxies. It
11925is enabled by when "option httplog" is specified in the frontend. It provides
11926the same level of information as the TCP format with additional features which
11927are specific to the HTTP protocol. Just like the TCP format, the log is usually
11928emitted at the end of the session, unless "option logasap" is specified, which
11929generally only makes sense for download sites. A session which matches the
11930"monitor" rules will never logged. It is also possible not to log sessions for
11931which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020011932frontend. Successful connections will not be logged if "option dontlog-normal"
11933is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011934
11935Most fields are shared with the TCP log, some being different. A few fields may
11936slightly vary depending on some configuration options. Those ones are marked
11937with a star ('*') after the field name below.
11938
11939 Example :
11940 frontend http-in
11941 mode http
11942 option httplog
11943 log global
11944 default_backend bck
11945
11946 backend static
11947 server srv1 127.0.0.1:8000
11948
11949 >>> Feb 6 12:14:14 localhost \
11950 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
11951 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 +010011952 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011953
11954 Field Format Extract from the example above
11955 1 process_name '[' pid ']:' haproxy[14389]:
11956 2 client_ip ':' client_port 10.0.1.2:33317
11957 3 '[' accept_date ']' [06/Feb/2009:12:14:14.655]
11958 4 frontend_name http-in
11959 5 backend_name '/' server_name static/srv1
11960 6 Tq '/' Tw '/' Tc '/' Tr '/' Tt* 10/0/30/69/109
11961 7 status_code 200
11962 8 bytes_read* 2750
11963 9 captured_request_cookie -
11964 10 captured_response_cookie -
11965 11 termination_state ----
11966 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
11967 13 srv_queue '/' backend_queue 0/0
11968 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
11969 15 '{' captured_response_headers* '}' {}
11970 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010011971
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011972
11973Detailed fields description :
11974 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010011975 connection to haproxy. If the connection was accepted on a UNIX socket
11976 instead, the IP address would be replaced with the word "unix". Note that
11977 when the connection is accepted on a socket configured with "accept-proxy"
11978 and the PROXY protocol is correctly used, then the logs will reflect the
11979 forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011980
11981 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010011982 If the connection was accepted on a UNIX socket instead, the port would be
11983 replaced with the ID of the accepting socket, which is also reported in the
11984 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010011985
11986 - "accept_date" is the exact date when the TCP connection was received by
11987 haproxy (which might be very slightly different from the date observed on
11988 the network if there was some queuing in the system's backlog). This is
11989 usually the same date which may appear in any upstream firewall's log. This
11990 does not depend on the fact that the client has sent the request or not.
11991
11992 - "frontend_name" is the name of the frontend (or listener) which received
11993 and processed the connection.
11994
11995 - "backend_name" is the name of the backend (or listener) which was selected
11996 to manage the connection to the server. This will be the same as the
11997 frontend if no switching rule has been applied.
11998
11999 - "server_name" is the name of the last server to which the connection was
12000 sent, which might differ from the first one if there were connection errors
12001 and a redispatch occurred. Note that this server belongs to the backend
12002 which processed the request. If the request was aborted before reaching a
12003 server, "<NOSRV>" is indicated instead of a server name. If the request was
12004 intercepted by the stats subsystem, "<STATS>" is indicated instead.
12005
12006 - "Tq" is the total time in milliseconds spent waiting for the client to send
12007 a full HTTP request, not counting data. It can be "-1" if the connection
12008 was aborted before a complete request could be received. It should always
12009 be very small because a request generally fits in one single packet. Large
12010 times here generally indicate network trouble between the client and
12011 haproxy. See "Timers" below for more details.
12012
12013 - "Tw" is the total time in milliseconds spent waiting in the various queues.
12014 It can be "-1" if the connection was aborted before reaching the queue.
12015 See "Timers" below for more details.
12016
12017 - "Tc" is the total time in milliseconds spent waiting for the connection to
12018 establish to the final server, including retries. It can be "-1" if the
12019 request was aborted before a connection could be established. See "Timers"
12020 below for more details.
12021
12022 - "Tr" is the total time in milliseconds spent waiting for the server to send
12023 a full HTTP response, not counting data. It can be "-1" if the request was
12024 aborted before a complete response could be received. It generally matches
12025 the server's processing time for the request, though it may be altered by
12026 the amount of data sent by the client to the server. Large times here on
12027 "GET" requests generally indicate an overloaded server. See "Timers" below
12028 for more details.
12029
12030 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012031 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012032 "option logasap" was specified, then the time counting stops at the moment
12033 the log is emitted. In this case, a '+' sign is prepended before the value,
12034 indicating that the final one will be larger. See "Timers" below for more
12035 details.
12036
12037 - "status_code" is the HTTP status code returned to the client. This status
12038 is generally set by the server, but it might also be set by haproxy when
12039 the server cannot be reached or when its response is blocked by haproxy.
12040
12041 - "bytes_read" is the total number of bytes transmitted to the client when
12042 the log is emitted. This does include HTTP headers. If "option logasap" is
12043 specified, the this value will be prefixed with a '+' sign indicating that
12044 the final one may be larger. Please note that this value is a 64-bit
12045 counter, so log analysis tools must be able to handle it without
12046 overflowing.
12047
12048 - "captured_request_cookie" is an optional "name=value" entry indicating that
12049 the client had this cookie in the request. The cookie name and its maximum
12050 length are defined by the "capture cookie" statement in the frontend
12051 configuration. The field is a single dash ('-') when the option is not
12052 set. Only one cookie may be captured, it is generally used to track session
12053 ID exchanges between a client and a server to detect session crossing
12054 between clients due to application bugs. For more details, please consult
12055 the section "Capturing HTTP headers and cookies" below.
12056
12057 - "captured_response_cookie" is an optional "name=value" entry indicating
12058 that the server has returned a cookie with its response. The cookie name
12059 and its maximum length are defined by the "capture cookie" statement in the
12060 frontend configuration. The field is a single dash ('-') when the option is
12061 not set. Only one cookie may be captured, it is generally used to track
12062 session ID exchanges between a client and a server to detect session
12063 crossing between clients due to application bugs. For more details, please
12064 consult the section "Capturing HTTP headers and cookies" below.
12065
12066 - "termination_state" is the condition the session was in when the session
12067 ended. This indicates the session state, which side caused the end of
12068 session to happen, for what reason (timeout, error, ...), just like in TCP
12069 logs, and information about persistence operations on cookies in the last
12070 two characters. The normal flags should begin with "--", indicating the
12071 session was closed by either end with no data remaining in buffers. See
12072 below "Session state at disconnection" for more details.
12073
12074 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012075 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012076 limits have been reached. For instance, if actconn is close to 512 or 1024
12077 when multiple connection errors occur, chances are high that the system
12078 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012079 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012080 system.
12081
12082 - "feconn" is the total number of concurrent connections on the frontend when
12083 the session was logged. It is useful to estimate the amount of resource
12084 required to sustain high loads, and to detect when the frontend's "maxconn"
12085 has been reached. Most often when this value increases by huge jumps, it is
12086 because there is congestion on the backend servers, but sometimes it can be
12087 caused by a denial of service attack.
12088
12089 - "beconn" is the total number of concurrent connections handled by the
12090 backend when the session was logged. It includes the total number of
12091 concurrent connections active on servers as well as the number of
12092 connections pending in queues. It is useful to estimate the amount of
12093 additional servers needed to support high loads for a given application.
12094 Most often when this value increases by huge jumps, it is because there is
12095 congestion on the backend servers, but sometimes it can be caused by a
12096 denial of service attack.
12097
12098 - "srv_conn" is the total number of concurrent connections still active on
12099 the server when the session was logged. It can never exceed the server's
12100 configured "maxconn" parameter. If this value is very often close or equal
12101 to the server's "maxconn", it means that traffic regulation is involved a
12102 lot, meaning that either the server's maxconn value is too low, or that
12103 there aren't enough servers to process the load with an optimal response
12104 time. When only one of the server's "srv_conn" is high, it usually means
12105 that this server has some trouble causing the requests to take longer to be
12106 processed than on other servers.
12107
12108 - "retries" is the number of connection retries experienced by this session
12109 when trying to connect to the server. It must normally be zero, unless a
12110 server is being stopped at the same moment the connection was attempted.
12111 Frequent retries generally indicate either a network problem between
12112 haproxy and the server, or a misconfigured system backlog on the server
12113 preventing new connections from being queued. This field may optionally be
12114 prefixed with a '+' sign, indicating that the session has experienced a
12115 redispatch after the maximal retry count has been reached on the initial
12116 server. In this case, the server name appearing in the log is the one the
12117 connection was redispatched to, and not the first one, though both may
12118 sometimes be the same in case of hashing for instance. So as a general rule
12119 of thumb, when a '+' is present in front of the retry count, this count
12120 should not be attributed to the logged server.
12121
12122 - "srv_queue" is the total number of requests which were processed before
12123 this one in the server queue. It is zero when the request has not gone
12124 through the server queue. It makes it possible to estimate the approximate
12125 server's response time by dividing the time spent in queue by the number of
12126 requests in the queue. It is worth noting that if a session experiences a
12127 redispatch and passes through two server queues, their positions will be
12128 cumulated. A request should not pass through both the server queue and the
12129 backend queue unless a redispatch occurs.
12130
12131 - "backend_queue" is the total number of requests which were processed before
12132 this one in the backend's global queue. It is zero when the request has not
12133 gone through the global queue. It makes it possible to estimate the average
12134 queue length, which easily translates into a number of missing servers when
12135 divided by a server's "maxconn" parameter. It is worth noting that if a
12136 session experiences a redispatch, it may pass twice in the backend's queue,
12137 and then both positions will be cumulated. A request should not pass
12138 through both the server queue and the backend queue unless a redispatch
12139 occurs.
12140
12141 - "captured_request_headers" is a list of headers captured in the request due
12142 to the presence of the "capture request header" statement in the frontend.
12143 Multiple headers can be captured, they will be delimited by a vertical bar
12144 ('|'). When no capture is enabled, the braces do not appear, causing a
12145 shift of remaining fields. It is important to note that this field may
12146 contain spaces, and that using it requires a smarter log parser than when
12147 it's not used. Please consult the section "Capturing HTTP headers and
12148 cookies" below for more details.
12149
12150 - "captured_response_headers" is a list of headers captured in the response
12151 due to the presence of the "capture response header" statement in the
12152 frontend. Multiple headers can be captured, they will be delimited by a
12153 vertical bar ('|'). When no capture is enabled, the braces do not appear,
12154 causing a shift of remaining fields. It is important to note that this
12155 field may contain spaces, and that using it requires a smarter log parser
12156 than when it's not used. Please consult the section "Capturing HTTP headers
12157 and cookies" below for more details.
12158
12159 - "http_request" is the complete HTTP request line, including the method,
12160 request and HTTP version string. Non-printable characters are encoded (see
12161 below the section "Non-printable characters"). This is always the last
12162 field, and it is always delimited by quotes and is the only one which can
12163 contain quotes. If new fields are added to the log format, they will be
12164 added before this field. This field might be truncated if the request is
12165 huge and does not fit in the standard syslog buffer (1024 characters). This
12166 is the reason why this field must always remain the last one.
12167
12168
Cyril Bontédc4d9032012-04-08 21:57:39 +0200121698.2.4. Custom log format
12170------------------------
William Lallemand48940402012-01-30 16:47:22 +010012171
Willy Tarreau2beef582012-12-20 17:22:52 +010012172The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010012173mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010012174
12175HAproxy understands some log format variables. % precedes log format variables.
12176Variables can take arguments using braces ('{}'), and multiple arguments are
12177separated by commas within the braces. Flags may be added or removed by
12178prefixing them with a '+' or '-' sign.
12179
12180Special variable "%o" may be used to propagate its flags to all other
12181variables on the same format string. This is particularly handy with quoted
12182string formats ("Q").
12183
Willy Tarreauc8368452012-12-21 00:09:23 +010012184If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020012185as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010012186less common information such as the client's SSL certificate's DN, or to log
12187the key that would be used to store an entry into a stick table.
12188
William Lallemand48940402012-01-30 16:47:22 +010012189Note: spaces must be escaped. A space character is considered as a separator.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012190In order to emit a verbatim '%', it must be preceded by another '%' resulting
Willy Tarreau06d97f92013-12-02 17:45:48 +010012191in '%%'. HAProxy will automatically merge consecutive separators.
William Lallemand48940402012-01-30 16:47:22 +010012192
12193Flags are :
12194 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012195 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
William Lallemand48940402012-01-30 16:47:22 +010012196
12197 Example:
12198
12199 log-format %T\ %t\ Some\ Text
12200 log-format %{+Q}o\ %t\ %s\ %{-Q}r
12201
12202At the moment, the default HTTP format is defined this way :
12203
Willy Tarreau2beef582012-12-20 17:22:52 +010012204 log-format %ci:%cp\ [%t]\ %ft\ %b/%s\ %Tq/%Tw/%Tc/%Tr/%Tt\ %ST\ %B\ %CC\ \
12205 %CS\ %tsc\ %ac/%fc/%bc/%sc/%rc\ %sq/%bq\ %hr\ %hs\ %{+Q}r
William Lallemand48940402012-01-30 16:47:22 +010012206
William Lallemandbddd4fd2012-02-27 11:23:10 +010012207the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010012208
Willy Tarreau2beef582012-12-20 17:22:52 +010012209 log-format %{+Q}o\ %{-Q}ci\ -\ -\ [%T]\ %r\ %ST\ %B\ \"\"\ \"\"\ %cp\ \
Willy Tarreau773d65f2012-10-12 14:56:11 +020012210 %ms\ %ft\ %b\ %s\ \%Tq\ %Tw\ %Tc\ %Tr\ %Tt\ %tsc\ %ac\ %fc\ \
Willy Tarreau2beef582012-12-20 17:22:52 +010012211 %bc\ %sc\ %rc\ %sq\ %bq\ %CC\ %CS\ \%hrl\ %hsl
William Lallemand48940402012-01-30 16:47:22 +010012212
William Lallemandbddd4fd2012-02-27 11:23:10 +010012213and the default TCP format is defined this way :
12214
Willy Tarreau2beef582012-12-20 17:22:52 +010012215 log-format %ci:%cp\ [%t]\ %ft\ %b/%s\ %Tw/%Tc/%Tt\ %B\ %ts\ \
William Lallemandbddd4fd2012-02-27 11:23:10 +010012216 %ac/%fc/%bc/%sc/%rc\ %sq/%bq
12217
William Lallemand48940402012-01-30 16:47:22 +010012218Please refer to the table below for currently defined variables :
12219
William Lallemandbddd4fd2012-02-27 11:23:10 +010012220 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012221 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012222 +---+------+-----------------------------------------------+-------------+
12223 | | %o | special variable, apply flags on all next var | |
12224 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010012225 | | %B | bytes_read (from server to client) | numeric |
12226 | H | %CC | captured_request_cookie | string |
12227 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020012228 | | %H | hostname | string |
William Lallemanda73203e2012-03-12 12:48:57 +010012229 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020012230 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020012231 | | %T | gmt_date_time | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012232 | | %Tc | Tc | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080012233 | | %Tl | local_date_time | date |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012234 | H | %Tq | Tq | numeric |
12235 | H | %Tr | Tr | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020012236 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012237 | | %Tt | Tt | numeric |
12238 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010012239 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012240 | | %ac | actconn | numeric |
12241 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010012242 | | %bc | beconn (backend concurrent connections) | numeric |
12243 | | %bi | backend_source_ip (connecting address) | IP |
12244 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012245 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010012246 | | %ci | client_ip (accepted address) | IP |
12247 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012248 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010012249 | | %fc | feconn (frontend concurrent connections) | numeric |
12250 | | %fi | frontend_ip (accepting address) | IP |
12251 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020012252 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020012253 | | %hr | captured_request_headers default style | string |
12254 | | %hrl | captured_request_headers CLF style | string list |
12255 | | %hs | captured_response_headers default style | string |
12256 | | %hsl | captured_response_headers CLF style | string list |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012257 | | %ms | accept date milliseconds | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020012258 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012259 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012260 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010012261 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012262 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010012263 | | %sc | srv_conn (server concurrent connections) | numeric |
12264 | | %si | server_IP (target address) | IP |
12265 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012266 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012267 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
12268 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010012269 | | %t | date_time (with millisecond resolution) | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012270 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012271 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010012272 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010012273
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020012274 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010012275
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010012276
122778.2.5. Error log format
12278-----------------------
12279
12280When an incoming connection fails due to an SSL handshake or an invalid PROXY
12281protocol header, haproxy will log the event using a shorter, fixed line format.
12282By default, logs are emitted at the LOG_INFO level, unless the option
12283"log-separate-errors" is set in the backend, in which case the LOG_ERR level
12284will be used. Connections on which no data are exchanged (eg: probes) are not
12285logged if the "dontlognull" option is set.
12286
12287The format looks like this :
12288
12289 >>> Dec 3 18:27:14 localhost \
12290 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
12291 Connection error during SSL handshake
12292
12293 Field Format Extract from the example above
12294 1 process_name '[' pid ']:' haproxy[6103]:
12295 2 client_ip ':' client_port 127.0.0.1:56059
12296 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
12297 4 frontend_name "/" bind_name ":" frt/f1:
12298 5 message Connection error during SSL handshake
12299
12300These fields just provide minimal information to help debugging connection
12301failures.
12302
12303
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123048.3. Advanced logging options
12305-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012306
12307Some advanced logging options are often looked for but are not easy to find out
12308just by looking at the various options. Here is an entry point for the few
12309options which can enable better logging. Please refer to the keywords reference
12310for more information about their usage.
12311
12312
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123138.3.1. Disabling logging of external tests
12314------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012315
12316It is quite common to have some monitoring tools perform health checks on
12317haproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
12318commercial load-balancer, and sometimes it will simply be a more complete
12319monitoring system such as Nagios. When the tests are very frequent, users often
12320ask how to disable logging for those checks. There are three possibilities :
12321
12322 - if connections come from everywhere and are just TCP probes, it is often
12323 desired to simply disable logging of connections without data exchange, by
12324 setting "option dontlognull" in the frontend. It also disables logging of
12325 port scans, which may or may not be desired.
12326
12327 - if the connection come from a known source network, use "monitor-net" to
12328 declare this network as monitoring only. Any host in this network will then
12329 only be able to perform health checks, and their requests will not be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012330 logged. This is generally appropriate to designate a list of equipment
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012331 such as other load-balancers.
12332
12333 - if the tests are performed on a known URI, use "monitor-uri" to declare
12334 this URI as dedicated to monitoring. Any host sending this request will
12335 only get the result of a health-check, and the request will not be logged.
12336
12337
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123388.3.2. Logging before waiting for the session to terminate
12339----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012340
12341The problem with logging at end of connection is that you have no clue about
12342what is happening during very long sessions, such as remote terminal sessions
12343or large file downloads. This problem can be worked around by specifying
12344"option logasap" in the frontend. Haproxy will then log as soon as possible,
12345just before data transfer begins. This means that in case of TCP, it will still
12346log the connection status to the server, and in case of HTTP, it will log just
12347after processing the server headers. In this case, the number of bytes reported
12348is the number of header bytes sent to the client. In order to avoid confusion
12349with normal logs, the total time field and the number of bytes are prefixed
12350with a '+' sign which means that real numbers are certainly larger.
12351
12352
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123538.3.3. Raising log level upon errors
12354------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020012355
12356Sometimes it is more convenient to separate normal traffic from errors logs,
12357for instance in order to ease error monitoring from log files. When the option
12358"log-separate-errors" is used, connections which experience errors, timeouts,
12359retries, redispatches or HTTP status codes 5xx will see their syslog level
12360raised from "info" to "err". This will help a syslog daemon store the log in
12361a separate file. It is very important to keep the errors in the normal traffic
12362file too, so that log ordering is not altered. You should also be careful if
12363you already have configured your syslog daemon to store all logs higher than
12364"notice" in an "admin" file, because the "err" level is higher than "notice".
12365
12366
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123678.3.4. Disabling logging of successful connections
12368--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020012369
12370Although this may sound strange at first, some large sites have to deal with
12371multiple thousands of logs per second and are experiencing difficulties keeping
12372them intact for a long time or detecting errors within them. If the option
12373"dontlog-normal" is set on the frontend, all normal connections will not be
12374logged. In this regard, a normal connection is defined as one without any
12375error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
12376and a response with a status 5xx is not considered normal and will be logged
12377too. Of course, doing is is really discouraged as it will remove most of the
12378useful information from the logs. Do this only if you have no other
12379alternative.
12380
12381
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123828.4. Timing events
12383------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012384
12385Timers provide a great help in troubleshooting network problems. All values are
12386reported in milliseconds (ms). These timers should be used in conjunction with
12387the session termination flags. In TCP mode with "option tcplog" set on the
12388frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
12389mode, 5 control points are reported under the form "Tq/Tw/Tc/Tr/Tt" :
12390
12391 - Tq: total time to get the client request (HTTP mode only). It's the time
12392 elapsed between the moment the client connection was accepted and the
12393 moment the proxy received the last HTTP header. The value "-1" indicates
12394 that the end of headers (empty line) has never been seen. This happens when
12395 the client closes prematurely or times out.
12396
12397 - Tw: total time spent in the queues waiting for a connection slot. It
12398 accounts for backend queue as well as the server queues, and depends on the
12399 queue size, and the time needed for the server to complete previous
12400 requests. The value "-1" means that the request was killed before reaching
12401 the queue, which is generally what happens with invalid or denied requests.
12402
12403 - Tc: total time to establish the TCP connection to the server. It's the time
12404 elapsed between the moment the proxy sent the connection request, and the
12405 moment it was acknowledged by the server, or between the TCP SYN packet and
12406 the matching SYN/ACK packet in return. The value "-1" means that the
12407 connection never established.
12408
12409 - Tr: server response time (HTTP mode only). It's the time elapsed between
12410 the moment the TCP connection was established to the server and the moment
12411 the server sent its complete response headers. It purely shows its request
12412 processing time, without the network overhead due to the data transmission.
12413 It is worth noting that when the client has data to send to the server, for
12414 instance during a POST request, the time already runs, and this can distort
12415 apparent response time. For this reason, it's generally wise not to trust
12416 too much this field for POST requests initiated from clients behind an
12417 untrusted network. A value of "-1" here means that the last the response
12418 header (empty line) was never seen, most likely because the server timeout
12419 stroke before the server managed to process the request.
12420
12421 - Tt: total session duration time, between the moment the proxy accepted it
12422 and the moment both ends were closed. The exception is when the "logasap"
12423 option is specified. In this case, it only equals (Tq+Tw+Tc+Tr), and is
12424 prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012425 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012426
12427 Td = Tt - (Tq + Tw + Tc + Tr)
12428
12429 Timers with "-1" values have to be excluded from this equation. In TCP
12430 mode, "Tq" and "Tr" have to be excluded too. Note that "Tt" can never be
12431 negative.
12432
12433These timers provide precious indications on trouble causes. Since the TCP
12434protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
12435that timers close to multiples of 3s are nearly always related to lost packets
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012436due to network problems (wires, negotiation, congestion). Moreover, if "Tt" is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012437close to a timeout value specified in the configuration, it often means that a
12438session has been aborted on timeout.
12439
12440Most common cases :
12441
12442 - If "Tq" is close to 3000, a packet has probably been lost between the
12443 client and the proxy. This is very rare on local networks but might happen
12444 when clients are on far remote networks and send large requests. It may
12445 happen that values larger than usual appear here without any network cause.
12446 Sometimes, during an attack or just after a resource starvation has ended,
12447 haproxy may accept thousands of connections in a few milliseconds. The time
12448 spent accepting these connections will inevitably slightly delay processing
12449 of other connections, and it can happen that request times in the order of
12450 a few tens of milliseconds are measured after a few thousands of new
Patrick Mezard105faca2010-06-12 17:02:46 +020012451 connections have been accepted at once. Setting "option http-server-close"
12452 may display larger request times since "Tq" also measures the time spent
12453 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012454
12455 - If "Tc" is close to 3000, a packet has probably been lost between the
12456 server and the proxy during the server connection phase. This value should
12457 always be very low, such as 1 ms on local networks and less than a few tens
12458 of ms on remote networks.
12459
Willy Tarreau55165fe2009-05-10 12:02:55 +020012460 - If "Tr" is nearly always lower than 3000 except some rare values which seem
12461 to be the average majored by 3000, there are probably some packets lost
12462 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012463
12464 - If "Tt" is large even for small byte counts, it generally is because
12465 neither the client nor the server decides to close the connection, for
12466 instance because both have agreed on a keep-alive connection mode. In order
12467 to solve this issue, it will be needed to specify "option httpclose" on
12468 either the frontend or the backend. If the problem persists, it means that
12469 the server ignores the "close" connection mode and expects the client to
12470 close. Then it will be required to use "option forceclose". Having the
12471 smallest possible 'Tt' is important when connection regulation is used with
12472 the "maxconn" option on the servers, since no new connection will be sent
12473 to the server until another one is released.
12474
12475Other noticeable HTTP log cases ('xx' means any value to be ignored) :
12476
12477 Tq/Tw/Tc/Tr/+Tt The "option logasap" is present on the frontend and the log
12478 was emitted before the data phase. All the timers are valid
12479 except "Tt" which is shorter than reality.
12480
12481 -1/xx/xx/xx/Tt The client was not able to send a complete request in time
12482 or it aborted too early. Check the session termination flags
12483 then "timeout http-request" and "timeout client" settings.
12484
12485 Tq/-1/xx/xx/Tt It was not possible to process the request, maybe because
12486 servers were out of order, because the request was invalid
12487 or forbidden by ACL rules. Check the session termination
12488 flags.
12489
12490 Tq/Tw/-1/xx/Tt The connection could not establish on the server. Either it
12491 actively refused it or it timed out after Tt-(Tq+Tw) ms.
12492 Check the session termination flags, then check the
12493 "timeout connect" setting. Note that the tarpit action might
12494 return similar-looking patterns, with "Tw" equal to the time
12495 the client connection was maintained open.
12496
12497 Tq/Tw/Tc/-1/Tt The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012498 a complete response in time, or it closed its connection
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012499 unexpectedly after Tt-(Tq+Tw+Tc) ms. Check the session
12500 termination flags, then check the "timeout server" setting.
12501
12502
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200125038.5. Session state at disconnection
12504-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012505
12506TCP and HTTP logs provide a session termination indicator in the
12507"termination_state" field, just before the number of active connections. It is
125082-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
12509each of which has a special meaning :
12510
12511 - On the first character, a code reporting the first event which caused the
12512 session to terminate :
12513
12514 C : the TCP session was unexpectedly aborted by the client.
12515
12516 S : the TCP session was unexpectedly aborted by the server, or the
12517 server explicitly refused it.
12518
12519 P : the session was prematurely aborted by the proxy, because of a
12520 connection limit enforcement, because a DENY filter was matched,
12521 because of a security check which detected and blocked a dangerous
12522 error in server response which might have caused information leak
Willy Tarreau570f2212013-06-10 16:42:09 +020012523 (eg: cacheable cookie).
12524
12525 L : the session was locally processed by haproxy and was not passed to
12526 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012527
12528 R : a resource on the proxy has been exhausted (memory, sockets, source
12529 ports, ...). Usually, this appears during the connection phase, and
12530 system logs should contain a copy of the precise error. If this
12531 happens, it must be considered as a very serious anomaly which
12532 should be fixed as soon as possible by any means.
12533
12534 I : an internal error was identified by the proxy during a self-check.
12535 This should NEVER happen, and you are encouraged to report any log
12536 containing this, because this would almost certainly be a bug. It
12537 would be wise to preventively restart the process after such an
12538 event too, in case it would be caused by memory corruption.
12539
Simon Horman752dc4a2011-06-21 14:34:59 +090012540 D : the session was killed by haproxy because the server was detected
12541 as down and was configured to kill all connections when going down.
12542
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070012543 U : the session was killed by haproxy on this backup server because an
12544 active server was detected as up and was configured to kill all
12545 backup connections when going up.
12546
Willy Tarreaua2a64e92011-09-07 23:01:56 +020012547 K : the session was actively killed by an admin operating on haproxy.
12548
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012549 c : the client-side timeout expired while waiting for the client to
12550 send or receive data.
12551
12552 s : the server-side timeout expired while waiting for the server to
12553 send or receive data.
12554
12555 - : normal session completion, both the client and the server closed
12556 with nothing left in the buffers.
12557
12558 - on the second character, the TCP or HTTP session state when it was closed :
12559
Willy Tarreauf7b30a92010-12-06 22:59:17 +010012560 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012561 (HTTP mode only). Nothing was sent to any server.
12562
12563 Q : the proxy was waiting in the QUEUE for a connection slot. This can
12564 only happen when servers have a 'maxconn' parameter set. It can
12565 also happen in the global queue after a redispatch consecutive to
12566 a failed attempt to connect to a dying server. If no redispatch is
12567 reported, then no connection attempt was made to any server.
12568
12569 C : the proxy was waiting for the CONNECTION to establish on the
12570 server. The server might at most have noticed a connection attempt.
12571
12572 H : the proxy was waiting for complete, valid response HEADERS from the
12573 server (HTTP only).
12574
12575 D : the session was in the DATA phase.
12576
12577 L : the proxy was still transmitting LAST data to the client while the
12578 server had already finished. This one is very rare as it can only
12579 happen when the client dies while receiving the last packets.
12580
12581 T : the request was tarpitted. It has been held open with the client
12582 during the whole "timeout tarpit" duration or until the client
12583 closed, both of which will be reported in the "Tw" timer.
12584
12585 - : normal session completion after end of data transfer.
12586
12587 - the third character tells whether the persistence cookie was provided by
12588 the client (only in HTTP mode) :
12589
12590 N : the client provided NO cookie. This is usually the case for new
12591 visitors, so counting the number of occurrences of this flag in the
12592 logs generally indicate a valid trend for the site frequentation.
12593
12594 I : the client provided an INVALID cookie matching no known server.
12595 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020012596 cookies between HTTP/HTTPS sites, persistence conditionally
12597 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012598
12599 D : the client provided a cookie designating a server which was DOWN,
12600 so either "option persist" was used and the client was sent to
12601 this server, or it was not set and the client was redispatched to
12602 another server.
12603
Willy Tarreau996a92c2010-10-13 19:30:47 +020012604 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012605 server.
12606
Willy Tarreau996a92c2010-10-13 19:30:47 +020012607 E : the client provided a valid cookie, but with a last date which was
12608 older than what is allowed by the "maxidle" cookie parameter, so
12609 the cookie is consider EXPIRED and is ignored. The request will be
12610 redispatched just as if there was no cookie.
12611
12612 O : the client provided a valid cookie, but with a first date which was
12613 older than what is allowed by the "maxlife" cookie parameter, so
12614 the cookie is consider too OLD and is ignored. The request will be
12615 redispatched just as if there was no cookie.
12616
Willy Tarreauc89ccb62012-04-05 21:18:22 +020012617 U : a cookie was present but was not used to select the server because
12618 some other server selection mechanism was used instead (typically a
12619 "use-server" rule).
12620
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012621 - : does not apply (no cookie set in configuration).
12622
12623 - the last character reports what operations were performed on the persistence
12624 cookie returned by the server (only in HTTP mode) :
12625
12626 N : NO cookie was provided by the server, and none was inserted either.
12627
12628 I : no cookie was provided by the server, and the proxy INSERTED one.
12629 Note that in "cookie insert" mode, if the server provides a cookie,
12630 it will still be overwritten and reported as "I" here.
12631
Willy Tarreau996a92c2010-10-13 19:30:47 +020012632 U : the proxy UPDATED the last date in the cookie that was presented by
12633 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012634 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020012635 date indicated in the cookie. If any other change happens, such as
12636 a redispatch, then the cookie will be marked as inserted instead.
12637
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012638 P : a cookie was PROVIDED by the server and transmitted as-is.
12639
12640 R : the cookie provided by the server was REWRITTEN by the proxy, which
12641 happens in "cookie rewrite" or "cookie prefix" modes.
12642
12643 D : the cookie provided by the server was DELETED by the proxy.
12644
12645 - : does not apply (no cookie set in configuration).
12646
Willy Tarreau996a92c2010-10-13 19:30:47 +020012647The combination of the two first flags gives a lot of information about what
12648was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012649helpful to detect server saturation, network troubles, local system resource
12650starvation, attacks, etc...
12651
12652The most common termination flags combinations are indicated below. They are
12653alphabetically sorted, with the lowercase set just after the upper case for
12654easier finding and understanding.
12655
12656 Flags Reason
12657
12658 -- Normal termination.
12659
12660 CC The client aborted before the connection could be established to the
12661 server. This can happen when haproxy tries to connect to a recently
12662 dead (or unchecked) server, and the client aborts while haproxy is
12663 waiting for the server to respond or for "timeout connect" to expire.
12664
12665 CD The client unexpectedly aborted during data transfer. This can be
12666 caused by a browser crash, by an intermediate equipment between the
12667 client and haproxy which decided to actively break the connection,
12668 by network routing issues between the client and haproxy, or by a
12669 keep-alive session between the server and the client terminated first
12670 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010012671
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012672 cD The client did not send nor acknowledge any data for as long as the
12673 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020012674 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012675
12676 CH The client aborted while waiting for the server to start responding.
12677 It might be the server taking too long to respond or the client
12678 clicking the 'Stop' button too fast.
12679
12680 cH The "timeout client" stroke while waiting for client data during a
12681 POST request. This is sometimes caused by too large TCP MSS values
12682 for PPPoE networks which cannot transport full-sized packets. It can
12683 also happen when client timeout is smaller than server timeout and
12684 the server takes too long to respond.
12685
12686 CQ The client aborted while its session was queued, waiting for a server
12687 with enough empty slots to accept it. It might be that either all the
12688 servers were saturated or that the assigned server was taking too
12689 long a time to respond.
12690
12691 CR The client aborted before sending a full HTTP request. Most likely
12692 the request was typed by hand using a telnet client, and aborted
12693 too early. The HTTP status code is likely a 400 here. Sometimes this
12694 might also be caused by an IDS killing the connection between haproxy
12695 and the client.
12696
12697 cR The "timeout http-request" stroke before the client sent a full HTTP
12698 request. This is sometimes caused by too large TCP MSS values on the
12699 client side for PPPoE networks which cannot transport full-sized
12700 packets, or by clients sending requests by hand and not typing fast
12701 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020012702 request. The HTTP status code is likely a 408 here. Note: recently,
12703 some browsers such as Google Chrome started to break the deployed Web
12704 infrastructure by aggressively implementing a new "pre-connect"
12705 feature, consisting in sending connections to sites recently visited
12706 without sending any request on them until the user starts to browse
12707 the site. This mechanism causes massive disruption among resource-
12708 limited servers, and causes a lot of 408 errors in HAProxy logs.
12709 Worse, some people report that sometimes the browser displays the 408
12710 error when the user expects to see the actual content (Mozilla fixed
12711 this bug in 2004, while Chrome users continue to report it in 2014),
12712 so in this case, using "errorfile 408 /dev/null" can be used as a
12713 workaround. More information on the subject is available here :
12714 https://bugzilla.mozilla.org/show_bug.cgi?id=248827
12715 https://code.google.com/p/chromium/issues/detail?id=85229
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012716
12717 CT The client aborted while its session was tarpitted. It is important to
12718 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020012719 wrong tarpit rules have been written. If a lot of them happen, it
12720 might make sense to lower the "timeout tarpit" value to something
12721 closer to the average reported "Tw" timer, in order not to consume
12722 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012723
Willy Tarreau570f2212013-06-10 16:42:09 +020012724 LR The request was intercepted and locally handled by haproxy. Generally
12725 it means that this was a redirect or a stats request.
12726
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012727 SC The server or an equipment between it and haproxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012728 the TCP connection (the proxy received a TCP RST or an ICMP message
12729 in return). Under some circumstances, it can also be the network
12730 stack telling the proxy that the server is unreachable (eg: no route,
12731 or no ARP response on local network). When this happens in HTTP mode,
12732 the status code is likely a 502 or 503 here.
12733
12734 sC The "timeout connect" stroke before a connection to the server could
12735 complete. When this happens in HTTP mode, the status code is likely a
12736 503 or 504 here.
12737
12738 SD The connection to the server died with an error during the data
12739 transfer. This usually means that haproxy has received an RST from
12740 the server or an ICMP message from an intermediate equipment while
12741 exchanging data with the server. This can be caused by a server crash
12742 or by a network issue on an intermediate equipment.
12743
12744 sD The server did not send nor acknowledge any data for as long as the
12745 "timeout server" setting during the data phase. This is often caused
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012746 by too short timeouts on L4 equipments before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012747 load-balancers, ...), as well as keep-alive sessions maintained
12748 between the client and the server expiring first on haproxy.
12749
12750 SH The server aborted before sending its full HTTP response headers, or
12751 it crashed while processing the request. Since a server aborting at
12752 this moment is very rare, it would be wise to inspect its logs to
12753 control whether it crashed and why. The logged request may indicate a
12754 small set of faulty requests, demonstrating bugs in the application.
12755 Sometimes this might also be caused by an IDS killing the connection
12756 between haproxy and the server.
12757
12758 sH The "timeout server" stroke before the server could return its
12759 response headers. This is the most common anomaly, indicating too
12760 long transactions, probably caused by server or database saturation.
12761 The immediate workaround consists in increasing the "timeout server"
12762 setting, but it is important to keep in mind that the user experience
12763 will suffer from these long response times. The only long term
12764 solution is to fix the application.
12765
12766 sQ The session spent too much time in queue and has been expired. See
12767 the "timeout queue" and "timeout connect" settings to find out how to
12768 fix this if it happens too often. If it often happens massively in
12769 short periods, it may indicate general problems on the affected
12770 servers due to I/O or database congestion, or saturation caused by
12771 external attacks.
12772
12773 PC The proxy refused to establish a connection to the server because the
12774 process' socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020012775 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012776 so that it does not happen anymore. This status is very rare and
12777 might happen when the global "ulimit-n" parameter is forced by hand.
12778
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010012779 PD The proxy blocked an incorrectly formatted chunked encoded message in
12780 a request or a response, after the server has emitted its headers. In
12781 most cases, this will indicate an invalid message from the server to
Willy Tarreauf3a3e132013-08-31 08:16:26 +020012782 the client. Haproxy supports chunk sizes of up to 2GB - 1 (2147483647
12783 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010012784
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012785 PH The proxy blocked the server's response, because it was invalid,
12786 incomplete, dangerous (cache control), or matched a security filter.
12787 In any case, an HTTP 502 error is sent to the client. One possible
12788 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010012789 containing unauthorized characters. It is also possible but quite
12790 rare, that the proxy blocked a chunked-encoding request from the
12791 client due to an invalid syntax, before the server responded. In this
12792 case, an HTTP 400 error is sent to the client and reported in the
12793 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012794
12795 PR The proxy blocked the client's HTTP request, either because of an
12796 invalid HTTP syntax, in which case it returned an HTTP 400 error to
12797 the client, or because a deny filter matched, in which case it
12798 returned an HTTP 403 error.
12799
12800 PT The proxy blocked the client's request and has tarpitted its
12801 connection before returning it a 500 server error. Nothing was sent
12802 to the server. The connection was maintained open for as long as
12803 reported by the "Tw" timer field.
12804
12805 RC A local resource has been exhausted (memory, sockets, source ports)
12806 preventing the connection to the server from establishing. The error
12807 logs will tell precisely what was missing. This is very rare and can
12808 only be solved by proper system tuning.
12809
Willy Tarreau996a92c2010-10-13 19:30:47 +020012810The combination of the two last flags gives a lot of information about how
12811persistence was handled by the client, the server and by haproxy. This is very
12812important to troubleshoot disconnections, when users complain they have to
12813re-authenticate. The commonly encountered flags are :
12814
12815 -- Persistence cookie is not enabled.
12816
12817 NN No cookie was provided by the client, none was inserted in the
12818 response. For instance, this can be in insert mode with "postonly"
12819 set on a GET request.
12820
12821 II A cookie designating an invalid server was provided by the client,
12822 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012823 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020012824 value can be presented by a client when no other server knows it.
12825
12826 NI No cookie was provided by the client, one was inserted in the
12827 response. This typically happens for first requests from every user
12828 in "insert" mode, which makes it an easy way to count real users.
12829
12830 VN A cookie was provided by the client, none was inserted in the
12831 response. This happens for most responses for which the client has
12832 already got a cookie.
12833
12834 VU A cookie was provided by the client, with a last visit date which is
12835 not completely up-to-date, so an updated cookie was provided in
12836 response. This can also happen if there was no date at all, or if
12837 there was a date but the "maxidle" parameter was not set, so that the
12838 cookie can be switched to unlimited time.
12839
12840 EI A cookie was provided by the client, with a last visit date which is
12841 too old for the "maxidle" parameter, so the cookie was ignored and a
12842 new cookie was inserted in the response.
12843
12844 OI A cookie was provided by the client, with a first visit date which is
12845 too old for the "maxlife" parameter, so the cookie was ignored and a
12846 new cookie was inserted in the response.
12847
12848 DI The server designated by the cookie was down, a new server was
12849 selected and a new cookie was emitted in the response.
12850
12851 VI The server designated by the cookie was not marked dead but could not
12852 be reached. A redispatch happened and selected another one, which was
12853 then advertised in the response.
12854
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012855
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200128568.6. Non-printable characters
12857-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012858
12859In order not to cause trouble to log analysis tools or terminals during log
12860consulting, non-printable characters are not sent as-is into log files, but are
12861converted to the two-digits hexadecimal representation of their ASCII code,
12862prefixed by the character '#'. The only characters that can be logged without
12863being escaped are comprised between 32 and 126 (inclusive). Obviously, the
12864escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
12865is the same for the character '"' which becomes "#22", as well as '{', '|' and
12866'}' when logging headers.
12867
12868Note that the space character (' ') is not encoded in headers, which can cause
12869issues for tools relying on space count to locate fields. A typical header
12870containing spaces is "User-Agent".
12871
12872Last, it has been observed that some syslog daemons such as syslog-ng escape
12873the quote ('"') with a backslash ('\'). The reverse operation can safely be
12874performed since no quote may appear anywhere else in the logs.
12875
12876
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200128778.7. Capturing HTTP cookies
12878---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012879
12880Cookie capture simplifies the tracking a complete user session. This can be
12881achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012882section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012883cookie will simultaneously be checked in the request ("Cookie:" header) and in
12884the response ("Set-Cookie:" header). The respective values will be reported in
12885the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012886locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012887not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
12888user switches to a new session for example, because the server will reassign it
12889a new cookie. It is also possible to detect if a server unexpectedly sets a
12890wrong cookie to a client, leading to session crossing.
12891
12892 Examples :
12893 # capture the first cookie whose name starts with "ASPSESSION"
12894 capture cookie ASPSESSION len 32
12895
12896 # capture the first cookie whose name is exactly "vgnvisitor"
12897 capture cookie vgnvisitor= len 32
12898
12899
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200129008.8. Capturing HTTP headers
12901---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012902
12903Header captures are useful to track unique request identifiers set by an upper
12904proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
12905the response, one can search for information about the response length, how the
12906server asked the cache to behave, or an object location during a redirection.
12907
12908Header captures are performed using the "capture request header" and "capture
12909response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012910section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012911
12912It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012913time. Non-existent headers are logged as empty strings, and if one header
12914appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012915are grouped within braces '{' and '}' in the same order as they were declared,
12916and delimited with a vertical bar '|' without any space. Response headers
12917follow the same representation, but are displayed after a space following the
12918request headers block. These blocks are displayed just before the HTTP request
12919in the logs.
12920
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020012921As a special case, it is possible to specify an HTTP header capture in a TCP
12922frontend. The purpose is to enable logging of headers which will be parsed in
12923an HTTP backend if the request is then switched to this HTTP backend.
12924
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012925 Example :
12926 # This instance chains to the outgoing proxy
12927 listen proxy-out
12928 mode http
12929 option httplog
12930 option logasap
12931 log global
12932 server cache1 192.168.1.1:3128
12933
12934 # log the name of the virtual server
12935 capture request header Host len 20
12936
12937 # log the amount of data uploaded during a POST
12938 capture request header Content-Length len 10
12939
12940 # log the beginning of the referrer
12941 capture request header Referer len 20
12942
12943 # server name (useful for outgoing proxies only)
12944 capture response header Server len 20
12945
12946 # logging the content-length is useful with "option logasap"
12947 capture response header Content-Length len 10
12948
12949 # log the expected cache behaviour on the response
12950 capture response header Cache-Control len 8
12951
12952 # the Via header will report the next proxy's name
12953 capture response header Via len 20
12954
12955 # log the URL location during a redirection
12956 capture response header Location len 20
12957
12958 >>> Aug 9 20:26:09 localhost \
12959 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
12960 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
12961 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
12962 "GET http://fr.adserver.yahoo.com/"
12963
12964 >>> Aug 9 20:30:46 localhost \
12965 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
12966 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
12967 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010012968 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012969
12970 >>> Aug 9 20:30:46 localhost \
12971 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
12972 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
12973 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
12974 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010012975 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012976
12977
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200129788.9. Examples of logs
12979---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010012980
12981These are real-world examples of logs accompanied with an explanation. Some of
12982them have been made up by hand. The syslog part has been removed for better
12983reading. Their sole purpose is to explain how to decipher them.
12984
12985 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
12986 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
12987 "HEAD / HTTP/1.0"
12988
12989 => long request (6.5s) entered by hand through 'telnet'. The server replied
12990 in 147 ms, and the session ended normally ('----')
12991
12992 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
12993 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
12994 0/9 "HEAD / HTTP/1.0"
12995
12996 => Idem, but the request was queued in the global queue behind 9 other
12997 requests, and waited there for 1230 ms.
12998
12999 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
13000 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
13001 "GET /image.iso HTTP/1.0"
13002
13003 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010013004 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010013005 14 ms, 243 bytes of headers were sent to the client, and total time from
13006 accept to first data byte is 30 ms.
13007
13008 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
13009 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
13010 "GET /cgi-bin/bug.cgi? HTTP/1.0"
13011
13012 => the proxy blocked a server response either because of an "rspdeny" or
13013 "rspideny" filter, or because the response was improperly formatted and
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020013014 not HTTP-compliant, or because it blocked sensitive information which
Willy Tarreaucc6c8912009-02-22 10:53:55 +010013015 risked being cached. In this case, the response is replaced with a "502
13016 bad gateway". The flags ("PH--") tell us that it was haproxy who decided
13017 to return the 502 and not the server.
13018
13019 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010013020 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 +010013021
13022 => the client never completed its request and aborted itself ("C---") after
13023 8.5s, while the proxy was waiting for the request headers ("-R--").
13024 Nothing was sent to any server.
13025
13026 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
13027 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
13028
13029 => The client never completed its request, which was aborted by the
13030 time-out ("c---") after 50s, while the proxy was waiting for the request
13031 headers ("-R--"). Nothing was sent to any server, but the proxy could
13032 send a 408 return code to the client.
13033
13034 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
13035 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
13036
13037 => This log was produced with "option tcplog". The client timed out after
13038 5 seconds ("c----").
13039
13040 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
13041 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010013042 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010013043
13044 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013045 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010013046 (config says 'retries 3'), and no redispatch (otherwise we would have
13047 seen "/+3"). Status code 503 was returned to the client. There were 115
13048 connections on this server, 202 connections on this proxy, and 205 on
13049 the global process. It is possible that the server refused the
13050 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013051
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013052
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200130539. Statistics and monitoring
13054----------------------------
13055
13056It is possible to query HAProxy about its status. The most commonly used
13057mechanism is the HTTP statistics page. This page also exposes an alternative
13058CSV output format for monitoring tools. The same format is provided on the
13059Unix socket.
13060
13061
130629.1. CSV format
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013063---------------
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +010013064
Willy Tarreau7f062c42009-03-05 18:43:00 +010013065The statistics may be consulted either from the unix socket or from the HTTP
Willy Tarreaua3310dc2014-06-16 15:43:21 +020013066page. Both means provide a CSV format whose fields follow. The first line
13067begins with a sharp ('#') and has one word per comma-delimited field which
13068represents the title of the column. All other lines starting at the second one
13069use a classical CSV format using a comma as the delimiter, and the double quote
13070('"') as an optional text delimiter, but only if the enclosed text is ambiguous
13071(if it contains a quote or a comma). The double-quote character ('"') in the
13072text is doubled ('""'), which is the format that most tools recognize. Please
13073do not insert any column before these ones in order not to break tools which
13074use hard-coded column positions.
Willy Tarreau7f062c42009-03-05 18:43:00 +010013075
James Westby75c98ad2014-07-08 10:14:57 -040013076In brackets after each field name are the types which may have a value for
13077that field. The types are L (Listeners), F (Frontends), B (Backends), and
13078S (Servers).
13079
13080 0. pxname [LFBS]: proxy name
13081 1. svname [LFBS]: service name (FRONTEND for frontend, BACKEND for backend,
13082 any name for server/listener)
13083 2. qcur [..BS]: current queued requests. For the backend this reports the
13084 number queued without a server assigned.
13085 3. qmax [..BS]: max value of qcur
13086 4. scur [LFBS]: current sessions
13087 5. smax [LFBS]: max sessions
13088 6. slim [LFBS]: configured session limit
13089 7. stot [LFBS]: cumulative number of connections
13090 8. bin [LFBS]: bytes in
13091 9. bout [LFBS]: bytes out
13092 10. dreq [LFB.]: requests denied because of security concerns.
13093 - For tcp this is because of a matched tcp-request content rule.
13094 - For http this is because of a matched http-request or tarpit rule.
13095 11. dresp [LFBS]: responses denied because of security concerns.
13096 - For http this is because of a matched http-request rule, or
13097 "option checkcache".
13098 12. ereq [LF..]: request errors. Some of the possible causes are:
13099 - early termination from the client, before the request has been sent.
13100 - read error from the client
13101 - client timeout
13102 - client closed connection
13103 - various bad requests from the client.
13104 - request was tarpitted.
13105 13. econ [..BS]: number of requests that encountered an error trying to
13106 connect to a backend server. The backend stat is the sum of the stat
13107 for all servers of that backend, plus any connection errors not
13108 associated with a particular server (such as the backend having no
13109 active servers).
13110 14. eresp [..BS]: response errors. srv_abrt will be counted here also.
13111 Some other errors are:
13112 - write error on the client socket (won't be counted for the server stat)
13113 - failure applying filters to the response.
13114 15. wretr [..BS]: number of times a connection to a server was retried.
13115 16. wredis [..BS]: number of times a request was redispatched to another
13116 server. The server value counts the number of times that server was
13117 switched away from.
13118 17. status [LFBS]: status (UP/DOWN/NOLB/MAINT/MAINT(via)...)
13119 18. weight [..BS]: server weight (server), total weight (backend)
13120 19. act [..BS]: server is active (server), number of active servers (backend)
13121 20. bck [..BS]: server is backup (server), number of backup servers (backend)
13122 21. chkfail [...S]: number of failed checks. (Only counts checks failed when
13123 the server is up.)
13124 22. chkdown [..BS]: number of UP->DOWN transitions. The backend counter counts
13125 transitions to the whole backend being down, rather than the sum of the
13126 counters for each server.
13127 23. lastchg [..BS]: number of seconds since the last UP<->DOWN transition
13128 24. downtime [..BS]: total downtime (in seconds). The value for the backend
13129 is the downtime for the whole backend, not the sum of the server downtime.
13130 25. qlimit [...S]: configured maxqueue for the server, or nothing in the
13131 value is 0 (default, meaning no limit)
13132 26. pid [LFBS]: process id (0 for first instance, 1 for second, ...)
13133 27. iid [LFBS]: unique proxy id
13134 28. sid [L..S]: server id (unique inside a proxy)
13135 29. throttle [...S]: current throttle percentage for the server, when
13136 slowstart is active, or no value if not in slowstart.
13137 30. lbtot [..BS]: total number of times a server was selected, either for new
13138 sessions, or when re-dispatching. The server counter is the number
13139 of times that server was selected.
13140 31. tracked [...S]: id of proxy/server if tracking is enabled.
13141 32. type [LFBS]: (0=frontend, 1=backend, 2=server, 3=socket/listener)
13142 33. rate [.FBS]: number of sessions per second over last elapsed second
13143 34. rate_lim [.F..]: configured limit on new sessions per second
13144 35. rate_max [.FBS]: max number of new sessions per second
13145 36. check_status [...S]: status of last health check, one of:
Cyril Bontéf0c60612010-02-06 14:44:47 +010013146 UNK -> unknown
13147 INI -> initializing
13148 SOCKERR -> socket error
13149 L4OK -> check passed on layer 4, no upper layers testing enabled
13150 L4TMOUT -> layer 1-4 timeout
13151 L4CON -> layer 1-4 connection problem, for example
13152 "Connection refused" (tcp rst) or "No route to host" (icmp)
13153 L6OK -> check passed on layer 6
13154 L6TOUT -> layer 6 (SSL) timeout
13155 L6RSP -> layer 6 invalid response - protocol error
13156 L7OK -> check passed on layer 7
13157 L7OKC -> check conditionally passed on layer 7, for example 404 with
13158 disable-on-404
13159 L7TOUT -> layer 7 (HTTP/SMTP) timeout
13160 L7RSP -> layer 7 invalid response - protocol error
13161 L7STS -> layer 7 response error, for example HTTP 5xx
James Westby75c98ad2014-07-08 10:14:57 -040013162 37. check_code [...S]: layer5-7 code, if available
13163 38. check_duration [...S]: time in ms took to finish last health check
13164 39. hrsp_1xx [.FBS]: http responses with 1xx code
13165 40. hrsp_2xx [.FBS]: http responses with 2xx code
13166 41. hrsp_3xx [.FBS]: http responses with 3xx code
13167 42. hrsp_4xx [.FBS]: http responses with 4xx code
13168 43. hrsp_5xx [.FBS]: http responses with 5xx code
13169 44. hrsp_other [.FBS]: http responses with other codes (protocol error)
13170 45. hanafail [...S]: failed health checks details
13171 46. req_rate [.F..]: HTTP requests per second over last elapsed second
13172 47. req_rate_max [.F..]: max number of HTTP requests per second observed
13173 48. req_tot [.F..]: total number of HTTP requests received
13174 49. cli_abrt [..BS]: number of data transfers aborted by the client
13175 50. srv_abrt [..BS]: number of data transfers aborted by the server
13176 (inc. in eresp)
13177 51. comp_in [.FB.]: number of HTTP response bytes fed to the compressor
13178 52. comp_out [.FB.]: number of HTTP response bytes emitted by the compressor
13179 53. comp_byp [.FB.]: number of bytes that bypassed the HTTP compressor
13180 (CPU/BW limit)
13181 54. comp_rsp [.FB.]: number of HTTP responses that were compressed
13182 55. lastsess [..BS]: number of seconds since last session assigned to
13183 server/backend
13184 56. last_chk [...S]: last health check contents or textual error
13185 57. last_agt [...S]: last agent check contents or textual error
13186 58. qtime [..BS]: the average queue time in ms over the 1024 last requests
13187 59. ctime [..BS]: the average connect time in ms over the 1024 last requests
13188 60. rtime [..BS]: the average response time in ms over the 1024 last requests
13189 (0 for TCP)
13190 61. ttime [..BS]: the average total session time in ms over the 1024 last
13191 requests
Willy Tarreau844e3c52008-01-11 16:28:18 +010013192
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013193
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200131949.2. Unix Socket commands
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013195-------------------------
Krzysztof Piotr Oledzki2c6962c2008-03-02 02:42:14 +010013196
Willy Tarreau468f4932013-08-01 16:50:16 +020013197The stats socket is not enabled by default. In order to enable it, it is
13198necessary to add one line in the global section of the haproxy configuration.
13199A second line is recommended to set a larger timeout, always appreciated when
13200issuing commands by hand :
13201
13202 global
13203 stats socket /var/run/haproxy.sock mode 600 level admin
13204 stats timeout 2m
13205
13206It is also possible to add multiple instances of the stats socket by repeating
13207the line, and make them listen to a TCP port instead of a UNIX socket. This is
13208never done by default because this is dangerous, but can be handy in some
13209situations :
13210
13211 global
13212 stats socket /var/run/haproxy.sock mode 600 level admin
13213 stats socket ipv4@192.168.0.1:9999 level admin
13214 stats timeout 2m
13215
13216To access the socket, an external utility such as "socat" is required. Socat is a
13217swiss-army knife to connect anything to anything. We use it to connect terminals
13218to the socket, or a couple of stdin/stdout pipes to it for scripts. The two main
13219syntaxes we'll use are the following :
13220
13221 # socat /var/run/haproxy.sock stdio
13222 # socat /var/run/haproxy.sock readline
13223
13224The first one is used with scripts. It is possible to send the output of a
13225script to haproxy, and pass haproxy's output to another script. That's useful
13226for retrieving counters or attack traces for example.
13227
13228The second one is only useful for issuing commands by hand. It has the benefit
13229that the terminal is handled by the readline library which supports line
13230editing and history, which is very convenient when issuing repeated commands
13231(eg: watch a counter).
13232
13233The socket supports two operation modes :
13234 - interactive
13235 - non-interactive
13236
13237The non-interactive mode is the default when socat connects to the socket. In
13238this mode, a single line may be sent. It is processed as a whole, responses are
13239sent back, and the connection closes after the end of the response. This is the
13240mode that scripts and monitoring tools use. It is possible to send multiple
13241commands in this mode, they need to be delimited by a semi-colon (';'). For
13242example :
13243
13244 # echo "show info;show stat;show table" | socat /var/run/haproxy stdio
13245
13246The interactive mode displays a prompt ('>') and waits for commands to be
13247entered on the line, then processes them, and displays the prompt again to wait
13248for a new command. This mode is entered via the "prompt" command which must be
13249sent on the first line in non-interactive mode. The mode is a flip switch, if
13250"prompt" is sent in interactive mode, it is disabled and the connection closes
13251after processing the last command of the same line.
13252
13253For this reason, when debugging by hand, it's quite common to start with the
13254"prompt" command :
13255
13256 # socat /var/run/haproxy readline
13257 prompt
13258 > show info
13259 ...
13260 >
13261
13262Since multiple commands may be issued at once, haproxy uses the empty line as a
13263delimiter to mark an end of output for each command, and takes care of ensuring
13264that no command can emit an empty line on output. A script can thus easily
13265parse the output even when multiple commands were pipelined on a single line.
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013266
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013267It is important to understand that when multiple haproxy processes are started
13268on the same sockets, any process may pick up the request and will output its
13269own stats.
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013270
Willy Tarreau468f4932013-08-01 16:50:16 +020013271The list of commands currently supported on the stats socket is provided below.
13272If an unknown command is sent, haproxy displays the usage message which reminds
13273all supported commands. Some commands support a more complex syntax, generally
13274it will explain what part of the command is invalid when this happens.
13275
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013276add acl <acl> <pattern>
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013277 Add an entry into the acl <acl>. <acl> is the #<id> or the <file> returned by
13278 "show acl". This command does not verify if the entry already exists. This
13279 command cannot be used if the reference <acl> is a file also used with a map.
13280 In this case, you must use the command "add map" in place of "add acl".
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013281
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013282add map <map> <key> <value>
13283 Add an entry into the map <map> to associate the value <value> to the key
13284 <key>. This command does not verify if the entry already exists. It is
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013285 mainly used to fill a map after a clear operation. Note that if the reference
13286 <map> is a file and is shared with a map, this map will contain also a new
13287 pattern entry.
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013288
Willy Tarreaud63335a2010-02-26 12:56:52 +010013289clear counters
13290 Clear the max values of the statistics counters in each proxy (frontend &
13291 backend) and in each server. The cumulated counters are not affected. This
13292 can be used to get clean counters after an incident, without having to
13293 restart nor to clear traffic counters. This command is restricted and can
13294 only be issued on sockets configured for levels "operator" or "admin".
13295
13296clear counters all
13297 Clear all statistics counters in each proxy (frontend & backend) and in each
13298 server. This has the same effect as restarting. This command is restricted
13299 and can only be issued on sockets configured for level "admin".
13300
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013301clear acl <acl>
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013302 Remove all entries from the acl <acl>. <acl> is the #<id> or the <file>
13303 returned by "show acl". Note that if the reference <acl> is a file and is
13304 shared with a map, this map will be also cleared.
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013305
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013306clear map <map>
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013307 Remove all entries from the map <map>. <map> is the #<id> or the <file>
13308 returned by "show map". Note that if the reference <map> is a file and is
13309 shared with a acl, this acl will be also cleared.
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013310
Simon Hormanc88b8872011-06-15 15:18:49 +090013311clear table <table> [ data.<type> <operator> <value> ] | [ key <key> ]
13312 Remove entries from the stick-table <table>.
13313
13314 This is typically used to unblock some users complaining they have been
13315 abusively denied access to a service, but this can also be used to clear some
13316 stickiness entries matching a server that is going to be replaced (see "show
13317 table" below for details). Note that sometimes, removal of an entry will be
13318 refused because it is currently tracked by a session. Retrying a few seconds
13319 later after the session ends is usual enough.
13320
13321 In the case where no options arguments are given all entries will be removed.
13322
13323 When the "data." form is used entries matching a filter applied using the
13324 stored data (see "stick-table" in section 4.2) are removed. A stored data
13325 type must be specified in <type>, and this data type must be stored in the
13326 table otherwise an error is reported. The data is compared according to
13327 <operator> with the 64-bit integer <value>. Operators are the same as with
13328 the ACLs :
13329
13330 - eq : match entries whose data is equal to this value
13331 - ne : match entries whose data is not equal to this value
13332 - le : match entries whose data is less than or equal to this value
13333 - ge : match entries whose data is greater than or equal to this value
13334 - lt : match entries whose data is less than this value
13335 - gt : match entries whose data is greater than this value
13336
13337 When the key form is used the entry <key> is removed. The key must be of the
Simon Horman619e3cc2011-06-15 15:18:52 +090013338 same type as the table, which currently is limited to IPv4, IPv6, integer and
13339 string.
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013340
13341 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013342 $ echo "show table http_proxy" | socat stdio /tmp/sock1
Emeric Brun7c6b82e2010-09-24 16:34:28 +020013343 >>> # table: http_proxy, type: ip, size:204800, used:2
Willy Tarreau62a36c42010-08-17 15:53:10 +020013344 >>> 0x80e6a4c: key=127.0.0.1 use=0 exp=3594729 gpc0=0 conn_rate(30000)=1 \
13345 bytes_out_rate(60000)=187
13346 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13347 bytes_out_rate(60000)=191
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013348
13349 $ echo "clear table http_proxy key 127.0.0.1" | socat stdio /tmp/sock1
13350
13351 $ echo "show table http_proxy" | socat stdio /tmp/sock1
Emeric Brun7c6b82e2010-09-24 16:34:28 +020013352 >>> # table: http_proxy, type: ip, size:204800, used:1
Willy Tarreau62a36c42010-08-17 15:53:10 +020013353 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13354 bytes_out_rate(60000)=191
Simon Hormanc88b8872011-06-15 15:18:49 +090013355 $ echo "clear table http_proxy data.gpc0 eq 1" | socat stdio /tmp/sock1
13356 $ echo "show table http_proxy" | socat stdio /tmp/sock1
13357 >>> # table: http_proxy, type: ip, size:204800, used:1
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013358
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013359del acl <acl> [<key>|#<ref>]
13360 Delete all the acl entries from the acl <acl> corresponding to the key <key>.
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013361 <acl> is the #<id> or the <file> returned by "show acl". If the <ref> is used,
13362 this command delete only the listed reference. The reference can be found with
13363 listing the content of the acl. Note that if the reference <acl> is a file and
13364 is shared with a map, the entry will be also deleted in the map.
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013365
13366del map <map> [<key>|#<ref>]
Thierry FOURNIER0b90f312014-01-29 20:40:18 +010013367 Delete all the map entries from the map <map> corresponding to the key <key>.
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013368 <map> is the #<id> or the <file> returned by "show map". If the <ref> is used,
13369 this command delete only the listed reference. The reference can be found with
13370 listing the content of the map. Note that if the reference <map> is a file and
13371 is shared with a acl, the entry will be also deleted in the map.
Thierry FOURNIER0b90f312014-01-29 20:40:18 +010013372
13373disable agent <backend>/<server>
Simon Horman671b6f02013-11-25 10:46:39 +090013374 Mark the auxiliary agent check as temporarily stopped.
13375
13376 In the case where an agent check is being run as a auxiliary check, due
13377 to the agent-check parameter of a server directive, new checks are only
13378 initialised when the agent is in the enabled. Thus, disable agent will
13379 prevent any new agent checks from begin initiated until the agent
13380 re-enabled using enable agent.
13381
13382 When an agent is disabled the processing of an auxiliary agent check that
13383 was initiated while the agent was set as enabled is as follows: All
13384 results that would alter the weight, specifically "drain" or a weight
13385 returned by the agent, are ignored. The processing of agent check is
13386 otherwise unchanged.
13387
13388 The motivation for this feature is to allow the weight changing effects
13389 of the agent checks to be paused to allow the weight of a server to be
13390 configured using set weight without being overridden by the agent.
13391
13392 This command is restricted and can only be issued on sockets configured for
13393 level "admin".
13394
Willy Tarreau532a4502011-09-07 22:37:44 +020013395disable frontend <frontend>
13396 Mark the frontend as temporarily stopped. This corresponds to the mode which
13397 is used during a soft restart : the frontend releases the port but can be
13398 enabled again if needed. This should be used with care as some non-Linux OSes
13399 are unable to enable it back. This is intended to be used in environments
13400 where stopping a proxy is not even imaginable but a misconfigured proxy must
13401 be fixed. That way it's possible to release the port and bind it into another
13402 process to restore operations. The frontend will appear with status "STOP"
13403 on the stats page.
13404
13405 The frontend may be specified either by its name or by its numeric ID,
13406 prefixed with a sharp ('#').
13407
13408 This command is restricted and can only be issued on sockets configured for
13409 level "admin".
13410
Willy Tarreau9b5aecd2014-05-23 11:53:10 +020013411disable health <backend>/<server>
13412 Mark the primary health check as temporarily stopped. This will disable
13413 sending of health checks, and the last health check result will be ignored.
13414 The server will be in unchecked state and considered UP unless an auxiliary
13415 agent check forces it down.
13416
13417 This command is restricted and can only be issued on sockets configured for
13418 level "admin".
13419
Willy Tarreaud63335a2010-02-26 12:56:52 +010013420disable server <backend>/<server>
13421 Mark the server DOWN for maintenance. In this mode, no more checks will be
13422 performed on the server until it leaves maintenance.
13423 If the server is tracked by other servers, those servers will be set to DOWN
13424 during the maintenance.
13425
13426 In the statistics page, a server DOWN for maintenance will appear with a
13427 "MAINT" status, its tracking servers with the "MAINT(via)" one.
13428
13429 Both the backend and the server may be specified either by their name or by
Willy Tarreauf5f31922011-08-02 11:32:07 +020013430 their numeric ID, prefixed with a sharp ('#').
Willy Tarreaud63335a2010-02-26 12:56:52 +010013431
13432 This command is restricted and can only be issued on sockets configured for
13433 level "admin".
13434
Simon Horman671b6f02013-11-25 10:46:39 +090013435enable agent <backend>/<server>
13436 Resume auxiliary agent check that was temporarily stopped.
13437
13438 See "disable agent" for details of the effect of temporarily starting
13439 and stopping an auxiliary agent.
13440
13441 This command is restricted and can only be issued on sockets configured for
13442 level "admin".
13443
Willy Tarreau532a4502011-09-07 22:37:44 +020013444enable frontend <frontend>
13445 Resume a frontend which was temporarily stopped. It is possible that some of
13446 the listening ports won't be able to bind anymore (eg: if another process
13447 took them since the 'disable frontend' operation). If this happens, an error
13448 is displayed. Some operating systems might not be able to resume a frontend
13449 which was disabled.
13450
13451 The frontend may be specified either by its name or by its numeric ID,
13452 prefixed with a sharp ('#').
13453
13454 This command is restricted and can only be issued on sockets configured for
13455 level "admin".
13456
Willy Tarreau9b5aecd2014-05-23 11:53:10 +020013457enable health <backend>/<server>
13458 Resume a primary health check that was temporarily stopped. This will enable
13459 sending of health checks again. Please see "disable health" for details.
13460
13461 This command is restricted and can only be issued on sockets configured for
13462 level "admin".
13463
Willy Tarreaud63335a2010-02-26 12:56:52 +010013464enable server <backend>/<server>
13465 If the server was previously marked as DOWN for maintenance, this marks the
13466 server UP and checks are re-enabled.
13467
13468 Both the backend and the server may be specified either by their name or by
Willy Tarreauf5f31922011-08-02 11:32:07 +020013469 their numeric ID, prefixed with a sharp ('#').
Willy Tarreaud63335a2010-02-26 12:56:52 +010013470
13471 This command is restricted and can only be issued on sockets configured for
13472 level "admin".
13473
Thierry FOURNIER5b16df72014-02-26 18:07:38 +010013474get map <map> <value>
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013475get acl <acl> <value>
13476 Lookup the value <value> in the map <map> or in the ACL <acl>. <map> or <acl>
13477 are the #<id> or the <file> returned by "show map" or "show acl". This command
13478 returns all the matching patterns associated with this map. This is useful for
13479 debugging maps and ACLs. The output format is composed by one line par
13480 matching type. Each line is composed by space-delimited series of words.
Thierry FOURNIER5b16df72014-02-26 18:07:38 +010013481
13482 The first two words are:
13483
13484 <match method>: The match method applied. It can be "found", "bool",
13485 "int", "ip", "bin", "len", "str", "beg", "sub", "dir",
13486 "dom", "end" or "reg".
13487
13488 <match result>: The result. Can be "match" or "no-match".
13489
13490 The following words are returned only if the pattern matches an entry.
13491
13492 <index type>: "tree" or "list". The internal lookup algorithm.
13493
13494 <case>: "case-insensitive" or "case-sensitive". The
13495 interpretation of the case.
13496
13497 <entry matched>: match="<entry>". Return the matched pattern. It is
13498 useful with regular expressions.
13499
13500 The two last word are used to show the returned value and its type. With the
13501 "acl" case, the pattern doesn't exist.
13502
13503 return=nothing: No return because there are no "map".
13504 return="<value>": The value returned in the string format.
13505 return=cannot-display: The value cannot be converted as string.
13506
13507 type="<type>": The type of the returned sample.
13508
Willy Tarreaud63335a2010-02-26 12:56:52 +010013509get weight <backend>/<server>
13510 Report the current weight and the initial weight of server <server> in
13511 backend <backend> or an error if either doesn't exist. The initial weight is
13512 the one that appears in the configuration file. Both are normally equal
13513 unless the current weight has been changed. Both the backend and the server
13514 may be specified either by their name or by their numeric ID, prefixed with a
Willy Tarreauf5f31922011-08-02 11:32:07 +020013515 sharp ('#').
Willy Tarreaud63335a2010-02-26 12:56:52 +010013516
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013517help
13518 Print the list of known keywords and their basic usage. The same help screen
13519 is also displayed for unknown commands.
Willy Tarreau3dfe6cd2008-12-07 22:29:48 +010013520
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013521prompt
13522 Toggle the prompt at the beginning of the line and enter or leave interactive
13523 mode. In interactive mode, the connection is not closed after a command
13524 completes. Instead, the prompt will appear again, indicating the user that
13525 the interpreter is waiting for a new command. The prompt consists in a right
13526 angle bracket followed by a space "> ". This mode is particularly convenient
13527 when one wants to periodically check information such as stats or errors.
13528 It is also a good idea to enter interactive mode before issuing a "help"
13529 command.
13530
13531quit
13532 Close the connection when in interactive mode.
Krzysztof Piotr Oledzki2c6962c2008-03-02 02:42:14 +010013533
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013534set map <map> [<key>|#<ref>] <value>
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013535 Modify the value corresponding to each key <key> in a map <map>. <map> is the
13536 #<id> or <file> returned by "show map". If the <ref> is used in place of
13537 <key>, only the entry pointed by <ref> is changed. The new value is <value>.
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013538
Willy Tarreau2a0f4d22011-08-02 11:49:05 +020013539set maxconn frontend <frontend> <value>
Willy Tarreau3c7a79d2012-09-26 21:07:15 +020013540 Dynamically change the specified frontend's maxconn setting. Any positive
13541 value is allowed including zero, but setting values larger than the global
13542 maxconn does not make much sense. If the limit is increased and connections
13543 were pending, they will immediately be accepted. If it is lowered to a value
13544 below the current number of connections, new connections acceptation will be
Willy Tarreau2a0f4d22011-08-02 11:49:05 +020013545 delayed until the threshold is reached. The frontend might be specified by
13546 either its name or its numeric ID prefixed with a sharp ('#').
13547
Willy Tarreau91886b62011-09-07 14:38:31 +020013548set maxconn global <maxconn>
13549 Dynamically change the global maxconn setting within the range defined by the
13550 initial global maxconn setting. If it is increased and connections were
13551 pending, they will immediately be accepted. If it is lowered to a value below
13552 the current number of connections, new connections acceptation will be
13553 delayed until the threshold is reached. A value of zero restores the initial
13554 setting.
13555
Willy Tarreauf5b22872011-09-07 16:13:44 +020013556set rate-limit connections global <value>
13557 Change the process-wide connection rate limit, which is set by the global
13558 'maxconnrate' setting. A value of zero disables the limitation. This limit
13559 applies to all frontends and the change has an immediate effect. The value
13560 is passed in number of connections per second.
13561
William Lallemandd85f9172012-11-09 17:05:39 +010013562set rate-limit http-compression global <value>
13563 Change the maximum input compression rate, which is set by the global
13564 'maxcomprate' setting. A value of zero disables the limitation. The value is
William Lallemand096f5542012-11-19 17:26:05 +010013565 passed in number of kilobytes per second. The value is available in the "show
13566 info" on the line "CompressBpsRateLim" in bytes.
William Lallemandd85f9172012-11-09 17:05:39 +010013567
Willy Tarreau93e7c002013-10-07 18:51:07 +020013568set rate-limit sessions global <value>
13569 Change the process-wide session rate limit, which is set by the global
13570 'maxsessrate' setting. A value of zero disables the limitation. This limit
13571 applies to all frontends and the change has an immediate effect. The value
13572 is passed in number of sessions per second.
13573
Willy Tarreaue43d5322013-10-07 20:01:52 +020013574set rate-limit ssl-sessions global <value>
13575 Change the process-wide SSL session rate limit, which is set by the global
13576 'maxsslrate' setting. A value of zero disables the limitation. This limit
13577 applies to all frontends and the change has an immediate effect. The value
13578 is passed in number of sessions per second sent to the SSL stack. It applies
13579 before the handshake in order to protect the stack against handshake abuses.
13580
Willy Tarreau2a4b70f2014-05-22 18:42:35 +020013581set server <backend>/<server> agent [ up | down ]
13582 Force a server's agent to a new state. This can be useful to immediately
13583 switch a server's state regardless of some slow agent checks for example.
13584 Note that the change is propagated to tracking servers if any.
13585
13586set server <backend>/<server> health [ up | stopping | down ]
13587 Force a server's health to a new state. This can be useful to immediately
13588 switch a server's state regardless of some slow health checks for example.
13589 Note that the change is propagated to tracking servers if any.
13590
13591set server <backend>/<server> state [ ready | drain | maint ]
13592 Force a server's administrative state to a new state. This can be useful to
13593 disable load balancing and/or any traffic to a server. Setting the state to
13594 "ready" puts the server in normal mode, and the command is the equivalent of
13595 the "enable server" command. Setting the state to "maint" disables any traffic
13596 to the server as well as any health checks. This is the equivalent of the
13597 "disable server" command. Setting the mode to "drain" only removes the server
13598 from load balancing but still allows it to be checked and to accept new
13599 persistent connections. Changes are propagated to tracking servers if any.
13600
13601set server <backend>/<server> weight <weight>[%]
13602 Change a server's weight to the value passed in argument. This is the exact
13603 equivalent of the "set weight" command below.
13604
Emeric Brun4147b2e2014-06-16 18:36:30 +020013605set ssl ocsp-response <response>
13606 This command is used to update an OCSP Response for a certificate (see "crt"
13607 on "bind" lines). Same controls are performed as during the initial loading of
13608 the response. The <response> must be passed as a base64 encoded string of the
13609 DER encoded response from the OCSP server.
13610
13611 Example:
13612 openssl ocsp -issuer issuer.pem -cert server.pem \
13613 -host ocsp.issuer.com:80 -respout resp.der
13614 echo "set ssl ocsp-response $(base64 -w 10000 resp.der)" | \
13615 socat stdio /var/run/haproxy.stat
13616
Willy Tarreau47060b62013-08-01 21:11:42 +020013617set table <table> key <key> [data.<data_type> <value>]*
Willy Tarreau654694e2012-06-07 01:03:16 +020013618 Create or update a stick-table entry in the table. If the key is not present,
13619 an entry is inserted. See stick-table in section 4.2 to find all possible
13620 values for <data_type>. The most likely use consists in dynamically entering
13621 entries for source IP addresses, with a flag in gpc0 to dynamically block an
Willy Tarreau47060b62013-08-01 21:11:42 +020013622 IP address or affect its quality of service. It is possible to pass multiple
13623 data_types in a single call.
Willy Tarreau654694e2012-06-07 01:03:16 +020013624
Willy Tarreaud63335a2010-02-26 12:56:52 +010013625set timeout cli <delay>
13626 Change the CLI interface timeout for current connection. This can be useful
13627 during long debugging sessions where the user needs to constantly inspect
13628 some indicators without being disconnected. The delay is passed in seconds.
13629
13630set weight <backend>/<server> <weight>[%]
13631 Change a server's weight to the value passed in argument. If the value ends
13632 with the '%' sign, then the new weight will be relative to the initially
Simon Horman58b5d292013-02-12 10:45:52 +090013633 configured weight. Absolute weights are permitted between 0 and 256.
13634 Relative weights must be positive with the resulting absolute weight is
13635 capped at 256. Servers which are part of a farm running a static
13636 load-balancing algorithm have stricter limitations because the weight
13637 cannot change once set. Thus for these servers, the only accepted values
13638 are 0 and 100% (or 0 and the initial weight). Changes take effect
13639 immediately, though certain LB algorithms require a certain amount of
13640 requests to consider changes. A typical usage of this command is to
13641 disable a server during an update by setting its weight to zero, then to
13642 enable it again after the update by setting it back to 100%. This command
13643 is restricted and can only be issued on sockets configured for level
13644 "admin". Both the backend and the server may be specified either by their
13645 name or by their numeric ID, prefixed with a sharp ('#').
Willy Tarreaud63335a2010-02-26 12:56:52 +010013646
Willy Tarreaue0c8a1a2009-03-04 16:33:10 +010013647show errors [<iid>]
13648 Dump last known request and response errors collected by frontends and
13649 backends. If <iid> is specified, the limit the dump to errors concerning
Willy Tarreau6162db22009-10-10 17:13:00 +020013650 either frontend or backend whose ID is <iid>. This command is restricted
13651 and can only be issued on sockets configured for levels "operator" or
13652 "admin".
Willy Tarreaue0c8a1a2009-03-04 16:33:10 +010013653
13654 The errors which may be collected are the last request and response errors
13655 caused by protocol violations, often due to invalid characters in header
13656 names. The report precisely indicates what exact character violated the
13657 protocol. Other important information such as the exact date the error was
13658 detected, frontend and backend names, the server name (when known), the
13659 internal session ID and the source address which has initiated the session
13660 are reported too.
13661
13662 All characters are returned, and non-printable characters are encoded. The
13663 most common ones (\t = 9, \n = 10, \r = 13 and \e = 27) are encoded as one
13664 letter following a backslash. The backslash itself is encoded as '\\' to
13665 avoid confusion. Other non-printable characters are encoded '\xNN' where
13666 NN is the two-digits hexadecimal representation of the character's ASCII
13667 code.
13668
13669 Lines are prefixed with the position of their first character, starting at 0
13670 for the beginning of the buffer. At most one input line is printed per line,
13671 and large lines will be broken into multiple consecutive output lines so that
13672 the output never goes beyond 79 characters wide. It is easy to detect if a
13673 line was broken, because it will not end with '\n' and the next line's offset
13674 will be followed by a '+' sign, indicating it is a continuation of previous
13675 line.
13676
13677 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013678 $ echo "show errors" | socat stdio /tmp/sock1
13679 >>> [04/Mar/2009:15:46:56.081] backend http-in (#2) : invalid response
Willy Tarreaue0c8a1a2009-03-04 16:33:10 +010013680 src 127.0.0.1, session #54, frontend fe-eth0 (#1), server s2 (#1)
13681 response length 213 bytes, error at position 23:
13682
13683 00000 HTTP/1.0 200 OK\r\n
13684 00017 header/bizarre:blah\r\n
13685 00038 Location: blah\r\n
13686 00054 Long-line: this is a very long line which should b
13687 00104+ e broken into multiple lines on the output buffer,
13688 00154+ otherwise it would be too large to print in a ter
13689 00204+ minal\r\n
13690 00211 \r\n
13691
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013692 In the example above, we see that the backend "http-in" which has internal
Willy Tarreaue0c8a1a2009-03-04 16:33:10 +010013693 ID 2 has blocked an invalid response from its server s2 which has internal
13694 ID 1. The request was on session 54 initiated by source 127.0.0.1 and
13695 received by frontend fe-eth0 whose ID is 1. The total response length was
13696 213 bytes when the error was detected, and the error was at byte 23. This
13697 is the slash ('/') in header name "header/bizarre", which is not a valid
13698 HTTP character for a header name.
Krzysztof Piotr Oledzki2c6962c2008-03-02 02:42:14 +010013699
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013700show info
13701 Dump info about haproxy status on current process.
13702
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013703show map [<map>]
13704 Dump info about map converters. Without argument, the list of all available
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013705 maps is returned. If a <map> is specified, its contents are dumped. <map> is
13706 the #<id> or <file>. The first column is a unique identifier. It can be used
13707 as reference for the operation "del map" and "set map". The second column is
13708 the pattern and the third column is the sample if available. The data returned
13709 are not directly a list of available maps, but are the list of all patterns
13710 composing any map. Many of these patterns can be shared with ACL.
Thierry FOURNIERd32079e2014-01-29 20:02:04 +010013711
13712show acl [<acl>]
13713 Dump info about acl converters. Without argument, the list of all available
Thierry FOURNIER65ce6132014-03-20 11:42:45 +010013714 acls is returned. If a <acl> is specified, its contents are dumped. <acl> if
13715 the #<id> or <file>. The dump format is the same than the map even for the
13716 sample value. The data returned are not a list of available ACL, but are the
13717 list of all patterns composing any ACL. Many of these patterns can be shared
13718 with maps.
Thierry FOURNIERc0e0d7b2013-12-11 16:55:52 +010013719
Willy Tarreau12833bb2014-01-28 16:49:56 +010013720show pools
13721 Dump the status of internal memory pools. This is useful to track memory
13722 usage when suspecting a memory leak for example. It does exactly the same
13723 as the SIGQUIT when running in foreground except that it does not flush
13724 the pools.
13725
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013726show sess
13727 Dump all known sessions. Avoid doing this on slow connections as this can
Willy Tarreau6162db22009-10-10 17:13:00 +020013728 be huge. This command is restricted and can only be issued on sockets
13729 configured for levels "operator" or "admin".
13730
Willy Tarreau66dc20a2010-03-05 17:53:32 +010013731show sess <id>
13732 Display a lot of internal information about the specified session identifier.
13733 This identifier is the first field at the beginning of the lines in the dumps
13734 of "show sess" (it corresponds to the session pointer). Those information are
13735 useless to most users but may be used by haproxy developers to troubleshoot a
13736 complex bug. The output format is intentionally not documented so that it can
Willy Tarreau76153662012-11-26 01:16:39 +010013737 freely evolve depending on demands. The special id "all" dumps the states of
13738 all sessions, which can be avoided as much as possible as it is highly CPU
13739 intensive and can take a lot of time.
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013740
13741show stat [<iid> <type> <sid>]
13742 Dump statistics in the CSV format. By passing <id>, <type> and <sid>, it is
13743 possible to dump only selected items :
13744 - <iid> is a proxy ID, -1 to dump everything
13745 - <type> selects the type of dumpable objects : 1 for frontends, 2 for
13746 backends, 4 for servers, -1 for everything. These values can be ORed,
13747 for example:
13748 1 + 2 = 3 -> frontend + backend.
13749 1 + 2 + 4 = 7 -> frontend + backend + server.
13750 - <sid> is a server ID, -1 to dump everything from the selected proxy.
13751
13752 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013753 $ echo "show info;show stat" | socat stdio unix-connect:/tmp/sock1
13754 >>> Name: HAProxy
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013755 Version: 1.4-dev2-49
13756 Release_date: 2009/09/23
13757 Nbproc: 1
13758 Process_num: 1
13759 (...)
13760
13761 # pxname,svname,qcur,qmax,scur,smax,slim,stot,bin,bout,dreq, (...)
13762 stats,FRONTEND,,,0,0,1000,0,0,0,0,0,0,,,,,OPEN,,,,,,,,,1,1,0, (...)
13763 stats,BACKEND,0,0,0,0,1000,0,0,0,0,0,,0,0,0,0,UP,0,0,0,,0,250,(...)
13764 (...)
13765 www1,BACKEND,0,0,0,0,1000,0,0,0,0,0,,0,0,0,0,UP,1,1,0,,0,250, (...)
13766
13767 $
13768
13769 Here, two commands have been issued at once. That way it's easy to find
13770 which process the stats apply to in multi-process mode. Notice the empty
13771 line after the information output which marks the end of the first block.
13772 A similar empty line appears at the end of the second block (stats) so that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010013773 the reader knows the output has not been truncated.
Willy Tarreau9a42c0d2009-09-22 19:31:03 +020013774
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013775show table
13776 Dump general information on all known stick-tables. Their name is returned
13777 (the name of the proxy which holds them), their type (currently zero, always
13778 IP), their size in maximum possible number of entries, and the number of
13779 entries currently in use.
13780
13781 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013782 $ echo "show table" | socat stdio /tmp/sock1
Simon Horman64b28d02011-08-13 08:03:50 +090013783 >>> # table: front_pub, type: ip, size:204800, used:171454
13784 >>> # table: back_rdp, type: ip, size:204800, used:0
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013785
Simon Horman17bce342011-06-15 15:18:47 +090013786show table <name> [ data.<type> <operator> <value> ] | [ key <key> ]
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013787 Dump contents of stick-table <name>. In this mode, a first line of generic
13788 information about the table is reported as with "show table", then all
13789 entries are dumped. Since this can be quite heavy, it is possible to specify
Simon Horman17bce342011-06-15 15:18:47 +090013790 a filter in order to specify what entries to display.
13791
13792 When the "data." form is used the filter applies to the stored data (see
13793 "stick-table" in section 4.2). A stored data type must be specified
13794 in <type>, and this data type must be stored in the table otherwise an
13795 error is reported. The data is compared according to <operator> with the
13796 64-bit integer <value>. Operators are the same as with the ACLs :
13797
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013798 - eq : match entries whose data is equal to this value
13799 - ne : match entries whose data is not equal to this value
13800 - le : match entries whose data is less than or equal to this value
13801 - ge : match entries whose data is greater than or equal to this value
13802 - lt : match entries whose data is less than this value
13803 - gt : match entries whose data is greater than this value
13804
Simon Hormanc88b8872011-06-15 15:18:49 +090013805
13806 When the key form is used the entry <key> is shown. The key must be of the
Simon Horman619e3cc2011-06-15 15:18:52 +090013807 same type as the table, which currently is limited to IPv4, IPv6, integer,
13808 and string.
Simon Horman17bce342011-06-15 15:18:47 +090013809
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013810 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013811 $ echo "show table http_proxy" | socat stdio /tmp/sock1
Simon Horman64b28d02011-08-13 08:03:50 +090013812 >>> # table: http_proxy, type: ip, size:204800, used:2
Willy Tarreau62a36c42010-08-17 15:53:10 +020013813 >>> 0x80e6a4c: key=127.0.0.1 use=0 exp=3594729 gpc0=0 conn_rate(30000)=1 \
13814 bytes_out_rate(60000)=187
13815 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13816 bytes_out_rate(60000)=191
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013817
Willy Tarreau62a36c42010-08-17 15:53:10 +020013818 $ echo "show table http_proxy data.gpc0 gt 0" | socat stdio /tmp/sock1
Simon Horman64b28d02011-08-13 08:03:50 +090013819 >>> # table: http_proxy, type: ip, size:204800, used:2
Willy Tarreau62a36c42010-08-17 15:53:10 +020013820 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13821 bytes_out_rate(60000)=191
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013822
Willy Tarreau62a36c42010-08-17 15:53:10 +020013823 $ echo "show table http_proxy data.conn_rate gt 5" | \
13824 socat stdio /tmp/sock1
Simon Horman64b28d02011-08-13 08:03:50 +090013825 >>> # table: http_proxy, type: ip, size:204800, used:2
Willy Tarreau62a36c42010-08-17 15:53:10 +020013826 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13827 bytes_out_rate(60000)=191
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013828
Simon Horman17bce342011-06-15 15:18:47 +090013829 $ echo "show table http_proxy key 127.0.0.2" | \
13830 socat stdio /tmp/sock1
Simon Horman64b28d02011-08-13 08:03:50 +090013831 >>> # table: http_proxy, type: ip, size:204800, used:2
Simon Horman17bce342011-06-15 15:18:47 +090013832 >>> 0x80e6a80: key=127.0.0.2 use=0 exp=3594740 gpc0=1 conn_rate(30000)=10 \
13833 bytes_out_rate(60000)=191
13834
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013835 When the data criterion applies to a dynamic value dependent on time such as
13836 a bytes rate, the value is dynamically computed during the evaluation of the
13837 entry in order to decide whether it has to be dumped or not. This means that
13838 such a filter could match for some time then not match anymore because as
13839 time goes, the average event rate drops.
13840
13841 It is possible to use this to extract lists of IP addresses abusing the
13842 service, in order to monitor them or even blacklist them in a firewall.
13843 Example :
Willy Tarreau62a36c42010-08-17 15:53:10 +020013844 $ echo "show table http_proxy data.gpc0 gt 0" \
13845 | socat stdio /tmp/sock1 \
Willy Tarreau88bc4ec2010-08-01 07:58:48 +020013846 | fgrep 'key=' | cut -d' ' -f2 | cut -d= -f2 > abusers-ip.txt
13847 ( or | awk '/key/{ print a[split($2,a,"=")]; }' )
Krzysztof Piotr Oledzki719e7262009-10-04 15:02:46 +020013848
Willy Tarreau532a4502011-09-07 22:37:44 +020013849shutdown frontend <frontend>
13850 Completely delete the specified frontend. All the ports it was bound to will
13851 be released. It will not be possible to enable the frontend anymore after
13852 this operation. This is intended to be used in environments where stopping a
13853 proxy is not even imaginable but a misconfigured proxy must be fixed. That
13854 way it's possible to release the port and bind it into another process to
13855 restore operations. The frontend will not appear at all on the stats page
13856 once it is terminated.
13857
13858 The frontend may be specified either by its name or by its numeric ID,
13859 prefixed with a sharp ('#').
13860
13861 This command is restricted and can only be issued on sockets configured for
13862 level "admin".
13863
Willy Tarreaua295edc2011-09-07 23:21:03 +020013864shutdown session <id>
13865 Immediately terminate the session matching the specified session identifier.
13866 This identifier is the first field at the beginning of the lines in the dumps
13867 of "show sess" (it corresponds to the session pointer). This can be used to
13868 terminate a long-running session without waiting for a timeout or when an
13869 endless transfer is ongoing. Such terminated sessions are reported with a 'K'
13870 flag in the logs.
13871
Cyril Bontéde9789b2014-07-12 18:22:42 +020013872shutdown sessions server <backend>/<server>
Willy Tarreau52b2d222011-09-07 23:48:48 +020013873 Immediately terminate all the sessions attached to the specified server. This
13874 can be used to terminate long-running sessions after a server is put into
13875 maintenance mode, for instance. Such terminated sessions are reported with a
13876 'K' flag in the logs.
13877
Willy Tarreau0ba27502007-12-24 16:55:16 +010013878/*
13879 * Local variables:
13880 * fill-column: 79
13881 * End:
13882 */