blob: c2ede71547a7e12c8a6250e78c3269439802f8f8 [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 Tarreau0e658fb2016-11-25 16:55:50 +01005 version 1.8
Willy Tarreau6a06a402007-07-15 20:15:28 +02006 willy tarreau
Willy Tarreaue59fcdd2016-11-25 16:39:17 +01007 2016/11/25
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
William Lallemandf9873ba2015-05-05 17:37:14 +0200422.2. Quoting and escaping
William Lallemandb2f07452015-05-12 14:27:13 +0200432.3. Environment variables
442.4. Time format
452.5. Examples
Willy Tarreauc57f0e22009-05-10 13:12:33 +020046
473. Global parameters
483.1. Process management and security
493.2. Performance tuning
503.3. Debugging
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +0100513.4. Userlists
Cyril Bontédc4d9032012-04-08 21:57:39 +0200523.5. Peers
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200533.6. Mailers
Willy Tarreauc57f0e22009-05-10 13:12:33 +020054
554. Proxies
564.1. Proxy keywords matrix
574.2. Alphabetically sorted keywords reference
58
Willy Tarreau086fbf52012-09-24 20:34:51 +0200595. Bind and Server options
605.1. Bind options
615.2. Server and default-server options
Baptiste Assmann1fa66662015-04-14 00:28:47 +0200625.3. Server DNS resolution
635.3.1. Global overview
645.3.2. The resolvers section
Willy Tarreauc57f0e22009-05-10 13:12:33 +020065
666. HTTP header manipulation
67
Willy Tarreau74ca5042013-06-11 23:12:07 +0200687. Using ACLs and fetching samples
697.1. ACL basics
707.1.1. Matching booleans
717.1.2. Matching integers
727.1.3. Matching strings
737.1.4. Matching regular expressions (regexes)
747.1.5. Matching arbitrary data blocks
757.1.6. Matching IPv4 and IPv6 addresses
767.2. Using ACLs to form conditions
777.3. Fetching samples
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200787.3.1. Converters
797.3.2. Fetching samples from internal states
807.3.3. Fetching samples at Layer 4
817.3.4. Fetching samples at Layer 5
827.3.5. Fetching samples from buffer contents (Layer 6)
837.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +0200847.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020085
868. Logging
878.1. Log levels
888.2. Log formats
898.2.1. Default log format
908.2.2. TCP log format
918.2.3. HTTP log format
William Lallemand48940402012-01-30 16:47:22 +0100928.2.4. Custom log format
Willy Tarreau5f51e1a2012-12-03 18:40:10 +0100938.2.5. Error log format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200948.3. Advanced logging options
958.3.1. Disabling logging of external tests
968.3.2. Logging before waiting for the session to terminate
978.3.3. Raising log level upon errors
988.3.4. Disabling logging of successful connections
998.4. Timing events
1008.5. Session state at disconnection
1018.6. Non-printable characters
1028.7. Capturing HTTP cookies
1038.8. Capturing HTTP headers
1048.9. Examples of logs
105
Christopher Fauletc3fe5332016-04-07 15:30:10 +02001069. Supported filters
1079.1. Trace
1089.2. HTTP compression
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +02001099.3. Stream Processing Offload Engine (SPOE)
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200110
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200111
1121. Quick reminder about HTTP
113----------------------------
114
115When haproxy is running in HTTP mode, both the request and the response are
116fully analyzed and indexed, thus it becomes possible to build matching criteria
117on almost anything found in the contents.
118
119However, it is important to understand how HTTP requests and responses are
120formed, and how HAProxy decomposes them. It will then become easier to write
121correct rules and to debug existing configurations.
122
123
1241.1. The HTTP transaction model
125-------------------------------
126
127The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100128to one and only one response. Traditionally, a TCP connection is established
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200129from the client to the server, a request is sent by the client on the
130connection, the server responds and the connection is closed. A new request
131will involve a new connection :
132
133 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
134
135In this mode, called the "HTTP close" mode, there are as many connection
136establishments as there are HTTP transactions. Since the connection is closed
137by the server after the response, the client does not need to know the content
138length.
139
140Due to the transactional nature of the protocol, it was possible to improve it
141to avoid closing a connection between two subsequent transactions. In this mode
142however, it is mandatory that the server indicates the content length for each
143response so that the client does not wait indefinitely. For this, a special
144header is used: "Content-length". This mode is called the "keep-alive" mode :
145
146 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
147
148Its advantages are a reduced latency between transactions, and less processing
149power required on the server side. It is generally better than the close mode,
150but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200151a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200152
153A last improvement in the communications is the pipelining mode. It still uses
154keep-alive, but the client does not wait for the first response to send the
155second request. This is useful for fetching large number of images composing a
156page :
157
158 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
159
160This can obviously have a tremendous benefit on performance because the network
161latency is eliminated between subsequent requests. Many HTTP agents do not
162correctly support pipelining since there is no way to associate a response with
163the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100164server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165
Willy Tarreau70dffda2014-01-30 03:07:23 +0100166By default HAProxy operates in keep-alive mode with regards to persistent
167connections: for each connection it processes each request and response, and
168leaves the connection idle on both sides between the end of a response and the
169start of a new request.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200170
Willy Tarreau70dffda2014-01-30 03:07:23 +0100171HAProxy supports 5 connection modes :
172 - keep alive : all requests and responses are processed (default)
173 - tunnel : only the first request and response are processed,
174 everything else is forwarded with no analysis.
175 - passive close : tunnel with "Connection: close" added in both directions.
176 - server close : the server-facing connection is closed after the response.
177 - forced close : the connection is actively closed after end of response.
178
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200179
1801.2. HTTP request
181-----------------
182
183First, let's consider this HTTP request :
184
185 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100186 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200187 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
188 2 Host: www.mydomain.com
189 3 User-agent: my small browser
190 4 Accept: image/jpeg, image/gif
191 5 Accept: image/png
192
193
1941.2.1. The Request line
195-----------------------
196
197Line 1 is the "request line". It is always composed of 3 fields :
198
199 - a METHOD : GET
200 - a URI : /serv/login.php?lang=en&profile=2
201 - a version tag : HTTP/1.1
202
203All of them are delimited by what the standard calls LWS (linear white spaces),
204which are commonly spaces, but can also be tabs or line feeds/carriage returns
205followed by spaces/tabs. The method itself cannot contain any colon (':') and
206is limited to alphabetic letters. All those various combinations make it
207desirable that HAProxy performs the splitting itself rather than leaving it to
208the user to write a complex or inaccurate regular expression.
209
210The URI itself can have several forms :
211
212 - A "relative URI" :
213
214 /serv/login.php?lang=en&profile=2
215
216 It is a complete URL without the host part. This is generally what is
217 received by servers, reverse proxies and transparent proxies.
218
219 - An "absolute URI", also called a "URL" :
220
221 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
222
223 It is composed of a "scheme" (the protocol name followed by '://'), a host
224 name or address, optionally a colon (':') followed by a port number, then
225 a relative URI beginning at the first slash ('/') after the address part.
226 This is generally what proxies receive, but a server supporting HTTP/1.1
227 must accept this form too.
228
229 - a star ('*') : this form is only accepted in association with the OPTIONS
230 method and is not relayable. It is used to inquiry a next hop's
231 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100232
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200233 - an address:port combination : 192.168.0.12:80
234 This is used with the CONNECT method, which is used to establish TCP
235 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
236 other protocols too.
237
238In a relative URI, two sub-parts are identified. The part before the question
239mark is called the "path". It is typically the relative path to static objects
240on the server. The part after the question mark is called the "query string".
241It is mostly used with GET requests sent to dynamic scripts and is very
242specific to the language, framework or application in use.
243
244
2451.2.2. The request headers
246--------------------------
247
248The headers start at the second line. They are composed of a name at the
249beginning of the line, immediately followed by a colon (':'). Traditionally,
250an LWS is added after the colon but that's not required. Then come the values.
251Multiple identical headers may be folded into one single line, delimiting the
252values with commas, provided that their order is respected. This is commonly
253encountered in the "Cookie:" field. A header may span over multiple lines if
254the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
255define a total of 3 values for the "Accept:" header.
256
257Contrary to a common mis-conception, header names are not case-sensitive, and
258their values are not either if they refer to other header names (such as the
259"Connection:" header).
260
261The end of the headers is indicated by the first empty line. People often say
262that it's a double line feed, which is not exact, even if a double line feed
263is one valid form of empty line.
264
265Fortunately, HAProxy takes care of all these complex combinations when indexing
266headers, checking values and counting them, so there is no reason to worry
267about the way they could be written, but it is important not to accuse an
268application of being buggy if it does unusual, valid things.
269
270Important note:
271 As suggested by RFC2616, HAProxy normalizes headers by replacing line breaks
272 in the middle of headers by LWS in order to join multi-line headers. This
273 is necessary for proper analysis and helps less capable HTTP parsers to work
274 correctly and not to be fooled by such complex constructs.
275
276
2771.3. HTTP response
278------------------
279
280An HTTP response looks very much like an HTTP request. Both are called HTTP
281messages. Let's consider this HTTP response :
282
283 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100284 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200285 1 HTTP/1.1 200 OK
286 2 Content-length: 350
287 3 Content-Type: text/html
288
Willy Tarreau816b9792009-09-15 21:25:21 +0200289As a special case, HTTP supports so called "Informational responses" as status
290codes 1xx. These messages are special in that they don't convey any part of the
291response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100292continue to post its request for instance. In the case of a status 100 response
293the requested information will be carried by the next non-100 response message
294following the informational one. This implies that multiple responses may be
295sent to a single request, and that this only works when keep-alive is enabled
296(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
297correctly forward and skip them, and only process the next non-100 response. As
298such, these messages are neither logged nor transformed, unless explicitly
299state otherwise. Status 101 messages indicate that the protocol is changing
300over the same connection and that haproxy must switch to tunnel mode, just as
301if a CONNECT had occurred. Then the Upgrade header would contain additional
302information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200303
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200304
3051.3.1. The Response line
306------------------------
307
308Line 1 is the "response line". It is always composed of 3 fields :
309
310 - a version tag : HTTP/1.1
311 - a status code : 200
312 - a reason : OK
313
314The status code is always 3-digit. The first digit indicates a general status :
Willy Tarreau816b9792009-09-15 21:25:21 +0200315 - 1xx = informational message to be skipped (eg: 100, 101)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200316 - 2xx = OK, content is following (eg: 200, 206)
317 - 3xx = OK, no content following (eg: 302, 304)
318 - 4xx = error caused by the client (eg: 401, 403, 404)
319 - 5xx = error caused by the server (eg: 500, 502, 503)
320
321Please refer to RFC2616 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100322"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200323found there, but it's a common practice to respect the well-established
324messages. It can be composed of one or multiple words, such as "OK", "Found",
325or "Authentication Required".
326
327Haproxy may emit the following status codes by itself :
328
329 Code When / reason
330 200 access to stats page, and when replying to monitoring requests
331 301 when performing a redirection, depending on the configured code
332 302 when performing a redirection, depending on the configured code
333 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100334 307 when performing a redirection, depending on the configured code
335 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200336 400 for an invalid or too large request
337 401 when an authentication is required to perform the action (when
338 accessing the stats page)
339 403 when a request is forbidden by a "block" ACL or "reqdeny" filter
340 408 when the request timeout strikes before the request is complete
341 500 when haproxy encounters an unrecoverable internal error, such as a
342 memory allocation failure, which should never happen
343 502 when the server returns an empty, invalid or incomplete response, or
344 when an "rspdeny" filter blocks the response.
345 503 when no server was available to handle the request, or in response to
346 monitoring requests which match the "monitor fail" condition
347 504 when the response timeout strikes before the server responds
348
349The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3504.2).
351
352
3531.3.2. The response headers
354---------------------------
355
356Response headers work exactly like request headers, and as such, HAProxy uses
357the same parsing function for both. Please refer to paragraph 1.2.2 for more
358details.
359
360
3612. Configuring HAProxy
362----------------------
363
3642.1. Configuration file format
365------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200366
367HAProxy's configuration process involves 3 major sources of parameters :
368
369 - the arguments from the command-line, which always take precedence
370 - the "global" section, which sets process-wide parameters
371 - the proxies sections which can take form of "defaults", "listen",
372 "frontend" and "backend".
373
Willy Tarreau0ba27502007-12-24 16:55:16 +0100374The configuration file syntax consists in lines beginning with a keyword
375referenced in this manual, optionally followed by one or several parameters
William Lallemandf9873ba2015-05-05 17:37:14 +0200376delimited by spaces.
Willy Tarreau0ba27502007-12-24 16:55:16 +0100377
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200378
William Lallemandf9873ba2015-05-05 17:37:14 +02003792.2. Quoting and escaping
380-------------------------
381
382HAProxy's configuration introduces a quoting and escaping system similar to
383many programming languages. The configuration file supports 3 types: escaping
384with a backslash, weak quoting with double quotes, and strong quoting with
385single quotes.
386
387If spaces have to be entered in strings, then they must be escaped by preceding
388them by a backslash ('\') or by quoting them. Backslashes also have to be
389escaped by doubling or strong quoting them.
390
391Escaping is achieved by preceding a special character by a backslash ('\'):
392
393 \ to mark a space and differentiate it from a delimiter
394 \# to mark a hash and differentiate it from a comment
395 \\ to use a backslash
396 \' to use a single quote and differentiate it from strong quoting
397 \" to use a double quote and differentiate it from weak quoting
398
399Weak quoting is achieved by using double quotes (""). Weak quoting prevents
400the interpretation of:
401
402 space as a parameter separator
403 ' single quote as a strong quoting delimiter
404 # hash as a comment start
405
William Lallemandb2f07452015-05-12 14:27:13 +0200406Weak quoting permits the interpretation of variables, if you want to use a non
407-interpreted dollar within a double quoted string, you should escape it with a
408backslash ("\$"), it does not work outside weak quoting.
409
410Interpretation of escaping and special characters are not prevented by weak
William Lallemandf9873ba2015-05-05 17:37:14 +0200411quoting.
412
413Strong quoting is achieved by using single quotes (''). Inside single quotes,
414nothing is interpreted, it's the efficient way to quote regexes.
415
416Quoted and escaped strings are replaced in memory by their interpreted
417equivalent, it allows you to perform concatenation.
418
419 Example:
420 # those are equivalents:
421 log-format %{+Q}o\ %t\ %s\ %{-Q}r
422 log-format "%{+Q}o %t %s %{-Q}r"
423 log-format '%{+Q}o %t %s %{-Q}r'
424 log-format "%{+Q}o %t"' %s %{-Q}r'
425 log-format "%{+Q}o %t"' %s'\ %{-Q}r
426
427 # those are equivalents:
428 reqrep "^([^\ :]*)\ /static/(.*)" \1\ /\2
429 reqrep "^([^ :]*)\ /static/(.*)" '\1 /\2'
430 reqrep "^([^ :]*)\ /static/(.*)" "\1 /\2"
431 reqrep "^([^ :]*)\ /static/(.*)" "\1\ /\2"
432
433
William Lallemandb2f07452015-05-12 14:27:13 +02004342.3. Environment variables
435--------------------------
436
437HAProxy's configuration supports environment variables. Those variables are
438interpreted only within double quotes. Variables are expanded during the
439configuration parsing. Variable names must be preceded by a dollar ("$") and
440optionally enclosed with braces ("{}") similarly to what is done in Bourne
441shell. Variable names can contain alphanumerical characters or the character
442underscore ("_") but should not start with a digit.
443
444 Example:
445
446 bind "fd@${FD_APP1}"
447
448 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
449
450 user "$HAPROXY_USER"
451
452
4532.4. Time format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200454----------------
455
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100456Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100457values are generally expressed in milliseconds (unless explicitly stated
458otherwise) but may be expressed in any other unit by suffixing the unit to the
459numeric value. It is important to consider this because it will not be repeated
460for every keyword. Supported units are :
461
462 - us : microseconds. 1 microsecond = 1/1000000 second
463 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
464 - s : seconds. 1s = 1000ms
465 - m : minutes. 1m = 60s = 60000ms
466 - h : hours. 1h = 60m = 3600s = 3600000ms
467 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
468
469
William Lallemandf9873ba2015-05-05 17:37:14 +02004702.4. Examples
Patrick Mezard35da19c2010-06-12 17:02:47 +0200471-------------
472
473 # Simple configuration for an HTTP proxy listening on port 80 on all
474 # interfaces and forwarding requests to a single backend "servers" with a
475 # single server "server1" listening on 127.0.0.1:8000
476 global
477 daemon
478 maxconn 256
479
480 defaults
481 mode http
482 timeout connect 5000ms
483 timeout client 50000ms
484 timeout server 50000ms
485
486 frontend http-in
487 bind *:80
488 default_backend servers
489
490 backend servers
491 server server1 127.0.0.1:8000 maxconn 32
492
493
494 # The same configuration defined with a single listen block. Shorter but
495 # less expressive, especially in HTTP mode.
496 global
497 daemon
498 maxconn 256
499
500 defaults
501 mode http
502 timeout connect 5000ms
503 timeout client 50000ms
504 timeout server 50000ms
505
506 listen http-in
507 bind *:80
508 server server1 127.0.0.1:8000 maxconn 32
509
510
511Assuming haproxy is in $PATH, test these configurations in a shell with:
512
Willy Tarreauccb289d2010-12-11 20:19:38 +0100513 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200514
515
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005163. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200517--------------------
518
519Parameters in the "global" section are process-wide and often OS-specific. They
520are generally set once for all and do not need being changed once correct. Some
521of them have command-line equivalents.
522
523The following keywords are supported in the "global" section :
524
525 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200526 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200527 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200528 - crt-base
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200529 - cpu-map
Willy Tarreau6a06a402007-07-15 20:15:28 +0200530 - daemon
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200531 - description
532 - deviceatlas-json-file
533 - deviceatlas-log-level
534 - deviceatlas-separator
535 - deviceatlas-properties-cookie
Simon Horman98637e52014-06-20 12:30:16 +0900536 - external-check
Willy Tarreau6a06a402007-07-15 20:15:28 +0200537 - gid
538 - group
539 - log
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200540 - log-tag
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100541 - log-send-hostname
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200542 - lua-load
Willy Tarreau6a06a402007-07-15 20:15:28 +0200543 - nbproc
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200544 - node
Willy Tarreau6a06a402007-07-15 20:15:28 +0200545 - pidfile
Willy Tarreau1d549722016-02-16 12:41:57 +0100546 - presetenv
547 - resetenv
Willy Tarreau6a06a402007-07-15 20:15:28 +0200548 - uid
549 - ulimit-n
550 - user
Willy Tarreau1d549722016-02-16 12:41:57 +0100551 - setenv
Willy Tarreaufbee7132007-10-18 13:53:22 +0200552 - stats
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200553 - ssl-default-bind-ciphers
554 - ssl-default-bind-options
555 - ssl-default-server-ciphers
556 - ssl-default-server-options
557 - ssl-dh-param-file
Emeric Brun850efd52014-01-29 12:24:34 +0100558 - ssl-server-verify
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100559 - unix-bind
Willy Tarreau1d549722016-02-16 12:41:57 +0100560 - unsetenv
Thomas Holmesdb04f192015-05-18 13:21:39 +0100561 - 51degrees-data-file
562 - 51degrees-property-name-list
Dragan Dosen93b38d92015-06-29 16:43:25 +0200563 - 51degrees-property-separator
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200564 - 51degrees-cache-size
scientiamobiled0027ed2016-11-04 10:55:08 +0100565 - wurfl-data-file
566 - wurfl-information-list
567 - wurfl-information-list-separator
568 - wurfl-engine-mode
569 - wurfl-cache-size
570 - wurfl-useragent-priority
Willy Tarreaud72758d2010-01-12 10:42:19 +0100571
Willy Tarreau6a06a402007-07-15 20:15:28 +0200572 * Performance tuning
Willy Tarreau1746eec2014-04-25 10:46:47 +0200573 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +0200574 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +0200575 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +0100576 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +0100577 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100578 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +0200579 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +0200580 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +0200581 - maxsslrate
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200582 - maxzlibmem
Willy Tarreau6a06a402007-07-15 20:15:28 +0200583 - noepoll
584 - nokqueue
585 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100586 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300587 - nogetaddrinfo
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +0000588 - noreuseport
Willy Tarreaufe255b72007-10-14 23:09:26 +0200589 - spread-checks
Baptiste Assmann5626f482015-08-23 10:00:10 +0200590 - server-state-base
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200591 - server-state-file
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200592 - tune.buffers.limit
593 - tune.buffers.reserve
Willy Tarreau27a674e2009-08-17 07:23:33 +0200594 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +0200595 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +0100596 - tune.comp.maxlevel
Willy Tarreau193b8c62012-11-22 00:17:38 +0100597 - tune.http.cookielen
Willy Tarreauac1932d2011-10-24 19:14:41 +0200598 - tune.http.maxhdr
Willy Tarreau7e312732014-02-12 16:35:14 +0100599 - tune.idletimer
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100600 - tune.lua.forced-yield
Willy Tarreau32f61e22015-03-18 17:54:59 +0100601 - tune.lua.maxmem
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100602 - tune.lua.session-timeout
603 - tune.lua.task-timeout
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +0200604 - tune.lua.service-timeout
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100605 - tune.maxaccept
606 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +0200607 - tune.maxrewrite
Willy Tarreauf3045d22015-04-29 16:24:50 +0200608 - tune.pattern.cache-size
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200609 - tune.pipesize
Willy Tarreaue803de22010-01-21 17:43:04 +0100610 - tune.rcvbuf.client
611 - tune.rcvbuf.server
Willy Tarreaub22fc302015-12-14 12:04:35 +0100612 - tune.recv_enough
Willy Tarreaue803de22010-01-21 17:43:04 +0100613 - tune.sndbuf.client
614 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +0100615 - tune.ssl.cachesize
Willy Tarreaubfd59462013-02-21 07:46:09 +0100616 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +0200617 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +0100618 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +0200619 - tune.ssl.default-dh-param
Christopher Faulet31af49d2015-06-09 17:29:50 +0200620 - tune.ssl.ssl-ctx-cache-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200621 - tune.vars.global-max-size
Christopher Fauletff2613e2016-11-09 11:36:17 +0100622 - tune.vars.proc-max-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200623 - tune.vars.reqres-max-size
624 - tune.vars.sess-max-size
625 - tune.vars.txn-max-size
William Lallemanda509e4c2012-11-07 16:54:34 +0100626 - tune.zlib.memlevel
627 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +0100628
Willy Tarreau6a06a402007-07-15 20:15:28 +0200629 * Debugging
630 - debug
631 - quiet
Willy Tarreau6a06a402007-07-15 20:15:28 +0200632
633
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006343.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +0200635------------------------------------
636
Emeric Brunc8e8d122012-10-02 18:42:10 +0200637ca-base <dir>
638 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emeric Brunfd33a262012-10-11 16:28:27 +0200639 relative path is used with "ca-file" or "crl-file" directives. Absolute
640 locations specified in "ca-file" and "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +0200641
Willy Tarreau6a06a402007-07-15 20:15:28 +0200642chroot <jail dir>
643 Changes current directory to <jail dir> and performs a chroot() there before
644 dropping privileges. This increases the security level in case an unknown
645 vulnerability would be exploited, since it would make it very hard for the
646 attacker to exploit the system. This only works when the process is started
647 with superuser privileges. It is important to ensure that <jail_dir> is both
648 empty and unwritable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100649
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100650cpu-map <"all"|"odd"|"even"|process_num> <cpu-set>...
651 On Linux 2.6 and above, it is possible to bind a process to a specific CPU
652 set. This means that the process will never run on other CPUs. The "cpu-map"
653 directive specifies CPU sets for process sets. The first argument is the
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100654 process number to bind. This process must have a number between 1 and 32 or
655 64, depending on the machine's word size, and any process IDs above nbproc
656 are ignored. It is possible to specify all processes at once using "all",
657 only odd numbers using "odd" or even numbers using "even", just like with the
658 "bind-process" directive. The second and forthcoming arguments are CPU sets.
659 Each CPU set is either a unique number between 0 and 31 or 63 or a range with
660 two such numbers delimited by a dash ('-'). Multiple CPU numbers or ranges
661 may be specified, and the processes will be allowed to bind to all of them.
662 Obviously, multiple "cpu-map" directives may be specified. Each "cpu-map"
663 directive will replace the previous ones when they overlap.
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100664
Emeric Brunc8e8d122012-10-02 18:42:10 +0200665crt-base <dir>
666 Assigns a default directory to fetch SSL certificates from when a relative
667 path is used with "crtfile" directives. Absolute locations specified after
668 "crtfile" prevail and ignore "crt-base".
669
Willy Tarreau6a06a402007-07-15 20:15:28 +0200670daemon
671 Makes the process fork into background. This is the recommended mode of
672 operation. It is equivalent to the command line "-D" argument. It can be
673 disabled by the command line "-db" argument.
674
David Carlier8167f302015-06-01 13:50:06 +0200675deviceatlas-json-file <path>
676 Sets the path of the DeviceAtlas JSON data file to be loaded by the API.
677 The path must be a valid JSON data file and accessible by Haproxy process.
678
679deviceatlas-log-level <value>
680 Sets the level of informations returned by the API. This directive is
681 optional and set to 0 by default if not set.
682
683deviceatlas-separator <char>
684 Sets the character separator for the API properties results. This directive
685 is optional and set to | by default if not set.
686
Cyril Bonté0306c4a2015-10-26 22:37:38 +0100687deviceatlas-properties-cookie <name>
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200688 Sets the client cookie's name used for the detection if the DeviceAtlas
689 Client-side component was used during the request. This directive is optional
690 and set to DAPROPS by default if not set.
David Carlier29b3ca32015-09-25 14:09:21 +0100691
Simon Horman98637e52014-06-20 12:30:16 +0900692external-check
693 Allows the use of an external agent to perform health checks.
694 This is disabled by default as a security precaution.
695 See "option external-check".
696
Willy Tarreau6a06a402007-07-15 20:15:28 +0200697gid <number>
698 Changes the process' group ID to <number>. It is recommended that the group
699 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
700 be started with a user belonging to this group, or with superuser privileges.
Michael Schererab012dd2013-01-12 18:35:19 +0100701 Note that if haproxy is started from a user having supplementary groups, it
702 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200703 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100704
Willy Tarreau6a06a402007-07-15 20:15:28 +0200705group <group name>
706 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
707 See also "gid" and "user".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100708
Dragan Dosen7ad31542015-09-28 17:16:47 +0200709log <address> [len <length>] [format <format>] <facility> [max level [min level]]
Willy Tarreau6a06a402007-07-15 20:15:28 +0200710 Adds a global syslog server. Up to two global servers can be defined. They
711 will receive logs for startups and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +0100712 configured with "log global".
713
714 <address> can be one of:
715
Willy Tarreau2769aa02007-12-27 18:26:09 +0100716 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +0100717 no port is specified, 514 is used by default (the standard syslog
718 port).
719
David du Colombier24bb5f52011-03-17 10:40:23 +0100720 - An IPv6 address followed by a colon and optionally a UDP port. If
721 no port is specified, 514 is used by default (the standard syslog
722 port).
723
Robert Tsai81ae1952007-12-05 10:47:29 +0100724 - A filesystem path to a UNIX domain socket, keeping in mind
725 considerations for chroot (be sure the path is accessible inside
726 the chroot) and uid/gid (be sure the path is appropriately
727 writeable).
728
William Lallemandb2f07452015-05-12 14:27:13 +0200729 You may want to reference some environment variables in the address
730 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +0100731
Willy Tarreau18324f52014-06-27 18:10:07 +0200732 <length> is an optional maximum line length. Log lines larger than this value
733 will be truncated before being sent. The reason is that syslog
734 servers act differently on log line length. All servers support the
735 default value of 1024, but some servers simply drop larger lines
736 while others do log them. If a server supports long lines, it may
737 make sense to set this value here in order to avoid truncating long
738 lines. Similarly, if a server drops long lines, it is preferable to
739 truncate them before sending them. Accepted values are 80 to 65535
740 inclusive. The default value of 1024 is generally fine for all
741 standard usages. Some specific cases of long captures or
742 JSON-formated logs may require larger values.
743
Dragan Dosen7ad31542015-09-28 17:16:47 +0200744 <format> is the log format used when generating syslog messages. It may be
745 one of the following :
746
747 rfc3164 The RFC3164 syslog message format. This is the default.
748 (https://tools.ietf.org/html/rfc3164)
749
750 rfc5424 The RFC5424 syslog message format.
751 (https://tools.ietf.org/html/rfc5424)
752
Robert Tsai81ae1952007-12-05 10:47:29 +0100753 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200754
755 kern user mail daemon auth syslog lpr news
756 uucp cron auth2 ftp ntp audit alert cron2
757 local0 local1 local2 local3 local4 local5 local6 local7
758
759 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +0200760 all messages are sent. If a maximum level is specified, only messages with a
761 severity at least as important as this level will be sent. An optional minimum
762 level can be specified. If it is set, logs emitted with a more severe level
763 than this one will be capped to this level. This is used to avoid sending
764 "emerg" messages on all terminals on some default syslog configurations.
765 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200766
Cyril Bontédc4d9032012-04-08 21:57:39 +0200767 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +0200768
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100769log-send-hostname [<string>]
770 Sets the hostname field in the syslog header. If optional "string" parameter
771 is set the header is set to the string contents, otherwise uses the hostname
772 of the system. Generally used if one is not relaying logs through an
773 intermediate syslog server or for simply customizing the hostname printed in
774 the logs.
775
Kevinm48936af2010-12-22 16:08:21 +0000776log-tag <string>
777 Sets the tag field in the syslog header to this string. It defaults to the
778 program name as launched from the command line, which usually is "haproxy".
779 Sometimes it can be useful to differentiate between multiple processes
Willy Tarreau094af4e2015-01-07 15:03:42 +0100780 running on the same host. See also the per-proxy "log-tag" directive.
Kevinm48936af2010-12-22 16:08:21 +0000781
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100782lua-load <file>
783 This global directive loads and executes a Lua file. This directive can be
784 used multiple times.
785
Willy Tarreau6a06a402007-07-15 20:15:28 +0200786nbproc <number>
787 Creates <number> processes when going daemon. This requires the "daemon"
788 mode. By default, only one process is created, which is the recommended mode
789 of operation. For systems limited to small sets of file descriptors per
790 process, it may be needed to fork multiple daemons. USING MULTIPLE PROCESSES
791 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. See also "daemon".
792
793pidfile <pidfile>
794 Writes pids of all daemons into file <pidfile>. This option is equivalent to
795 the "-p" command line argument. The file must be accessible to the user
796 starting the process. See also "daemon".
797
Willy Tarreau1d549722016-02-16 12:41:57 +0100798presetenv <name> <value>
799 Sets environment variable <name> to value <value>. If the variable exists, it
800 is NOT overwritten. The changes immediately take effect so that the next line
801 in the configuration file sees the new value. See also "setenv", "resetenv",
802 and "unsetenv".
803
804resetenv [<name> ...]
805 Removes all environment variables except the ones specified in argument. It
806 allows to use a clean controlled environment before setting new values with
807 setenv or unsetenv. Please note that some internal functions may make use of
808 some environment variables, such as time manipulation functions, but also
809 OpenSSL or even external checks. This must be used with extreme care and only
810 after complete validation. The changes immediately take effect so that the
811 next line in the configuration file sees the new environment. See also
812 "setenv", "presetenv", and "unsetenv".
813
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100814stats bind-process [ all | odd | even | <number 1-64>[-<number 1-64>] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +0200815 Limits the stats socket to a certain set of processes numbers. By default the
816 stats socket is bound to all processes, causing a warning to be emitted when
817 nbproc is greater than 1 because there is no way to select the target process
818 when connecting. However, by using this setting, it becomes possible to pin
819 the stats socket to a specific set of processes, typically the first one. The
820 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +0100821 the number of processes used. The maximum process ID depends on the machine's
Willy Tarreauae302532014-05-07 19:22:24 +0200822 word size (32 or 64). A better option consists in using the "process" setting
823 of the "stats socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +0200824
Baptiste Assmann5626f482015-08-23 10:00:10 +0200825server-state-base <directory>
826 Specifies the directory prefix to be prepended in front of all servers state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +0200827 file names which do not start with a '/'. See also "server-state-file",
828 "load-server-state-from-file" and "server-state-file-name".
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200829
830server-state-file <file>
831 Specifies the path to the file containing state of servers. If the path starts
832 with a slash ('/'), it is considered absolute, otherwise it is considered
833 relative to the directory specified using "server-state-base" (if set) or to
834 the current directory. Before reloading HAProxy, it is possible to save the
835 servers' current state using the stats command "show servers state". The
836 output of this command must be written in the file pointed by <file>. When
837 starting up, before handling traffic, HAProxy will read, load and apply state
838 for each server found in the file and available in its current running
Baptiste Assmann01c6cc32015-08-23 11:45:29 +0200839 configuration. See also "server-state-base" and "show servers state",
840 "load-server-state-from-file" and "server-state-file-name"
Baptiste Assmann5626f482015-08-23 10:00:10 +0200841
Willy Tarreau1d549722016-02-16 12:41:57 +0100842setenv <name> <value>
843 Sets environment variable <name> to value <value>. If the variable exists, it
844 is overwritten. The changes immediately take effect so that the next line in
845 the configuration file sees the new value. See also "presetenv", "resetenv",
846 and "unsetenv".
847
Willy Tarreau610f04b2014-02-13 11:36:41 +0100848ssl-default-bind-ciphers <ciphers>
849 This setting is only available when support for OpenSSL was built in. It sets
850 the default string describing the list of cipher algorithms ("cipher suite")
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300851 that are negotiated during the SSL/TLS handshake for all "bind" lines which
Willy Tarreau610f04b2014-02-13 11:36:41 +0100852 do not explicitly define theirs. The format of the string is defined in
853 "man 1 ciphers" from OpenSSL man pages, and can be for instance a string such
854 as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes). Please check the
855 "bind" keyword for more information.
856
Emeric Brun2c86cbf2014-10-30 15:56:50 +0100857ssl-default-bind-options [<option>]...
858 This setting is only available when support for OpenSSL was built in. It sets
859 default ssl-options to force on all "bind" lines. Please check the "bind"
860 keyword to see available options.
861
862 Example:
863 global
864 ssl-default-bind-options no-sslv3 no-tls-tickets
865
Willy Tarreau610f04b2014-02-13 11:36:41 +0100866ssl-default-server-ciphers <ciphers>
867 This setting is only available when support for OpenSSL was built in. It
868 sets the default string describing the list of cipher algorithms that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300869 negotiated during the SSL/TLS handshake with the server, for all "server"
Willy Tarreau610f04b2014-02-13 11:36:41 +0100870 lines which do not explicitly define theirs. The format of the string is
871 defined in "man 1 ciphers". Please check the "server" keyword for more
872 information.
873
Emeric Brun2c86cbf2014-10-30 15:56:50 +0100874ssl-default-server-options [<option>]...
875 This setting is only available when support for OpenSSL was built in. It sets
876 default ssl-options to force on all "server" lines. Please check the "server"
877 keyword to see available options.
878
Remi Gacogne47783ef2015-05-29 15:53:22 +0200879ssl-dh-param-file <file>
880 This setting is only available when support for OpenSSL was built in. It sets
881 the default DH parameters that are used during the SSL/TLS handshake when
882 ephemeral Diffie-Hellman (DHE) key exchange is used, for all "bind" lines
883 which do not explicitely define theirs. It will be overridden by custom DH
884 parameters found in a bind certificate file if any. If custom DH parameters
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200885 are not specified either by using ssl-dh-param-file or by setting them
886 directly in the certificate file, pre-generated DH parameters of the size
887 specified by tune.ssl.default-dh-param will be used. Custom parameters are
888 known to be more secure and therefore their use is recommended.
Remi Gacogne47783ef2015-05-29 15:53:22 +0200889 Custom DH parameters may be generated by using the OpenSSL command
890 "openssl dhparam <size>", where size should be at least 2048, as 1024-bit DH
891 parameters should not be considered secure anymore.
892
Emeric Brun850efd52014-01-29 12:24:34 +0100893ssl-server-verify [none|required]
894 The default behavior for SSL verify on servers side. If specified to 'none',
895 servers certificates are not verified. The default is 'required' except if
896 forced using cmdline option '-dV'.
897
Willy Tarreauabb175f2012-09-24 12:43:26 +0200898stats socket [<address:port>|<path>] [param*]
899 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
900 Connections to this socket will return various statistics outputs and even
901 allow some commands to be issued to change some runtime settings. Please
Kevin Decherf949c7202015-10-13 23:26:44 +0200902 consult section 9.2 "Unix Socket commands" of Management Guide for more
903 details.
Willy Tarreau6162db22009-10-10 17:13:00 +0200904
Willy Tarreauabb175f2012-09-24 12:43:26 +0200905 All parameters supported by "bind" lines are supported, for instance to
906 restrict access to some users or their access rights. Please consult
907 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +0200908
909stats timeout <timeout, in milliseconds>
910 The default timeout on the stats socket is set to 10 seconds. It is possible
911 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +0100912 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +0200913
914stats maxconn <connections>
915 By default, the stats socket is limited to 10 concurrent connections. It is
916 possible to change this value with "stats maxconn".
917
Willy Tarreau6a06a402007-07-15 20:15:28 +0200918uid <number>
919 Changes the process' user ID to <number>. It is recommended that the user ID
920 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
921 be started with superuser privileges in order to be able to switch to another
922 one. See also "gid" and "user".
923
924ulimit-n <number>
925 Sets the maximum number of per-process file-descriptors to <number>. By
926 default, it is automatically computed, so it is recommended not to use this
927 option.
928
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100929unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
930 [ group <group> ] [ gid <gid> ]
931
932 Fixes common settings to UNIX listening sockets declared in "bind" statements.
933 This is mainly used to simplify declaration of those UNIX sockets and reduce
934 the risk of errors, since those settings are most commonly required but are
935 also process-specific. The <prefix> setting can be used to force all socket
936 path to be relative to that directory. This might be needed to access another
937 component's chroot. Note that those paths are resolved before haproxy chroots
938 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
939 all have the same meaning as their homonyms used by the "bind" statement. If
940 both are specified, the "bind" statement has priority, meaning that the
941 "unix-bind" settings may be seen as process-wide default settings.
942
Willy Tarreau1d549722016-02-16 12:41:57 +0100943unsetenv [<name> ...]
944 Removes environment variables specified in arguments. This can be useful to
945 hide some sensitive information that are occasionally inherited from the
946 user's environment during some operations. Variables which did not exist are
947 silently ignored so that after the operation, it is certain that none of
948 these variables remain. The changes immediately take effect so that the next
949 line in the configuration file will not see these variables. See also
950 "setenv", "presetenv", and "resetenv".
951
Willy Tarreau6a06a402007-07-15 20:15:28 +0200952user <user name>
953 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
954 See also "uid" and "group".
955
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +0200956node <name>
957 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
958
959 This statement is useful in HA configurations where two or more processes or
960 servers share the same IP address. By setting a different node-name on all
961 nodes, it becomes easy to immediately spot what server is handling the
962 traffic.
963
964description <text>
965 Add a text that describes the instance.
966
967 Please note that it is required to escape certain characters (# for example)
968 and this text is inserted into a html page so you should avoid using
969 "<" and ">" characters.
970
Thomas Holmesdb04f192015-05-18 13:21:39 +010097151degrees-data-file <file path>
972 The path of the 51Degrees data file to provide device detection services. The
973 file should be unzipped and accessible by HAProxy with relevavnt permissions.
974
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200975 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +0100976 compiled with USE_51DEGREES.
977
Ben Shillitof25e8e52016-12-02 14:25:37 +000097851degrees-property-name-list [<string> ...]
Thomas Holmesdb04f192015-05-18 13:21:39 +0100979 A list of 51Degrees property names to be load from the dataset. A full list
980 of names is available on the 51Degrees website:
981 https://51degrees.com/resources/property-dictionary
982
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200983 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +0100984 compiled with USE_51DEGREES.
985
Dragan Dosen93b38d92015-06-29 16:43:25 +020098651degrees-property-separator <char>
Thomas Holmesdb04f192015-05-18 13:21:39 +0100987 A char that will be appended to every property value in a response header
988 containing 51Degrees results. If not set that will be set as ','.
989
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200990 Please note that this option is only available when haproxy has been
991 compiled with USE_51DEGREES.
992
99351degrees-cache-size <number>
994 Sets the size of the 51Degrees converter cache to <number> entries. This
995 is an LRU cache which reminds previous device detections and their results.
996 By default, this cache is disabled.
997
998 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +0100999 compiled with USE_51DEGREES.
1000
scientiamobiled0027ed2016-11-04 10:55:08 +01001001wurfl-data-file <file path>
1002 The path of the WURFL data file to provide device detection services. The
1003 file should be accessible by HAProxy with relevant permissions.
1004
1005 Please note that this option is only available when haproxy has been compiled
1006 with USE_WURFL=1.
1007
1008wurfl-information-list [<capability>]*
1009 A space-delimited list of WURFL capabilities, virtual capabilities, property
1010 names we plan to use in injected headers. A full list of capability and
1011 virtual capability names is available on the Scientiamobile website :
1012
1013 https://www.scientiamobile.com/wurflCapability
1014
1015 Valid WURFL properties are:
1016 - wurfl_id Contains the device ID of the matched device.
1017
1018 - wurfl_root_id Contains the device root ID of the matched
1019 device.
1020
1021 - wurfl_isdevroot Tells if the matched device is a root device.
1022 Possible values are "TRUE" or "FALSE".
1023
1024 - wurfl_useragent The original useragent coming with this
1025 particular web request.
1026
1027 - wurfl_api_version Contains a string representing the currently
1028 used Libwurfl API version.
1029
1030 - wurfl_engine_target Contains a string representing the currently
1031 set WURFL Engine Target. Possible values are
1032 "HIGH_ACCURACY", "HIGH_PERFORMANCE", "INVALID".
1033
1034 - wurfl_info A string containing information on the parsed
1035 wurfl.xml and its full path.
1036
1037 - wurfl_last_load_time Contains the UNIX timestamp of the last time
1038 WURFL has been loaded successfully.
1039
1040 - wurfl_normalized_useragent The normalized useragent.
1041
1042 - wurfl_useragent_priority The user agent priority used by WURFL.
1043
1044 Please note that this option is only available when haproxy has been compiled
1045 with USE_WURFL=1.
1046
1047wurfl-information-list-separator <char>
1048 A char that will be used to separate values in a response header containing
1049 WURFL results. If not set that a comma (',') will be used by default.
1050
1051 Please note that this option is only available when haproxy has been compiled
1052 with USE_WURFL=1.
1053
1054wurfl-patch-file [<file path>]
1055 A list of WURFL patch file paths. Note that patches are loaded during startup
1056 thus before the chroot.
1057
1058 Please note that this option is only available when haproxy has been compiled
1059 with USE_WURFL=1.
1060
1061wurfl-engine-mode { accuracy | performance }
1062 Sets the WURFL engine target. You can choose between 'accuracy' or
1063 'performance' targets. In performance mode, desktop web browser detection is
1064 done programmatically without referencing the WURFL data. As a result, most
1065 desktop web browsers are returned as generic_web_browser WURFL ID for
1066 performance. If either performance or accuracy are not defined, performance
1067 mode is enabled by default.
1068
1069 Please note that this option is only available when haproxy has been compiled
1070 with USE_WURFL=1.
1071
1072wurfl-cache-size <U>[,<D>]
1073 Sets the WURFL caching strategy. Here <U> is the Useragent cache size, and
1074 <D> is the internal device cache size. There are three possibilities here :
1075 - "0" : no cache is used.
1076 - <U> : the Single LRU cache is used, the size is expressed in elements.
1077 - <U>,<D> : the Double LRU cache is used, both sizes are in elements. This is
1078 the highest performing option.
1079
1080 Please note that this option is only available when haproxy has been compiled
1081 with USE_WURFL=1.
1082
1083wurfl-useragent-priority { plain | sideloaded_browser }
1084 Tells WURFL if it should prioritize use of the plain user agent ('plain')
1085 over the default sideloaded browser user agent ('sideloaded_browser').
1086
1087 Please note that this option is only available when haproxy has been compiled
1088 with USE_WURFL=1.
1089
Willy Tarreau6a06a402007-07-15 20:15:28 +02001090
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010913.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +02001092-----------------------
1093
Willy Tarreau1746eec2014-04-25 10:46:47 +02001094max-spread-checks <delay in milliseconds>
1095 By default, haproxy tries to spread the start of health checks across the
1096 smallest health check interval of all the servers in a farm. The principle is
1097 to avoid hammering services running on the same server. But when using large
1098 check intervals (10 seconds or more), the last servers in the farm take some
1099 time before starting to be tested, which can be a problem. This parameter is
1100 used to enforce an upper bound on delay between the first and the last check,
1101 even if the servers' check intervals are larger. When servers run with
1102 shorter intervals, their intervals will be respected though.
1103
Willy Tarreau6a06a402007-07-15 20:15:28 +02001104maxconn <number>
1105 Sets the maximum per-process number of concurrent connections to <number>. It
1106 is equivalent to the command-line argument "-n". Proxies will stop accepting
1107 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +02001108 automatically adjusted according to this value. See also "ulimit-n". Note:
1109 the "select" poller cannot reliably use more than 1024 file descriptors on
1110 some platforms. If your platform only supports select and reports "select
1111 FAILED" on startup, you need to reduce maxconn until it works (slightly
Willy Tarreaud0256482015-01-15 21:45:22 +01001112 below 500 in general). If this value is not set, it will default to the value
1113 set in DEFAULT_MAXCONN at build time (reported in haproxy -vv) if no memory
1114 limit is enforced, or will be computed based on the memory limit, the buffer
1115 size, memory allocated to compression, SSL cache size, and use or not of SSL
1116 and the associated maxsslconn (which can also be automatic).
Willy Tarreau6a06a402007-07-15 20:15:28 +02001117
Willy Tarreau81c25d02011-09-07 15:17:21 +02001118maxconnrate <number>
1119 Sets the maximum per-process number of connections per second to <number>.
1120 Proxies will stop accepting connections when this limit is reached. It can be
1121 used to limit the global capacity regardless of each frontend capacity. It is
1122 important to note that this can only be used as a service protection measure,
1123 as there will not necessarily be a fair share between frontends when the
1124 limit is reached, so it's a good idea to also limit each frontend to some
1125 value close to its expected share. Also, lowering tune.maxaccept can improve
1126 fairness.
1127
William Lallemandd85f9172012-11-09 17:05:39 +01001128maxcomprate <number>
1129 Sets the maximum per-process input compression rate to <number> kilobytes
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001130 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +01001131 level will be decreased during the session. If the maximum is reached at the
1132 beginning of a session, the session will not compress at all. If the maximum
1133 is not reached, the compression level will be increased up to
1134 tune.comp.maxlevel. A value of zero means there is no limit, this is the
1135 default value.
1136
William Lallemand072a2bf2012-11-20 17:01:01 +01001137maxcompcpuusage <number>
1138 Sets the maximum CPU usage HAProxy can reach before stopping the compression
1139 for new requests or decreasing the compression level of current requests.
1140 It works like 'maxcomprate' but measures CPU usage instead of incoming data
1141 bandwidth. The value is expressed in percent of the CPU used by haproxy. In
1142 case of multiple processes (nbproc > 1), each process manages its individual
1143 usage. A value of 100 disable the limit. The default value is 100. Setting
1144 a lower value will prevent the compression work from slowing the whole
1145 process down and from introducing high latencies.
1146
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001147maxpipes <number>
1148 Sets the maximum per-process number of pipes to <number>. Currently, pipes
1149 are only used by kernel-based tcp splicing. Since a pipe contains two file
1150 descriptors, the "ulimit-n" value will be increased accordingly. The default
1151 value is maxconn/4, which seems to be more than enough for most heavy usages.
1152 The splice code dynamically allocates and releases pipes, and can fall back
1153 to standard copy, so setting this value too low may only impact performance.
1154
Willy Tarreau93e7c002013-10-07 18:51:07 +02001155maxsessrate <number>
1156 Sets the maximum per-process number of sessions per second to <number>.
1157 Proxies will stop accepting connections when this limit is reached. It can be
1158 used to limit the global capacity regardless of each frontend capacity. It is
1159 important to note that this can only be used as a service protection measure,
1160 as there will not necessarily be a fair share between frontends when the
1161 limit is reached, so it's a good idea to also limit each frontend to some
1162 value close to its expected share. Also, lowering tune.maxaccept can improve
1163 fairness.
1164
Willy Tarreau403edff2012-09-06 11:58:37 +02001165maxsslconn <number>
1166 Sets the maximum per-process number of concurrent SSL connections to
1167 <number>. By default there is no SSL-specific limit, which means that the
1168 global maxconn setting will apply to all connections. Setting this limit
1169 avoids having openssl use too much memory and crash when malloc returns NULL
1170 (since it unfortunately does not reliably check for such conditions). Note
1171 that the limit applies both to incoming and outgoing connections, so one
1172 connection which is deciphered then ciphered accounts for 2 SSL connections.
Willy Tarreaud0256482015-01-15 21:45:22 +01001173 If this value is not set, but a memory limit is enforced, this value will be
1174 automatically computed based on the memory limit, maxconn, the buffer size,
1175 memory allocated to compression, SSL cache size, and use of SSL in either
1176 frontends, backends or both. If neither maxconn nor maxsslconn are specified
1177 when there is a memory limit, haproxy will automatically adjust these values
1178 so that 100% of the connections can be made over SSL with no risk, and will
1179 consider the sides where it is enabled (frontend, backend, both).
Willy Tarreau403edff2012-09-06 11:58:37 +02001180
Willy Tarreaue43d5322013-10-07 20:01:52 +02001181maxsslrate <number>
1182 Sets the maximum per-process number of SSL sessions per second to <number>.
1183 SSL listeners will stop accepting connections when this limit is reached. It
1184 can be used to limit the global SSL CPU usage regardless of each frontend
1185 capacity. It is important to note that this can only be used as a service
1186 protection measure, as there will not necessarily be a fair share between
1187 frontends when the limit is reached, so it's a good idea to also limit each
1188 frontend to some value close to its expected share. It is also important to
1189 note that the sessions are accounted before they enter the SSL stack and not
1190 after, which also protects the stack against bad handshakes. Also, lowering
1191 tune.maxaccept can improve fairness.
1192
William Lallemand9d5f5482012-11-07 16:12:57 +01001193maxzlibmem <number>
1194 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
1195 When the maximum amount is reached, future sessions will not compress as long
1196 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +01001197 The default value is 0. The value is available in bytes on the UNIX socket
1198 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
1199 "ZlibMemUsage" in bytes.
1200
Willy Tarreau6a06a402007-07-15 20:15:28 +02001201noepoll
1202 Disables the use of the "epoll" event polling system on Linux. It is
1203 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +01001204 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001205
1206nokqueue
1207 Disables the use of the "kqueue" event polling system on BSD. It is
1208 equivalent to the command-line argument "-dk". The next polling system
1209 used will generally be "poll". See also "nopoll".
1210
1211nopoll
1212 Disables the use of the "poll" event polling system. It is equivalent to the
1213 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001214 It should never be needed to disable "poll" since it's available on all
Willy Tarreaue9f49e72012-11-11 17:42:00 +01001215 platforms supported by HAProxy. See also "nokqueue" and "noepoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001216
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001217nosplice
1218 Disables the use of kernel tcp splicing between sockets on Linux. It is
1219 equivalent to the command line argument "-dS". Data will then be copied
1220 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001221 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001222 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
1223 be used. This option makes it easier to globally disable kernel splicing in
1224 case of doubt. See also "option splice-auto", "option splice-request" and
1225 "option splice-response".
1226
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001227nogetaddrinfo
1228 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
1229 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
1230
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00001231noreuseport
1232 Disables the use of SO_REUSEPORT - see socket(7). It is equivalent to the
1233 command line argument "-dR".
1234
Willy Tarreaufe255b72007-10-14 23:09:26 +02001235spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +09001236 Sometimes it is desirable to avoid sending agent and health checks to
1237 servers at exact intervals, for instance when many logical servers are
1238 located on the same physical server. With the help of this parameter, it
1239 becomes possible to add some randomness in the check interval between 0
1240 and +/- 50%. A value between 2 and 5 seems to show good results. The
1241 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +02001242
Willy Tarreau33cb0652014-12-23 22:52:37 +01001243tune.buffers.limit <number>
1244 Sets a hard limit on the number of buffers which may be allocated per process.
1245 The default value is zero which means unlimited. The minimum non-zero value
1246 will always be greater than "tune.buffers.reserve" and should ideally always
1247 be about twice as large. Forcing this value can be particularly useful to
1248 limit the amount of memory a process may take, while retaining a sane
1249 behaviour. When this limit is reached, sessions which need a buffer wait for
1250 another one to be released by another session. Since buffers are dynamically
1251 allocated and released, the waiting time is very short and not perceptible
1252 provided that limits remain reasonable. In fact sometimes reducing the limit
1253 may even increase performance by increasing the CPU cache's efficiency. Tests
1254 have shown good results on average HTTP traffic with a limit to 1/10 of the
1255 expected global maxconn setting, which also significantly reduces memory
1256 usage. The memory savings come from the fact that a number of connections
1257 will not allocate 2*tune.bufsize. It is best not to touch this value unless
1258 advised to do so by an haproxy core developer.
1259
Willy Tarreau1058ae72014-12-23 22:40:40 +01001260tune.buffers.reserve <number>
1261 Sets the number of buffers which are pre-allocated and reserved for use only
1262 during memory shortage conditions resulting in failed memory allocations. The
1263 minimum value is 2 and is also the default. There is no reason a user would
1264 want to change this value, it's mostly aimed at haproxy core developers.
1265
Willy Tarreau27a674e2009-08-17 07:23:33 +02001266tune.bufsize <number>
1267 Sets the buffer size to this size (in bytes). Lower values allow more
1268 sessions to coexist in the same amount of RAM, and higher values allow some
1269 applications with very large cookies to work. The default value is 16384 and
1270 can be changed at build time. It is strongly recommended not to change this
1271 from the default value, as very low values will break some services such as
1272 statistics, and values larger than default size will increase memory usage,
1273 possibly causing the system to run out of memory. At least the global maxconn
1274 parameter should be decreased by the same factor as this one is increased.
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04001275 If HTTP request is larger than (tune.bufsize - tune.maxrewrite), haproxy will
1276 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
1277 than this size, haproxy will return HTTP 502 (Bad Gateway).
Willy Tarreau27a674e2009-08-17 07:23:33 +02001278
Willy Tarreau43961d52010-10-04 20:39:20 +02001279tune.chksize <number>
1280 Sets the check buffer size to this size (in bytes). Higher values may help
1281 find string or regex patterns in very large pages, though doing so may imply
1282 more memory and CPU usage. The default value is 16384 and can be changed at
1283 build time. It is not recommended to change this value, but to use better
1284 checks whenever possible.
1285
William Lallemandf3747832012-11-09 12:33:10 +01001286tune.comp.maxlevel <number>
1287 Sets the maximum compression level. The compression level affects CPU
1288 usage during compression. This value affects CPU usage during compression.
1289 Each session using compression initializes the compression algorithm with
1290 this value. The default value is 1.
1291
Willy Tarreau193b8c62012-11-22 00:17:38 +01001292tune.http.cookielen <number>
1293 Sets the maximum length of captured cookies. This is the maximum value that
1294 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
1295 will automatically be truncated to this one. It is important not to set too
1296 high a value because all cookie captures still allocate this size whatever
1297 their configured value (they share a same pool). This value is per request
1298 per response, so the memory allocated is twice this value per connection.
1299 When not specified, the limit is set to 63 characters. It is recommended not
1300 to change this value.
1301
Willy Tarreauac1932d2011-10-24 19:14:41 +02001302tune.http.maxhdr <number>
1303 Sets the maximum number of headers in a request. When a request comes with a
1304 number of headers greater than this value (including the first line), it is
1305 rejected with a "400 Bad Request" status code. Similarly, too large responses
1306 are blocked with "502 Bad Gateway". The default value is 101, which is enough
1307 for all usages, considering that the widely deployed Apache server uses the
1308 same limit. It can be useful to push this limit further to temporarily allow
1309 a buggy application to work by the time it gets fixed. Keep in mind that each
1310 new header consumes 32bits of memory for each session, so don't push this
1311 limit too high.
1312
Willy Tarreau7e312732014-02-12 16:35:14 +01001313tune.idletimer <timeout>
1314 Sets the duration after which haproxy will consider that an empty buffer is
1315 probably associated with an idle stream. This is used to optimally adjust
1316 some packet sizes while forwarding large and small data alternatively. The
1317 decision to use splice() or to send large buffers in SSL is modulated by this
1318 parameter. The value is in milliseconds between 0 and 65535. A value of zero
1319 means that haproxy will not try to detect idle streams. The default is 1000,
1320 which seems to correctly detect end user pauses (eg: read a page before
1321 clicking). There should be not reason for changing this value. Please check
1322 tune.ssl.maxrecord below.
1323
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001324tune.lua.forced-yield <number>
1325 This directive forces the Lua engine to execute a yield each <number> of
Tim Düsterhus4896c442016-11-29 02:15:19 +01001326 instructions executed. This permits interrupting a long script and allows the
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001327 HAProxy scheduler to process other tasks like accepting connections or
1328 forwarding traffic. The default value is 10000 instructions. If HAProxy often
1329 executes some Lua code but more reactivity is required, this value can be
1330 lowered. If the Lua code is quite long and its result is absolutely required
1331 to process the data, the <number> can be increased.
1332
Willy Tarreau32f61e22015-03-18 17:54:59 +01001333tune.lua.maxmem
1334 Sets the maximum amount of RAM in megabytes per process usable by Lua. By
1335 default it is zero which means unlimited. It is important to set a limit to
1336 ensure that a bug in a script will not result in the system running out of
1337 memory.
1338
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001339tune.lua.session-timeout <timeout>
1340 This is the execution timeout for the Lua sessions. This is useful for
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001341 preventing infinite loops or spending too much time in Lua. This timeout
1342 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
1343 not taked in account. The default timeout is 4s.
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001344
1345tune.lua.task-timeout <timeout>
1346 Purpose is the same as "tune.lua.session-timeout", but this timeout is
1347 dedicated to the tasks. By default, this timeout isn't set because a task may
1348 remain alive during of the lifetime of HAProxy. For example, a task used to
1349 check servers.
1350
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001351tune.lua.service-timeout <timeout>
1352 This is the execution timeout for the Lua services. This is useful for
1353 preventing infinite loops or spending too much time in Lua. This timeout
1354 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
1355 not taked in account. The default timeout is 4s.
1356
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001357tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +01001358 Sets the maximum number of consecutive connections a process may accept in a
1359 row before switching to other work. In single process mode, higher numbers
1360 give better performance at high connection rates. However in multi-process
1361 modes, keeping a bit of fairness between processes generally is better to
1362 increase performance. This value applies individually to each listener, so
1363 that the number of processes a listener is bound to is taken into account.
1364 This value defaults to 64. In multi-process mode, it is divided by twice
1365 the number of processes the listener is bound to. Setting this value to -1
1366 completely disables the limitation. It should normally not be needed to tweak
1367 this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001368
1369tune.maxpollevents <number>
1370 Sets the maximum amount of events that can be processed at once in a call to
1371 the polling system. The default value is adapted to the operating system. It
1372 has been noticed that reducing it below 200 tends to slightly decrease
1373 latency at the expense of network bandwidth, and increasing it above 200
1374 tends to trade latency for slightly increased bandwidth.
1375
Willy Tarreau27a674e2009-08-17 07:23:33 +02001376tune.maxrewrite <number>
1377 Sets the reserved buffer space to this size in bytes. The reserved space is
1378 used for header rewriting or appending. The first reads on sockets will never
1379 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
1380 bufsize, though that does not make much sense since there are rarely large
1381 numbers of headers to add. Setting it too high prevents processing of large
1382 requests or responses. Setting it too low prevents addition of new headers
1383 to already large requests or to POST requests. It is generally wise to set it
1384 to about 1024. It is automatically readjusted to half of bufsize if it is
1385 larger than that. This means you don't have to worry about it when changing
1386 bufsize.
1387
Willy Tarreauf3045d22015-04-29 16:24:50 +02001388tune.pattern.cache-size <number>
1389 Sets the size of the pattern lookup cache to <number> entries. This is an LRU
1390 cache which reminds previous lookups and their results. It is used by ACLs
1391 and maps on slow pattern lookups, namely the ones using the "sub", "reg",
1392 "dir", "dom", "end", "bin" match methods as well as the case-insensitive
1393 strings. It applies to pattern expressions which means that it will be able
1394 to memorize the result of a lookup among all the patterns specified on a
1395 configuration line (including all those loaded from files). It automatically
1396 invalidates entries which are updated using HTTP actions or on the CLI. The
1397 default cache size is set to 10000 entries, which limits its footprint to
1398 about 5 MB on 32-bit systems and 8 MB on 64-bit systems. There is a very low
1399 risk of collision in this cache, which is in the order of the size of the
1400 cache divided by 2^64. Typically, at 10000 requests per second with the
1401 default cache size of 10000 entries, there's 1% chance that a brute force
1402 attack could cause a single collision after 60 years, or 0.1% after 6 years.
1403 This is considered much lower than the risk of a memory corruption caused by
1404 aging components. If this is not acceptable, the cache can be disabled by
1405 setting this parameter to 0.
1406
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02001407tune.pipesize <number>
1408 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
1409 are the default size for the system. But sometimes when using TCP splicing,
1410 it can improve performance to increase pipe sizes, especially if it is
1411 suspected that pipes are not filled and that many calls to splice() are
1412 performed. This has an impact on the kernel's memory footprint, so this must
1413 not be changed if impacts are not understood.
1414
Willy Tarreaue803de22010-01-21 17:43:04 +01001415tune.rcvbuf.client <number>
1416tune.rcvbuf.server <number>
1417 Forces the kernel socket receive buffer size on the client or the server side
1418 to the specified value in bytes. This value applies to all TCP/HTTP frontends
1419 and backends. It should normally never be set, and the default size (0) lets
1420 the kernel autotune this value depending on the amount of available memory.
1421 However it can sometimes help to set it to very low values (eg: 4096) in
1422 order to save kernel memory by preventing it from buffering too large amounts
1423 of received data. Lower values will significantly increase CPU usage though.
1424
Willy Tarreaub22fc302015-12-14 12:04:35 +01001425tune.recv_enough <number>
1426 Haproxy uses some hints to detect that a short read indicates the end of the
1427 socket buffers. One of them is that a read returns more than <recv_enough>
1428 bytes, which defaults to 10136 (7 segments of 1448 each). This default value
1429 may be changed by this setting to better deal with workloads involving lots
1430 of short messages such as telnet or SSH sessions.
1431
Willy Tarreaue803de22010-01-21 17:43:04 +01001432tune.sndbuf.client <number>
1433tune.sndbuf.server <number>
1434 Forces the kernel socket send buffer size on the client or the server side to
1435 the specified value in bytes. This value applies to all TCP/HTTP frontends
1436 and backends. It should normally never be set, and the default size (0) lets
1437 the kernel autotune this value depending on the amount of available memory.
1438 However it can sometimes help to set it to very low values (eg: 4096) in
1439 order to save kernel memory by preventing it from buffering too large amounts
1440 of received data. Lower values will significantly increase CPU usage though.
1441 Another use case is to prevent write timeouts with extremely slow clients due
1442 to the kernel waiting for a large part of the buffer to be read before
1443 notifying haproxy again.
1444
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001445tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +01001446 Sets the size of the global SSL session cache, in a number of blocks. A block
1447 is large enough to contain an encoded session without peer certificate.
1448 An encoded session with peer certificate is stored in multiple blocks
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001449 depending on the size of the peer certificate. A block uses approximately
Emeric Brunaf9619d2012-11-28 18:47:52 +01001450 200 bytes of memory. The default value may be forced at build time, otherwise
1451 defaults to 20000. When the cache is full, the most idle entries are purged
1452 and reassigned. Higher values reduce the occurrence of such a purge, hence
1453 the number of CPU-intensive SSL handshakes by ensuring that all users keep
1454 their session as long as possible. All entries are pre-allocated upon startup
Emeric Brun22890a12012-12-28 14:41:32 +01001455 and are shared between all processes if "nbproc" is greater than 1. Setting
1456 this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001457
Emeric Brun8dc60392014-05-09 13:52:00 +02001458tune.ssl.force-private-cache
1459 This boolean disables SSL session cache sharing between all processes. It
1460 should normally not be used since it will force many renegotiations due to
1461 clients hitting a random process. But it may be required on some operating
1462 systems where none of the SSL cache synchronization method may be used. In
1463 this case, adding a first layer of hash-based load balancing before the SSL
1464 layer might limit the impact of the lack of session sharing.
1465
Emeric Brun4f65bff2012-11-16 15:11:00 +01001466tune.ssl.lifetime <timeout>
1467 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001468 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01001469 does not guarantee that sessions will last that long, because if the cache is
1470 full, the longest idle sessions will be purged despite their configured
1471 lifetime. The real usefulness of this setting is to prevent sessions from
1472 being used for too long.
1473
Willy Tarreaubfd59462013-02-21 07:46:09 +01001474tune.ssl.maxrecord <number>
1475 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
1476 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
1477 data only once it has received a full record. With large records, it means
1478 that clients might have to download up to 16kB of data before starting to
1479 process them. Limiting the value can improve page load times on browsers
1480 located over high latency or low bandwidth networks. It is suggested to find
1481 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
1482 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
1483 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
1484 2859 gave good results during tests. Use "strace -e trace=write" to find the
Willy Tarreau7e312732014-02-12 16:35:14 +01001485 best value. Haproxy will automatically switch to this setting after an idle
1486 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01001487
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001488tune.ssl.default-dh-param <number>
1489 Sets the maximum size of the Diffie-Hellman parameters used for generating
1490 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
1491 final size will try to match the size of the server's RSA (or DSA) key (e.g,
1492 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
1493 this maximum value. Default value if 1024. Only 1024 or higher values are
1494 allowed. Higher values will increase the CPU load, and values greater than
1495 1024 bits are not supported by Java 7 and earlier clients. This value is not
Remi Gacogne47783ef2015-05-29 15:53:22 +02001496 used if static Diffie-Hellman parameters are supplied either directly
1497 in the certificate file or by using the ssl-dh-param-file parameter.
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001498
Christopher Faulet31af49d2015-06-09 17:29:50 +02001499tune.ssl.ssl-ctx-cache-size <number>
1500 Sets the size of the cache used to store generated certificates to <number>
1501 entries. This is a LRU cache. Because generating a SSL certificate
1502 dynamically is expensive, they are cached. The default cache size is set to
1503 1000 entries.
1504
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001505tune.vars.global-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001506tune.vars.proc-max-size <size>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001507tune.vars.reqres-max-size <size>
1508tune.vars.sess-max-size <size>
1509tune.vars.txn-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001510 These five tunes help to manage the maximum amount of memory used by the
1511 variables system. "global" limits the overall amount of memory available for
1512 all scopes. "proc" limits the memory for the process scope, "sess" limits the
1513 memory for the session scope, "txn" for the transaction scope, and "reqres"
1514 limits the memory for each request or response processing.
1515 Memory accounting is hierarchical, meaning more coarse grained limits include
1516 the finer grained ones: "proc" includes "sess", "sess" includes "txn", and
1517 "txn" includes "reqres".
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001518
Daniel Schneller0b547052016-03-21 20:46:57 +01001519 For example, when "tune.vars.sess-max-size" is limited to 100,
1520 "tune.vars.txn-max-size" and "tune.vars.reqres-max-size" cannot exceed
1521 100 either. If we create a variable "txn.var" that contains 100 bytes,
1522 all available space is consumed.
1523 Notice that exceeding the limits at runtime will not result in an error
1524 message, but values might be cut off or corrupted. So make sure to accurately
1525 plan for the amount of space needed to store all your variables.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001526
William Lallemanda509e4c2012-11-07 16:54:34 +01001527tune.zlib.memlevel <number>
1528 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001529 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01001530 state. A value of 1 uses minimum memory but is slow and reduces compression
1531 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
1532 between 1 and 9. The default value is 8.
1533
1534tune.zlib.windowsize <number>
1535 Sets the window size (the size of the history buffer) as a parameter of the
1536 zlib initialization for each session. Larger values of this parameter result
1537 in better compression at the expense of memory usage. Can be a value between
1538 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001539
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015403.3. Debugging
1541--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02001542
1543debug
1544 Enables debug mode which dumps to stdout all exchanges, and disables forking
1545 into background. It is the equivalent of the command-line argument "-d". It
1546 should never be used in a production configuration since it may prevent full
1547 system startup.
1548
1549quiet
1550 Do not display any message during startup. It is equivalent to the command-
1551 line argument "-q".
1552
Emeric Brunf099e792010-09-27 12:05:28 +02001553
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010015543.4. Userlists
1555--------------
1556It is possible to control access to frontend/backend/listen sections or to
1557http stats by allowing only authenticated and authorized users. To do this,
1558it is required to create at least one userlist and to define users.
1559
1560userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01001561 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001562 used to store authentication & authorization data for independent customers.
1563
1564group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01001565 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001566 attach users to this group by using a comma separated list of names
1567 proceeded by "users" keyword.
1568
Cyril Bontéf0c60612010-02-06 14:44:47 +01001569user <username> [password|insecure-password <password>]
1570 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001571 Adds user <username> to the current userlist. Both secure (encrypted) and
1572 insecure (unencrypted) passwords can be used. Encrypted passwords are
Cyril Bonté78caf842010-03-10 22:41:43 +01001573 evaluated using the crypt(3) function so depending of the system's
1574 capabilities, different algorithms are supported. For example modern Glibc
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001575 based Linux system supports MD5, SHA-256, SHA-512 and of course classic,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001576 DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001577
1578
1579 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01001580 userlist L1
1581 group G1 users tiger,scott
1582 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001583
Cyril Bontéf0c60612010-02-06 14:44:47 +01001584 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
1585 user scott insecure-password elgato
1586 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001587
Cyril Bontéf0c60612010-02-06 14:44:47 +01001588 userlist L2
1589 group G1
1590 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001591
Cyril Bontéf0c60612010-02-06 14:44:47 +01001592 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
1593 user scott insecure-password elgato groups G1,G2
1594 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001595
1596 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001597
Emeric Brunf099e792010-09-27 12:05:28 +02001598
15993.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02001600----------
Emeric Brun94900952015-06-11 18:25:54 +02001601It is possible to propagate entries of any data-types in stick-tables between
1602several haproxy instances over TCP connections in a multi-master fashion. Each
1603instance pushes its local updates and insertions to remote peers. The pushed
1604values overwrite remote ones without aggregation. Interrupted exchanges are
1605automatically detected and recovered from the last known point.
1606In addition, during a soft restart, the old process connects to the new one
1607using such a TCP connection to push all its entries before the new process
1608tries to connect to other peers. That ensures very fast replication during a
1609reload, it typically takes a fraction of a second even for large tables.
1610Note that Server IDs are used to identify servers remotely, so it is important
1611that configurations look similar or at least that the same IDs are forced on
1612each server on all participants.
Emeric Brunf099e792010-09-27 12:05:28 +02001613
1614peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001615 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02001616 which is referenced by one or more stick-tables.
1617
Willy Tarreau77e4bd12015-05-01 20:02:17 +02001618disabled
1619 Disables a peers section. It disables both listening and any synchronization
1620 related to this section. This is provided to disable synchronization of stick
1621 tables without having to comment out all "peers" references.
1622
1623enable
1624 This re-enables a disabled peers section which was previously disabled.
1625
Emeric Brunf099e792010-09-27 12:05:28 +02001626peer <peername> <ip>:<port>
1627 Defines a peer inside a peers section.
1628 If <peername> is set to the local peer name (by default hostname, or forced
1629 using "-L" command line option), haproxy will listen for incoming remote peer
1630 connection on <ip>:<port>. Otherwise, <ip>:<port> defines where to connect to
1631 to join the remote peer, and <peername> is used at the protocol level to
1632 identify and validate the remote peer on the server side.
1633
1634 During a soft restart, local peer <ip>:<port> is used by the old instance to
1635 connect the new one and initiate a complete replication (teaching process).
1636
1637 It is strongly recommended to have the exact same peers declaration on all
1638 peers and to only rely on the "-L" command line argument to change the local
1639 peer name. This makes it easier to maintain coherent configuration files
1640 across all peers.
1641
William Lallemandb2f07452015-05-12 14:27:13 +02001642 You may want to reference some environment variables in the address
1643 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01001644
Cyril Bontédc4d9032012-04-08 21:57:39 +02001645 Example:
Emeric Brunf099e792010-09-27 12:05:28 +02001646 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001647 peer haproxy1 192.168.0.1:1024
1648 peer haproxy2 192.168.0.2:1024
1649 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02001650
1651 backend mybackend
1652 mode tcp
1653 balance roundrobin
1654 stick-table type ip size 20k peers mypeers
1655 stick on src
1656
Willy Tarreauf7b30a92010-12-06 22:59:17 +01001657 server srv1 192.168.0.30:80
1658 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02001659
1660
Simon Horman51a1cf62015-02-03 13:00:44 +090016613.6. Mailers
1662------------
1663It is possible to send email alerts when the state of servers changes.
1664If configured email alerts are sent to each mailer that is configured
1665in a mailers section. Email is sent to mailers using SMTP.
1666
Pieter Baauw386a1272015-08-16 15:26:24 +02001667mailers <mailersect>
Simon Horman51a1cf62015-02-03 13:00:44 +09001668 Creates a new mailer list with the name <mailersect>. It is an
1669 independent section which is referenced by one or more proxies.
1670
1671mailer <mailername> <ip>:<port>
1672 Defines a mailer inside a mailers section.
1673
1674 Example:
1675 mailers mymailers
1676 mailer smtp1 192.168.0.1:587
1677 mailer smtp2 192.168.0.2:587
1678
1679 backend mybackend
1680 mode tcp
1681 balance roundrobin
1682
1683 email-alert mailers mymailers
1684 email-alert from test1@horms.org
1685 email-alert to test2@horms.org
1686
1687 server srv1 192.168.0.30:80
1688 server srv2 192.168.0.31:80
1689
Pieter Baauw235fcfc2016-02-13 15:33:40 +01001690timeout mail <time>
1691 Defines the time available for a mail/connection to be made and send to
1692 the mail-server. If not defined the default value is 10 seconds. To allow
1693 for at least two SYN-ACK packets to be send during initial TCP handshake it
1694 is advised to keep this value above 4 seconds.
1695
1696 Example:
1697 mailers mymailers
1698 timeout mail 20s
1699 mailer smtp1 192.168.0.1:587
Simon Horman51a1cf62015-02-03 13:00:44 +09001700
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017014. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02001702----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001703
Willy Tarreau6a06a402007-07-15 20:15:28 +02001704Proxy configuration can be located in a set of sections :
William Lallemand6e62fb62015-04-28 16:55:23 +02001705 - defaults [<name>]
Willy Tarreau6a06a402007-07-15 20:15:28 +02001706 - frontend <name>
1707 - backend <name>
1708 - listen <name>
1709
1710A "defaults" section sets default parameters for all other sections following
1711its declaration. Those default parameters are reset by the next "defaults"
1712section. See below for the list of parameters which can be set in a "defaults"
Willy Tarreau0ba27502007-12-24 16:55:16 +01001713section. The name is optional but its use is encouraged for better readability.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001714
1715A "frontend" section describes a set of listening sockets accepting client
1716connections.
1717
1718A "backend" section describes a set of servers to which the proxy will connect
1719to forward incoming connections.
1720
1721A "listen" section defines a complete proxy with its frontend and backend
1722parts combined in one section. It is generally useful for TCP-only traffic.
1723
Willy Tarreau0ba27502007-12-24 16:55:16 +01001724All proxy names must be formed from upper and lower case letters, digits,
1725'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
1726case-sensitive, which means that "www" and "WWW" are two different proxies.
1727
1728Historically, all proxy names could overlap, it just caused troubles in the
1729logs. Since the introduction of content switching, it is mandatory that two
1730proxies with overlapping capabilities (frontend/backend) have different names.
1731However, it is still permitted that a frontend and a backend share the same
1732name, as this configuration seems to be commonly encountered.
1733
1734Right now, two major proxy modes are supported : "tcp", also known as layer 4,
1735and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001736bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01001737protocol, and can interact with it by allowing, blocking, switching, adding,
1738modifying, or removing arbitrary contents in requests or responses, based on
1739arbitrary criteria.
1740
Willy Tarreau70dffda2014-01-30 03:07:23 +01001741In HTTP mode, the processing applied to requests and responses flowing over
1742a connection depends in the combination of the frontend's HTTP options and
1743the backend's. HAProxy supports 5 connection modes :
1744
1745 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
1746 requests and responses are processed, and connections remain open but idle
1747 between responses and new requests.
1748
1749 - TUN: tunnel ("option http-tunnel") : this was the default mode for versions
1750 1.0 to 1.5-dev21 : only the first request and response are processed, and
1751 everything else is forwarded with no analysis at all. This mode should not
1752 be used as it creates lots of trouble with logging and HTTP processing.
1753
1754 - PCL: passive close ("option httpclose") : exactly the same as tunnel mode,
1755 but with "Connection: close" appended in both directions to try to make
1756 both ends close after the first request/response exchange.
1757
1758 - SCL: server close ("option http-server-close") : the server-facing
1759 connection is closed after the end of the response is received, but the
1760 client-facing connection remains open.
1761
1762 - FCL: forced close ("option forceclose") : the connection is actively closed
1763 after the end of the response.
1764
1765The effective mode that will be applied to a connection passing through a
1766frontend and a backend can be determined by both proxy modes according to the
1767following matrix, but in short, the modes are symmetric, keep-alive is the
1768weakest option and force close is the strongest.
1769
1770 Backend mode
1771
1772 | KAL | TUN | PCL | SCL | FCL
1773 ----+-----+-----+-----+-----+----
1774 KAL | KAL | TUN | PCL | SCL | FCL
1775 ----+-----+-----+-----+-----+----
1776 TUN | TUN | TUN | PCL | SCL | FCL
1777 Frontend ----+-----+-----+-----+-----+----
1778 mode PCL | PCL | PCL | PCL | FCL | FCL
1779 ----+-----+-----+-----+-----+----
1780 SCL | SCL | SCL | FCL | SCL | FCL
1781 ----+-----+-----+-----+-----+----
1782 FCL | FCL | FCL | FCL | FCL | FCL
1783
Willy Tarreau0ba27502007-12-24 16:55:16 +01001784
Willy Tarreau70dffda2014-01-30 03:07:23 +01001785
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017864.1. Proxy keywords matrix
1787--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001788
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001789The following list of keywords is supported. Most of them may only be used in a
1790limited set of section types. Some of them are marked as "deprecated" because
1791they are inherited from an old syntax which may be confusing or functionally
1792limited, and there are new recommended keywords to replace them. Keywords
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001793marked with "(*)" can be optionally inverted using the "no" prefix, eg. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001794option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02001795and must be disabled for a specific instance. Such options may also be prefixed
1796with "default" in order to restore default settings regardless of what has been
1797specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01001798
Willy Tarreau6a06a402007-07-15 20:15:28 +02001799
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001800 keyword defaults frontend listen backend
1801------------------------------------+----------+----------+---------+---------
1802acl - X X X
Willy Tarreau294d0f02015-08-10 19:40:12 +02001803appsession - - - -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001804backlog X X X -
1805balance X - X X
1806bind - X X -
1807bind-process X X X X
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02001808block (deprecated) - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001809capture cookie - X X -
1810capture request header - X X -
1811capture response header - X X -
1812clitimeout (deprecated) X X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02001813compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001814contimeout (deprecated) X - X X
1815cookie X - X X
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02001816declare capture - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001817default-server X - X X
1818default_backend X X X -
1819description - X X X
1820disabled X X X X
1821dispatch - - X X
Simon Horman51a1cf62015-02-03 13:00:44 +09001822email-alert from X X X X
Simon Horman64e34162015-02-06 11:11:57 +09001823email-alert level X X X X
Simon Horman51a1cf62015-02-03 13:00:44 +09001824email-alert mailers X X X X
1825email-alert myhostname X X X X
1826email-alert to X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001827enabled X X X X
1828errorfile X X X X
1829errorloc X X X X
1830errorloc302 X X X X
1831-- keyword -------------------------- defaults - frontend - listen -- backend -
1832errorloc303 X X X X
Cyril Bonté0d4bf012010-04-25 23:21:46 +02001833force-persist - X X X
Christopher Fauletc3fe5332016-04-07 15:30:10 +02001834filter - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001835fullconn X - X X
1836grace X X X X
1837hash-type X - X X
1838http-check disable-on-404 X - X X
Willy Tarreaubd741542010-03-16 18:46:54 +01001839http-check expect - - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02001840http-check send-state X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001841http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02001842http-response - X X X
Willy Tarreau30631952015-08-06 15:05:24 +02001843http-reuse X - X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02001844http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001845id - X X X
Cyril Bonté0d4bf012010-04-25 23:21:46 +02001846ignore-persist - X X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001847load-server-state-from-file X - X X
William Lallemand0f99e342011-10-12 17:50:54 +02001848log (*) X X X X
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01001849log-format X X X -
Dragan Dosen7ad31542015-09-28 17:16:47 +02001850log-format-sd X X X -
Willy Tarreau094af4e2015-01-07 15:03:42 +01001851log-tag X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02001852max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001853maxconn X X X -
1854mode X X X X
1855monitor fail - X X -
1856monitor-net X X X -
1857monitor-uri X X X -
1858option abortonclose (*) X - X X
1859option accept-invalid-http-request (*) X X X -
1860option accept-invalid-http-response (*) X - X X
1861option allbackups (*) X - X X
1862option checkcache (*) X - X X
1863option clitcpka (*) X X X -
1864option contstats (*) X X X -
1865option dontlog-normal (*) X X X -
1866option dontlognull (*) X X X -
1867option forceclose (*) X X X X
1868-- keyword -------------------------- defaults - frontend - listen -- backend -
1869option forwardfor X X X X
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02001870option http-buffer-request (*) X X X X
Willy Tarreau82649f92015-05-01 22:40:51 +02001871option http-ignore-probes (*) X X X -
Willy Tarreau16bfb022010-01-16 19:48:41 +01001872option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02001873option http-no-delay (*) X X X X
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02001874option http-pretend-keepalive (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001875option http-server-close (*) X X X X
Willy Tarreau02bce8b2014-01-30 00:15:28 +01001876option http-tunnel (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001877option http-use-proxy-header (*) X X X -
1878option httpchk X - X X
1879option httpclose (*) X X X X
1880option httplog X X X X
1881option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04001882option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02001883option ldap-check X - X X
Simon Horman98637e52014-06-20 12:30:16 +09001884option external-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001885option log-health-checks (*) X - X X
1886option log-separate-errors (*) X X X -
1887option logasap (*) X X X -
1888option mysql-check X - X X
1889option nolinger (*) X X X X
1890option originalto X X X X
1891option persist (*) X - X X
Baptiste Assmann809e22a2015-10-12 20:22:55 +02001892option pgsql-check X - X X
1893option prefer-last-server (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001894option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02001895option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001896option smtpchk X - X X
1897option socket-stats (*) X X X -
1898option splice-auto (*) X X X X
1899option splice-request (*) X X X X
1900option splice-response (*) X X X X
Christopher Fauletba7bc162016-11-07 21:07:38 +01001901option spop-check - - - X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001902option srvtcpka (*) X - X X
1903option ssl-hello-chk X - X X
1904-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01001905option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001906option tcp-smart-accept (*) X X X -
1907option tcp-smart-connect (*) X - X X
1908option tcpka X X X X
1909option tcplog X X X X
1910option transparent (*) X - X X
Simon Horman98637e52014-06-20 12:30:16 +09001911external-check command X - X X
1912external-check path X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001913persist rdp-cookie X - X X
1914rate-limit sessions X X X -
1915redirect - X X X
1916redisp (deprecated) X - X X
1917redispatch (deprecated) X - X X
1918reqadd - X X X
1919reqallow - X X X
1920reqdel - X X X
1921reqdeny - X X X
1922reqiallow - X X X
1923reqidel - X X X
1924reqideny - X X X
1925reqipass - X X X
1926reqirep - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001927reqitarpit - X X X
1928reqpass - X X X
1929reqrep - X X X
1930-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001931reqtarpit - X X X
1932retries X - X X
1933rspadd - X X X
1934rspdel - X X X
1935rspdeny - X X X
1936rspidel - X X X
1937rspideny - X X X
1938rspirep - X X X
1939rsprep - X X X
1940server - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001941server-state-file-name X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001942source X - X X
1943srvtimeout (deprecated) X - X X
Baptiste Assmann5a549212015-10-12 20:30:24 +02001944stats admin - X X X
1945stats auth X X X X
1946stats enable X X X X
1947stats hide-version X X X X
1948stats http-request - X X X
1949stats realm X X X X
1950stats refresh X X X X
1951stats scope X X X X
1952stats show-desc X X X X
1953stats show-legends X X X X
1954stats show-node X X X X
1955stats uri X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001956-- keyword -------------------------- defaults - frontend - listen -- backend -
1957stick match - - X X
1958stick on - - X X
1959stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02001960stick store-response - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001961stick-table - - X X
Willy Tarreau938c7fe2014-04-25 14:21:39 +02001962tcp-check connect - - X X
1963tcp-check expect - - X X
1964tcp-check send - - X X
1965tcp-check send-binary - - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02001966tcp-request connection - X X -
1967tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02001968tcp-request inspect-delay - X X X
Willy Tarreau4f614292016-10-21 17:49:36 +02001969tcp-request session - X X -
Emeric Brun0a3b67f2010-09-24 15:34:53 +02001970tcp-response content - - X X
1971tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001972timeout check X - X X
1973timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02001974timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001975timeout clitimeout (deprecated) X X X -
1976timeout connect X - X X
1977timeout contimeout (deprecated) X - X X
1978timeout http-keep-alive X X X X
1979timeout http-request X X X X
1980timeout queue X - X X
1981timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02001982timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001983timeout srvtimeout (deprecated) X - X X
1984timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02001985timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001986transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01001987unique-id-format X X X -
1988unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001989use_backend - X X -
Willy Tarreau4a5cade2012-04-05 21:09:48 +02001990use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01001991------------------------------------+----------+----------+---------+---------
1992 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02001993
Willy Tarreau0ba27502007-12-24 16:55:16 +01001994
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019954.2. Alphabetically sorted keywords reference
1996---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01001997
1998This section provides a description of each keyword and its usage.
1999
2000
2001acl <aclname> <criterion> [flags] [operator] <value> ...
2002 Declare or complete an access list.
2003 May be used in sections : defaults | frontend | listen | backend
2004 no | yes | yes | yes
2005 Example:
2006 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
2007 acl invalid_src src_port 0:1023
2008 acl local_dst hdr(host) -i localhost
2009
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002010 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002011
2012
Cyril Bontéb21570a2009-11-29 20:04:48 +01002013appsession <cookie> len <length> timeout <holdtime>
2014 [request-learn] [prefix] [mode <path-parameters|query-string>]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002015 Define session stickiness on an existing application cookie.
2016 May be used in sections : defaults | frontend | listen | backend
2017 no | no | yes | yes
2018 Arguments :
2019 <cookie> this is the name of the cookie used by the application and which
2020 HAProxy will have to learn for each new session.
2021
Cyril Bontéb21570a2009-11-29 20:04:48 +01002022 <length> this is the max number of characters that will be memorized and
Willy Tarreau0ba27502007-12-24 16:55:16 +01002023 checked in each cookie value.
2024
2025 <holdtime> this is the time after which the cookie will be removed from
2026 memory if unused. If no unit is specified, this time is in
2027 milliseconds.
2028
Cyril Bontébf47aeb2009-10-15 00:15:40 +02002029 request-learn
2030 If this option is specified, then haproxy will be able to learn
2031 the cookie found in the request in case the server does not
2032 specify any in response. This is typically what happens with
2033 PHPSESSID cookies, or when haproxy's session expires before
2034 the application's session and the correct server is selected.
2035 It is recommended to specify this option to improve reliability.
2036
Cyril Bontéb21570a2009-11-29 20:04:48 +01002037 prefix When this option is specified, haproxy will match on the cookie
2038 prefix (or URL parameter prefix). The appsession value is the
2039 data following this prefix.
2040
2041 Example :
2042 appsession ASPSESSIONID len 64 timeout 3h prefix
2043
2044 This will match the cookie ASPSESSIONIDXXXX=XXXXX,
2045 the appsession value will be XXXX=XXXXX.
2046
2047 mode This option allows to change the URL parser mode.
2048 2 modes are currently supported :
2049 - path-parameters :
2050 The parser looks for the appsession in the path parameters
2051 part (each parameter is separated by a semi-colon), which is
2052 convenient for JSESSIONID for example.
2053 This is the default mode if the option is not set.
2054 - query-string :
2055 In this mode, the parser will look for the appsession in the
2056 query string.
2057
Willy Tarreau294d0f02015-08-10 19:40:12 +02002058 As of version 1.6, appsessions was removed. It is more flexible and more
2059 convenient to use stick-tables instead, and stick-tables support multi-master
2060 replication and data conservation across reloads, which appsessions did not.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002061
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01002062 See also : "cookie", "capture cookie", "balance", "stick", "stick-table",
2063 "ignore-persist", "nbproc" and "bind-process".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002064
2065
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002066backlog <conns>
2067 Give hints to the system about the approximate listen backlog desired size
2068 May be used in sections : defaults | frontend | listen | backend
2069 yes | yes | yes | no
2070 Arguments :
2071 <conns> is the number of pending connections. Depending on the operating
2072 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02002073 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002074
2075 In order to protect against SYN flood attacks, one solution is to increase
2076 the system's SYN backlog size. Depending on the system, sometimes it is just
2077 tunable via a system parameter, sometimes it is not adjustable at all, and
2078 sometimes the system relies on hints given by the application at the time of
2079 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
2080 to the listen() syscall. On systems which can make use of this value, it can
2081 sometimes be useful to be able to specify a different value, hence this
2082 backlog parameter.
2083
2084 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
2085 used as a hint and the system accepts up to the smallest greater power of
2086 two, and never more than some limits (usually 32768).
2087
2088 See also : "maxconn" and the target operating system's tuning guide.
2089
2090
Willy Tarreau0ba27502007-12-24 16:55:16 +01002091balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02002092balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002093 Define the load balancing algorithm to be used in a backend.
2094 May be used in sections : defaults | frontend | listen | backend
2095 yes | no | yes | yes
2096 Arguments :
2097 <algorithm> is the algorithm used to select a server when doing load
2098 balancing. This only applies when no persistence information
2099 is available, or when a connection is redispatched to another
2100 server. <algorithm> may be one of the following :
2101
2102 roundrobin Each server is used in turns, according to their weights.
2103 This is the smoothest and fairest algorithm when the server's
2104 processing time remains equally distributed. This algorithm
2105 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02002106 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08002107 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02002108 large farms, when a server becomes up after having been down
2109 for a very short time, it may sometimes take a few hundreds
2110 requests for it to be re-integrated into the farm and start
2111 receiving traffic. This is normal, though very rare. It is
2112 indicated here in case you would have the chance to observe
2113 it, so that you don't worry.
2114
2115 static-rr Each server is used in turns, according to their weights.
2116 This algorithm is as similar to roundrobin except that it is
2117 static, which means that changing a server's weight on the
2118 fly will have no effect. On the other hand, it has no design
2119 limitation on the number of servers, and when a server goes
2120 up, it is always immediately reintroduced into the farm, once
2121 the full map is recomputed. It also uses slightly less CPU to
2122 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01002123
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01002124 leastconn The server with the lowest number of connections receives the
2125 connection. Round-robin is performed within groups of servers
2126 of the same load to ensure that all servers will be used. Use
2127 of this algorithm is recommended where very long sessions are
2128 expected, such as LDAP, SQL, TSE, etc... but is not very well
2129 suited for protocols using short sessions such as HTTP. This
2130 algorithm is dynamic, which means that server weights may be
2131 adjusted on the fly for slow starts for instance.
2132
Willy Tarreauf09c6602012-02-13 17:12:08 +01002133 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002134 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01002135 identifier to the highest (see server parameter "id"), which
2136 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02002137 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01002138 not make sense to use this algorithm without setting maxconn.
2139 The purpose of this algorithm is to always use the smallest
2140 number of servers so that extra servers can be powered off
2141 during non-intensive hours. This algorithm ignores the server
2142 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02002143 or IMAP than HTTP, though it can be useful there too. In
2144 order to use this algorithm efficiently, it is recommended
2145 that a cloud controller regularly checks server usage to turn
2146 them off when unused, and regularly checks backend queue to
2147 turn new servers on when the queue inflates. Alternatively,
2148 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01002149
Willy Tarreau0ba27502007-12-24 16:55:16 +01002150 source The source IP address is hashed and divided by the total
2151 weight of the running servers to designate which server will
2152 receive the request. This ensures that the same client IP
2153 address will always reach the same server as long as no
2154 server goes down or up. If the hash result changes due to the
2155 number of running servers changing, many clients will be
2156 directed to a different server. This algorithm is generally
2157 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002158 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01002159 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002160 static by default, which means that changing a server's
2161 weight on the fly will have no effect, but this can be
2162 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002163
Oskar Stolc8dc41842012-05-19 10:19:54 +01002164 uri This algorithm hashes either the left part of the URI (before
2165 the question mark) or the whole URI (if the "whole" parameter
2166 is present) and divides the hash value by the total weight of
2167 the running servers. The result designates which server will
2168 receive the request. This ensures that the same URI will
2169 always be directed to the same server as long as no server
2170 goes up or down. This is used with proxy caches and
2171 anti-virus proxies in order to maximize the cache hit rate.
2172 Note that this algorithm may only be used in an HTTP backend.
2173 This algorithm is static by default, which means that
2174 changing a server's weight on the fly will have no effect,
2175 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002176
Oskar Stolc8dc41842012-05-19 10:19:54 +01002177 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002178 "depth", both followed by a positive integer number. These
2179 options may be helpful when it is needed to balance servers
2180 based on the beginning of the URI only. The "len" parameter
2181 indicates that the algorithm should only consider that many
2182 characters at the beginning of the URI to compute the hash.
2183 Note that having "len" set to 1 rarely makes sense since most
2184 URIs start with a leading "/".
2185
2186 The "depth" parameter indicates the maximum directory depth
2187 to be used to compute the hash. One level is counted for each
2188 slash in the request. If both parameters are specified, the
2189 evaluation stops when either is reached.
2190
Willy Tarreau0ba27502007-12-24 16:55:16 +01002191 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002192 the query string of each HTTP GET request.
2193
2194 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02002195 request entity will be searched for the parameter argument,
2196 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02002197 ('?') in the URL. The message body will only start to be
2198 analyzed once either the advertised amount of data has been
2199 received or the request buffer is full. In the unlikely event
2200 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02002201 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02002202 be randomly balanced if at all. This keyword used to support
2203 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002204
2205 If the parameter is found followed by an equal sign ('=') and
2206 a value, then the value is hashed and divided by the total
2207 weight of the running servers. The result designates which
2208 server will receive the request.
2209
2210 This is used to track user identifiers in requests and ensure
2211 that a same user ID will always be sent to the same server as
2212 long as no server goes up or down. If no value is found or if
2213 the parameter is not found, then a round robin algorithm is
2214 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002215 backend. This algorithm is static by default, which means
2216 that changing a server's weight on the fly will have no
2217 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002218
Cyril Bontédc4d9032012-04-08 21:57:39 +02002219 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
2220 request. Just as with the equivalent ACL 'hdr()' function,
2221 the header name in parenthesis is not case sensitive. If the
2222 header is absent or if it does not contain any value, the
2223 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01002224
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002225 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01002226 reducing the hash algorithm to the main domain part with some
2227 specific headers such as 'Host'. For instance, in the Host
2228 value "haproxy.1wt.eu", only "1wt" will be considered.
2229
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002230 This algorithm is static by default, which means that
2231 changing a server's weight on the fly will have no effect,
2232 but this can be changed using "hash-type".
2233
Emeric Brun736aa232009-06-30 17:56:00 +02002234 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02002235 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02002236 The RDP cookie <name> (or "mstshash" if omitted) will be
2237 looked up and hashed for each incoming TCP request. Just as
2238 with the equivalent ACL 'req_rdp_cookie()' function, the name
2239 is not case-sensitive. This mechanism is useful as a degraded
2240 persistence mode, as it makes it possible to always send the
2241 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002242 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02002243 used instead.
2244
2245 Note that for this to work, the frontend must ensure that an
2246 RDP cookie is already present in the request buffer. For this
2247 you must use 'tcp-request content accept' rule combined with
2248 a 'req_rdp_cookie_cnt' ACL.
2249
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002250 This algorithm is static by default, which means that
2251 changing a server's weight on the fly will have no effect,
2252 but this can be changed using "hash-type".
2253
Cyril Bontédc4d9032012-04-08 21:57:39 +02002254 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09002255
Willy Tarreau0ba27502007-12-24 16:55:16 +01002256 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002257 algorithms. Right now, only "url_param" and "uri" support an
2258 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002259
Willy Tarreau3cd9af22009-03-15 14:06:41 +01002260 The load balancing algorithm of a backend is set to roundrobin when no other
2261 algorithm, mode nor option have been set. The algorithm may only be set once
2262 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002263
2264 Examples :
2265 balance roundrobin
2266 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002267 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01002268 balance hdr(User-Agent)
2269 balance hdr(host)
2270 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002271
2272 Note: the following caveats and limitations on using the "check_post"
2273 extension with "url_param" must be considered :
2274
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002275 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002276 to determine if the parameters will be found in the body or entity which
2277 may contain binary data. Therefore another method may be required to
2278 restrict consideration of POST requests that have no URL parameters in
2279 the body. (see acl reqideny http_end)
2280
2281 - using a <max_wait> value larger than the request buffer size does not
2282 make sense and is useless. The buffer size is set at build time, and
2283 defaults to 16 kB.
2284
2285 - Content-Encoding is not supported, the parameter search will probably
2286 fail; and load balancing will fall back to Round Robin.
2287
2288 - Expect: 100-continue is not supported, load balancing will fall back to
2289 Round Robin.
2290
2291 - Transfer-Encoding (RFC2616 3.6.1) is only supported in the first chunk.
2292 If the entire parameter value is not present in the first chunk, the
2293 selection of server is undefined (actually, defined by how little
2294 actually appeared in the first chunk).
2295
2296 - This feature does not support generation of a 100, 411 or 501 response.
2297
2298 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002299 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002300 white space or control characters are found, indicating the end of what
2301 might be a URL parameter list. This is probably not a concern with SGML
2302 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002303
Willy Tarreau294d0f02015-08-10 19:40:12 +02002304 See also : "dispatch", "cookie", "transparent", "hash-type" and "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002305
2306
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002307bind [<address>]:<port_range> [, ...] [param*]
2308bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002309 Define one or several listening addresses and/or ports in a frontend.
2310 May be used in sections : defaults | frontend | listen | backend
2311 no | yes | yes | no
2312 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002313 <address> is optional and can be a host name, an IPv4 address, an IPv6
2314 address, or '*'. It designates the address the frontend will
2315 listen on. If unset, all IPv4 addresses of the system will be
2316 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01002317 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01002318 Optionally, an address family prefix may be used before the
2319 address to force the family regardless of the address format,
2320 which can be useful to specify a path to a unix socket with
2321 no slash ('/'). Currently supported prefixes are :
2322 - 'ipv4@' -> address is always IPv4
2323 - 'ipv6@' -> address is always IPv6
2324 - 'unix@' -> address is a path to a local unix socket
Willy Tarreau70f72e02014-07-08 00:37:50 +02002325 - 'abns@' -> address is in abstract namespace (Linux only).
2326 Note: since abstract sockets are not "rebindable", they
2327 do not cope well with multi-process mode during
2328 soft-restart, so it is better to avoid them if
2329 nbproc is greater than 1. The effect is that if the
2330 new process fails to start, only one of the old ones
2331 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01002332 - 'fd@<n>' -> use file descriptor <n> inherited from the
2333 parent. The fd must be bound and may or may not already
2334 be listening.
William Lallemandb2f07452015-05-12 14:27:13 +02002335 You may want to reference some environment variables in the
2336 address parameter, see section 2.3 about environment
2337 variables.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002338
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002339 <port_range> is either a unique TCP port, or a port range for which the
2340 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002341 above. The port is mandatory for TCP listeners. Note that in
2342 the case of an IPv6 address, the port is always the number
2343 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002344 - a numerical port (ex: '80')
2345 - a dash-delimited ports range explicitly stating the lower
2346 and upper bounds (ex: '2000-2100') which are included in
2347 the range.
2348
2349 Particular care must be taken against port ranges, because
2350 every <address:port> couple consumes one socket (= a file
2351 descriptor), so it's easy to consume lots of descriptors
2352 with a simple range, and to run out of sockets. Also, each
2353 <address:port> couple must be used only once among all
2354 instances running on a same system. Please note that binding
2355 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002356 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002357 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002358
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002359 <path> is a UNIX socket path beginning with a slash ('/'). This is
2360 alternative to the TCP listening port. Haproxy will then
2361 receive UNIX connections on the socket located at this place.
2362 The path must begin with a slash and by default is absolute.
2363 It can be relative to the prefix defined by "unix-bind" in
2364 the global section. Note that the total length of the prefix
2365 followed by the socket path cannot exceed some system limits
2366 for UNIX sockets, which commonly are set to 107 characters.
2367
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002368 <param*> is a list of parameters common to all sockets declared on the
2369 same line. These numerous parameters depend on OS and build
2370 options and have a complete section dedicated to them. Please
2371 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002372
Willy Tarreau0ba27502007-12-24 16:55:16 +01002373 It is possible to specify a list of address:port combinations delimited by
2374 commas. The frontend will then listen on all of these addresses. There is no
2375 fixed limit to the number of addresses and ports which can be listened on in
2376 a frontend, as well as there is no limit to the number of "bind" statements
2377 in a frontend.
2378
2379 Example :
2380 listen http_proxy
2381 bind :80,:443
2382 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002383 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01002384
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002385 listen http_https_proxy
2386 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02002387 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002388
Willy Tarreau24709282013-03-10 21:32:12 +01002389 listen http_https_proxy_explicit
2390 bind ipv6@:80
2391 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
2392 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
2393
Willy Tarreaudad36a32013-03-11 01:20:04 +01002394 listen external_bind_app1
William Lallemandb2f07452015-05-12 14:27:13 +02002395 bind "fd@${FD_APP1}"
Willy Tarreaudad36a32013-03-11 01:20:04 +01002396
Willy Tarreau55dcaf62015-09-27 15:03:15 +02002397 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
2398 sun_path length is used for the address length. Some other programs
2399 such as socat use the string length only by default. Pass the option
2400 ",unix-tightsocklen=0" to any abstract socket definition in socat to
2401 make it compatible with HAProxy's.
2402
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002403 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002404 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002405
2406
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002407bind-process [ all | odd | even | <number 1-64>[-<number 1-64>] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002408 Limit visibility of an instance to a certain set of processes numbers.
2409 May be used in sections : defaults | frontend | listen | backend
2410 yes | yes | yes | yes
2411 Arguments :
2412 all All process will see this instance. This is the default. It
2413 may be used to override a default value.
2414
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002415 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002416 option may be combined with other numbers.
2417
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002418 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002419 option may be combined with other numbers. Do not use it
2420 with less than 2 processes otherwise some instances might be
2421 missing from all processes.
2422
Willy Tarreau110ecc12012-11-15 17:50:01 +01002423 number The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002424 whose values must all be between 1 and 32 or 64 depending on
Willy Tarreau102df612014-05-07 23:56:38 +02002425 the machine's word size. If a proxy is bound to process
2426 numbers greater than the configured global.nbproc, it will
2427 either be forced to process #1 if a single process was
2428 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002429
2430 This keyword limits binding of certain instances to certain processes. This
2431 is useful in order not to have too many processes listening to the same
2432 ports. For instance, on a dual-core machine, it might make sense to set
2433 'nbproc 2' in the global section, then distributes the listeners among 'odd'
2434 and 'even' instances.
2435
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002436 At the moment, it is not possible to reference more than 32 or 64 processes
2437 using this keyword, but this should be more than enough for most setups.
2438 Please note that 'all' really means all processes regardless of the machine's
2439 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002440
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002441 Each "bind" line may further be limited to a subset of the proxy's processes,
2442 please consult the "process" bind keyword in section 5.1.
2443
Willy Tarreaub369a042014-09-16 13:21:03 +02002444 When a frontend has no explicit "bind-process" line, it tries to bind to all
2445 the processes referenced by its "bind" lines. That means that frontends can
2446 easily adapt to their listeners' processes.
2447
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002448 If some backends are referenced by frontends bound to other processes, the
2449 backend automatically inherits the frontend's processes.
2450
2451 Example :
2452 listen app_ip1
2453 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002454 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002455
2456 listen app_ip2
2457 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002458 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002459
2460 listen management
2461 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002462 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002463
Willy Tarreau110ecc12012-11-15 17:50:01 +01002464 listen management
2465 bind 10.0.0.4:80
2466 bind-process 1-4
2467
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002468 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002469
2470
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02002471block { if | unless } <condition> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002472 Block a layer 7 request if/unless a condition is matched
2473 May be used in sections : defaults | frontend | listen | backend
2474 no | yes | yes | yes
2475
2476 The HTTP request will be blocked very early in the layer 7 processing
2477 if/unless <condition> is matched. A 403 error will be returned if the request
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002478 is blocked. The condition has to reference ACLs (see section 7). This is
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02002479 typically used to deny access to certain sensitive resources if some
Willy Tarreau0ba27502007-12-24 16:55:16 +01002480 conditions are met or not met. There is no fixed limit to the number of
2481 "block" statements per instance.
2482
Jarno Huuskonen8c8c3492016-12-28 18:50:29 +02002483 This form is deprecated, do not use it in any new configuration, use the new
2484 "http-request deny" instead.
2485
Willy Tarreau0ba27502007-12-24 16:55:16 +01002486 Example:
2487 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
2488 acl invalid_src src_port 0:1023
2489 acl local_dst hdr(host) -i localhost
2490 block if invalid_src || local_dst
2491
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002492 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002493
2494
2495capture cookie <name> len <length>
2496 Capture and log a cookie in the request and in the response.
2497 May be used in sections : defaults | frontend | listen | backend
2498 no | yes | yes | no
2499 Arguments :
2500 <name> is the beginning of the name of the cookie to capture. In order
2501 to match the exact name, simply suffix the name with an equal
2502 sign ('='). The full name will appear in the logs, which is
2503 useful with application servers which adjust both the cookie name
2504 and value (eg: ASPSESSIONXXXXX).
2505
2506 <length> is the maximum number of characters to report in the logs, which
2507 include the cookie name, the equal sign and the value, all in the
2508 standard "name=value" form. The string will be truncated on the
2509 right if it exceeds <length>.
2510
2511 Only the first cookie is captured. Both the "cookie" request headers and the
2512 "set-cookie" response headers are monitored. This is particularly useful to
2513 check for application bugs causing session crossing or stealing between
2514 users, because generally the user's cookies can only change on a login page.
2515
2516 When the cookie was not presented by the client, the associated log column
2517 will report "-". When a request does not cause a cookie to be assigned by the
2518 server, a "-" is reported in the response column.
2519
2520 The capture is performed in the frontend only because it is necessary that
2521 the log format does not change for a given frontend depending on the
2522 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01002523 "capture cookie" statement in a frontend. The maximum capture length is set
2524 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
2525 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002526
2527 Example:
2528 capture cookie ASPSESSION len 32
2529
2530 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002531 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002532
2533
2534capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01002535 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002536 May be used in sections : defaults | frontend | listen | backend
2537 no | yes | yes | no
2538 Arguments :
2539 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002540 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01002541 appear in the requests, with the first letter of each word in
2542 upper case. The header name will not appear in the logs, only the
2543 value is reported, but the position in the logs is respected.
2544
2545 <length> is the maximum number of characters to extract from the value and
2546 report in the logs. The string will be truncated on the right if
2547 it exceeds <length>.
2548
Willy Tarreau4460d032012-11-21 23:37:37 +01002549 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002550 value will be added to the logs between braces ('{}'). If multiple headers
2551 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002552 in the same order they were declared in the configuration. Non-existent
2553 headers will be logged just as an empty string. Common uses for request
2554 header captures include the "Host" field in virtual hosting environments, the
2555 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002556 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002557 environments to find where the request came from.
2558
2559 Note that when capturing headers such as "User-agent", some spaces may be
2560 logged, making the log analysis more difficult. Thus be careful about what
2561 you log if you know your log parser is not smart enough to rely on the
2562 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002563
Willy Tarreau0900abb2012-11-22 00:21:46 +01002564 There is no limit to the number of captured request headers nor to their
2565 length, though it is wise to keep them low to limit memory usage per session.
2566 In order to keep log format consistent for a same frontend, header captures
2567 can only be declared in a frontend. It is not possible to specify a capture
2568 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002569
2570 Example:
2571 capture request header Host len 15
2572 capture request header X-Forwarded-For len 15
Cyril Bontéd1b0f7c2015-10-26 22:37:39 +01002573 capture request header Referer len 15
Willy Tarreau0ba27502007-12-24 16:55:16 +01002574
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002575 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002576 about logging.
2577
2578
2579capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01002580 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002581 May be used in sections : defaults | frontend | listen | backend
2582 no | yes | yes | no
2583 Arguments :
2584 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002585 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01002586 appear in the response, with the first letter of each word in
2587 upper case. The header name will not appear in the logs, only the
2588 value is reported, but the position in the logs is respected.
2589
2590 <length> is the maximum number of characters to extract from the value and
2591 report in the logs. The string will be truncated on the right if
2592 it exceeds <length>.
2593
Willy Tarreau4460d032012-11-21 23:37:37 +01002594 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01002595 result will be added to the logs between braces ('{}') after the captured
2596 request headers. If multiple headers are captured, they will be delimited by
2597 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01002598 the configuration. Non-existent headers will be logged just as an empty
2599 string. Common uses for response header captures include the "Content-length"
2600 header which indicates how many bytes are expected to be returned, the
2601 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002602
Willy Tarreau0900abb2012-11-22 00:21:46 +01002603 There is no limit to the number of captured response headers nor to their
2604 length, though it is wise to keep them low to limit memory usage per session.
2605 In order to keep log format consistent for a same frontend, header captures
2606 can only be declared in a frontend. It is not possible to specify a capture
2607 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002608
2609 Example:
2610 capture response header Content-length len 9
2611 capture response header Location len 15
2612
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002613 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01002614 about logging.
2615
2616
Cyril Bontéf0c60612010-02-06 14:44:47 +01002617clitimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002618 Set the maximum inactivity time on the client side.
2619 May be used in sections : defaults | frontend | listen | backend
2620 yes | yes | yes | no
2621 Arguments :
2622 <timeout> is the timeout value is specified in milliseconds by default, but
2623 can be in any other unit if the number is suffixed by the unit,
2624 as explained at the top of this document.
2625
2626 The inactivity timeout applies when the client is expected to acknowledge or
2627 send data. In HTTP mode, this timeout is particularly important to consider
2628 during the first phase, when the client sends the request, and during the
2629 response while it is reading data sent by the server. The value is specified
2630 in milliseconds by default, but can be in any other unit if the number is
2631 suffixed by the unit, as specified at the top of this document. In TCP mode
2632 (and to a lesser extent, in HTTP mode), it is highly recommended that the
2633 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002634 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01002635 losses by specifying timeouts that are slightly above multiples of 3 seconds
2636 (eg: 4 or 5 seconds).
2637
2638 This parameter is specific to frontends, but can be specified once for all in
2639 "defaults" sections. This is in fact one of the easiest solutions not to
2640 forget about it. An unspecified timeout results in an infinite timeout, which
2641 is not recommended. Such a usage is accepted and works but reports a warning
2642 during startup because it may results in accumulation of expired sessions in
2643 the system if the system's timeouts are not configured either.
2644
2645 This parameter is provided for compatibility but is currently deprecated.
2646 Please use "timeout client" instead.
2647
Willy Tarreau036fae02008-01-06 13:24:40 +01002648 See also : "timeout client", "timeout http-request", "timeout server", and
2649 "srvtimeout".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002650
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002651compression algo <algorithm> ...
2652compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02002653compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02002654 Enable HTTP compression.
2655 May be used in sections : defaults | frontend | listen | backend
2656 yes | yes | yes | yes
2657 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002658 algo is followed by the list of supported compression algorithms.
2659 type is followed by the list of MIME types that will be compressed.
2660 offload makes haproxy work as a compression offloader only (see notes).
2661
2662 The currently supported algorithms are :
Willy Tarreauc91840a2015-03-28 17:00:39 +01002663 identity this is mostly for debugging, and it was useful for developing
2664 the compression feature. Identity does not apply any change on
2665 data.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002666
Willy Tarreauc91840a2015-03-28 17:00:39 +01002667 gzip applies gzip compression. This setting is only available when
Baptiste Assmannf085d632015-12-21 17:57:32 +01002668 support for zlib or libslz was built in.
Willy Tarreauc91840a2015-03-28 17:00:39 +01002669
2670 deflate same as "gzip", but with deflate algorithm and zlib format.
2671 Note that this algorithm has ambiguous support on many
2672 browsers and no support at all from recent ones. It is
2673 strongly recommended not to use it for anything else than
2674 experimentation. This setting is only available when support
Baptiste Assmannf085d632015-12-21 17:57:32 +01002675 for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002676
Willy Tarreauc91840a2015-03-28 17:00:39 +01002677 raw-deflate same as "deflate" without the zlib wrapper, and used as an
2678 alternative when the browser wants "deflate". All major
2679 browsers understand it and despite violating the standards,
2680 it is known to work better than "deflate", at least on MSIE
2681 and some versions of Safari. Do not use it in conjunction
2682 with "deflate", use either one or the other since both react
2683 to the same Accept-Encoding token. This setting is only
Baptiste Assmannf085d632015-12-21 17:57:32 +01002684 available when support for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002685
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04002686 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01002687 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002688 If backend servers support HTTP compression, these directives
2689 will be no-op: haproxy will see the compressed response and will not
2690 compress again. If backend servers do not support HTTP compression and
2691 there is Accept-Encoding header in request, haproxy will compress the
2692 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02002693
2694 The "offload" setting makes haproxy remove the Accept-Encoding header to
2695 prevent backend servers from compressing responses. It is strongly
2696 recommended not to do this because this means that all the compression work
2697 will be done on the single point where haproxy is located. However in some
2698 deployment scenarios, haproxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04002699 with broken HTTP compression implementation which can't be turned off.
2700 In that case haproxy can be used to prevent that gateway from emitting
2701 invalid payloads. In this case, simply removing the header in the
2702 configuration does not work because it applies before the header is parsed,
2703 so that prevents haproxy from compressing. The "offload" setting should
Willy Tarreauffea9fd2014-07-12 16:37:02 +02002704 then be used for such scenarios. Note: for now, the "offload" setting is
2705 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02002706
William Lallemand05097442012-11-20 12:14:28 +01002707 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002708 * the request does not advertise a supported compression algorithm in the
2709 "Accept-Encoding" header
2710 * the response message is not HTTP/1.1
William Lallemandd3002612012-11-26 14:34:47 +01002711 * HTTP status code is not 200
William Lallemand8bb4e342013-12-10 17:28:48 +01002712 * response header "Transfer-Encoding" contains "chunked" (Temporary
2713 Workaround)
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002714 * response contain neither a "Content-Length" header nor a
2715 "Transfer-Encoding" whose last value is "chunked"
2716 * response contains a "Content-Type" header whose first value starts with
2717 "multipart"
2718 * the response contains the "no-transform" value in the "Cache-control"
2719 header
2720 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
2721 and later
2722 * The response contains a "Content-Encoding" header, indicating that the
2723 response is already compressed (see compression offload)
William Lallemand05097442012-11-20 12:14:28 +01002724
Baptiste Assmann650d53d2013-01-05 15:44:44 +01002725 Note: The compression does not rewrite Etag headers, and does not emit the
2726 Warning header.
William Lallemand05097442012-11-20 12:14:28 +01002727
William Lallemand82fe75c2012-10-23 10:25:10 +02002728 Examples :
2729 compression algo gzip
2730 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01002731
Christopher Fauletc3fe5332016-04-07 15:30:10 +02002732
Cyril Bontéf0c60612010-02-06 14:44:47 +01002733contimeout <timeout> (deprecated)
Willy Tarreau0ba27502007-12-24 16:55:16 +01002734 Set the maximum time to wait for a connection attempt to a server to succeed.
2735 May be used in sections : defaults | frontend | listen | backend
2736 yes | no | yes | yes
2737 Arguments :
2738 <timeout> is the timeout value is specified in milliseconds by default, but
2739 can be in any other unit if the number is suffixed by the unit,
2740 as explained at the top of this document.
2741
2742 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01002743 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +01002744 cover one or several TCP packet losses by specifying timeouts that are
Willy Tarreau0ba27502007-12-24 16:55:16 +01002745 slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the
2746 connect timeout also presets the queue timeout to the same value if this one
2747 has not been specified. Historically, the contimeout was also used to set the
2748 tarpit timeout in a listen section, which is not possible in a pure frontend.
2749
2750 This parameter is specific to backends, but can be specified once for all in
2751 "defaults" sections. This is in fact one of the easiest solutions not to
2752 forget about it. An unspecified timeout results in an infinite timeout, which
2753 is not recommended. Such a usage is accepted and works but reports a warning
2754 during startup because it may results in accumulation of failed sessions in
2755 the system if the system's timeouts are not configured either.
2756
2757 This parameter is provided for backwards compatibility but is currently
2758 deprecated. Please use "timeout connect", "timeout queue" or "timeout tarpit"
2759 instead.
2760
2761 See also : "timeout connect", "timeout queue", "timeout tarpit",
2762 "timeout server", "contimeout".
2763
2764
Willy Tarreau55165fe2009-05-10 12:02:55 +02002765cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02002766 [ postonly ] [ preserve ] [ httponly ] [ secure ]
2767 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002768 Enable cookie-based persistence in a backend.
2769 May be used in sections : defaults | frontend | listen | backend
2770 yes | no | yes | yes
2771 Arguments :
2772 <name> is the name of the cookie which will be monitored, modified or
2773 inserted in order to bring persistence. This cookie is sent to
2774 the client via a "Set-Cookie" header in the response, and is
2775 brought back by the client in a "Cookie" header in all requests.
2776 Special care should be taken to choose a name which does not
2777 conflict with any likely application cookie. Also, if the same
2778 backends are subject to be used by the same clients (eg:
2779 HTTP/HTTPS), care should be taken to use different cookie names
2780 between all backends if persistence between them is not desired.
2781
2782 rewrite This keyword indicates that the cookie will be provided by the
2783 server and that haproxy will have to modify its value to set the
2784 server's identifier in it. This mode is handy when the management
2785 of complex combinations of "Set-cookie" and "Cache-control"
2786 headers is left to the application. The application can then
2787 decide whether or not it is appropriate to emit a persistence
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01002788 cookie. Since all responses should be monitored, this mode
2789 doesn't work in HTTP tunnel mode. Unless the application
2790 behaviour is very complex and/or broken, it is advised not to
2791 start with this mode for new deployments. This keyword is
2792 incompatible with "insert" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002793
2794 insert This keyword indicates that the persistence cookie will have to
Willy Tarreaua79094d2010-08-31 22:54:15 +02002795 be inserted by haproxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002796
Willy Tarreaua79094d2010-08-31 22:54:15 +02002797 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002798 server. When used without the "preserve" option, if the server
2799 emits a cookie with the same name, it will be remove before
2800 processing. For this reason, this mode can be used to upgrade
2801 existing configurations running in the "rewrite" mode. The cookie
2802 will only be a session cookie and will not be stored on the
2803 client's disk. By default, unless the "indirect" option is added,
2804 the server will see the cookies emitted by the client. Due to
2805 caching effects, it is generally wise to add the "nocache" or
2806 "postonly" keywords (see below). The "insert" keyword is not
2807 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002808
2809 prefix This keyword indicates that instead of relying on a dedicated
2810 cookie for the persistence, an existing one will be completed.
2811 This may be needed in some specific environments where the client
2812 does not support more than one single cookie and the application
2813 already needs it. In this case, whenever the server sets a cookie
2814 named <name>, it will be prefixed with the server's identifier
2815 and a delimiter. The prefix will be removed from all client
2816 requests so that the server still finds the cookie it emitted.
2817 Since all requests and responses are subject to being modified,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01002818 this mode doesn't work with tunnel mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02002819 not compatible with "rewrite" and "insert". Note: it is highly
2820 recommended not to use "indirect" with "prefix", otherwise server
2821 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002822
Willy Tarreaua79094d2010-08-31 22:54:15 +02002823 indirect When this option is specified, no cookie will be emitted to a
2824 client which already has a valid one for the server which has
2825 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002826 it will be removed, unless the "preserve" option is also set. In
2827 "insert" mode, this will additionally remove cookies from the
2828 requests transmitted to the server, making the persistence
2829 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02002830 Note: it is highly recommended not to use "indirect" with
2831 "prefix", otherwise server cookie updates would not be sent to
2832 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002833
2834 nocache This option is recommended in conjunction with the insert mode
2835 when there is a cache between the client and HAProxy, as it
2836 ensures that a cacheable response will be tagged non-cacheable if
2837 a cookie needs to be inserted. This is important because if all
2838 persistence cookies are added on a cacheable home page for
2839 instance, then all customers will then fetch the page from an
2840 outer cache and will all share the same persistence cookie,
2841 leading to one server receiving much more traffic than others.
2842 See also the "insert" and "postonly" options.
2843
2844 postonly This option ensures that cookie insertion will only be performed
2845 on responses to POST requests. It is an alternative to the
2846 "nocache" option, because POST responses are not cacheable, so
2847 this ensures that the persistence cookie will never get cached.
2848 Since most sites do not need any sort of persistence before the
2849 first POST which generally is a login request, this is a very
2850 efficient method to optimize caching without risking to find a
2851 persistence cookie in the cache.
2852 See also the "insert" and "nocache" options.
2853
Willy Tarreauba4c5be2010-10-23 12:46:42 +02002854 preserve This option may only be used with "insert" and/or "indirect". It
2855 allows the server to emit the persistence cookie itself. In this
2856 case, if a cookie is found in the response, haproxy will leave it
2857 untouched. This is useful in order to end persistence after a
2858 logout request for instance. For this, the server just has to
2859 emit a cookie with an invalid value (eg: empty) or with a date in
2860 the past. By combining this mechanism with the "disable-on-404"
2861 check option, it is possible to perform a completely graceful
2862 shutdown because users will definitely leave the server after
2863 they logout.
2864
Willy Tarreau4992dd22012-05-31 21:02:17 +02002865 httponly This option tells haproxy to add an "HttpOnly" cookie attribute
2866 when a cookie is inserted. This attribute is used so that a
2867 user agent doesn't share the cookie with non-HTTP components.
2868 Please check RFC6265 for more information on this attribute.
2869
2870 secure This option tells haproxy to add a "Secure" cookie attribute when
2871 a cookie is inserted. This attribute is used so that a user agent
2872 never emits this cookie over non-secure channels, which means
2873 that a cookie learned with this flag will be presented only over
2874 SSL/TLS connections. Please check RFC6265 for more information on
2875 this attribute.
2876
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02002877 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002878 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01002879 name. If the domain begins with a dot, the browser is allowed to
2880 use it for any host ending with that name. It is also possible to
2881 specify several domain names by invoking this option multiple
2882 times. Some browsers might have small limits on the number of
2883 domains, so be careful when doing that. For the record, sending
2884 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02002885
Willy Tarreau996a92c2010-10-13 19:30:47 +02002886 maxidle This option allows inserted cookies to be ignored after some idle
2887 time. It only works with insert-mode cookies. When a cookie is
2888 sent to the client, the date this cookie was emitted is sent too.
2889 Upon further presentations of this cookie, if the date is older
2890 than the delay indicated by the parameter (in seconds), it will
2891 be ignored. Otherwise, it will be refreshed if needed when the
2892 response is sent to the client. This is particularly useful to
2893 prevent users who never close their browsers from remaining for
2894 too long on the same server (eg: after a farm size change). When
2895 this option is set and a cookie has no date, it is always
2896 accepted, but gets refreshed in the response. This maintains the
2897 ability for admins to access their sites. Cookies that have a
2898 date in the future further than 24 hours are ignored. Doing so
2899 lets admins fix timezone issues without risking kicking users off
2900 the site.
2901
2902 maxlife This option allows inserted cookies to be ignored after some life
2903 time, whether they're in use or not. It only works with insert
2904 mode cookies. When a cookie is first sent to the client, the date
2905 this cookie was emitted is sent too. Upon further presentations
2906 of this cookie, if the date is older than the delay indicated by
2907 the parameter (in seconds), it will be ignored. If the cookie in
2908 the request has no date, it is accepted and a date will be set.
2909 Cookies that have a date in the future further than 24 hours are
2910 ignored. Doing so lets admins fix timezone issues without risking
2911 kicking users off the site. Contrary to maxidle, this value is
2912 not refreshed, only the first visit date counts. Both maxidle and
2913 maxlife may be used at the time. This is particularly useful to
2914 prevent users who never close their browsers from remaining for
2915 too long on the same server (eg: after a farm size change). This
2916 is stronger than the maxidle method in that it forces a
2917 redispatch after some absolute delay.
2918
Willy Tarreau0ba27502007-12-24 16:55:16 +01002919 There can be only one persistence cookie per HTTP backend, and it can be
2920 declared in a defaults section. The value of the cookie will be the value
2921 indicated after the "cookie" keyword in a "server" statement. If no cookie
2922 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002923
Willy Tarreau0ba27502007-12-24 16:55:16 +01002924 Examples :
2925 cookie JSESSIONID prefix
2926 cookie SRV insert indirect nocache
2927 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02002928 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01002929
Willy Tarreau294d0f02015-08-10 19:40:12 +02002930 See also : "balance source", "capture cookie", "server" and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002931
Willy Tarreau983e01e2010-01-11 18:42:06 +01002932
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02002933declare capture [ request | response ] len <length>
2934 Declares a capture slot.
2935 May be used in sections : defaults | frontend | listen | backend
2936 no | yes | yes | no
2937 Arguments:
2938 <length> is the length allowed for the capture.
2939
2940 This declaration is only available in the frontend or listen section, but the
2941 reserved slot can be used in the backends. The "request" keyword allocates a
2942 capture slot for use in the request, and "response" allocates a capture slot
2943 for use in the response.
2944
2945 See also: "capture-req", "capture-res" (sample converters),
Baptiste Assmann5ac425c2015-10-21 23:13:46 +02002946 "capture.req.hdr", "capture.res.hdr" (sample fetches),
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02002947 "http-request capture" and "http-response capture".
2948
2949
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002950default-server [param*]
2951 Change default options for a server in a backend
2952 May be used in sections : defaults | frontend | listen | backend
2953 yes | no | yes | yes
2954 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01002955 <param*> is a list of parameters for this server. The "default-server"
2956 keyword accepts an important number of options and has a complete
2957 section dedicated to it. Please refer to section 5 for more
2958 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002959
Willy Tarreau983e01e2010-01-11 18:42:06 +01002960 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01002961 default-server inter 1000 weight 13
2962
2963 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01002964
Willy Tarreau983e01e2010-01-11 18:42:06 +01002965
Willy Tarreau0ba27502007-12-24 16:55:16 +01002966default_backend <backend>
2967 Specify the backend to use when no "use_backend" rule has been matched.
2968 May be used in sections : defaults | frontend | listen | backend
2969 yes | yes | yes | no
2970 Arguments :
2971 <backend> is the name of the backend to use.
2972
2973 When doing content-switching between frontend and backends using the
2974 "use_backend" keyword, it is often useful to indicate which backend will be
2975 used when no rule has matched. It generally is the dynamic backend which
2976 will catch all undetermined requests.
2977
Willy Tarreau0ba27502007-12-24 16:55:16 +01002978 Example :
2979
2980 use_backend dynamic if url_dyn
2981 use_backend static if url_css url_img extension_img
2982 default_backend dynamic
2983
Willy Tarreau98d04852015-05-26 12:18:29 +02002984 See also : "use_backend"
Willy Tarreau2769aa02007-12-27 18:26:09 +01002985
Willy Tarreau0ba27502007-12-24 16:55:16 +01002986
Baptiste Assmann27f51342013-10-09 06:51:49 +02002987description <string>
2988 Describe a listen, frontend or backend.
2989 May be used in sections : defaults | frontend | listen | backend
2990 no | yes | yes | yes
2991 Arguments : string
2992
2993 Allows to add a sentence to describe the related object in the HAProxy HTML
2994 stats page. The description will be printed on the right of the object name
2995 it describes.
2996 No need to backslash spaces in the <string> arguments.
2997
2998
Willy Tarreau0ba27502007-12-24 16:55:16 +01002999disabled
3000 Disable a proxy, frontend or backend.
3001 May be used in sections : defaults | frontend | listen | backend
3002 yes | yes | yes | yes
3003 Arguments : none
3004
3005 The "disabled" keyword is used to disable an instance, mainly in order to
3006 liberate a listening port or to temporarily disable a service. The instance
3007 will still be created and its configuration will be checked, but it will be
3008 created in the "stopped" state and will appear as such in the statistics. It
3009 will not receive any traffic nor will it send any health-checks or logs. It
3010 is possible to disable many instances at once by adding the "disabled"
3011 keyword in a "defaults" section.
3012
3013 See also : "enabled"
3014
3015
Willy Tarreau5ce94572010-06-07 14:35:41 +02003016dispatch <address>:<port>
3017 Set a default server address
3018 May be used in sections : defaults | frontend | listen | backend
3019 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003020 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02003021
3022 <address> is the IPv4 address of the default server. Alternatively, a
3023 resolvable hostname is supported, but this name will be resolved
3024 during start-up.
3025
3026 <ports> is a mandatory port specification. All connections will be sent
3027 to this port, and it is not permitted to use port offsets as is
3028 possible with normal servers.
3029
Willy Tarreau787aed52011-04-15 06:45:37 +02003030 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02003031 server can take the connection. In the past it was used to forward non
3032 persistent connections to an auxiliary load balancer. Due to its simple
3033 syntax, it has also been used for simple TCP relays. It is recommended not to
3034 use it for more clarity, and to use the "server" directive instead.
3035
3036 See also : "server"
3037
3038
Willy Tarreau0ba27502007-12-24 16:55:16 +01003039enabled
3040 Enable a proxy, frontend or backend.
3041 May be used in sections : defaults | frontend | listen | backend
3042 yes | yes | yes | yes
3043 Arguments : none
3044
3045 The "enabled" keyword is used to explicitly enable an instance, when the
3046 defaults has been set to "disabled". This is very rarely used.
3047
3048 See also : "disabled"
3049
3050
3051errorfile <code> <file>
3052 Return a file contents instead of errors generated by HAProxy
3053 May be used in sections : defaults | frontend | listen | backend
3054 yes | yes | yes | yes
3055 Arguments :
3056 <code> is the HTTP status code. Currently, HAProxy is capable of
CJ Ess108b1dd2015-04-07 12:03:37 -04003057 generating codes 200, 400, 403, 405, 408, 429, 500, 502, 503, and
3058 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003059
3060 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003061 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01003062 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01003063 error pages, and to use absolute paths, since files are read
3064 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003065
3066 It is important to understand that this keyword is not meant to rewrite
3067 errors returned by the server, but errors detected and returned by HAProxy.
3068 This is why the list of supported errors is limited to a small set.
3069
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003070 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3071
Willy Tarreau0ba27502007-12-24 16:55:16 +01003072 The files are returned verbatim on the TCP socket. This allows any trick such
3073 as redirections to another URL or site, as well as tricks to clean cookies,
3074 force enable or disable caching, etc... The package provides default error
3075 files returning the same contents as default errors.
3076
Willy Tarreau59140a22009-02-22 12:02:30 +01003077 The files should not exceed the configured buffer size (BUFSIZE), which
3078 generally is 8 or 16 kB, otherwise they will be truncated. It is also wise
3079 not to put any reference to local contents (eg: images) in order to avoid
3080 loops between the client and HAProxy when all servers are down, causing an
3081 error to be returned instead of an image. For better HTTP compliance, it is
3082 recommended that all header lines end with CR-LF and not LF alone.
3083
Willy Tarreau0ba27502007-12-24 16:55:16 +01003084 The files are read at the same time as the configuration and kept in memory.
3085 For this reason, the errors continue to be returned even when the process is
3086 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01003087 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01003088 403 status code and interrogating a blocked URL.
3089
3090 See also : "errorloc", "errorloc302", "errorloc303"
3091
Willy Tarreau59140a22009-02-22 12:02:30 +01003092 Example :
3093 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau989222a2016-01-15 10:26:26 +01003094 errorfile 408 /dev/null # work around Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01003095 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
3096 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
3097
Willy Tarreau2769aa02007-12-27 18:26:09 +01003098
3099errorloc <code> <url>
3100errorloc302 <code> <url>
3101 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3102 May be used in sections : defaults | frontend | listen | backend
3103 yes | yes | yes | yes
3104 Arguments :
3105 <code> is the HTTP status code. Currently, HAProxy is capable of
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003106 generating codes 200, 400, 403, 408, 500, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003107
3108 <url> it is the exact contents of the "Location" header. It may contain
3109 either a relative URI to an error page hosted on the same site,
3110 or an absolute URI designating an error page on another site.
3111 Special care should be given to relative URIs to avoid redirect
3112 loops if the URI itself may generate the same error (eg: 500).
3113
3114 It is important to understand that this keyword is not meant to rewrite
3115 errors returned by the server, but errors detected and returned by HAProxy.
3116 This is why the list of supported errors is limited to a small set.
3117
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003118 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3119
Willy Tarreau2769aa02007-12-27 18:26:09 +01003120 Note that both keyword return the HTTP 302 status code, which tells the
3121 client to fetch the designated URL using the same HTTP method. This can be
3122 quite problematic in case of non-GET methods such as POST, because the URL
3123 sent to the client might not be allowed for something other than GET. To
Willy Tarreau989222a2016-01-15 10:26:26 +01003124 work around this problem, please use "errorloc303" which send the HTTP 303
Willy Tarreau2769aa02007-12-27 18:26:09 +01003125 status code, indicating to the client that the URL must be fetched with a GET
3126 request.
3127
3128 See also : "errorfile", "errorloc303"
3129
3130
3131errorloc303 <code> <url>
3132 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3133 May be used in sections : defaults | frontend | listen | backend
3134 yes | yes | yes | yes
3135 Arguments :
3136 <code> is the HTTP status code. Currently, HAProxy is capable of
3137 generating codes 400, 403, 408, 500, 502, 503, and 504.
3138
3139 <url> it is the exact contents of the "Location" header. It may contain
3140 either a relative URI to an error page hosted on the same site,
3141 or an absolute URI designating an error page on another site.
3142 Special care should be given to relative URIs to avoid redirect
3143 loops if the URI itself may generate the same error (eg: 500).
3144
3145 It is important to understand that this keyword is not meant to rewrite
3146 errors returned by the server, but errors detected and returned by HAProxy.
3147 This is why the list of supported errors is limited to a small set.
3148
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003149 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3150
Willy Tarreau2769aa02007-12-27 18:26:09 +01003151 Note that both keyword return the HTTP 303 status code, which tells the
3152 client to fetch the designated URL using the same HTTP GET method. This
3153 solves the usual problems associated with "errorloc" and the 302 code. It is
3154 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003155 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003156
3157 See also : "errorfile", "errorloc", "errorloc302"
3158
3159
Simon Horman51a1cf62015-02-03 13:00:44 +09003160email-alert from <emailaddr>
3161 Declare the from email address to be used in both the envelope and header
3162 of email alerts. This is the address that email alerts are sent from.
3163 May be used in sections: defaults | frontend | listen | backend
3164 yes | yes | yes | yes
3165
3166 Arguments :
3167
3168 <emailaddr> is the from email address to use when sending email alerts
3169
3170 Also requires "email-alert mailers" and "email-alert to" to be set
3171 and if so sending email alerts is enabled for the proxy.
3172
Simon Horman64e34162015-02-06 11:11:57 +09003173 See also : "email-alert level", "email-alert mailers",
Cyril Bonté307ee1e2015-09-28 23:16:06 +02003174 "email-alert myhostname", "email-alert to", section 3.6 about
3175 mailers.
Simon Horman64e34162015-02-06 11:11:57 +09003176
3177
3178email-alert level <level>
3179 Declare the maximum log level of messages for which email alerts will be
3180 sent. This acts as a filter on the sending of email alerts.
3181 May be used in sections: defaults | frontend | listen | backend
3182 yes | yes | yes | yes
3183
3184 Arguments :
3185
3186 <level> One of the 8 syslog levels:
3187 emerg alert crit err warning notice info debug
3188 The above syslog levels are ordered from lowest to highest.
3189
3190 By default level is alert
3191
3192 Also requires "email-alert from", "email-alert mailers" and
3193 "email-alert to" to be set and if so sending email alerts is enabled
3194 for the proxy.
3195
Simon Horman1421e212015-04-30 13:10:35 +09003196 Alerts are sent when :
3197
3198 * An un-paused server is marked as down and <level> is alert or lower
3199 * A paused server is marked as down and <level> is notice or lower
3200 * A server is marked as up or enters the drain state and <level>
3201 is notice or lower
3202 * "option log-health-checks" is enabled, <level> is info or lower,
3203 and a health check status update occurs
3204
Simon Horman64e34162015-02-06 11:11:57 +09003205 See also : "email-alert from", "email-alert mailers",
3206 "email-alert myhostname", "email-alert to",
Simon Horman51a1cf62015-02-03 13:00:44 +09003207 section 3.6 about mailers.
3208
3209
3210email-alert mailers <mailersect>
3211 Declare the mailers to be used when sending email alerts
3212 May be used in sections: defaults | frontend | listen | backend
3213 yes | yes | yes | yes
3214
3215 Arguments :
3216
3217 <mailersect> is the name of the mailers section to send email alerts.
3218
3219 Also requires "email-alert from" and "email-alert to" to be set
3220 and if so sending email alerts is enabled for the proxy.
3221
Simon Horman64e34162015-02-06 11:11:57 +09003222 See also : "email-alert from", "email-alert level", "email-alert myhostname",
3223 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003224
3225
3226email-alert myhostname <hostname>
3227 Declare the to hostname address to be used when communicating with
3228 mailers.
3229 May be used in sections: defaults | frontend | listen | backend
3230 yes | yes | yes | yes
3231
3232 Arguments :
3233
Baptiste Assmann738bad92015-12-21 15:27:53 +01003234 <hostname> is the hostname to use when communicating with mailers
Simon Horman51a1cf62015-02-03 13:00:44 +09003235
3236 By default the systems hostname is used.
3237
3238 Also requires "email-alert from", "email-alert mailers" and
3239 "email-alert to" to be set and if so sending email alerts is enabled
3240 for the proxy.
3241
Simon Horman64e34162015-02-06 11:11:57 +09003242 See also : "email-alert from", "email-alert level", "email-alert mailers",
3243 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003244
3245
3246email-alert to <emailaddr>
3247 Declare both the recipent address in the envelope and to address in the
3248 header of email alerts. This is the address that email alerts are sent to.
3249 May be used in sections: defaults | frontend | listen | backend
3250 yes | yes | yes | yes
3251
3252 Arguments :
3253
3254 <emailaddr> is the to email address to use when sending email alerts
3255
3256 Also requires "email-alert mailers" and "email-alert to" to be set
3257 and if so sending email alerts is enabled for the proxy.
3258
Simon Horman64e34162015-02-06 11:11:57 +09003259 See also : "email-alert from", "email-alert level", "email-alert mailers",
Simon Horman51a1cf62015-02-03 13:00:44 +09003260 "email-alert myhostname", section 3.6 about mailers.
3261
3262
Willy Tarreau4de91492010-01-22 19:10:05 +01003263force-persist { if | unless } <condition>
3264 Declare a condition to force persistence on down servers
3265 May be used in sections: defaults | frontend | listen | backend
3266 no | yes | yes | yes
3267
3268 By default, requests are not dispatched to down servers. It is possible to
3269 force this using "option persist", but it is unconditional and redispatches
3270 to a valid server if "option redispatch" is set. That leaves with very little
3271 possibilities to force some requests to reach a server which is artificially
3272 marked down for maintenance operations.
3273
3274 The "force-persist" statement allows one to declare various ACL-based
3275 conditions which, when met, will cause a request to ignore the down status of
3276 a server and still try to connect to it. That makes it possible to start a
3277 server, still replying an error to the health checks, and run a specially
3278 configured browser to test the service. Among the handy methods, one could
3279 use a specific source IP address, or a specific cookie. The cookie also has
3280 the advantage that it can easily be added/removed on the browser from a test
3281 page. Once the service is validated, it is then possible to open the service
3282 to the world by returning a valid response to health checks.
3283
3284 The forced persistence is enabled when an "if" condition is met, or unless an
3285 "unless" condition is met. The final redispatch is always disabled when this
3286 is used.
3287
Cyril Bonté0d4bf012010-04-25 23:21:46 +02003288 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02003289 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01003290
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003291
3292filter <name> [param*]
3293 Add the filter <name> in the filter list attached to the proxy.
3294 May be used in sections : defaults | frontend | listen | backend
3295 no | yes | yes | yes
3296 Arguments :
3297 <name> is the name of the filter. Officially supported filters are
3298 referenced in section 9.
3299
Tim Düsterhus4896c442016-11-29 02:15:19 +01003300 <param*> is a list of parameters accepted by the filter <name>. The
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003301 parsing of these parameters are the responsibility of the
Tim Düsterhus4896c442016-11-29 02:15:19 +01003302 filter. Please refer to the documentation of the corresponding
3303 filter (section 9) for all details on the supported parameters.
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003304
3305 Multiple occurrences of the filter line can be used for the same proxy. The
3306 same filter can be referenced many times if needed.
3307
3308 Example:
3309 listen
3310 bind *:80
3311
3312 filter trace name BEFORE-HTTP-COMP
3313 filter compression
3314 filter trace name AFTER-HTTP-COMP
3315
3316 compression algo gzip
3317 compression offload
3318
3319 server srv1 192.168.0.1:80
3320
3321 See also : section 9.
3322
Willy Tarreau4de91492010-01-22 19:10:05 +01003323
Willy Tarreau2769aa02007-12-27 18:26:09 +01003324fullconn <conns>
3325 Specify at what backend load the servers will reach their maxconn
3326 May be used in sections : defaults | frontend | listen | backend
3327 yes | no | yes | yes
3328 Arguments :
3329 <conns> is the number of connections on the backend which will make the
3330 servers use the maximal number of connections.
3331
Willy Tarreau198a7442008-01-17 12:05:32 +01003332 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01003333 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01003334 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01003335 load. The server will then always accept at least <minconn> connections,
3336 never more than <maxconn>, and the limit will be on the ramp between both
3337 values when the backend has less than <conns> concurrent connections. This
3338 makes it possible to limit the load on the servers during normal loads, but
3339 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003340 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003341
Willy Tarreaufbb78422011-06-05 15:38:35 +02003342 Since it's hard to get this value right, haproxy automatically sets it to
3343 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01003344 backend (based on "use_backend" and "default_backend" rules). That way it's
3345 safe to leave it unset. However, "use_backend" involving dynamic names are
3346 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02003347
Willy Tarreau2769aa02007-12-27 18:26:09 +01003348 Example :
3349 # The servers will accept between 100 and 1000 concurrent connections each
3350 # and the maximum of 1000 will be reached when the backend reaches 10000
3351 # connections.
3352 backend dynamic
3353 fullconn 10000
3354 server srv1 dyn1:80 minconn 100 maxconn 1000
3355 server srv2 dyn2:80 minconn 100 maxconn 1000
3356
3357 See also : "maxconn", "server"
3358
3359
3360grace <time>
3361 Maintain a proxy operational for some time after a soft stop
3362 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01003363 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01003364 Arguments :
3365 <time> is the time (by default in milliseconds) for which the instance
3366 will remain operational with the frontend sockets still listening
3367 when a soft-stop is received via the SIGUSR1 signal.
3368
3369 This may be used to ensure that the services disappear in a certain order.
3370 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003371 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01003372 needed by the equipment to detect the failure.
3373
3374 Note that currently, there is very little benefit in using this parameter,
3375 and it may in fact complicate the soft-reconfiguration process more than
3376 simplify it.
3377
Willy Tarreau0ba27502007-12-24 16:55:16 +01003378
Andrew Rodland17be45e2016-10-25 17:04:12 -04003379hash-balance-factor <factor>
3380 Specify the balancing factor for bounded-load consistent hashing
3381 May be used in sections : defaults | frontend | listen | backend
3382 yes | no | no | yes
3383 Arguments :
3384 <factor> is the control for the maximum number of concurrent requests to
3385 send to a server, expressed as a percentage of the average number
3386 of concurrent requests across all of the active servers.
3387
3388 Specifying a "hash-balance-factor" for a server with "hash-type consistent"
3389 enables an algorithm that prevents any one server from getting too many
3390 requests at once, even if some hash buckets receive many more requests than
3391 others. Setting <factor> to 0 (the default) disables the feature. Otherwise,
3392 <factor> is a percentage greater than 100. For example, if <factor> is 150,
3393 then no server will be allowed to have a load more than 1.5 times the average.
3394 If server weights are used, they will be respected.
3395
3396 If the first-choice server is disqualified, the algorithm will choose another
3397 server based on the request hash, until a server with additional capacity is
3398 found. A higher <factor> allows more imbalance between the servers, while a
3399 lower <factor> means that more servers will be checked on average, affecting
3400 performance. Reasonable values are from 125 to 200.
3401
3402 See also : "balance" and "hash-type".
3403
3404
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003405hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003406 Specify a method to use for mapping hashes to servers
3407 May be used in sections : defaults | frontend | listen | backend
3408 yes | no | yes | yes
3409 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04003410 <method> is the method used to select a server from the hash computed by
3411 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003412
Bhaskar98634f02013-10-29 23:30:51 -04003413 map-based the hash table is a static array containing all alive servers.
3414 The hashes will be very smooth, will consider weights, but
3415 will be static in that weight changes while a server is up
3416 will be ignored. This means that there will be no slow start.
3417 Also, since a server is selected by its position in the array,
3418 most mappings are changed when the server count changes. This
3419 means that when a server goes up or down, or when a server is
3420 added to a farm, most connections will be redistributed to
3421 different servers. This can be inconvenient with caches for
3422 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01003423
Bhaskar98634f02013-10-29 23:30:51 -04003424 consistent the hash table is a tree filled with many occurrences of each
3425 server. The hash key is looked up in the tree and the closest
3426 server is chosen. This hash is dynamic, it supports changing
3427 weights while the servers are up, so it is compatible with the
3428 slow start feature. It has the advantage that when a server
3429 goes up or down, only its associations are moved. When a
3430 server is added to the farm, only a few part of the mappings
3431 are redistributed, making it an ideal method for caches.
3432 However, due to its principle, the distribution will never be
3433 very smooth and it may sometimes be necessary to adjust a
3434 server's weight or its ID to get a more balanced distribution.
3435 In order to get the same distribution on multiple load
3436 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003437 same IDs. Note: consistent hash uses sdbm and avalanche if no
3438 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04003439
3440 <function> is the hash function to be used :
3441
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003442 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04003443 reimplementation of ndbm) database library. It was found to do
3444 well in scrambling bits, causing better distribution of the keys
3445 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003446 function with good distribution, unless the total server weight
3447 is a multiple of 64, in which case applying the avalanche
3448 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04003449
3450 djb2 this function was first proposed by Dan Bernstein many years ago
3451 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003452 function provides a better distribution than sdbm. It generally
3453 works well with text-based inputs though it can perform extremely
3454 poorly with numeric-only input or when the total server weight is
3455 a multiple of 33, unless the avalanche modifier is also used.
3456
Willy Tarreaua0f42712013-11-14 14:30:35 +01003457 wt6 this function was designed for haproxy while testing other
3458 functions in the past. It is not as smooth as the other ones, but
3459 is much less sensible to the input data set or to the number of
3460 servers. It can make sense as an alternative to sdbm+avalanche or
3461 djb2+avalanche for consistent hashing or when hashing on numeric
3462 data such as a source IP address or a visitor identifier in a URL
3463 parameter.
3464
Willy Tarreau324f07f2015-01-20 19:44:50 +01003465 crc32 this is the most common CRC32 implementation as used in Ethernet,
3466 gzip, PNG, etc. It is slower than the other ones but may provide
3467 a better distribution or less predictable results especially when
3468 used on strings.
3469
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003470 <modifier> indicates an optional method applied after hashing the key :
3471
3472 avalanche This directive indicates that the result from the hash
3473 function above should not be used in its raw form but that
3474 a 4-byte full avalanche hash must be applied first. The
3475 purpose of this step is to mix the resulting bits from the
3476 previous hash in order to avoid any undesired effect when
3477 the input contains some limited values or when the number of
3478 servers is a multiple of one of the hash's components (64
3479 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
3480 result less predictable, but it's also not as smooth as when
3481 using the original function. Some testing might be needed
3482 with some workloads. This hash is one of the many proposed
3483 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003484
Bhaskar98634f02013-10-29 23:30:51 -04003485 The default hash type is "map-based" and is recommended for most usages. The
3486 default function is "sdbm", the selection of a function should be based on
3487 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003488
Andrew Rodland17be45e2016-10-25 17:04:12 -04003489 See also : "balance", "hash-balance-factor", "server"
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003490
3491
Willy Tarreau0ba27502007-12-24 16:55:16 +01003492http-check disable-on-404
3493 Enable a maintenance mode upon HTTP/404 response to health-checks
3494 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01003495 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01003496 Arguments : none
3497
3498 When this option is set, a server which returns an HTTP code 404 will be
3499 excluded from further load-balancing, but will still receive persistent
3500 connections. This provides a very convenient method for Web administrators
3501 to perform a graceful shutdown of their servers. It is also important to note
3502 that a server which is detected as failed while it was in this mode will not
3503 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
3504 will immediately be reinserted into the farm. The status on the stats page
3505 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01003506 option only works in conjunction with the "httpchk" option. If this option
3507 is used with "http-check expect", then it has precedence over it so that 404
3508 responses will still be considered as soft-stop.
3509
3510 See also : "option httpchk", "http-check expect"
3511
3512
3513http-check expect [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003514 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01003515 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02003516 yes | no | yes | yes
Willy Tarreaubd741542010-03-16 18:46:54 +01003517 Arguments :
3518 <match> is a keyword indicating how to look for a specific pattern in the
3519 response. The keyword may be one of "status", "rstatus",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003520 "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01003521 exclamation mark ("!") to negate the match. Spaces are allowed
3522 between the exclamation mark and the keyword. See below for more
3523 details on the supported keywords.
3524
3525 <pattern> is the pattern to look for. It may be a string or a regular
3526 expression. If the pattern contains spaces, they must be escaped
3527 with the usual backslash ('\').
3528
3529 By default, "option httpchk" considers that response statuses 2xx and 3xx
3530 are valid, and that others are invalid. When "http-check expect" is used,
3531 it defines what is considered valid or invalid. Only one "http-check"
3532 statement is supported in a backend. If a server fails to respond or times
3533 out, the check obviously fails. The available matches are :
3534
3535 status <string> : test the exact string match for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003536 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003537 response's status code is exactly this string. If the
3538 "status" keyword is prefixed with "!", then the response
3539 will be considered invalid if the status code matches.
3540
3541 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003542 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003543 response's status code matches the expression. If the
3544 "rstatus" keyword is prefixed with "!", then the response
3545 will be considered invalid if the status code matches.
3546 This is mostly used to check for multiple codes.
3547
3548 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003549 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003550 response's body contains this exact string. If the
3551 "string" keyword is prefixed with "!", then the response
3552 will be considered invalid if the body contains this
3553 string. This can be used to look for a mandatory word at
3554 the end of a dynamic page, or to detect a failure when a
3555 specific error appears on the check page (eg: a stack
3556 trace).
3557
3558 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003559 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003560 response's body matches this expression. If the "rstring"
3561 keyword is prefixed with "!", then the response will be
3562 considered invalid if the body matches the expression.
3563 This can be used to look for a mandatory word at the end
3564 of a dynamic page, or to detect a failure when a specific
3565 error appears on the check page (eg: a stack trace).
3566
3567 It is important to note that the responses will be limited to a certain size
3568 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
3569 Thus, too large responses may not contain the mandatory pattern when using
3570 "string" or "rstring". If a large response is absolutely required, it is
3571 possible to change the default max size by setting the global variable.
3572 However, it is worth keeping in mind that parsing very large responses can
3573 waste some CPU cycles, especially when regular expressions are used, and that
3574 it is always better to focus the checks on smaller resources.
3575
Cyril Bonté32602d22015-01-30 00:07:07 +01003576 Also "http-check expect" doesn't support HTTP keep-alive. Keep in mind that it
3577 will automatically append a "Connection: close" header, meaning that this
3578 header should not be present in the request provided by "option httpchk".
3579
Willy Tarreaubd741542010-03-16 18:46:54 +01003580 Last, if "http-check expect" is combined with "http-check disable-on-404",
3581 then this last one has precedence when the server responds with 404.
3582
3583 Examples :
3584 # only accept status 200 as valid
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003585 http-check expect status 200
Willy Tarreaubd741542010-03-16 18:46:54 +01003586
3587 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003588 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01003589
3590 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003591 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01003592
3593 # check that we have a correct hexadecimal tag before /html
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01003594 http-check expect rstring <!--tag:[0-9a-f]*</html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01003595
Willy Tarreaubd741542010-03-16 18:46:54 +01003596 See also : "option httpchk", "http-check disable-on-404"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003597
3598
Willy Tarreauef781042010-01-27 11:53:01 +01003599http-check send-state
3600 Enable emission of a state header with HTTP health checks
3601 May be used in sections : defaults | frontend | listen | backend
3602 yes | no | yes | yes
3603 Arguments : none
3604
3605 When this option is set, haproxy will systematically send a special header
3606 "X-Haproxy-Server-State" with a list of parameters indicating to each server
3607 how they are seen by haproxy. This can be used for instance when a server is
3608 manipulated without access to haproxy and the operator needs to know whether
3609 haproxy still sees it up or not, or if the server is the last one in a farm.
3610
3611 The header is composed of fields delimited by semi-colons, the first of which
3612 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
3613 checks on the total number before transition, just as appears in the stats
3614 interface. Next headers are in the form "<variable>=<value>", indicating in
3615 no specific order some values available in the stats interface :
Joseph Lynch514061c2015-01-15 17:52:59 -08003616 - a variable "address", containing the address of the backend server.
3617 This corresponds to the <address> field in the server declaration. For
3618 unix domain sockets, it will read "unix".
3619
3620 - a variable "port", containing the port of the backend server. This
3621 corresponds to the <port> field in the server declaration. For unix
3622 domain sockets, it will read "unix".
3623
Willy Tarreauef781042010-01-27 11:53:01 +01003624 - a variable "name", containing the name of the backend followed by a slash
3625 ("/") then the name of the server. This can be used when a server is
3626 checked in multiple backends.
3627
3628 - a variable "node" containing the name of the haproxy node, as set in the
3629 global "node" variable, otherwise the system's hostname if unspecified.
3630
3631 - a variable "weight" indicating the weight of the server, a slash ("/")
3632 and the total weight of the farm (just counting usable servers). This
3633 helps to know if other servers are available to handle the load when this
3634 one fails.
3635
3636 - a variable "scur" indicating the current number of concurrent connections
3637 on the server, followed by a slash ("/") then the total number of
3638 connections on all servers of the same backend.
3639
3640 - a variable "qcur" indicating the current number of requests in the
3641 server's queue.
3642
3643 Example of a header received by the application server :
3644 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
3645 scur=13/22; qcur=0
3646
3647 See also : "option httpchk", "http-check disable-on-404"
3648
Willy Tarreaube1d34d2016-06-26 19:37:59 +02003649http-request { allow | tarpit | auth [realm <realm>] | redirect <rule> |
3650 deny [deny_status <status>] |
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003651 add-header <name> <fmt> | set-header <name> <fmt> |
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02003652 capture <sample> [ len <length> | id <id> ] |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02003653 del-header <name> | set-nice <nice> | set-log-level <level> |
Sasha Pachev218f0642014-06-16 12:05:59 -06003654 replace-header <name> <match-regex> <replace-fmt> |
3655 replace-value <name> <match-regex> <replace-fmt> |
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01003656 set-method <fmt> | set-path <fmt> | set-query <fmt> |
3657 set-uri <fmt> | set-tos <tos> | set-mark <mark> |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003658 add-acl(<file name>) <key fmt> |
3659 del-acl(<file name>) <key fmt> |
3660 del-map(<file name>) <key fmt> |
Baptiste Assmannbb7e86a2014-09-03 18:29:47 +02003661 set-map(<file name>) <key fmt> <value fmt> |
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003662 set-var(<var name>) <expr> |
Christopher Faulet85d79c92016-11-09 16:54:56 +01003663 unset-var(<var name>) |
Thierry FOURNIER90da1912015-03-05 11:17:06 +01003664 { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] |
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02003665 sc-inc-gpc0(<sc-id>) |
Thierry FOURNIER236657b2015-08-19 08:25:14 +02003666 sc-set-gpt0(<sc-id>) <int> |
Willy Tarreau2d392c22015-08-24 01:43:45 +02003667 silent-drop |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003668 }
Cyril Bontéf0c60612010-02-06 14:44:47 +01003669 [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003670 Access control for Layer 7 requests
3671
3672 May be used in sections: defaults | frontend | listen | backend
3673 no | yes | yes | yes
3674
Willy Tarreau20b0de52012-12-24 15:45:22 +01003675 The http-request statement defines a set of rules which apply to layer 7
3676 processing. The rules are evaluated in their declaration order when they are
3677 met in a frontend, listen or backend section. Any rule may optionally be
3678 followed by an ACL-based condition, in which case it will only be evaluated
3679 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01003680
Willy Tarreau20b0de52012-12-24 15:45:22 +01003681 The first keyword is the rule's action. Currently supported actions include :
3682 - "allow" : this stops the evaluation of the rules and lets the request
3683 pass the check. No further "http-request" rules are evaluated.
3684
3685 - "deny" : this stops the evaluation of the rules and immediately rejects
Willy Tarreaube1d34d2016-06-26 19:37:59 +02003686 the request and emits an HTTP 403 error, or optionally the status code
3687 specified as an argument to "deny_status". The list of permitted status
3688 codes is limited to those that can be overridden by the "errorfile"
3689 directive. No further "http-request" rules are evaluated.
Willy Tarreau20b0de52012-12-24 15:45:22 +01003690
Willy Tarreauccbcc372012-12-27 12:37:57 +01003691 - "tarpit" : this stops the evaluation of the rules and immediately blocks
3692 the request without responding for a delay specified by "timeout tarpit"
3693 or "timeout connect" if the former is not set. After that delay, if the
3694 client is still connected, an HTTP error 500 is returned so that the
3695 client does not suspect it has been tarpitted. Logs will report the flags
3696 "PT". The goal of the tarpit rule is to slow down robots during an attack
3697 when they're limited on the number of concurrent requests. It can be very
3698 efficient against very dumb robots, and will significantly reduce the
3699 load on firewalls compared to a "deny" rule. But when facing "correctly"
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003700 developed robots, it can make things worse by forcing haproxy and the
Willy Tarreau2d392c22015-08-24 01:43:45 +02003701 front firewall to support insane number of concurrent connections. See
3702 also the "silent-drop" action below.
Willy Tarreauccbcc372012-12-27 12:37:57 +01003703
Willy Tarreau20b0de52012-12-24 15:45:22 +01003704 - "auth" : this stops the evaluation of the rules and immediately responds
3705 with an HTTP 401 or 407 error code to invite the user to present a valid
3706 user name and password. No further "http-request" rules are evaluated. An
3707 optional "realm" parameter is supported, it sets the authentication realm
3708 that is returned with the response (typically the application's name).
3709
Willy Tarreau81499eb2012-12-27 12:19:02 +01003710 - "redirect" : this performs an HTTP redirection based on a redirect rule.
3711 This is exactly the same as the "redirect" statement except that it
3712 inserts a redirect rule which can be processed in the middle of other
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01003713 "http-request" rules and that these rules use the "log-format" strings.
3714 See the "redirect" keyword for the rule's syntax.
Willy Tarreau81499eb2012-12-27 12:19:02 +01003715
Willy Tarreau20b0de52012-12-24 15:45:22 +01003716 - "add-header" appends an HTTP header field whose name is specified in
3717 <name> and whose value is defined by <fmt> which follows the log-format
3718 rules (see Custom Log Format in section 8.2.4). This is particularly
3719 useful to pass connection-specific information to the server (eg: the
3720 client's SSL certificate), or to combine several headers into one. This
3721 rule is not final, so it is possible to add other similar rules. Note
3722 that header addition is performed immediately, so one rule might reuse
3723 the resulting header from a previous rule.
3724
3725 - "set-header" does the same as "add-header" except that the header name
3726 is first removed if it existed. This is useful when passing security
3727 information to the server, where the header must not be manipulated by
Willy Tarreau85603282015-01-21 20:39:27 +01003728 external users. Note that the new value is computed before the removal so
3729 it is possible to concatenate a value to an existing header.
Willy Tarreau20b0de52012-12-24 15:45:22 +01003730
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02003731 - "del-header" removes all HTTP header fields whose name is specified in
3732 <name>.
3733
Sasha Pachev218f0642014-06-16 12:05:59 -06003734 - "replace-header" matches the regular expression in all occurrences of
3735 header field <name> according to <match-regex>, and replaces them with
3736 the <replace-fmt> argument. Format characters are allowed in replace-fmt
3737 and work like in <fmt> arguments in "add-header". The match is only
3738 case-sensitive. It is important to understand that this action only
3739 considers whole header lines, regardless of the number of values they
3740 may contain. This usage is suited to headers naturally containing commas
3741 in their value, such as If-Modified-Since and so on.
3742
3743 Example:
3744
3745 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
3746
3747 applied to:
3748
3749 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
3750
3751 outputs:
3752
3753 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
3754
3755 assuming the backend IP is 192.168.1.20
3756
3757 - "replace-value" works like "replace-header" except that it matches the
3758 regex against every comma-delimited value of the header field <name>
3759 instead of the entire header. This is suited for all headers which are
3760 allowed to carry more than one value. An example could be the Accept
3761 header.
3762
3763 Example:
3764
3765 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
3766
3767 applied to:
3768
3769 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
3770
3771 outputs:
3772
3773 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
3774
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01003775 - "set-method" rewrites the request method with the result of the
3776 evaluation of format string <fmt>. There should be very few valid reasons
3777 for having to do so as this is more likely to break something than to fix
3778 it.
3779
3780 - "set-path" rewrites the request path with the result of the evaluation of
3781 format string <fmt>. The query string, if any, is left intact. If a
3782 scheme and authority is found before the path, they are left intact as
3783 well. If the request doesn't have a path ("*"), this one is replaced with
3784 the format. This can be used to prepend a directory component in front of
3785 a path for example. See also "set-query" and "set-uri".
3786
3787 Example :
3788 # prepend the host name before the path
3789 http-request set-path /%[hdr(host)]%[path]
3790
3791 - "set-query" rewrites the request's query string which appears after the
3792 first question mark ("?") with the result of the evaluation of format
3793 string <fmt>. The part prior to the question mark is left intact. If the
3794 request doesn't contain a question mark and the new value is not empty,
3795 then one is added at the end of the URI, followed by the new value. If
3796 a question mark was present, it will never be removed even if the value
3797 is empty. This can be used to add or remove parameters from the query
3798 string. See also "set-query" and "set-uri".
3799
3800 Example :
3801 # replace "%3D" with "=" in the query string
3802 http-request set-query %[query,regsub(%3D,=,g)]
3803
3804 - "set-uri" rewrites the request URI with the result of the evaluation of
3805 format string <fmt>. The scheme, authority, path and query string are all
3806 replaced at once. This can be used to rewrite hosts in front of proxies,
3807 or to perform complex modifications to the URI such as moving parts
3808 between the path and the query string. See also "set-path" and
3809 "set-query".
3810
Willy Tarreauf4c43c12013-06-11 17:01:13 +02003811 - "set-nice" sets the "nice" factor of the current request being processed.
3812 It only has effect against the other requests being processed at the same
3813 time. The default value is 0, unless altered by the "nice" setting on the
3814 "bind" line. The accepted range is -1024..1024. The higher the value, the
3815 nicest the request will be. Lower values will make the request more
3816 important than other ones. This can be useful to improve the speed of
3817 some requests, or lower the priority of non-important requests. Using
3818 this setting without prior experimentation can cause some major slowdown.
3819
Willy Tarreau9a355ec2013-06-11 17:45:46 +02003820 - "set-log-level" is used to change the log level of the current request
3821 when a certain condition is met. Valid levels are the 8 syslog levels
3822 (see the "log" keyword) plus the special level "silent" which disables
3823 logging for this request. This rule is not final so the last matching
3824 rule wins. This rule can be useful to disable health checks coming from
3825 another equipment.
3826
Willy Tarreau42cf39e2013-06-11 18:51:32 +02003827 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
3828 the client to the value passed in <tos> on platforms which support this.
3829 This value represents the whole 8 bits of the IP TOS field, and can be
3830 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
3831 that only the 6 higher bits are used in DSCP or TOS, and the two lower
3832 bits are always 0. This can be used to adjust some routing behaviour on
3833 border routers based on some information from the request. See RFC 2474,
3834 2597, 3260 and 4594 for more information.
3835
Willy Tarreau51347ed2013-06-11 19:34:13 +02003836 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
3837 client to the value passed in <mark> on platforms which support it. This
3838 value is an unsigned 32 bit value which can be matched by netfilter and
3839 by the routing table. It can be expressed both in decimal or hexadecimal
3840 format (prefixed by "0x"). This can be useful to force certain packets to
3841 take a different route (for example a cheaper network path for bulk
3842 downloads). This works on Linux kernels 2.6.32 and above and requires
3843 admin privileges.
3844
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02003845 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
3846 from a file (even a dummy empty file). The file name of the ACL to be
3847 updated is passed between parentheses. It takes one argument: <key fmt>,
3848 which follows log-format rules, to collect content of the new entry. It
3849 performs a lookup in the ACL before insertion, to avoid duplicated (or
3850 more) values. This lookup is done by a linear search and can be expensive
3851 with large lists! It is the equivalent of the "add acl" command from the
3852 stats socket, but can be triggered by an HTTP request.
3853
3854 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
3855 from a file (even a dummy empty file). The file name of the ACL to be
3856 updated is passed between parentheses. It takes one argument: <key fmt>,
3857 which follows log-format rules, to collect content of the entry to delete.
3858 It is the equivalent of the "del acl" command from the stats socket, but
3859 can be triggered by an HTTP request.
3860
3861 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
3862 from a file (even a dummy empty file). The file name of the MAP to be
3863 updated is passed between parentheses. It takes one argument: <key fmt>,
3864 which follows log-format rules, to collect content of the entry to delete.
3865 It takes one argument: "file name" It is the equivalent of the "del map"
3866 command from the stats socket, but can be triggered by an HTTP request.
3867
3868 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
3869 from a file (even a dummy empty file). The file name of the MAP to be
3870 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
3871 which follows log-format rules, used to collect MAP key, and <value fmt>,
3872 which follows log-format rules, used to collect content for the new entry.
3873 It performs a lookup in the MAP before insertion, to avoid duplicated (or
3874 more) values. This lookup is done by a linear search and can be expensive
3875 with large lists! It is the equivalent of the "set map" command from the
3876 stats socket, but can be triggered by an HTTP request.
3877
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02003878 - capture <sample> [ len <length> | id <id> ] :
Willy Tarreaua9083d02015-05-08 15:27:59 +02003879 captures sample expression <sample> from the request buffer, and converts
3880 it to a string of at most <len> characters. The resulting string is
3881 stored into the next request "capture" slot, so it will possibly appear
3882 next to some captured HTTP headers. It will then automatically appear in
3883 the logs, and it will be possible to extract it using sample fetch rules
3884 to feed it into headers or anything. The length should be limited given
3885 that this size will be allocated for each capture during the whole
3886 session life. Please check section 7.3 (Fetching samples) and "capture
3887 request header" for more information.
3888
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02003889 If the keyword "id" is used instead of "len", the action tries to store
3890 the captured string in a previously declared capture slot. This is useful
3891 to run captures in backends. The slot id can be declared by a previous
3892 directive "http-request capture" or with the "declare capture" keyword.
Baptiste Assmanne9544932015-11-03 23:31:35 +01003893 If the slot <id> doesn't exist, then HAProxy fails parsing the
3894 configuration to prevent unexpected behavior at run time.
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02003895
Willy Tarreau09448f72014-06-25 18:12:15 +02003896 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
3897 enables tracking of sticky counters from current request. These rules
3898 do not stop evaluation and do not change default action. Three sets of
3899 counters may be simultaneously tracked by the same connection. The first
3900 "track-sc0" rule executed enables tracking of the counters of the
3901 specified table as the first set. The first "track-sc1" rule executed
3902 enables tracking of the counters of the specified table as the second
3903 set. The first "track-sc2" rule executed enables tracking of the
3904 counters of the specified table as the third set. It is a recommended
3905 practice to use the first set of counters for the per-frontend counters
3906 and the second set for the per-backend ones. But this is just a
3907 guideline, all may be used everywhere.
3908
3909 These actions take one or two arguments :
3910 <key> is mandatory, and is a sample expression rule as described
3911 in section 7.3. It describes what elements of the incoming
3912 request or connection will be analysed, extracted, combined,
3913 and used to select which table entry to update the counters.
3914
3915 <table> is an optional table to be used instead of the default one,
3916 which is the stick-table declared in the current proxy. All
3917 the counters for the matches and updates for the key will
3918 then be performed in that table until the session ends.
3919
3920 Once a "track-sc*" rule is executed, the key is looked up in the table
3921 and if it is not found, an entry is allocated for it. Then a pointer to
3922 that entry is kept during all the session's life, and this entry's
3923 counters are updated as often as possible, every time the session's
3924 counters are updated, and also systematically when the session ends.
3925 Counters are only updated for events that happen after the tracking has
3926 been started. As an exception, connection counters and request counters
3927 are systematically updated so that they reflect useful information.
3928
3929 If the entry tracks concurrent connection counters, one connection is
3930 counted for as long as the entry is tracked, and the entry will not
3931 expire during that time. Tracking counters also provides a performance
3932 advantage over just checking the keys, because only one table lookup is
3933 performed for all ACL checks that make use of it.
3934
Thierry FOURNIER236657b2015-08-19 08:25:14 +02003935 - sc-set-gpt0(<sc-id>) <int> :
3936 This action sets the GPT0 tag according to the sticky counter designated
3937 by <sc-id> and the value of <int>. The expected result is a boolean. If
3938 an error occurs, this action silently fails and the actions evaluation
3939 continues.
3940
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02003941 - sc-inc-gpc0(<sc-id>):
3942 This action increments the GPC0 counter according with the sticky counter
3943 designated by <sc-id>. If an error occurs, this action silently fails and
3944 the actions evaluation continues.
3945
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003946 - set-var(<var-name>) <expr> :
3947 Is used to set the contents of a variable. The variable is declared
3948 inline.
3949
Daniel Schneller0b547052016-03-21 20:46:57 +01003950 <var-name> The name of the variable starts with an indication about
3951 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01003952 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01003953 "sess" : the variable is shared with the whole session
3954 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003955 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01003956 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003957 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01003958 "res" : the variable is shared only during response
3959 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003960 This prefix is followed by a name. The separator is a '.'.
Daniel Schneller0b547052016-03-21 20:46:57 +01003961 The name may only contain characters 'a-z', 'A-Z', '0-9'
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02003962 and '_'.
3963
3964 <expr> Is a standard HAProxy expression formed by a sample-fetch
3965 followed by some converters.
3966
3967 Example:
3968
3969 http-request set-var(req.my_var) req.fhdr(user-agent),lower
3970
Christopher Faulet85d79c92016-11-09 16:54:56 +01003971 - unset-var(<var-name>) :
3972 Is used to unset a variable. See above for details about <var-name>.
3973
3974 Example:
3975
3976 http-request unset-var(req.my_var)
3977
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02003978 - set-src <expr> :
3979 Is used to set the source IP address to the value of specified
3980 expression. Useful when a proxy in front of HAProxy rewrites source IP,
3981 but provides the correct IP in a HTTP header; or you want to mask
3982 source IP for privacy.
3983
3984 <expr> Is a standard HAProxy expression formed by a sample-fetch
3985 followed by some converters.
3986
3987 Example:
3988
3989 http-request set-src hdr(x-forwarded-for)
3990 http-request set-src src,ipmask(24)
3991
Willy Tarreau00005ce2016-10-21 15:07:45 +02003992 When possible, set-src preserves the original source port as long as the
3993 address family allows it, otherwise the source port is set to 0.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02003994
William Lallemand44be6402016-05-25 01:51:35 +02003995 - set-src-port <expr> :
3996 Is used to set the source port address to the value of specified
3997 expression.
3998
3999 <expr> Is a standard HAProxy expression formed by a sample-fetch
4000 followed by some converters.
4001
4002 Example:
4003
4004 http-request set-src-port hdr(x-port)
4005 http-request set-src-port int(4000)
4006
Willy Tarreau00005ce2016-10-21 15:07:45 +02004007 When possible, set-src-port preserves the original source address as long
4008 as the address family supports a port, otherwise it forces the source
4009 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02004010
William Lallemand13e9b0c2016-05-25 02:34:07 +02004011 - set-dst <expr> :
4012 Is used to set the destination IP address to the value of specified
4013 expression. Useful when a proxy in front of HAProxy rewrites destination
4014 IP, but provides the correct IP in a HTTP header; or you want to mask
4015 the IP for privacy. If you want to connect to the new address/port, use
4016 '0.0.0.0:0' as a server address in the backend.
4017
4018 <expr> Is a standard HAProxy expression formed by a sample-fetch
4019 followed by some converters.
4020
4021 Example:
4022
4023 http-request set-dst hdr(x-dst)
4024 http-request set-dst dst,ipmask(24)
4025
Willy Tarreau00005ce2016-10-21 15:07:45 +02004026 When possible, set-dst preserves the original destination port as long as
4027 the address family allows it, otherwise the destination port is set to 0.
4028
William Lallemand13e9b0c2016-05-25 02:34:07 +02004029 - set-dst-port <expr> :
4030 Is used to set the destination port address to the value of specified
4031 expression. If you want to connect to the new address/port, use
4032 '0.0.0.0:0' as a server address in the backend.
4033
4034 <expr> Is a standard HAProxy expression formed by a sample-fetch
4035 followed by some converters.
4036
4037 Example:
4038
4039 http-request set-dst-port hdr(x-port)
4040 http-request set-dst-port int(4000)
4041
Willy Tarreau00005ce2016-10-21 15:07:45 +02004042 When possible, set-dst-port preserves the original destination address as
4043 long as the address family supports a port, otherwise it forces the
4044 destination address to IPv4 "0.0.0.0" before rewriting the port.
4045
Willy Tarreau2d392c22015-08-24 01:43:45 +02004046 - "silent-drop" : this stops the evaluation of the rules and makes the
4047 client-facing connection suddenly disappear using a system-dependant way
4048 that tries to prevent the client from being notified. The effect it then
4049 that the client still sees an established connection while there's none
4050 on HAProxy. The purpose is to achieve a comparable effect to "tarpit"
4051 except that it doesn't use any local resource at all on the machine
4052 running HAProxy. It can resist much higher loads than "tarpit", and slow
4053 down stronger attackers. It is important to undestand the impact of using
4054 this mechanism. All stateful equipments placed between the client and
4055 HAProxy (firewalls, proxies, load balancers) will also keep the
4056 established connection for a long time and may suffer from this action.
4057 On modern Linux systems running with enough privileges, the TCP_REPAIR
4058 socket option is used to block the emission of a TCP reset. On other
4059 systems, the socket's TTL is reduced to 1 so that the TCP reset doesn't
4060 pass the first router, though it's still delivered to local networks. Do
4061 not use it unless you fully understand how it works.
4062
Willy Tarreau20b0de52012-12-24 15:45:22 +01004063 There is no limit to the number of http-request statements per instance.
4064
4065 It is important to know that http-request rules are processed very early in
4066 the HTTP processing, just after "block" rules and before "reqdel" or "reqrep"
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004067 or "reqadd" rules. That way, headers added by "add-header"/"set-header" are
4068 visible by almost all further ACL rules.
4069
4070 Using "reqadd"/"reqdel"/"reqrep" to manipulate request headers is discouraged
4071 in newer versions (>= 1.5). But if you need to use regular expression to
4072 delete headers, you can still use "reqdel". Also please use
4073 "http-request deny/allow/tarpit" instead of "reqdeny"/"reqpass"/"reqtarpit".
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004074
4075 Example:
Cyril Bonté78caf842010-03-10 22:41:43 +01004076 acl nagios src 192.168.129.3
4077 acl local_net src 192.168.0.0/16
4078 acl auth_ok http_auth(L1)
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004079
Cyril Bonté78caf842010-03-10 22:41:43 +01004080 http-request allow if nagios
4081 http-request allow if local_net auth_ok
4082 http-request auth realm Gimme if local_net auth_ok
4083 http-request deny
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004084
Cyril Bonté78caf842010-03-10 22:41:43 +01004085 Example:
4086 acl auth_ok http_auth_group(L1) G1
Cyril Bonté78caf842010-03-10 22:41:43 +01004087 http-request auth unless auth_ok
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004088
Willy Tarreau20b0de52012-12-24 15:45:22 +01004089 Example:
4090 http-request set-header X-Haproxy-Current-Date %T
4091 http-request set-header X-SSL %[ssl_fc]
Willy Tarreaufca42612015-08-27 17:15:05 +02004092 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id,hex]
Willy Tarreau20b0de52012-12-24 15:45:22 +01004093 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
4094 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
4095 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
4096 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
4097 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
4098 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
4099
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004100 Example:
4101 acl key req.hdr(X-Add-Acl-Key) -m found
4102 acl add path /addacl
4103 acl del path /delacl
4104
4105 acl myhost hdr(Host) -f myhost.lst
4106
4107 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
4108 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
4109
4110 Example:
4111 acl value req.hdr(X-Value) -m found
4112 acl setmap path /setmap
4113 acl delmap path /delmap
4114
4115 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
4116
4117 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
4118 http-request del-map(map.lst) %[src] if delmap
4119
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02004120 See also : "stats http-request", section 3.4 about userlists and section 7
4121 about ACL usage.
Willy Tarreauef781042010-01-27 11:53:01 +01004122
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004123http-response { allow | deny | add-header <name> <fmt> | set-nice <nice> |
Willy Tarreau51d861a2015-05-22 17:30:48 +02004124 capture <sample> id <id> | redirect <rule> |
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004125 set-header <name> <fmt> | del-header <name> |
Sasha Pachev218f0642014-06-16 12:05:59 -06004126 replace-header <name> <regex-match> <replace-fmt> |
4127 replace-value <name> <regex-match> <replace-fmt> |
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004128 set-status <status> [reason <str>] |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004129 set-log-level <level> | set-mark <mark> | set-tos <tos> |
4130 add-acl(<file name>) <key fmt> |
4131 del-acl(<file name>) <key fmt> |
4132 del-map(<file name>) <key fmt> |
Thierry FOURNIER90da1912015-03-05 11:17:06 +01004133 set-map(<file name>) <key fmt> <value fmt> |
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004134 set-var(<var-name>) <expr> |
Christopher Faulet85d79c92016-11-09 16:54:56 +01004135 unset-var(<var-name>) |
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004136 { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] |
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004137 sc-inc-gpc0(<sc-id>) |
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004138 sc-set-gpt0(<sc-id>) <int> |
Willy Tarreau2d392c22015-08-24 01:43:45 +02004139 silent-drop |
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004140 }
Lukas Tribus2dd1d1a2013-06-19 23:34:41 +02004141 [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004142 Access control for Layer 7 responses
4143
4144 May be used in sections: defaults | frontend | listen | backend
4145 no | yes | yes | yes
4146
4147 The http-response statement defines a set of rules which apply to layer 7
4148 processing. The rules are evaluated in their declaration order when they are
4149 met in a frontend, listen or backend section. Any rule may optionally be
4150 followed by an ACL-based condition, in which case it will only be evaluated
4151 if the condition is true. Since these rules apply on responses, the backend
4152 rules are applied first, followed by the frontend's rules.
4153
4154 The first keyword is the rule's action. Currently supported actions include :
4155 - "allow" : this stops the evaluation of the rules and lets the response
4156 pass the check. No further "http-response" rules are evaluated for the
4157 current section.
4158
4159 - "deny" : this stops the evaluation of the rules and immediately rejects
4160 the response and emits an HTTP 502 error. No further "http-response"
4161 rules are evaluated.
4162
4163 - "add-header" appends an HTTP header field whose name is specified in
4164 <name> and whose value is defined by <fmt> which follows the log-format
4165 rules (see Custom Log Format in section 8.2.4). This may be used to send
4166 a cookie to a client for example, or to pass some internal information.
4167 This rule is not final, so it is possible to add other similar rules.
4168 Note that header addition is performed immediately, so one rule might
4169 reuse the resulting header from a previous rule.
4170
4171 - "set-header" does the same as "add-header" except that the header name
4172 is first removed if it existed. This is useful when passing security
4173 information to the server, where the header must not be manipulated by
4174 external users.
4175
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004176 - "del-header" removes all HTTP header fields whose name is specified in
4177 <name>.
4178
Sasha Pachev218f0642014-06-16 12:05:59 -06004179 - "replace-header" matches the regular expression in all occurrences of
4180 header field <name> according to <match-regex>, and replaces them with
4181 the <replace-fmt> argument. Format characters are allowed in replace-fmt
4182 and work like in <fmt> arguments in "add-header". The match is only
4183 case-sensitive. It is important to understand that this action only
4184 considers whole header lines, regardless of the number of values they
4185 may contain. This usage is suited to headers naturally containing commas
4186 in their value, such as Set-Cookie, Expires and so on.
4187
4188 Example:
4189
4190 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
4191
4192 applied to:
4193
4194 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
4195
4196 outputs:
4197
4198 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
4199
4200 assuming the backend IP is 192.168.1.20.
4201
4202 - "replace-value" works like "replace-header" except that it matches the
4203 regex against every comma-delimited value of the header field <name>
4204 instead of the entire header. This is suited for all headers which are
4205 allowed to carry more than one value. An example could be the Accept
4206 header.
4207
4208 Example:
4209
4210 http-response replace-value Cache-control ^public$ private
4211
4212 applied to:
4213
4214 Cache-Control: max-age=3600, public
4215
4216 outputs:
4217
4218 Cache-Control: max-age=3600, private
4219
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004220 - "set-status" replaces the response status code with <status> which must
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004221 be an integer between 100 and 999. Optionally, a custom reason text can be
4222 provided defined by <str>, or the default reason for the specified code
4223 will be used as a fallback.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004224
4225 Example:
4226
4227 # return "431 Request Header Fields Too Large"
4228 http-response set-status 431
Robin H. Johnson52f5db22017-01-01 13:10:52 -08004229 # return "503 Slow Down", custom reason
4230 http-response set-status 503 reason "Slow Down".
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004231
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004232 - "set-nice" sets the "nice" factor of the current request being processed.
4233 It only has effect against the other requests being processed at the same
4234 time. The default value is 0, unless altered by the "nice" setting on the
4235 "bind" line. The accepted range is -1024..1024. The higher the value, the
4236 nicest the request will be. Lower values will make the request more
4237 important than other ones. This can be useful to improve the speed of
4238 some requests, or lower the priority of non-important requests. Using
4239 this setting without prior experimentation can cause some major slowdown.
4240
Willy Tarreau9a355ec2013-06-11 17:45:46 +02004241 - "set-log-level" is used to change the log level of the current request
4242 when a certain condition is met. Valid levels are the 8 syslog levels
4243 (see the "log" keyword) plus the special level "silent" which disables
4244 logging for this request. This rule is not final so the last matching
4245 rule wins. This rule can be useful to disable health checks coming from
4246 another equipment.
4247
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004248 - "set-tos" is used to set the TOS or DSCP field value of packets sent to
4249 the client to the value passed in <tos> on platforms which support this.
4250 This value represents the whole 8 bits of the IP TOS field, and can be
4251 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note
4252 that only the 6 higher bits are used in DSCP or TOS, and the two lower
4253 bits are always 0. This can be used to adjust some routing behaviour on
4254 border routers based on some information from the request. See RFC 2474,
4255 2597, 3260 and 4594 for more information.
4256
Willy Tarreau51347ed2013-06-11 19:34:13 +02004257 - "set-mark" is used to set the Netfilter MARK on all packets sent to the
4258 client to the value passed in <mark> on platforms which support it. This
4259 value is an unsigned 32 bit value which can be matched by netfilter and
4260 by the routing table. It can be expressed both in decimal or hexadecimal
4261 format (prefixed by "0x"). This can be useful to force certain packets to
4262 take a different route (for example a cheaper network path for bulk
4263 downloads). This works on Linux kernels 2.6.32 and above and requires
4264 admin privileges.
4265
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004266 - "add-acl" is used to add a new entry into an ACL. The ACL must be loaded
4267 from a file (even a dummy empty file). The file name of the ACL to be
4268 updated is passed between parentheses. It takes one argument: <key fmt>,
4269 which follows log-format rules, to collect content of the new entry. It
4270 performs a lookup in the ACL before insertion, to avoid duplicated (or
4271 more) values. This lookup is done by a linear search and can be expensive
4272 with large lists! It is the equivalent of the "add acl" command from the
4273 stats socket, but can be triggered by an HTTP response.
4274
4275 - "del-acl" is used to delete an entry from an ACL. The ACL must be loaded
4276 from a file (even a dummy empty file). The file name of the ACL to be
4277 updated is passed between parentheses. It takes one argument: <key fmt>,
4278 which follows log-format rules, to collect content of the entry to delete.
4279 It is the equivalent of the "del acl" command from the stats socket, but
4280 can be triggered by an HTTP response.
4281
4282 - "del-map" is used to delete an entry from a MAP. The MAP must be loaded
4283 from a file (even a dummy empty file). The file name of the MAP to be
4284 updated is passed between parentheses. It takes one argument: <key fmt>,
4285 which follows log-format rules, to collect content of the entry to delete.
4286 It takes one argument: "file name" It is the equivalent of the "del map"
4287 command from the stats socket, but can be triggered by an HTTP response.
4288
4289 - "set-map" is used to add a new entry into a MAP. The MAP must be loaded
4290 from a file (even a dummy empty file). The file name of the MAP to be
4291 updated is passed between parentheses. It takes 2 arguments: <key fmt>,
4292 which follows log-format rules, used to collect MAP key, and <value fmt>,
4293 which follows log-format rules, used to collect content for the new entry.
4294 It performs a lookup in the MAP before insertion, to avoid duplicated (or
4295 more) values. This lookup is done by a linear search and can be expensive
4296 with large lists! It is the equivalent of the "set map" command from the
4297 stats socket, but can be triggered by an HTTP response.
4298
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004299 - capture <sample> id <id> :
4300 captures sample expression <sample> from the response buffer, and converts
4301 it to a string. The resulting string is stored into the next request
4302 "capture" slot, so it will possibly appear next to some captured HTTP
4303 headers. It will then automatically appear in the logs, and it will be
4304 possible to extract it using sample fetch rules to feed it into headers or
4305 anything. Please check section 7.3 (Fetching samples) and "capture
4306 response header" for more information.
4307
4308 The keyword "id" is the id of the capture slot which is used for storing
4309 the string. The capture slot must be defined in an associated frontend.
4310 This is useful to run captures in backends. The slot id can be declared by
4311 a previous directive "http-response capture" or with the "declare capture"
4312 keyword.
Baptiste Assmanne9544932015-11-03 23:31:35 +01004313 If the slot <id> doesn't exist, then HAProxy fails parsing the
4314 configuration to prevent unexpected behavior at run time.
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004315
Willy Tarreau51d861a2015-05-22 17:30:48 +02004316 - "redirect" : this performs an HTTP redirection based on a redirect rule.
4317 This supports a format string similarly to "http-request redirect" rules,
4318 with the exception that only the "location" type of redirect is possible
4319 on the response. See the "redirect" keyword for the rule's syntax. When
4320 a redirect rule is applied during a response, connections to the server
4321 are closed so that no data can be forwarded from the server to the client.
4322
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004323 - set-var(<var-name>) expr:
4324 Is used to set the contents of a variable. The variable is declared
4325 inline.
4326
Daniel Schneller0b547052016-03-21 20:46:57 +01004327 <var-name> The name of the variable starts with an indication about
4328 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01004329 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01004330 "sess" : the variable is shared with the whole session
4331 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004332 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01004333 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004334 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01004335 "res" : the variable is shared only during response
4336 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004337 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01004338 The name may only contain characters 'a-z', 'A-Z', '0-9',
4339 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004340
4341 <expr> Is a standard HAProxy expression formed by a sample-fetch
4342 followed by some converters.
4343
4344 Example:
4345
4346 http-response set-var(sess.last_redir) res.hdr(location)
4347
Christopher Faulet85d79c92016-11-09 16:54:56 +01004348 - unset-var(<var-name>) :
4349 Is used to unset a variable. See above for details about <var-name>.
4350
4351 Example:
4352
4353 http-response unset-var(sess.last_redir)
4354
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004355 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
4356 enables tracking of sticky counters from current response. Please refer to
4357 "http-request track-sc" for a complete description. The only difference
4358 from "http-request track-sc" is the <key> sample expression can only make
4359 use of samples in response (eg. res.*, status etc.) and samples below
4360 Layer 6 (eg. ssl related samples, see section 7.3.4). If the sample is
4361 not supported, haproxy will fail and warn while parsing the config.
4362
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004363 - sc-set-gpt0(<sc-id>) <int> :
4364 This action sets the GPT0 tag according to the sticky counter designated
4365 by <sc-id> and the value of <int>. The expected result is a boolean. If
4366 an error occurs, this action silently fails and the actions evaluation
4367 continues.
4368
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004369 - sc-inc-gpc0(<sc-id>):
4370 This action increments the GPC0 counter according with the sticky counter
4371 designated by <sc-id>. If an error occurs, this action silently fails and
4372 the actions evaluation continues.
4373
Willy Tarreau2d392c22015-08-24 01:43:45 +02004374 - "silent-drop" : this stops the evaluation of the rules and makes the
4375 client-facing connection suddenly disappear using a system-dependant way
4376 that tries to prevent the client from being notified. The effect it then
4377 that the client still sees an established connection while there's none
4378 on HAProxy. The purpose is to achieve a comparable effect to "tarpit"
4379 except that it doesn't use any local resource at all on the machine
4380 running HAProxy. It can resist much higher loads than "tarpit", and slow
4381 down stronger attackers. It is important to undestand the impact of using
4382 this mechanism. All stateful equipments placed between the client and
4383 HAProxy (firewalls, proxies, load balancers) will also keep the
4384 established connection for a long time and may suffer from this action.
4385 On modern Linux systems running with enough privileges, the TCP_REPAIR
4386 socket option is used to block the emission of a TCP reset. On other
4387 systems, the socket's TTL is reduced to 1 so that the TCP reset doesn't
4388 pass the first router, though it's still delivered to local networks. Do
4389 not use it unless you fully understand how it works.
4390
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004391 There is no limit to the number of http-response statements per instance.
4392
Godbach09250262013-07-02 01:19:15 +08004393 It is important to know that http-response rules are processed very early in
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004394 the HTTP processing, before "rspdel" or "rsprep" or "rspadd" rules. That way,
4395 headers added by "add-header"/"set-header" are visible by almost all further ACL
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004396 rules.
4397
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004398 Using "rspadd"/"rspdel"/"rsprep" to manipulate request headers is discouraged
4399 in newer versions (>= 1.5). But if you need to use regular expression to
4400 delete headers, you can still use "rspdel". Also please use
4401 "http-response deny" instead of "rspdeny".
4402
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004403 Example:
4404 acl key_acl res.hdr(X-Acl-Key) -m found
4405
4406 acl myhost hdr(Host) -f myhost.lst
4407
4408 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
4409 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
4410
4411 Example:
4412 acl value res.hdr(X-Value) -m found
4413
4414 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
4415
4416 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
4417 http-response del-map(map.lst) %[src] if ! value
4418
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004419 See also : "http-request", section 3.4 about userlists and section 7 about
4420 ACL usage.
4421
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02004422
Willy Tarreau30631952015-08-06 15:05:24 +02004423http-reuse { never | safe | aggressive | always }
4424 Declare how idle HTTP connections may be shared between requests
4425
4426 May be used in sections: defaults | frontend | listen | backend
4427 yes | no | yes | yes
4428
4429 By default, a connection established between haproxy and the backend server
4430 belongs to the session that initiated it. The downside is that between the
4431 response and the next request, the connection remains idle and is not used.
4432 In many cases for performance reasons it is desirable to make it possible to
4433 reuse these idle connections to serve other requests from different sessions.
4434 This directive allows to tune this behaviour.
4435
4436 The argument indicates the desired connection reuse strategy :
4437
4438 - "never" : idle connections are never shared between sessions. This is
4439 the default choice. It may be enforced to cancel a different
4440 strategy inherited from a defaults section or for
4441 troubleshooting. For example, if an old bogus application
4442 considers that multiple requests over the same connection come
4443 from the same client and it is not possible to fix the
4444 application, it may be desirable to disable connection sharing
4445 in a single backend. An example of such an application could
4446 be an old haproxy using cookie insertion in tunnel mode and
4447 not checking any request past the first one.
4448
4449 - "safe" : this is the recommended strategy. The first request of a
4450 session is always sent over its own connection, and only
4451 subsequent requests may be dispatched over other existing
4452 connections. This ensures that in case the server closes the
4453 connection when the request is being sent, the browser can
4454 decide to silently retry it. Since it is exactly equivalent to
4455 regular keep-alive, there should be no side effects.
4456
4457 - "aggressive" : this mode may be useful in webservices environments where
4458 all servers are not necessarily known and where it would be
4459 appreciable to deliver most first requests over existing
4460 connections. In this case, first requests are only delivered
4461 over existing connections that have been reused at least once,
4462 proving that the server correctly supports connection reuse.
4463 It should only be used when it's sure that the client can
4464 retry a failed request once in a while and where the benefit
4465 of aggressive connection reuse significantly outweights the
4466 downsides of rare connection failures.
4467
4468 - "always" : this mode is only recommended when the path to the server is
4469 known for never breaking existing connections quickly after
4470 releasing them. It allows the first request of a session to be
4471 sent to an existing connection. This can provide a significant
4472 performance increase over the "safe" strategy when the backend
4473 is a cache farm, since such components tend to show a
4474 consistent behaviour and will benefit from the connection
4475 sharing. It is recommended that the "http-keep-alive" timeout
4476 remains low in this mode so that no dead connections remain
4477 usable. In most cases, this will lead to the same performance
4478 gains as "aggressive" but with more risks. It should only be
4479 used when it improves the situation over "aggressive".
4480
4481 When http connection sharing is enabled, a great care is taken to respect the
4482 connection properties and compatiblities. Specifically :
4483 - connections made with "usesrc" followed by a client-dependant value
4484 ("client", "clientip", "hdr_ip") are marked private and never shared ;
4485
4486 - connections sent to a server with a TLS SNI extension are marked private
4487 and are never shared ;
4488
4489 - connections receiving a status code 401 or 407 expect some authentication
4490 to be sent in return. Due to certain bogus authentication schemes (such
4491 as NTLM) relying on the connection, these connections are marked private
4492 and are never shared ;
4493
4494 No connection pool is involved, once a session dies, the last idle connection
4495 it was attached to is deleted at the same time. This ensures that connections
4496 may not last after all sessions are closed.
4497
4498 Note: connection reuse improves the accuracy of the "server maxconn" setting,
4499 because almost no new connection will be established while idle connections
4500 remain available. This is particularly true with the "always" strategy.
4501
4502 See also : "option http-keep-alive", "server maxconn"
4503
4504
Mark Lamourinec2247f02012-01-04 13:02:01 -05004505http-send-name-header [<header>]
4506 Add the server name to a request. Use the header string given by <header>
4507
4508 May be used in sections: defaults | frontend | listen | backend
4509 yes | no | yes | yes
4510
4511 Arguments :
4512
4513 <header> The header string to use to send the server name
4514
4515 The "http-send-name-header" statement causes the name of the target
4516 server to be added to the headers of an HTTP request. The name
4517 is added with the header string proved.
4518
4519 See also : "server"
4520
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01004521id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02004522 Set a persistent ID to a proxy.
4523 May be used in sections : defaults | frontend | listen | backend
4524 no | yes | yes | yes
4525 Arguments : none
4526
4527 Set a persistent ID for the proxy. This ID must be unique and positive.
4528 An unused ID will automatically be assigned if unset. The first assigned
4529 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01004530
4531
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004532ignore-persist { if | unless } <condition>
4533 Declare a condition to ignore persistence
4534 May be used in sections: defaults | frontend | listen | backend
4535 no | yes | yes | yes
4536
4537 By default, when cookie persistence is enabled, every requests containing
4538 the cookie are unconditionally persistent (assuming the target server is up
4539 and running).
4540
4541 The "ignore-persist" statement allows one to declare various ACL-based
4542 conditions which, when met, will cause a request to ignore persistence.
4543 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03004544 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004545 persistence for a specific User-Agent (for example, some web crawler bots).
4546
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004547 The persistence is ignored when an "if" condition is met, or unless an
4548 "unless" condition is met.
4549
4550 See also : "force-persist", "cookie", and section 7 about ACL usage.
4551
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004552load-server-state-from-file { global | local | none }
4553 Allow seamless reload of HAProxy
4554 May be used in sections: defaults | frontend | listen | backend
4555 yes | no | yes | yes
4556
4557 This directive points HAProxy to a file where server state from previous
4558 running process has been saved. That way, when starting up, before handling
4559 traffic, the new process can apply old states to servers exactly has if no
4560 reload occured. The purpose of the "load-server-state-from-file" directive is
4561 to tell haproxy which file to use. For now, only 2 arguments to either prevent
4562 loading state or load states from a file containing all backends and servers.
4563 The state file can be generated by running the command "show servers state"
4564 over the stats socket and redirect output.
4565
4566 The format of the file is versionned and is very specific. To understand it,
4567 please read the documentation of the "show servers state" command (chapter
Kevin Decherf949c7202015-10-13 23:26:44 +02004568 9.2 of Management Guide).
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004569
4570 Arguments:
4571 global load the content of the file pointed by the global directive
4572 named "server-state-file".
4573
4574 local load the content of the file pointed by the directive
4575 "server-state-file-name" if set. If not set, then the backend
4576 name is used as a file name.
4577
4578 none don't load any stat for this backend
4579
4580 Notes:
Willy Tarreaue5a60682016-11-09 14:54:53 +01004581 - server's IP address is preserved across reloads by default, but the
4582 order can be changed thanks to the server's "init-addr" setting. This
4583 means that an IP address change performed on the CLI at run time will
4584 be preserved, and that any change to the local resolver (eg: /etc/hosts)
4585 will possibly not have any effect if the state file is in use.
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004586
4587 - server's weight is applied from previous running process unless it has
4588 has changed between previous and new configuration files.
4589
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004590 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004591
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004592 global
4593 stats socket /tmp/socket
4594 server-state-file /tmp/server_state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004595
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004596 defaults
4597 load-server-state-from-file global
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004598
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004599 backend bk
4600 server s1 127.0.0.1:22 check weight 11
4601 server s2 127.0.0.1:22 check weight 12
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004602
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004603
4604 Then one can run :
4605
4606 socat /tmp/socket - <<< "show servers state" > /tmp/server_state
4607
4608 Content of the file /tmp/server_state would be like this:
4609
4610 1
4611 # <field names skipped for the doc example>
4612 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
4613 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
4614
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004615 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004616
4617 global
4618 stats socket /tmp/socket
4619 server-state-base /etc/haproxy/states
4620
4621 defaults
4622 load-server-state-from-file local
4623
4624 backend bk
4625 server s1 127.0.0.1:22 check weight 11
4626 server s2 127.0.0.1:22 check weight 12
4627
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02004628
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02004629 Then one can run :
4630
4631 socat /tmp/socket - <<< "show servers state bk" > /etc/haproxy/states/bk
4632
4633 Content of the file /etc/haproxy/states/bk would be like this:
4634
4635 1
4636 # <field names skipped for the doc example>
4637 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
4638 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
4639
4640 See also: "server-state-file", "server-state-file-name", and
4641 "show servers state"
4642
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004643
Willy Tarreau2769aa02007-12-27 18:26:09 +01004644log global
Willy Tarreau18324f52014-06-27 18:10:07 +02004645log <address> [len <length>] <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02004646no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01004647 Enable per-instance logging of events and traffic.
4648 May be used in sections : defaults | frontend | listen | backend
4649 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02004650
4651 Prefix :
4652 no should be used when the logger list must be flushed. For example,
4653 if you don't want to inherit from the default logger list. This
4654 prefix does not allow arguments.
4655
Willy Tarreau2769aa02007-12-27 18:26:09 +01004656 Arguments :
4657 global should be used when the instance's logging parameters are the
4658 same as the global ones. This is the most common usage. "global"
4659 replaces <address>, <facility> and <level> with those of the log
4660 entries found in the "global" section. Only one "log global"
4661 statement may be used per instance, and this form takes no other
4662 parameter.
4663
4664 <address> indicates where to send the logs. It takes the same format as
4665 for the "global" section's logs, and can be one of :
4666
4667 - An IPv4 address optionally followed by a colon (':') and a UDP
4668 port. If no port is specified, 514 is used by default (the
4669 standard syslog port).
4670
David du Colombier24bb5f52011-03-17 10:40:23 +01004671 - An IPv6 address followed by a colon (':') and optionally a UDP
4672 port. If no port is specified, 514 is used by default (the
4673 standard syslog port).
4674
Willy Tarreau2769aa02007-12-27 18:26:09 +01004675 - A filesystem path to a UNIX domain socket, keeping in mind
4676 considerations for chroot (be sure the path is accessible
4677 inside the chroot) and uid/gid (be sure the path is
4678 appropriately writeable).
4679
William Lallemandb2f07452015-05-12 14:27:13 +02004680 You may want to reference some environment variables in the
4681 address parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01004682
Willy Tarreau18324f52014-06-27 18:10:07 +02004683 <length> is an optional maximum line length. Log lines larger than this
4684 value will be truncated before being sent. The reason is that
4685 syslog servers act differently on log line length. All servers
4686 support the default value of 1024, but some servers simply drop
4687 larger lines while others do log them. If a server supports long
4688 lines, it may make sense to set this value here in order to avoid
4689 truncating long lines. Similarly, if a server drops long lines,
4690 it is preferable to truncate them before sending them. Accepted
4691 values are 80 to 65535 inclusive. The default value of 1024 is
4692 generally fine for all standard usages. Some specific cases of
4693 long captures or JSON-formated logs may require larger values.
4694
Willy Tarreau2769aa02007-12-27 18:26:09 +01004695 <facility> must be one of the 24 standard syslog facilities :
4696
4697 kern user mail daemon auth syslog lpr news
4698 uucp cron auth2 ftp ntp audit alert cron2
4699 local0 local1 local2 local3 local4 local5 local6 local7
4700
4701 <level> is optional and can be specified to filter outgoing messages. By
4702 default, all messages are sent. If a level is specified, only
4703 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02004704 will be sent. An optional minimum level can be specified. If it
4705 is set, logs emitted with a more severe level than this one will
4706 be capped to this level. This is used to avoid sending "emerg"
4707 messages on all terminals on some default syslog configurations.
4708 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01004709
4710 emerg alert crit err warning notice info debug
4711
William Lallemand0f99e342011-10-12 17:50:54 +02004712 It is important to keep in mind that it is the frontend which decides what to
4713 log from a connection, and that in case of content switching, the log entries
4714 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004715
4716 However, backend log declaration define how and where servers status changes
4717 will be logged. Level "notice" will be used to indicate a server going up,
4718 "warning" will be used for termination signals and definitive service
4719 termination, and "alert" will be used for when a server goes down.
4720
4721 Note : According to RFC3164, messages are truncated to 1024 bytes before
4722 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004723
4724 Example :
4725 log global
Willy Tarreauf7edefa2009-05-10 17:20:05 +02004726 log 127.0.0.1:514 local0 notice # only send important events
4727 log 127.0.0.1:514 local0 notice notice # same but limit output level
William Lallemandb2f07452015-05-12 14:27:13 +02004728 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
Willy Tarreaudad36a32013-03-11 01:20:04 +01004729
Willy Tarreau2769aa02007-12-27 18:26:09 +01004730
William Lallemand48940402012-01-30 16:47:22 +01004731log-format <string>
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01004732 Specifies the log format string to use for traffic logs
4733 May be used in sections: defaults | frontend | listen | backend
4734 yes | yes | yes | no
William Lallemand48940402012-01-30 16:47:22 +01004735
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01004736 This directive specifies the log format string that will be used for all logs
4737 resulting from traffic passing through the frontend using this line. If the
4738 directive is used in a defaults section, all subsequent frontends will use
4739 the same log format. Please see section 8.2.4 which covers the log format
4740 string in depth.
William Lallemand48940402012-01-30 16:47:22 +01004741
Dragan Dosen7ad31542015-09-28 17:16:47 +02004742log-format-sd <string>
4743 Specifies the RFC5424 structured-data log format string
4744 May be used in sections: defaults | frontend | listen | backend
4745 yes | yes | yes | no
4746
4747 This directive specifies the RFC5424 structured-data log format string that
4748 will be used for all logs resulting from traffic passing through the frontend
4749 using this line. If the directive is used in a defaults section, all
4750 subsequent frontends will use the same log format. Please see section 8.2.4
4751 which covers the log format string in depth.
4752
4753 See https://tools.ietf.org/html/rfc5424#section-6.3 for more information
4754 about the RFC5424 structured-data part.
4755
4756 Note : This log format string will be used only for loggers that have set
4757 log format to "rfc5424".
4758
4759 Example :
4760 log-format-sd [exampleSDID@1234\ bytes=\"%B\"\ status=\"%ST\"]
4761
4762
Willy Tarreau094af4e2015-01-07 15:03:42 +01004763log-tag <string>
4764 Specifies the log tag to use for all outgoing logs
4765 May be used in sections: defaults | frontend | listen | backend
4766 yes | yes | yes | yes
4767
4768 Sets the tag field in the syslog header to this string. It defaults to the
4769 log-tag set in the global section, otherwise the program name as launched
4770 from the command line, which usually is "haproxy". Sometimes it can be useful
4771 to differentiate between multiple processes running on the same host, or to
4772 differentiate customer instances running in the same process. In the backend,
4773 logs about servers up/down will use this tag. As a hint, it can be convenient
4774 to set a log-tag related to a hosted customer in a defaults section then put
4775 all the frontends and backends for that customer, then start another customer
4776 in a new defaults section. See also the global "log-tag" directive.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004777
Willy Tarreauc35362a2014-04-25 13:58:37 +02004778max-keep-alive-queue <value>
4779 Set the maximum server queue size for maintaining keep-alive connections
4780 May be used in sections: defaults | frontend | listen | backend
4781 yes | no | yes | yes
4782
4783 HTTP keep-alive tries to reuse the same server connection whenever possible,
4784 but sometimes it can be counter-productive, for example if a server has a lot
4785 of connections while other ones are idle. This is especially true for static
4786 servers.
4787
4788 The purpose of this setting is to set a threshold on the number of queued
4789 connections at which haproxy stops trying to reuse the same server and prefers
4790 to find another one. The default value, -1, means there is no limit. A value
4791 of zero means that keep-alive requests will never be queued. For very close
4792 servers which can be reached with a low latency and which are not sensible to
4793 breaking keep-alive, a low value is recommended (eg: local static server can
4794 use a value of 10 or less). For remote servers suffering from a high latency,
4795 higher values might be needed to cover for the latency and/or the cost of
4796 picking a different server.
4797
4798 Note that this has no impact on responses which are maintained to the same
4799 server consecutively to a 401 response. They will still go to the same server
4800 even if they have to be queued.
4801
4802 See also : "option http-server-close", "option prefer-last-server", server
4803 "maxconn" and cookie persistence.
4804
4805
Willy Tarreau2769aa02007-12-27 18:26:09 +01004806maxconn <conns>
4807 Fix the maximum number of concurrent connections on a frontend
4808 May be used in sections : defaults | frontend | listen | backend
4809 yes | yes | yes | no
4810 Arguments :
4811 <conns> is the maximum number of concurrent connections the frontend will
4812 accept to serve. Excess connections will be queued by the system
4813 in the socket's listen queue and will be served once a connection
4814 closes.
4815
4816 If the system supports it, it can be useful on big sites to raise this limit
4817 very high so that haproxy manages connection queues, instead of leaving the
4818 clients with unanswered connection attempts. This value should not exceed the
4819 global maxconn. Also, keep in mind that a connection contains two buffers
Baptiste Assmann79fb45d2016-03-06 23:34:31 +01004820 of tune.bufsize (16kB by default) each, as well as some other data resulting
4821 in about 33 kB of RAM being consumed per established connection. That means
4822 that a medium system equipped with 1GB of RAM can withstand around
4823 20000-25000 concurrent connections if properly tuned.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004824
4825 Also, when <conns> is set to large values, it is possible that the servers
4826 are not sized to accept such loads, and for this reason it is generally wise
4827 to assign them some reasonable connection limits.
4828
Vincent Bernat6341be52012-06-27 17:18:30 +02004829 By default, this value is set to 2000.
4830
Willy Tarreau2769aa02007-12-27 18:26:09 +01004831 See also : "server", global section's "maxconn", "fullconn"
4832
4833
4834mode { tcp|http|health }
4835 Set the running mode or protocol of the instance
4836 May be used in sections : defaults | frontend | listen | backend
4837 yes | yes | yes | yes
4838 Arguments :
4839 tcp The instance will work in pure TCP mode. A full-duplex connection
4840 will be established between clients and servers, and no layer 7
4841 examination will be performed. This is the default mode. It
4842 should be used for SSL, SSH, SMTP, ...
4843
4844 http The instance will work in HTTP mode. The client request will be
4845 analyzed in depth before connecting to any server. Any request
4846 which is not RFC-compliant will be rejected. Layer 7 filtering,
4847 processing and switching will be possible. This is the mode which
4848 brings HAProxy most of its value.
4849
4850 health The instance will work in "health" mode. It will just reply "OK"
Willy Tarreau82569f92012-09-27 23:48:56 +02004851 to incoming connections and close the connection. Alternatively,
4852 If the "httpchk" option is set, "HTTP/1.0 200 OK" will be sent
4853 instead. Nothing will be logged in either case. This mode is used
4854 to reply to external components health checks. This mode is
4855 deprecated and should not be used anymore as it is possible to do
4856 the same and even better by combining TCP or HTTP modes with the
4857 "monitor" keyword.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004858
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004859 When doing content switching, it is mandatory that the frontend and the
4860 backend are in the same mode (generally HTTP), otherwise the configuration
4861 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004862
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004863 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01004864 defaults http_instances
4865 mode http
4866
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004867 See also : "monitor", "monitor-net"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004868
Willy Tarreau0ba27502007-12-24 16:55:16 +01004869
Cyril Bontéf0c60612010-02-06 14:44:47 +01004870monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01004871 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004872 May be used in sections : defaults | frontend | listen | backend
4873 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01004874 Arguments :
4875 if <cond> the monitor request will fail if the condition is satisfied,
4876 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004877 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01004878 are met, for instance a low number of servers both in a
4879 backend and its backup.
4880
4881 unless <cond> the monitor request will succeed only if the condition is
4882 satisfied, and will fail otherwise. Such a condition may be
4883 based on a test on the presence of a minimum number of active
4884 servers in a list of backends.
4885
4886 This statement adds a condition which can force the response to a monitor
4887 request to report a failure. By default, when an external component queries
4888 the URI dedicated to monitoring, a 200 response is returned. When one of the
4889 conditions above is met, haproxy will return 503 instead of 200. This is
4890 very useful to report a site failure to an external component which may base
4891 routing advertisements between multiple sites on the availability reported by
4892 haproxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004893 criterion. Note that "monitor fail" only works in HTTP mode. Both status
4894 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004895
4896 Example:
4897 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01004898 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01004899 acl site_dead nbsrv(dynamic) lt 2
4900 acl site_dead nbsrv(static) lt 2
4901 monitor-uri /site_alive
4902 monitor fail if site_dead
4903
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004904 See also : "monitor-net", "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004905
4906
4907monitor-net <source>
4908 Declare a source network which is limited to monitor requests
4909 May be used in sections : defaults | frontend | listen | backend
4910 yes | yes | yes | no
4911 Arguments :
4912 <source> is the source IPv4 address or network which will only be able to
4913 get monitor responses to any request. It can be either an IPv4
4914 address, a host name, or an address followed by a slash ('/')
4915 followed by a mask.
4916
4917 In TCP mode, any connection coming from a source matching <source> will cause
4918 the connection to be immediately closed without any log. This allows another
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004919 equipment to probe the port and verify that it is still listening, without
Willy Tarreau2769aa02007-12-27 18:26:09 +01004920 forwarding the connection to a remote server.
4921
4922 In HTTP mode, a connection coming from a source matching <source> will be
4923 accepted, the following response will be sent without waiting for a request,
4924 then the connection will be closed : "HTTP/1.0 200 OK". This is normally
4925 enough for any front-end HTTP probe to detect that the service is UP and
Willy Tarreau82569f92012-09-27 23:48:56 +02004926 running without forwarding the request to a backend server. Note that this
4927 response is sent in raw format, without any transformation. This is important
4928 as it means that it will not be SSL-encrypted on SSL listeners.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004929
Willy Tarreau82569f92012-09-27 23:48:56 +02004930 Monitor requests are processed very early, just after tcp-request connection
4931 ACLs which are the only ones able to block them. These connections are short
4932 lived and never wait for any data from the client. They cannot be logged, and
4933 it is the intended purpose. They are only used to report HAProxy's health to
4934 an upper component, nothing more. Please note that "monitor fail" rules do
4935 not apply to connections intercepted by "monitor-net".
Willy Tarreau2769aa02007-12-27 18:26:09 +01004936
Willy Tarreau95cd2832010-03-04 23:36:33 +01004937 Last, please note that only one "monitor-net" statement can be specified in
4938 a frontend. If more than one is found, only the last one will be considered.
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004939
Willy Tarreau2769aa02007-12-27 18:26:09 +01004940 Example :
4941 # addresses .252 and .253 are just probing us.
4942 frontend www
4943 monitor-net 192.168.0.252/31
4944
4945 See also : "monitor fail", "monitor-uri"
4946
4947
4948monitor-uri <uri>
4949 Intercept a URI used by external components' monitor requests
4950 May be used in sections : defaults | frontend | listen | backend
4951 yes | yes | yes | no
4952 Arguments :
4953 <uri> is the exact URI which we want to intercept to return HAProxy's
4954 health status instead of forwarding the request.
4955
4956 When an HTTP request referencing <uri> will be received on a frontend,
4957 HAProxy will not forward it nor log it, but instead will return either
4958 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
4959 conditions defined with "monitor fail". This is normally enough for any
4960 front-end HTTP probe to detect that the service is UP and running without
4961 forwarding the request to a backend server. Note that the HTTP method, the
4962 version and all headers are ignored, but the request must at least be valid
4963 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
4964
4965 Monitor requests are processed very early. It is not possible to block nor
4966 divert them using ACLs. They cannot be logged either, and it is the intended
4967 purpose. They are only used to report HAProxy's health to an upper component,
4968 nothing more. However, it is possible to add any number of conditions using
4969 "monitor fail" and ACLs so that the result can be adjusted to whatever check
4970 can be imagined (most often the number of available servers in a backend).
4971
4972 Example :
4973 # Use /haproxy_test to report haproxy's status
4974 frontend www
4975 mode http
4976 monitor-uri /haproxy_test
4977
4978 See also : "monitor fail", "monitor-net"
4979
Willy Tarreau0ba27502007-12-24 16:55:16 +01004980
Willy Tarreaubf1f8162007-12-28 17:42:56 +01004981option abortonclose
4982no option abortonclose
4983 Enable or disable early dropping of aborted requests pending in queues.
4984 May be used in sections : defaults | frontend | listen | backend
4985 yes | no | yes | yes
4986 Arguments : none
4987
4988 In presence of very high loads, the servers will take some time to respond.
4989 The per-instance connection queue will inflate, and the response time will
4990 increase respective to the size of the queue times the average per-session
4991 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01004992 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01004993 the queue, and slowing down other users, and the servers as well, because the
4994 request will eventually be served, then aborted at the first error
4995 encountered while delivering the response.
4996
4997 As there is no way to distinguish between a full STOP and a simple output
4998 close on the client side, HTTP agents should be conservative and consider
4999 that the client might only have closed its output channel while waiting for
5000 the response. However, this introduces risks of congestion when lots of users
5001 do the same, and is completely useless nowadays because probably no client at
5002 all will close the session while waiting for the response. Some HTTP agents
Willy Tarreaud72758d2010-01-12 10:42:19 +01005003 support this behaviour (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005004 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01005005 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005006 of being the single component to break rare but valid traffic is extremely
5007 low, which adds to the temptation to be able to abort a session early while
5008 still not served and not pollute the servers.
5009
5010 In HAProxy, the user can choose the desired behaviour using the option
5011 "abortonclose". By default (without the option) the behaviour is HTTP
5012 compliant and aborted requests will be served. But when the option is
5013 specified, a session with an incoming channel closed will be aborted while
5014 it is still possible, either pending in the queue for a connection slot, or
5015 during the connection establishment if the server has not yet acknowledged
5016 the connection request. This considerably reduces the queue size and the load
5017 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01005018 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005019
5020 If this option has been enabled in a "defaults" section, it can be disabled
5021 in a specific instance by prepending the "no" keyword before it.
5022
5023 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
5024
5025
Willy Tarreau4076a152009-04-02 15:18:36 +02005026option accept-invalid-http-request
5027no option accept-invalid-http-request
5028 Enable or disable relaxing of HTTP request parsing
5029 May be used in sections : defaults | frontend | listen | backend
5030 yes | yes | yes | no
5031 Arguments : none
5032
Willy Tarreau91852eb2015-05-01 13:26:00 +02005033 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005034 means that invalid characters in header names are not permitted and cause an
5035 error to be returned to the client. This is the desired behaviour as such
5036 forbidden characters are essentially used to build attacks exploiting server
5037 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5038 server will emit invalid header names for whatever reason (configuration,
5039 implementation) and the issue will not be immediately fixed. In such a case,
5040 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01005041 even if that does not make sense, by specifying this option. Similarly, the
5042 list of characters allowed to appear in a URI is well defined by RFC3986, and
5043 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
5044 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
5045 not allowed at all. Haproxy always blocks a number of them (0..32, 127). The
Willy Tarreau91852eb2015-05-01 13:26:00 +02005046 remaining ones are blocked by default unless this option is enabled. This
Willy Tarreau13317662015-05-01 13:47:08 +02005047 option also relaxes the test on the HTTP version, it allows HTTP/0.9 requests
5048 to pass through (no version specified) and multiple digits for both the major
5049 and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005050
5051 This option should never be enabled by default as it hides application bugs
5052 and open security breaches. It should only be deployed after a problem has
5053 been confirmed.
5054
5055 When this option is enabled, erroneous header names will still be accepted in
5056 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01005057 analysis using the "show errors" request on the UNIX stats socket. Similarly,
5058 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02005059 also helps confirming that the issue has been solved.
5060
5061 If this option has been enabled in a "defaults" section, it can be disabled
5062 in a specific instance by prepending the "no" keyword before it.
5063
5064 See also : "option accept-invalid-http-response" and "show errors" on the
5065 stats socket.
5066
5067
5068option accept-invalid-http-response
5069no option accept-invalid-http-response
5070 Enable or disable relaxing of HTTP response parsing
5071 May be used in sections : defaults | frontend | listen | backend
5072 yes | no | yes | yes
5073 Arguments : none
5074
Willy Tarreau91852eb2015-05-01 13:26:00 +02005075 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005076 means that invalid characters in header names are not permitted and cause an
5077 error to be returned to the client. This is the desired behaviour as such
5078 forbidden characters are essentially used to build attacks exploiting server
5079 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5080 server will emit invalid header names for whatever reason (configuration,
5081 implementation) and the issue will not be immediately fixed. In such a case,
5082 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau91852eb2015-05-01 13:26:00 +02005083 even if that does not make sense, by specifying this option. This option also
5084 relaxes the test on the HTTP version format, it allows multiple digits for
5085 both the major and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005086
5087 This option should never be enabled by default as it hides application bugs
5088 and open security breaches. It should only be deployed after a problem has
5089 been confirmed.
5090
5091 When this option is enabled, erroneous header names will still be accepted in
5092 responses, but the complete response will be captured in order to permit
5093 later analysis using the "show errors" request on the UNIX stats socket.
5094 Doing this also helps confirming that the issue has been solved.
5095
5096 If this option has been enabled in a "defaults" section, it can be disabled
5097 in a specific instance by prepending the "no" keyword before it.
5098
5099 See also : "option accept-invalid-http-request" and "show errors" on the
5100 stats socket.
5101
5102
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005103option allbackups
5104no option allbackups
5105 Use either all backup servers at a time or only the first one
5106 May be used in sections : defaults | frontend | listen | backend
5107 yes | no | yes | yes
5108 Arguments : none
5109
5110 By default, the first operational backup server gets all traffic when normal
5111 servers are all down. Sometimes, it may be preferred to use multiple backups
5112 at once, because one will not be enough. When "option allbackups" is enabled,
5113 the load balancing will be performed among all backup servers when all normal
5114 ones are unavailable. The same load balancing algorithm will be used and the
5115 servers' weights will be respected. Thus, there will not be any priority
5116 order between the backup servers anymore.
5117
5118 This option is mostly used with static server farms dedicated to return a
5119 "sorry" page when an application is completely offline.
5120
5121 If this option has been enabled in a "defaults" section, it can be disabled
5122 in a specific instance by prepending the "no" keyword before it.
5123
5124
5125option checkcache
5126no option checkcache
Godbach7056a352013-12-11 20:01:07 +08005127 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005128 May be used in sections : defaults | frontend | listen | backend
5129 yes | no | yes | yes
5130 Arguments : none
5131
5132 Some high-level frameworks set application cookies everywhere and do not
5133 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005134 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005135 high risk of session crossing or stealing between users traversing the same
5136 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02005137 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005138
5139 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005140 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01005141 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005142 response to check if there's a risk of caching a cookie on a client-side
5143 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01005144 to the client are :
5145 - all those without "Set-Cookie" header ;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005146 - all those with a return code other than 200, 203, 206, 300, 301, 410,
Willy Tarreau198a7442008-01-17 12:05:32 +01005147 provided that the server has not set a "Cache-control: public" header ;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005148 - all those that come from a POST request, provided that the server has not
5149 set a 'Cache-Control: public' header ;
5150 - those with a 'Pragma: no-cache' header
5151 - those with a 'Cache-control: private' header
5152 - those with a 'Cache-control: no-store' header
5153 - those with a 'Cache-control: max-age=0' header
5154 - those with a 'Cache-control: s-maxage=0' header
5155 - those with a 'Cache-control: no-cache' header
5156 - those with a 'Cache-control: no-cache="set-cookie"' header
5157 - those with a 'Cache-control: no-cache="set-cookie,' header
5158 (allowing other fields after set-cookie)
5159
5160 If a response doesn't respect these requirements, then it will be blocked
Willy Tarreau198a7442008-01-17 12:05:32 +01005161 just as if it was from an "rspdeny" filter, with an "HTTP 502 bad gateway".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005162 The session state shows "PH--" meaning that the proxy blocked the response
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005163 during headers processing. Additionally, an alert will be sent in the logs so
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005164 that admins are informed that there's something to be fixed.
5165
5166 Due to the high impact on the application, the application should be tested
5167 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01005168 good practice to always activate it during tests, even if it is not used in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005169 production, as it will report potentially dangerous application behaviours.
5170
5171 If this option has been enabled in a "defaults" section, it can be disabled
5172 in a specific instance by prepending the "no" keyword before it.
5173
5174
5175option clitcpka
5176no option clitcpka
5177 Enable or disable the sending of TCP keepalive packets on the client side
5178 May be used in sections : defaults | frontend | listen | backend
5179 yes | yes | yes | no
5180 Arguments : none
5181
5182 When there is a firewall or any session-aware component between a client and
5183 a server, and when the protocol involves very long sessions with long idle
5184 periods (eg: remote desktops), there is a risk that one of the intermediate
5185 components decides to expire a session which has remained idle for too long.
5186
5187 Enabling socket-level TCP keep-alives makes the system regularly send packets
5188 to the other end of the connection, leaving it active. The delay between
5189 keep-alive probes is controlled by the system only and depends both on the
5190 operating system and its tuning parameters.
5191
5192 It is important to understand that keep-alive packets are neither emitted nor
5193 received at the application level. It is only the network stacks which sees
5194 them. For this reason, even if one side of the proxy already uses keep-alives
5195 to maintain its connection alive, those keep-alive packets will not be
5196 forwarded to the other side of the proxy.
5197
5198 Please note that this has nothing to do with HTTP keep-alive.
5199
5200 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
5201 client side of a connection, which should help when session expirations are
5202 noticed between HAProxy and a client.
5203
5204 If this option has been enabled in a "defaults" section, it can be disabled
5205 in a specific instance by prepending the "no" keyword before it.
5206
5207 See also : "option srvtcpka", "option tcpka"
5208
5209
Willy Tarreau0ba27502007-12-24 16:55:16 +01005210option contstats
5211 Enable continuous traffic statistics updates
5212 May be used in sections : defaults | frontend | listen | backend
5213 yes | yes | yes | no
5214 Arguments : none
5215
5216 By default, counters used for statistics calculation are incremented
5217 only when a session finishes. It works quite well when serving small
5218 objects, but with big ones (for example large images or archives) or
5219 with A/V streaming, a graph generated from haproxy counters looks like
Willy Tarreaudef0d222016-11-08 22:03:00 +01005220 a hedgehog. With this option enabled counters get incremented frequently
5221 along the session, typically every 5 seconds, which is often enough to
5222 produce clean graphs. Recounting touches a hotpath directly so it is not
5223 not enabled by default, as it can cause a lot of wakeups for very large
5224 session counts and cause a small performance drop.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005225
5226
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005227option dontlog-normal
5228no option dontlog-normal
5229 Enable or disable logging of normal, successful connections
5230 May be used in sections : defaults | frontend | listen | backend
5231 yes | yes | yes | no
5232 Arguments : none
5233
5234 There are large sites dealing with several thousand connections per second
5235 and for which logging is a major pain. Some of them are even forced to turn
5236 logs off and cannot debug production issues. Setting this option ensures that
5237 normal connections, those which experience no error, no timeout, no retry nor
5238 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
5239 mode, the response status code is checked and return codes 5xx will still be
5240 logged.
5241
5242 It is strongly discouraged to use this option as most of the time, the key to
5243 complex issues is in the normal logs which will not be logged here. If you
5244 need to separate logs, see the "log-separate-errors" option instead.
5245
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005246 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005247 logging.
5248
5249
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005250option dontlognull
5251no option dontlognull
5252 Enable or disable logging of null connections
5253 May be used in sections : defaults | frontend | listen | backend
5254 yes | yes | yes | no
5255 Arguments : none
5256
5257 In certain environments, there are components which will regularly connect to
5258 various systems to ensure that they are still alive. It can be the case from
5259 another load balancer as well as from monitoring systems. By default, even a
5260 simple port probe or scan will produce a log. If those connections pollute
5261 the logs too much, it is possible to enable option "dontlognull" to indicate
5262 that a connection on which no data has been transferred will not be logged,
Willy Tarreau0f228a02015-05-01 15:37:53 +02005263 which typically corresponds to those probes. Note that errors will still be
5264 returned to the client and accounted for in the stats. If this is not what is
5265 desired, option http-ignore-probes can be used instead.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005266
5267 It is generally recommended not to use this option in uncontrolled
5268 environments (eg: internet), otherwise scans and other malicious activities
5269 would not be logged.
5270
5271 If this option has been enabled in a "defaults" section, it can be disabled
5272 in a specific instance by prepending the "no" keyword before it.
5273
Willy Tarreau0f228a02015-05-01 15:37:53 +02005274 See also : "log", "http-ignore-probes", "monitor-net", "monitor-uri", and
5275 section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005276
5277
5278option forceclose
5279no option forceclose
5280 Enable or disable active connection closing after response is transferred.
5281 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaua31e5df2009-12-30 01:10:35 +01005282 yes | yes | yes | yes
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005283 Arguments : none
5284
5285 Some HTTP servers do not necessarily close the connections when they receive
5286 the "Connection: close" set by "option httpclose", and if the client does not
5287 close either, then the connection remains open till the timeout expires. This
5288 causes high number of simultaneous connections on the servers and shows high
5289 global session times in the logs.
5290
5291 When this happens, it is possible to use "option forceclose". It will
Willy Tarreau82eeaf22009-12-29 12:09:05 +01005292 actively close the outgoing server channel as soon as the server has finished
Cyril Bonté653dcd62014-02-20 00:13:15 +01005293 to respond and release some resources earlier than with "option httpclose".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005294
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005295 This option may also be combined with "option http-pretend-keepalive", which
5296 will disable sending of the "Connection: close" header, but will still cause
5297 the connection to be closed once the whole response is received.
5298
Cyril Bonté653dcd62014-02-20 00:13:15 +01005299 This option disables and replaces any previous "option httpclose", "option
5300 http-server-close", "option http-keep-alive", or "option http-tunnel".
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005301
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005302 If this option has been enabled in a "defaults" section, it can be disabled
5303 in a specific instance by prepending the "no" keyword before it.
5304
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005305 See also : "option httpclose" and "option http-pretend-keepalive"
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005306
5307
Willy Tarreau87cf5142011-08-19 22:57:24 +02005308option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01005309 Enable insertion of the X-Forwarded-For header to requests sent to servers
5310 May be used in sections : defaults | frontend | listen | backend
5311 yes | yes | yes | yes
5312 Arguments :
5313 <network> is an optional argument used to disable this option for sources
5314 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02005315 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01005316 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005317
5318 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
5319 their client address. This is sometimes annoying when the client's IP address
5320 is expected in server logs. To solve this problem, the well-known HTTP header
5321 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
5322 This header contains a value representing the client's IP address. Since this
5323 header is always appended at the end of the existing header list, the server
5324 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02005325 the server's manual to find how to enable use of this standard header. Note
5326 that only the last occurrence of the header must be used, since it is really
5327 possible that the client has already brought one.
5328
Willy Tarreaud72758d2010-01-12 10:42:19 +01005329 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02005330 the default "X-Forwarded-For". This can be useful where you might already
Willy Tarreaud72758d2010-01-12 10:42:19 +01005331 have a "X-Forwarded-For" header from a different application (eg: stunnel),
5332 and you need preserve it. Also if your backend server doesn't use the
Ross Westaf72a1d2008-08-03 10:51:45 +02005333 "X-Forwarded-For" header and requires different one (eg: Zeus Web Servers
5334 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01005335
5336 Sometimes, a same HAProxy instance may be shared between a direct client
5337 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
5338 used to decrypt HTTPS traffic). It is possible to disable the addition of the
5339 header for a known source address or network by adding the "except" keyword
5340 followed by the network address. In this case, any source IP matching the
5341 network will not cause an addition of this header. Most common uses are with
5342 private networks or 127.0.0.1.
5343
Willy Tarreau87cf5142011-08-19 22:57:24 +02005344 Alternatively, the keyword "if-none" states that the header will only be
5345 added if it is not present. This should only be used in perfectly trusted
5346 environment, as this might cause a security issue if headers reaching haproxy
5347 are under the control of the end-user.
5348
Willy Tarreauc27debf2008-01-06 08:57:02 +01005349 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02005350 least one of them uses it, the header will be added. Note that the backend's
5351 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02005352 both are defined. In the case of the "if-none" argument, if at least one of
5353 the frontend or the backend does not specify it, it wants the addition to be
5354 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005355
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005356 Example :
Willy Tarreauc27debf2008-01-06 08:57:02 +01005357 # Public HTTP address also used by stunnel on the same machine
5358 frontend www
5359 mode http
5360 option forwardfor except 127.0.0.1 # stunnel already adds the header
5361
Ross Westaf72a1d2008-08-03 10:51:45 +02005362 # Those servers want the IP Address in X-Client
5363 backend www
5364 mode http
5365 option forwardfor header X-Client
5366
Willy Tarreau87cf5142011-08-19 22:57:24 +02005367 See also : "option httpclose", "option http-server-close",
Willy Tarreau70dffda2014-01-30 03:07:23 +01005368 "option forceclose", "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01005369
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005370
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005371option http-buffer-request
5372no option http-buffer-request
5373 Enable or disable waiting for whole HTTP request body before proceeding
5374 May be used in sections : defaults | frontend | listen | backend
5375 yes | yes | yes | yes
5376 Arguments : none
5377
5378 It is sometimes desirable to wait for the body of an HTTP request before
5379 taking a decision. This is what is being done by "balance url_param" for
5380 example. The first use case is to buffer requests from slow clients before
5381 connecting to the server. Another use case consists in taking the routing
5382 decision based on the request body's contents. This option placed in a
5383 frontend or backend forces the HTTP processing to wait until either the whole
5384 body is received, or the request buffer is full, or the first chunk is
5385 complete in case of chunked encoding. It can have undesired side effects with
Tim Düsterhus4896c442016-11-29 02:15:19 +01005386 some applications abusing HTTP by expecting unbuffered transmissions between
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005387 the frontend and the backend, so this should definitely not be used by
5388 default.
5389
Baptiste Assmanneccdf432015-10-28 13:49:01 +01005390 See also : "option http-no-delay", "timeout http-request"
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005391
5392
Willy Tarreau0f228a02015-05-01 15:37:53 +02005393option http-ignore-probes
5394no option http-ignore-probes
5395 Enable or disable logging of null connections and request timeouts
5396 May be used in sections : defaults | frontend | listen | backend
5397 yes | yes | yes | no
5398 Arguments : none
5399
5400 Recently some browsers started to implement a "pre-connect" feature
5401 consisting in speculatively connecting to some recently visited web sites
5402 just in case the user would like to visit them. This results in many
5403 connections being established to web sites, which end up in 408 Request
5404 Timeout if the timeout strikes first, or 400 Bad Request when the browser
5405 decides to close them first. These ones pollute the log and feed the error
5406 counters. There was already "option dontlognull" but it's insufficient in
5407 this case. Instead, this option does the following things :
5408 - prevent any 400/408 message from being sent to the client if nothing
5409 was received over a connection before it was closed ;
5410 - prevent any log from being emitted in this situation ;
5411 - prevent any error counter from being incremented
5412
5413 That way the empty connection is silently ignored. Note that it is better
5414 not to use this unless it is clear that it is needed, because it will hide
5415 real problems. The most common reason for not receiving a request and seeing
5416 a 408 is due to an MTU inconsistency between the client and an intermediary
5417 element such as a VPN, which blocks too large packets. These issues are
5418 generally seen with POST requests as well as GET with large cookies. The logs
5419 are often the only way to detect them.
5420
5421 If this option has been enabled in a "defaults" section, it can be disabled
5422 in a specific instance by prepending the "no" keyword before it.
5423
5424 See also : "log", "dontlognull", "errorfile", and section 8 about logging.
5425
5426
Willy Tarreau16bfb022010-01-16 19:48:41 +01005427option http-keep-alive
5428no option http-keep-alive
5429 Enable or disable HTTP keep-alive from client to server
5430 May be used in sections : defaults | frontend | listen | backend
5431 yes | yes | yes | yes
5432 Arguments : none
5433
Willy Tarreau70dffda2014-01-30 03:07:23 +01005434 By default HAProxy operates in keep-alive mode with regards to persistent
5435 connections: for each connection it processes each request and response, and
5436 leaves the connection idle on both sides between the end of a response and the
5437 start of a new request. This mode may be changed by several options such as
5438 "option http-server-close", "option forceclose", "option httpclose" or
5439 "option http-tunnel". This option allows to set back the keep-alive mode,
5440 which can be useful when another mode was used in a defaults section.
5441
5442 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
5443 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01005444 network) and the fastest session reuse on the server side at the expense
5445 of maintaining idle connections to the servers. In general, it is possible
5446 with this option to achieve approximately twice the request rate that the
5447 "http-server-close" option achieves on small objects. There are mainly two
5448 situations where this option may be useful :
5449
5450 - when the server is non-HTTP compliant and authenticates the connection
5451 instead of requests (eg: NTLM authentication)
5452
5453 - when the cost of establishing the connection to the server is significant
5454 compared to the cost of retrieving the associated object from the server.
5455
5456 This last case can happen when the server is a fast static server of cache.
5457 In this case, the server will need to be properly tuned to support high enough
5458 connection counts because connections will last until the client sends another
5459 request.
5460
5461 If the client request has to go to another backend or another server due to
5462 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01005463 immediately be closed and a new one re-opened. Option "prefer-last-server" is
5464 available to try optimize server selection so that if the server currently
5465 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01005466
5467 In general it is preferred to use "option http-server-close" with application
5468 servers, and some static servers might benefit from "option http-keep-alive".
5469
5470 At the moment, logs will not indicate whether requests came from the same
5471 session or not. The accept date reported in the logs corresponds to the end
5472 of the previous request, and the request time corresponds to the time spent
5473 waiting for a new request. The keep-alive request time is still bound to the
5474 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
5475 not set.
5476
Cyril Bonté653dcd62014-02-20 00:13:15 +01005477 This option disables and replaces any previous "option httpclose", "option
5478 http-server-close", "option forceclose" or "option http-tunnel". When backend
Willy Tarreau70dffda2014-01-30 03:07:23 +01005479 and frontend options differ, all of these 4 options have precedence over
Cyril Bonté653dcd62014-02-20 00:13:15 +01005480 "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01005481
5482 See also : "option forceclose", "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01005483 "option prefer-last-server", "option http-pretend-keepalive",
5484 "option httpclose", and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01005485
5486
Willy Tarreau96e31212011-05-30 18:10:30 +02005487option http-no-delay
5488no option http-no-delay
5489 Instruct the system to favor low interactive delays over performance in HTTP
5490 May be used in sections : defaults | frontend | listen | backend
5491 yes | yes | yes | yes
5492 Arguments : none
5493
5494 In HTTP, each payload is unidirectional and has no notion of interactivity.
5495 Any agent is expected to queue data somewhat for a reasonably low delay.
5496 There are some very rare server-to-server applications that abuse the HTTP
5497 protocol and expect the payload phase to be highly interactive, with many
5498 interleaved data chunks in both directions within a single request. This is
5499 absolutely not supported by the HTTP specification and will not work across
5500 most proxies or servers. When such applications attempt to do this through
5501 haproxy, it works but they will experience high delays due to the network
5502 optimizations which favor performance by instructing the system to wait for
5503 enough data to be available in order to only send full packets. Typical
5504 delays are around 200 ms per round trip. Note that this only happens with
5505 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
5506 affected.
5507
5508 When "option http-no-delay" is present in either the frontend or the backend
5509 used by a connection, all such optimizations will be disabled in order to
5510 make the exchanges as fast as possible. Of course this offers no guarantee on
5511 the functionality, as it may break at any other place. But if it works via
5512 HAProxy, it will work as fast as possible. This option should never be used
5513 by default, and should never be used at all unless such a buggy application
5514 is discovered. The impact of using this option is an increase of bandwidth
5515 usage and CPU usage, which may significantly lower performance in high
5516 latency environments.
5517
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02005518 See also : "option http-buffer-request"
5519
Willy Tarreau96e31212011-05-30 18:10:30 +02005520
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005521option http-pretend-keepalive
5522no option http-pretend-keepalive
5523 Define whether haproxy will announce keepalive to the server or not
5524 May be used in sections : defaults | frontend | listen | backend
5525 yes | yes | yes | yes
5526 Arguments : none
5527
5528 When running with "option http-server-close" or "option forceclose", haproxy
5529 adds a "Connection: close" header to the request forwarded to the server.
5530 Unfortunately, when some servers see this header, they automatically refrain
5531 from using the chunked encoding for responses of unknown length, while this
5532 is totally unrelated. The immediate effect is that this prevents haproxy from
5533 maintaining the client connection alive. A second effect is that a client or
5534 a cache could receive an incomplete response without being aware of it, and
5535 consider the response complete.
5536
5537 By setting "option http-pretend-keepalive", haproxy will make the server
5538 believe it will keep the connection alive. The server will then not fall back
5539 to the abnormal undesired above. When haproxy gets the whole response, it
5540 will close the connection with the server just as it would do with the
5541 "forceclose" option. That way the client gets a normal response and the
5542 connection is correctly closed on the server side.
5543
5544 It is recommended not to enable this option by default, because most servers
5545 will more efficiently close the connection themselves after the last packet,
5546 and release its buffers slightly earlier. Also, the added packet on the
5547 network could slightly reduce the overall peak performance. However it is
5548 worth noting that when this option is enabled, haproxy will have slightly
5549 less work to do. So if haproxy is the bottleneck on the whole architecture,
5550 enabling this option might save a few CPU cycles.
5551
5552 This option may be set both in a frontend and in a backend. It is enabled if
5553 at least one of the frontend or backend holding a connection has it enabled.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005554 This option may be combined with "option httpclose", which will cause
Willy Tarreau22a95342010-09-29 14:31:41 +02005555 keepalive to be announced to the server and close to be announced to the
5556 client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005557
5558 If this option has been enabled in a "defaults" section, it can be disabled
5559 in a specific instance by prepending the "no" keyword before it.
5560
Willy Tarreau16bfb022010-01-16 19:48:41 +01005561 See also : "option forceclose", "option http-server-close", and
5562 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02005563
Willy Tarreauc27debf2008-01-06 08:57:02 +01005564
Willy Tarreaub608feb2010-01-02 22:47:18 +01005565option http-server-close
5566no option http-server-close
5567 Enable or disable HTTP connection closing on the server side
5568 May be used in sections : defaults | frontend | listen | backend
5569 yes | yes | yes | yes
5570 Arguments : none
5571
Willy Tarreau70dffda2014-01-30 03:07:23 +01005572 By default HAProxy operates in keep-alive mode with regards to persistent
5573 connections: for each connection it processes each request and response, and
5574 leaves the connection idle on both sides between the end of a response and
5575 the start of a new request. This mode may be changed by several options such
5576 as "option http-server-close", "option forceclose", "option httpclose" or
5577 "option http-tunnel". Setting "option http-server-close" enables HTTP
5578 connection-close mode on the server side while keeping the ability to support
5579 HTTP keep-alive and pipelining on the client side. This provides the lowest
5580 latency on the client side (slow network) and the fastest session reuse on
5581 the server side to save server resources, similarly to "option forceclose".
5582 It also permits non-keepalive capable servers to be served in keep-alive mode
5583 to the clients if they conform to the requirements of RFC2616. Please note
5584 that some servers do not always conform to those requirements when they see
5585 "Connection: close" in the request. The effect will be that keep-alive will
5586 never be used. A workaround consists in enabling "option
5587 http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01005588
5589 At the moment, logs will not indicate whether requests came from the same
5590 session or not. The accept date reported in the logs corresponds to the end
5591 of the previous request, and the request time corresponds to the time spent
5592 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01005593 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
5594 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01005595
5596 This option may be set both in a frontend and in a backend. It is enabled if
5597 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01005598 It disables and replaces any previous "option httpclose", "option forceclose",
5599 "option http-tunnel" or "option http-keep-alive". Please check section 4
Willy Tarreau70dffda2014-01-30 03:07:23 +01005600 ("Proxies") to see how this option combines with others when frontend and
5601 backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01005602
5603 If this option has been enabled in a "defaults" section, it can be disabled
5604 in a specific instance by prepending the "no" keyword before it.
5605
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02005606 See also : "option forceclose", "option http-pretend-keepalive",
Willy Tarreau16bfb022010-01-16 19:48:41 +01005607 "option httpclose", "option http-keep-alive", and
5608 "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01005609
5610
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005611option http-tunnel
5612no option http-tunnel
5613 Disable or enable HTTP connection processing after first transaction
5614 May be used in sections : defaults | frontend | listen | backend
5615 yes | yes | yes | yes
5616 Arguments : none
5617
Willy Tarreau70dffda2014-01-30 03:07:23 +01005618 By default HAProxy operates in keep-alive mode with regards to persistent
5619 connections: for each connection it processes each request and response, and
5620 leaves the connection idle on both sides between the end of a response and
5621 the start of a new request. This mode may be changed by several options such
5622 as "option http-server-close", "option forceclose", "option httpclose" or
5623 "option http-tunnel".
5624
5625 Option "http-tunnel" disables any HTTP processing past the first request and
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005626 the first response. This is the mode which was used by default in versions
Willy Tarreau70dffda2014-01-30 03:07:23 +01005627 1.0 to 1.5-dev21. It is the mode with the lowest processing overhead, which
5628 is normally not needed anymore unless in very specific cases such as when
5629 using an in-house protocol that looks like HTTP but is not compatible, or
5630 just to log one request per client in order to reduce log size. Note that
5631 everything which works at the HTTP level, including header parsing/addition,
5632 cookie processing or content switching will only work for the first request
5633 and will be ignored after the first response.
Willy Tarreau02bce8b2014-01-30 00:15:28 +01005634
5635 If this option has been enabled in a "defaults" section, it can be disabled
5636 in a specific instance by prepending the "no" keyword before it.
5637
5638 See also : "option forceclose", "option http-server-close",
5639 "option httpclose", "option http-keep-alive", and
5640 "1.1. The HTTP transaction model".
5641
5642
Willy Tarreau88d349d2010-01-25 12:15:43 +01005643option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01005644no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01005645 Make use of non-standard Proxy-Connection header instead of Connection
5646 May be used in sections : defaults | frontend | listen | backend
5647 yes | yes | yes | no
5648 Arguments : none
5649
5650 While RFC2616 explicitly states that HTTP/1.1 agents must use the
5651 Connection header to indicate their wish of persistent or non-persistent
5652 connections, both browsers and proxies ignore this header for proxied
5653 connections and make use of the undocumented, non-standard Proxy-Connection
5654 header instead. The issue begins when trying to put a load balancer between
5655 browsers and such proxies, because there will be a difference between what
5656 haproxy understands and what the client and the proxy agree on.
5657
5658 By setting this option in a frontend, haproxy can automatically switch to use
5659 that non-standard header if it sees proxied requests. A proxied request is
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01005660 defined here as one where the URI begins with neither a '/' nor a '*'. This
5661 is incompatible with the HTTP tunnel mode. Note that this option can only be
5662 specified in a frontend and will affect the request along its whole life.
Willy Tarreau88d349d2010-01-25 12:15:43 +01005663
Willy Tarreau844a7e72010-01-31 21:46:18 +01005664 Also, when this option is set, a request which requires authentication will
5665 automatically switch to use proxy authentication headers if it is itself a
5666 proxied request. That makes it possible to check or enforce authentication in
5667 front of an existing proxy.
5668
Willy Tarreau88d349d2010-01-25 12:15:43 +01005669 This option should normally never be used, except in front of a proxy.
5670
5671 See also : "option httpclose", "option forceclose" and "option
5672 http-server-close".
5673
5674
Willy Tarreaud63335a2010-02-26 12:56:52 +01005675option httpchk
5676option httpchk <uri>
5677option httpchk <method> <uri>
5678option httpchk <method> <uri> <version>
5679 Enable HTTP protocol to check on the servers health
5680 May be used in sections : defaults | frontend | listen | backend
5681 yes | no | yes | yes
5682 Arguments :
5683 <method> is the optional HTTP method used with the requests. When not set,
5684 the "OPTIONS" method is used, as it generally requires low server
5685 processing and is easy to filter out from the logs. Any method
5686 may be used, though it is not recommended to invent non-standard
5687 ones.
5688
5689 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
5690 which is accessible by default on almost any server, but may be
5691 changed to any other URI. Query strings are permitted.
5692
5693 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
5694 but some servers might behave incorrectly in HTTP 1.0, so turning
5695 it to HTTP/1.1 may sometimes help. Note that the Host field is
5696 mandatory in HTTP/1.1, and as a trick, it is possible to pass it
5697 after "\r\n" following the version string.
5698
5699 By default, server health checks only consist in trying to establish a TCP
5700 connection. When "option httpchk" is specified, a complete HTTP request is
5701 sent once the TCP connection is established, and responses 2xx and 3xx are
5702 considered valid, while all other ones indicate a server failure, including
5703 the lack of any response.
5704
5705 The port and interval are specified in the server configuration.
5706
5707 This option does not necessarily require an HTTP backend, it also works with
5708 plain TCP backends. This is particularly useful to check simple scripts bound
5709 to some dedicated ports using the inetd daemon.
5710
5711 Examples :
5712 # Relay HTTPS traffic to Apache instance and check service availability
5713 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
5714 backend https_relay
5715 mode tcp
5716 option httpchk OPTIONS * HTTP/1.1\r\nHost:\ www
5717 server apache1 192.168.1.1:443 check port 80
5718
Simon Hormanafc47ee2013-11-25 10:46:35 +09005719 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
5720 "option pgsql-check", "http-check" and the "check", "port" and
5721 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01005722
5723
Willy Tarreauc27debf2008-01-06 08:57:02 +01005724option httpclose
5725no option httpclose
5726 Enable or disable passive HTTP connection closing
5727 May be used in sections : defaults | frontend | listen | backend
5728 yes | yes | yes | yes
5729 Arguments : none
5730
Willy Tarreau70dffda2014-01-30 03:07:23 +01005731 By default HAProxy operates in keep-alive mode with regards to persistent
5732 connections: for each connection it processes each request and response, and
5733 leaves the connection idle on both sides between the end of a response and
5734 the start of a new request. This mode may be changed by several options such
Cyril Bonté653dcd62014-02-20 00:13:15 +01005735 as "option http-server-close", "option forceclose", "option httpclose" or
Willy Tarreau70dffda2014-01-30 03:07:23 +01005736 "option http-tunnel".
5737
5738 If "option httpclose" is set, HAProxy will work in HTTP tunnel mode and check
5739 if a "Connection: close" header is already set in each direction, and will
5740 add one if missing. Each end should react to this by actively closing the TCP
5741 connection after each transfer, thus resulting in a switch to the HTTP close
5742 mode. Any "Connection" header different from "close" will also be removed.
5743 Note that this option is deprecated since what it does is very cheap but not
5744 reliable. Using "option http-server-close" or "option forceclose" is strongly
5745 recommended instead.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005746
5747 It seldom happens that some servers incorrectly ignore this header and do not
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005748 close the connection even though they reply "Connection: close". For this
Willy Tarreau0dfdf192010-01-05 11:33:11 +01005749 reason, they are not compatible with older HTTP 1.0 browsers. If this happens
5750 it is possible to use the "option forceclose" which actively closes the
5751 request connection once the server responds. Option "forceclose" also
5752 releases the server connection earlier because it does not have to wait for
5753 the client to acknowledge it.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005754
5755 This option may be set both in a frontend and in a backend. It is enabled if
5756 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01005757 It disables and replaces any previous "option http-server-close",
5758 "option forceclose", "option http-keep-alive" or "option http-tunnel". Please
Willy Tarreau70dffda2014-01-30 03:07:23 +01005759 check section 4 ("Proxies") to see how this option combines with others when
5760 frontend and backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005761
5762 If this option has been enabled in a "defaults" section, it can be disabled
5763 in a specific instance by prepending the "no" keyword before it.
5764
Patrick Mezard9ec2ec42010-06-12 17:02:45 +02005765 See also : "option forceclose", "option http-server-close" and
5766 "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01005767
5768
Emeric Brun3a058f32009-06-30 18:26:00 +02005769option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01005770 Enable logging of HTTP request, session state and timers
5771 May be used in sections : defaults | frontend | listen | backend
5772 yes | yes | yes | yes
Emeric Brun3a058f32009-06-30 18:26:00 +02005773 Arguments :
5774 clf if the "clf" argument is added, then the output format will be
5775 the CLF format instead of HAProxy's default HTTP format. You can
5776 use this when you need to feed HAProxy's logs through a specific
5777 log analyser which only support the CLF format and which is not
5778 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005779
5780 By default, the log output format is very poor, as it only contains the
5781 source and destination addresses, and the instance name. By specifying
5782 "option httplog", each log line turns into a much richer format including,
5783 but not limited to, the HTTP request, the connection timers, the session
5784 status, the connections numbers, the captured headers and cookies, the
5785 frontend, backend and server name, and of course the source address and
5786 ports.
5787
5788 This option may be set either in the frontend or the backend.
5789
PiBa-NLbd556bf2014-12-11 21:31:54 +01005790 Specifying only "option httplog" will automatically clear the 'clf' mode
5791 if it was set by default.
Emeric Brun3a058f32009-06-30 18:26:00 +02005792
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005793 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01005794
Willy Tarreau55165fe2009-05-10 12:02:55 +02005795
5796option http_proxy
5797no option http_proxy
5798 Enable or disable plain HTTP proxy mode
5799 May be used in sections : defaults | frontend | listen | backend
5800 yes | yes | yes | yes
5801 Arguments : none
5802
5803 It sometimes happens that people need a pure HTTP proxy which understands
5804 basic proxy requests without caching nor any fancy feature. In this case,
5805 it may be worth setting up an HAProxy instance with the "option http_proxy"
5806 set. In this mode, no server is declared, and the connection is forwarded to
5807 the IP address and port found in the URL after the "http://" scheme.
5808
5809 No host address resolution is performed, so this only works when pure IP
5810 addresses are passed. Since this option's usage perimeter is rather limited,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01005811 it will probably be used only by experts who know they need exactly it. This
5812 is incompatible with the HTTP tunnel mode.
Willy Tarreau55165fe2009-05-10 12:02:55 +02005813
5814 If this option has been enabled in a "defaults" section, it can be disabled
5815 in a specific instance by prepending the "no" keyword before it.
5816
5817 Example :
5818 # this backend understands HTTP proxy requests and forwards them directly.
5819 backend direct_forward
5820 option httpclose
5821 option http_proxy
5822
5823 See also : "option httpclose"
5824
Willy Tarreau211ad242009-10-03 21:45:07 +02005825
Jamie Gloudon801a0a32012-08-25 00:18:33 -04005826option independent-streams
5827no option independent-streams
5828 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02005829 May be used in sections : defaults | frontend | listen | backend
5830 yes | yes | yes | yes
5831 Arguments : none
5832
5833 By default, when data is sent over a socket, both the write timeout and the
5834 read timeout for that socket are refreshed, because we consider that there is
5835 activity on that socket, and we have no other means of guessing if we should
5836 receive data or not.
5837
5838 While this default behaviour is desirable for almost all applications, there
5839 exists a situation where it is desirable to disable it, and only refresh the
5840 read timeout if there are incoming data. This happens on sessions with large
5841 timeouts and low amounts of exchanged data such as telnet session. If the
5842 server suddenly disappears, the output data accumulates in the system's
5843 socket buffers, both timeouts are correctly refreshed, and there is no way
5844 to know the server does not receive them, so we don't timeout. However, when
5845 the underlying protocol always echoes sent data, it would be enough by itself
5846 to detect the issue using the read timeout. Note that this problem does not
5847 happen with more verbose protocols because data won't accumulate long in the
5848 socket buffers.
5849
5850 When this option is set on the frontend, it will disable read timeout updates
5851 on data sent to the client. There probably is little use of this case. When
5852 the option is set on the backend, it will disable read timeout updates on
5853 data sent to the server. Doing so will typically break large HTTP posts from
5854 slow lines, so use it with caution.
5855
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005856 Note: older versions used to call this setting "option independent-streams"
Jamie Gloudon801a0a32012-08-25 00:18:33 -04005857 with a spelling mistake. This spelling is still supported but
5858 deprecated.
5859
Willy Tarreauce887fd2012-05-12 12:50:00 +02005860 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02005861
5862
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02005863option ldap-check
5864 Use LDAPv3 health checks for server testing
5865 May be used in sections : defaults | frontend | listen | backend
5866 yes | no | yes | yes
5867 Arguments : none
5868
5869 It is possible to test that the server correctly talks LDAPv3 instead of just
5870 testing that it accepts the TCP connection. When this option is set, an
5871 LDAPv3 anonymous simple bind message is sent to the server, and the response
5872 is analyzed to find an LDAPv3 bind response message.
5873
5874 The server is considered valid only when the LDAP response contains success
5875 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
5876
5877 Logging of bind requests is server dependent see your documentation how to
5878 configure it.
5879
5880 Example :
5881 option ldap-check
5882
5883 See also : "option httpchk"
5884
5885
Simon Horman98637e52014-06-20 12:30:16 +09005886option external-check
5887 Use external processes for server health checks
5888 May be used in sections : defaults | frontend | listen | backend
5889 yes | no | yes | yes
5890
5891 It is possible to test the health of a server using an external command.
5892 This is achieved by running the executable set using "external-check
5893 command".
5894
5895 Requires the "external-check" global to be set.
5896
5897 See also : "external-check", "external-check command", "external-check path"
5898
5899
Willy Tarreau211ad242009-10-03 21:45:07 +02005900option log-health-checks
5901no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02005902 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02005903 May be used in sections : defaults | frontend | listen | backend
5904 yes | no | yes | yes
5905 Arguments : none
5906
Willy Tarreaubef1b322014-05-13 21:01:39 +02005907 By default, failed health check are logged if server is UP and successful
5908 health checks are logged if server is DOWN, so the amount of additional
5909 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02005910
Willy Tarreaubef1b322014-05-13 21:01:39 +02005911 When this option is enabled, any change of the health check status or to
5912 the server's health will be logged, so that it becomes possible to know
5913 that a server was failing occasional checks before crashing, or exactly when
5914 it failed to respond a valid HTTP status, then when the port started to
5915 reject connections, then when the server stopped responding at all.
5916
5917 Note that status changes not caused by health checks (eg: enable/disable on
5918 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02005919
Willy Tarreaubef1b322014-05-13 21:01:39 +02005920 See also: "option httpchk", "option ldap-check", "option mysql-check",
5921 "option pgsql-check", "option redis-check", "option smtpchk",
5922 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02005923
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005924
5925option log-separate-errors
5926no option log-separate-errors
5927 Change log level for non-completely successful connections
5928 May be used in sections : defaults | frontend | listen | backend
5929 yes | yes | yes | no
5930 Arguments : none
5931
5932 Sometimes looking for errors in logs is not easy. This option makes haproxy
5933 raise the level of logs containing potentially interesting information such
5934 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
5935 level changes from "info" to "err". This makes it possible to log them
5936 separately to a different file with most syslog daemons. Be careful not to
5937 remove them from the original file, otherwise you would lose ordering which
5938 provides very important information.
5939
5940 Using this option, large sites dealing with several thousand connections per
5941 second may log normal traffic to a rotating buffer and only archive smaller
5942 error logs.
5943
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005944 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005945 logging.
5946
Willy Tarreauc27debf2008-01-06 08:57:02 +01005947
5948option logasap
5949no option logasap
5950 Enable or disable early logging of HTTP requests
5951 May be used in sections : defaults | frontend | listen | backend
5952 yes | yes | yes | no
5953 Arguments : none
5954
5955 By default, HTTP requests are logged upon termination so that the total
5956 transfer time and the number of bytes appear in the logs. When large objects
5957 are being transferred, it may take a while before the request appears in the
5958 logs. Using "option logasap", the request gets logged as soon as the server
5959 sends the complete headers. The only missing information in the logs will be
5960 the total number of bytes which will indicate everything except the amount
5961 of data transferred, and the total time which will not take the transfer
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01005962 time into account. In such a situation, it's a good practice to capture the
Willy Tarreauc27debf2008-01-06 08:57:02 +01005963 "Content-Length" response header so that the logs at least indicate how many
5964 bytes are expected to be transferred.
5965
Willy Tarreaucc6c8912009-02-22 10:53:55 +01005966 Examples :
5967 listen http_proxy 0.0.0.0:80
5968 mode http
5969 option httplog
5970 option logasap
5971 log 192.168.2.200 local3
5972
5973 >>> Feb 6 12:14:14 localhost \
5974 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
5975 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
5976 "GET /image.iso HTTP/1.0"
5977
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005978 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01005979 logging.
5980
5981
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02005982option mysql-check [ user <username> [ post-41 ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02005983 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01005984 May be used in sections : defaults | frontend | listen | backend
5985 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02005986 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005987 <username> This is the username which will be used when connecting to MySQL
5988 server.
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02005989 post-41 Send post v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02005990
5991 If you specify a username, the check consists of sending two MySQL packet,
5992 one Client Authentication packet, and one QUIT packet, to correctly close
5993 MySQL session. We then parse the MySQL Handshake Initialisation packet and/or
5994 Error packet. It is a basic but useful test which does not produce error nor
5995 aborted connect on the server. However, it requires adding an authorization
5996 in the MySQL table, like this :
5997
5998 USE mysql;
5999 INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');
6000 FLUSH PRIVILEGES;
6001
6002 If you don't specify a username (it is deprecated and not recommended), the
6003 check only consists in parsing the Mysql Handshake Initialisation packet or
6004 Error packet, we don't send anything in this mode. It was reported that it
6005 can generate lockout if check is too frequent and/or if there is not enough
6006 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
6007 value as if a connection is established successfully within fewer than MySQL
6008 "max_connect_errors" attempts after a previous connection was interrupted,
6009 the error count for the host is cleared to zero. If HAProxy's server get
6010 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
6011
6012 Remember that this does not check database presence nor database consistency.
6013 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006014
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02006015 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006016
6017 Most often, an incoming MySQL server needs to see the client's IP address for
6018 various purposes, including IP privilege matching and connection logging.
6019 When possible, it is often wise to masquerade the client's IP address when
6020 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02006021 which requires the transparent proxy feature to be compiled in, and the MySQL
6022 server to route the client via the machine hosting haproxy.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006023
6024 See also: "option httpchk"
6025
6026
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006027option nolinger
6028no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006029 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006030 May be used in sections: defaults | frontend | listen | backend
6031 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006032 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006033
6034 When clients or servers abort connections in a dirty way (eg: they are
6035 physically disconnected), the session timeouts triggers and the session is
6036 closed. But it will remain in FIN_WAIT1 state for some time in the system,
6037 using some resources and possibly limiting the ability to establish newer
6038 connections.
6039
6040 When this happens, it is possible to activate "option nolinger" which forces
6041 the system to immediately remove any socket's pending data on close. Thus,
6042 the session is instantly purged from the system's tables. This usually has
6043 side effects such as increased number of TCP resets due to old retransmits
6044 getting immediately rejected. Some firewalls may sometimes complain about
6045 this too.
6046
6047 For this reason, it is not recommended to use this option when not absolutely
6048 needed. You know that you need it when you have thousands of FIN_WAIT1
6049 sessions on your system (TIME_WAIT ones do not count).
6050
6051 This option may be used both on frontends and backends, depending on the side
6052 where it is required. Use it on the frontend for clients, and on the backend
6053 for servers.
6054
6055 If this option has been enabled in a "defaults" section, it can be disabled
6056 in a specific instance by prepending the "no" keyword before it.
6057
6058
Willy Tarreau55165fe2009-05-10 12:02:55 +02006059option originalto [ except <network> ] [ header <name> ]
6060 Enable insertion of the X-Original-To header to requests sent to servers
6061 May be used in sections : defaults | frontend | listen | backend
6062 yes | yes | yes | yes
6063 Arguments :
6064 <network> is an optional argument used to disable this option for sources
6065 matching <network>
6066 <name> an optional argument to specify a different "X-Original-To"
6067 header name.
6068
6069 Since HAProxy can work in transparent mode, every request from a client can
6070 be redirected to the proxy and HAProxy itself can proxy every request to a
6071 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
6072 be lost. This is annoying when you want access rules based on destination ip
6073 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
6074 added by HAProxy to all requests sent to the server. This header contains a
6075 value representing the original destination IP address. Since this must be
6076 configured to always use the last occurrence of this header only. Note that
6077 only the last occurrence of the header must be used, since it is really
6078 possible that the client has already brought one.
6079
6080 The keyword "header" may be used to supply a different header name to replace
6081 the default "X-Original-To". This can be useful where you might already
6082 have a "X-Original-To" header from a different application, and you need
6083 preserve it. Also if your backend server doesn't use the "X-Original-To"
6084 header and requires different one.
6085
6086 Sometimes, a same HAProxy instance may be shared between a direct client
6087 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
6088 used to decrypt HTTPS traffic). It is possible to disable the addition of the
6089 header for a known source address or network by adding the "except" keyword
6090 followed by the network address. In this case, any source IP matching the
6091 network will not cause an addition of this header. Most common uses are with
6092 private networks or 127.0.0.1.
6093
6094 This option may be specified either in the frontend or in the backend. If at
6095 least one of them uses it, the header will be added. Note that the backend's
6096 setting of the header subargument takes precedence over the frontend's if
6097 both are defined.
6098
Willy Tarreau55165fe2009-05-10 12:02:55 +02006099 Examples :
6100 # Original Destination address
6101 frontend www
6102 mode http
6103 option originalto except 127.0.0.1
6104
6105 # Those servers want the IP Address in X-Client-Dst
6106 backend www
6107 mode http
6108 option originalto header X-Client-Dst
6109
Willy Tarreau87cf5142011-08-19 22:57:24 +02006110 See also : "option httpclose", "option http-server-close",
6111 "option forceclose"
Willy Tarreau55165fe2009-05-10 12:02:55 +02006112
6113
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006114option persist
6115no option persist
6116 Enable or disable forced persistence on down servers
6117 May be used in sections: defaults | frontend | listen | backend
6118 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006119 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006120
6121 When an HTTP request reaches a backend with a cookie which references a dead
6122 server, by default it is redispatched to another server. It is possible to
6123 force the request to be sent to the dead server first using "option persist"
6124 if absolutely needed. A common use case is when servers are under extreme
6125 load and spend their time flapping. In this case, the users would still be
6126 directed to the server they opened the session on, in the hope they would be
6127 correctly served. It is recommended to use "option redispatch" in conjunction
6128 with this option so that in the event it would not be possible to connect to
6129 the server at all (server definitely dead), the client would finally be
6130 redirected to another valid server.
6131
6132 If this option has been enabled in a "defaults" section, it can be disabled
6133 in a specific instance by prepending the "no" keyword before it.
6134
Willy Tarreau4de91492010-01-22 19:10:05 +01006135 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006136
6137
Willy Tarreau0c122822013-12-15 18:49:01 +01006138option pgsql-check [ user <username> ]
6139 Use PostgreSQL health checks for server testing
6140 May be used in sections : defaults | frontend | listen | backend
6141 yes | no | yes | yes
6142 Arguments :
6143 <username> This is the username which will be used when connecting to
6144 PostgreSQL server.
6145
6146 The check sends a PostgreSQL StartupMessage and waits for either
6147 Authentication request or ErrorResponse message. It is a basic but useful
6148 test which does not produce error nor aborted connect on the server.
6149 This check is identical with the "mysql-check".
6150
6151 See also: "option httpchk"
6152
6153
Willy Tarreau9420b122013-12-15 18:58:25 +01006154option prefer-last-server
6155no option prefer-last-server
6156 Allow multiple load balanced requests to remain on the same server
6157 May be used in sections: defaults | frontend | listen | backend
6158 yes | no | yes | yes
6159 Arguments : none
6160
6161 When the load balancing algorithm in use is not deterministic, and a previous
6162 request was sent to a server to which haproxy still holds a connection, it is
6163 sometimes desirable that subsequent requests on a same session go to the same
6164 server as much as possible. Note that this is different from persistence, as
6165 we only indicate a preference which haproxy tries to apply without any form
6166 of warranty. The real use is for keep-alive connections sent to servers. When
6167 this option is used, haproxy will try to reuse the same connection that is
6168 attached to the server instead of rebalancing to another server, causing a
6169 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01006170 not make much sense to use this in combination with hashing algorithms. Note,
6171 haproxy already automatically tries to stick to a server which sends a 401 or
6172 to a proxy which sends a 407 (authentication required). This is mandatory for
6173 use with the broken NTLM authentication challenge, and significantly helps in
6174 troubleshooting some faulty applications. Option prefer-last-server might be
6175 desirable in these environments as well, to avoid redistributing the traffic
6176 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01006177
6178 If this option has been enabled in a "defaults" section, it can be disabled
6179 in a specific instance by prepending the "no" keyword before it.
6180
6181 See also: "option http-keep-alive"
6182
6183
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006184option redispatch
Joseph Lynch726ab712015-05-11 23:25:34 -07006185option redispatch <interval>
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006186no option redispatch
6187 Enable or disable session redistribution in case of connection failure
6188 May be used in sections: defaults | frontend | listen | backend
6189 yes | no | yes | yes
Joseph Lynch726ab712015-05-11 23:25:34 -07006190 Arguments :
6191 <interval> The optional integer value that controls how often redispatches
6192 occur when retrying connections. Positive value P indicates a
6193 redispatch is desired on every Pth retry, and negative value
6194 N indicate a redispath is desired on the Nth retry prior to the
6195 last retry. For example, the default of -1 preserves the
6196 historical behaviour of redispatching on the last retry, a
6197 positive value of 1 would indicate a redispatch on every retry,
6198 and a positive value of 3 would indicate a redispatch on every
6199 third retry. You can disable redispatches with a value of 0.
6200
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006201
6202 In HTTP mode, if a server designated by a cookie is down, clients may
6203 definitely stick to it because they cannot flush the cookie, so they will not
6204 be able to access the service anymore.
6205
6206 Specifying "option redispatch" will allow the proxy to break their
6207 persistence and redistribute them to a working server.
6208
Joseph Lynch726ab712015-05-11 23:25:34 -07006209 It also allows to retry connections to another server in case of multiple
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006210 connection failures. Of course, it requires having "retries" set to a nonzero
6211 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01006212
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006213 This form is the preferred form, which replaces both the "redispatch" and
6214 "redisp" keywords.
6215
6216 If this option has been enabled in a "defaults" section, it can be disabled
6217 in a specific instance by prepending the "no" keyword before it.
6218
Willy Tarreau4de91492010-01-22 19:10:05 +01006219 See also : "redispatch", "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006220
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006221
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02006222option redis-check
6223 Use redis health checks for server testing
6224 May be used in sections : defaults | frontend | listen | backend
6225 yes | no | yes | yes
6226 Arguments : none
6227
6228 It is possible to test that the server correctly talks REDIS protocol instead
6229 of just testing that it accepts the TCP connection. When this option is set,
6230 a PING redis command is sent to the server, and the response is analyzed to
6231 find the "+PONG" response message.
6232
6233 Example :
6234 option redis-check
6235
6236 See also : "option httpchk"
6237
6238
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006239option smtpchk
6240option smtpchk <hello> <domain>
6241 Use SMTP health checks for server testing
6242 May be used in sections : defaults | frontend | listen | backend
6243 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01006244 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006245 <hello> is an optional argument. It is the "hello" command to use. It can
6246 be either "HELO" (for SMTP) or "EHLO" (for ESTMP). All other
6247 values will be turned into the default command ("HELO").
6248
6249 <domain> is the domain name to present to the server. It may only be
6250 specified (and is mandatory) if the hello command has been
6251 specified. By default, "localhost" is used.
6252
6253 When "option smtpchk" is set, the health checks will consist in TCP
6254 connections followed by an SMTP command. By default, this command is
6255 "HELO localhost". The server's return code is analyzed and only return codes
6256 starting with a "2" will be considered as valid. All other responses,
6257 including a lack of response will constitute an error and will indicate a
6258 dead server.
6259
6260 This test is meant to be used with SMTP servers or relays. Depending on the
6261 request, it is possible that some servers do not log each connection attempt,
6262 so you may want to experiment to improve the behaviour. Using telnet on port
6263 25 is often easier than adjusting the configuration.
6264
6265 Most often, an incoming SMTP server needs to see the client's IP address for
6266 various purposes, including spam filtering, anti-spoofing and logging. When
6267 possible, it is often wise to masquerade the client's IP address when
6268 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02006269 which requires the transparent proxy feature to be compiled in.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006270
6271 Example :
6272 option smtpchk HELO mydomain.org
6273
6274 See also : "option httpchk", "source"
6275
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006276
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02006277option socket-stats
6278no option socket-stats
6279
6280 Enable or disable collecting & providing separate statistics for each socket.
6281 May be used in sections : defaults | frontend | listen | backend
6282 yes | yes | yes | no
6283
6284 Arguments : none
6285
6286
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006287option splice-auto
6288no option splice-auto
6289 Enable or disable automatic kernel acceleration on sockets in both directions
6290 May be used in sections : defaults | frontend | listen | backend
6291 yes | yes | yes | yes
6292 Arguments : none
6293
6294 When this option is enabled either on a frontend or on a backend, haproxy
6295 will automatically evaluate the opportunity to use kernel tcp splicing to
6296 forward data between the client and the server, in either direction. Haproxy
6297 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006298 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006299 are not much aggressive in order to limit excessive use of splicing. This
6300 option requires splicing to be enabled at compile time, and may be globally
6301 disabled with the global option "nosplice". Since splice uses pipes, using it
6302 requires that there are enough spare pipes.
6303
6304 Important note: kernel-based TCP splicing is a Linux-specific feature which
6305 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
6306 transfer data between sockets without copying these data to user-space, thus
6307 providing noticeable performance gains and CPU cycles savings. Since many
6308 early implementations are buggy, corrupt data and/or are inefficient, this
6309 feature is not enabled by default, and it should be used with extreme care.
6310 While it is not possible to detect the correctness of an implementation,
6311 2.6.29 is the first version offering a properly working implementation. In
6312 case of doubt, splicing may be globally disabled using the global "nosplice"
6313 keyword.
6314
6315 Example :
6316 option splice-auto
6317
6318 If this option has been enabled in a "defaults" section, it can be disabled
6319 in a specific instance by prepending the "no" keyword before it.
6320
6321 See also : "option splice-request", "option splice-response", and global
6322 options "nosplice" and "maxpipes"
6323
6324
6325option splice-request
6326no option splice-request
6327 Enable or disable automatic kernel acceleration on sockets for requests
6328 May be used in sections : defaults | frontend | listen | backend
6329 yes | yes | yes | yes
6330 Arguments : none
6331
6332 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04006333 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006334 the client to the server. It might still use the recv/send scheme if there
6335 are no spare pipes left. This option requires splicing to be enabled at
6336 compile time, and may be globally disabled with the global option "nosplice".
6337 Since splice uses pipes, using it requires that there are enough spare pipes.
6338
6339 Important note: see "option splice-auto" for usage limitations.
6340
6341 Example :
6342 option splice-request
6343
6344 If this option has been enabled in a "defaults" section, it can be disabled
6345 in a specific instance by prepending the "no" keyword before it.
6346
6347 See also : "option splice-auto", "option splice-response", and global options
6348 "nosplice" and "maxpipes"
6349
6350
6351option splice-response
6352no option splice-response
6353 Enable or disable automatic kernel acceleration on sockets for responses
6354 May be used in sections : defaults | frontend | listen | backend
6355 yes | yes | yes | yes
6356 Arguments : none
6357
6358 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04006359 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01006360 the server to the client. It might still use the recv/send scheme if there
6361 are no spare pipes left. This option requires splicing to be enabled at
6362 compile time, and may be globally disabled with the global option "nosplice".
6363 Since splice uses pipes, using it requires that there are enough spare pipes.
6364
6365 Important note: see "option splice-auto" for usage limitations.
6366
6367 Example :
6368 option splice-response
6369
6370 If this option has been enabled in a "defaults" section, it can be disabled
6371 in a specific instance by prepending the "no" keyword before it.
6372
6373 See also : "option splice-auto", "option splice-request", and global options
6374 "nosplice" and "maxpipes"
6375
6376
Christopher Fauletba7bc162016-11-07 21:07:38 +01006377option spop-check
6378 Use SPOP health checks for server testing
6379 May be used in sections : defaults | frontend | listen | backend
6380 no | no | no | yes
6381 Arguments : none
6382
6383 It is possible to test that the server correctly talks SPOP protocol instead
6384 of just testing that it accepts the TCP connection. When this option is set,
6385 a HELLO handshake is performed between HAProxy and the server, and the
6386 response is analyzed to check no error is reported.
6387
6388 Example :
6389 option spop-check
6390
6391 See also : "option httpchk"
6392
6393
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006394option srvtcpka
6395no option srvtcpka
6396 Enable or disable the sending of TCP keepalive packets on the server side
6397 May be used in sections : defaults | frontend | listen | backend
6398 yes | no | yes | yes
6399 Arguments : none
6400
6401 When there is a firewall or any session-aware component between a client and
6402 a server, and when the protocol involves very long sessions with long idle
6403 periods (eg: remote desktops), there is a risk that one of the intermediate
6404 components decides to expire a session which has remained idle for too long.
6405
6406 Enabling socket-level TCP keep-alives makes the system regularly send packets
6407 to the other end of the connection, leaving it active. The delay between
6408 keep-alive probes is controlled by the system only and depends both on the
6409 operating system and its tuning parameters.
6410
6411 It is important to understand that keep-alive packets are neither emitted nor
6412 received at the application level. It is only the network stacks which sees
6413 them. For this reason, even if one side of the proxy already uses keep-alives
6414 to maintain its connection alive, those keep-alive packets will not be
6415 forwarded to the other side of the proxy.
6416
6417 Please note that this has nothing to do with HTTP keep-alive.
6418
6419 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
6420 server side of a connection, which should help when session expirations are
6421 noticed between HAProxy and a server.
6422
6423 If this option has been enabled in a "defaults" section, it can be disabled
6424 in a specific instance by prepending the "no" keyword before it.
6425
6426 See also : "option clitcpka", "option tcpka"
6427
6428
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006429option ssl-hello-chk
6430 Use SSLv3 client hello health checks for server testing
6431 May be used in sections : defaults | frontend | listen | backend
6432 yes | no | yes | yes
6433 Arguments : none
6434
6435 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
6436 possible to test that the server correctly talks SSL instead of just testing
6437 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
6438 SSLv3 client hello messages are sent once the connection is established to
6439 the server, and the response is analyzed to find an SSL server hello message.
6440 The server is considered valid only when the response contains this server
6441 hello message.
6442
6443 All servers tested till there correctly reply to SSLv3 client hello messages,
6444 and most servers tested do not even log the requests containing only hello
6445 messages, which is appreciable.
6446
Willy Tarreau763a95b2012-10-04 23:15:39 +02006447 Note that this check works even when SSL support was not built into haproxy
6448 because it forges the SSL message. When SSL support is available, it is best
6449 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006450
Willy Tarreau763a95b2012-10-04 23:15:39 +02006451 See also: "option httpchk", "check-ssl"
6452
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006453
Willy Tarreaued179852013-12-16 01:07:00 +01006454option tcp-check
6455 Perform health checks using tcp-check send/expect sequences
6456 May be used in sections: defaults | frontend | listen | backend
6457 yes | no | yes | yes
6458
6459 This health check method is intended to be combined with "tcp-check" command
6460 lists in order to support send/expect types of health check sequences.
6461
6462 TCP checks currently support 4 modes of operations :
6463 - no "tcp-check" directive : the health check only consists in a connection
6464 attempt, which remains the default mode.
6465
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006466 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01006467 used to send a string along with a connection opening. With some
6468 protocols, it helps sending a "QUIT" message for example that prevents
6469 the server from logging a connection error for each health check. The
6470 check result will still be based on the ability to open the connection
6471 only.
6472
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006473 - "tcp-check expect" only is mentioned : this is used to test a banner.
Willy Tarreaued179852013-12-16 01:07:00 +01006474 The connection is opened and haproxy waits for the server to present some
6475 contents which must validate some rules. The check result will be based
6476 on the matching between the contents and the rules. This is suited for
6477 POP, IMAP, SMTP, FTP, SSH, TELNET.
6478
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006479 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01006480 used to test a hello-type protocol. Haproxy sends a message, the server
6481 responds and its response is analysed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006482 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01006483 suited for protocols which require a binding or a request/response model.
6484 LDAP, MySQL, Redis and SSL are example of such protocols, though they
6485 already all have their dedicated checks with a deeper understanding of
6486 the respective protocols.
6487 In this mode, many questions may be sent and many answers may be
6488 analysed.
6489
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006490 A fifth mode can be used to insert comments in different steps of the
6491 script.
6492
6493 For each tcp-check rule you create, you can add a "comment" directive,
6494 followed by a string. This string will be reported in the log and stderr
6495 in debug mode. It is useful to make user-friendly error reporting.
6496 The "comment" is of course optional.
6497
6498
Willy Tarreaued179852013-12-16 01:07:00 +01006499 Examples :
6500 # perform a POP check (analyse only server's banner)
6501 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006502 tcp-check expect string +OK\ POP3\ ready comment POP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01006503
6504 # perform an IMAP check (analyse only server's banner)
6505 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006506 tcp-check expect string *\ OK\ IMAP4\ ready comment IMAP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01006507
6508 # look for the redis master server after ensuring it speaks well
6509 # redis protocol, then it exits properly.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006510 # (send a command then analyse the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01006511 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006512 tcp-check comment PING\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01006513 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02006514 tcp-check expect string +PONG
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006515 tcp-check comment role\ check
Willy Tarreaued179852013-12-16 01:07:00 +01006516 tcp-check send info\ replication\r\n
6517 tcp-check expect string role:master
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006518 tcp-check comment QUIT\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01006519 tcp-check send QUIT\r\n
6520 tcp-check expect string +OK
6521
6522 forge a HTTP request, then analyse the response
6523 (send many headers before analyzing)
6524 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006525 tcp-check comment forge\ and\ send\ HTTP\ request
Willy Tarreaued179852013-12-16 01:07:00 +01006526 tcp-check send HEAD\ /\ HTTP/1.1\r\n
6527 tcp-check send Host:\ www.mydomain.com\r\n
6528 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
6529 tcp-check send \r\n
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02006530 tcp-check expect rstring HTTP/1\..\ (2..|3..) comment check\ HTTP\ response
Willy Tarreaued179852013-12-16 01:07:00 +01006531
6532
6533 See also : "tcp-check expect", "tcp-check send"
6534
6535
Willy Tarreau9ea05a72009-06-14 12:07:01 +02006536option tcp-smart-accept
6537no option tcp-smart-accept
6538 Enable or disable the saving of one ACK packet during the accept sequence
6539 May be used in sections : defaults | frontend | listen | backend
6540 yes | yes | yes | no
6541 Arguments : none
6542
6543 When an HTTP connection request comes in, the system acknowledges it on
6544 behalf of HAProxy, then the client immediately sends its request, and the
6545 system acknowledges it too while it is notifying HAProxy about the new
6546 connection. HAProxy then reads the request and responds. This means that we
6547 have one TCP ACK sent by the system for nothing, because the request could
6548 very well be acknowledged by HAProxy when it sends its response.
6549
6550 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
6551 sending this useless ACK on platforms which support it (currently at least
6552 Linux). It must not cause any problem, because the system will send it anyway
6553 after 40 ms if the response takes more time than expected to come.
6554
6555 During complex network debugging sessions, it may be desirable to disable
6556 this optimization because delayed ACKs can make troubleshooting more complex
6557 when trying to identify where packets are delayed. It is then possible to
6558 fall back to normal behaviour by specifying "no option tcp-smart-accept".
6559
6560 It is also possible to force it for non-HTTP proxies by simply specifying
6561 "option tcp-smart-accept". For instance, it can make sense with some services
6562 such as SMTP where the server speaks first.
6563
6564 It is recommended to avoid forcing this option in a defaults section. In case
6565 of doubt, consider setting it back to automatic values by prepending the
6566 "default" keyword before it, or disabling it using the "no" keyword.
6567
Willy Tarreaud88edf22009-06-14 15:48:17 +02006568 See also : "option tcp-smart-connect"
6569
6570
6571option tcp-smart-connect
6572no option tcp-smart-connect
6573 Enable or disable the saving of one ACK packet during the connect sequence
6574 May be used in sections : defaults | frontend | listen | backend
6575 yes | no | yes | yes
6576 Arguments : none
6577
6578 On certain systems (at least Linux), HAProxy can ask the kernel not to
6579 immediately send an empty ACK upon a connection request, but to directly
6580 send the buffer request instead. This saves one packet on the network and
6581 thus boosts performance. It can also be useful for some servers, because they
6582 immediately get the request along with the incoming connection.
6583
6584 This feature is enabled when "option tcp-smart-connect" is set in a backend.
6585 It is not enabled by default because it makes network troubleshooting more
6586 complex.
6587
6588 It only makes sense to enable it with protocols where the client speaks first
6589 such as HTTP. In other situations, if there is no data to send in place of
6590 the ACK, a normal ACK is sent.
6591
6592 If this option has been enabled in a "defaults" section, it can be disabled
6593 in a specific instance by prepending the "no" keyword before it.
6594
6595 See also : "option tcp-smart-accept"
6596
Willy Tarreau9ea05a72009-06-14 12:07:01 +02006597
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006598option tcpka
6599 Enable or disable the sending of TCP keepalive packets on both sides
6600 May be used in sections : defaults | frontend | listen | backend
6601 yes | yes | yes | yes
6602 Arguments : none
6603
6604 When there is a firewall or any session-aware component between a client and
6605 a server, and when the protocol involves very long sessions with long idle
6606 periods (eg: remote desktops), there is a risk that one of the intermediate
6607 components decides to expire a session which has remained idle for too long.
6608
6609 Enabling socket-level TCP keep-alives makes the system regularly send packets
6610 to the other end of the connection, leaving it active. The delay between
6611 keep-alive probes is controlled by the system only and depends both on the
6612 operating system and its tuning parameters.
6613
6614 It is important to understand that keep-alive packets are neither emitted nor
6615 received at the application level. It is only the network stacks which sees
6616 them. For this reason, even if one side of the proxy already uses keep-alives
6617 to maintain its connection alive, those keep-alive packets will not be
6618 forwarded to the other side of the proxy.
6619
6620 Please note that this has nothing to do with HTTP keep-alive.
6621
6622 Using option "tcpka" enables the emission of TCP keep-alive probes on both
6623 the client and server sides of a connection. Note that this is meaningful
6624 only in "defaults" or "listen" sections. If this option is used in a
6625 frontend, only the client side will get keep-alives, and if this option is
6626 used in a backend, only the server side will get keep-alives. For this
6627 reason, it is strongly recommended to explicitly use "option clitcpka" and
6628 "option srvtcpka" when the configuration is split between frontends and
6629 backends.
6630
6631 See also : "option clitcpka", "option srvtcpka"
6632
Willy Tarreau844e3c52008-01-11 16:28:18 +01006633
6634option tcplog
6635 Enable advanced logging of TCP connections with session state and timers
6636 May be used in sections : defaults | frontend | listen | backend
6637 yes | yes | yes | yes
6638 Arguments : none
6639
6640 By default, the log output format is very poor, as it only contains the
6641 source and destination addresses, and the instance name. By specifying
6642 "option tcplog", each log line turns into a much richer format including, but
6643 not limited to, the connection timers, the session status, the connections
6644 numbers, the frontend, backend and server name, and of course the source
6645 address and ports. This option is useful for pure TCP proxies in order to
6646 find which of the client or server disconnects or times out. For normal HTTP
6647 proxies, it's better to use "option httplog" which is even more complete.
6648
6649 This option may be set either in the frontend or the backend.
6650
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006651 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01006652
6653
Willy Tarreau844e3c52008-01-11 16:28:18 +01006654option transparent
6655no option transparent
6656 Enable client-side transparent proxying
6657 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01006658 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01006659 Arguments : none
6660
6661 This option was introduced in order to provide layer 7 persistence to layer 3
6662 load balancers. The idea is to use the OS's ability to redirect an incoming
6663 connection for a remote address to a local process (here HAProxy), and let
6664 this process know what address was initially requested. When this option is
6665 used, sessions without cookies will be forwarded to the original destination
6666 IP address of the incoming request (which should match that of another
6667 equipment), while requests with cookies will still be forwarded to the
6668 appropriate server.
6669
6670 Note that contrary to a common belief, this option does NOT make HAProxy
6671 present the client's IP to the server when establishing the connection.
6672
Willy Tarreaua1146052011-03-01 09:51:54 +01006673 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006674 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01006675
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006676
Simon Horman98637e52014-06-20 12:30:16 +09006677external-check command <command>
6678 Executable to run when performing an external-check
6679 May be used in sections : defaults | frontend | listen | backend
6680 yes | no | yes | yes
6681
6682 Arguments :
6683 <command> is the external command to run
6684
Simon Horman98637e52014-06-20 12:30:16 +09006685 The arguments passed to the to the command are:
6686
Cyril Bonté777be862014-12-02 21:21:35 +01006687 <proxy_address> <proxy_port> <server_address> <server_port>
Simon Horman98637e52014-06-20 12:30:16 +09006688
Cyril Bonté777be862014-12-02 21:21:35 +01006689 The <proxy_address> and <proxy_port> are derived from the first listener
6690 that is either IPv4, IPv6 or a UNIX socket. In the case of a UNIX socket
6691 listener the proxy_address will be the path of the socket and the
6692 <proxy_port> will be the string "NOT_USED". In a backend section, it's not
6693 possible to determine a listener, and both <proxy_address> and <proxy_port>
6694 will have the string value "NOT_USED".
Simon Horman98637e52014-06-20 12:30:16 +09006695
Cyril Bonté72cda2a2014-12-27 22:28:39 +01006696 Some values are also provided through environment variables.
6697
6698 Environment variables :
6699 HAPROXY_PROXY_ADDR The first bind address if available (or empty if not
6700 applicable, for example in a "backend" section).
6701
6702 HAPROXY_PROXY_ID The backend id.
6703
6704 HAPROXY_PROXY_NAME The backend name.
6705
6706 HAPROXY_PROXY_PORT The first bind port if available (or empty if not
6707 applicable, for example in a "backend" section or
6708 for a UNIX socket).
6709
6710 HAPROXY_SERVER_ADDR The server address.
6711
6712 HAPROXY_SERVER_CURCONN The current number of connections on the server.
6713
6714 HAPROXY_SERVER_ID The server id.
6715
6716 HAPROXY_SERVER_MAXCONN The server max connections.
6717
6718 HAPROXY_SERVER_NAME The server name.
6719
6720 HAPROXY_SERVER_PORT The server port if available (or empty for a UNIX
6721 socket).
6722
6723 PATH The PATH environment variable used when executing
6724 the command may be set using "external-check path".
6725
Simon Horman98637e52014-06-20 12:30:16 +09006726 If the command executed and exits with a zero status then the check is
6727 considered to have passed, otherwise the check is considered to have
6728 failed.
6729
6730 Example :
6731 external-check command /bin/true
6732
6733 See also : "external-check", "option external-check", "external-check path"
6734
6735
6736external-check path <path>
6737 The value of the PATH environment variable used when running an external-check
6738 May be used in sections : defaults | frontend | listen | backend
6739 yes | no | yes | yes
6740
6741 Arguments :
6742 <path> is the path used when executing external command to run
6743
6744 The default path is "".
6745
6746 Example :
6747 external-check path "/usr/bin:/bin"
6748
6749 See also : "external-check", "option external-check",
6750 "external-check command"
6751
6752
Emeric Brun647caf12009-06-30 17:57:00 +02006753persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02006754persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02006755 Enable RDP cookie-based persistence
6756 May be used in sections : defaults | frontend | listen | backend
6757 yes | no | yes | yes
6758 Arguments :
6759 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02006760 default cookie name "msts" will be used. There currently is no
6761 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02006762
6763 This statement enables persistence based on an RDP cookie. The RDP cookie
6764 contains all information required to find the server in the list of known
6765 servers. So when this option is set in the backend, the request is analysed
6766 and if an RDP cookie is found, it is decoded. If it matches a known server
6767 which is still UP (or if "option persist" is set), then the connection is
6768 forwarded to this server.
6769
6770 Note that this only makes sense in a TCP backend, but for this to work, the
6771 frontend must have waited long enough to ensure that an RDP cookie is present
6772 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006773 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02006774 a single "listen" section.
6775
Willy Tarreau61e28f22010-05-16 22:31:05 +02006776 Also, it is important to understand that the terminal server will emit this
6777 RDP cookie only if it is configured for "token redirection mode", which means
6778 that the "IP address redirection" option is disabled.
6779
Emeric Brun647caf12009-06-30 17:57:00 +02006780 Example :
6781 listen tse-farm
6782 bind :3389
6783 # wait up to 5s for an RDP cookie in the request
6784 tcp-request inspect-delay 5s
6785 tcp-request content accept if RDP_COOKIE
6786 # apply RDP cookie persistence
6787 persist rdp-cookie
6788 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02006789 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02006790 balance rdp-cookie
6791 server srv1 1.1.1.1:3389
6792 server srv2 1.1.1.2:3389
6793
Simon Hormanab814e02011-06-24 14:50:20 +09006794 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
6795 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02006796
6797
Willy Tarreau3a7d2072009-03-05 23:48:25 +01006798rate-limit sessions <rate>
6799 Set a limit on the number of new sessions accepted per second on a frontend
6800 May be used in sections : defaults | frontend | listen | backend
6801 yes | yes | yes | no
6802 Arguments :
6803 <rate> The <rate> parameter is an integer designating the maximum number
6804 of new sessions per second to accept on the frontend.
6805
6806 When the frontend reaches the specified number of new sessions per second, it
6807 stops accepting new connections until the rate drops below the limit again.
6808 During this time, the pending sessions will be kept in the socket's backlog
6809 (in system buffers) and haproxy will not even be aware that sessions are
6810 pending. When applying very low limit on a highly loaded service, it may make
6811 sense to increase the socket's backlog using the "backlog" keyword.
6812
6813 This feature is particularly efficient at blocking connection-based attacks
6814 or service abuse on fragile servers. Since the session rate is measured every
6815 millisecond, it is extremely accurate. Also, the limit applies immediately,
6816 no delay is needed at all to detect the threshold.
6817
6818 Example : limit the connection rate on SMTP to 10 per second max
6819 listen smtp
6820 mode tcp
6821 bind :25
6822 rate-limit sessions 10
Panagiotis Panagiotopoulos7282d8e2016-02-11 16:37:15 +02006823 server smtp1 127.0.0.1:1025
Willy Tarreau3a7d2072009-03-05 23:48:25 +01006824
Willy Tarreaua17c2d92011-07-25 08:16:20 +02006825 Note : when the maximum rate is reached, the frontend's status is not changed
6826 but its sockets appear as "WAITING" in the statistics if the
6827 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01006828
6829 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
6830
6831
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006832redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
6833redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
6834redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02006835 Return an HTTP redirection if/unless a condition is matched
6836 May be used in sections : defaults | frontend | listen | backend
6837 no | yes | yes | yes
6838
6839 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +01006840 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02006841
Willy Tarreau0140f252008-11-19 21:07:09 +01006842 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006843 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01006844 the HTTP "Location" header. When used in an "http-request" rule,
6845 <loc> value follows the log-format rules and can include some
6846 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006847
6848 <pfx> With "redirect prefix", the "Location" header is built from the
6849 concatenation of <pfx> and the complete URI path, including the
6850 query string, unless the "drop-query" option is specified (see
6851 below). As a special case, if <pfx> equals exactly "/", then
6852 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01006853 redirect to the same URL (for instance, to insert a cookie). When
6854 used in an "http-request" rule, <pfx> value follows the log-format
6855 rules and can include some dynamic values (see Custom Log Format
6856 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006857
6858 <sch> With "redirect scheme", then the "Location" header is built by
6859 concatenating <sch> with "://" then the first occurrence of the
6860 "Host" header, and then the URI path, including the query string
6861 unless the "drop-query" option is specified (see below). If no
6862 path is found or if the path is "*", then "/" is used instead. If
6863 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006864 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006865 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01006866 HTTPS. When used in an "http-request" rule, <sch> value follows
6867 the log-format rules and can include some dynamic values (see
6868 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +01006869
6870 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +01006871 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
6872 with 302 used by default if no code is specified. 301 means
6873 "Moved permanently", and a browser may cache the Location. 302
Baptiste Assmannea849c02015-08-03 11:42:50 +02006874 means "Moved temporarily" and means that the browser should not
Willy Tarreaub67fdc42013-03-29 19:28:11 +01006875 cache the redirection. 303 is equivalent to 302 except that the
6876 browser will fetch the location with a GET method. 307 is just
6877 like 302 but makes it clear that the same method must be reused.
6878 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +01006879
6880 <option> There are several options which can be specified to adjust the
6881 expected behaviour of a redirection :
6882
6883 - "drop-query"
6884 When this keyword is used in a prefix-based redirection, then the
6885 location will be set without any possible query-string, which is useful
6886 for directing users to a non-secure page for instance. It has no effect
6887 with a location-type redirect.
6888
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01006889 - "append-slash"
6890 This keyword may be used in conjunction with "drop-query" to redirect
6891 users who use a URL not ending with a '/' to the same one with the '/'.
6892 It can be useful to ensure that search engines will only see one URL.
6893 For this, a return code 301 is preferred.
6894
Willy Tarreau0140f252008-11-19 21:07:09 +01006895 - "set-cookie NAME[=value]"
6896 A "Set-Cookie" header will be added with NAME (and optionally "=value")
6897 to the response. This is sometimes used to indicate that a user has
6898 been seen, for instance to protect against some types of DoS. No other
6899 cookie option is added, so the cookie will be a session cookie. Note
6900 that for a browser, a sole cookie name without an equal sign is
6901 different from a cookie with an equal sign.
6902
6903 - "clear-cookie NAME[=]"
6904 A "Set-Cookie" header will be added with NAME (and optionally "="), but
6905 with the "Max-Age" attribute set to zero. This will tell the browser to
6906 delete this cookie. It is useful for instance on logout pages. It is
6907 important to note that clearing the cookie "NAME" will not remove a
6908 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
6909 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02006910
6911 Example: move the login URL only to HTTPS.
6912 acl clear dst_port 80
6913 acl secure dst_port 8080
6914 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +01006915 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +01006916 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +01006917 acl cookie_set hdr_sub(cookie) SEEN=1
6918
6919 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +01006920 redirect prefix https://mysite.com if login_page !secure
6921 redirect prefix http://mysite.com drop-query if login_page !uid_given
6922 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +01006923 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +02006924
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01006925 Example: send redirects for request for articles without a '/'.
6926 acl missing_slash path_reg ^/article/[^/]*$
6927 redirect code 301 prefix / drop-query append-slash if missing_slash
6928
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006929 Example: redirect all HTTP traffic to HTTPS when SSL is handled by haproxy.
David BERARDe7153042012-11-03 00:11:31 +01006930 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +02006931
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01006932 Example: append 'www.' prefix in front of all hosts not having it
Coen Rosdorff596659b2016-04-11 11:33:49 +02006933 http-request redirect code 301 location \
6934 http://www.%[hdr(host)]%[capture.req.uri] \
6935 unless { hdr_beg(host) -i www }
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01006936
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006937 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02006938
6939
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006940redisp (deprecated)
6941redispatch (deprecated)
6942 Enable or disable session redistribution in case of connection failure
6943 May be used in sections: defaults | frontend | listen | backend
6944 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006945 Arguments : none
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006946
6947 In HTTP mode, if a server designated by a cookie is down, clients may
6948 definitely stick to it because they cannot flush the cookie, so they will not
6949 be able to access the service anymore.
6950
6951 Specifying "redispatch" will allow the proxy to break their persistence and
6952 redistribute them to a working server.
6953
6954 It also allows to retry last connection to another server in case of multiple
6955 connection failures. Of course, it requires having "retries" set to a nonzero
6956 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01006957
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006958 This form is deprecated, do not use it in any new configuration, use the new
6959 "option redispatch" instead.
6960
6961 See also : "option redispatch"
6962
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006963
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01006964reqadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01006965 Add a header at the end of the HTTP request
6966 May be used in sections : defaults | frontend | listen | backend
6967 no | yes | yes | yes
6968 Arguments :
6969 <string> is the complete line to be added. Any space or known delimiter
6970 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006971 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01006972
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01006973 <cond> is an optional matching condition built from ACLs. It makes it
6974 possible to ignore this rule when other conditions are not met.
6975
Willy Tarreau303c0352008-01-17 19:01:39 +01006976 A new line consisting in <string> followed by a line feed will be added after
6977 the last header of an HTTP request.
6978
6979 Header transformations only apply to traffic which passes through HAProxy,
6980 and not to traffic generated by HAProxy, such as health-checks or error
6981 responses.
6982
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01006983 Example : add "X-Proto: SSL" to requests coming via port 81
6984 acl is-ssl dst_port 81
6985 reqadd X-Proto:\ SSL if is-ssl
6986
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08006987 See also: "rspadd", "http-request", section 6 about HTTP header manipulation,
6988 and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01006989
6990
Willy Tarreau5321c422010-01-28 20:35:13 +01006991reqallow <search> [{if | unless} <cond>]
6992reqiallow <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01006993 Definitely allow an HTTP request if a line matches a regular expression
6994 May be used in sections : defaults | frontend | listen | backend
6995 no | yes | yes | yes
6996 Arguments :
6997 <search> is the regular expression applied to HTTP headers and to the
6998 request line. This is an extended regular expression. Parenthesis
6999 grouping is supported and no preliminary backslash is required.
7000 Any space or known delimiter must be escaped using a backslash
7001 ('\'). The pattern applies to a full line at a time. The
7002 "reqallow" keyword strictly matches case while "reqiallow"
7003 ignores case.
7004
Willy Tarreau5321c422010-01-28 20:35:13 +01007005 <cond> is an optional matching condition built from ACLs. It makes it
7006 possible to ignore this rule when other conditions are not met.
7007
Willy Tarreau303c0352008-01-17 19:01:39 +01007008 A request containing any line which matches extended regular expression
7009 <search> will mark the request as allowed, even if any later test would
7010 result in a deny. The test applies both to the request line and to request
7011 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007012 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007013
7014 It is easier, faster and more powerful to use ACLs to write access policies.
7015 Reqdeny, reqallow and reqpass should be avoided in new designs.
7016
7017 Example :
7018 # allow www.* but refuse *.local
7019 reqiallow ^Host:\ www\.
7020 reqideny ^Host:\ .*\.local
7021
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007022 See also: "reqdeny", "block", "http-request", section 6 about HTTP header
7023 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007024
7025
Willy Tarreau5321c422010-01-28 20:35:13 +01007026reqdel <search> [{if | unless} <cond>]
7027reqidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007028 Delete all headers matching a regular expression in an HTTP request
7029 May be used in sections : defaults | frontend | listen | backend
7030 no | yes | yes | yes
7031 Arguments :
7032 <search> is the regular expression applied to HTTP headers and to the
7033 request line. This is an extended regular expression. Parenthesis
7034 grouping is supported and no preliminary backslash is required.
7035 Any space or known delimiter must be escaped using a backslash
7036 ('\'). The pattern applies to a full line at a time. The "reqdel"
7037 keyword strictly matches case while "reqidel" ignores case.
7038
Willy Tarreau5321c422010-01-28 20:35:13 +01007039 <cond> is an optional matching condition built from ACLs. It makes it
7040 possible to ignore this rule when other conditions are not met.
7041
Willy Tarreau303c0352008-01-17 19:01:39 +01007042 Any header line matching extended regular expression <search> in the request
7043 will be completely deleted. Most common use of this is to remove unwanted
7044 and/or dangerous headers or cookies from a request before passing it to the
7045 next servers.
7046
7047 Header transformations only apply to traffic which passes through HAProxy,
7048 and not to traffic generated by HAProxy, such as health-checks or error
7049 responses. Keep in mind that header names are not case-sensitive.
7050
7051 Example :
7052 # remove X-Forwarded-For header and SERVER cookie
7053 reqidel ^X-Forwarded-For:.*
7054 reqidel ^Cookie:.*SERVER=
7055
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007056 See also: "reqadd", "reqrep", "rspdel", "http-request", section 6 about
7057 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007058
7059
Willy Tarreau5321c422010-01-28 20:35:13 +01007060reqdeny <search> [{if | unless} <cond>]
7061reqideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007062 Deny an HTTP request if a line matches a regular expression
7063 May be used in sections : defaults | frontend | listen | backend
7064 no | yes | yes | yes
7065 Arguments :
7066 <search> is the regular expression applied to HTTP headers and to the
7067 request line. This is an extended regular expression. Parenthesis
7068 grouping is supported and no preliminary backslash is required.
7069 Any space or known delimiter must be escaped using a backslash
7070 ('\'). The pattern applies to a full line at a time. The
7071 "reqdeny" keyword strictly matches case while "reqideny" ignores
7072 case.
7073
Willy Tarreau5321c422010-01-28 20:35:13 +01007074 <cond> is an optional matching condition built from ACLs. It makes it
7075 possible to ignore this rule when other conditions are not met.
7076
Willy Tarreau303c0352008-01-17 19:01:39 +01007077 A request containing any line which matches extended regular expression
7078 <search> will mark the request as denied, even if any later test would
7079 result in an allow. The test applies both to the request line and to request
7080 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007081 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007082
Willy Tarreauced27012008-01-17 20:35:34 +01007083 A denied request will generate an "HTTP 403 forbidden" response once the
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007084 complete request has been parsed. This is consistent with what is practiced
Willy Tarreaud72758d2010-01-12 10:42:19 +01007085 using ACLs.
Willy Tarreauced27012008-01-17 20:35:34 +01007086
Willy Tarreau303c0352008-01-17 19:01:39 +01007087 It is easier, faster and more powerful to use ACLs to write access policies.
7088 Reqdeny, reqallow and reqpass should be avoided in new designs.
7089
7090 Example :
7091 # refuse *.local, then allow www.*
7092 reqideny ^Host:\ .*\.local
7093 reqiallow ^Host:\ www\.
7094
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007095 See also: "reqallow", "rspdeny", "block", "http-request", section 6 about
7096 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007097
7098
Willy Tarreau5321c422010-01-28 20:35:13 +01007099reqpass <search> [{if | unless} <cond>]
7100reqipass <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007101 Ignore any HTTP request line matching a regular expression in next rules
7102 May be used in sections : defaults | frontend | listen | backend
7103 no | yes | yes | yes
7104 Arguments :
7105 <search> is the regular expression applied to HTTP headers and to the
7106 request line. This is an extended regular expression. Parenthesis
7107 grouping is supported and no preliminary backslash is required.
7108 Any space or known delimiter must be escaped using a backslash
7109 ('\'). The pattern applies to a full line at a time. The
7110 "reqpass" keyword strictly matches case while "reqipass" ignores
7111 case.
7112
Willy Tarreau5321c422010-01-28 20:35:13 +01007113 <cond> is an optional matching condition built from ACLs. It makes it
7114 possible to ignore this rule when other conditions are not met.
7115
Willy Tarreau303c0352008-01-17 19:01:39 +01007116 A request containing any line which matches extended regular expression
7117 <search> will skip next rules, without assigning any deny or allow verdict.
7118 The test applies both to the request line and to request headers. Keep in
7119 mind that URLs in request line are case-sensitive while header names are not.
7120
7121 It is easier, faster and more powerful to use ACLs to write access policies.
7122 Reqdeny, reqallow and reqpass should be avoided in new designs.
7123
7124 Example :
7125 # refuse *.local, then allow www.*, but ignore "www.private.local"
7126 reqipass ^Host:\ www.private\.local
7127 reqideny ^Host:\ .*\.local
7128 reqiallow ^Host:\ www\.
7129
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007130 See also: "reqallow", "reqdeny", "block", "http-request", section 6 about
7131 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007132
7133
Willy Tarreau5321c422010-01-28 20:35:13 +01007134reqrep <search> <string> [{if | unless} <cond>]
7135reqirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007136 Replace a regular expression with a string in an HTTP request line
7137 May be used in sections : defaults | frontend | listen | backend
7138 no | yes | yes | yes
7139 Arguments :
7140 <search> is the regular expression applied to HTTP headers and to the
7141 request line. This is an extended regular expression. Parenthesis
7142 grouping is supported and no preliminary backslash is required.
7143 Any space or known delimiter must be escaped using a backslash
7144 ('\'). The pattern applies to a full line at a time. The "reqrep"
7145 keyword strictly matches case while "reqirep" ignores case.
7146
7147 <string> is the complete line to be added. Any space or known delimiter
7148 must be escaped using a backslash ('\'). References to matched
7149 pattern groups are possible using the common \N form, with N
7150 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007151 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007152
Willy Tarreau5321c422010-01-28 20:35:13 +01007153 <cond> is an optional matching condition built from ACLs. It makes it
7154 possible to ignore this rule when other conditions are not met.
7155
Willy Tarreau303c0352008-01-17 19:01:39 +01007156 Any line matching extended regular expression <search> in the request (both
7157 the request line and header lines) will be completely replaced with <string>.
7158 Most common use of this is to rewrite URLs or domain names in "Host" headers.
7159
7160 Header transformations only apply to traffic which passes through HAProxy,
7161 and not to traffic generated by HAProxy, such as health-checks or error
7162 responses. Note that for increased readability, it is suggested to add enough
7163 spaces between the request and the response. Keep in mind that URLs in
7164 request line are case-sensitive while header names are not.
7165
7166 Example :
7167 # replace "/static/" with "/" at the beginning of any request path.
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04007168 reqrep ^([^\ :]*)\ /static/(.*) \1\ /\2
Willy Tarreau303c0352008-01-17 19:01:39 +01007169 # replace "www.mydomain.com" with "www" in the host name.
7170 reqirep ^Host:\ www.mydomain.com Host:\ www
7171
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007172 See also: "reqadd", "reqdel", "rsprep", "tune.bufsize", "http-request",
7173 section 6 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007174
7175
Willy Tarreau5321c422010-01-28 20:35:13 +01007176reqtarpit <search> [{if | unless} <cond>]
7177reqitarpit <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007178 Tarpit an HTTP request containing a line matching a regular expression
7179 May be used in sections : defaults | frontend | listen | backend
7180 no | yes | yes | yes
7181 Arguments :
7182 <search> is the regular expression applied to HTTP headers and to the
7183 request line. This is an extended regular expression. Parenthesis
7184 grouping is supported and no preliminary backslash is required.
7185 Any space or known delimiter must be escaped using a backslash
7186 ('\'). The pattern applies to a full line at a time. The
7187 "reqtarpit" keyword strictly matches case while "reqitarpit"
7188 ignores case.
7189
Willy Tarreau5321c422010-01-28 20:35:13 +01007190 <cond> is an optional matching condition built from ACLs. It makes it
7191 possible to ignore this rule when other conditions are not met.
7192
Willy Tarreau303c0352008-01-17 19:01:39 +01007193 A request containing any line which matches extended regular expression
7194 <search> will be tarpitted, which means that it will connect to nowhere, will
Willy Tarreauced27012008-01-17 20:35:34 +01007195 be kept open for a pre-defined time, then will return an HTTP error 500 so
7196 that the attacker does not suspect it has been tarpitted. The status 500 will
7197 be reported in the logs, but the completion flags will indicate "PT". The
Willy Tarreau303c0352008-01-17 19:01:39 +01007198 delay is defined by "timeout tarpit", or "timeout connect" if the former is
7199 not set.
7200
7201 The goal of the tarpit is to slow down robots attacking servers with
7202 identifiable requests. Many robots limit their outgoing number of connections
7203 and stay connected waiting for a reply which can take several minutes to
7204 come. Depending on the environment and attack, it may be particularly
7205 efficient at reducing the load on the network and firewalls.
7206
Willy Tarreau5321c422010-01-28 20:35:13 +01007207 Examples :
Willy Tarreau303c0352008-01-17 19:01:39 +01007208 # ignore user-agents reporting any flavour of "Mozilla" or "MSIE", but
7209 # block all others.
7210 reqipass ^User-Agent:\.*(Mozilla|MSIE)
7211 reqitarpit ^User-Agent:
7212
Willy Tarreau5321c422010-01-28 20:35:13 +01007213 # block bad guys
7214 acl badguys src 10.1.0.3 172.16.13.20/28
7215 reqitarpit . if badguys
7216
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007217 See also: "reqallow", "reqdeny", "reqpass", "http-request", section 6
7218 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007219
7220
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007221retries <value>
7222 Set the number of retries to perform on a server after a connection failure
7223 May be used in sections: defaults | frontend | listen | backend
7224 yes | no | yes | yes
7225 Arguments :
7226 <value> is the number of times a connection attempt should be retried on
7227 a server when a connection either is refused or times out. The
7228 default value is 3.
7229
7230 It is important to understand that this value applies to the number of
7231 connection attempts, not full requests. When a connection has effectively
7232 been established to a server, there will be no more retry.
7233
7234 In order to avoid immediate reconnections to a server which is restarting,
Joseph Lynch726ab712015-05-11 23:25:34 -07007235 a turn-around timer of min("timeout connect", one second) is applied before
7236 a retry occurs.
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007237
7238 When "option redispatch" is set, the last retry may be performed on another
7239 server even if a cookie references a different server.
7240
7241 See also : "option redispatch"
7242
7243
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007244rspadd <string> [{if | unless} <cond>]
Willy Tarreau303c0352008-01-17 19:01:39 +01007245 Add a header at the end of the HTTP response
7246 May be used in sections : defaults | frontend | listen | backend
7247 no | yes | yes | yes
7248 Arguments :
7249 <string> is the complete line to be added. Any space or known delimiter
7250 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007251 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007252
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007253 <cond> is an optional matching condition built from ACLs. It makes it
7254 possible to ignore this rule when other conditions are not met.
7255
Willy Tarreau303c0352008-01-17 19:01:39 +01007256 A new line consisting in <string> followed by a line feed will be added after
7257 the last header of an HTTP response.
7258
7259 Header transformations only apply to traffic which passes through HAProxy,
7260 and not to traffic generated by HAProxy, such as health-checks or error
7261 responses.
7262
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007263 See also: "rspdel" "reqadd", "http-response", section 6 about HTTP header
7264 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007265
7266
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007267rspdel <search> [{if | unless} <cond>]
7268rspidel <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007269 Delete all headers matching a regular expression in an HTTP response
7270 May be used in sections : defaults | frontend | listen | backend
7271 no | yes | yes | yes
7272 Arguments :
7273 <search> is the regular expression applied to HTTP headers and to the
7274 response line. This is an extended regular expression, so
7275 parenthesis grouping is supported and no preliminary backslash
7276 is required. Any space or known delimiter must be escaped using
7277 a backslash ('\'). The pattern applies to a full line at a time.
7278 The "rspdel" keyword strictly matches case while "rspidel"
7279 ignores case.
7280
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007281 <cond> is an optional matching condition built from ACLs. It makes it
7282 possible to ignore this rule when other conditions are not met.
7283
Willy Tarreau303c0352008-01-17 19:01:39 +01007284 Any header line matching extended regular expression <search> in the response
7285 will be completely deleted. Most common use of this is to remove unwanted
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02007286 and/or sensitive headers or cookies from a response before passing it to the
Willy Tarreau303c0352008-01-17 19:01:39 +01007287 client.
7288
7289 Header transformations only apply to traffic which passes through HAProxy,
7290 and not to traffic generated by HAProxy, such as health-checks or error
7291 responses. Keep in mind that header names are not case-sensitive.
7292
7293 Example :
7294 # remove the Server header from responses
Willy Tarreau5e80e022013-05-25 08:31:25 +02007295 rspidel ^Server:.*
Willy Tarreau303c0352008-01-17 19:01:39 +01007296
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007297 See also: "rspadd", "rsprep", "reqdel", "http-response", section 6 about
7298 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007299
7300
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007301rspdeny <search> [{if | unless} <cond>]
7302rspideny <search> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007303 Block an HTTP response if a line matches a regular expression
7304 May be used in sections : defaults | frontend | listen | backend
7305 no | yes | yes | yes
7306 Arguments :
7307 <search> is the regular expression applied to HTTP headers and to the
7308 response line. This is an extended regular expression, so
7309 parenthesis grouping is supported and no preliminary backslash
7310 is required. Any space or known delimiter must be escaped using
7311 a backslash ('\'). The pattern applies to a full line at a time.
7312 The "rspdeny" keyword strictly matches case while "rspideny"
7313 ignores case.
7314
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007315 <cond> is an optional matching condition built from ACLs. It makes it
7316 possible to ignore this rule when other conditions are not met.
7317
Willy Tarreau303c0352008-01-17 19:01:39 +01007318 A response containing any line which matches extended regular expression
7319 <search> will mark the request as denied. The test applies both to the
7320 response line and to response headers. Keep in mind that header names are not
7321 case-sensitive.
7322
7323 Main use of this keyword is to prevent sensitive information leak and to
Willy Tarreauced27012008-01-17 20:35:34 +01007324 block the response before it reaches the client. If a response is denied, it
7325 will be replaced with an HTTP 502 error so that the client never retrieves
7326 any sensitive data.
Willy Tarreau303c0352008-01-17 19:01:39 +01007327
7328 It is easier, faster and more powerful to use ACLs to write access policies.
7329 Rspdeny should be avoided in new designs.
7330
7331 Example :
7332 # Ensure that no content type matching ms-word will leak
7333 rspideny ^Content-type:\.*/ms-word
7334
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007335 See also: "reqdeny", "acl", "block", "http-response", section 6 about
7336 HTTP header manipulation and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007337
7338
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007339rsprep <search> <string> [{if | unless} <cond>]
7340rspirep <search> <string> [{if | unless} <cond>] (ignore case)
Willy Tarreau303c0352008-01-17 19:01:39 +01007341 Replace a regular expression with a string in an HTTP response line
7342 May be used in sections : defaults | frontend | listen | backend
7343 no | yes | yes | yes
7344 Arguments :
7345 <search> is the regular expression applied to HTTP headers and to the
7346 response line. This is an extended regular expression, so
7347 parenthesis grouping is supported and no preliminary backslash
7348 is required. Any space or known delimiter must be escaped using
7349 a backslash ('\'). The pattern applies to a full line at a time.
7350 The "rsprep" keyword strictly matches case while "rspirep"
7351 ignores case.
7352
7353 <string> is the complete line to be added. Any space or known delimiter
7354 must be escaped using a backslash ('\'). References to matched
7355 pattern groups are possible using the common \N form, with N
7356 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007357 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007358
Willy Tarreaufdb563c2010-01-31 15:43:27 +01007359 <cond> is an optional matching condition built from ACLs. It makes it
7360 possible to ignore this rule when other conditions are not met.
7361
Willy Tarreau303c0352008-01-17 19:01:39 +01007362 Any line matching extended regular expression <search> in the response (both
7363 the response line and header lines) will be completely replaced with
7364 <string>. Most common use of this is to rewrite Location headers.
7365
7366 Header transformations only apply to traffic which passes through HAProxy,
7367 and not to traffic generated by HAProxy, such as health-checks or error
7368 responses. Note that for increased readability, it is suggested to add enough
7369 spaces between the request and the response. Keep in mind that header names
7370 are not case-sensitive.
7371
7372 Example :
7373 # replace "Location: 127.0.0.1:8080" with "Location: www.mydomain.com"
7374 rspirep ^Location:\ 127.0.0.1:8080 Location:\ www.mydomain.com
7375
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007376 See also: "rspadd", "rspdel", "reqrep", "http-response", section 6 about
7377 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007378
7379
David du Colombier486df472011-03-17 10:40:26 +01007380server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007381 Declare a server in a backend
7382 May be used in sections : defaults | frontend | listen | backend
7383 no | no | yes | yes
7384 Arguments :
7385 <name> is the internal name assigned to this server. This name will
Cyril Bonté941a0c62012-10-15 19:44:24 +02007386 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -05007387 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007388
David du Colombier486df472011-03-17 10:40:26 +01007389 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
7390 resolvable hostname is supported, but this name will be resolved
7391 during start-up. Address "0.0.0.0" or "*" has a special meaning.
7392 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +02007393 address as the one from the client connection. This is useful in
7394 transparent proxy architectures where the client's connection is
7395 intercepted and haproxy must forward to the original destination
7396 address. This is more or less what the "transparent" keyword does
7397 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +01007398 to report statistics. Optionally, an address family prefix may be
7399 used before the address to force the family regardless of the
7400 address format, which can be useful to specify a path to a unix
7401 socket with no slash ('/'). Currently supported prefixes are :
7402 - 'ipv4@' -> address is always IPv4
7403 - 'ipv6@' -> address is always IPv6
7404 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02007405 - 'abns@' -> address is in abstract namespace (Linux only)
William Lallemandb2f07452015-05-12 14:27:13 +02007406 You may want to reference some environment variables in the
7407 address parameter, see section 2.3 about environment
Willy Tarreau6a031d12016-11-07 19:42:35 +01007408 variables. The "init-addr" setting can be used to modify the way
7409 IP addresses should be resolved upon startup.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007410
Willy Tarreaub6205fd2012-09-24 12:27:33 +02007411 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007412 be sent to this port. If unset, the same port the client
7413 connected to will be used. The port may also be prefixed by a "+"
7414 or a "-". In this case, the server's port will be determined by
7415 adding this value to the client's port.
7416
7417 <param*> is a list of parameters for this server. The "server" keywords
7418 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007419 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007420
7421 Examples :
7422 server first 10.1.1.1:1080 cookie first check inter 1000
7423 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +01007424 server transp ipv4@
William Lallemandb2f07452015-05-12 14:27:13 +02007425 server backup "${SRV_BACKUP}:1080" backup
7426 server www1_dc1 "${LAN_DC1}.101:80"
7427 server www1_dc2 "${LAN_DC2}.101:80"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007428
Willy Tarreau55dcaf62015-09-27 15:03:15 +02007429 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
7430 sun_path length is used for the address length. Some other programs
7431 such as socat use the string length only by default. Pass the option
7432 ",unix-tightsocklen=0" to any abstract socket definition in socat to
7433 make it compatible with HAProxy's.
7434
Mark Lamourinec2247f02012-01-04 13:02:01 -05007435 See also: "default-server", "http-send-name-header" and section 5 about
7436 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007437
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007438server-state-file-name [<file>]
7439 Set the server state file to read, load and apply to servers available in
7440 this backend. It only applies when the directive "load-server-state-from-file"
7441 is set to "local". When <file> is not provided or if this directive is not
7442 set, then backend name is used. If <file> starts with a slash '/', then it is
7443 considered as an absolute path. Otherwise, <file> is concatenated to the
7444 global directive "server-state-file-base".
7445
7446 Example: the minimal configuration below would make HAProxy look for the
7447 state server file '/etc/haproxy/states/bk':
7448
7449 global
7450 server-state-file-base /etc/haproxy/states
7451
7452 backend bk
7453 load-server-state-from-file
7454
7455 See also: "server-state-file-base", "load-server-state-from-file", and
7456 "show servers state"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007457
7458source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02007459source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +01007460source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007461 Set the source address for outgoing connections
7462 May be used in sections : defaults | frontend | listen | backend
7463 yes | no | yes | yes
7464 Arguments :
7465 <addr> is the IPv4 address HAProxy will bind to before connecting to a
7466 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +01007467
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007468 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +01007469 the most appropriate address to reach its destination. Optionally
7470 an address family prefix may be used before the address to force
7471 the family regardless of the address format, which can be useful
7472 to specify a path to a unix socket with no slash ('/'). Currently
7473 supported prefixes are :
7474 - 'ipv4@' -> address is always IPv4
7475 - 'ipv6@' -> address is always IPv6
7476 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02007477 - 'abns@' -> address is in abstract namespace (Linux only)
Cyril Bonté307ee1e2015-09-28 23:16:06 +02007478 You may want to reference some environment variables in the
7479 address parameter, see section 2.3 about environment variables.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007480
7481 <port> is an optional port. It is normally not needed but may be useful
7482 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02007483 the system will select a free port. Note that port ranges are not
7484 supported in the backend. If you want to force port ranges, you
7485 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007486
7487 <addr2> is the IP address to present to the server when connections are
7488 forwarded in full transparent proxy mode. This is currently only
7489 supported on some patched Linux kernels. When this address is
7490 specified, clients connecting to the server will be presented
7491 with this address, while health checks will still use the address
7492 <addr>.
7493
7494 <port2> is the optional port to present to the server when connections
7495 are forwarded in full transparent proxy mode (see <addr2> above).
7496 The default value of zero means the system will select a free
7497 port.
7498
Willy Tarreaubce70882009-09-07 11:51:47 +02007499 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
7500 This is the name of a comma-separated header list which can
7501 contain multiple IP addresses. By default, the last occurrence is
7502 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +01007503 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +02007504 by previous proxy, typically Stunnel. In order to use another
7505 occurrence from the last one, please see the <occ> parameter
7506 below. When the header (or occurrence) is not found, no binding
7507 is performed so that the proxy's default IP address is used. Also
7508 keep in mind that the header name is case insensitive, as for any
7509 HTTP header.
7510
7511 <occ> is the occurrence number of a value to be used in a multi-value
7512 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007513 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +02007514 address. Positive values indicate a position from the first
7515 occurrence, 1 being the first one. Negative values indicate
7516 positions relative to the last one, -1 being the last one. This
7517 is helpful for situations where an X-Forwarded-For header is set
7518 at the entry point of an infrastructure and must be used several
7519 proxy layers away. When this value is not specified, -1 is
7520 assumed. Passing a zero here disables the feature.
7521
Willy Tarreaud53f96b2009-02-04 18:46:54 +01007522 <name> is an optional interface name to which to bind to for outgoing
7523 traffic. On systems supporting this features (currently, only
7524 Linux), this allows one to bind all traffic to the server to
7525 this interface even if it is not the one the system would select
7526 based on routing tables. This should be used with extreme care.
7527 Note that using this option requires root privileges.
7528
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007529 The "source" keyword is useful in complex environments where a specific
7530 address only is allowed to connect to the servers. It may be needed when a
7531 private address must be used through a public gateway for instance, and it is
7532 known that the system cannot determine the adequate source address by itself.
7533
7534 An extension which is available on certain patched Linux kernels may be used
7535 through the "usesrc" optional keyword. It makes it possible to connect to the
7536 servers with an IP address which does not belong to the system itself. This
7537 is called "full transparent proxy mode". For this to work, the destination
7538 servers have to route their traffic back to this address through the machine
7539 running HAProxy, and IP forwarding must generally be enabled on this machine.
7540
7541 In this "full transparent proxy" mode, it is possible to force a specific IP
7542 address to be presented to the servers. This is not much used in fact. A more
7543 common use is to tell HAProxy to present the client's IP address. For this,
7544 there are two methods :
7545
7546 - present the client's IP and port addresses. This is the most transparent
7547 mode, but it can cause problems when IP connection tracking is enabled on
7548 the machine, because a same connection may be seen twice with different
7549 states. However, this solution presents the huge advantage of not
7550 limiting the system to the 64k outgoing address+port couples, because all
7551 of the client ranges may be used.
7552
7553 - present only the client's IP address and select a spare port. This
7554 solution is still quite elegant but slightly less transparent (downstream
7555 firewalls logs will not match upstream's). It also presents the downside
7556 of limiting the number of concurrent connections to the usual 64k ports.
7557 However, since the upstream and downstream ports are different, local IP
7558 connection tracking on the machine will not be upset by the reuse of the
7559 same session.
7560
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007561 This option sets the default source for all servers in the backend. It may
7562 also be specified in a "defaults" section. Finer source address specification
7563 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007564 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007565
Baptiste Assmann91bd3372015-07-17 21:59:42 +02007566 In order to work, "usesrc" requires root privileges.
7567
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007568 Examples :
7569 backend private
7570 # Connect to the servers using our 192.168.1.200 source address
7571 source 192.168.1.200
7572
7573 backend transparent_ssl1
7574 # Connect to the SSL farm from the client's source address
7575 source 192.168.1.200 usesrc clientip
7576
7577 backend transparent_ssl2
7578 # Connect to the SSL farm from the client's source address and port
7579 # not recommended if IP conntrack is present on the local machine.
7580 source 192.168.1.200 usesrc client
7581
7582 backend transparent_ssl3
7583 # Connect to the SSL farm from the client's source address. It
7584 # is more conntrack-friendly.
7585 source 192.168.1.200 usesrc clientip
7586
7587 backend transparent_smtp
7588 # Connect to the SMTP farm from the client's source address/port
7589 # with Tproxy version 4.
7590 source 0.0.0.0 usesrc clientip
7591
Willy Tarreaubce70882009-09-07 11:51:47 +02007592 backend transparent_http
7593 # Connect to the servers using the client's IP as seen by previous
7594 # proxy.
7595 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
7596
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007597 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007598 the Linux kernel on www.balabit.com, the "bind" keyword.
7599
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007600
Willy Tarreau844e3c52008-01-11 16:28:18 +01007601srvtimeout <timeout> (deprecated)
7602 Set the maximum inactivity time on the server side.
7603 May be used in sections : defaults | frontend | listen | backend
7604 yes | no | yes | yes
7605 Arguments :
7606 <timeout> is the timeout value specified in milliseconds by default, but
7607 can be in any other unit if the number is suffixed by the unit,
7608 as explained at the top of this document.
7609
7610 The inactivity timeout applies when the server is expected to acknowledge or
7611 send data. In HTTP mode, this timeout is particularly important to consider
7612 during the first phase of the server's response, when it has to send the
7613 headers, as it directly represents the server's processing time for the
7614 request. To find out what value to put there, it's often good to start with
7615 what would be considered as unacceptable response times, then check the logs
7616 to observe the response time distribution, and adjust the value accordingly.
7617
7618 The value is specified in milliseconds by default, but can be in any other
7619 unit if the number is suffixed by the unit, as specified at the top of this
7620 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
7621 recommended that the client timeout remains equal to the server timeout in
7622 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007623 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +01007624 packet losses by specifying timeouts that are slightly above multiples of 3
Willy Tarreaud72758d2010-01-12 10:42:19 +01007625 seconds (eg: 4 or 5 seconds minimum).
Willy Tarreau844e3c52008-01-11 16:28:18 +01007626
7627 This parameter is specific to backends, but can be specified once for all in
7628 "defaults" sections. This is in fact one of the easiest solutions not to
7629 forget about it. An unspecified timeout results in an infinite timeout, which
7630 is not recommended. Such a usage is accepted and works but reports a warning
7631 during startup because it may results in accumulation of expired sessions in
7632 the system if the system's timeouts are not configured either.
7633
7634 This parameter is provided for compatibility but is currently deprecated.
7635 Please use "timeout server" instead.
7636
Willy Tarreauce887fd2012-05-12 12:50:00 +02007637 See also : "timeout server", "timeout tunnel", "timeout client" and
7638 "clitimeout".
Willy Tarreau844e3c52008-01-11 16:28:18 +01007639
7640
Cyril Bonté66c327d2010-10-12 00:14:37 +02007641stats admin { if | unless } <cond>
7642 Enable statistics admin level if/unless a condition is matched
7643 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007644 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +02007645
7646 This statement enables the statistics admin level if/unless a condition is
7647 matched.
7648
7649 The admin level allows to enable/disable servers from the web interface. By
7650 default, statistics page is read-only for security reasons.
7651
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01007652 Note : Consider not using this feature in multi-process mode (nbproc > 1)
7653 unless you know what you do : memory is not shared between the
7654 processes, which can result in random behaviours.
7655
Cyril Bonté23b39d92011-02-10 22:54:44 +01007656 Currently, the POST request is limited to the buffer size minus the reserved
7657 buffer space, which means that if the list of servers is too long, the
7658 request won't be processed. It is recommended to alter few servers at a
7659 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +02007660
7661 Example :
7662 # statistics admin level only for localhost
7663 backend stats_localhost
7664 stats enable
7665 stats admin if LOCALHOST
7666
7667 Example :
7668 # statistics admin level always enabled because of the authentication
7669 backend stats_auth
7670 stats enable
7671 stats auth admin:AdMiN123
7672 stats admin if TRUE
7673
7674 Example :
7675 # statistics admin level depends on the authenticated user
7676 userlist stats-auth
7677 group admin users admin
7678 user admin insecure-password AdMiN123
7679 group readonly users haproxy
7680 user haproxy insecure-password haproxy
7681
7682 backend stats_auth
7683 stats enable
7684 acl AUTH http_auth(stats-auth)
7685 acl AUTH_ADMIN http_auth_group(stats-auth) admin
7686 stats http-request auth unless AUTH
7687 stats admin if AUTH_ADMIN
7688
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01007689 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
7690 "bind-process", section 3.4 about userlists and section 7 about
7691 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +02007692
7693
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007694stats auth <user>:<passwd>
7695 Enable statistics with authentication and grant access to an account
7696 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007697 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007698 Arguments :
7699 <user> is a user name to grant access to
7700
7701 <passwd> is the cleartext password associated to this user
7702
7703 This statement enables statistics with default settings, and restricts access
7704 to declared users only. It may be repeated as many times as necessary to
7705 allow as many users as desired. When a user tries to access the statistics
7706 without a valid account, a "401 Forbidden" response will be returned so that
7707 the browser asks the user to provide a valid user and password. The real
7708 which will be returned to the browser is configurable using "stats realm".
7709
7710 Since the authentication method is HTTP Basic Authentication, the passwords
7711 circulate in cleartext on the network. Thus, it was decided that the
7712 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02007713 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007714
7715 It is also possible to reduce the scope of the proxies which appear in the
7716 report using "stats scope".
7717
7718 Though this statement alone is enough to enable statistics reporting, it is
7719 recommended to set all other settings in order to avoid relying on default
7720 unobvious parameters.
7721
7722 Example :
7723 # public access (limited to this backend only)
7724 backend public_www
7725 server srv1 192.168.0.1:80
7726 stats enable
7727 stats hide-version
7728 stats scope .
7729 stats uri /admin?stats
7730 stats realm Haproxy\ Statistics
7731 stats auth admin1:AdMiN123
7732 stats auth admin2:AdMiN321
7733
7734 # internal monitoring access (unlimited)
7735 backend private_monitoring
7736 stats enable
7737 stats uri /admin?stats
7738 stats refresh 5s
7739
7740 See also : "stats enable", "stats realm", "stats scope", "stats uri"
7741
7742
7743stats enable
7744 Enable statistics reporting with default settings
7745 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007746 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007747 Arguments : none
7748
7749 This statement enables statistics reporting with default settings defined
7750 at build time. Unless stated otherwise, these settings are used :
7751 - stats uri : /haproxy?stats
7752 - stats realm : "HAProxy Statistics"
7753 - stats auth : no authentication
7754 - stats scope : no restriction
7755
7756 Though this statement alone is enough to enable statistics reporting, it is
7757 recommended to set all other settings in order to avoid relying on default
7758 unobvious parameters.
7759
7760 Example :
7761 # public access (limited to this backend only)
7762 backend public_www
7763 server srv1 192.168.0.1:80
7764 stats enable
7765 stats hide-version
7766 stats scope .
7767 stats uri /admin?stats
7768 stats realm Haproxy\ Statistics
7769 stats auth admin1:AdMiN123
7770 stats auth admin2:AdMiN321
7771
7772 # internal monitoring access (unlimited)
7773 backend private_monitoring
7774 stats enable
7775 stats uri /admin?stats
7776 stats refresh 5s
7777
7778 See also : "stats auth", "stats realm", "stats uri"
7779
7780
Willy Tarreaud63335a2010-02-26 12:56:52 +01007781stats hide-version
7782 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007783 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007784 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01007785 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007786
Willy Tarreaud63335a2010-02-26 12:56:52 +01007787 By default, the stats page reports some useful status information along with
7788 the statistics. Among them is HAProxy's version. However, it is generally
7789 considered dangerous to report precise version to anyone, as it can help them
7790 target known weaknesses with specific attacks. The "stats hide-version"
7791 statement removes the version from the statistics report. This is recommended
7792 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007793
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02007794 Though this statement alone is enough to enable statistics reporting, it is
7795 recommended to set all other settings in order to avoid relying on default
7796 unobvious parameters.
7797
Willy Tarreaud63335a2010-02-26 12:56:52 +01007798 Example :
7799 # public access (limited to this backend only)
7800 backend public_www
7801 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02007802 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01007803 stats hide-version
7804 stats scope .
7805 stats uri /admin?stats
7806 stats realm Haproxy\ Statistics
7807 stats auth admin1:AdMiN123
7808 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007809
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007810 # internal monitoring access (unlimited)
7811 backend private_monitoring
7812 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01007813 stats uri /admin?stats
7814 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +01007815
Willy Tarreaud63335a2010-02-26 12:56:52 +01007816 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02007817
Willy Tarreau983e01e2010-01-11 18:42:06 +01007818
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02007819stats http-request { allow | deny | auth [realm <realm>] }
7820 [ { if | unless } <condition> ]
7821 Access control for statistics
7822
7823 May be used in sections: defaults | frontend | listen | backend
7824 no | no | yes | yes
7825
7826 As "http-request", these set of options allow to fine control access to
7827 statistics. Each option may be followed by if/unless and acl.
7828 First option with matched condition (or option without condition) is final.
7829 For "deny" a 403 error will be returned, for "allow" normal processing is
7830 performed, for "auth" a 401/407 error code is returned so the client
7831 should be asked to enter a username and password.
7832
7833 There is no fixed limit to the number of http-request statements per
7834 instance.
7835
7836 See also : "http-request", section 3.4 about userlists and section 7
7837 about ACL usage.
7838
7839
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007840stats realm <realm>
7841 Enable statistics and set authentication realm
7842 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007843 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007844 Arguments :
7845 <realm> is the name of the HTTP Basic Authentication realm reported to
7846 the browser. The browser uses it to display it in the pop-up
7847 inviting the user to enter a valid username and password.
7848
7849 The realm is read as a single word, so any spaces in it should be escaped
7850 using a backslash ('\').
7851
7852 This statement is useful only in conjunction with "stats auth" since it is
7853 only related to authentication.
7854
7855 Though this statement alone is enough to enable statistics reporting, it is
7856 recommended to set all other settings in order to avoid relying on default
7857 unobvious parameters.
7858
7859 Example :
7860 # public access (limited to this backend only)
7861 backend public_www
7862 server srv1 192.168.0.1:80
7863 stats enable
7864 stats hide-version
7865 stats scope .
7866 stats uri /admin?stats
7867 stats realm Haproxy\ Statistics
7868 stats auth admin1:AdMiN123
7869 stats auth admin2:AdMiN321
7870
7871 # internal monitoring access (unlimited)
7872 backend private_monitoring
7873 stats enable
7874 stats uri /admin?stats
7875 stats refresh 5s
7876
7877 See also : "stats auth", "stats enable", "stats uri"
7878
7879
7880stats refresh <delay>
7881 Enable statistics with automatic refresh
7882 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007883 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007884 Arguments :
7885 <delay> is the suggested refresh delay, specified in seconds, which will
7886 be returned to the browser consulting the report page. While the
7887 browser is free to apply any delay, it will generally respect it
7888 and refresh the page this every seconds. The refresh interval may
7889 be specified in any other non-default time unit, by suffixing the
7890 unit after the value, as explained at the top of this document.
7891
7892 This statement is useful on monitoring displays with a permanent page
7893 reporting the load balancer's activity. When set, the HTML report page will
7894 include a link "refresh"/"stop refresh" so that the user can select whether
7895 he wants automatic refresh of the page or not.
7896
7897 Though this statement alone is enough to enable statistics reporting, it is
7898 recommended to set all other settings in order to avoid relying on default
7899 unobvious parameters.
7900
7901 Example :
7902 # public access (limited to this backend only)
7903 backend public_www
7904 server srv1 192.168.0.1:80
7905 stats enable
7906 stats hide-version
7907 stats scope .
7908 stats uri /admin?stats
7909 stats realm Haproxy\ Statistics
7910 stats auth admin1:AdMiN123
7911 stats auth admin2:AdMiN321
7912
7913 # internal monitoring access (unlimited)
7914 backend private_monitoring
7915 stats enable
7916 stats uri /admin?stats
7917 stats refresh 5s
7918
7919 See also : "stats auth", "stats enable", "stats realm", "stats uri"
7920
7921
7922stats scope { <name> | "." }
7923 Enable statistics and limit access scope
7924 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007925 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007926 Arguments :
7927 <name> is the name of a listen, frontend or backend section to be
7928 reported. The special name "." (a single dot) designates the
7929 section in which the statement appears.
7930
7931 When this statement is specified, only the sections enumerated with this
7932 statement will appear in the report. All other ones will be hidden. This
7933 statement may appear as many times as needed if multiple sections need to be
7934 reported. Please note that the name checking is performed as simple string
7935 comparisons, and that it is never checked that a give section name really
7936 exists.
7937
7938 Though this statement alone is enough to enable statistics reporting, it is
7939 recommended to set all other settings in order to avoid relying on default
7940 unobvious parameters.
7941
7942 Example :
7943 # public access (limited to this backend only)
7944 backend public_www
7945 server srv1 192.168.0.1:80
7946 stats enable
7947 stats hide-version
7948 stats scope .
7949 stats uri /admin?stats
7950 stats realm Haproxy\ Statistics
7951 stats auth admin1:AdMiN123
7952 stats auth admin2:AdMiN321
7953
7954 # internal monitoring access (unlimited)
7955 backend private_monitoring
7956 stats enable
7957 stats uri /admin?stats
7958 stats refresh 5s
7959
7960 See also : "stats auth", "stats enable", "stats realm", "stats uri"
7961
Willy Tarreaud63335a2010-02-26 12:56:52 +01007962
Willy Tarreauc9705a12010-07-27 20:05:50 +02007963stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +01007964 Enable reporting of a description on the statistics page.
7965 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02007966 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01007967
Willy Tarreauc9705a12010-07-27 20:05:50 +02007968 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +01007969 description from global section is automatically used instead.
7970
7971 This statement is useful for users that offer shared services to their
7972 customers, where node or description should be different for each customer.
7973
7974 Though this statement alone is enough to enable statistics reporting, it is
7975 recommended to set all other settings in order to avoid relying on default
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04007976 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +01007977
7978 Example :
7979 # internal monitoring access (unlimited)
7980 backend private_monitoring
7981 stats enable
7982 stats show-desc Master node for Europe, Asia, Africa
7983 stats uri /admin?stats
7984 stats refresh 5s
7985
7986 See also: "show-node", "stats enable", "stats uri" and "description" in
7987 global section.
7988
7989
7990stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +02007991 Enable reporting additional information on the statistics page
7992 May be used in sections : defaults | frontend | listen | backend
7993 yes | yes | yes | yes
7994 Arguments : none
7995
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007996 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +01007997 - cap: capabilities (proxy)
7998 - mode: one of tcp, http or health (proxy)
7999 - id: SNMP ID (proxy, socket, server)
8000 - IP (socket, server)
8001 - cookie (backend, server)
8002
8003 Though this statement alone is enough to enable statistics reporting, it is
8004 recommended to set all other settings in order to avoid relying on default
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04008005 unobvious parameters. Default behaviour is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008006
8007 See also: "stats enable", "stats uri".
8008
8009
8010stats show-node [ <name> ]
8011 Enable reporting of a host name on the statistics page.
8012 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008013 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008014 Arguments:
8015 <name> is an optional name to be reported. If unspecified, the
8016 node name from global section is automatically used instead.
8017
8018 This statement is useful for users that offer shared services to their
8019 customers, where node or description might be different on a stats page
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04008020 provided for each customer. Default behaviour is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008021
8022 Though this statement alone is enough to enable statistics reporting, it is
8023 recommended to set all other settings in order to avoid relying on default
8024 unobvious parameters.
8025
8026 Example:
8027 # internal monitoring access (unlimited)
8028 backend private_monitoring
8029 stats enable
8030 stats show-node Europe-1
8031 stats uri /admin?stats
8032 stats refresh 5s
8033
8034 See also: "show-desc", "stats enable", "stats uri", and "node" in global
8035 section.
8036
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008037
8038stats uri <prefix>
8039 Enable statistics and define the URI prefix to access them
8040 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008041 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008042 Arguments :
8043 <prefix> is the prefix of any URI which will be redirected to stats. This
8044 prefix may contain a question mark ('?') to indicate part of a
8045 query string.
8046
8047 The statistics URI is intercepted on the relayed traffic, so it appears as a
8048 page within the normal application. It is strongly advised to ensure that the
8049 selected URI will never appear in the application, otherwise it will never be
8050 possible to reach it in the application.
8051
8052 The default URI compiled in haproxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008053 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008054 It is generally a good idea to include a question mark in the URI so that
8055 intermediate proxies refrain from caching the results. Also, since any string
8056 beginning with the prefix will be accepted as a stats request, the question
8057 mark helps ensuring that no valid URI will begin with the same words.
8058
8059 It is sometimes very convenient to use "/" as the URI prefix, and put that
8060 statement in a "listen" instance of its own. That makes it easy to dedicate
8061 an address or a port to statistics only.
8062
8063 Though this statement alone is enough to enable statistics reporting, it is
8064 recommended to set all other settings in order to avoid relying on default
8065 unobvious parameters.
8066
8067 Example :
8068 # public access (limited to this backend only)
8069 backend public_www
8070 server srv1 192.168.0.1:80
8071 stats enable
8072 stats hide-version
8073 stats scope .
8074 stats uri /admin?stats
8075 stats realm Haproxy\ Statistics
8076 stats auth admin1:AdMiN123
8077 stats auth admin2:AdMiN321
8078
8079 # internal monitoring access (unlimited)
8080 backend private_monitoring
8081 stats enable
8082 stats uri /admin?stats
8083 stats refresh 5s
8084
8085 See also : "stats auth", "stats enable", "stats realm"
8086
8087
Willy Tarreaud63335a2010-02-26 12:56:52 +01008088stick match <pattern> [table <table>] [{if | unless} <cond>]
8089 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008090 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +01008091 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008092
8093 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008094 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008095 describes what elements of the incoming request or connection
8096 will be analysed in the hope to find a matching entry in a
8097 stickiness table. This rule is mandatory.
8098
8099 <table> is an optional stickiness table name. If unspecified, the same
8100 backend's table is used. A stickiness table is declared using
8101 the "stick-table" statement.
8102
8103 <cond> is an optional matching condition. It makes it possible to match
8104 on a certain criterion only when other conditions are met (or
8105 not met). For instance, it could be used to match on a source IP
8106 address except when a request passes through a known proxy, in
8107 which case we'd match on a header containing that IP address.
8108
8109 Some protocols or applications require complex stickiness rules and cannot
8110 always simply rely on cookies nor hashing. The "stick match" statement
8111 describes a rule to extract the stickiness criterion from an incoming request
8112 or connection. See section 7 for a complete list of possible patterns and
8113 transformation rules.
8114
8115 The table has to be declared using the "stick-table" statement. It must be of
8116 a type compatible with the pattern. By default it is the one which is present
8117 in the same backend. It is possible to share a table with other backends by
8118 referencing it using the "table" keyword. If another table is referenced,
8119 the server's ID inside the backends are used. By default, all server IDs
8120 start at 1 in each backend, so the server ordering is enough. But in case of
8121 doubt, it is highly recommended to force server IDs using their "id" setting.
8122
8123 It is possible to restrict the conditions where a "stick match" statement
8124 will apply, using "if" or "unless" followed by a condition. See section 7 for
8125 ACL based conditions.
8126
8127 There is no limit on the number of "stick match" statements. The first that
8128 applies and matches will cause the request to be directed to the same server
8129 as was used for the request which created the entry. That way, multiple
8130 matches can be used as fallbacks.
8131
8132 The stick rules are checked after the persistence cookies, so they will not
8133 affect stickiness if a cookie has already been used to select a server. That
8134 way, it becomes very easy to insert cookies and match on IP addresses in
8135 order to maintain stickiness between HTTP and HTTPS.
8136
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008137 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8138 unless you know what you do : memory is not shared between the
8139 processes, which can result in random behaviours.
8140
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008141 Example :
8142 # forward SMTP users to the same server they just used for POP in the
8143 # last 30 minutes
8144 backend pop
8145 mode tcp
8146 balance roundrobin
8147 stick store-request src
8148 stick-table type ip size 200k expire 30m
8149 server s1 192.168.1.1:110
8150 server s2 192.168.1.1:110
8151
8152 backend smtp
8153 mode tcp
8154 balance roundrobin
8155 stick match src table pop
8156 server s1 192.168.1.1:25
8157 server s2 192.168.1.1:25
8158
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008159 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02008160 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008161
8162
8163stick on <pattern> [table <table>] [{if | unless} <condition>]
8164 Define a request pattern to associate a user to a server
8165 May be used in sections : defaults | frontend | listen | backend
8166 no | no | yes | yes
8167
8168 Note : This form is exactly equivalent to "stick match" followed by
8169 "stick store-request", all with the same arguments. Please refer
8170 to both keywords for details. It is only provided as a convenience
8171 for writing more maintainable configurations.
8172
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008173 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8174 unless you know what you do : memory is not shared between the
8175 processes, which can result in random behaviours.
8176
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008177 Examples :
8178 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +01008179 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008180
8181 # ...is strictly equivalent to this one :
8182 stick match src table pop if !localhost
8183 stick store-request src table pop if !localhost
8184
8185
8186 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
8187 # well as HTTP without cookie. Share the same table between both accesses.
8188 backend http
8189 mode http
8190 balance roundrobin
8191 stick on src table https
8192 cookie SRV insert indirect nocache
8193 server s1 192.168.1.1:80 cookie s1
8194 server s2 192.168.1.1:80 cookie s2
8195
8196 backend https
8197 mode tcp
8198 balance roundrobin
8199 stick-table type ip size 200k expire 30m
8200 stick on src
8201 server s1 192.168.1.1:443
8202 server s2 192.168.1.1:443
8203
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008204 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008205
8206
8207stick store-request <pattern> [table <table>] [{if | unless} <condition>]
8208 Define a request pattern used to create an entry in a stickiness table
8209 May be used in sections : defaults | frontend | listen | backend
8210 no | no | yes | yes
8211
8212 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008213 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008214 describes what elements of the incoming request or connection
8215 will be analysed, extracted and stored in the table once a
8216 server is selected.
8217
8218 <table> is an optional stickiness table name. If unspecified, the same
8219 backend's table is used. A stickiness table is declared using
8220 the "stick-table" statement.
8221
8222 <cond> is an optional storage condition. It makes it possible to store
8223 certain criteria only when some conditions are met (or not met).
8224 For instance, it could be used to store the source IP address
8225 except when the request passes through a known proxy, in which
8226 case we'd store a converted form of a header containing that IP
8227 address.
8228
8229 Some protocols or applications require complex stickiness rules and cannot
8230 always simply rely on cookies nor hashing. The "stick store-request" statement
8231 describes a rule to decide what to extract from the request and when to do
8232 it, in order to store it into a stickiness table for further requests to
8233 match it using the "stick match" statement. Obviously the extracted part must
8234 make sense and have a chance to be matched in a further request. Storing a
8235 client's IP address for instance often makes sense. Storing an ID found in a
8236 URL parameter also makes sense. Storing a source port will almost never make
8237 any sense because it will be randomly matched. See section 7 for a complete
8238 list of possible patterns and transformation rules.
8239
8240 The table has to be declared using the "stick-table" statement. It must be of
8241 a type compatible with the pattern. By default it is the one which is present
8242 in the same backend. It is possible to share a table with other backends by
8243 referencing it using the "table" keyword. If another table is referenced,
8244 the server's ID inside the backends are used. By default, all server IDs
8245 start at 1 in each backend, so the server ordering is enough. But in case of
8246 doubt, it is highly recommended to force server IDs using their "id" setting.
8247
8248 It is possible to restrict the conditions where a "stick store-request"
8249 statement will apply, using "if" or "unless" followed by a condition. This
8250 condition will be evaluated while parsing the request, so any criteria can be
8251 used. See section 7 for ACL based conditions.
8252
8253 There is no limit on the number of "stick store-request" statements, but
8254 there is a limit of 8 simultaneous stores per request or response. This
8255 makes it possible to store up to 8 criteria, all extracted from either the
8256 request or the response, regardless of the number of rules. Only the 8 first
8257 ones which match will be kept. Using this, it is possible to feed multiple
8258 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01008259 another protocol or access method. Using multiple store-request rules with
8260 the same table is possible and may be used to find the best criterion to rely
8261 on, by arranging the rules by decreasing preference order. Only the first
8262 extracted criterion for a given table will be stored. All subsequent store-
8263 request rules referencing the same table will be skipped and their ACLs will
8264 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008265
8266 The "store-request" rules are evaluated once the server connection has been
8267 established, so that the table will contain the real server that processed
8268 the request.
8269
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008270 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8271 unless you know what you do : memory is not shared between the
8272 processes, which can result in random behaviours.
8273
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008274 Example :
8275 # forward SMTP users to the same server they just used for POP in the
8276 # last 30 minutes
8277 backend pop
8278 mode tcp
8279 balance roundrobin
8280 stick store-request src
8281 stick-table type ip size 200k expire 30m
8282 server s1 192.168.1.1:110
8283 server s2 192.168.1.1:110
8284
8285 backend smtp
8286 mode tcp
8287 balance roundrobin
8288 stick match src table pop
8289 server s1 192.168.1.1:25
8290 server s2 192.168.1.1:25
8291
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008292 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02008293 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008294
8295
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008296stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Emeric Brunf099e792010-09-27 12:05:28 +02008297 size <size> [expire <expire>] [nopurge] [peers <peersect>]
8298 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +08008299 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008300 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +02008301 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008302
8303 Arguments :
8304 ip a table declared with "type ip" will only store IPv4 addresses.
8305 This form is very compact (about 50 bytes per entry) and allows
8306 very fast entry lookup and stores with almost no overhead. This
8307 is mainly used to store client source IP addresses.
8308
David du Colombier9a6d3c92011-03-17 10:40:24 +01008309 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
8310 This form is very compact (about 60 bytes per entry) and allows
8311 very fast entry lookup and stores with almost no overhead. This
8312 is mainly used to store client source IP addresses.
8313
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008314 integer a table declared with "type integer" will store 32bit integers
8315 which can represent a client identifier found in a request for
8316 instance.
8317
8318 string a table declared with "type string" will store substrings of up
8319 to <len> characters. If the string provided by the pattern
8320 extractor is larger than <len>, it will be truncated before
8321 being stored. During matching, at most <len> characters will be
8322 compared between the string in the table and the extracted
8323 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008324 to 32 characters.
8325
8326 binary a table declared with "type binary" will store binary blocks
8327 of <len> bytes. If the block provided by the pattern
8328 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +02008329 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008330 is shorter than <len>, it will be padded by 0. When not
8331 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008332
8333 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +02008334 "string" type table (See type "string" above). Or the number
8335 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008336 changing this parameter as memory usage will proportionally
8337 increase.
8338
8339 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +01008340 value directly impacts memory usage. Count approximately
8341 50 bytes per entry, plus the size of a string if any. The size
8342 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008343
8344 [nopurge] indicates that we refuse to purge older entries when the table
8345 is full. When not specified and the table is full when haproxy
8346 wants to store an entry in it, it will flush a few of the oldest
8347 entries in order to release some space for the new ones. This is
8348 most often the desired behaviour. In some specific cases, it
8349 be desirable to refuse new entries instead of purging the older
8350 ones. That may be the case when the amount of data to store is
8351 far above the hardware limits and we prefer not to offer access
8352 to new clients than to reject the ones already connected. When
8353 using this parameter, be sure to properly set the "expire"
8354 parameter (see below).
8355
Emeric Brunf099e792010-09-27 12:05:28 +02008356 <peersect> is the name of the peers section to use for replication. Entries
8357 which associate keys to server IDs are kept synchronized with
8358 the remote peers declared in this section. All entries are also
8359 automatically learned from the local peer (old process) during a
8360 soft restart.
8361
Willy Tarreau1abc6732015-05-01 19:21:02 +02008362 NOTE : each peers section may be referenced only by tables
8363 belonging to the same unique process.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008364
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008365 <expire> defines the maximum duration of an entry in the table since it
8366 was last created, refreshed or matched. The expiration delay is
8367 defined using the standard time format, similarly as the various
8368 timeouts. The maximum duration is slightly above 24 days. See
8369 section 2.2 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +02008370 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008371 be removed once full. Be sure not to use the "nopurge" parameter
8372 if not expiration delay is specified.
8373
Willy Tarreau08d5f982010-06-06 13:34:54 +02008374 <data_type> is used to store additional information in the stick-table. This
8375 may be used by ACLs in order to control various criteria related
8376 to the activity of the client matching the stick-table. For each
8377 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +02008378 that the additional data can fit. Several data types may be
8379 stored with an entry. Multiple data types may be specified after
8380 the "store" keyword, as a comma-separated list. Alternatively,
8381 it is possible to repeat the "store" keyword followed by one or
8382 several data types. Except for the "server_id" type which is
8383 automatically detected and enabled, all data types must be
8384 explicitly declared to be stored. If an ACL references a data
8385 type which is not stored, the ACL will simply not match. Some
8386 data types require an argument which must be passed just after
8387 the type between parenthesis. See below for the supported data
8388 types and their arguments.
8389
8390 The data types that can be stored with an entry are the following :
8391 - server_id : this is an integer which holds the numeric ID of the server a
8392 request was assigned to. It is used by the "stick match", "stick store",
8393 and "stick on" rules. It is automatically enabled when referenced.
8394
8395 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
8396 integer which may be used for anything. Most of the time it will be used
8397 to put a special tag on some entries, for instance to note that a
8398 specific behaviour was detected and must be known for future matches.
8399
Willy Tarreauba2ffd12013-05-29 15:54:14 +02008400 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
8401 over a period. It is a positive 32-bit integer integer which may be used
8402 for anything. Just like <gpc0>, it counts events, but instead of keeping
8403 a cumulative count, it maintains the rate at which the counter is
8404 incremented. Most of the time it will be used to measure the frequency of
8405 occurrence of certain events (eg: requests to a specific URL).
8406
Willy Tarreauc9705a12010-07-27 20:05:50 +02008407 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
8408 the absolute number of connections received from clients which matched
8409 this entry. It does not mean the connections were accepted, just that
8410 they were received.
8411
8412 - conn_cur : Current Connections. It is a positive 32-bit integer which
8413 stores the concurrent connection counts for the entry. It is incremented
8414 once an incoming connection matches the entry, and decremented once the
8415 connection leaves. That way it is possible to know at any time the exact
8416 number of concurrent connections for an entry.
8417
8418 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8419 integer parameter <period> which indicates in milliseconds the length
8420 of the period over which the average is measured. It reports the average
8421 incoming connection rate over that period, in connections per period. The
8422 result is an integer which can be matched using ACLs.
8423
8424 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
8425 the absolute number of sessions received from clients which matched this
8426 entry. A session is a connection that was accepted by the layer 4 rules.
8427
8428 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8429 integer parameter <period> which indicates in milliseconds the length
8430 of the period over which the average is measured. It reports the average
8431 incoming session rate over that period, in sessions per period. The
8432 result is an integer which can be matched using ACLs.
8433
8434 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
8435 counts the absolute number of HTTP requests received from clients which
8436 matched this entry. It does not matter whether they are valid requests or
8437 not. Note that this is different from sessions when keep-alive is used on
8438 the client side.
8439
8440 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8441 integer parameter <period> which indicates in milliseconds the length
8442 of the period over which the average is measured. It reports the average
8443 HTTP request rate over that period, in requests per period. The result is
8444 an integer which can be matched using ACLs. It does not matter whether
8445 they are valid requests or not. Note that this is different from sessions
8446 when keep-alive is used on the client side.
8447
8448 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
8449 counts the absolute number of HTTP requests errors induced by clients
8450 which matched this entry. Errors are counted on invalid and truncated
8451 requests, as well as on denied or tarpitted requests, and on failed
8452 authentications. If the server responds with 4xx, then the request is
8453 also counted as an error since it's an error triggered by the client
8454 (eg: vulnerability scan).
8455
8456 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8457 integer parameter <period> which indicates in milliseconds the length
8458 of the period over which the average is measured. It reports the average
8459 HTTP request error rate over that period, in requests per period (see
8460 http_err_cnt above for what is accounted as an error). The result is an
8461 integer which can be matched using ACLs.
8462
8463 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
8464 integer which counts the cumulated amount of bytes received from clients
8465 which matched this entry. Headers are included in the count. This may be
8466 used to limit abuse of upload features on photo or video servers.
8467
8468 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
8469 integer parameter <period> which indicates in milliseconds the length
8470 of the period over which the average is measured. It reports the average
8471 incoming bytes rate over that period, in bytes per period. It may be used
8472 to detect users which upload too much and too fast. Warning: with large
8473 uploads, it is possible that the amount of uploaded data will be counted
8474 once upon termination, thus causing spikes in the average transfer speed
8475 instead of having a smooth one. This may partially be smoothed with
8476 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
8477 recommended for better fairness.
8478
8479 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
8480 integer which counts the cumulated amount of bytes sent to clients which
8481 matched this entry. Headers are included in the count. This may be used
8482 to limit abuse of bots sucking the whole site.
8483
8484 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
8485 an integer parameter <period> which indicates in milliseconds the length
8486 of the period over which the average is measured. It reports the average
8487 outgoing bytes rate over that period, in bytes per period. It may be used
8488 to detect users which download too much and too fast. Warning: with large
8489 transfers, it is possible that the amount of transferred data will be
8490 counted once upon termination, thus causing spikes in the average
8491 transfer speed instead of having a smooth one. This may partially be
8492 smoothed with "option contstats" though this is not perfect yet. Use of
8493 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +02008494
Willy Tarreauc00cdc22010-06-06 16:48:26 +02008495 There is only one stick-table per proxy. At the moment of writing this doc,
8496 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008497 to be required, simply create a dummy backend with a stick-table in it and
8498 reference it.
8499
8500 It is important to understand that stickiness based on learning information
8501 has some limitations, including the fact that all learned associations are
Baptiste Assmann123ff042016-03-06 23:29:28 +01008502 lost upon restart unless peers are properly configured to transfer such
8503 information upon restart (recommended). In general it can be good as a
8504 complement but not always as an exclusive stickiness.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008505
Willy Tarreauc9705a12010-07-27 20:05:50 +02008506 Last, memory requirements may be important when storing many data types.
8507 Indeed, storing all indicators above at once in each entry requires 116 bytes
8508 per entry, or 116 MB for a 1-million entries table. This is definitely not
8509 something that can be ignored.
8510
8511 Example:
8512 # Keep track of counters of up to 1 million IP addresses over 5 minutes
8513 # and store a general purpose counter and the average connection rate
8514 # computed over a sliding window of 30 seconds.
8515 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
8516
8517 See also : "stick match", "stick on", "stick store-request", section 2.2
David du Colombiera13d1b92011-03-17 10:40:22 +01008518 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008519
8520
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008521stick store-response <pattern> [table <table>] [{if | unless} <condition>]
Baptiste Assmann2f2d2ec2016-03-06 23:27:24 +01008522 Define a response pattern used to create an entry in a stickiness table
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008523 May be used in sections : defaults | frontend | listen | backend
8524 no | no | yes | yes
8525
8526 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008527 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008528 describes what elements of the response or connection will
8529 be analysed, extracted and stored in the table once a
8530 server is selected.
8531
8532 <table> is an optional stickiness table name. If unspecified, the same
8533 backend's table is used. A stickiness table is declared using
8534 the "stick-table" statement.
8535
8536 <cond> is an optional storage condition. It makes it possible to store
8537 certain criteria only when some conditions are met (or not met).
8538 For instance, it could be used to store the SSL session ID only
8539 when the response is a SSL server hello.
8540
8541 Some protocols or applications require complex stickiness rules and cannot
8542 always simply rely on cookies nor hashing. The "stick store-response"
8543 statement describes a rule to decide what to extract from the response and
8544 when to do it, in order to store it into a stickiness table for further
8545 requests to match it using the "stick match" statement. Obviously the
8546 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008547 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008548 See section 7 for a complete list of possible patterns and transformation
8549 rules.
8550
8551 The table has to be declared using the "stick-table" statement. It must be of
8552 a type compatible with the pattern. By default it is the one which is present
8553 in the same backend. It is possible to share a table with other backends by
8554 referencing it using the "table" keyword. If another table is referenced,
8555 the server's ID inside the backends are used. By default, all server IDs
8556 start at 1 in each backend, so the server ordering is enough. But in case of
8557 doubt, it is highly recommended to force server IDs using their "id" setting.
8558
8559 It is possible to restrict the conditions where a "stick store-response"
8560 statement will apply, using "if" or "unless" followed by a condition. This
8561 condition will be evaluated while parsing the response, so any criteria can
8562 be used. See section 7 for ACL based conditions.
8563
8564 There is no limit on the number of "stick store-response" statements, but
8565 there is a limit of 8 simultaneous stores per request or response. This
8566 makes it possible to store up to 8 criteria, all extracted from either the
8567 request or the response, regardless of the number of rules. Only the 8 first
8568 ones which match will be kept. Using this, it is possible to feed multiple
8569 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01008570 another protocol or access method. Using multiple store-response rules with
8571 the same table is possible and may be used to find the best criterion to rely
8572 on, by arranging the rules by decreasing preference order. Only the first
8573 extracted criterion for a given table will be stored. All subsequent store-
8574 response rules referencing the same table will be skipped and their ACLs will
8575 not be evaluated. However, even if a store-request rule references a table, a
8576 store-response rule may also use the same table. This means that each table
8577 may learn exactly one element from the request and one element from the
8578 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008579
8580 The table will contain the real server that processed the request.
8581
8582 Example :
8583 # Learn SSL session ID from both request and response and create affinity.
8584 backend https
8585 mode tcp
8586 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +02008587 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008588 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008589
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008590 acl clienthello req_ssl_hello_type 1
8591 acl serverhello rep_ssl_hello_type 2
8592
8593 # use tcp content accepts to detects ssl client and server hello.
8594 tcp-request inspect-delay 5s
8595 tcp-request content accept if clienthello
8596
8597 # no timeout on response inspect delay by default.
8598 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008599
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008600 # SSL session ID (SSLID) may be present on a client or server hello.
8601 # Its length is coded on 1 byte at offset 43 and its value starts
8602 # at offset 44.
8603
8604 # Match and learn on request if client hello.
8605 stick on payload_lv(43,1) if clienthello
8606
8607 # Learn on response if server hello.
8608 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +02008609
Emeric Brun6a1cefa2010-09-24 18:15:17 +02008610 server s1 192.168.1.1:443
8611 server s2 192.168.1.1:443
8612
8613 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
8614 extraction.
8615
8616
Willy Tarreau938c7fe2014-04-25 14:21:39 +02008617tcp-check connect [params*]
8618 Opens a new connection
8619 May be used in sections: defaults | frontend | listen | backend
8620 no | no | yes | yes
8621
8622 When an application lies on more than a single TCP port or when HAProxy
8623 load-balance many services in a single backend, it makes sense to probe all
8624 the services individually before considering a server as operational.
8625
8626 When there are no TCP port configured on the server line neither server port
8627 directive, then the 'tcp-check connect port <port>' must be the first step
8628 of the sequence.
8629
8630 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
8631 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
8632 do.
8633
8634 Parameters :
8635 They are optional and can be used to describe how HAProxy should open and
8636 use the TCP connection.
8637
8638 port if not set, check port or server port is used.
8639 It tells HAProxy where to open the connection to.
8640 <port> must be a valid TCP port source integer, from 1 to 65535.
8641
8642 send-proxy send a PROXY protocol string
8643
8644 ssl opens a ciphered connection
8645
8646 Examples:
8647 # check HTTP and HTTPs services on a server.
8648 # first open port 80 thanks to server line port directive, then
8649 # tcp-check opens port 443, ciphered and run a request on it:
8650 option tcp-check
8651 tcp-check connect
8652 tcp-check send GET\ /\ HTTP/1.0\r\n
8653 tcp-check send Host:\ haproxy.1wt.eu\r\n
8654 tcp-check send \r\n
8655 tcp-check expect rstring (2..|3..)
8656 tcp-check connect port 443 ssl
8657 tcp-check send GET\ /\ HTTP/1.0\r\n
8658 tcp-check send Host:\ haproxy.1wt.eu\r\n
8659 tcp-check send \r\n
8660 tcp-check expect rstring (2..|3..)
8661 server www 10.0.0.1 check port 80
8662
8663 # check both POP and IMAP from a single server:
8664 option tcp-check
8665 tcp-check connect port 110
8666 tcp-check expect string +OK\ POP3\ ready
8667 tcp-check connect port 143
8668 tcp-check expect string *\ OK\ IMAP4\ ready
8669 server mail 10.0.0.1 check
8670
8671 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
8672
8673
8674tcp-check expect [!] <match> <pattern>
8675 Specify data to be collected and analysed during a generic health check
8676 May be used in sections: defaults | frontend | listen | backend
8677 no | no | yes | yes
8678
8679 Arguments :
8680 <match> is a keyword indicating how to look for a specific pattern in the
8681 response. The keyword may be one of "string", "rstring" or
8682 binary.
8683 The keyword may be preceded by an exclamation mark ("!") to negate
8684 the match. Spaces are allowed between the exclamation mark and the
8685 keyword. See below for more details on the supported keywords.
8686
8687 <pattern> is the pattern to look for. It may be a string or a regular
8688 expression. If the pattern contains spaces, they must be escaped
8689 with the usual backslash ('\').
8690 If the match is set to binary, then the pattern must be passed as
8691 a serie of hexadecimal digits in an even number. Each sequence of
8692 two digits will represent a byte. The hexadecimal digits may be
8693 used upper or lower case.
8694
8695
8696 The available matches are intentionally similar to their http-check cousins :
8697
8698 string <string> : test the exact string matches in the response buffer.
8699 A health check response will be considered valid if the
8700 response's buffer contains this exact string. If the
8701 "string" keyword is prefixed with "!", then the response
8702 will be considered invalid if the body contains this
8703 string. This can be used to look for a mandatory pattern
8704 in a protocol response, or to detect a failure when a
8705 specific error appears in a protocol banner.
8706
8707 rstring <regex> : test a regular expression on the response buffer.
8708 A health check response will be considered valid if the
8709 response's buffer matches this expression. If the
8710 "rstring" keyword is prefixed with "!", then the response
8711 will be considered invalid if the body matches the
8712 expression.
8713
8714 binary <hexstring> : test the exact string in its hexadecimal form matches
8715 in the response buffer. A health check response will
8716 be considered valid if the response's buffer contains
8717 this exact hexadecimal string.
8718 Purpose is to match data on binary protocols.
8719
8720 It is important to note that the responses will be limited to a certain size
8721 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
8722 Thus, too large responses may not contain the mandatory pattern when using
8723 "string", "rstring" or binary. If a large response is absolutely required, it
8724 is possible to change the default max size by setting the global variable.
8725 However, it is worth keeping in mind that parsing very large responses can
8726 waste some CPU cycles, especially when regular expressions are used, and that
8727 it is always better to focus the checks on smaller resources. Also, in its
8728 current state, the check will not find any string nor regex past a null
8729 character in the response. Similarly it is not possible to request matching
8730 the null character.
8731
8732 Examples :
8733 # perform a POP check
8734 option tcp-check
8735 tcp-check expect string +OK\ POP3\ ready
8736
8737 # perform an IMAP check
8738 option tcp-check
8739 tcp-check expect string *\ OK\ IMAP4\ ready
8740
8741 # look for the redis master server
8742 option tcp-check
8743 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02008744 tcp-check expect string +PONG
Willy Tarreau938c7fe2014-04-25 14:21:39 +02008745 tcp-check send info\ replication\r\n
8746 tcp-check expect string role:master
8747 tcp-check send QUIT\r\n
8748 tcp-check expect string +OK
8749
8750
8751 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
8752 "tcp-check send-binary", "http-check expect", tune.chksize
8753
8754
8755tcp-check send <data>
8756 Specify a string to be sent as a question during a generic health check
8757 May be used in sections: defaults | frontend | listen | backend
8758 no | no | yes | yes
8759
8760 <data> : the data to be sent as a question during a generic health check
8761 session. For now, <data> must be a string.
8762
8763 Examples :
8764 # look for the redis master server
8765 option tcp-check
8766 tcp-check send info\ replication\r\n
8767 tcp-check expect string role:master
8768
8769 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
8770 "tcp-check send-binary", tune.chksize
8771
8772
8773tcp-check send-binary <hexastring>
8774 Specify an hexa digits string to be sent as a binary question during a raw
8775 tcp health check
8776 May be used in sections: defaults | frontend | listen | backend
8777 no | no | yes | yes
8778
8779 <data> : the data to be sent as a question during a generic health check
8780 session. For now, <data> must be a string.
8781 <hexastring> : test the exact string in its hexadecimal form matches in the
8782 response buffer. A health check response will be considered
8783 valid if the response's buffer contains this exact
8784 hexadecimal string.
8785 Purpose is to send binary data to ask on binary protocols.
8786
8787 Examples :
8788 # redis check in binary
8789 option tcp-check
8790 tcp-check send-binary 50494e470d0a # PING\r\n
8791 tcp-check expect binary 2b504F4e47 # +PONG
8792
8793
8794 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
8795 "tcp-check send", tune.chksize
8796
8797
Willy Tarreaue9656522010-08-17 15:40:09 +02008798tcp-request connection <action> [{if | unless} <condition>]
8799 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +02008800 May be used in sections : defaults | frontend | listen | backend
8801 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +02008802 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02008803 <action> defines the action to perform if the condition applies. See
8804 below.
Willy Tarreau1a687942010-05-23 22:40:30 +02008805
Willy Tarreaue9656522010-08-17 15:40:09 +02008806 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008807
8808 Immediately after acceptance of a new incoming connection, it is possible to
8809 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +02008810 or dropped or have its counters tracked. Those conditions cannot make use of
8811 any data contents because the connection has not been read from yet, and the
8812 buffers are not yet allocated. This is used to selectively and very quickly
8813 accept or drop connections from various sources with a very low overhead. If
8814 some contents need to be inspected in order to take the decision, the
8815 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008816
Willy Tarreaue9656522010-08-17 15:40:09 +02008817 The "tcp-request connection" rules are evaluated in their exact declaration
8818 order. If no rule matches or if there is no rule, the default action is to
8819 accept the incoming connection. There is no specific limit to the number of
8820 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008821
Willy Tarreaua9083d02015-05-08 15:27:59 +02008822 Four types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +02008823 - accept :
8824 accepts the connection if the condition is true (when used with "if")
8825 or false (when used with "unless"). The first such rule executed ends
8826 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008827
Willy Tarreaue9656522010-08-17 15:40:09 +02008828 - reject :
8829 rejects the connection if the condition is true (when used with "if")
8830 or false (when used with "unless"). The first such rule executed ends
8831 the rules evaluation. Rejected connections do not even become a
8832 session, which is why they are accounted separately for in the stats,
8833 as "denied connections". They are not considered for the session
8834 rate-limit and are not logged either. The reason is that these rules
8835 should only be used to filter extremely high connection rates such as
8836 the ones encountered during a massive DDoS attack. Under these extreme
8837 conditions, the simple action of logging each event would make the
8838 system collapse and would considerably lower the filtering capacity. If
8839 logging is absolutely desired, then "tcp-request content" rules should
Willy Tarreau4f614292016-10-21 17:49:36 +02008840 be used instead, as "tcp-request session" rules will not log either.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008841
Willy Tarreau4f0d9192013-06-11 20:40:55 +02008842 - expect-proxy layer4 :
8843 configures the client-facing connection to receive a PROXY protocol
8844 header before any byte is read from the socket. This is equivalent to
8845 having the "accept-proxy" keyword on the "bind" line, except that using
8846 the TCP rule allows the PROXY protocol to be accepted only for certain
8847 IP address ranges using an ACL. This is convenient when multiple layers
8848 of load balancers are passed through by traffic coming from public
8849 hosts.
8850
Bertrand Jacquin90759682016-06-06 15:35:39 +01008851 - expect-netscaler-cip layer4 :
8852 configures the client-facing connection to receive a NetScaler Client
8853 IP insertion protocol header before any byte is read from the socket.
8854 This is equivalent to having the "accept-netscaler-cip" keyword on the
8855 "bind" line, except that using the TCP rule allows the PROXY protocol
8856 to be accepted only for certain IP address ranges using an ACL. This
8857 is convenient when multiple layers of load balancers are passed
8858 through by traffic coming from public hosts.
8859
Willy Tarreau18bf01e2014-06-13 16:18:52 +02008860 - capture <sample> len <length> :
8861 This only applies to "tcp-request content" rules. It captures sample
8862 expression <sample> from the request buffer, and converts it to a
8863 string of at most <len> characters. The resulting string is stored into
8864 the next request "capture" slot, so it will possibly appear next to
8865 some captured HTTP headers. It will then automatically appear in the
8866 logs, and it will be possible to extract it using sample fetch rules to
8867 feed it into headers or anything. The length should be limited given
8868 that this size will be allocated for each capture during the whole
Willy Tarreaua9083d02015-05-08 15:27:59 +02008869 session life. Please check section 7.3 (Fetching samples) and "capture
8870 request header" for more information.
Willy Tarreau18bf01e2014-06-13 16:18:52 +02008871
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02008872 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +02008873 enables tracking of sticky counters from current connection. These
Willy Tarreau09448f72014-06-25 18:12:15 +02008874 rules do not stop evaluation and do not change default action. 3 sets
Willy Tarreaue9656522010-08-17 15:40:09 +02008875 of counters may be simultaneously tracked by the same connection. The
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02008876 first "track-sc0" rule executed enables tracking of the counters of the
8877 specified table as the first set. The first "track-sc1" rule executed
Willy Tarreaue9656522010-08-17 15:40:09 +02008878 enables tracking of the counters of the specified table as the second
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02008879 set. The first "track-sc2" rule executed enables tracking of the
8880 counters of the specified table as the third set. It is a recommended
8881 practice to use the first set of counters for the per-frontend counters
8882 and the second set for the per-backend ones. But this is just a
8883 guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008884
Willy Tarreaue9656522010-08-17 15:40:09 +02008885 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008886 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +02008887 in section 7.3. It describes what elements of the incoming
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01008888 request or connection will be analysed, extracted, combined,
8889 and used to select which table entry to update the counters.
8890 Note that "tcp-request connection" cannot use content-based
8891 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008892
Willy Tarreaue9656522010-08-17 15:40:09 +02008893 <table> is an optional table to be used instead of the default one,
8894 which is the stick-table declared in the current proxy. All
8895 the counters for the matches and updates for the key will
8896 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008897
Willy Tarreaue9656522010-08-17 15:40:09 +02008898 Once a "track-sc*" rule is executed, the key is looked up in the table
8899 and if it is not found, an entry is allocated for it. Then a pointer to
8900 that entry is kept during all the session's life, and this entry's
8901 counters are updated as often as possible, every time the session's
8902 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01008903 Counters are only updated for events that happen after the tracking has
8904 been started. For example, connection counters will not be updated when
8905 tracking layer 7 information, since the connection event happens before
8906 layer7 information is extracted.
8907
Willy Tarreaue9656522010-08-17 15:40:09 +02008908 If the entry tracks concurrent connection counters, one connection is
8909 counted for as long as the entry is tracked, and the entry will not
8910 expire during that time. Tracking counters also provides a performance
8911 advantage over just checking the keys, because only one table lookup is
8912 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02008913
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02008914 - sc-inc-gpc0(<sc-id>):
8915 The "sc-inc-gpc0" increments the GPC0 counter according to the sticky
8916 counter designated by <sc-id>. If an error occurs, this action silently
8917 fails and the actions evaluation continues.
8918
Thierry FOURNIER236657b2015-08-19 08:25:14 +02008919 - sc-set-gpt0(<sc-id>) <int>:
8920 This action sets the GPT0 tag according to the sticky counter designated
8921 by <sc-id> and the value of <int>. The expected result is a boolean. If
8922 an error occurs, this action silently fails and the actions evaluation
8923 continues.
8924
William Lallemand2e785f22016-05-25 01:48:42 +02008925 - set-src <expr> :
8926 Is used to set the source IP address to the value of specified
8927 expression. Useful if you want to mask source IP for privacy.
8928 If you want to provide an IP from a HTTP header use "http-request
8929 set-src"
8930
8931 <expr> Is a standard HAProxy expression formed by a sample-fetch
8932 followed by some converters.
8933
8934 Example:
8935
8936 tcp-request connection set-src src,ipmask(24)
8937
Willy Tarreau0c630532016-10-21 17:52:58 +02008938 When possible, set-src preserves the original source port as long as the
8939 address family allows it, otherwise the source port is set to 0.
William Lallemand2e785f22016-05-25 01:48:42 +02008940
William Lallemand44be6402016-05-25 01:51:35 +02008941 - set-src-port <expr> :
8942 Is used to set the source port address to the value of specified
8943 expression.
8944
8945 <expr> Is a standard HAProxy expression formed by a sample-fetch
8946 followed by some converters.
8947
8948 Example:
8949
8950 tcp-request connection set-src-port int(4000)
8951
Willy Tarreau0c630532016-10-21 17:52:58 +02008952 When possible, set-src-port preserves the original source address as long
8953 as the address family supports a port, otherwise it forces the source
8954 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02008955
William Lallemand13e9b0c2016-05-25 02:34:07 +02008956 - set-dst <expr> :
8957 Is used to set the destination IP address to the value of specified
8958 expression. Useful if you want to mask IP for privacy in log.
8959 If you want to provide an IP from a HTTP header use "http-request
8960 set-dst". If you want to connect to the new address/port, use
8961 '0.0.0.0:0' as a server address in the backend.
8962
8963 <expr> Is a standard HAProxy expression formed by a sample-fetch
8964 followed by some converters.
8965
8966 Example:
8967
8968 tcp-request connection set-dst dst,ipmask(24)
8969 tcp-request connection set-dst ipv4(10.0.0.1)
8970
Willy Tarreau0c630532016-10-21 17:52:58 +02008971 When possible, set-dst preserves the original destination port as long as
8972 the address family allows it, otherwise the destination port is set to 0.
8973
William Lallemand13e9b0c2016-05-25 02:34:07 +02008974 - set-dst-port <expr> :
8975 Is used to set the destination port address to the value of specified
8976 expression. If you want to connect to the new address/port, use
8977 '0.0.0.0:0' as a server address in the backend.
8978
8979
8980 <expr> Is a standard HAProxy expression formed by a sample-fetch
8981 followed by some converters.
8982
8983 Example:
8984
8985 tcp-request connection set-dst-port int(4000)
8986
Willy Tarreau0c630532016-10-21 17:52:58 +02008987 When possible, set-dst-port preserves the original destination address as
8988 long as the address family supports a port, otherwise it forces the
8989 destination address to IPv4 "0.0.0.0" before rewriting the port.
8990
Willy Tarreau2d392c22015-08-24 01:43:45 +02008991 - "silent-drop" :
8992 This stops the evaluation of the rules and makes the client-facing
8993 connection suddenly disappear using a system-dependant way that tries
8994 to prevent the client from being notified. The effect it then that the
8995 client still sees an established connection while there's none on
8996 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
8997 except that it doesn't use any local resource at all on the machine
8998 running HAProxy. It can resist much higher loads than "tarpit", and
8999 slow down stronger attackers. It is important to undestand the impact
9000 of using this mechanism. All stateful equipments placed between the
9001 client and HAProxy (firewalls, proxies, load balancers) will also keep
9002 the established connection for a long time and may suffer from this
9003 action. On modern Linux systems running with enough privileges, the
9004 TCP_REPAIR socket option is used to block the emission of a TCP
9005 reset. On other systems, the socket's TTL is reduced to 1 so that the
9006 TCP reset doesn't pass the first router, though it's still delivered to
9007 local networks. Do not use it unless you fully understand how it works.
9008
Willy Tarreaue9656522010-08-17 15:40:09 +02009009 Note that the "if/unless" condition is optional. If no condition is set on
9010 the action, it is simply performed unconditionally. That can be useful for
9011 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009012
Willy Tarreaue9656522010-08-17 15:40:09 +02009013 Example: accept all connections from white-listed hosts, reject too fast
9014 connection without counting them, and track accepted connections.
9015 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009016
Willy Tarreaue9656522010-08-17 15:40:09 +02009017 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009018 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009019 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009020
Willy Tarreaue9656522010-08-17 15:40:09 +02009021 Example: accept all connections from white-listed hosts, count all other
9022 connections and reject too fast ones. This results in abusive ones
9023 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009024
Willy Tarreaue9656522010-08-17 15:40:09 +02009025 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009026 tcp-request connection track-sc0 src
9027 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009028
Willy Tarreau4f0d9192013-06-11 20:40:55 +02009029 Example: enable the PROXY protocol for traffic coming from all known proxies.
9030
9031 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
9032
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009033 See section 7 about ACL usage.
9034
Willy Tarreau4f614292016-10-21 17:49:36 +02009035 See also : "tcp-request session", "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009036
9037
Willy Tarreaue9656522010-08-17 15:40:09 +02009038tcp-request content <action> [{if | unless} <condition>]
9039 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +02009040 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02009041 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +02009042 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009043 <action> defines the action to perform if the condition applies. See
9044 below.
Willy Tarreau62644772008-07-16 18:36:06 +02009045
Willy Tarreaue9656522010-08-17 15:40:09 +02009046 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +02009047
Willy Tarreaue9656522010-08-17 15:40:09 +02009048 A request's contents can be analysed at an early stage of request processing
9049 called "TCP content inspection". During this stage, ACL-based rules are
9050 evaluated every time the request contents are updated, until either an
9051 "accept" or a "reject" rule matches, or the TCP request inspection delay
9052 expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +02009053
Willy Tarreaue9656522010-08-17 15:40:09 +02009054 The first difference between these rules and "tcp-request connection" rules
9055 is that "tcp-request content" rules can make use of contents to take a
9056 decision. Most often, these decisions will consider a protocol recognition or
9057 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +01009058 both frontends and backends. In case of HTTP keep-alive with the client, all
9059 tcp-request content rules are evaluated again, so haproxy keeps a record of
9060 what sticky counters were assigned by a "tcp-request connection" versus a
9061 "tcp-request content" rule, and flushes all the content-related ones after
9062 processing an HTTP request, so that they may be evaluated again by the rules
9063 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009064 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +01009065 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +02009066
Willy Tarreaue9656522010-08-17 15:40:09 +02009067 Content-based rules are evaluated in their exact declaration order. If no
9068 rule matches or if there is no rule, the default action is to accept the
9069 contents. There is no specific limit to the number of rules which may be
9070 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +02009071
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009072 Several types of actions are supported :
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009073 - accept : the request is accepted
9074 - reject : the request is rejected and the connection is closed
9075 - capture : the specified sample expression is captured
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009076 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009077 - sc-inc-gpc0(<sc-id>)
Thierry Fournierb9125672016-03-29 19:34:37 +02009078 - sc-set-gpt0(<sc-id>) <int>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009079 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +01009080 - unset-var(<var-name>)
Willy Tarreau2d392c22015-08-24 01:43:45 +02009081 - silent-drop
Willy Tarreau62644772008-07-16 18:36:06 +02009082
Willy Tarreaue9656522010-08-17 15:40:09 +02009083 They have the same meaning as their counter-parts in "tcp-request connection"
9084 so please refer to that section for a complete description.
Willy Tarreau62644772008-07-16 18:36:06 +02009085
Willy Tarreauf3338342014-01-28 21:40:28 +01009086 While there is nothing mandatory about it, it is recommended to use the
9087 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
9088 content" rules in the frontend, and track-sc2 for "tcp-request content"
9089 rules in the backend, because that makes the configuration more readable
9090 and easier to troubleshoot, but this is just a guideline and all counters
9091 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +02009092
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009093 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +02009094 the action, it is simply performed unconditionally. That can be useful for
9095 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +02009096
Willy Tarreaue9656522010-08-17 15:40:09 +02009097 It is perfectly possible to match layer 7 contents with "tcp-request content"
Willy Tarreauc0239e02012-04-16 14:42:55 +02009098 rules, since HTTP-specific ACL matches are able to preliminarily parse the
9099 contents of a buffer before extracting the required data. If the buffered
9100 contents do not parse as a valid HTTP message, then the ACL does not match.
9101 The parser which is involved there is exactly the same as for all other HTTP
Willy Tarreauf3338342014-01-28 21:40:28 +01009102 processing, so there is no risk of parsing something differently. In an HTTP
9103 backend connected to from an HTTP frontend, it is guaranteed that HTTP
9104 contents will always be immediately present when the rule is evaluated first.
Willy Tarreau62644772008-07-16 18:36:06 +02009105
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009106 Tracking layer7 information is also possible provided that the information
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009107 are present when the rule is processed. The rule processing engine is able to
9108 wait until the inspect delay expires when the data to be tracked is not yet
9109 available.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009110
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009111 The "set-var" is used to set the content of a variable. The variable is
Willy Tarreau4f614292016-10-21 17:49:36 +02009112 declared inline. For "tcp-request session" rules, only session-level
9113 variables can be used, without any layer7 contents.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009114
Daniel Schneller0b547052016-03-21 20:46:57 +01009115 <var-name> The name of the variable starts with an indication about
9116 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01009117 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01009118 "sess" : the variable is shared with the whole session
9119 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009120 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01009121 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009122 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01009123 "res" : the variable is shared only during response
9124 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009125 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01009126 The name may only contain characters 'a-z', 'A-Z', '0-9',
9127 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009128
9129 <expr> Is a standard HAProxy expression formed by a sample-fetch
9130 followed by some converters.
9131
Christopher Faulet85d79c92016-11-09 16:54:56 +01009132 The "unset-var" is used to unset a variable. See above for details about
9133 <var-name>.
9134
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009135 Example:
9136
9137 tcp-request content set-var(sess.my_var) src
Christopher Faulet85d79c92016-11-09 16:54:56 +01009138 tcp-request content unset-var(sess.my_var2)
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009139
Willy Tarreau62644772008-07-16 18:36:06 +02009140 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +02009141 # Accept HTTP requests containing a Host header saying "example.com"
9142 # and reject everything else.
9143 acl is_host_com hdr(Host) -i example.com
9144 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +02009145 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +02009146 tcp-request content reject
9147
9148 Example:
Willy Tarreau62644772008-07-16 18:36:06 +02009149 # reject SMTP connection if client speaks first
9150 tcp-request inspect-delay 30s
9151 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009152 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +02009153
9154 # Forward HTTPS connection only if client speaks
9155 tcp-request inspect-delay 30s
9156 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009157 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +02009158 tcp-request content reject
9159
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009160 Example:
9161 # Track the last IP from X-Forwarded-For
9162 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009163 tcp-request content track-sc0 hdr(x-forwarded-for,-1)
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009164
9165 Example:
9166 # track request counts per "base" (concatenation of Host+URL)
9167 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009168 tcp-request content track-sc0 base table req-rate
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009169
Willy Tarreaue9656522010-08-17 15:40:09 +02009170 Example: track per-frontend and per-backend counters, block abusers at the
9171 frontend when the backend detects abuse.
9172
9173 frontend http
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009174 # Use General Purpose Couter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +02009175 # protecting all our sites
9176 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009177 tcp-request connection track-sc0 src
9178 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +02009179 ...
9180 use_backend http_dynamic if { path_end .php }
9181
9182 backend http_dynamic
9183 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009184 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +02009185 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009186 acl click_too_fast sc1_http_req_rate gt 10
9187 acl mark_as_abuser sc0_inc_gpc0 gt 0
9188 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +02009189 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +02009190
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009191 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +02009192
Willy Tarreau4f614292016-10-21 17:49:36 +02009193 See also : "tcp-request connection", "tcp-request session", and
9194 "tcp-request inspect-delay"
Willy Tarreau62644772008-07-16 18:36:06 +02009195
9196
9197tcp-request inspect-delay <timeout>
9198 Set the maximum allowed time to wait for data during content inspection
9199 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02009200 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +02009201 Arguments :
9202 <timeout> is the timeout value specified in milliseconds by default, but
9203 can be in any other unit if the number is suffixed by the unit,
9204 as explained at the top of this document.
9205
9206 People using haproxy primarily as a TCP relay are often worried about the
9207 risk of passing any type of protocol to a server without any analysis. In
9208 order to be able to analyze the request contents, we must first withhold
9209 the data then analyze them. This statement simply enables withholding of
9210 data for at most the specified amount of time.
9211
Willy Tarreaufb356202010-08-03 14:02:05 +02009212 TCP content inspection applies very early when a connection reaches a
9213 frontend, then very early when the connection is forwarded to a backend. This
9214 means that a connection may experience a first delay in the frontend and a
9215 second delay in the backend if both have tcp-request rules.
9216
Willy Tarreau62644772008-07-16 18:36:06 +02009217 Note that when performing content inspection, haproxy will evaluate the whole
9218 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009219 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +02009220 a last check is performed upon expiration, this time considering that the
Willy Tarreaud869b242009-03-15 14:43:58 +01009221 contents are definitive. If no delay is set, haproxy will not wait at all
9222 and will immediately apply a verdict based on the available information.
9223 Obviously this is unlikely to be very useful and might even be racy, so such
9224 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +02009225
9226 As soon as a rule matches, the request is released and continues as usual. If
9227 the timeout is reached and no rule matches, the default policy will be to let
9228 it pass through unaffected.
9229
9230 For most protocols, it is enough to set it to a few seconds, as most clients
9231 send the full request immediately upon connection. Add 3 or more seconds to
9232 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +01009233 to use large values, for instance to ensure that the client never talks
Willy Tarreau62644772008-07-16 18:36:06 +02009234 before the server (eg: SMTP), or to wait for a client to talk before passing
9235 data to the server (eg: SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +02009236 least the inspection delay, otherwise it will expire first. If the client
9237 closes the connection or if the buffer is full, the delay immediately expires
9238 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +02009239
Willy Tarreau55165fe2009-05-10 12:02:55 +02009240 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +02009241 "timeout client".
9242
9243
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009244tcp-response content <action> [{if | unless} <condition>]
9245 Perform an action on a session response depending on a layer 4-7 condition
9246 May be used in sections : defaults | frontend | listen | backend
9247 no | no | yes | yes
9248 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009249 <action> defines the action to perform if the condition applies. See
9250 below.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009251
9252 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
9253
9254 Response contents can be analysed at an early stage of response processing
9255 called "TCP content inspection". During this stage, ACL-based rules are
9256 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009257 "accept", "close" or a "reject" rule matches, or a TCP response inspection
9258 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009259
9260 Most often, these decisions will consider a protocol recognition or validity.
9261
9262 Content-based rules are evaluated in their exact declaration order. If no
9263 rule matches or if there is no rule, the default action is to accept the
9264 contents. There is no specific limit to the number of rules which may be
9265 inserted.
9266
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009267 Several types of actions are supported :
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009268 - accept :
9269 accepts the response if the condition is true (when used with "if")
9270 or false (when used with "unless"). The first such rule executed ends
9271 the rules evaluation.
9272
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009273 - close :
9274 immediately closes the connection with the server if the condition is
9275 true (when used with "if"), or false (when used with "unless"). The
9276 first such rule executed ends the rules evaluation. The main purpose of
9277 this action is to force a connection to be finished between a client
9278 and a server after an exchange when the application protocol expects
9279 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009280 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +02009281 protocols.
9282
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009283 - reject :
9284 rejects the response if the condition is true (when used with "if")
9285 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009286 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009287
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009288 - set-var(<var-name>) <expr>
9289 Sets a variable.
9290
Christopher Faulet85d79c92016-11-09 16:54:56 +01009291 - unset-var(<var-name>)
9292 Unsets a variable.
9293
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009294 - sc-inc-gpc0(<sc-id>):
9295 This action increments the GPC0 counter according to the sticky
9296 counter designated by <sc-id>. If an error occurs, this action fails
9297 silently and the actions evaluation continues.
9298
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009299 - sc-set-gpt0(<sc-id>) <int> :
9300 This action sets the GPT0 tag according to the sticky counter designated
9301 by <sc-id> and the value of <int>. The expected result is a boolean. If
9302 an error occurs, this action silently fails and the actions evaluation
9303 continues.
9304
Willy Tarreau2d392c22015-08-24 01:43:45 +02009305 - "silent-drop" :
9306 This stops the evaluation of the rules and makes the client-facing
9307 connection suddenly disappear using a system-dependant way that tries
9308 to prevent the client from being notified. The effect it then that the
9309 client still sees an established connection while there's none on
9310 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
9311 except that it doesn't use any local resource at all on the machine
9312 running HAProxy. It can resist much higher loads than "tarpit", and
9313 slow down stronger attackers. It is important to undestand the impact
9314 of using this mechanism. All stateful equipments placed between the
9315 client and HAProxy (firewalls, proxies, load balancers) will also keep
9316 the established connection for a long time and may suffer from this
9317 action. On modern Linux systems running with enough privileges, the
9318 TCP_REPAIR socket option is used to block the emission of a TCP
9319 reset. On other systems, the socket's TTL is reduced to 1 so that the
9320 TCP reset doesn't pass the first router, though it's still delivered to
9321 local networks. Do not use it unless you fully understand how it works.
9322
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009323 Note that the "if/unless" condition is optional. If no condition is set on
9324 the action, it is simply performed unconditionally. That can be useful for
9325 for changing the default action to a reject.
9326
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009327 It is perfectly possible to match layer 7 contents with "tcp-response
9328 content" rules, but then it is important to ensure that a full response has
9329 been buffered, otherwise no contents will match. In order to achieve this,
9330 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009331 period.
9332
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009333 The "set-var" is used to set the content of a variable. The variable is
9334 declared inline.
9335
Daniel Schneller0b547052016-03-21 20:46:57 +01009336 <var-name> The name of the variable starts with an indication about
9337 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01009338 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01009339 "sess" : the variable is shared with the whole session
9340 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009341 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01009342 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009343 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01009344 "res" : the variable is shared only during response
9345 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009346 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01009347 The name may only contain characters 'a-z', 'A-Z', '0-9',
9348 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009349
9350 <expr> Is a standard HAProxy expression formed by a sample-fetch
9351 followed by some converters.
9352
9353 Example:
9354
9355 tcp-request content set-var(sess.my_var) src
9356
Christopher Faulet85d79c92016-11-09 16:54:56 +01009357 The "unset-var" is used to unset a variable. See above for details about
9358 <var-name>.
9359
9360 Example:
9361
9362 tcp-request content unset-var(sess.my_var)
9363
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009364 See section 7 about ACL usage.
9365
9366 See also : "tcp-request content", "tcp-response inspect-delay"
9367
9368
Willy Tarreau4f614292016-10-21 17:49:36 +02009369tcp-request session <action> [{if | unless} <condition>]
9370 Perform an action on a validated session depending on a layer 5 condition
9371 May be used in sections : defaults | frontend | listen | backend
9372 no | yes | yes | no
9373 Arguments :
9374 <action> defines the action to perform if the condition applies. See
9375 below.
9376
9377 <condition> is a standard layer5-only ACL-based condition (see section 7).
9378
9379 Once a session is validated, (ie. after all handshakes have been completed),
9380 it is possible to evaluate some conditions to decide whether this session
9381 must be accepted or dropped or have its counters tracked. Those conditions
9382 cannot make use of any data contents because no buffers are allocated yet and
9383 the processing cannot wait at this stage. The main use case it to copy some
9384 early information into variables (since variables are accessible in the
9385 session), or to keep track of some information collected after the handshake,
9386 such as SSL-level elements (SNI, ciphers, client cert's CN) or information
9387 from the PROXY protocol header (eg: track a source forwarded this way). The
9388 extracted information can thus be copied to a variable or tracked using
9389 "track-sc" rules. Of course it is also possible to decide to accept/reject as
9390 with other rulesets. Most operations performed here could also be performed
9391 in "tcp-request content" rules, except that in HTTP these rules are evaluated
9392 for each new request, and that might not always be acceptable. For example a
9393 rule might increment a counter on each evaluation. It would also be possible
9394 that a country is resolved by geolocation from the source IP address,
9395 assigned to a session-wide variable, then the source address rewritten from
9396 an HTTP header for all requests. If some contents need to be inspected in
9397 order to take the decision, the "tcp-request content" statements must be used
9398 instead.
9399
9400 The "tcp-request session" rules are evaluated in their exact declaration
9401 order. If no rule matches or if there is no rule, the default action is to
9402 accept the incoming session. There is no specific limit to the number of
9403 rules which may be inserted.
9404
9405 Several types of actions are supported :
9406 - accept : the request is accepted
9407 - reject : the request is rejected and the connection is closed
9408 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
9409 - sc-inc-gpc0(<sc-id>)
9410 - sc-set-gpt0(<sc-id>) <int>
9411 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +01009412 - unset-var(<var-name>)
Willy Tarreau4f614292016-10-21 17:49:36 +02009413 - silent-drop
9414
9415 These actions have the same meaning as their respective counter-parts in
9416 "tcp-request connection" and "tcp-request content", so please refer to these
9417 sections for a complete description.
9418
9419 Note that the "if/unless" condition is optional. If no condition is set on
9420 the action, it is simply performed unconditionally. That can be useful for
9421 "track-sc*" actions as well as for changing the default action to a reject.
9422
9423 Example: track the original source address by default, or the one advertised
9424 in the PROXY protocol header for connection coming from the local
9425 proxies. The first connection-level rule enables receipt of the
9426 PROXY protocol for these ones, the second rule tracks whatever
9427 address we decide to keep after optional decoding.
9428
9429 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
9430 tcp-request session track-sc0 src
9431
9432 Example: accept all sessions from white-listed hosts, reject too fast
9433 sessions without counting them, and track accepted sessions.
9434 This results in session rate being capped from abusive sources.
9435
9436 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
9437 tcp-request session reject if { src_sess_rate gt 10 }
9438 tcp-request session track-sc0 src
9439
9440 Example: accept all sessions from white-listed hosts, count all other
9441 sessions and reject too fast ones. This results in abusive ones
9442 being blocked as long as they don't slow down.
9443
9444 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
9445 tcp-request session track-sc0 src
9446 tcp-request session reject if { sc0_sess_rate gt 10 }
9447
9448 See section 7 about ACL usage.
9449
9450 See also : "tcp-request connection", "tcp-request content", "stick-table"
9451
9452
Emeric Brun0a3b67f2010-09-24 15:34:53 +02009453tcp-response inspect-delay <timeout>
9454 Set the maximum allowed time to wait for a response during content inspection
9455 May be used in sections : defaults | frontend | listen | backend
9456 no | no | yes | yes
9457 Arguments :
9458 <timeout> is the timeout value specified in milliseconds by default, but
9459 can be in any other unit if the number is suffixed by the unit,
9460 as explained at the top of this document.
9461
9462 See also : "tcp-response content", "tcp-request inspect-delay".
9463
9464
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009465timeout check <timeout>
9466 Set additional check timeout, but only after a connection has been already
9467 established.
9468
9469 May be used in sections: defaults | frontend | listen | backend
9470 yes | no | yes | yes
9471 Arguments:
9472 <timeout> is the timeout value specified in milliseconds by default, but
9473 can be in any other unit if the number is suffixed by the unit,
9474 as explained at the top of this document.
9475
9476 If set, haproxy uses min("timeout connect", "inter") as a connect timeout
9477 for check and "timeout check" as an additional read timeout. The "min" is
9478 used so that people running with *very* long "timeout connect" (eg. those
9479 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +01009480 (Please also note that there is no valid reason to have such long connect
9481 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
9482 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009483
9484 If "timeout check" is not set haproxy uses "inter" for complete check
9485 timeout (connect + read) exactly like all <1.3.15 version.
9486
9487 In most cases check request is much simpler and faster to handle than normal
9488 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +01009489 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009490
9491 This parameter is specific to backends, but can be specified once for all in
9492 "defaults" sections. This is in fact one of the easiest solutions not to
9493 forget about it.
9494
Willy Tarreau41a340d2008-01-22 12:25:31 +01009495 See also: "timeout connect", "timeout queue", "timeout server",
9496 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009497
9498
Willy Tarreau0ba27502007-12-24 16:55:16 +01009499timeout client <timeout>
9500timeout clitimeout <timeout> (deprecated)
9501 Set the maximum inactivity time on the client side.
9502 May be used in sections : defaults | frontend | listen | backend
9503 yes | yes | yes | no
9504 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01009505 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +01009506 can be in any other unit if the number is suffixed by the unit,
9507 as explained at the top of this document.
9508
9509 The inactivity timeout applies when the client is expected to acknowledge or
9510 send data. In HTTP mode, this timeout is particularly important to consider
9511 during the first phase, when the client sends the request, and during the
Baptiste Assmann2e1941e2016-03-06 23:24:12 +01009512 response while it is reading data sent by the server. That said, for the
9513 first phase, it is preferable to set the "timeout http-request" to better
9514 protect HAProxy from Slowloris like attacks. The value is specified in
9515 milliseconds by default, but can be in any other unit if the number is
Willy Tarreau0ba27502007-12-24 16:55:16 +01009516 suffixed by the unit, as specified at the top of this document. In TCP mode
9517 (and to a lesser extent, in HTTP mode), it is highly recommended that the
9518 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01009519 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +01009520 losses by specifying timeouts that are slightly above multiples of 3 seconds
Willy Tarreauce887fd2012-05-12 12:50:00 +02009521 (eg: 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
9522 sessions (eg: WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +02009523 which overrides "timeout client" and "timeout server" for tunnels, as well as
9524 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01009525
9526 This parameter is specific to frontends, but can be specified once for all in
9527 "defaults" sections. This is in fact one of the easiest solutions not to
9528 forget about it. An unspecified timeout results in an infinite timeout, which
9529 is not recommended. Such a usage is accepted and works but reports a warning
9530 during startup because it may results in accumulation of expired sessions in
9531 the system if the system's timeouts are not configured either.
9532
9533 This parameter replaces the old, deprecated "clitimeout". It is recommended
9534 to use it to write new configurations. The form "timeout clitimeout" is
9535 provided only by backwards compatibility but its use is strongly discouraged.
9536
Baptiste Assmann2e1941e2016-03-06 23:24:12 +01009537 See also : "clitimeout", "timeout server", "timeout tunnel",
9538 "timeout http-request".
Willy Tarreau0ba27502007-12-24 16:55:16 +01009539
9540
Willy Tarreau05cdd962014-05-10 14:30:07 +02009541timeout client-fin <timeout>
9542 Set the inactivity timeout on the client side for half-closed connections.
9543 May be used in sections : defaults | frontend | listen | backend
9544 yes | yes | yes | no
9545 Arguments :
9546 <timeout> is the timeout value specified in milliseconds by default, but
9547 can be in any other unit if the number is suffixed by the unit,
9548 as explained at the top of this document.
9549
9550 The inactivity timeout applies when the client is expected to acknowledge or
9551 send data while one direction is already shut down. This timeout is different
9552 from "timeout client" in that it only applies to connections which are closed
9553 in one direction. This is particularly useful to avoid keeping connections in
9554 FIN_WAIT state for too long when clients do not disconnect cleanly. This
9555 problem is particularly common long connections such as RDP or WebSocket.
9556 Note that this timeout can override "timeout tunnel" when a connection shuts
9557 down in one direction.
9558
9559 This parameter is specific to frontends, but can be specified once for all in
9560 "defaults" sections. By default it is not set, so half-closed connections
9561 will use the other timeouts (timeout.client or timeout.tunnel).
9562
9563 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
9564
9565
Willy Tarreau0ba27502007-12-24 16:55:16 +01009566timeout connect <timeout>
9567timeout contimeout <timeout> (deprecated)
9568 Set the maximum time to wait for a connection attempt to a server to succeed.
9569 May be used in sections : defaults | frontend | listen | backend
9570 yes | no | yes | yes
9571 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01009572 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +01009573 can be in any other unit if the number is suffixed by the unit,
9574 as explained at the top of this document.
9575
9576 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01009577 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +01009578 cover one or several TCP packet losses by specifying timeouts that are
Willy Tarreau0ba27502007-12-24 16:55:16 +01009579 slightly above multiples of 3 seconds (eg: 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +01009580 connect timeout also presets both queue and tarpit timeouts to the same value
9581 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +01009582
9583 This parameter is specific to backends, but can be specified once for all in
9584 "defaults" sections. This is in fact one of the easiest solutions not to
9585 forget about it. An unspecified timeout results in an infinite timeout, which
9586 is not recommended. Such a usage is accepted and works but reports a warning
9587 during startup because it may results in accumulation of failed sessions in
9588 the system if the system's timeouts are not configured either.
9589
9590 This parameter replaces the old, deprecated "contimeout". It is recommended
9591 to use it to write new configurations. The form "timeout contimeout" is
9592 provided only by backwards compatibility but its use is strongly discouraged.
9593
Willy Tarreau41a340d2008-01-22 12:25:31 +01009594 See also: "timeout check", "timeout queue", "timeout server", "contimeout",
9595 "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +01009596
9597
Willy Tarreaub16a5742010-01-10 14:46:16 +01009598timeout http-keep-alive <timeout>
9599 Set the maximum allowed time to wait for a new HTTP request to appear
9600 May be used in sections : defaults | frontend | listen | backend
9601 yes | yes | yes | yes
9602 Arguments :
9603 <timeout> is the timeout value specified in milliseconds by default, but
9604 can be in any other unit if the number is suffixed by the unit,
9605 as explained at the top of this document.
9606
9607 By default, the time to wait for a new request in case of keep-alive is set
9608 by "timeout http-request". However this is not always convenient because some
9609 people want very short keep-alive timeouts in order to release connections
9610 faster, and others prefer to have larger ones but still have short timeouts
9611 once the request has started to present itself.
9612
9613 The "http-keep-alive" timeout covers these needs. It will define how long to
9614 wait for a new HTTP request to start coming after a response was sent. Once
9615 the first byte of request has been seen, the "http-request" timeout is used
9616 to wait for the complete request to come. Note that empty lines prior to a
9617 new request do not refresh the timeout and are not counted as a new request.
9618
9619 There is also another difference between the two timeouts : when a connection
9620 expires during timeout http-keep-alive, no error is returned, the connection
9621 just closes. If the connection expires in "http-request" while waiting for a
9622 connection to complete, a HTTP 408 error is returned.
9623
9624 In general it is optimal to set this value to a few tens to hundreds of
9625 milliseconds, to allow users to fetch all objects of a page at once but
9626 without waiting for further clicks. Also, if set to a very small value (eg:
9627 1 millisecond) it will probably only accept pipelined requests but not the
9628 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +02009629 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +01009630
9631 If this parameter is not set, the "http-request" timeout applies, and if both
9632 are not set, "timeout client" still applies at the lower level. It should be
9633 set in the frontend to take effect, unless the frontend is in TCP mode, in
9634 which case the HTTP backend's timeout will be used.
9635
9636 See also : "timeout http-request", "timeout client".
9637
9638
Willy Tarreau036fae02008-01-06 13:24:40 +01009639timeout http-request <timeout>
9640 Set the maximum allowed time to wait for a complete HTTP request
9641 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +02009642 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +01009643 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +01009644 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +01009645 can be in any other unit if the number is suffixed by the unit,
9646 as explained at the top of this document.
9647
9648 In order to offer DoS protection, it may be required to lower the maximum
9649 accepted time to receive a complete HTTP request without affecting the client
9650 timeout. This helps protecting against established connections on which
9651 nothing is sent. The client timeout cannot offer a good protection against
9652 this abuse because it is an inactivity timeout, which means that if the
9653 attacker sends one character every now and then, the timeout will not
9654 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +02009655 types, the request will be aborted if it does not complete in time. When the
9656 timeout expires, an HTTP 408 response is sent to the client to inform it
9657 about the problem, and the connection is closed. The logs will report
9658 termination codes "cR". Some recent browsers are having problems with this
9659 standard, well-documented behaviour, so it might be needed to hide the 408
Willy Tarreau0f228a02015-05-01 15:37:53 +02009660 code using "option http-ignore-probes" or "errorfile 408 /dev/null". See
9661 more details in the explanations of the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +01009662
Baptiste Assmanneccdf432015-10-28 13:49:01 +01009663 By default, this timeout only applies to the header part of the request,
9664 and not to any data. As soon as the empty line is received, this timeout is
9665 not used anymore. When combined with "option http-buffer-request", this
9666 timeout also applies to the body of the request..
9667 It is used again on keep-alive connections to wait for a second
Willy Tarreaub16a5742010-01-10 14:46:16 +01009668 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +01009669
9670 Generally it is enough to set it to a few seconds, as most clients send the
9671 full request immediately upon connection. Add 3 or more seconds to cover TCP
9672 retransmits but that's all. Setting it to very low values (eg: 50 ms) will
9673 generally work on local networks as long as there are no packet losses. This
9674 will prevent people from sending bare HTTP requests using telnet.
9675
9676 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +02009677 chunk of the incoming request. It should be set in the frontend to take
9678 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
9679 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +01009680
Willy Tarreau0f228a02015-05-01 15:37:53 +02009681 See also : "errorfile", "http-ignore-probes", "timeout http-keep-alive", and
Baptiste Assmanneccdf432015-10-28 13:49:01 +01009682 "timeout client", "option http-buffer-request".
Willy Tarreau036fae02008-01-06 13:24:40 +01009683
Willy Tarreau844e3c52008-01-11 16:28:18 +01009684
9685timeout queue <timeout>
9686 Set the maximum time to wait in the queue for a connection slot to be free
9687 May be used in sections : defaults | frontend | listen | backend
9688 yes | no | yes | yes
9689 Arguments :
9690 <timeout> is the timeout value specified in milliseconds by default, but
9691 can be in any other unit if the number is suffixed by the unit,
9692 as explained at the top of this document.
9693
9694 When a server's maxconn is reached, connections are left pending in a queue
9695 which may be server-specific or global to the backend. In order not to wait
9696 indefinitely, a timeout is applied to requests pending in the queue. If the
9697 timeout is reached, it is considered that the request will almost never be
9698 served, so it is dropped and a 503 error is returned to the client.
9699
9700 The "timeout queue" statement allows to fix the maximum time for a request to
9701 be left pending in a queue. If unspecified, the same value as the backend's
9702 connection timeout ("timeout connect") is used, for backwards compatibility
9703 with older versions with no "timeout queue" parameter.
9704
9705 See also : "timeout connect", "contimeout".
9706
9707
9708timeout server <timeout>
9709timeout srvtimeout <timeout> (deprecated)
9710 Set the maximum inactivity time on the server side.
9711 May be used in sections : defaults | frontend | listen | backend
9712 yes | no | yes | yes
9713 Arguments :
9714 <timeout> is the timeout value specified in milliseconds by default, but
9715 can be in any other unit if the number is suffixed by the unit,
9716 as explained at the top of this document.
9717
9718 The inactivity timeout applies when the server is expected to acknowledge or
9719 send data. In HTTP mode, this timeout is particularly important to consider
9720 during the first phase of the server's response, when it has to send the
9721 headers, as it directly represents the server's processing time for the
9722 request. To find out what value to put there, it's often good to start with
9723 what would be considered as unacceptable response times, then check the logs
9724 to observe the response time distribution, and adjust the value accordingly.
9725
9726 The value is specified in milliseconds by default, but can be in any other
9727 unit if the number is suffixed by the unit, as specified at the top of this
9728 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
9729 recommended that the client timeout remains equal to the server timeout in
9730 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01009731 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +01009732 packet losses by specifying timeouts that are slightly above multiples of 3
Willy Tarreauce887fd2012-05-12 12:50:00 +02009733 seconds (eg: 4 or 5 seconds minimum). If some long-lived sessions are mixed
9734 with short-lived sessions (eg: WebSocket and HTTP), it's worth considering
9735 "timeout tunnel", which overrides "timeout client" and "timeout server" for
9736 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009737
9738 This parameter is specific to backends, but can be specified once for all in
9739 "defaults" sections. This is in fact one of the easiest solutions not to
9740 forget about it. An unspecified timeout results in an infinite timeout, which
9741 is not recommended. Such a usage is accepted and works but reports a warning
9742 during startup because it may results in accumulation of expired sessions in
9743 the system if the system's timeouts are not configured either.
9744
9745 This parameter replaces the old, deprecated "srvtimeout". It is recommended
9746 to use it to write new configurations. The form "timeout srvtimeout" is
9747 provided only by backwards compatibility but its use is strongly discouraged.
9748
Willy Tarreauce887fd2012-05-12 12:50:00 +02009749 See also : "srvtimeout", "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +01009750
Willy Tarreau05cdd962014-05-10 14:30:07 +02009751
9752timeout server-fin <timeout>
9753 Set the inactivity timeout on the server side for half-closed connections.
9754 May be used in sections : defaults | frontend | listen | backend
9755 yes | no | yes | yes
9756 Arguments :
9757 <timeout> is the timeout value specified in milliseconds by default, but
9758 can be in any other unit if the number is suffixed by the unit,
9759 as explained at the top of this document.
9760
9761 The inactivity timeout applies when the server is expected to acknowledge or
9762 send data while one direction is already shut down. This timeout is different
9763 from "timeout server" in that it only applies to connections which are closed
9764 in one direction. This is particularly useful to avoid keeping connections in
9765 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
9766 This problem is particularly common long connections such as RDP or WebSocket.
9767 Note that this timeout can override "timeout tunnel" when a connection shuts
9768 down in one direction. This setting was provided for completeness, but in most
9769 situations, it should not be needed.
9770
9771 This parameter is specific to backends, but can be specified once for all in
9772 "defaults" sections. By default it is not set, so half-closed connections
9773 will use the other timeouts (timeout.server or timeout.tunnel).
9774
9775 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
9776
Willy Tarreau844e3c52008-01-11 16:28:18 +01009777
9778timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +01009779 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +01009780 May be used in sections : defaults | frontend | listen | backend
9781 yes | yes | yes | yes
9782 Arguments :
9783 <timeout> is the tarpit duration specified in milliseconds by default, but
9784 can be in any other unit if the number is suffixed by the unit,
9785 as explained at the top of this document.
9786
9787 When a connection is tarpitted using "reqtarpit", it is maintained open with
9788 no activity for a certain amount of time, then closed. "timeout tarpit"
9789 defines how long it will be maintained open.
9790
9791 The value is specified in milliseconds by default, but can be in any other
9792 unit if the number is suffixed by the unit, as specified at the top of this
9793 document. If unspecified, the same value as the backend's connection timeout
9794 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +01009795 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009796
9797 See also : "timeout connect", "contimeout".
9798
9799
Willy Tarreauce887fd2012-05-12 12:50:00 +02009800timeout tunnel <timeout>
9801 Set the maximum inactivity time on the client and server side for tunnels.
9802 May be used in sections : defaults | frontend | listen | backend
9803 yes | no | yes | yes
9804 Arguments :
9805 <timeout> is the timeout value specified in milliseconds by default, but
9806 can be in any other unit if the number is suffixed by the unit,
9807 as explained at the top of this document.
9808
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009809 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +02009810 between a client and a server, and the connection remains inactive in both
9811 directions. This timeout supersedes both the client and server timeouts once
9812 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
9813 analyser remains attached to either connection (eg: tcp content rules are
9814 accepted). In HTTP, this timeout is used when a connection is upgraded (eg:
9815 when switching to the WebSocket protocol, or forwarding a CONNECT request
9816 to a proxy), or after the first response when no keepalive/close option is
9817 specified.
9818
Willy Tarreau05cdd962014-05-10 14:30:07 +02009819 Since this timeout is usually used in conjunction with long-lived connections,
9820 it usually is a good idea to also set "timeout client-fin" to handle the
9821 situation where a client suddenly disappears from the net and does not
9822 acknowledge a close, or sends a shutdown and does not acknowledge pending
9823 data anymore. This can happen in lossy networks where firewalls are present,
9824 and is detected by the presence of large amounts of sessions in a FIN_WAIT
9825 state.
9826
Willy Tarreauce887fd2012-05-12 12:50:00 +02009827 The value is specified in milliseconds by default, but can be in any other
9828 unit if the number is suffixed by the unit, as specified at the top of this
9829 document. Whatever the expected normal idle time, it is a good practice to
9830 cover at least one or several TCP packet losses by specifying timeouts that
9831 are slightly above multiples of 3 seconds (eg: 4 or 5 seconds minimum).
9832
9833 This parameter is specific to backends, but can be specified once for all in
9834 "defaults" sections. This is in fact one of the easiest solutions not to
9835 forget about it.
9836
9837 Example :
9838 defaults http
9839 option http-server-close
9840 timeout connect 5s
9841 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +02009842 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +02009843 timeout server 30s
9844 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
9845
Willy Tarreau05cdd962014-05-10 14:30:07 +02009846 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +02009847
9848
Willy Tarreau844e3c52008-01-11 16:28:18 +01009849transparent (deprecated)
9850 Enable client-side transparent proxying
9851 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01009852 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01009853 Arguments : none
9854
9855 This keyword was introduced in order to provide layer 7 persistence to layer
9856 3 load balancers. The idea is to use the OS's ability to redirect an incoming
9857 connection for a remote address to a local process (here HAProxy), and let
9858 this process know what address was initially requested. When this option is
9859 used, sessions without cookies will be forwarded to the original destination
9860 IP address of the incoming request (which should match that of another
9861 equipment), while requests with cookies will still be forwarded to the
9862 appropriate server.
9863
9864 The "transparent" keyword is deprecated, use "option transparent" instead.
9865
9866 Note that contrary to a common belief, this option does NOT make HAProxy
9867 present the client's IP to the server when establishing the connection.
9868
Willy Tarreau844e3c52008-01-11 16:28:18 +01009869 See also: "option transparent"
9870
William Lallemanda73203e2012-03-12 12:48:57 +01009871unique-id-format <string>
9872 Generate a unique ID for each request.
9873 May be used in sections : defaults | frontend | listen | backend
9874 yes | yes | yes | no
9875 Arguments :
9876 <string> is a log-format string.
9877
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009878 This keyword creates a ID for each request using the custom log format. A
9879 unique ID is useful to trace a request passing through many components of
9880 a complex infrastructure. The newly created ID may also be logged using the
9881 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +01009882
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009883 The format should be composed from elements that are guaranteed to be
9884 unique when combined together. For instance, if multiple haproxy instances
9885 are involved, it might be important to include the node name. It is often
9886 needed to log the incoming connection's source and destination addresses
9887 and ports. Note that since multiple requests may be performed over the same
9888 connection, including a request counter may help differentiate them.
9889 Similarly, a timestamp may protect against a rollover of the counter.
9890 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +01009891
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009892 It is recommended to use hexadecimal notation for many fields since it
9893 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +01009894
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009895 Example:
William Lallemanda73203e2012-03-12 12:48:57 +01009896
Julien Vehentf21be322014-03-07 08:27:34 -05009897 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +01009898
9899 will generate:
9900
9901 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
9902
9903 See also: "unique-id-header"
9904
9905unique-id-header <name>
9906 Add a unique ID header in the HTTP request.
9907 May be used in sections : defaults | frontend | listen | backend
9908 yes | yes | yes | no
9909 Arguments :
9910 <name> is the name of the header.
9911
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009912 Add a unique-id header in the HTTP request sent to the server, using the
9913 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +01009914
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009915 Example:
William Lallemanda73203e2012-03-12 12:48:57 +01009916
Julien Vehentf21be322014-03-07 08:27:34 -05009917 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +01009918 unique-id-header X-Unique-ID
9919
9920 will generate:
9921
9922 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
9923
9924 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +01009925
Willy Tarreauf51658d2014-04-23 01:21:56 +02009926use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +02009927 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009928 May be used in sections : defaults | frontend | listen | backend
9929 no | yes | yes | no
9930 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01009931 <backend> is the name of a valid backend or "listen" section, or a
9932 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009933
Willy Tarreauf51658d2014-04-23 01:21:56 +02009934 <condition> is a condition composed of ACLs, as described in section 7. If
9935 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009936
9937 When doing content-switching, connections arrive on a frontend and are then
9938 dispatched to various backends depending on a number of conditions. The
9939 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +02009940 "use_backend" keyword. While it is normally used with HTTP processing, it can
9941 also be used in pure TCP, either without content using stateless ACLs (eg:
9942 source address validation) or combined with a "tcp-request" rule to wait for
9943 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009944
9945 There may be as many "use_backend" rules as desired. All of these rules are
9946 evaluated in their declaration order, and the first one which matches will
9947 assign the backend.
9948
9949 In the first form, the backend will be used if the condition is met. In the
9950 second form, the backend will be used if the condition is not met. If no
9951 condition is valid, the backend defined with "default_backend" will be used.
9952 If no default backend is defined, either the servers in the same section are
9953 used (in case of a "listen" section) or, in case of a frontend, no server is
9954 used and a 503 service unavailable response is returned.
9955
Willy Tarreau51aecc72009-07-12 09:47:04 +02009956 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009957 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +02009958 and backend processing will immediately follow, or the backend will wait for
9959 a complete HTTP request to get in. This feature is useful when a frontend
9960 must decode several protocols on a unique port, one of them being HTTP.
9961
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01009962 When <backend> is a simple name, it is resolved at configuration time, and an
9963 error is reported if the specified backend does not exist. If <backend> is
9964 a log-format string instead, no check may be done at configuration time, so
9965 the backend name is resolved dynamically at run time. If the resulting
9966 backend name does not correspond to any valid backend, no other rule is
9967 evaluated, and the default_backend directive is applied instead. Note that
9968 when using dynamic backend names, it is highly recommended to use a prefix
9969 that no other backend uses in order to ensure that an unauthorized backend
9970 cannot be forced from the request.
9971
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009972 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01009973 used to detect the association between frontends and backends to compute the
9974 backend's "fullconn" setting. This cannot be done for dynamic names.
9975
9976 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
9977 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +01009978
Willy Tarreau036fae02008-01-06 13:24:40 +01009979
Willy Tarreau4a5cade2012-04-05 21:09:48 +02009980use-server <server> if <condition>
9981use-server <server> unless <condition>
9982 Only use a specific server if/unless an ACL-based condition is matched.
9983 May be used in sections : defaults | frontend | listen | backend
9984 no | no | yes | yes
9985 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009986 <server> is the name of a valid server in the same backend section.
Willy Tarreau4a5cade2012-04-05 21:09:48 +02009987
9988 <condition> is a condition composed of ACLs, as described in section 7.
9989
9990 By default, connections which arrive to a backend are load-balanced across
9991 the available servers according to the configured algorithm, unless a
9992 persistence mechanism such as a cookie is used and found in the request.
9993
9994 Sometimes it is desirable to forward a particular request to a specific
9995 server without having to declare a dedicated backend for this server. This
9996 can be achieved using the "use-server" rules. These rules are evaluated after
9997 the "redirect" rules and before evaluating cookies, and they have precedence
9998 on them. There may be as many "use-server" rules as desired. All of these
9999 rules are evaluated in their declaration order, and the first one which
10000 matches will assign the server.
10001
10002 If a rule designates a server which is down, and "option persist" is not used
10003 and no force-persist rule was validated, it is ignored and evaluation goes on
10004 with the next rules until one matches.
10005
10006 In the first form, the server will be used if the condition is met. In the
10007 second form, the server will be used if the condition is not met. If no
10008 condition is valid, the processing continues and the server will be assigned
10009 according to other persistence mechanisms.
10010
10011 Note that even if a rule is matched, cookie processing is still performed but
10012 does not assign the server. This allows prefixed cookies to have their prefix
10013 stripped.
10014
10015 The "use-server" statement works both in HTTP and TCP mode. This makes it
10016 suitable for use with content-based inspection. For instance, a server could
10017 be selected in a farm according to the TLS SNI field. And if these servers
10018 have their weight set to zero, they will not be used for other traffic.
10019
10020 Example :
10021 # intercept incoming TLS requests based on the SNI field
10022 use-server www if { req_ssl_sni -i www.example.com }
10023 server www 192.168.0.1:443 weight 0
10024 use-server mail if { req_ssl_sni -i mail.example.com }
10025 server mail 192.168.0.1:587 weight 0
10026 use-server imap if { req_ssl_sni -i imap.example.com }
10027 server mail 192.168.0.1:993 weight 0
10028 # all the rest is forwarded to this server
10029 server default 192.168.0.2:443 check
10030
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010031 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010032
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010033
100345. Bind and Server options
10035--------------------------
10036
10037The "bind", "server" and "default-server" keywords support a number of settings
10038depending on some build options and on the system HAProxy was built on. These
10039settings generally each consist in one word sometimes followed by a value,
10040written on the same line as the "bind" or "server" line. All these options are
10041described in this section.
10042
10043
100445.1. Bind options
10045-----------------
10046
10047The "bind" keyword supports a certain number of settings which are all passed
10048as arguments on the same line. The order in which those arguments appear makes
10049no importance, provided that they appear after the bind address. All of these
10050parameters are optional. Some of them consist in a single words (booleans),
10051while other ones expect a value after them. In this case, the value must be
10052provided immediately after the setting name.
10053
10054The currently supported settings are the following ones.
10055
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010056accept-netscaler-cip <magic number>
10057 Enforces the use of the NetScaler Client IP insertion protocol over any
10058 connection accepted by any of the TCP sockets declared on the same line. The
10059 NetScaler Client IP insertion protocol dictates the layer 3/4 addresses of
10060 the incoming connection to be used everywhere an address is used, with the
10061 only exception of "tcp-request connection" rules which will only see the
10062 real connection address. Logs will reflect the addresses indicated in the
10063 protocol, unless it is violated, in which case the real address will still
10064 be used. This keyword combined with support from external components can be
10065 used as an efficient and reliable alternative to the X-Forwarded-For
Bertrand Jacquin90759682016-06-06 15:35:39 +010010066 mechanism which is not always reliable and not even always usable. See also
10067 "tcp-request connection expect-netscaler-cip" for a finer-grained setting of
10068 which client is allowed to use the protocol.
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010069
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010070accept-proxy
10071 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +020010072 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
10073 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010074 3/4 addresses of the incoming connection to be used everywhere an address is
10075 used, with the only exception of "tcp-request connection" rules which will
10076 only see the real connection address. Logs will reflect the addresses
10077 indicated in the protocol, unless it is violated, in which case the real
10078 address will still be used. This keyword combined with support from external
10079 components can be used as an efficient and reliable alternative to the
10080 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +020010081 usable. See also "tcp-request connection expect-proxy" for a finer-grained
10082 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010083
Willy Tarreauab861d32013-04-02 02:30:41 +020010084alpn <protocols>
10085 This enables the TLS ALPN extension and advertises the specified protocol
10086 list as supported on top of ALPN. The protocol list consists in a comma-
10087 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
10088 quotes). This requires that the SSL library is build with support for TLS
10089 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
10090 initial NPN extension.
10091
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010092backlog <backlog>
10093 Sets the socket's backlog to this value. If unspecified, the frontend's
10094 backlog is used instead, which generally defaults to the maxconn value.
10095
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010010096curves <curves>
10097 This setting is only available when support for OpenSSL was built in. It sets
10098 the string describing the list of elliptic curves algorithms ("curve suite")
10099 that are negotiated during the SSL/TLS handshake with ECDHE. The format of the
10100 string is a colon-delimited list of curve name.
10101 Example: "X25519:P-256" (without quote)
10102 When "curves" is set, "ecdhe" parameter is ignored.
10103
Emeric Brun7fb34422012-09-28 15:26:15 +020010104ecdhe <named curve>
10105 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +010010106 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
10107 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +020010108
Emeric Brunfd33a262012-10-11 16:28:27 +020010109ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +020010110 This setting is only available when support for OpenSSL was built in. It
10111 designates a PEM file from which to load CA certificates used to verify
10112 client's certificate.
10113
Emeric Brunb6dc9342012-09-28 17:55:37 +020010114ca-ignore-err [all|<errorID>,...]
10115 This setting is only available when support for OpenSSL was built in.
10116 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
10117 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
10118 error is ignored.
10119
Christopher Faulet31af49d2015-06-09 17:29:50 +020010120ca-sign-file <cafile>
10121 This setting is only available when support for OpenSSL was built in. It
10122 designates a PEM file containing both the CA certificate and the CA private
10123 key used to create and sign server's certificates. This is a mandatory
10124 setting when the dynamic generation of certificates is enabled. See
10125 'generate-certificates' for details.
10126
Bertrand Jacquind4d0a232016-11-13 16:37:12 +000010127ca-sign-pass <passphrase>
Christopher Faulet31af49d2015-06-09 17:29:50 +020010128 This setting is only available when support for OpenSSL was built in. It is
10129 the CA private key passphrase. This setting is optional and used only when
10130 the dynamic generation of certificates is enabled. See
10131 'generate-certificates' for details.
10132
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010133ciphers <ciphers>
10134 This setting is only available when support for OpenSSL was built in. It sets
10135 the string describing the list of cipher algorithms ("cipher suite") that are
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010136 negotiated during the SSL/TLS handshake. The format of the string is defined
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010137 in "man 1 ciphers" from OpenSSL man pages, and can be for instance a string
10138 such as "AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH" (without quotes).
10139
Emeric Brunfd33a262012-10-11 16:28:27 +020010140crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +020010141 This setting is only available when support for OpenSSL was built in. It
10142 designates a PEM file from which to load certificate revocation list used
10143 to verify client's certificate.
10144
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010145crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +000010146 This setting is only available when support for OpenSSL was built in. It
10147 designates a PEM file containing both the required certificates and any
10148 associated private keys. This file can be built by concatenating multiple
10149 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
10150 requires an intermediate certificate, this can also be concatenated into this
10151 file.
10152
10153 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
10154 are loaded.
10155
10156 If a directory name is used instead of a PEM file, then all files found in
Cyril Bonté3180f7b2015-01-25 00:16:08 +010010157 that directory will be loaded in alphabetic order unless their name ends with
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010010158 '.issuer', '.ocsp' or '.sctl' (reserved extensions). This directive may be
10159 specified multiple times in order to load certificates from multiple files or
10160 directories. The certificates will be presented to clients who provide a
10161 valid TLS Server Name Indication field matching one of their CN or alt
10162 subjects. Wildcards are supported, where a wildcard character '*' is used
10163 instead of the first hostname component (eg: *.example.org matches
10164 www.example.org but not www.sub.example.org).
Alex Davies0fbf0162013-03-02 16:04:50 +000010165
10166 If no SNI is provided by the client or if the SSL library does not support
10167 TLS extensions, or if the client provides an SNI hostname which does not
10168 match any certificate, then the first loaded certificate will be presented.
10169 This means that when loading certificates from a directory, it is highly
Cyril Bonté3180f7b2015-01-25 00:16:08 +010010170 recommended to load the default one first as a file or to ensure that it will
10171 always be the first one in the directory.
Alex Davies0fbf0162013-03-02 16:04:50 +000010172
Emeric Brune032bfa2012-09-28 13:01:45 +020010173 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010174
Alex Davies0fbf0162013-03-02 16:04:50 +000010175 Some CAs (such as Godaddy) offer a drop down list of server types that do not
10176 include HAProxy when obtaining a certificate. If this happens be sure to
Godbach8bf60a12014-04-21 21:42:41 +080010177 choose a webserver that the CA believes requires an intermediate CA (for
Alex Davies0fbf0162013-03-02 16:04:50 +000010178 Godaddy, selection Apache Tomcat will get the correct bundle, but many
10179 others, e.g. nginx, result in a wrong bundle that will not work for some
10180 clients).
10181
Emeric Brun4147b2e2014-06-16 18:36:30 +020010182 For each PEM file, haproxy checks for the presence of file at the same path
10183 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
10184 Status Request extension (also known as "OCSP stapling") is automatically
10185 enabled. The content of this file is optional. If not empty, it must contain
10186 a valid OCSP Response in DER format. In order to be valid an OCSP Response
10187 must comply with the following rules: it has to indicate a good status,
10188 it has to be a single response for the certificate of the PEM file, and it
10189 has to be valid at the moment of addition. If these rules are not respected
10190 the OCSP Response is ignored and a warning is emitted. In order to identify
10191 which certificate an OCSP Response applies to, the issuer's certificate is
10192 necessary. If the issuer's certificate is not found in the PEM file, it will
10193 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
10194 if it exists otherwise it will fail with an error.
10195
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010010196 For each PEM file, haproxy also checks for the presence of file at the same
10197 path suffixed by ".sctl". If such file is found, support for Certificate
10198 Transparency (RFC6962) TLS extension is enabled. The file must contain a
10199 valid Signed Certificate Timestamp List, as described in RFC. File is parsed
10200 to check basic syntax, but no signatures are verified.
10201
yanbzhu6c25e9e2016-01-05 12:52:02 -050010202 There are cases where it is desirable to support multiple key types, e.g. RSA
10203 and ECDSA in the cipher suites offered to the clients. This allows clients
10204 that support EC certificates to be able to use EC ciphers, while
10205 simultaneously supporting older, RSA only clients.
yanbzhud19630c2015-12-14 15:10:25 -050010206
10207 In order to provide this functionality, multiple PEM files, each with a
10208 different key type, are required. To associate these PEM files into a
10209 "cert bundle" that is recognized by haproxy, they must be named in the
10210 following way: All PEM files that are to be bundled must have the same base
10211 name, with a suffix indicating the key type. Currently, three suffixes are
10212 supported: rsa, dsa and ecdsa. For example, if www.example.com has two PEM
10213 files, an RSA file and an ECDSA file, they must be named: "example.pem.rsa"
10214 and "example.pem.ecdsa". The first part of the filename is arbitrary; only the
10215 suffix matters. To load this bundle into haproxy, specify the base name only:
10216
10217 Example : bind :8443 ssl crt example.pem
10218
yanbzhu6c25e9e2016-01-05 12:52:02 -050010219 Note that the suffix is not given to haproxy; this tells haproxy to look for
yanbzhud19630c2015-12-14 15:10:25 -050010220 a cert bundle.
10221
10222 Haproxy will load all PEM files in the bundle at the same time to try to
10223 support multiple key types. PEM files are combined based on Common Name
10224 (CN) and Subject Alternative Name (SAN) to support SNI lookups. This means
10225 that even if you give haproxy a cert bundle, if there are no shared CN/SAN
10226 entries in the certificates in that bundle, haproxy will not be able to
10227 provide multi-cert support.
10228
10229 Assuming bundle in the example above contained the following:
10230
10231 Filename | CN | SAN
10232 -------------------+-----------------+-------------------
10233 example.pem.rsa | www.example.com | rsa.example.com
yanbzhu6c25e9e2016-01-05 12:52:02 -050010234 -------------------+-----------------+-------------------
yanbzhud19630c2015-12-14 15:10:25 -050010235 example.pem.ecdsa | www.example.com | ecdsa.example.com
10236 -------------------+-----------------+-------------------
10237
10238 Users connecting with an SNI of "www.example.com" will be able
10239 to use both RSA and ECDSA cipher suites. Users connecting with an SNI of
10240 "rsa.example.com" will only be able to use RSA cipher suites, and users
10241 connecting with "ecdsa.example.com" will only be able to use ECDSA cipher
10242 suites.
10243
10244 If a directory name is given as the <cert> argument, haproxy will
10245 automatically search and load bundled files in that directory.
10246
10247 OSCP files (.ocsp) and issuer files (.issuer) are supported with multi-cert
10248 bundling. Each certificate can have its own .ocsp and .issuer file. At this
10249 time, sctl is not supported in multi-certificate bundling.
10250
Emeric Brunb6dc9342012-09-28 17:55:37 +020010251crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +000010252 This setting is only available when support for OpenSSL was built in. Sets a
10253 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010254 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +000010255 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +020010256
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010257crt-list <file>
10258 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010259 designates a list of PEM file with an optional ssl configuration and a SNI
10260 filter per certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010261
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010262 <crtfile> [\[<sslbindconf> ...\]] [[!]<snifilter> ...]
10263
10264 sslbindconf support "npn", "alpn", "verify", "ca_file", "crl_file", "ecdhe",
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010010265 "curves", "ciphers" configuration and ssl options (see "ssl-default-bind-options").
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010266 It override the configuration set in bind line for the certificate.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010267
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020010268 Wildcards are supported in the SNI filter. Negative filter are also supported,
10269 only useful in combination with a wildcard filter to exclude a particular SNI.
10270 The certificates will be presented to clients who provide a valid TLS Server
10271 Name Indication field matching one of the SNI filters. If no SNI filter is
10272 specified, the CN and alt subjects are used. This directive may be specified
10273 multiple times. See the "crt" option for more information. The default
10274 certificate is still needed to meet OpenSSL expectations. If it is not used,
10275 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010010276
yanbzhu6c25e9e2016-01-05 12:52:02 -050010277 Multi-cert bundling (see "crt") is supported with crt-list, as long as only
Emmanuel Hocdetd294aea2016-05-13 11:14:06 +020010278 the base name is given in the crt-list. SNI filter will do the same work on
10279 all bundled certificates.
yanbzhud19630c2015-12-14 15:10:25 -050010280
Emmanuel Hocdet98263292016-12-29 18:26:15 +010010281 crt-list file example:
10282 cert1.pem
10283 cert2.pem [npn h2,http/1.1]
10284 certW.pem *.domain.tld !secure.domain.tld
10285 certS.pem [ecdhe secp521r1 ciphers ECDHE-ECDSA-AES256-GCM-SHA384] secure.domain.tld
10286
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010287defer-accept
10288 Is an optional keyword which is supported only on certain Linux kernels. It
10289 states that a connection will only be accepted once some data arrive on it,
10290 or at worst after the first retransmit. This should be used only on protocols
10291 for which the client talks first (eg: HTTP). It can slightly improve
10292 performance by ensuring that most of the request is already available when
10293 the connection is accepted. On the other hand, it will not be able to detect
10294 connections which don't talk. It is important to note that this option is
10295 broken in all kernels up to 2.6.31, as the connection is never accepted until
10296 the client talks. This can cause issues with front firewalls which would see
10297 an established connection while the proxy will only see it in SYN_RECV. This
10298 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
10299
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010300force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010301 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010302 this listener. SSLv3 is generally less expensive than the TLS counterparts
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010303 for high connection rates. This option is also available on global statement
10304 "ssl-default-bind-options". See also "no-tlsv*" and "no-sslv3".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010305
10306force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010307 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010308 this listener. This option is also available on global statement
10309 "ssl-default-bind-options". See also "no-tlsv*" and "no-sslv3".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010310
10311force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010312 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010313 this listener. This option is also available on global statement
10314 "ssl-default-bind-options". See also "no-tlsv*", and "no-sslv3".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010315
10316force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010317 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010318 this listener. This option is also available on global statement
10319 "ssl-default-bind-options". See also "no-tlsv*", and "no-sslv3".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010320
Christopher Faulet31af49d2015-06-09 17:29:50 +020010321generate-certificates
10322 This setting is only available when support for OpenSSL was built in. It
10323 enables the dynamic SSL certificates generation. A CA certificate and its
10324 private key are necessary (see 'ca-sign-file'). When HAProxy is configured as
10325 a transparent forward proxy, SSL requests generate errors because of a common
10326 name mismatch on the certificate presented to the client. With this option
10327 enabled, HAProxy will try to forge a certificate using the SNI hostname
10328 indicated by the client. This is done only if no certificate matches the SNI
10329 hostname (see 'crt-list'). If an error occurs, the default certificate is
10330 used, else the 'strict-sni' option is set.
10331 It can also be used when HAProxy is configured as a reverse proxy to ease the
10332 deployment of an architecture with many backends.
10333
10334 Creating a SSL certificate is an expensive operation, so a LRU cache is used
10335 to store forged certificates (see 'tune.ssl.ssl-ctx-cache-size'). It
10336 increases the HAProxy's memroy footprint to reduce latency when the same
10337 certificate is used many times.
10338
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010339gid <gid>
10340 Sets the group of the UNIX sockets to the designated system gid. It can also
10341 be set by default in the global section's "unix-bind" statement. Note that
10342 some platforms simply ignore this. This setting is equivalent to the "group"
10343 setting except that the group ID is used instead of its name. This setting is
10344 ignored by non UNIX sockets.
10345
10346group <group>
10347 Sets the group of the UNIX sockets to the designated system group. It can
10348 also be set by default in the global section's "unix-bind" statement. Note
10349 that some platforms simply ignore this. This setting is equivalent to the
10350 "gid" setting except that the group name is used instead of its gid. This
10351 setting is ignored by non UNIX sockets.
10352
10353id <id>
10354 Fixes the socket ID. By default, socket IDs are automatically assigned, but
10355 sometimes it is more convenient to fix them to ease monitoring. This value
10356 must be strictly positive and unique within the listener/frontend. This
10357 option can only be used when defining only a single socket.
10358
10359interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +010010360 Restricts the socket to a specific interface. When specified, only packets
10361 received from that particular interface are processed by the socket. This is
10362 currently only supported on Linux. The interface must be a primary system
10363 interface, not an aliased interface. It is also possible to bind multiple
10364 frontends to the same address if they are bound to different interfaces. Note
10365 that binding to a network interface requires root privileges. This parameter
10366 is only compatible with TCPv4/TCPv6 sockets.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010367
Willy Tarreauabb175f2012-09-24 12:43:26 +020010368level <level>
10369 This setting is used with the stats sockets only to restrict the nature of
10370 the commands that can be issued on the socket. It is ignored by other
10371 sockets. <level> can be one of :
10372 - "user" is the least privileged level ; only non-sensitive stats can be
10373 read, and no change is allowed. It would make sense on systems where it
10374 is not easy to restrict access to the socket.
10375 - "operator" is the default level and fits most common uses. All data can
10376 be read, and only non-sensitive changes are permitted (eg: clear max
10377 counters).
10378 - "admin" should be used with care, as everything is permitted (eg: clear
10379 all counters).
10380
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010381maxconn <maxconn>
10382 Limits the sockets to this number of concurrent connections. Extraneous
10383 connections will remain in the system's backlog until a connection is
10384 released. If unspecified, the limit will be the same as the frontend's
10385 maxconn. Note that in case of port ranges or multiple addresses, the same
10386 value will be applied to each socket. This setting enables different
10387 limitations on expensive sockets, for instance SSL entries which may easily
10388 eat all memory.
10389
10390mode <mode>
10391 Sets the octal mode used to define access permissions on the UNIX socket. It
10392 can also be set by default in the global section's "unix-bind" statement.
10393 Note that some platforms simply ignore this. This setting is ignored by non
10394 UNIX sockets.
10395
10396mss <maxseg>
10397 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
10398 connections. This can be used to force a lower MSS for certain specific
10399 ports, for instance for connections passing through a VPN. Note that this
10400 relies on a kernel feature which is theoretically supported under Linux but
10401 was buggy in all versions prior to 2.6.28. It may or may not work on other
10402 operating systems. It may also not change the advertised value but change the
10403 effective size of outgoing segments. The commonly advertised value for TCPv4
10404 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
10405 positive, it will be used as the advertised MSS. If it is negative, it will
10406 indicate by how much to reduce the incoming connection's advertised MSS for
10407 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
10408
10409name <name>
10410 Sets an optional name for these sockets, which will be reported on the stats
10411 page.
10412
Willy Tarreaud72f0f32015-10-13 14:50:22 +020010413namespace <name>
10414 On Linux, it is possible to specify which network namespace a socket will
10415 belong to. This directive makes it possible to explicitly bind a listener to
10416 a namespace different from the default one. Please refer to your operating
10417 system's documentation to find more details about network namespaces.
10418
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010419nice <nice>
10420 Sets the 'niceness' of connections initiated from the socket. Value must be
10421 in the range -1024..1024 inclusive, and defaults to zero. Positive values
10422 means that such connections are more friendly to others and easily offer
10423 their place in the scheduler. On the opposite, negative values mean that
10424 connections want to run with a higher priority than others. The difference
10425 only happens under high loads when the system is close to saturation.
10426 Negative values are appropriate for low-latency or administration services,
10427 and high values are generally recommended for CPU intensive tasks such as SSL
10428 processing or bulk transfers which are less sensible to latency. For example,
10429 it may make sense to use a positive value for an SMTP socket and a negative
10430 one for an RDP socket.
10431
Emeric Brun9b3009b2012-10-05 11:55:06 +020010432no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010433 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010434 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010435 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010436 be enabled using any configuration option. This option is also available on
10437 global statement "ssl-default-bind-options". See also "force-tls*",
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010438 and "force-sslv3".
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010439
Emeric Brun90ad8722012-10-02 14:00:59 +020010440no-tls-tickets
10441 This setting is only available when support for OpenSSL was built in. It
10442 disables the stateless session resumption (RFC 5077 TLS Ticket
10443 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010444 session resumption is more expensive in CPU usage. This option is also
10445 available on global statement "ssl-default-bind-options".
Emeric Brun90ad8722012-10-02 14:00:59 +020010446
Emeric Brun9b3009b2012-10-05 11:55:06 +020010447no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010448 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010449 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010450 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010451 cannot be enabled using any configuration option. This option is also
10452 available on global statement "ssl-default-bind-options". See also
10453 "force-tlsv*", and "force-sslv3".
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010454
Emeric Brun9b3009b2012-10-05 11:55:06 +020010455no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +020010456 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010457 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010458 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010459 cannot be enabled using any configuration option. This option is also
10460 available on global statement "ssl-default-bind-options". See also
10461 "force-tlsv*", and "force-sslv3".
Emeric Brunf5da4932012-09-28 19:42:54 +020010462
Emeric Brun9b3009b2012-10-05 11:55:06 +020010463no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +020010464 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010465 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020010466 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010467 cannot be enabled using any configuration option. This option is also
10468 available on global statement "ssl-default-bind-options". See also
10469 "force-tlsv*", and "force-sslv3".
Emeric Brunf5da4932012-09-28 19:42:54 +020010470
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020010471npn <protocols>
10472 This enables the NPN TLS extension and advertises the specified protocol list
10473 as supported on top of NPN. The protocol list consists in a comma-delimited
10474 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
10475 This requires that the SSL library is build with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +020010476 enabled (check with haproxy -vv). Note that the NPN extension has been
10477 replaced with the ALPN extension (see the "alpn" keyword).
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020010478
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020010479process [ all | odd | even | <number 1-64>[-<number 1-64>] ]
10480 This restricts the list of processes on which this listener is allowed to
10481 run. It does not enforce any process but eliminates those which do not match.
10482 If the frontend uses a "bind-process" setting, the intersection between the
10483 two is applied. If in the end the listener is not allowed to run on any
10484 remaining process, a warning is emitted, and the listener will either run on
10485 the first process of the listener if a single process was specified, or on
10486 all of its processes if multiple processes were specified. For the unlikely
Willy Tarreauae302532014-05-07 19:22:24 +020010487 case where several ranges are needed, this directive may be repeated. The
10488 main purpose of this directive is to be used with the stats sockets and have
10489 one different socket per process. The second purpose is to have multiple bind
10490 lines sharing the same IP:port but not the same process in a listener, so
10491 that the system can distribute the incoming connections into multiple queues
10492 and allow a smoother inter-process load balancing. Currently Linux 3.9 and
10493 above is known for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020010494
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010495ssl
10496 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010497 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010498 certificate is necessary (see "crt" above). All contents in the buffers will
10499 appear in clear text, so that ACLs and HTTP processing will only have access
10500 to deciphered contents.
10501
Emmanuel Hocdet65623372013-01-24 17:17:15 +010010502strict-sni
10503 This setting is only available when support for OpenSSL was built in. The
10504 SSL/TLS negotiation is allow only if the client provided an SNI which match
10505 a certificate. The default certificate is not used.
10506 See the "crt" option for more information.
10507
Willy Tarreau2af207a2015-02-04 00:45:58 +010010508tcp-ut <delay>
Tim Düsterhus4896c442016-11-29 02:15:19 +010010509 Sets the TCP User Timeout for all incoming connections instantiated from this
Willy Tarreau2af207a2015-02-04 00:45:58 +010010510 listening socket. This option is available on Linux since version 2.6.37. It
10511 allows haproxy to configure a timeout for sockets which contain data not
Tim Düsterhus4896c442016-11-29 02:15:19 +010010512 receiving an acknowledgement for the configured delay. This is especially
Willy Tarreau2af207a2015-02-04 00:45:58 +010010513 useful on long-lived connections experiencing long idle periods such as
10514 remote terminals or database connection pools, where the client and server
10515 timeouts must remain high to allow a long period of idle, but where it is
10516 important to detect that the client has disappeared in order to release all
10517 resources associated with its connection (and the server's session). The
10518 argument is a delay expressed in milliseconds by default. This only works
10519 for regular TCP connections, and is ignored for other protocols.
10520
Willy Tarreau1c862c52012-10-05 16:21:00 +020010521tfo
Lukas Tribus0defb902013-02-13 23:35:39 +010010522 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +020010523 enables TCP Fast Open on the listening socket, which means that clients which
10524 support this feature will be able to send a request and receive a response
10525 during the 3-way handshake starting from second connection, thus saving one
10526 round-trip after the first connection. This only makes sense with protocols
10527 that use high connection rates and where each round trip matters. This can
10528 possibly cause issues with many firewalls which do not accept data on SYN
10529 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +020010530 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
10531 need to build HAProxy with USE_TFO=1 if your libc doesn't define
10532 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +020010533
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010010534tls-ticket-keys <keyfile>
10535 Sets the TLS ticket keys file to load the keys from. The keys need to be 48
10536 bytes long, encoded with base64 (ex. openssl rand -base64 48). Number of keys
10537 is specified by the TLS_TICKETS_NO build option (default 3) and at least as
10538 many keys need to be present in the file. Last TLS_TICKETS_NO keys will be
10539 used for decryption and the penultimate one for encryption. This enables easy
10540 key rotation by just appending new key to the file and reloading the process.
10541 Keys must be periodically rotated (ex. every 12h) or Perfect Forward Secrecy
10542 is compromised. It is also a good idea to keep the keys off any permanent
10543 storage such as hard drives (hint: use tmpfs and don't swap those files).
10544 Lifetime hint can be changed using tune.ssl.timeout.
10545
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010546transparent
10547 Is an optional keyword which is supported only on certain Linux kernels. It
10548 indicates that the addresses will be bound even if they do not belong to the
10549 local machine, and that packets targeting any of these addresses will be
10550 intercepted just as if the addresses were locally configured. This normally
10551 requires that IP forwarding is enabled. Caution! do not use this with the
10552 default address '*', as it would redirect any traffic for the specified port.
10553 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
10554 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
10555 kernel version. Some distribution kernels include backports of the feature,
10556 so check for support with your vendor.
10557
Willy Tarreau77e3af92012-11-24 15:07:23 +010010558v4v6
10559 Is an optional keyword which is supported only on most recent systems
10560 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
10561 and IPv6 when it uses the default address. Doing so is sometimes necessary
10562 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010563 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +010010564
Willy Tarreau9b6700f2012-11-24 11:55:28 +010010565v6only
10566 Is an optional keyword which is supported only on most recent systems
10567 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
10568 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +010010569 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
10570 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +010010571
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010572uid <uid>
10573 Sets the owner of the UNIX sockets to the designated system uid. It can also
10574 be set by default in the global section's "unix-bind" statement. Note that
10575 some platforms simply ignore this. This setting is equivalent to the "user"
10576 setting except that the user numeric ID is used instead of its name. This
10577 setting is ignored by non UNIX sockets.
10578
10579user <user>
10580 Sets the owner of the UNIX sockets to the designated system user. It can also
10581 be set by default in the global section's "unix-bind" statement. Note that
10582 some platforms simply ignore this. This setting is equivalent to the "uid"
10583 setting except that the user name is used instead of its uid. This setting is
10584 ignored by non UNIX sockets.
10585
Emeric Brun1a073b42012-09-28 17:07:34 +020010586verify [none|optional|required]
10587 This setting is only available when support for OpenSSL was built in. If set
10588 to 'none', client certificate is not requested. This is the default. In other
10589 cases, a client certificate is requested. If the client does not provide a
10590 certificate after the request and if 'verify' is set to 'required', then the
10591 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +020010592 certificate provided by the client is always verified using CAs from
10593 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
10594 is aborted, regardless of the 'verify' option, unless the error code exactly
10595 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010596
Willy Tarreaub6205fd2012-09-24 12:27:33 +0200105975.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +010010598------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020010599
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010010600The "server" and "default-server" keywords support a certain number of settings
10601which are all passed as arguments on the server line. The order in which those
10602arguments appear does not count, and they are all optional. Some of those
10603settings are single words (booleans) while others expect one or several values
10604after them. In this case, the values must immediately follow the setting name.
10605Except default-server, all those settings must be specified after the server's
10606address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +020010607
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010608 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010010609 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +020010610
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010611The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +010010612
Willy Tarreauceb4ac92012-04-28 00:41:46 +020010613addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010614 Using the "addr" parameter, it becomes possible to use a different IP address
Baptiste Assmann13f83532016-03-06 23:14:36 +010010615 to send health-checks or to probe the agent-check. On some servers, it may be
10616 desirable to dedicate an IP address to specific component able to perform
10617 complex tests which are more suitable to health-checks than the application.
10618 This parameter is ignored if the "check" parameter is not set. See also the
10619 "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +020010620
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010621 Supported in default-server: No
10622
Simon Hormand60d6912013-11-25 10:46:36 +090010623agent-check
10624 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +010010625 health check. An agent health check is performed by making a TCP connection
10626 to the port set by the "agent-port" parameter and reading an ASCII string.
10627 The string is made of a series of words delimited by spaces, tabs or commas
10628 in any order, optionally terminated by '\r' and/or '\n', each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +090010629
Willy Tarreau81f5d942013-12-09 20:51:51 +010010630 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +090010631 Values in this format will set the weight proportional to the initial
Willy Tarreauc5af3a62014-10-07 15:27:33 +020010632 weight of a server as configured when haproxy starts. Note that a zero
10633 weight is reported on the stats page as "DRAIN" since it has the same
10634 effect on the server (it's removed from the LB farm).
Simon Hormand60d6912013-11-25 10:46:36 +090010635
Nenad Merdanovic174dd372016-04-24 23:10:06 +020010636 - The string "maxconn:" followed by an integer (no space between). Values in
10637 this format will set the maxconn of a server. The maximum number of
10638 connections advertised needs to be multipled by the number of load balancers
10639 and different backends that use this health check to get the total number
10640 of connections the server might receive. Example: maxconn:30
10641
Willy Tarreau81f5d942013-12-09 20:51:51 +010010642 - The word "ready". This will turn the server's administrative state to the
10643 READY mode, thus cancelling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +090010644
Willy Tarreau81f5d942013-12-09 20:51:51 +010010645 - The word "drain". This will turn the server's administrative state to the
10646 DRAIN mode, thus it will not accept any new connections other than those
10647 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +090010648
Willy Tarreau81f5d942013-12-09 20:51:51 +010010649 - The word "maint". This will turn the server's administrative state to the
10650 MAINT mode, thus it will not accept any new connections at all, and health
10651 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +090010652
Willy Tarreau81f5d942013-12-09 20:51:51 +010010653 - The words "down", "failed", or "stopped", optionally followed by a
10654 description string after a sharp ('#'). All of these mark the server's
10655 operating state as DOWN, but since the word itself is reported on the stats
10656 page, the difference allows an administrator to know if the situation was
10657 expected or not : the service may intentionally be stopped, may appear up
10658 but fail some validity tests, or may be seen as down (eg: missing process,
10659 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +090010660
Willy Tarreau81f5d942013-12-09 20:51:51 +010010661 - The word "up" sets back the server's operating state as UP if health checks
10662 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +090010663
Willy Tarreau81f5d942013-12-09 20:51:51 +010010664 Parameters which are not advertised by the agent are not changed. For
10665 example, an agent might be designed to monitor CPU usage and only report a
10666 relative weight and never interact with the operating status. Similarly, an
10667 agent could be designed as an end-user interface with 3 radio buttons
10668 allowing an administrator to change only the administrative state. However,
10669 it is important to consider that only the agent may revert its own actions,
10670 so if a server is set to DRAIN mode or to DOWN state using the agent, the
10671 agent must implement the other equivalent actions to bring the service into
10672 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +090010673
Simon Horman2f1f9552013-11-25 10:46:37 +090010674 Failure to connect to the agent is not considered an error as connectivity
10675 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +010010676 parameter. Warning though, it is not a good idea to stop an agent after it
10677 reports "down", since only an agent reporting "up" will be able to turn the
10678 server up again. Note that the CLI on the Unix stats socket is also able to
Willy Tarreau989222a2016-01-15 10:26:26 +010010679 force an agent's result in order to work around a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +090010680
Willy Tarreau81f5d942013-12-09 20:51:51 +010010681 Requires the "agent-port" parameter to be set. See also the "agent-inter"
10682 parameter.
Simon Hormand60d6912013-11-25 10:46:36 +090010683
10684 Supported in default-server: No
10685
James Brown55f9ff12015-10-21 18:19:05 -070010686agent-send <string>
10687 If this option is specified, haproxy will send the given string (verbatim)
10688 to the agent server upon connection. You could, for example, encode
10689 the backend name into this string, which would enable your agent to send
10690 different responses based on the backend. Make sure to include a '\n' if
10691 you want to terminate your request with a newline.
10692
Simon Hormand60d6912013-11-25 10:46:36 +090010693agent-inter <delay>
10694 The "agent-inter" parameter sets the interval between two agent checks
10695 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
10696
10697 Just as with every other time-based parameter, it may be entered in any
10698 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
10699 parameter also serves as a timeout for agent checks "timeout check" is
10700 not set. In order to reduce "resonance" effects when multiple servers are
10701 hosted on the same hardware, the agent and health checks of all servers
10702 are started with a small time offset between them. It is also possible to
10703 add some random noise in the agent and health checks interval using the
10704 global "spread-checks" keyword. This makes sense for instance when a lot
10705 of backends use the same servers.
10706
10707 See also the "agent-check" and "agent-port" parameters.
10708
10709 Supported in default-server: Yes
10710
Misiek768d8602017-01-09 09:52:43 +010010711agent-addr <addr>
10712 The "agent-addr" parameter sets address for agent check.
10713
10714 You can offload agent-check to another target, so you can make single place
10715 managing status and weights of servers defined in haproxy in case you can't
10716 make self-aware and self-managing services. You can specify both IP or
10717 hostname, it will be resolved.
10718
Simon Hormand60d6912013-11-25 10:46:36 +090010719agent-port <port>
10720 The "agent-port" parameter sets the TCP port used for agent checks.
10721
10722 See also the "agent-check" and "agent-inter" parameters.
10723
10724 Supported in default-server: Yes
10725
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010726backup
10727 When "backup" is present on a server line, the server is only used in load
10728 balancing when all other non-backup servers are unavailable. Requests coming
10729 with a persistence cookie referencing the server will always be served
10730 though. By default, only the first operational backup server is used, unless
10731 the "allbackups" option is set in the backend. See also the "allbackups"
10732 option.
Willy Tarreau6a06a402007-07-15 20:15:28 +020010733
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010734 Supported in default-server: No
10735
Emeric Brunef42d922012-10-11 16:11:36 +020010736ca-file <cafile>
10737 This setting is only available when support for OpenSSL was built in. It
10738 designates a PEM file from which to load CA certificates used to verify
10739 server's certificate.
10740
10741 Supported in default-server: No
10742
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010743check
10744 This option enables health checks on the server. By default, a server is
Patrick Mézardb7aeec62012-01-22 16:01:22 +010010745 always considered available. If "check" is set, the server is available when
10746 accepting periodic TCP connections, to ensure that it is really able to serve
10747 requests. The default address and port to send the tests to are those of the
10748 server, and the default source is the same as the one defined in the
10749 backend. It is possible to change the address using the "addr" parameter, the
10750 port using the "port" parameter, the source address using the "source"
10751 address, and the interval and timers using the "inter", "rise" and "fall"
Simon Hormanafc47ee2013-11-25 10:46:35 +090010752 parameters. The request method is define in the backend using the "httpchk",
10753 "smtpchk", "mysql-check", "pgsql-check" and "ssl-hello-chk" options. Please
10754 refer to those options and parameters for more information.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010755
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010756 Supported in default-server: No
10757
Willy Tarreau6c16adc2012-10-05 00:04:16 +020010758check-send-proxy
10759 This option forces emission of a PROXY protocol line with outgoing health
10760 checks, regardless of whether the server uses send-proxy or not for the
10761 normal traffic. By default, the PROXY protocol is enabled for health checks
10762 if it is already enabled for normal traffic and if no "port" nor "addr"
10763 directive is present. However, if such a directive is present, the
10764 "check-send-proxy" option needs to be used to force the use of the
10765 protocol. See also the "send-proxy" option for more information.
10766
10767 Supported in default-server: No
10768
Willy Tarreau763a95b2012-10-04 23:15:39 +020010769check-ssl
10770 This option forces encryption of all health checks over SSL, regardless of
10771 whether the server uses SSL or not for the normal traffic. This is generally
10772 used when an explicit "port" or "addr" directive is specified and SSL health
10773 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010774 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +020010775 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
10776 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
10777 All SSL settings are common to health checks and traffic (eg: ciphers).
10778 See the "ssl" option for more information.
10779
10780 Supported in default-server: No
10781
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020010782ciphers <ciphers>
10783 This option sets the string describing the list of cipher algorithms that is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010784 is negotiated during the SSL/TLS handshake with the server. The format of the
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020010785 string is defined in "man 1 ciphers". When SSL is used to communicate with
10786 servers on the local network, it is common to see a weaker set of algorithms
10787 than what is used over the internet. Doing so reduces CPU usage on both the
10788 server and haproxy while still keeping it compatible with deployed software.
10789 Some algorithms such as RC4-SHA1 are reasonably cheap. If no security at all
10790 is needed and just connectivity, using DES can be appropriate.
10791
Willy Tarreau763a95b2012-10-04 23:15:39 +020010792 Supported in default-server: No
10793
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010794cookie <value>
10795 The "cookie" parameter sets the cookie value assigned to the server to
10796 <value>. This value will be checked in incoming requests, and the first
10797 operational server possessing the same value will be selected. In return, in
10798 cookie insertion or rewrite modes, this value will be assigned to the cookie
10799 sent to the client. There is nothing wrong in having several servers sharing
10800 the same cookie value, and it is in fact somewhat common between normal and
10801 backup servers. See also the "cookie" keyword in backend section.
10802
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010803 Supported in default-server: No
10804
Emeric Brunef42d922012-10-11 16:11:36 +020010805crl-file <crlfile>
10806 This setting is only available when support for OpenSSL was built in. It
10807 designates a PEM file from which to load certificate revocation list used
10808 to verify server's certificate.
10809
10810 Supported in default-server: No
10811
Emeric Bruna7aa3092012-10-26 12:58:00 +020010812crt <cert>
10813 This setting is only available when support for OpenSSL was built in.
10814 It designates a PEM file from which to load both a certificate and the
10815 associated private key. This file can be built by concatenating both PEM
10816 files into one. This certificate will be sent if the server send a client
10817 certificate request.
10818
10819 Supported in default-server: No
10820
Willy Tarreau96839092010-03-29 10:02:24 +020010821disabled
10822 The "disabled" keyword starts the server in the "disabled" state. That means
10823 that it is marked down in maintenance mode, and no connection other than the
10824 ones allowed by persist mode will reach it. It is very well suited to setup
10825 new servers, because normal traffic will never reach them, while it is still
10826 possible to test the service by making use of the force-persist mechanism.
10827
10828 Supported in default-server: No
10829
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010830error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +010010831 If health observing is enabled, the "error-limit" parameter specifies the
10832 number of consecutive errors that triggers event selected by the "on-error"
10833 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010010834
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010835 Supported in default-server: Yes
10836
10837 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010010838
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010839fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010840 The "fall" parameter states that a server will be considered as dead after
10841 <count> consecutive unsuccessful health checks. This value defaults to 3 if
10842 unspecified. See also the "check", "inter" and "rise" parameters.
10843
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010844 Supported in default-server: Yes
10845
Emeric Brun8694b9a2012-10-05 14:39:07 +020010846force-sslv3
10847 This option enforces use of SSLv3 only when SSL is used to communicate with
10848 the server. SSLv3 is generally less expensive than the TLS counterparts for
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010849 high connection rates. This option is also available on global statement
10850 "ssl-default-server-options". See also "no-tlsv*", "no-sslv3".
Emeric Brun8694b9a2012-10-05 14:39:07 +020010851
10852 Supported in default-server: No
10853
10854force-tlsv10
10855 This option enforces use of TLSv1.0 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010856 the server. This option is also available on global statement
10857 "ssl-default-server-options". See also "no-tlsv*", "no-sslv3".
Emeric Brun8694b9a2012-10-05 14:39:07 +020010858
10859 Supported in default-server: No
10860
10861force-tlsv11
10862 This option enforces use of TLSv1.1 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010863 the server. This option is also available on global statement
10864 "ssl-default-server-options". See also "no-tlsv*", "no-sslv3".
Emeric Brun8694b9a2012-10-05 14:39:07 +020010865
10866 Supported in default-server: No
10867
10868force-tlsv12
10869 This option enforces use of TLSv1.2 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010010870 the server. This option is also available on global statement
10871 "ssl-default-server-options". See also "no-tlsv*", "no-sslv3".
Emeric Brun8694b9a2012-10-05 14:39:07 +020010872
10873 Supported in default-server: No
10874
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010875id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +020010876 Set a persistent ID for the server. This ID must be positive and unique for
10877 the proxy. An unused ID will automatically be assigned if unset. The first
10878 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010879
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010880 Supported in default-server: No
10881
Willy Tarreau6a031d12016-11-07 19:42:35 +010010882init-addr {last | libc | none | <ip>},[...]*
10883 Indicate in what order the server's address should be resolved upon startup
10884 if it uses an FQDN. Attempts are made to resolve the address by applying in
10885 turn each of the methods mentionned in the comma-delimited list. The first
10886 method which succeeds is used. If the end of the list is reached without
10887 finding a working method, an error is thrown. Method "last" suggests to pick
10888 the address which appears in the state file (see "server-state-file"). Method
10889 "libc" uses the libc's internal resolver (gethostbyname() or getaddrinfo()
10890 depending on the operating system and build options). Method "none"
10891 specifically indicates that the server should start without any valid IP
10892 address in a down state. It can be useful to ignore some DNS issues upon
10893 startup, waiting for the situation to get fixed later. Finally, an IP address
10894 (IPv4 or IPv6) may be provided. It can be the currently known address of the
10895 server (eg: filled by a configuration generator), or the address of a dummy
10896 server used to catch old sessions and present them with a decent error
10897 message for example. When the "first" load balancing algorithm is used, this
10898 IP address could point to a fake server used to trigger the creation of new
10899 instances on the fly. This option defaults to "last,libc" indicating that the
10900 previous address found in the state file (if any) is used first, otherwise
10901 the libc's resolver is used. This ensures continued compatibility with the
10902 historic behaviour.
10903
10904 Example:
10905 defaults
10906 # never fail on address resolution
10907 default-server init-addr last,libc,none
10908
10909 Supported in default-server: Yes
10910
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010911inter <delay>
10912fastinter <delay>
10913downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010914 The "inter" parameter sets the interval between two consecutive health checks
10915 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
10916 It is also possible to use "fastinter" and "downinter" to optimize delays
10917 between checks depending on the server state :
10918
Pieter Baauw44fc9df2015-09-17 21:30:46 +020010919 Server state | Interval used
10920 ----------------------------------------+----------------------------------
10921 UP 100% (non-transitional) | "inter"
10922 ----------------------------------------+----------------------------------
10923 Transitionally UP (going down "fall"), | "fastinter" if set,
10924 Transitionally DOWN (going up "rise"), | "inter" otherwise.
10925 or yet unchecked. |
10926 ----------------------------------------+----------------------------------
10927 DOWN 100% (non-transitional) | "downinter" if set,
10928 | "inter" otherwise.
10929 ----------------------------------------+----------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +010010930
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010931 Just as with every other time-based parameter, they can be entered in any
10932 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
10933 serves as a timeout for health checks sent to servers if "timeout check" is
10934 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +090010935 hosted on the same hardware, the agent and health checks of all servers
10936 are started with a small time offset between them. It is also possible to
10937 add some random noise in the agent and health checks interval using the
10938 global "spread-checks" keyword. This makes sense for instance when a lot
10939 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010940
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010941 Supported in default-server: Yes
10942
10943maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010944 The "maxconn" parameter specifies the maximal number of concurrent
10945 connections that will be sent to this server. If the number of incoming
10946 concurrent requests goes higher than this value, they will be queued, waiting
10947 for a connection to be released. This parameter is very important as it can
10948 save fragile servers from going down under extreme loads. If a "minconn"
10949 parameter is specified, the limit becomes dynamic. The default value is "0"
10950 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
10951 the backend's "fullconn" keyword.
10952
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010953 Supported in default-server: Yes
10954
10955maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010956 The "maxqueue" parameter specifies the maximal number of connections which
10957 will wait in the queue for this server. If this limit is reached, next
10958 requests will be redispatched to other servers instead of indefinitely
10959 waiting to be served. This will break persistence but may allow people to
10960 quickly re-log in when the server they try to connect to is dying. The
10961 default value is "0" which means the queue is unlimited. See also the
10962 "maxconn" and "minconn" parameters.
10963
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010964 Supported in default-server: Yes
10965
10966minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010967 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
10968 limit following the backend's load. The server will always accept at least
10969 <minconn> connections, never more than <maxconn>, and the limit will be on
10970 the ramp between both values when the backend has less than <fullconn>
10971 concurrent connections. This makes it possible to limit the load on the
10972 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010973 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010974 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010010975
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010010976 Supported in default-server: Yes
10977
Willy Tarreaud72f0f32015-10-13 14:50:22 +020010978namespace <name>
10979 On Linux, it is possible to specify which network namespace a socket will
10980 belong to. This directive makes it possible to explicitly bind a server to
10981 a namespace different from the default one. Please refer to your operating
10982 system's documentation to find more details about network namespaces.
10983
Willy Tarreau2a3fb1c2015-02-05 16:47:07 +010010984no-ssl-reuse
10985 This option disables SSL session reuse when SSL is used to communicate with
10986 the server. It will force the server to perform a full handshake for every
10987 new connection. It's probably only useful for benchmarking, troubleshooting,
10988 and for paranoid users.
10989
10990 Supported in default-server: No
10991
Emeric Brun9b3009b2012-10-05 11:55:06 +020010992no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020010993 This option disables support for SSLv3 when SSL is used to communicate with
10994 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emeric Brun8694b9a2012-10-05 14:39:07 +020010995 using any configuration option. See also "force-sslv3", "force-tlsv*".
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020010996
Willy Tarreau763a95b2012-10-04 23:15:39 +020010997 Supported in default-server: No
10998
Emeric Brunf9c5c472012-10-11 15:28:34 +020010999no-tls-tickets
11000 This setting is only available when support for OpenSSL was built in. It
11001 disables the stateless session resumption (RFC 5077 TLS Ticket
11002 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011003 session resumption is more expensive in CPU usage for servers. This option
11004 is also available on global statement "ssl-default-server-options".
Emeric Brunf9c5c472012-10-11 15:28:34 +020011005
11006 Supported in default-server: No
11007
Emeric Brun9b3009b2012-10-05 11:55:06 +020011008no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +020011009 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020011010 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11011 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011012 often makes sense to disable it when communicating with local servers. This
11013 option is also available on global statement "ssl-default-server-options".
11014 See also "force-sslv3", "force-tlsv*".
Emeric Brunf5da4932012-09-28 19:42:54 +020011015
Willy Tarreau763a95b2012-10-04 23:15:39 +020011016 Supported in default-server: No
11017
Emeric Brun9b3009b2012-10-05 11:55:06 +020011018no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +020011019 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020011020 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11021 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011022 often makes sense to disable it when communicating with local servers. This
11023 option is also available on global statement "ssl-default-server-options".
11024 See also "force-sslv3", "force-tlsv*".
Emeric Brunf5da4932012-09-28 19:42:54 +020011025
Willy Tarreau763a95b2012-10-04 23:15:39 +020011026 Supported in default-server: No
11027
Emeric Brun9b3009b2012-10-05 11:55:06 +020011028no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +020011029 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011030 the server. Note that SSLv2 is disabled in the code and cannot be enabled
11031 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011032 often makes sense to disable it when communicating with local servers. This
11033 option is also available on global statement "ssl-default-server-options".
11034 See also "force-sslv3", "force-tlsv*".
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011035
Willy Tarreau763a95b2012-10-04 23:15:39 +020011036 Supported in default-server: No
11037
Simon Hormanfa461682011-06-25 09:39:49 +090011038non-stick
11039 Never add connections allocated to this sever to a stick-table.
11040 This may be used in conjunction with backup to ensure that
11041 stick-table persistence is disabled for backup servers.
11042
Willy Tarreau763a95b2012-10-04 23:15:39 +020011043 Supported in default-server: No
11044
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011045observe <mode>
11046 This option enables health adjusting based on observing communication with
11047 the server. By default this functionality is disabled and enabling it also
11048 requires to enable health checks. There are two supported modes: "layer4" and
11049 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
11050 significant. In layer7, which is only allowed for http proxies, responses
11051 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +010011052 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011053
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011054 Supported in default-server: No
11055
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011056 See also the "check", "on-error" and "error-limit".
11057
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011058on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011059 Select what should happen when enough consecutive errors are detected.
11060 Currently, four modes are available:
11061 - fastinter: force fastinter
11062 - fail-check: simulate a failed check, also forces fastinter (default)
11063 - sudden-death: simulate a pre-fatal failed health check, one more failed
11064 check will mark a server down, forces fastinter
11065 - mark-down: mark the server immediately down and force fastinter
11066
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011067 Supported in default-server: Yes
11068
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011069 See also the "check", "observe" and "error-limit".
11070
Simon Hormane0d1bfb2011-06-21 14:34:58 +090011071on-marked-down <action>
11072 Modify what occurs when a server is marked down.
11073 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070011074 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
11075 all connections to the server are immediately terminated when the server
11076 goes down. It might be used if the health check detects more complex cases
11077 than a simple connection status, and long timeouts would cause the service
11078 to remain unresponsive for too long a time. For instance, a health check
11079 might detect that a database is stuck and that there's no chance to reuse
11080 existing connections anymore. Connections killed this way are logged with
11081 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +090011082
11083 Actions are disabled by default
11084
11085 Supported in default-server: Yes
11086
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070011087on-marked-up <action>
11088 Modify what occurs when a server is marked up.
11089 Currently one action is available:
11090 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
11091 done only if the server is not in backup state and if it is not disabled
11092 (it must have an effective weight > 0). This can be used sometimes to force
11093 an active server to take all the traffic back after recovery when dealing
11094 with long sessions (eg: LDAP, SQL, ...). Doing this can cause more trouble
11095 than it tries to solve (eg: incomplete transactions), so use this feature
11096 with extreme care. Sessions killed because a server comes up are logged
11097 with an 'U' termination code (for "Up").
11098
11099 Actions are disabled by default
11100
11101 Supported in default-server: Yes
11102
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011103port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011104 Using the "port" parameter, it becomes possible to use a different port to
11105 send health-checks. On some servers, it may be desirable to dedicate a port
11106 to a specific component able to perform complex tests which are more suitable
11107 to health-checks than the application. It is common to run a simple script in
11108 inetd for instance. This parameter is ignored if the "check" parameter is not
11109 set. See also the "addr" parameter.
11110
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011111 Supported in default-server: Yes
11112
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011113redir <prefix>
11114 The "redir" parameter enables the redirection mode for all GET and HEAD
11115 requests addressing this server. This means that instead of having HAProxy
11116 forward the request to the server, it will send an "HTTP 302" response with
11117 the "Location" header composed of this prefix immediately followed by the
11118 requested URI beginning at the leading '/' of the path component. That means
11119 that no trailing slash should be used after <prefix>. All invalid requests
11120 will be rejected, and all non-GET or HEAD requests will be normally served by
11121 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011122 mangling nor cookie insertion is possible in the response. However, cookies in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011123 requests are still analysed, making this solution completely usable to direct
11124 users to a remote location in case of local disaster. Main use consists in
11125 increasing bandwidth for static servers by having the clients directly
11126 connect to them. Note: never use a relative location here, it would cause a
11127 loop between the client and HAProxy!
11128
11129 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
11130
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011131 Supported in default-server: No
11132
11133rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011134 The "rise" parameter states that a server will be considered as operational
11135 after <count> consecutive successful health checks. This value defaults to 2
11136 if unspecified. See also the "check", "inter" and "fall" parameters.
11137
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011138 Supported in default-server: Yes
11139
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011140resolve-prefer <family>
11141 When DNS resolution is enabled for a server and multiple IP addresses from
11142 different families are returned, HAProxy will prefer using an IP address
11143 from the family mentioned in the "resolve-prefer" parameter.
11144 Available families: "ipv4" and "ipv6"
11145
Baptiste Assmannc4aabae2015-08-04 22:43:06 +020011146 Default value: ipv6
11147
11148 Supported in default-server: Yes
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011149
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011150 Example:
11151
11152 server s1 app1.domain.com:80 resolvers mydns resolve-prefer ipv6
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011153
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011154resolve-net <network>[,<network[,...]]
11155 This options prioritize th choice of an ip address matching a network. This is
11156 useful with clouds to prefer a local ip. In some cases, a cloud high
Tim Düsterhus4896c442016-11-29 02:15:19 +010011157 availability service can be announced with many ip addresses on many
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011158 differents datacenters. The latency between datacenter is not negligible, so
11159 this patch permitsto prefers a local datacenter. If none address matchs the
11160 configured network, another address is selected.
11161
11162 Supported in default-server: Yes
11163
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011164 Example:
11165
11166 server s1 app1.domain.com:80 resolvers mydns resolve-net 10.0.0.0/8
Thierry Fournierac88cfe2016-02-17 22:05:30 +010011167
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011168resolvers <id>
11169 Points to an existing "resolvers" section to resolve current server's
11170 hostname.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011171 In order to be operational, DNS resolution requires that health check is
11172 enabled on the server. Actually, health checks triggers the DNS resolution.
11173 You must precise one 'resolvers' parameter on each server line where DNS
11174 resolution is required.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011175
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011176 Supported in default-server: No
11177
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011178 Example:
11179
11180 server s1 app1.domain.com:80 check resolvers mydns
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011181
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011182 See also section 5.3
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011183
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010011184send-proxy
11185 The "send-proxy" parameter enforces use of the PROXY protocol over any
11186 connection established to this server. The PROXY protocol informs the other
11187 end about the layer 3/4 addresses of the incoming connection, so that it can
11188 know the client's address or the public address it accessed to, whatever the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010011189 upper layer protocol. For connections accepted by an "accept-proxy" or
11190 "accept-netscaler-cip" listener, the advertised address will be used. Only
11191 TCPv4 and TCPv6 address families are supported. Other families such as
11192 Unix sockets, will report an UNKNOWN family. Servers using this option can
11193 fully be chained to another instance of haproxy listening with an
11194 "accept-proxy" setting. This setting must not be used if the server isn't
11195 aware of the protocol. When health checks are sent to the server, the PROXY
11196 protocol is automatically used when this option is set, unless there is an
11197 explicit "port" or "addr" directive, in which case an explicit
11198 "check-send-proxy" directive would also be needed to use the PROXY protocol.
11199 See also the "accept-proxy" and "accept-netscaler-cip" option of the "bind"
11200 keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010011201
11202 Supported in default-server: No
11203
David Safb76832014-05-08 23:42:08 -040011204send-proxy-v2
11205 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
11206 over any connection established to this server. The PROXY protocol informs
11207 the other end about the layer 3/4 addresses of the incoming connection, so
11208 that it can know the client's address or the public address it accessed to,
11209 whatever the upper layer protocol. This setting must not be used if the
11210 server isn't aware of this version of the protocol. See also the "send-proxy"
11211 option of the "bind" keyword.
11212
11213 Supported in default-server: No
11214
11215send-proxy-v2-ssl
11216 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
11217 2 over any connection established to this server. The PROXY protocol informs
11218 the other end about the layer 3/4 addresses of the incoming connection, so
11219 that it can know the client's address or the public address it accessed to,
11220 whatever the upper layer protocol. In addition, the SSL information extension
11221 of the PROXY protocol is added to the PROXY protocol header. This setting
11222 must not be used if the server isn't aware of this version of the protocol.
11223 See also the "send-proxy-v2" option of the "bind" keyword.
11224
11225 Supported in default-server: No
11226
11227send-proxy-v2-ssl-cn
11228 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
11229 2 over any connection established to this server. The PROXY protocol informs
11230 the other end about the layer 3/4 addresses of the incoming connection, so
11231 that it can know the client's address or the public address it accessed to,
11232 whatever the upper layer protocol. In addition, the SSL information extension
11233 of the PROXY protocol, along along with the Common Name from the subject of
11234 the client certificate (if any), is added to the PROXY protocol header. This
11235 setting must not be used if the server isn't aware of this version of the
11236 protocol. See also the "send-proxy-v2" option of the "bind" keyword.
11237
11238 Supported in default-server: No
11239
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011240slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011241 The "slowstart" parameter for a server accepts a value in milliseconds which
11242 indicates after how long a server which has just come back up will run at
11243 full speed. Just as with every other time-based parameter, it can be entered
11244 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
11245 linearly from 0 to 100% during this time. The limitation applies to two
11246 parameters :
11247
11248 - maxconn: the number of connections accepted by the server will grow from 1
11249 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
11250
11251 - weight: when the backend uses a dynamic weighted algorithm, the weight
11252 grows linearly from 1 to 100%. In this case, the weight is updated at every
11253 health-check. For this reason, it is important that the "inter" parameter
11254 is smaller than the "slowstart", in order to maximize the number of steps.
11255
11256 The slowstart never applies when haproxy starts, otherwise it would cause
11257 trouble to running servers. It only applies when a server has been previously
11258 seen as failed.
11259
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011260 Supported in default-server: Yes
11261
Willy Tarreau732eac42015-07-09 11:40:25 +020011262sni <expression>
11263 The "sni" parameter evaluates the sample fetch expression, converts it to a
11264 string and uses the result as the host name sent in the SNI TLS extension to
11265 the server. A typical use case is to send the SNI received from the client in
11266 a bridged HTTPS scenario, using the "ssl_fc_sni" sample fetch for the
11267 expression, though alternatives such as req.hdr(host) can also make sense.
11268
11269 Supported in default-server: no
11270
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011271source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020011272source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011273source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011274 The "source" parameter sets the source address which will be used when
11275 connecting to the server. It follows the exact same parameters and principle
11276 as the backend "source" keyword, except that it only applies to the server
11277 referencing it. Please consult the "source" keyword for details.
11278
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020011279 Additionally, the "source" statement on a server line allows one to specify a
11280 source port range by indicating the lower and higher bounds delimited by a
11281 dash ('-'). Some operating systems might require a valid IP address when a
11282 source port range is specified. It is permitted to have the same IP/range for
11283 several servers. Doing so makes it possible to bypass the maximum of 64k
11284 total concurrent connections. The limit will then reach 64k connections per
11285 server.
11286
Lukas Tribus7d56c6d2016-09-13 09:51:15 +000011287 Since Linux 4.2/libc 2.23 IP_BIND_ADDRESS_NO_PORT is set for connections
11288 specifying the source address without port(s).
11289
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011290 Supported in default-server: No
11291
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011292ssl
Willy Tarreau44f65392013-06-25 07:56:20 +020011293 This option enables SSL ciphering on outgoing connections to the server. It
11294 is critical to verify server certificates using "verify" when using SSL to
11295 connect to servers, otherwise the communication is prone to trivial man in
11296 the-middle attacks rendering SSL useless. When this option is used, health
11297 checks are automatically sent in SSL too unless there is a "port" or an
11298 "addr" directive indicating the check should be sent to a different location.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011299 See the "check-ssl" option to force SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011300
11301 Supported in default-server: No
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011302
Willy Tarreau163d4622015-10-13 16:16:41 +020011303tcp-ut <delay>
11304 Sets the TCP User Timeout for all outgoing connections to this server. This
11305 option is available on Linux since version 2.6.37. It allows haproxy to
11306 configure a timeout for sockets which contain data not receiving an
Tim Düsterhus4896c442016-11-29 02:15:19 +010011307 acknowledgement for the configured delay. This is especially useful on
Willy Tarreau163d4622015-10-13 16:16:41 +020011308 long-lived connections experiencing long idle periods such as remote
11309 terminals or database connection pools, where the client and server timeouts
11310 must remain high to allow a long period of idle, but where it is important to
11311 detect that the server has disappeared in order to release all resources
11312 associated with its connection (and the client's session). One typical use
11313 case is also to force dead server connections to die when health checks are
11314 too slow or during a soft reload since health checks are then disabled. The
11315 argument is a delay expressed in milliseconds by default. This only works for
11316 regular TCP connections, and is ignored for other protocols.
11317
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011318track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +020011319 This option enables ability to set the current state of the server by tracking
11320 another one. It is possible to track a server which itself tracks another
11321 server, provided that at the end of the chain, a server has health checks
11322 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011323 used, it has to be enabled on both proxies.
11324
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011325 Supported in default-server: No
11326
Emeric Brunef42d922012-10-11 16:11:36 +020011327verify [none|required]
11328 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +010011329 to 'none', server certificate is not verified. In the other case, The
11330 certificate provided by the server is verified using CAs from 'ca-file'
11331 and optional CRLs from 'crl-file'. If 'ssl_server_verify' is not specified
11332 in global section, this is the default. On verify failure the handshake
Willy Tarreau44f65392013-06-25 07:56:20 +020011333 is aborted. It is critically important to verify server certificates when
11334 using SSL to connect to servers, otherwise the communication is prone to
11335 trivial man-in-the-middle attacks rendering SSL totally useless.
Emeric Brunef42d922012-10-11 16:11:36 +020011336
11337 Supported in default-server: No
11338
Evan Broderbe554312013-06-27 00:05:25 -070011339verifyhost <hostname>
11340 This setting is only available when support for OpenSSL was built in, and
11341 only takes effect if 'verify required' is also specified. When set, the
11342 hostnames in the subject and subjectAlternateNames of the certificate
11343 provided by the server are checked. If none of the hostnames in the
11344 certificate match the specified hostname, the handshake is aborted. The
11345 hostnames in the server-provided certificate may include wildcards.
11346
11347 Supported in default-server: No
11348
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011349weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011350 The "weight" parameter is used to adjust the server's weight relative to
11351 other servers. All servers will receive a load proportional to their weight
11352 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +020011353 load. The default weight is 1, and the maximal value is 256. A value of 0
11354 means the server will not participate in load-balancing but will still accept
11355 persistent connections. If this parameter is used to distribute the load
11356 according to server's capacity, it is recommended to start with values which
11357 can both grow and shrink, for instance between 10 and 100 to leave enough
11358 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011359
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011360 Supported in default-server: Yes
11361
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011362
Cyril Bonté46175dd2015-07-02 22:45:32 +0200113635.3. Server IP address resolution using DNS
11364-------------------------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011365
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011366HAProxy allows using a host name on the server line to retrieve its IP address
11367using name servers. By default, HAProxy resolves the name when parsing the
11368configuration file, at startup and cache the result for the process' life.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011369This is not sufficient in some cases, such as in Amazon where a server's IP
11370can change after a reboot or an ELB Virtual IP can change based on current
11371workload.
11372This chapter describes how HAProxy can be configured to process server's name
11373resolution at run time.
11374Whether run time server name resolution has been enable or not, HAProxy will
11375carry on doing the first resolution when parsing the configuration.
11376
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011377Bear in mind that DNS resolution is triggered by health checks. This makes
11378health checks mandatory to allow DNS resolution.
11379
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011380
Cyril Bonté46175dd2015-07-02 22:45:32 +0200113815.3.1. Global overview
11382----------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011383
11384As we've seen in introduction, name resolution in HAProxy occurs at two
11385different steps of the process life:
11386
11387 1. when starting up, HAProxy parses the server line definition and matches a
11388 host name. It uses libc functions to get the host name resolved. This
11389 resolution relies on /etc/resolv.conf file.
11390
11391 2. at run time, when HAProxy gets prepared to run a health check on a server,
11392 it verifies if the current name resolution is still considered as valid.
11393 If not, it processes a new resolution, in parallel of the health check.
11394
11395A few other events can trigger a name resolution at run time:
11396 - when a server's health check ends up in a connection timeout: this may be
11397 because the server has a new IP address. So we need to trigger a name
11398 resolution to know this new IP.
11399
11400A few things important to notice:
11401 - all the name servers are queried in the mean time. HAProxy will process the
11402 first valid response.
11403
11404 - a resolution is considered as invalid (NX, timeout, refused), when all the
11405 servers return an error.
11406
11407
Cyril Bonté46175dd2015-07-02 22:45:32 +0200114085.3.2. The resolvers section
11409----------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011410
11411This section is dedicated to host information related to name resolution in
11412HAProxy.
11413There can be as many as resolvers section as needed. Each section can contain
11414many name servers.
11415
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011416When multiple name servers are configured in a resolvers section, then HAProxy
11417uses the first valid response. In case of invalid responses, only the last one
11418is treated. Purpose is to give the chance to a slow server to deliver a valid
11419answer after a fast faulty or outdated server.
11420
11421When each server returns a different error type, then only the last error is
11422used by HAProxy to decide what type of behavior to apply.
11423
11424Two types of behavior can be applied:
11425 1. stop DNS resolution
11426 2. replay the DNS query with a new query type
11427 In such case, the following types are applied in this exact order:
11428 1. ANY query type
11429 2. query type corresponding to family pointed by resolve-prefer
11430 server's parameter
11431 3. remaining family type
11432
11433HAProxy stops DNS resolution when the following errors occur:
11434 - invalid DNS response packet
11435 - wrong name in the query section of the response
11436 - NX domain
11437 - Query refused by server
11438 - CNAME not pointing to an IP address
11439
11440HAProxy tries a new query type when the following errors occur:
11441 - no Answer records in the response
11442 - DNS response truncated
11443 - Error in DNS response
11444 - No expected DNS records found in the response
11445 - name server timeout
11446
11447For example, with 2 name servers configured in a resolvers section:
11448 - first response is valid and is applied directly, second response is ignored
11449 - first response is invalid and second one is valid, then second response is
11450 applied;
11451 - first response is a NX domain and second one a truncated response, then
11452 HAProxy replays the query with a new type;
11453 - first response is truncated and second one is a NX Domain, then HAProxy
11454 stops resolution.
11455
11456
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011457resolvers <resolvers id>
11458 Creates a new name server list labelled <resolvers id>
11459
11460A resolvers section accept the following parameters:
11461
11462nameserver <id> <ip>:<port>
11463 DNS server description:
11464 <id> : label of the server, should be unique
11465 <ip> : IP address of the server
11466 <port> : port where the DNS service actually runs
11467
11468hold <status> <period>
11469 Defines <period> during which the last name resolution should be kept based
11470 on last resolution <status>
Baptiste Assmann987e16d2016-11-02 22:23:31 +010011471 <status> : last name resolution status. Acceptable values are "nx",
11472 "other", "refused", "timeout", "valid".
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011473 <period> : interval between two successive name resolution when the last
11474 answer was in <status>. It follows the HAProxy time format.
11475 <period> is in milliseconds by default.
11476
Baptiste Assmann987e16d2016-11-02 22:23:31 +010011477 Default value is 10s for "valid" and 30s for others.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011478
11479 Note: since the name resolution is triggered by the health checks, a new
11480 resolution is triggered after <period> modulo the <inter> parameter of
11481 the healch check.
11482
11483resolve_retries <nb>
11484 Defines the number <nb> of queries to send to resolve a server name before
11485 giving up.
11486 Default value: 3
11487
Baptiste Assmann62b75b42015-09-09 01:11:36 +020011488 A retry occurs on name server timeout or when the full sequence of DNS query
11489 type failover is over and we need to start up from the default ANY query
11490 type.
11491
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011492timeout <event> <time>
11493 Defines timeouts related to name resolution
11494 <event> : the event on which the <time> timeout period applies to.
11495 events available are:
11496 - retry: time between two DNS queries, when no response have
11497 been received.
11498 Default value: 1s
11499 <time> : time related to the event. It follows the HAProxy time format.
11500 <time> is expressed in milliseconds.
11501
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020011502 Example:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011503
11504 resolvers mydns
11505 nameserver dns1 10.0.0.1:53
11506 nameserver dns2 10.0.0.2:53
11507 resolve_retries 3
11508 timeout retry 1s
Baptiste Assmann987e16d2016-11-02 22:23:31 +010011509 hold other 30s
11510 hold refused 30s
11511 hold nx 30s
11512 hold timeout 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020011513 hold valid 10s
11514
11515
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200115166. HTTP header manipulation
11517---------------------------
11518
11519In HTTP mode, it is possible to rewrite, add or delete some of the request and
11520response headers based on regular expressions. It is also possible to block a
11521request or a response if a particular header matches a regular expression,
11522which is enough to stop most elementary protocol attacks, and to protect
Willy Tarreau70dffda2014-01-30 03:07:23 +010011523against information leak from the internal network.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011524
Willy Tarreau70dffda2014-01-30 03:07:23 +010011525If HAProxy encounters an "Informational Response" (status code 1xx), it is able
11526to process all rsp* rules which can allow, deny, rewrite or delete a header,
11527but it will refuse to add a header to any such messages as this is not
11528HTTP-compliant. The reason for still processing headers in such responses is to
11529stop and/or fix any possible information leak which may happen, for instance
11530because another downstream equipment would unconditionally add a header, or if
11531a server name appears there. When such messages are seen, normal processing
11532still occurs on the next non-informational messages.
Willy Tarreau816b9792009-09-15 21:25:21 +020011533
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011534This section covers common usage of the following keywords, described in detail
11535in section 4.2 :
11536
11537 - reqadd <string>
11538 - reqallow <search>
11539 - reqiallow <search>
11540 - reqdel <search>
11541 - reqidel <search>
11542 - reqdeny <search>
11543 - reqideny <search>
11544 - reqpass <search>
11545 - reqipass <search>
11546 - reqrep <search> <replace>
11547 - reqirep <search> <replace>
11548 - reqtarpit <search>
11549 - reqitarpit <search>
11550 - rspadd <string>
11551 - rspdel <search>
11552 - rspidel <search>
11553 - rspdeny <search>
11554 - rspideny <search>
11555 - rsprep <search> <replace>
11556 - rspirep <search> <replace>
11557
11558With all these keywords, the same conventions are used. The <search> parameter
11559is a POSIX extended regular expression (regex) which supports grouping through
11560parenthesis (without the backslash). Spaces and other delimiters must be
11561prefixed with a backslash ('\') to avoid confusion with a field delimiter.
11562Other characters may be prefixed with a backslash to change their meaning :
11563
11564 \t for a tab
11565 \r for a carriage return (CR)
11566 \n for a new line (LF)
11567 \ to mark a space and differentiate it from a delimiter
11568 \# to mark a sharp and differentiate it from a comment
11569 \\ to use a backslash in a regex
11570 \\\\ to use a backslash in the text (*2 for regex, *2 for haproxy)
11571 \xXX to write the ASCII hex code XX as in the C language
11572
11573The <replace> parameter contains the string to be used to replace the largest
11574portion of text matching the regex. It can make use of the special characters
11575above, and can reference a substring which is delimited by parenthesis in the
11576regex, by writing a backslash ('\') immediately followed by one digit from 0 to
115779 indicating the group position (0 designating the entire line). This practice
11578is very common to users of the "sed" program.
11579
11580The <string> parameter represents the string which will systematically be added
11581after the last header line. It can also use special character sequences above.
11582
11583Notes related to these keywords :
11584---------------------------------
11585 - these keywords are not always convenient to allow/deny based on header
11586 contents. It is strongly recommended to use ACLs with the "block" keyword
11587 instead, resulting in far more flexible and manageable rules.
11588
11589 - lines are always considered as a whole. It is not possible to reference
11590 a header name only or a value only. This is important because of the way
11591 headers are written (notably the number of spaces after the colon).
11592
11593 - the first line is always considered as a header, which makes it possible to
11594 rewrite or filter HTTP requests URIs or response codes, but in turn makes
11595 it harder to distinguish between headers and request line. The regex prefix
11596 ^[^\ \t]*[\ \t] matches any HTTP method followed by a space, and the prefix
11597 ^[^ \t:]*: matches any header name followed by a colon.
11598
11599 - for performances reasons, the number of characters added to a request or to
11600 a response is limited at build time to values between 1 and 4 kB. This
11601 should normally be far more than enough for most usages. If it is too short
11602 on occasional usages, it is possible to gain some space by removing some
11603 useless headers before adding new ones.
11604
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011605 - keywords beginning with "reqi" and "rspi" are the same as their counterpart
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011606 without the 'i' letter except that they ignore case when matching patterns.
11607
11608 - when a request passes through a frontend then a backend, all req* rules
11609 from the frontend will be evaluated, then all req* rules from the backend
11610 will be evaluated. The reverse path is applied to responses.
11611
11612 - req* statements are applied after "block" statements, so that "block" is
11613 always the first one, but before "use_backend" in order to permit rewriting
Willy Tarreaud72758d2010-01-12 10:42:19 +010011614 before switching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011615
11616
Willy Tarreau74ca5042013-06-11 23:12:07 +0200116177. Using ACLs and fetching samples
11618----------------------------------
11619
11620Haproxy is capable of extracting data from request or response streams, from
11621client or server information, from tables, environmental information etc...
11622The action of extracting such data is called fetching a sample. Once retrieved,
11623these samples may be used for various purposes such as a key to a stick-table,
11624but most common usages consist in matching them against predefined constant
11625data called patterns.
11626
11627
116287.1. ACL basics
11629---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011630
11631The use of Access Control Lists (ACL) provides a flexible solution to perform
11632content switching and generally to take decisions based on content extracted
11633from the request, the response or any environmental status. The principle is
11634simple :
11635
Willy Tarreau74ca5042013-06-11 23:12:07 +020011636 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +010011637 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +020011638 - apply one or multiple pattern matching methods on this sample
11639 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011640
Willy Tarreau74ca5042013-06-11 23:12:07 +020011641The actions generally consist in blocking a request, selecting a backend, or
11642adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011643
11644In order to define a test, the "acl" keyword is used. The syntax is :
11645
Willy Tarreau74ca5042013-06-11 23:12:07 +020011646 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011647
11648This creates a new ACL <aclname> or completes an existing one with new tests.
11649Those tests apply to the portion of request/response specified in <criterion>
11650and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +010011651an operator which may be specified before the set of values. Optionally some
11652conversion operators may be applied to the sample, and they will be specified
11653as a comma-delimited list of keywords just after the first keyword. The values
11654are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011655
11656ACL names must be formed from upper and lower case letters, digits, '-' (dash),
11657'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
11658which means that "my_acl" and "My_Acl" are two different ACLs.
11659
11660There is no enforced limit to the number of ACLs. The unused ones do not affect
11661performance, they just consume a small amount of memory.
11662
Willy Tarreau74ca5042013-06-11 23:12:07 +020011663The criterion generally is the name of a sample fetch method, or one of its ACL
11664specific declinations. The default test method is implied by the output type of
11665this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +010011666methods of a same sample fetch method. The sample fetch methods are the only
11667ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +020011668
11669Sample fetch methods return data which can be of the following types :
11670 - boolean
11671 - integer (signed or unsigned)
11672 - IPv4 or IPv6 address
11673 - string
11674 - data block
11675
Willy Tarreaue6b11e42013-11-26 19:02:32 +010011676Converters transform any of these data into any of these. For example, some
11677converters might convert a string to a lower-case string while other ones
11678would turn a string to an IPv4 address, or apply a netmask to an IP address.
11679The resulting sample is of the type of the last converter applied to the list,
11680which defaults to the type of the sample fetch method.
11681
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011682Each sample or converter returns data of a specific type, specified with its
11683keyword in this documentation. When an ACL is declared using a standard sample
11684fetch method, certain types automatically involved a default matching method
11685which are summarized in the table below :
11686
11687 +---------------------+-----------------+
11688 | Sample or converter | Default |
11689 | output type | matching method |
11690 +---------------------+-----------------+
11691 | boolean | bool |
11692 +---------------------+-----------------+
11693 | integer | int |
11694 +---------------------+-----------------+
11695 | ip | ip |
11696 +---------------------+-----------------+
11697 | string | str |
11698 +---------------------+-----------------+
11699 | binary | none, use "-m" |
11700 +---------------------+-----------------+
11701
11702Note that in order to match a binary samples, it is mandatory to specify a
11703matching method, see below.
11704
Willy Tarreau74ca5042013-06-11 23:12:07 +020011705The ACL engine can match these types against patterns of the following types :
11706 - boolean
11707 - integer or integer range
11708 - IP address / network
11709 - string (exact, substring, suffix, prefix, subdir, domain)
11710 - regular expression
11711 - hex block
11712
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011713The following ACL flags are currently supported :
11714
Willy Tarreau2b5285d2010-05-09 23:45:24 +020011715 -i : ignore case during matching of all subsequent patterns.
11716 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +020011717 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010011718 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +010011719 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +010011720 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +020011721 -- : force end of flags. Useful when a string looks like one of the flags.
11722
Willy Tarreau74ca5042013-06-11 23:12:07 +020011723The "-f" flag is followed by the name of a file from which all lines will be
11724read as individual values. It is even possible to pass multiple "-f" arguments
11725if the patterns are to be loaded from multiple files. Empty lines as well as
11726lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
11727will be stripped. If it is absolutely necessary to insert a valid pattern
11728beginning with a sharp, just prefix it with a space so that it is not taken for
11729a comment. Depending on the data type and match method, haproxy may load the
11730lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
11731exact string matching. In this case, duplicates will automatically be removed.
11732
Thierry FOURNIER9860c412014-01-29 14:23:29 +010011733The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
11734parsed as two column file. The first column contains the patterns used by the
11735ACL, and the second column contain the samples. The sample can be used later by
11736a map. This can be useful in some rare cases where an ACL would just be used to
11737check for the existence of a pattern in a map before a mapping is applied.
11738
Thierry FOURNIER3534d882014-01-20 17:01:44 +010011739The "-u" flag forces the unique id of the ACL. This unique id is used with the
11740socket interface to identify ACL and dynamically change its values. Note that a
11741file is always identified by its name even if an id is set.
11742
Willy Tarreau74ca5042013-06-11 23:12:07 +020011743Also, note that the "-i" flag applies to subsequent entries and not to entries
11744loaded from files preceding it. For instance :
11745
11746 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
11747
11748In this example, each line of "exact-ua.lst" will be exactly matched against
11749the "user-agent" header of the request. Then each line of "generic-ua" will be
11750case-insensitively matched. Then the word "test" will be insensitively matched
11751as well.
11752
11753The "-m" flag is used to select a specific pattern matching method on the input
11754sample. All ACL-specific criteria imply a pattern matching method and generally
11755do not need this flag. However, this flag is useful with generic sample fetch
11756methods to describe how they're going to be matched against the patterns. This
11757is required for sample fetches which return data type for which there is no
11758obvious matching method (eg: string or binary). When "-m" is specified and
11759followed by a pattern matching method name, this method is used instead of the
11760default one for the criterion. This makes it possible to match contents in ways
11761that were not initially planned, or with sample fetch methods which return a
11762string. The matching method also affects the way the patterns are parsed.
11763
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010011764The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
11765By default, if the parser cannot parse ip address it considers that the parsed
11766string is maybe a domain name and try dns resolution. The flag "-n" disable this
11767resolution. It is useful for detecting malformed ip lists. Note that if the DNS
11768server is not reachable, the haproxy configuration parsing may last many minutes
11769waiting fir the timeout. During this time no error messages are displayed. The
11770flag "-n" disable this behavior. Note also that during the runtime, this
11771function is disabled for the dynamic acl modifications.
11772
Willy Tarreau74ca5042013-06-11 23:12:07 +020011773There are some restrictions however. Not all methods can be used with all
11774sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
11775be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +020011776
11777 - "found" : only check if the requested sample could be found in the stream,
11778 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +020011779 to pass any pattern to avoid confusion. This matching method is
11780 particularly useful to detect presence of certain contents such
11781 as headers, cookies, etc... even if they are empty and without
11782 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011783
11784 - "bool" : check the value as a boolean. It can only be applied to fetches
11785 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +020011786 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011787
11788 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +020011789 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011790
11791 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +020011792 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011793
11794 - "bin" : match the contents against an hexadecimal string representing a
11795 binary sequence. This may be used with binary or string samples.
11796
11797 - "len" : match the sample's length as an integer. This may be used with
11798 binary or string samples.
11799
Willy Tarreau74ca5042013-06-11 23:12:07 +020011800 - "str" : exact match : match the contents against a string. This may be
11801 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011802
Willy Tarreau74ca5042013-06-11 23:12:07 +020011803 - "sub" : substring match : check that the contents contain at least one of
11804 the provided string patterns. This may be used with binary or
11805 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011806
Willy Tarreau74ca5042013-06-11 23:12:07 +020011807 - "reg" : regex match : match the contents against a list of regular
11808 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011809
Willy Tarreau74ca5042013-06-11 23:12:07 +020011810 - "beg" : prefix match : check that the contents begin like the provided
11811 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011812
Willy Tarreau74ca5042013-06-11 23:12:07 +020011813 - "end" : suffix match : check that the contents end like the provided
11814 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011815
Willy Tarreau74ca5042013-06-11 23:12:07 +020011816 - "dir" : subdir match : check that a slash-delimited portion of the
11817 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011818 This may be used with binary or string samples.
11819
Willy Tarreau74ca5042013-06-11 23:12:07 +020011820 - "dom" : domain match : check that a dot-delimited portion of the contents
11821 exactly match one of the provided string patterns. This may be
11822 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020011823
11824For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
11825request, it is possible to do :
11826
11827 acl jsess_present cook(JSESSIONID) -m found
11828
11829In order to apply a regular expression on the 500 first bytes of data in the
11830buffer, one would use the following acl :
11831
11832 acl script_tag payload(0,500) -m reg -i <script>
11833
Willy Tarreaue6b11e42013-11-26 19:02:32 +010011834On systems where the regex library is much slower when using "-i", it is
11835possible to convert the sample to lowercase before matching, like this :
11836
11837 acl script_tag payload(0,500),lower -m reg <script>
11838
Willy Tarreau74ca5042013-06-11 23:12:07 +020011839All ACL-specific criteria imply a default matching method. Most often, these
11840criteria are composed by concatenating the name of the original sample fetch
11841method and the matching method. For example, "hdr_beg" applies the "beg" match
11842to samples retrieved using the "hdr" fetch method. Since all ACL-specific
11843criteria rely on a sample fetch method, it is always possible instead to use
11844the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +020011845
Willy Tarreau74ca5042013-06-11 23:12:07 +020011846If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011847the matching method is simply applied to the underlying sample fetch method.
11848For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +020011849
Willy Tarreau74ca5042013-06-11 23:12:07 +020011850 acl short_form hdr_beg(host) www.
11851 acl alternate1 hdr_beg(host) -m beg www.
11852 acl alternate2 hdr_dom(host) -m beg www.
11853 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +020011854
Willy Tarreau2b5285d2010-05-09 23:45:24 +020011855
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011856The table below summarizes the compatibility matrix between sample or converter
11857types and the pattern types to fetch against. It indicates for each compatible
11858combination the name of the matching method to be used, surrounded with angle
11859brackets ">" and "<" when the method is the default one and will work by
11860default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +010011861
Willy Tarreau74ca5042013-06-11 23:12:07 +020011862 +-------------------------------------------------+
11863 | Input sample type |
11864 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011865 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011866 +----------------------+---------+---------+---------+---------+---------+
11867 | none (presence only) | found | found | found | found | found |
11868 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011869 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011870 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011871 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011872 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011873 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011874 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011875 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011876 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020011877 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011878 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011879 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011880 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011881 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011882 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011883 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011884 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011885 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011886 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011887 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011888 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010011889 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020011890 +----------------------+---------+---------+---------+---------+---------+
11891 | hex block | | | | bin | bin |
11892 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +020011893
11894
Willy Tarreau74ca5042013-06-11 23:12:07 +0200118957.1.1. Matching booleans
11896------------------------
11897
11898In order to match a boolean, no value is needed and all values are ignored.
11899Boolean matching is used by default for all fetch methods of type "boolean".
11900When boolean matching is used, the fetched value is returned as-is, which means
11901that a boolean "true" will always match and a boolean "false" will never match.
11902
11903Boolean matching may also be enforced using "-m bool" on fetch methods which
11904return an integer value. Then, integer value 0 is converted to the boolean
11905"false" and all other values are converted to "true".
11906
Willy Tarreau6a06a402007-07-15 20:15:28 +020011907
Willy Tarreau74ca5042013-06-11 23:12:07 +0200119087.1.2. Matching integers
11909------------------------
11910
11911Integer matching applies by default to integer fetch methods. It can also be
11912enforced on boolean fetches using "-m int". In this case, "false" is converted
11913to the integer 0, and "true" is converted to the integer 1.
11914
11915Integer matching also supports integer ranges and operators. Note that integer
11916matching only applies to positive values. A range is a value expressed with a
11917lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011918
11919For instance, "1024:65535" is a valid range to represent a range of
11920unprivileged ports, and "1024:" would also work. "0:1023" is a valid
11921representation of privileged ports, and ":1023" would also work.
11922
Willy Tarreau62644772008-07-16 18:36:06 +020011923As a special case, some ACL functions support decimal numbers which are in fact
11924two integers separated by a dot. This is used with some version checks for
11925instance. All integer properties apply to those decimal numbers, including
11926ranges and operators.
11927
Willy Tarreau6a06a402007-07-15 20:15:28 +020011928For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +010011929operators with ranges does not make much sense and is strongly discouraged.
11930Similarly, it does not make much sense to perform order comparisons with a set
11931of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011932
Willy Tarreau0ba27502007-12-24 16:55:16 +010011933Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +020011934
11935 eq : true if the tested value equals at least one value
11936 ge : true if the tested value is greater than or equal to at least one value
11937 gt : true if the tested value is greater than at least one value
11938 le : true if the tested value is less than or equal to at least one value
11939 lt : true if the tested value is less than at least one value
11940
Willy Tarreau0ba27502007-12-24 16:55:16 +010011941For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +020011942
11943 acl negative-length hdr_val(content-length) lt 0
11944
Willy Tarreau62644772008-07-16 18:36:06 +020011945This one matches SSL versions between 3.0 and 3.1 (inclusive) :
11946
11947 acl sslv3 req_ssl_ver 3:3.1
11948
Willy Tarreau6a06a402007-07-15 20:15:28 +020011949
Willy Tarreau74ca5042013-06-11 23:12:07 +0200119507.1.3. Matching strings
11951-----------------------
11952
11953String matching applies to string or binary fetch methods, and exists in 6
11954different forms :
11955
11956 - exact match (-m str) : the extracted string must exactly match the
11957 patterns ;
11958
11959 - substring match (-m sub) : the patterns are looked up inside the
11960 extracted string, and the ACL matches if any of them is found inside ;
11961
11962 - prefix match (-m beg) : the patterns are compared with the beginning of
11963 the extracted string, and the ACL matches if any of them matches.
11964
11965 - suffix match (-m end) : the patterns are compared with the end of the
11966 extracted string, and the ACL matches if any of them matches.
11967
Baptiste Assmann33db6002016-03-06 23:32:10 +010011968 - subdir match (-m dir) : the patterns are looked up inside the extracted
Willy Tarreau74ca5042013-06-11 23:12:07 +020011969 string, delimited with slashes ("/"), and the ACL matches if any of them
11970 matches.
11971
11972 - domain match (-m dom) : the patterns are looked up inside the extracted
11973 string, delimited with dots ("."), and the ACL matches if any of them
11974 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011975
11976String matching applies to verbatim strings as they are passed, with the
11977exception of the backslash ("\") which makes it possible to escape some
11978characters such as the space. If the "-i" flag is passed before the first
11979string, then the matching will be performed ignoring the case. In order
11980to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +010011981before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020011982
11983
Willy Tarreau74ca5042013-06-11 23:12:07 +0200119847.1.4. Matching regular expressions (regexes)
11985---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020011986
11987Just like with string matching, regex matching applies to verbatim strings as
11988they are passed, with the exception of the backslash ("\") which makes it
11989possible to escape some characters such as the space. If the "-i" flag is
11990passed before the first regex, then the matching will be performed ignoring
11991the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +010011992the "--" flag before the first string. Same principle applies of course to
11993match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020011994
11995
Willy Tarreau74ca5042013-06-11 23:12:07 +0200119967.1.5. Matching arbitrary data blocks
11997-------------------------------------
11998
11999It is possible to match some extracted samples against a binary block which may
12000not safely be represented as a string. For this, the patterns must be passed as
12001a series of hexadecimal digits in an even number, when the match method is set
12002to binary. Each sequence of two digits will represent a byte. The hexadecimal
12003digits may be used upper or lower case.
12004
12005Example :
12006 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
12007 acl hello payload(0,6) -m bin 48656c6c6f0a
12008
12009
120107.1.6. Matching IPv4 and IPv6 addresses
12011---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020012012
12013IPv4 addresses values can be specified either as plain addresses or with a
12014netmask appended, in which case the IPv4 address matches whenever it is
12015within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012016host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +010012017difficult to read and debug configurations. If hostnames are used, you should
12018at least ensure that they are present in /etc/hosts so that the configuration
12019does not depend on any random DNS match at the moment the configuration is
12020parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012021
Daniel Schnellereba56342016-04-13 00:26:52 +020012022The dotted IPv4 address notation is supported in both regular as well as the
12023abbreviated form with all-0-octets omitted:
12024
12025 +------------------+------------------+------------------+
12026 | Example 1 | Example 2 | Example 3 |
12027 +------------------+------------------+------------------+
12028 | 192.168.0.1 | 10.0.0.12 | 127.0.0.1 |
12029 | 192.168.1 | 10.12 | 127.1 |
12030 | 192.168.0.1/22 | 10.0.0.12/8 | 127.0.0.1/8 |
12031 | 192.168.1/22 | 10.12/8 | 127.1/8 |
12032 +------------------+------------------+------------------+
12033
12034Notice that this is different from RFC 4632 CIDR address notation in which
12035192.168.42/24 would be equivalent to 192.168.42.0/24.
12036
Willy Tarreauceb4ac92012-04-28 00:41:46 +020012037IPv6 may be entered in their usual form, with or without a netmask appended.
12038Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
12039trouble with randomly resolved IP addresses, host names are never allowed in
12040IPv6 patterns.
12041
12042HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
12043following situations :
12044 - tested address is IPv4, pattern address is IPv4, the match applies
12045 in IPv4 using the supplied mask if any.
12046 - tested address is IPv6, pattern address is IPv6, the match applies
12047 in IPv6 using the supplied mask if any.
12048 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
12049 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
12050 ::IPV4 or ::ffff:IPV4, otherwise it fails.
12051 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
12052 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
12053 applied in IPv6 using the supplied IPv6 mask.
12054
Willy Tarreau74ca5042013-06-11 23:12:07 +020012055
120567.2. Using ACLs to form conditions
12057----------------------------------
12058
12059Some actions are only performed upon a valid condition. A condition is a
12060combination of ACLs with operators. 3 operators are supported :
12061
12062 - AND (implicit)
12063 - OR (explicit with the "or" keyword or the "||" operator)
12064 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +020012065
Willy Tarreau74ca5042013-06-11 23:12:07 +020012066A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +020012067
Willy Tarreau74ca5042013-06-11 23:12:07 +020012068 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +020012069
Willy Tarreau74ca5042013-06-11 23:12:07 +020012070Such conditions are generally used after an "if" or "unless" statement,
12071indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +020012072
Willy Tarreau74ca5042013-06-11 23:12:07 +020012073For instance, to block HTTP requests to the "*" URL with methods other than
12074"OPTIONS", as well as POST requests without content-length, and GET or HEAD
12075requests with a content-length greater than 0, and finally every request which
12076is not either GET/HEAD/POST/OPTIONS !
12077
12078 acl missing_cl hdr_cnt(Content-length) eq 0
12079 block if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
12080 block if METH_GET HTTP_CONTENT
12081 block unless METH_GET or METH_POST or METH_OPTIONS
12082
12083To select a different backend for requests to static contents on the "www" site
12084and to every request on the "img", "video", "download" and "ftp" hosts :
12085
12086 acl url_static path_beg /static /images /img /css
12087 acl url_static path_end .gif .png .jpg .css .js
12088 acl host_www hdr_beg(host) -i www
12089 acl host_static hdr_beg(host) -i img. video. download. ftp.
12090
12091 # now use backend "static" for all static-only hosts, and for static urls
12092 # of host "www". Use backend "www" for the rest.
12093 use_backend static if host_static or host_www url_static
12094 use_backend www if host_www
12095
12096It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
12097expressions that are built on the fly without needing to be declared. They must
12098be enclosed between braces, with a space before and after each brace (because
12099the braces must be seen as independent words). Example :
12100
12101 The following rule :
12102
12103 acl missing_cl hdr_cnt(Content-length) eq 0
12104 block if METH_POST missing_cl
12105
12106 Can also be written that way :
12107
12108 block if METH_POST { hdr_cnt(Content-length) eq 0 }
12109
12110It is generally not recommended to use this construct because it's a lot easier
12111to leave errors in the configuration when written that way. However, for very
12112simple rules matching only one source IP address for instance, it can make more
12113sense to use them than to declare ACLs with random names. Another example of
12114good use is the following :
12115
12116 With named ACLs :
12117
12118 acl site_dead nbsrv(dynamic) lt 2
12119 acl site_dead nbsrv(static) lt 2
12120 monitor fail if site_dead
12121
12122 With anonymous ACLs :
12123
12124 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
12125
12126See section 4.2 for detailed help on the "block" and "use_backend" keywords.
12127
12128
121297.3. Fetching samples
12130---------------------
12131
12132Historically, sample fetch methods were only used to retrieve data to match
12133against patterns using ACLs. With the arrival of stick-tables, a new class of
12134sample fetch methods was created, most often sharing the same syntax as their
12135ACL counterpart. These sample fetch methods are also known as "fetches". As
12136of now, ACLs and fetches have converged. All ACL fetch methods have been made
12137available as fetch methods, and ACLs may use any sample fetch method as well.
12138
12139This section details all available sample fetch methods and their output type.
12140Some sample fetch methods have deprecated aliases that are used to maintain
12141compatibility with existing configurations. They are then explicitly marked as
12142deprecated and should not be used in new setups.
12143
12144The ACL derivatives are also indicated when available, with their respective
12145matching methods. These ones all have a well defined default pattern matching
12146method, so it is never necessary (though allowed) to pass the "-m" option to
12147indicate how the sample will be matched using ACLs.
12148
12149As indicated in the sample type versus matching compatibility matrix above,
12150when using a generic sample fetch method in an ACL, the "-m" option is
12151mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
12152the same keyword exists as an ACL keyword and as a standard fetch method, the
12153ACL engine will automatically pick the ACL-only one by default.
12154
12155Some of these keywords support one or multiple mandatory arguments, and one or
12156multiple optional arguments. These arguments are strongly typed and are checked
12157when the configuration is parsed so that there is no risk of running with an
12158incorrect argument (eg: an unresolved backend name). Fetch function arguments
12159are passed between parenthesis and are delimited by commas. When an argument
12160is optional, it will be indicated below between square brackets ('[ ]'). When
12161all arguments are optional, the parenthesis may be omitted.
12162
12163Thus, the syntax of a standard sample fetch method is one of the following :
12164 - name
12165 - name(arg1)
12166 - name(arg1,arg2)
12167
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012168
121697.3.1. Converters
12170-----------------
12171
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012172Sample fetch methods may be combined with transformations to be applied on top
12173of the fetched sample (also called "converters"). These combinations form what
12174is called "sample expressions" and the result is a "sample". Initially this
12175was only supported by "stick on" and "stick store-request" directives but this
12176has now be extended to all places where samples may be used (acls, log-format,
12177unique-id-format, add-header, ...).
12178
12179These transformations are enumerated as a series of specific keywords after the
12180sample fetch method. These keywords may equally be appended immediately after
12181the fetch keyword's argument, delimited by a comma. These keywords can also
12182support some arguments (eg: a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012183
Willy Tarreau97707872015-01-27 15:12:13 +010012184A certain category of converters are bitwise and arithmetic operators which
12185support performing basic operations on integers. Some bitwise operations are
12186supported (and, or, xor, cpl) and some arithmetic operations are supported
12187(add, sub, mul, div, mod, neg). Some comparators are provided (odd, even, not,
12188bool) which make it possible to report a match without having to write an ACL.
12189
Willy Tarreau74ca5042013-06-11 23:12:07 +020012190The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +010012191
Ben Shillitof25e8e52016-12-02 14:25:37 +00001219251d.single(<prop>[,<prop>*])
12193 Returns values for the properties requested as a string, where values are
12194 separated by the delimiter specified with "51degrees-property-separator".
12195 The device is identified using the User-Agent header passed to the
12196 converter. The function can be passed up to five property names, and if a
12197 property name can't be found, the value "NoData" is returned.
12198
12199 Example :
12200 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
12201 # containg values for the three properties requested by using the
12202 # User-Agent passed to the converter.
12203 frontend http-in
12204 bind *:8081
12205 default_backend servers
12206 http-request set-header X-51D-DeviceTypeMobileTablet \
12207 %[req.fhdr(User-Agent),51d.single(DeviceType,IsMobile,IsTablet)]
12208
Willy Tarreau97707872015-01-27 15:12:13 +010012209add(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012210 Adds <value> to the input value of type signed integer, and returns the
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012211 result as a signed integer. <value> can be a numeric value or a variable
Daniel Schneller0b547052016-03-21 20:46:57 +010012212 name. The name of the variable starts with an indication about its scope. The
12213 scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012214 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012215 "sess" : the variable is shared with the whole session
12216 "txn" : the variable is shared with the transaction (request and response)
12217 "req" : the variable is shared only during request processing
12218 "res" : the variable is shared only during response processing
12219 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012220 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012221
12222and(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012223 Performs a bitwise "AND" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012224 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010012225 numeric value or a variable name. The name of the variable starts with an
12226 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012227 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012228 "sess" : the variable is shared with the whole session
12229 "txn" : the variable is shared with the transaction (request and response)
12230 "req" : the variable is shared only during request processing
12231 "res" : the variable is shared only during response processing
12232 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012233 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012234
Emeric Brun53d1a982014-04-30 18:21:37 +020012235base64
12236 Converts a binary input sample to a base64 string. It is used to log or
12237 transfer binary content in a way that can be reliably transferred (eg:
12238 an SSL ID can be copied in a header).
12239
Willy Tarreau97707872015-01-27 15:12:13 +010012240bool
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012241 Returns a boolean TRUE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010012242 non-null, otherwise returns FALSE. Used in conjunction with and(), it can be
12243 used to report true/false for bit testing on input values (eg: verify the
12244 presence of a flag).
12245
Emeric Brun54c4ac82014-11-03 15:32:43 +010012246bytes(<offset>[,<length>])
12247 Extracts some bytes from an input binary sample. The result is a binary
12248 sample starting at an offset (in bytes) of the original sample and
Tim Düsterhus4896c442016-11-29 02:15:19 +010012249 optionally truncated at the given length.
Emeric Brun54c4ac82014-11-03 15:32:43 +010012250
Willy Tarreau97707872015-01-27 15:12:13 +010012251cpl
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012252 Takes the input value of type signed integer, applies a ones-complement
12253 (flips all bits) and returns the result as an signed integer.
Willy Tarreau97707872015-01-27 15:12:13 +010012254
Willy Tarreau80599772015-01-20 19:35:24 +010012255crc32([<avalanche>])
12256 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32
12257 hash function. Optionally, it is possible to apply a full avalanche hash
12258 function to the output if the optional <avalanche> argument equals 1. This
12259 converter uses the same functions as used by the various hash-based load
12260 balancing algorithms, so it will provide exactly the same results. It is
12261 provided for compatibility with other software which want a CRC32 to be
12262 computed on some input keys, so it follows the most common implementation as
12263 found in Ethernet, Gzip, PNG, etc... It is slower than the other algorithms
12264 but may provide a better or at least less predictable distribution. It must
12265 not be used for security purposes as a 32-bit hash is trivial to break. See
12266 also "djb2", "sdbm", "wt6" and the "hash-type" directive.
12267
David Carlier29b3ca32015-09-25 14:09:21 +010012268da-csv-conv(<prop>[,<prop>*])
David Carlier4542b102015-06-01 13:54:29 +020012269 Asks the DeviceAtlas converter to identify the User Agent string passed on
12270 input, and to emit a string made of the concatenation of the properties
12271 enumerated in argument, delimited by the separator defined by the global
12272 keyword "deviceatlas-property-separator", or by default the pipe character
David Carlier840b0242016-03-16 10:09:55 +000012273 ('|'). There's a limit of 12 different properties imposed by the haproxy
David Carlier4542b102015-06-01 13:54:29 +020012274 configuration language.
12275
12276 Example:
12277 frontend www
Cyril Bonté307ee1e2015-09-28 23:16:06 +020012278 bind *:8881
12279 default_backend servers
David Carlier840b0242016-03-16 10:09:55 +000012280 http-request set-header X-DeviceAtlas-Data %[req.fhdr(User-Agent),da-csv(primaryHardwareType,osName,osVersion,browserName,browserVersion,browserRenderingEngine)]
David Carlier4542b102015-06-01 13:54:29 +020012281
Thierry FOURNIER9687c772015-05-07 15:46:29 +020012282debug
12283 This converter is used as debug tool. It dumps on screen the content and the
12284 type of the input sample. The sample is returned as is on its output. This
12285 converter only exists when haproxy was built with debugging enabled.
12286
Willy Tarreau97707872015-01-27 15:12:13 +010012287div(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012288 Divides the input value of type signed integer by <value>, and returns the
12289 result as an signed integer. If <value> is null, the largest unsigned
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012290 integer is returned (typically 2^63-1). <value> can be a numeric value or a
Daniel Schneller0b547052016-03-21 20:46:57 +010012291 variable name. The name of the variable starts with an indication about its
12292 scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012293 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012294 "sess" : the variable is shared with the whole session
12295 "txn" : the variable is shared with the transaction (request and response)
12296 "req" : the variable is shared only during request processing
12297 "res" : the variable is shared only during response processing
12298 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012299 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012300
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012301djb2([<avalanche>])
12302 Hashes a binary input sample into an unsigned 32-bit quantity using the DJB2
12303 hash function. Optionally, it is possible to apply a full avalanche hash
12304 function to the output if the optional <avalanche> argument equals 1. This
12305 converter uses the same functions as used by the various hash-based load
12306 balancing algorithms, so it will provide exactly the same results. It is
12307 mostly intended for debugging, but can be used as a stick-table entry to
12308 collect rough statistics. It must not be used for security purposes as a
Willy Tarreau80599772015-01-20 19:35:24 +010012309 32-bit hash is trivial to break. See also "crc32", "sdbm", "wt6" and the
12310 "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012311
Willy Tarreau97707872015-01-27 15:12:13 +010012312even
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012313 Returns a boolean TRUE if the input value of type signed integer is even
Willy Tarreau97707872015-01-27 15:12:13 +010012314 otherwise returns FALSE. It is functionally equivalent to "not,and(1),bool".
12315
Emeric Brunf399b0d2014-11-03 17:07:03 +010012316field(<index>,<delimiters>)
12317 Extracts the substring at the given index considering given delimiters from
12318 an input string. Indexes start at 1 and delimiters are a string formatted
12319 list of chars.
12320
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012321hex
12322 Converts a binary input sample to an hex string containing two hex digits per
12323 input byte. It is used to log or transfer hex dumps of some binary input data
12324 in a way that can be reliably transferred (eg: an SSL ID can be copied in a
12325 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +010012326
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012327http_date([<offset>])
12328 Converts an integer supposed to contain a date since epoch to a string
12329 representing this date in a format suitable for use in HTTP header fields. If
12330 an offset value is specified, then it is a number of seconds that is added to
12331 the date before the conversion is operated. This is particularly useful to
12332 emit Date header fields, Expires values in responses when combined with a
12333 positive offset, or Last-Modified values when the offset is negative.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012334
Willy Tarreaud9f316a2014-07-10 14:03:38 +020012335in_table(<table>)
12336 Uses the string representation of the input sample to perform a look up in
12337 the specified table. If the key is not found in the table, a boolean false
12338 is returned. Otherwise a boolean true is returned. This can be used to verify
12339 the presence of a certain key in a table tracking some elements (eg: whether
12340 or not a source IP address or an Authorization header was already seen).
12341
Willy Tarreauffcb2e42014-07-10 16:29:08 +020012342ipmask(<mask>)
12343 Apply a mask to an IPv4 address, and use the result for lookups and storage.
12344 This can be used to make all hosts within a certain mask to share the same
12345 table entries and as such use the same server. The mask can be passed in
12346 dotted form (eg: 255.255.255.0) or in CIDR form (eg: 24).
12347
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020012348json([<input-code>])
12349 Escapes the input string and produces an ASCII ouput string ready to use as a
12350 JSON string. The converter tries to decode the input string according to the
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020012351 <input-code> parameter. It can be "ascii", "utf8", "utf8s", "utf8p" or
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020012352 "utf8ps". The "ascii" decoder never fails. The "utf8" decoder detects 3 types
12353 of errors:
12354 - bad UTF-8 sequence (lone continuation byte, bad number of continuation
12355 bytes, ...)
12356 - invalid range (the decoded value is within a UTF-8 prohibited range),
12357 - code overlong (the value is encoded with more bytes than necessary).
12358
12359 The UTF-8 JSON encoding can produce a "too long value" error when the UTF-8
12360 character is greater than 0xffff because the JSON string escape specification
12361 only authorizes 4 hex digits for the value encoding. The UTF-8 decoder exists
12362 in 4 variants designated by a combination of two suffix letters : "p" for
12363 "permissive" and "s" for "silently ignore". The behaviors of the decoders
12364 are :
12365 - "ascii" : never fails ;
12366 - "utf8" : fails on any detected errors ;
12367 - "utf8s" : never fails, but removes characters corresponding to errors ;
12368 - "utf8p" : accepts and fixes the overlong errors, but fails on any other
12369 error ;
12370 - "utf8ps" : never fails, accepts and fixes the overlong errors, but removes
12371 characters corresponding to the other errors.
12372
12373 This converter is particularly useful for building properly escaped JSON for
12374 logging to servers which consume JSON-formated traffic logs.
12375
12376 Example:
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020012377 capture request header Host len 15
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020012378 capture request header user-agent len 150
12379 log-format '{"ip":"%[src]","user-agent":"%[capture.req.hdr(1),json(utf8s)]"}'
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020012380
12381 Input request from client 127.0.0.1:
12382 GET / HTTP/1.0
12383 User-Agent: Very "Ugly" UA 1/2
12384
12385 Output log:
12386 {"ip":"127.0.0.1","user-agent":"Very \"Ugly\" UA 1\/2"}
12387
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012388language(<value>[,<default>])
12389 Returns the value with the highest q-factor from a list as extracted from the
12390 "accept-language" header using "req.fhdr". Values with no q-factor have a
12391 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
12392 belong to the list of semi-colon delimited <values> will be considered. The
12393 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
12394 given list and a default value is provided, it is returned. Note that language
12395 names may have a variant after a dash ('-'). If this variant is present in the
12396 list, it will be matched, but if it is not, only the base language is checked.
12397 The match is case-sensitive, and the output string is always one of those
12398 provided in arguments. The ordering of arguments is meaningless, only the
12399 ordering of the values in the request counts, as the first value among
12400 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +020012401
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012402 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +020012403
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012404 # this configuration switches to the backend matching a
12405 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +020012406
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012407 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
12408 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
12409 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
12410 use_backend spanish if es
12411 use_backend french if fr
12412 use_backend english if en
12413 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +020012414
Willy Tarreauffcb2e42014-07-10 16:29:08 +020012415lower
12416 Convert a string sample to lower case. This can only be placed after a string
12417 sample fetch function or after a transformation keyword returning a string
12418 type. The result is of type string.
12419
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020012420ltime(<format>[,<offset>])
12421 Converts an integer supposed to contain a date since epoch to a string
12422 representing this date in local time using a format defined by the <format>
12423 string using strftime(3). The purpose is to allow any date format to be used
12424 in logs. An optional <offset> in seconds may be applied to the input date
12425 (positive or negative). See the strftime() man page for the format supported
12426 by your operating system. See also the utime converter.
12427
12428 Example :
12429
12430 # Emit two colons, one with the local time and another with ip:port
12431 # Eg: 20140710162350 127.0.0.1:57325
12432 log-format %[date,ltime(%Y%m%d%H%M%S)]\ %ci:%cp
12433
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012434map(<map_file>[,<default_value>])
12435map_<match_type>(<map_file>[,<default_value>])
12436map_<match_type>_<output_type>(<map_file>[,<default_value>])
12437 Search the input value from <map_file> using the <match_type> matching method,
12438 and return the associated value converted to the type <output_type>. If the
12439 input value cannot be found in the <map_file>, the converter returns the
12440 <default_value>. If the <default_value> is not set, the converter fails and
12441 acts as if no input value could be fetched. If the <match_type> is not set, it
12442 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
12443 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
12444 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012445
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012446 It is important to avoid overlapping between the keys : IP addresses and
12447 strings are stored in trees, so the first of the finest match will be used.
12448 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012449
Tim Düsterhus4896c442016-11-29 02:15:19 +010012450 The following array contains the list of all map functions available sorted by
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012451 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012452
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012453 input type | match method | output type str | output type int | output type ip
12454 -----------+--------------+-----------------+-----------------+---------------
12455 str | str | map_str | map_str_int | map_str_ip
12456 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +020012457 str | beg | map_beg | map_beg_int | map_end_ip
12458 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012459 str | sub | map_sub | map_sub_int | map_sub_ip
12460 -----------+--------------+-----------------+-----------------+---------------
12461 str | dir | map_dir | map_dir_int | map_dir_ip
12462 -----------+--------------+-----------------+-----------------+---------------
12463 str | dom | map_dom | map_dom_int | map_dom_ip
12464 -----------+--------------+-----------------+-----------------+---------------
12465 str | end | map_end | map_end_int | map_end_ip
12466 -----------+--------------+-----------------+-----------------+---------------
Ruoshan Huang3c5e3742016-12-02 16:25:31 +080012467 str | reg | map_reg | map_reg_int | map_reg_ip
12468 -----------+--------------+-----------------+-----------------+---------------
12469 str | reg | map_regm | map_reg_int | map_reg_ip
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012470 -----------+--------------+-----------------+-----------------+---------------
12471 int | int | map_int | map_int_int | map_int_ip
12472 -----------+--------------+-----------------+-----------------+---------------
12473 ip | ip | map_ip | map_ip_int | map_ip_ip
12474 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012475
Thierry Fournier8feaa662016-02-10 22:55:20 +010012476 The special map called "map_regm" expect matching zone in the regular
12477 expression and modify the output replacing back reference (like "\1") by
12478 the corresponding match text.
12479
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012480 The file contains one key + value per line. Lines which start with '#' are
12481 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
12482 is then the first "word" (series of non-space/tabs characters), and the value
12483 is what follows this series of space/tab till the end of the line excluding
12484 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012485
Thierry FOURNIER060762e2014-04-23 13:29:15 +020012486 Example :
12487
12488 # this is a comment and is ignored
12489 2.22.246.0/23 United Kingdom \n
12490 <-><-----------><--><------------><---->
12491 | | | | `- trailing spaces ignored
12492 | | | `---------- value
12493 | | `-------------------- middle spaces ignored
12494 | `---------------------------- key
12495 `------------------------------------ leading spaces ignored
12496
Willy Tarreau97707872015-01-27 15:12:13 +010012497mod(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012498 Divides the input value of type signed integer by <value>, and returns the
12499 remainder as an signed integer. If <value> is null, then zero is returned.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012500 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010012501 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012502 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012503 "sess" : the variable is shared with the whole session
12504 "txn" : the variable is shared with the transaction (request and response)
12505 "req" : the variable is shared only during request processing
12506 "res" : the variable is shared only during response processing
12507 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012508 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012509
12510mul(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012511 Multiplies the input value of type signed integer by <value>, and returns
Thierry FOURNIER00c005c2015-07-08 01:10:21 +020012512 the product as an signed integer. In case of overflow, the largest possible
12513 value for the sign is returned so that the operation doesn't wrap around.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012514 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010012515 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012516 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012517 "sess" : the variable is shared with the whole session
12518 "txn" : the variable is shared with the transaction (request and response)
12519 "req" : the variable is shared only during request processing
12520 "res" : the variable is shared only during response processing
12521 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012522 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012523
12524neg
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012525 Takes the input value of type signed integer, computes the opposite value,
12526 and returns the remainder as an signed integer. 0 is identity. This operator
12527 is provided for reversed subtracts : in order to subtract the input from a
12528 constant, simply perform a "neg,add(value)".
Willy Tarreau97707872015-01-27 15:12:13 +010012529
12530not
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012531 Returns a boolean FALSE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010012532 non-null, otherwise returns TRUE. Used in conjunction with and(), it can be
12533 used to report true/false for bit testing on input values (eg: verify the
12534 absence of a flag).
12535
12536odd
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012537 Returns a boolean TRUE if the input value of type signed integer is odd
Willy Tarreau97707872015-01-27 15:12:13 +010012538 otherwise returns FALSE. It is functionally equivalent to "and(1),bool".
12539
12540or(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012541 Performs a bitwise "OR" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012542 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010012543 numeric value or a variable name. The name of the variable starts with an
12544 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012545 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012546 "sess" : the variable is shared with the whole session
12547 "txn" : the variable is shared with the transaction (request and response)
12548 "req" : the variable is shared only during request processing
12549 "res" : the variable is shared only during response processing
12550 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012551 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012552
Willy Tarreauc4dc3502015-01-23 20:39:28 +010012553regsub(<regex>,<subst>[,<flags>])
Willy Tarreau7eda8492015-01-20 19:47:06 +010012554 Applies a regex-based substitution to the input string. It does the same
12555 operation as the well-known "sed" utility with "s/<regex>/<subst>/". By
12556 default it will replace in the input string the first occurrence of the
12557 largest part matching the regular expression <regex> with the substitution
12558 string <subst>. It is possible to replace all occurrences instead by adding
12559 the flag "g" in the third argument <flags>. It is also possible to make the
12560 regex case insensitive by adding the flag "i" in <flags>. Since <flags> is a
12561 string, it is made up from the concatenation of all desired flags. Thus if
12562 both "i" and "g" are desired, using "gi" or "ig" will have the same effect.
12563 It is important to note that due to the current limitations of the
Baptiste Assmann66025d82016-03-06 23:36:48 +010012564 configuration parser, some characters such as closing parenthesis, closing
12565 square brackets or comma are not possible to use in the arguments. The first
12566 use of this converter is to replace certain characters or sequence of
12567 characters with other ones.
Willy Tarreau7eda8492015-01-20 19:47:06 +010012568
12569 Example :
12570
12571 # de-duplicate "/" in header "x-path".
12572 # input: x-path: /////a///b/c/xzxyz/
12573 # output: x-path: /a/b/c/xzxyz/
12574 http-request set-header x-path %[hdr(x-path),regsub(/+,/,g)]
12575
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020012576capture-req(<id>)
12577 Capture the string entry in the request slot <id> and returns the entry as
12578 is. If the slot doesn't exist, the capture fails silently.
12579
12580 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020012581 "http-response capture", "capture.req.hdr" and
12582 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020012583
12584capture-res(<id>)
12585 Capture the string entry in the response slot <id> and returns the entry as
12586 is. If the slot doesn't exist, the capture fails silently.
12587
12588 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020012589 "http-response capture", "capture.req.hdr" and
12590 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020012591
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012592sdbm([<avalanche>])
12593 Hashes a binary input sample into an unsigned 32-bit quantity using the SDBM
12594 hash function. Optionally, it is possible to apply a full avalanche hash
12595 function to the output if the optional <avalanche> argument equals 1. This
12596 converter uses the same functions as used by the various hash-based load
12597 balancing algorithms, so it will provide exactly the same results. It is
12598 mostly intended for debugging, but can be used as a stick-table entry to
12599 collect rough statistics. It must not be used for security purposes as a
Willy Tarreau80599772015-01-20 19:35:24 +010012600 32-bit hash is trivial to break. See also "crc32", "djb2", "wt6" and the
12601 "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012602
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012603set-var(<var name>)
Daniel Schneller0b547052016-03-21 20:46:57 +010012604 Sets a variable with the input content and returns the content on the output as
12605 is. The variable keeps the value and the associated input type. The name of the
12606 variable starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012607 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012608 "sess" : the variable is shared with the whole session
12609 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012610 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010012611 "req" : the variable is shared only during request processing,
12612 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012613 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012614 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012615
Willy Tarreau97707872015-01-27 15:12:13 +010012616sub(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012617 Subtracts <value> from the input value of type signed integer, and returns
12618 the result as an signed integer. Note: in order to subtract the input from
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012619 a constant, simply perform a "neg,add(value)". <value> can be a numeric value
Daniel Schneller0b547052016-03-21 20:46:57 +010012620 or a variable name. The name of the variable starts with an indication about
12621 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012622 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012623 "sess" : the variable is shared with the whole session
12624 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012625 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010012626 "req" : the variable is shared only during request processing,
12627 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012628 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012629 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012630
Willy Tarreaud9f316a2014-07-10 14:03:38 +020012631table_bytes_in_rate(<table>)
12632 Uses the string representation of the input sample to perform a look up in
12633 the specified table. If the key is not found in the table, integer value zero
12634 is returned. Otherwise the converter returns the average client-to-server
12635 bytes rate associated with the input sample in the designated table, measured
12636 in amount of bytes over the period configured in the table. See also the
12637 sc_bytes_in_rate sample fetch keyword.
12638
12639
12640table_bytes_out_rate(<table>)
12641 Uses the string representation of the input sample to perform a look up in
12642 the specified table. If the key is not found in the table, integer value zero
12643 is returned. Otherwise the converter returns the average server-to-client
12644 bytes rate associated with the input sample in the designated table, measured
12645 in amount of bytes over the period configured in the table. See also the
12646 sc_bytes_out_rate sample fetch keyword.
12647
12648table_conn_cnt(<table>)
12649 Uses the string representation of the input sample to perform a look up in
12650 the specified table. If the key is not found in the table, integer value zero
12651 is returned. Otherwise the converter returns the cumulated amount of incoming
12652 connections associated with the input sample in the designated table. See
12653 also the sc_conn_cnt sample fetch keyword.
12654
12655table_conn_cur(<table>)
12656 Uses the string representation of the input sample to perform a look up in
12657 the specified table. If the key is not found in the table, integer value zero
12658 is returned. Otherwise the converter returns the current amount of concurrent
12659 tracked connections associated with the input sample in the designated table.
12660 See also the sc_conn_cur sample fetch keyword.
12661
12662table_conn_rate(<table>)
12663 Uses the string representation of the input sample to perform a look up in
12664 the specified table. If the key is not found in the table, integer value zero
12665 is returned. Otherwise the converter returns the average incoming connection
12666 rate associated with the input sample in the designated table. See also the
12667 sc_conn_rate sample fetch keyword.
12668
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012669table_gpt0(<table>)
12670 Uses the string representation of the input sample to perform a look up in
12671 the specified table. If the key is not found in the table, boolean value zero
12672 is returned. Otherwise the converter returns the current value of the first
12673 general purpose tag associated with the input sample in the designated table.
12674 See also the sc_get_gpt0 sample fetch keyword.
12675
Willy Tarreaud9f316a2014-07-10 14:03:38 +020012676table_gpc0(<table>)
12677 Uses the string representation of the input sample to perform a look up in
12678 the specified table. If the key is not found in the table, integer value zero
12679 is returned. Otherwise the converter returns the current value of the first
12680 general purpose counter associated with the input sample in the designated
12681 table. See also the sc_get_gpc0 sample fetch keyword.
12682
12683table_gpc0_rate(<table>)
12684 Uses the string representation of the input sample to perform a look up in
12685 the specified table. If the key is not found in the table, integer value zero
12686 is returned. Otherwise the converter returns the frequency which the gpc0
12687 counter was incremented over the configured period in the table, associated
12688 with the input sample in the designated table. See also the sc_get_gpc0_rate
12689 sample fetch keyword.
12690
12691table_http_err_cnt(<table>)
12692 Uses the string representation of the input sample to perform a look up in
12693 the specified table. If the key is not found in the table, integer value zero
12694 is returned. Otherwise the converter returns the cumulated amount of HTTP
12695 errors associated with the input sample in the designated table. See also the
12696 sc_http_err_cnt sample fetch keyword.
12697
12698table_http_err_rate(<table>)
12699 Uses the string representation of the input sample to perform a look up in
12700 the specified table. If the key is not found in the table, integer value zero
12701 is returned. Otherwise the average rate of HTTP errors associated with the
12702 input sample in the designated table, measured in amount of errors over the
12703 period configured in the table. See also the sc_http_err_rate sample fetch
12704 keyword.
12705
12706table_http_req_cnt(<table>)
12707 Uses the string representation of the input sample to perform a look up in
12708 the specified table. If the key is not found in the table, integer value zero
12709 is returned. Otherwise the converter returns the cumulated amount of HTTP
12710 requests associated with the input sample in the designated table. See also
12711 the sc_http_req_cnt sample fetch keyword.
12712
12713table_http_req_rate(<table>)
12714 Uses the string representation of the input sample to perform a look up in
12715 the specified table. If the key is not found in the table, integer value zero
12716 is returned. Otherwise the average rate of HTTP requests associated with the
12717 input sample in the designated table, measured in amount of requests over the
12718 period configured in the table. See also the sc_http_req_rate sample fetch
12719 keyword.
12720
12721table_kbytes_in(<table>)
12722 Uses the string representation of the input sample to perform a look up in
12723 the specified table. If the key is not found in the table, integer value zero
12724 is returned. Otherwise the converter returns the cumulated amount of client-
12725 to-server data associated with the input sample in the designated table,
12726 measured in kilobytes. The test is currently performed on 32-bit integers,
12727 which limits values to 4 terabytes. See also the sc_kbytes_in sample fetch
12728 keyword.
12729
12730table_kbytes_out(<table>)
12731 Uses the string representation of the input sample to perform a look up in
12732 the specified table. If the key is not found in the table, integer value zero
12733 is returned. Otherwise the converter returns the cumulated amount of server-
12734 to-client data associated with the input sample in the designated table,
12735 measured in kilobytes. The test is currently performed on 32-bit integers,
12736 which limits values to 4 terabytes. See also the sc_kbytes_out sample fetch
12737 keyword.
12738
12739table_server_id(<table>)
12740 Uses the string representation of the input sample to perform a look up in
12741 the specified table. If the key is not found in the table, integer value zero
12742 is returned. Otherwise the converter returns the server ID associated with
12743 the input sample in the designated table. A server ID is associated to a
12744 sample by a "stick" rule when a connection to a server succeeds. A server ID
12745 zero means that no server is associated with this key.
12746
12747table_sess_cnt(<table>)
12748 Uses the string representation of the input sample to perform a look up in
12749 the specified table. If the key is not found in the table, integer value zero
12750 is returned. Otherwise the converter returns the cumulated amount of incoming
12751 sessions associated with the input sample in the designated table. Note that
12752 a session here refers to an incoming connection being accepted by the
12753 "tcp-request connection" rulesets. See also the sc_sess_cnt sample fetch
12754 keyword.
12755
12756table_sess_rate(<table>)
12757 Uses the string representation of the input sample to perform a look up in
12758 the specified table. If the key is not found in the table, integer value zero
12759 is returned. Otherwise the converter returns the average incoming session
12760 rate associated with the input sample in the designated table. Note that a
12761 session here refers to an incoming connection being accepted by the
12762 "tcp-request connection" rulesets. See also the sc_sess_rate sample fetch
12763 keyword.
12764
12765table_trackers(<table>)
12766 Uses the string representation of the input sample to perform a look up in
12767 the specified table. If the key is not found in the table, integer value zero
12768 is returned. Otherwise the converter returns the current amount of concurrent
12769 connections tracking the same key as the input sample in the designated
12770 table. It differs from table_conn_cur in that it does not rely on any stored
12771 information but on the table's reference count (the "use" value which is
12772 returned by "show table" on the CLI). This may sometimes be more suited for
12773 layer7 tracking. It can be used to tell a server how many concurrent
12774 connections there are from a given address for example. See also the
12775 sc_trackers sample fetch keyword.
12776
Willy Tarreauffcb2e42014-07-10 16:29:08 +020012777upper
12778 Convert a string sample to upper case. This can only be placed after a string
12779 sample fetch function or after a transformation keyword returning a string
12780 type. The result is of type string.
12781
Thierry FOURNIER82ff3c92015-05-07 15:46:20 +020012782url_dec
12783 Takes an url-encoded string provided as input and returns the decoded
12784 version as output. The input and the output are of type string.
12785
Christopher Faulet85d79c92016-11-09 16:54:56 +010012786unset-var(<var name>)
12787 Unsets a variable if the input content is defined. The name of the variable
12788 starts with an indication about its scope. The scopes allowed are:
12789 "proc" : the variable is shared with the whole process
12790 "sess" : the variable is shared with the whole session
12791 "txn" : the variable is shared with the transaction (request and
12792 response),
12793 "req" : the variable is shared only during request processing,
12794 "res" : the variable is shared only during response processing.
12795 This prefix is followed by a name. The separator is a '.'. The name may only
12796 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
12797
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020012798utime(<format>[,<offset>])
12799 Converts an integer supposed to contain a date since epoch to a string
12800 representing this date in UTC time using a format defined by the <format>
12801 string using strftime(3). The purpose is to allow any date format to be used
12802 in logs. An optional <offset> in seconds may be applied to the input date
12803 (positive or negative). See the strftime() man page for the format supported
12804 by your operating system. See also the ltime converter.
12805
12806 Example :
12807
12808 # Emit two colons, one with the UTC time and another with ip:port
12809 # Eg: 20140710162350 127.0.0.1:57325
12810 log-format %[date,utime(%Y%m%d%H%M%S)]\ %ci:%cp
12811
Emeric Brunc9a0f6d2014-11-25 14:09:01 +010012812word(<index>,<delimiters>)
12813 Extracts the nth word considering given delimiters from an input string.
12814 Indexes start at 1 and delimiters are a string formatted list of chars.
12815
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012816wt6([<avalanche>])
12817 Hashes a binary input sample into an unsigned 32-bit quantity using the WT6
12818 hash function. Optionally, it is possible to apply a full avalanche hash
12819 function to the output if the optional <avalanche> argument equals 1. This
12820 converter uses the same functions as used by the various hash-based load
12821 balancing algorithms, so it will provide exactly the same results. It is
12822 mostly intended for debugging, but can be used as a stick-table entry to
12823 collect rough statistics. It must not be used for security purposes as a
Willy Tarreau80599772015-01-20 19:35:24 +010012824 32-bit hash is trivial to break. See also "crc32", "djb2", "sdbm", and the
12825 "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020012826
Willy Tarreau97707872015-01-27 15:12:13 +010012827xor(<value>)
12828 Performs a bitwise "XOR" (exclusive OR) between <value> and the input value
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020012829 of type signed integer, and returns the result as an signed integer.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012830 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010012831 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012832 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012833 "sess" : the variable is shared with the whole session
12834 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012835 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010012836 "req" : the variable is shared only during request processing,
12837 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020012838 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010012839 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010012840
Thierry FOURNIER01e09742016-12-26 11:46:11 +010012841xxh32([<seed>])
12842 Hashes a binary input sample into an unsigned 32-bit quantity using the 32-bit
12843 variant of the XXHash hash function. This hash supports a seed which defaults
12844 to zero but a different value maybe passed as the <seed> argument. This hash
12845 is known to be very good and very fast so it can be used to hash URLs and/or
12846 URL parameters for use as stick-table keys to collect statistics with a low
12847 collision rate, though care must be taken as the algorithm is not considered
12848 as cryptographically secure.
12849
12850xxh64([<seed>])
12851 Hashes a binary input sample into a signed 64-bit quantity using the 64-bit
12852 variant of the XXHash hash function. This hash supports a seed which defaults
12853 to zero but a different value maybe passed as the <seed> argument. This hash
12854 is known to be very good and very fast so it can be used to hash URLs and/or
12855 URL parameters for use as stick-table keys to collect statistics with a low
12856 collision rate, though care must be taken as the algorithm is not considered
12857 as cryptographically secure.
12858
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010012859
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200128607.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020012861--------------------------------------------
12862
12863A first set of sample fetch methods applies to internal information which does
12864not even relate to any client information. These ones are sometimes used with
12865"monitor-fail" directives to report an internal status to external watchers.
12866The sample fetch methods described in this section are usable anywhere.
12867
12868always_false : boolean
12869 Always returns the boolean "false" value. It may be used with ACLs as a
12870 temporary replacement for another one when adjusting configurations.
12871
12872always_true : boolean
12873 Always returns the boolean "true" value. It may be used with ACLs as a
12874 temporary replacement for another one when adjusting configurations.
12875
12876avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010012877 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020012878 divided by the number of active servers. The current backend is used if no
12879 backend is specified. This is very similar to "queue" except that the size of
12880 the farm is considered, in order to give a more accurate measurement of the
12881 time it may take for a new connection to be processed. The main usage is with
12882 ACL to return a sorry page to new users when it becomes certain they will get
12883 a degraded service, or to pass to the backend servers in a header so that
12884 they decide to work in degraded mode or to disable some functions to speed up
12885 the processing a bit. Note that in the event there would not be any active
12886 server anymore, twice the number of queued connections would be considered as
12887 the measured value. This is a fair estimate, as we expect one server to get
12888 back soon anyway, but we still prefer to send new traffic to another backend
12889 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
12890 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010012891
Willy Tarreau74ca5042013-06-11 23:12:07 +020012892be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020012893 Applies to the number of currently established connections on the backend,
12894 possibly including the connection being evaluated. If no backend name is
12895 specified, the current one is used. But it is also possible to check another
12896 backend. It can be used to use a specific farm when the nominal one is full.
12897 See also the "fe_conn", "queue" and "be_sess_rate" criteria.
Willy Tarreau6a06a402007-07-15 20:15:28 +020012898
Willy Tarreau74ca5042013-06-11 23:12:07 +020012899be_sess_rate([<backend>]) : integer
12900 Returns an integer value corresponding to the sessions creation rate on the
12901 backend, in number of new sessions per second. This is used with ACLs to
12902 switch to an alternate backend when an expensive or fragile one reaches too
12903 high a session rate, or to limit abuse of service (eg. prevent sucking of an
12904 online dictionary). It can also be useful to add this element to logs using a
12905 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010012906
12907 Example :
12908 # Redirect to an error page if the dictionary is requested too often
12909 backend dynamic
12910 mode http
12911 acl being_scanned be_sess_rate gt 100
12912 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010012913
Thierry FOURNIERcc103292015-06-06 19:30:17 +020012914bin(<hexa>) : bin
12915 Returns a binary chain. The input is the hexadecimal representation
12916 of the string.
12917
12918bool(<bool>) : bool
12919 Returns a boolean value. <bool> can be 'true', 'false', '1' or '0'.
12920 'false' and '0' are the same. 'true' and '1' are the same.
12921
Willy Tarreau74ca5042013-06-11 23:12:07 +020012922connslots([<backend>]) : integer
12923 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012924 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020012925 connections on all servers and the maximum queue size. This is probably only
12926 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050012927
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080012928 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020012929 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080012930 usage; see "use_backend" keyword) can be redirected to a different backend.
12931
Willy Tarreau55165fe2009-05-10 12:02:55 +020012932 'connslots' = number of available server connection slots, + number of
12933 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080012934
Willy Tarreaua36af912009-10-10 12:02:45 +020012935 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020012936 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020012937 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020012938 you want to be able to differentiate between different backends, and their
12939 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020012940 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020012941 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080012942
Willy Tarreau55165fe2009-05-10 12:02:55 +020012943 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
12944 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020012945 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020012946 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080012947
Willy Tarreau6236d3a2013-07-25 14:28:25 +020012948date([<offset>]) : integer
12949 Returns the current date as the epoch (number of seconds since 01/01/1970).
12950 If an offset value is specified, then it is a number of seconds that is added
12951 to the current date before returning the value. This is particularly useful
12952 to compute relative dates, as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020012953 It is useful combined with the http_date converter.
12954
12955 Example :
12956
12957 # set an expires header to now+1 hour in every response
12958 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020012959
Willy Tarreau595ec542013-06-12 21:34:28 +020012960env(<name>) : string
12961 Returns a string containing the value of environment variable <name>. As a
12962 reminder, environment variables are per-process and are sampled when the
12963 process starts. This can be useful to pass some information to a next hop
12964 server, or with ACLs to take specific action when the process is started a
12965 certain way.
12966
12967 Examples :
12968 # Pass the Via header to next hop with the local hostname in it
12969 http-request add-header Via 1.1\ %[env(HOSTNAME)]
12970
12971 # reject cookie-less requests when the STOP environment variable is set
12972 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
12973
Willy Tarreau74ca5042013-06-11 23:12:07 +020012974fe_conn([<frontend>]) : integer
12975 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010012976 possibly including the connection being evaluated. If no frontend name is
12977 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020012978 frontend. It can be used to return a sorry page before hard-blocking, or to
12979 use a specific backend to drain new requests when the farm is considered
12980 full. This is mostly used with ACLs but can also be used to pass some
12981 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
12982 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020012983
Nenad Merdanovicad9a7e92016-10-03 04:57:37 +020012984fe_req_rate([<frontend>]) : integer
12985 Returns an integer value corresponding to the number of HTTP requests per
12986 second sent to a frontend. This number can differ from "fe_sess_rate" in
12987 situations where client-side keep-alive is enabled.
12988
Willy Tarreau74ca5042013-06-11 23:12:07 +020012989fe_sess_rate([<frontend>]) : integer
12990 Returns an integer value corresponding to the sessions creation rate on the
12991 frontend, in number of new sessions per second. This is used with ACLs to
12992 limit the incoming session rate to an acceptable range in order to prevent
12993 abuse of service at the earliest moment, for example when combined with other
12994 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
12995 down below the limit. It can also be useful to add this element to logs using
12996 a log-format directive. See also the "rate-limit sessions" directive for use
12997 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010012998
12999 Example :
13000 # This frontend limits incoming mails to 10/s with a max of 100
13001 # concurrent connections. We accept any connection below 10/s, and
13002 # force excess clients to wait for 100 ms. Since clients are limited to
13003 # 100 max, there cannot be more than 10 incoming mails per second.
13004 frontend mail
13005 bind :25
13006 mode tcp
13007 maxconn 100
13008 acl too_fast fe_sess_rate ge 10
13009 tcp-request inspect-delay 100ms
13010 tcp-request content accept if ! too_fast
13011 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010013012
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013013int(<integer>) : signed integer
13014 Returns a signed integer.
13015
Thierry FOURNIERcc103292015-06-06 19:30:17 +020013016ipv4(<ipv4>) : ipv4
13017 Returns an ipv4.
13018
13019ipv6(<ipv6>) : ipv6
13020 Returns an ipv6.
13021
13022meth(<method>) : method
13023 Returns a method.
13024
Willy Tarreau0f30d262014-11-24 16:02:05 +010013025nbproc : integer
13026 Returns an integer value corresponding to the number of processes that were
13027 started (it equals the global "nbproc" setting). This is useful for logging
13028 and debugging purposes.
13029
Willy Tarreau74ca5042013-06-11 23:12:07 +020013030nbsrv([<backend>]) : integer
13031 Returns an integer value corresponding to the number of usable servers of
13032 either the current backend or the named backend. This is mostly used with
13033 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010013034 switch to an alternate backend when the number of servers is too low to
13035 to handle some load. It is useful to report a failure when combined with
13036 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010013037
Willy Tarreau0f30d262014-11-24 16:02:05 +010013038proc : integer
13039 Returns an integer value corresponding to the position of the process calling
13040 the function, between 1 and global.nbproc. This is useful for logging and
13041 debugging purposes.
13042
Willy Tarreau74ca5042013-06-11 23:12:07 +020013043queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010013044 Returns the total number of queued connections of the designated backend,
13045 including all the connections in server queues. If no backend name is
13046 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020013047 one. This is useful with ACLs or to pass statistics to backend servers. This
13048 can be used to take actions when queuing goes above a known level, generally
13049 indicating a surge of traffic or a massive slowdown on the servers. One
13050 possible action could be to reject new users but still accept old ones. See
13051 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
13052
Willy Tarreau84310e22014-02-14 11:59:04 +010013053rand([<range>]) : integer
13054 Returns a random integer value within a range of <range> possible values,
13055 starting at zero. If the range is not specified, it defaults to 2^32, which
13056 gives numbers between 0 and 4294967295. It can be useful to pass some values
13057 needed to take some routing decisions for example, or just for debugging
13058 purposes. This random must not be used for security purposes.
13059
Willy Tarreau74ca5042013-06-11 23:12:07 +020013060srv_conn([<backend>/]<server>) : integer
13061 Returns an integer value corresponding to the number of currently established
13062 connections on the designated server, possibly including the connection being
13063 evaluated. If <backend> is omitted, then the server is looked up in the
13064 current backend. It can be used to use a specific farm when one server is
13065 full, or to inform the server about our view of the number of active
13066 connections with it. See also the "fe_conn", "be_conn" and "queue" fetch
13067 methods.
13068
13069srv_is_up([<backend>/]<server>) : boolean
13070 Returns true when the designated server is UP, and false when it is either
13071 DOWN or in maintenance mode. If <backend> is omitted, then the server is
13072 looked up in the current backend. It is mainly used to take action based on
13073 an external status reported via a health check (eg: a geographical site's
13074 availability). Another possible use which is more of a hack consists in
13075 using dummy servers as boolean variables that can be enabled or disabled from
13076 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
13077
13078srv_sess_rate([<backend>/]<server>) : integer
13079 Returns an integer corresponding to the sessions creation rate on the
13080 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013081 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020013082 used with ACLs but can make sense with logs too. This is used to switch to an
13083 alternate backend when an expensive or fragile one reaches too high a session
13084 rate, or to limit abuse of service (eg. prevent latent requests from
13085 overloading servers).
13086
13087 Example :
13088 # Redirect to a separate back
13089 acl srv1_full srv_sess_rate(be1/srv1) gt 50
13090 acl srv2_full srv_sess_rate(be1/srv2) gt 50
13091 use_backend be2 if srv1_full or srv2_full
13092
Willy Tarreau0f30d262014-11-24 16:02:05 +010013093stopping : boolean
13094 Returns TRUE if the process calling the function is currently stopping. This
13095 can be useful for logging, or for relaxing certain checks or helping close
13096 certain connections upon graceful shutdown.
13097
Thierry FOURNIERcc103292015-06-06 19:30:17 +020013098str(<string>) : string
13099 Returns a string.
13100
Willy Tarreau74ca5042013-06-11 23:12:07 +020013101table_avl([<table>]) : integer
13102 Returns the total number of available entries in the current proxy's
13103 stick-table or in the designated stick-table. See also table_cnt.
13104
13105table_cnt([<table>]) : integer
13106 Returns the total number of entries currently in use in the current proxy's
13107 stick-table or in the designated stick-table. See also src_conn_cnt and
13108 table_avl for other entry counting methods.
13109
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013110var(<var-name>) : undefined
13111 Returns a variable with the stored type. If the variable is not set, the
Daniel Schneller0b547052016-03-21 20:46:57 +010013112 sample fetch fails. The name of the variable starts with an indication
13113 about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013114 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013115 "sess" : the variable is shared with the whole session
13116 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013117 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013118 "req" : the variable is shared only during request processing,
13119 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013120 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013121 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013122
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200131237.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020013124----------------------------------
13125
13126The layer 4 usually describes just the transport layer which in haproxy is
13127closest to the connection, where no content is yet made available. The fetch
13128methods described here are usable as low as the "tcp-request connection" rule
13129sets unless they require some future information. Those generally include
13130TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013131the incoming connection. For retrieving a value from a sticky counters, the
13132counter number can be explicitly set as 0, 1, or 2 using the pre-defined
13133"sc0_", "sc1_", or "sc2_" prefix, or it can be specified as the first integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013134argument when using the "sc_" prefix. An optional table may be specified with
13135the "sc*" form, in which case the currently tracked key will be looked up into
13136this alternate table instead of the table currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013137
13138be_id : integer
13139 Returns an integer containing the current backend's id. It can be used in
13140 frontends with responses to check which backend processed the request.
13141
Marcin Deranekd2471c22016-12-12 14:08:05 +010013142be_name : string
13143 Returns a string containing the current backend's name. It can be used in
13144 frontends with responses to check which backend processed the request.
13145
Willy Tarreau74ca5042013-06-11 23:12:07 +020013146dst : ip
13147 This is the destination IPv4 address of the connection on the client side,
13148 which is the address the client connected to. It can be useful when running
13149 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
13150 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
13151 RFC 4291.
13152
13153dst_conn : integer
13154 Returns an integer value corresponding to the number of currently established
13155 connections on the same socket including the one being evaluated. It is
13156 normally used with ACLs but can as well be used to pass the information to
13157 servers in an HTTP header or in logs. It can be used to either return a sorry
13158 page before hard-blocking, or to use a specific backend to drain new requests
13159 when the socket is considered saturated. This offers the ability to assign
13160 different limits to different listening ports or addresses. See also the
13161 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010013162
Willy Tarreau16e01562016-08-09 16:46:18 +020013163dst_is_local : boolean
13164 Returns true if the destination address of the incoming connection is local
13165 to the system, or false if the address doesn't exist on the system, meaning
13166 that it was intercepted in transparent mode. It can be useful to apply
13167 certain rules by default to forwarded traffic and other rules to the traffic
13168 targetting the real address of the machine. For example the stats page could
13169 be delivered only on this address, or SSH access could be locally redirected.
13170 Please note that the check involves a few system calls, so it's better to do
13171 it only once per connection.
13172
Willy Tarreau74ca5042013-06-11 23:12:07 +020013173dst_port : integer
13174 Returns an integer value corresponding to the destination TCP port of the
13175 connection on the client side, which is the port the client connected to.
13176 This might be used when running in transparent mode, when assigning dynamic
13177 ports to some clients for a whole application session, to stick all users to
13178 a same server, or to pass the destination port information to a server using
13179 an HTTP header.
13180
Emeric Brun4f603012017-01-05 15:11:44 +010013181fc_rcvd_proxy : boolean
13182 Returns true if the client initiated the connection with a PROXY protocol
13183 header.
13184
Thierry Fournier / OZON.IO6310bef2016-07-24 20:16:50 +020013185fc_rtt(<unit>) : integer
13186 Returns the Round Trip Time (RTT) measured by the kernel for the client
13187 connection. <unit> is facultative, by default the unit is milliseconds. <unit>
13188 can be set to "ms" for milliseconds or "us" for microseconds. If the server
13189 connection is not established, if the connection is not TCP or if the
13190 operating system does not support TCP_INFO, for example Linux kernels before
13191 2.4, the sample fetch fails.
13192
13193fc_rttvar(<unit>) : integer
13194 Returns the Round Trip Time (RTT) variance measured by the kernel for the
13195 client connection. <unit> is facultative, by default the unit is milliseconds.
13196 <unit> can be set to "ms" for milliseconds or "us" for microseconds. If the
13197 server connection is not established, if the connection is not TCP or if the
13198 operating system does not support TCP_INFO, for example Linux kernels before
13199 2.4, the sample fetch fails.
13200
Joe Williams30fcd392016-08-10 07:06:44 -070013201fc_unacked(<unit>) : integer
13202 Returns the unacked counter measured by the kernel for the client connection.
13203 If the server connection is not established, if the connection is not TCP or
13204 if the operating system does not support TCP_INFO, for example Linux kernels
13205 before 2.4, the sample fetch fails.
13206
13207fc_sacked(<unit>) : integer
13208 Returns the sacked counter measured by the kernel for the client connection.
13209 If the server connection is not established, if the connection is not TCP or
13210 if the operating system does not support TCP_INFO, for example Linux kernels
13211 before 2.4, the sample fetch fails.
13212
13213fc_retrans(<unit>) : integer
13214 Returns the retransmits counter measured by the kernel for the client
13215 connection. If the server connection is not established, if the connection is
13216 not TCP or if the operating system does not support TCP_INFO, for example
13217 Linux kernels before 2.4, the sample fetch fails.
13218
13219fc_fackets(<unit>) : integer
13220 Returns the fack counter measured by the kernel for the client
13221 connection. If the server connection is not established, if the connection is
13222 not TCP or if the operating system does not support TCP_INFO, for example
13223 Linux kernels before 2.4, the sample fetch fails.
13224
13225fc_lost(<unit>) : integer
13226 Returns the lost counter measured by the kernel for the client
13227 connection. If the server connection is not established, if the connection is
13228 not TCP or if the operating system does not support TCP_INFO, for example
13229 Linux kernels before 2.4, the sample fetch fails.
13230
13231fc_reordering(<unit>) : integer
13232 Returns the reordering counter measured by the kernel for the client
13233 connection. If the server connection is not established, if the connection is
13234 not TCP or if the operating system does not support TCP_INFO, for example
13235 Linux kernels before 2.4, the sample fetch fails.
13236
Willy Tarreau74ca5042013-06-11 23:12:07 +020013237fe_id : integer
13238 Returns an integer containing the current frontend's id. It can be used in
Marcin Deranek6e413ed2016-12-13 12:40:01 +010013239 backends to check from which frontend it was called, or to stick all users
Willy Tarreau74ca5042013-06-11 23:12:07 +020013240 coming via a same frontend to the same server.
13241
Marcin Deranekd2471c22016-12-12 14:08:05 +010013242fe_name : string
13243 Returns a string containing the current frontend's name. It can be used in
13244 backends to check from which frontend it was called, or to stick all users
13245 coming via a same frontend to the same server.
13246
Cyril Bonté62ba8702014-04-22 23:52:25 +020013247sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013248sc0_bytes_in_rate([<table>]) : integer
13249sc1_bytes_in_rate([<table>]) : integer
13250sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013251 Returns the average client-to-server bytes rate from the currently tracked
13252 counters, measured in amount of bytes over the period configured in the
13253 table. See also src_bytes_in_rate.
13254
Cyril Bonté62ba8702014-04-22 23:52:25 +020013255sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013256sc0_bytes_out_rate([<table>]) : integer
13257sc1_bytes_out_rate([<table>]) : integer
13258sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013259 Returns the average server-to-client bytes rate from the currently tracked
13260 counters, measured in amount of bytes over the period configured in the
13261 table. See also src_bytes_out_rate.
13262
Cyril Bonté62ba8702014-04-22 23:52:25 +020013263sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013264sc0_clr_gpc0([<table>]) : integer
13265sc1_clr_gpc0([<table>]) : integer
13266sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020013267 Clears the first General Purpose Counter associated to the currently tracked
13268 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010013269 stored value is zero, so first invocation will always return zero. This is
13270 typically used as a second ACL in an expression in order to mark a connection
13271 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020013272
13273 # block if 5 consecutive requests continue to come faster than 10 sess
13274 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020013275 acl abuse sc0_http_req_rate gt 10
13276 acl kill sc0_inc_gpc0 gt 5
13277 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020013278 tcp-request connection accept if !abuse save
13279 tcp-request connection reject if abuse kill
13280
Cyril Bonté62ba8702014-04-22 23:52:25 +020013281sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013282sc0_conn_cnt([<table>]) : integer
13283sc1_conn_cnt([<table>]) : integer
13284sc2_conn_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013285 Returns the cumulated number of incoming connections from currently tracked
13286 counters. See also src_conn_cnt.
13287
Cyril Bonté62ba8702014-04-22 23:52:25 +020013288sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013289sc0_conn_cur([<table>]) : integer
13290sc1_conn_cur([<table>]) : integer
13291sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013292 Returns the current amount of concurrent connections tracking the same
13293 tracked counters. This number is automatically incremented when tracking
13294 begins and decremented when tracking stops. See also src_conn_cur.
13295
Cyril Bonté62ba8702014-04-22 23:52:25 +020013296sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013297sc0_conn_rate([<table>]) : integer
13298sc1_conn_rate([<table>]) : integer
13299sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013300 Returns the average connection rate from the currently tracked counters,
13301 measured in amount of connections over the period configured in the table.
13302 See also src_conn_rate.
13303
Cyril Bonté62ba8702014-04-22 23:52:25 +020013304sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013305sc0_get_gpc0([<table>]) : integer
13306sc1_get_gpc0([<table>]) : integer
13307sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013308 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013309 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020013310
Thierry FOURNIER236657b2015-08-19 08:25:14 +020013311sc_get_gpt0(<ctr>[,<table>]) : integer
13312sc0_get_gpt0([<table>]) : integer
13313sc1_get_gpt0([<table>]) : integer
13314sc2_get_gpt0([<table>]) : integer
13315 Returns the value of the first General Purpose Tag associated to the
13316 currently tracked counters. See also src_get_gpt0.
13317
Cyril Bonté62ba8702014-04-22 23:52:25 +020013318sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013319sc0_gpc0_rate([<table>]) : integer
13320sc1_gpc0_rate([<table>]) : integer
13321sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020013322 Returns the average increment rate of the first General Purpose Counter
13323 associated to the currently tracked counters. It reports the frequency
13324 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013325 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
13326 that the "gpc0_rate" counter must be stored in the stick-table for a value to
13327 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020013328
Cyril Bonté62ba8702014-04-22 23:52:25 +020013329sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013330sc0_http_err_cnt([<table>]) : integer
13331sc1_http_err_cnt([<table>]) : integer
13332sc2_http_err_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013333 Returns the cumulated number of HTTP errors from the currently tracked
13334 counters. This includes the both request errors and 4xx error responses.
13335 See also src_http_err_cnt.
13336
Cyril Bonté62ba8702014-04-22 23:52:25 +020013337sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013338sc0_http_err_rate([<table>]) : integer
13339sc1_http_err_rate([<table>]) : integer
13340sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013341 Returns the average rate of HTTP errors from the currently tracked counters,
13342 measured in amount of errors over the period configured in the table. This
13343 includes the both request errors and 4xx error responses. See also
13344 src_http_err_rate.
13345
Cyril Bonté62ba8702014-04-22 23:52:25 +020013346sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013347sc0_http_req_cnt([<table>]) : integer
13348sc1_http_req_cnt([<table>]) : integer
13349sc2_http_req_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013350 Returns the cumulated number of HTTP requests from the currently tracked
13351 counters. This includes every started request, valid or not. See also
13352 src_http_req_cnt.
13353
Cyril Bonté62ba8702014-04-22 23:52:25 +020013354sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013355sc0_http_req_rate([<table>]) : integer
13356sc1_http_req_rate([<table>]) : integer
13357sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013358 Returns the average rate of HTTP requests from the currently tracked
13359 counters, measured in amount of requests over the period configured in
13360 the table. This includes every started request, valid or not. See also
13361 src_http_req_rate.
13362
Cyril Bonté62ba8702014-04-22 23:52:25 +020013363sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013364sc0_inc_gpc0([<table>]) : integer
13365sc1_inc_gpc0([<table>]) : integer
13366sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013367 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010013368 tracked counters, and returns its new value. Before the first invocation,
13369 the stored value is zero, so first invocation will increase it to 1 and will
13370 return 1. This is typically used as a second ACL in an expression in order
13371 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020013372
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020013373 acl abuse sc0_http_req_rate gt 10
13374 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020013375 tcp-request connection reject if abuse kill
13376
Cyril Bonté62ba8702014-04-22 23:52:25 +020013377sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013378sc0_kbytes_in([<table>]) : integer
13379sc1_kbytes_in([<table>]) : integer
13380sc2_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020013381 Returns the total amount of client-to-server data from the currently tracked
13382 counters, measured in kilobytes. The test is currently performed on 32-bit
13383 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020013384
Cyril Bonté62ba8702014-04-22 23:52:25 +020013385sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013386sc0_kbytes_out([<table>]) : integer
13387sc1_kbytes_out([<table>]) : integer
13388sc2_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020013389 Returns the total amount of server-to-client data from the currently tracked
13390 counters, measured in kilobytes. The test is currently performed on 32-bit
13391 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020013392
Cyril Bonté62ba8702014-04-22 23:52:25 +020013393sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013394sc0_sess_cnt([<table>]) : integer
13395sc1_sess_cnt([<table>]) : integer
13396sc2_sess_cnt([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013397 Returns the cumulated number of incoming connections that were transformed
13398 into sessions, which means that they were accepted by a "tcp-request
13399 connection" rule, from the currently tracked counters. A backend may count
13400 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040013401 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020013402 with the client. See also src_sess_cnt.
13403
Cyril Bonté62ba8702014-04-22 23:52:25 +020013404sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013405sc0_sess_rate([<table>]) : integer
13406sc1_sess_rate([<table>]) : integer
13407sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020013408 Returns the average session rate from the currently tracked counters,
13409 measured in amount of sessions over the period configured in the table. A
13410 session is a connection that got past the early "tcp-request connection"
13411 rules. A backend may count more sessions than connections because each
13412 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040013413 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020013414
Cyril Bonté62ba8702014-04-22 23:52:25 +020013415sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020013416sc0_tracked([<table>]) : boolean
13417sc1_tracked([<table>]) : boolean
13418sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020013419 Returns true if the designated session counter is currently being tracked by
13420 the current session. This can be useful when deciding whether or not we want
13421 to set some values in a header passed to the server.
13422
Cyril Bonté62ba8702014-04-22 23:52:25 +020013423sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020013424sc0_trackers([<table>]) : integer
13425sc1_trackers([<table>]) : integer
13426sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010013427 Returns the current amount of concurrent connections tracking the same
13428 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020013429 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010013430 that it does not rely on any stored information but on the table's reference
13431 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020013432 may sometimes be more suited for layer7 tracking. It can be used to tell a
13433 server how many concurrent connections there are from a given address for
13434 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010013435
Willy Tarreau74ca5042013-06-11 23:12:07 +020013436so_id : integer
13437 Returns an integer containing the current listening socket's id. It is useful
13438 in frontends involving many "bind" lines, or to stick all users coming via a
13439 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010013440
Willy Tarreau74ca5042013-06-11 23:12:07 +020013441src : ip
13442 This is the source IPv4 address of the client of the session. It is of type
13443 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
13444 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
13445 TCP-level source address which is used, and not the address of a client
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010013446 behind a proxy. However if the "accept-proxy" or "accept-netscaler-cip" bind
13447 directive is used, it can be the address of a client behind another
13448 PROXY-protocol compatible component for all rule sets except
13449 "tcp-request connection" which sees the real address.
Willy Tarreaud63335a2010-02-26 12:56:52 +010013450
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013451 Example:
13452 # add an HTTP header in requests with the originating address' country
13453 http-request set-header X-Country %[src,map_ip(geoip.lst)]
13454
Willy Tarreau74ca5042013-06-11 23:12:07 +020013455src_bytes_in_rate([<table>]) : integer
13456 Returns the average bytes rate from the incoming connection's source address
13457 in the current proxy's stick-table or in the designated stick-table, measured
13458 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013459 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013460
Willy Tarreau74ca5042013-06-11 23:12:07 +020013461src_bytes_out_rate([<table>]) : integer
13462 Returns the average bytes rate to the incoming connection's source address in
13463 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020013464 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013465 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013466
Willy Tarreau74ca5042013-06-11 23:12:07 +020013467src_clr_gpc0([<table>]) : integer
13468 Clears the first General Purpose Counter associated to the incoming
13469 connection's source address in the current proxy's stick-table or in the
13470 designated stick-table, and returns its previous value. If the address is not
13471 found, an entry is created and 0 is returned. This is typically used as a
13472 second ACL in an expression in order to mark a connection when a first ACL
13473 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020013474
13475 # block if 5 consecutive requests continue to come faster than 10 sess
13476 # per second, and reset the counter as soon as the traffic slows down.
13477 acl abuse src_http_req_rate gt 10
13478 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010013479 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020013480 tcp-request connection accept if !abuse save
13481 tcp-request connection reject if abuse kill
13482
Willy Tarreau74ca5042013-06-11 23:12:07 +020013483src_conn_cnt([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020013484 Returns the cumulated number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020013485 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020013486 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013487 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013488
Willy Tarreau74ca5042013-06-11 23:12:07 +020013489src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020013490 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020013491 current incoming connection's source address in the current proxy's
13492 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013493 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013494
Willy Tarreau74ca5042013-06-11 23:12:07 +020013495src_conn_rate([<table>]) : integer
13496 Returns the average connection rate from the incoming connection's source
13497 address in the current proxy's stick-table or in the designated stick-table,
13498 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013499 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013500
Willy Tarreau74ca5042013-06-11 23:12:07 +020013501src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020013502 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020013503 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020013504 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013505 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013506
Thierry FOURNIER236657b2015-08-19 08:25:14 +020013507src_get_gpt0([<table>]) : integer
13508 Returns the value of the first General Purpose Tag associated to the
13509 incoming connection's source address in the current proxy's stick-table or in
13510 the designated stick-table. If the address is not found, zero is returned.
13511 See also sc/sc0/sc1/sc2_get_gpt0.
13512
Willy Tarreau74ca5042013-06-11 23:12:07 +020013513src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020013514 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020013515 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020013516 stick-table or in the designated stick-table. It reports the frequency
13517 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013518 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
13519 that the "gpc0_rate" counter must be stored in the stick-table for a value to
13520 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020013521
Willy Tarreau74ca5042013-06-11 23:12:07 +020013522src_http_err_cnt([<table>]) : integer
13523 Returns the cumulated number of HTTP errors from the incoming connection's
13524 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020013525 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013526 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020013527 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013528
Willy Tarreau74ca5042013-06-11 23:12:07 +020013529src_http_err_rate([<table>]) : integer
13530 Returns the average rate of HTTP errors from the incoming connection's source
13531 address in the current proxy's stick-table or in the designated stick-table,
13532 measured in amount of errors over the period configured in the table. This
13533 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013534 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013535
Willy Tarreau74ca5042013-06-11 23:12:07 +020013536src_http_req_cnt([<table>]) : integer
13537 Returns the cumulated number of HTTP requests from the incoming connection's
13538 source address in the current proxy's stick-table or in the designated stick-
13539 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013540 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013541
Willy Tarreau74ca5042013-06-11 23:12:07 +020013542src_http_req_rate([<table>]) : integer
13543 Returns the average rate of HTTP requests from the incoming connection's
13544 source address in the current proxy's stick-table or in the designated stick-
13545 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020013546 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013547 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013548
Willy Tarreau74ca5042013-06-11 23:12:07 +020013549src_inc_gpc0([<table>]) : integer
13550 Increments the first General Purpose Counter associated to the incoming
13551 connection's source address in the current proxy's stick-table or in the
13552 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020013553 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013554 This is typically used as a second ACL in an expression in order to mark a
13555 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020013556
13557 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010013558 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020013559 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020013560
Willy Tarreau16e01562016-08-09 16:46:18 +020013561src_is_local : boolean
13562 Returns true if the source address of the incoming connection is local to the
13563 system, or false if the address doesn't exist on the system, meaning that it
13564 comes from a remote machine. Note that UNIX addresses are considered local.
13565 It can be useful to apply certain access restrictions based on where the
13566 client comes from (eg: require auth or https for remote machines). Please
13567 note that the check involves a few system calls, so it's better to do it only
13568 once per connection.
13569
Willy Tarreau74ca5042013-06-11 23:12:07 +020013570src_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020013571 Returns the total amount of data received from the incoming connection's
13572 source address in the current proxy's stick-table or in the designated
13573 stick-table, measured in kilobytes. If the address is not found, zero is
13574 returned. The test is currently performed on 32-bit integers, which limits
13575 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013576
Willy Tarreau74ca5042013-06-11 23:12:07 +020013577src_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020013578 Returns the total amount of data sent to the incoming connection's source
13579 address in the current proxy's stick-table or in the designated stick-table,
13580 measured in kilobytes. If the address is not found, zero is returned. The
13581 test is currently performed on 32-bit integers, which limits values to 4
13582 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020013583
Willy Tarreau74ca5042013-06-11 23:12:07 +020013584src_port : integer
13585 Returns an integer value corresponding to the TCP source port of the
13586 connection on the client side, which is the port the client connected from.
13587 Usage of this function is very limited as modern protocols do not care much
13588 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010013589
Willy Tarreau74ca5042013-06-11 23:12:07 +020013590src_sess_cnt([<table>]) : integer
13591 Returns the cumulated number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020013592 connection's source IPv4 address in the current proxy's stick-table or in the
13593 designated stick-table, that were transformed into sessions, which means that
13594 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013595 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013596
Willy Tarreau74ca5042013-06-11 23:12:07 +020013597src_sess_rate([<table>]) : integer
13598 Returns the average session rate from the incoming connection's source
13599 address in the current proxy's stick-table or in the designated stick-table,
13600 measured in amount of sessions over the period configured in the table. A
13601 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020013602 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020013603
Willy Tarreau74ca5042013-06-11 23:12:07 +020013604src_updt_conn_cnt([<table>]) : integer
13605 Creates or updates the entry associated to the incoming connection's source
13606 address in the current proxy's stick-table or in the designated stick-table.
13607 This table must be configured to store the "conn_cnt" data type, otherwise
13608 the match will be ignored. The current count is incremented by one, and the
13609 expiration timer refreshed. The updated count is returned, so this match
13610 can't return zero. This was used to reject service abusers based on their
13611 source address. Note: it is recommended to use the more complete "track-sc*"
13612 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020013613
13614 Example :
13615 # This frontend limits incoming SSH connections to 3 per 10 second for
13616 # each source address, and rejects excess connections until a 10 second
13617 # silence is observed. At most 20 addresses are tracked.
13618 listen ssh
13619 bind :22
13620 mode tcp
13621 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020013622 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020013623 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020013624 server local 127.0.0.1:22
13625
Willy Tarreau74ca5042013-06-11 23:12:07 +020013626srv_id : integer
13627 Returns an integer containing the server's id when processing the response.
13628 While it's almost only used with ACLs, it may be used for logging or
13629 debugging.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020013630
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200136317.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020013632----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020013633
Willy Tarreau74ca5042013-06-11 23:12:07 +020013634The layer 5 usually describes just the session layer which in haproxy is
13635closest to the session once all the connection handshakes are finished, but
13636when no content is yet made available. The fetch methods described here are
13637usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013638future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020013639
Ben Shillitof25e8e52016-12-02 14:25:37 +00001364051d.all(<prop>[,<prop>*]) : string
13641 Returns values for the properties requested as a string, where values are
13642 separated by the delimiter specified with "51degrees-property-separator".
13643 The device is identified using all the important HTTP headers from the
13644 request. The function can be passed up to five property names, and if a
13645 property name can't be found, the value "NoData" is returned.
13646
13647 Example :
13648 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
13649 # containing the three properties requested using all relevant headers from
13650 # the request.
13651 frontend http-in
13652 bind *:8081
13653 default_backend servers
13654 http-request set-header X-51D-DeviceTypeMobileTablet \
13655 %[51d.all(DeviceType,IsMobile,IsTablet)]
13656
Emeric Brun645ae792014-04-30 14:21:06 +020013657ssl_bc : boolean
13658 Returns true when the back connection was made via an SSL/TLS transport
13659 layer and is locally deciphered. This means the outgoing connection was made
13660 other a server with the "ssl" option.
13661
13662ssl_bc_alg_keysize : integer
13663 Returns the symmetric cipher key size supported in bits when the outgoing
13664 connection was made over an SSL/TLS transport layer.
13665
13666ssl_bc_cipher : string
13667 Returns the name of the used cipher when the outgoing connection was made
13668 over an SSL/TLS transport layer.
13669
13670ssl_bc_protocol : string
13671 Returns the name of the used protocol when the outgoing connection was made
13672 over an SSL/TLS transport layer.
13673
Emeric Brunb73a9b02014-04-30 18:49:19 +020013674ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020013675 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020013676 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
13677 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
Emeric Brun645ae792014-04-30 14:21:06 +020013678
13679ssl_bc_session_id : binary
13680 Returns the SSL ID of the back connection when the outgoing connection was
13681 made over an SSL/TLS transport layer. It is useful to log if we want to know
13682 if session was reused or not.
13683
13684ssl_bc_use_keysize : integer
13685 Returns the symmetric cipher key size used in bits when the outgoing
13686 connection was made over an SSL/TLS transport layer.
13687
Willy Tarreau74ca5042013-06-11 23:12:07 +020013688ssl_c_ca_err : integer
13689 When the incoming connection was made over an SSL/TLS transport layer,
13690 returns the ID of the first error detected during verification of the client
13691 certificate at depth > 0, or 0 if no error was encountered during this
13692 verification process. Please refer to your SSL library's documentation to
13693 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020013694
Willy Tarreau74ca5042013-06-11 23:12:07 +020013695ssl_c_ca_err_depth : integer
13696 When the incoming connection was made over an SSL/TLS transport layer,
13697 returns the depth in the CA chain of the first error detected during the
13698 verification of the client certificate. If no error is encountered, 0 is
13699 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010013700
Emeric Brun43e79582014-10-29 19:03:26 +010013701ssl_c_der : binary
13702 Returns the DER formatted certificate presented by the client when the
13703 incoming connection was made over an SSL/TLS transport layer. When used for
13704 an ACL, the value(s) to match against can be passed in hexadecimal form.
13705
Willy Tarreau74ca5042013-06-11 23:12:07 +020013706ssl_c_err : integer
13707 When the incoming connection was made over an SSL/TLS transport layer,
13708 returns the ID of the first error detected during verification at depth 0, or
13709 0 if no error was encountered during this verification process. Please refer
13710 to your SSL library's documentation to find the exhaustive list of error
13711 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020013712
Willy Tarreau74ca5042013-06-11 23:12:07 +020013713ssl_c_i_dn([<entry>[,<occ>]]) : string
13714 When the incoming connection was made over an SSL/TLS transport layer,
13715 returns the full distinguished name of the issuer of the certificate
13716 presented by the client when no <entry> is specified, or the value of the
13717 first given entry found from the beginning of the DN. If a positive/negative
13718 occurrence number is specified as the optional second argument, it returns
13719 the value of the nth given entry value from the beginning/end of the DN.
13720 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
13721 "ssl_c_i_dn(CN)" retrieves the common name.
Willy Tarreau62644772008-07-16 18:36:06 +020013722
Willy Tarreau74ca5042013-06-11 23:12:07 +020013723ssl_c_key_alg : string
13724 Returns the name of the algorithm used to generate the key of the certificate
13725 presented by the client when the incoming connection was made over an SSL/TLS
13726 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020013727
Willy Tarreau74ca5042013-06-11 23:12:07 +020013728ssl_c_notafter : string
13729 Returns the end date presented by the client as a formatted string
13730 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
13731 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020013732
Willy Tarreau74ca5042013-06-11 23:12:07 +020013733ssl_c_notbefore : string
13734 Returns the start date presented by the client as a formatted string
13735 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
13736 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010013737
Willy Tarreau74ca5042013-06-11 23:12:07 +020013738ssl_c_s_dn([<entry>[,<occ>]]) : string
13739 When the incoming connection was made over an SSL/TLS transport layer,
13740 returns the full distinguished name of the subject of the certificate
13741 presented by the client when no <entry> is specified, or the value of the
13742 first given entry found from the beginning of the DN. If a positive/negative
13743 occurrence number is specified as the optional second argument, it returns
13744 the value of the nth given entry value from the beginning/end of the DN.
13745 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
13746 "ssl_c_s_dn(CN)" retrieves the common name.
Willy Tarreaub6672b52011-12-12 17:23:41 +010013747
Willy Tarreau74ca5042013-06-11 23:12:07 +020013748ssl_c_serial : binary
13749 Returns the serial of the certificate presented by the client when the
13750 incoming connection was made over an SSL/TLS transport layer. When used for
13751 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020013752
Willy Tarreau74ca5042013-06-11 23:12:07 +020013753ssl_c_sha1 : binary
13754 Returns the SHA-1 fingerprint of the certificate presented by the client when
13755 the incoming connection was made over an SSL/TLS transport layer. This can be
13756 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau2d0caa32014-07-02 19:01:22 +020013757 Note that the output is binary, so if you want to pass that signature to the
13758 server, you need to encode it in hex or base64, such as in the example below:
13759
13760 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020013761
Willy Tarreau74ca5042013-06-11 23:12:07 +020013762ssl_c_sig_alg : string
13763 Returns the name of the algorithm used to sign the certificate presented by
13764 the client when the incoming connection was made over an SSL/TLS transport
13765 layer.
Emeric Brun87855892012-10-17 17:39:35 +020013766
Willy Tarreau74ca5042013-06-11 23:12:07 +020013767ssl_c_used : boolean
13768 Returns true if current SSL session uses a client certificate even if current
13769 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020013770
Willy Tarreau74ca5042013-06-11 23:12:07 +020013771ssl_c_verify : integer
13772 Returns the verify result error ID when the incoming connection was made over
13773 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
13774 refer to your SSL library's documentation for an exhaustive list of error
13775 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020013776
Willy Tarreau74ca5042013-06-11 23:12:07 +020013777ssl_c_version : integer
13778 Returns the version of the certificate presented by the client when the
13779 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020013780
Emeric Brun43e79582014-10-29 19:03:26 +010013781ssl_f_der : binary
13782 Returns the DER formatted certificate presented by the frontend when the
13783 incoming connection was made over an SSL/TLS transport layer. When used for
13784 an ACL, the value(s) to match against can be passed in hexadecimal form.
13785
Willy Tarreau74ca5042013-06-11 23:12:07 +020013786ssl_f_i_dn([<entry>[,<occ>]]) : string
13787 When the incoming connection was made over an SSL/TLS transport layer,
13788 returns the full distinguished name of the issuer of the certificate
13789 presented by the frontend when no <entry> is specified, or the value of the
13790 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020013791 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020013792 the value of the nth given entry value from the beginning/end of the DN.
13793 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
13794 "ssl_f_i_dn(CN)" retrieves the common name.
Emeric Brun87855892012-10-17 17:39:35 +020013795
Willy Tarreau74ca5042013-06-11 23:12:07 +020013796ssl_f_key_alg : string
13797 Returns the name of the algorithm used to generate the key of the certificate
13798 presented by the frontend when the incoming connection was made over an
13799 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020013800
Willy Tarreau74ca5042013-06-11 23:12:07 +020013801ssl_f_notafter : string
13802 Returns the end date presented by the frontend as a formatted string
13803 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
13804 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020013805
Willy Tarreau74ca5042013-06-11 23:12:07 +020013806ssl_f_notbefore : string
13807 Returns the start date presented by the frontend as a formatted string
13808 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
13809 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020013810
Willy Tarreau74ca5042013-06-11 23:12:07 +020013811ssl_f_s_dn([<entry>[,<occ>]]) : string
13812 When the incoming connection was made over an SSL/TLS transport layer,
13813 returns the full distinguished name of the subject of the certificate
13814 presented by the frontend when no <entry> is specified, or the value of the
13815 first given entry found from the beginning of the DN. If a positive/negative
13816 occurrence number is specified as the optional second argument, it returns
13817 the value of the nth given entry value from the beginning/end of the DN.
13818 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
13819 "ssl_f_s_dn(CN)" retrieves the common name.
Emeric Brunce5ad802012-10-22 14:11:22 +020013820
Willy Tarreau74ca5042013-06-11 23:12:07 +020013821ssl_f_serial : binary
13822 Returns the serial of the certificate presented by the frontend when the
13823 incoming connection was made over an SSL/TLS transport layer. When used for
13824 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020013825
Emeric Brun55f4fa82014-04-30 17:11:25 +020013826ssl_f_sha1 : binary
13827 Returns the SHA-1 fingerprint of the certificate presented by the frontend
13828 when the incoming connection was made over an SSL/TLS transport layer. This
13829 can be used to know which certificate was chosen using SNI.
13830
Willy Tarreau74ca5042013-06-11 23:12:07 +020013831ssl_f_sig_alg : string
13832 Returns the name of the algorithm used to sign the certificate presented by
13833 the frontend when the incoming connection was made over an SSL/TLS transport
13834 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020013835
Willy Tarreau74ca5042013-06-11 23:12:07 +020013836ssl_f_version : integer
13837 Returns the version of the certificate presented by the frontend when the
13838 incoming connection was made over an SSL/TLS transport layer.
13839
13840ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020013841 Returns true when the front connection was made via an SSL/TLS transport
13842 layer and is locally deciphered. This means it has matched a socket declared
13843 with a "bind" line having the "ssl" option.
13844
Willy Tarreau74ca5042013-06-11 23:12:07 +020013845 Example :
13846 # This passes "X-Proto: https" to servers when client connects over SSL
13847 listen http-https
13848 bind :80
13849 bind :443 ssl crt /etc/haproxy.pem
13850 http-request add-header X-Proto https if { ssl_fc }
13851
13852ssl_fc_alg_keysize : integer
13853 Returns the symmetric cipher key size supported in bits when the incoming
13854 connection was made over an SSL/TLS transport layer.
13855
13856ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013857 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020013858 incoming connection made via a TLS transport layer and locally deciphered by
13859 haproxy. The result is a string containing the protocol name advertised by
13860 the client. The SSL library must have been built with support for TLS
13861 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
13862 not advertised unless the "alpn" keyword on the "bind" line specifies a
13863 protocol list. Also, nothing forces the client to pick a protocol from this
13864 list, any other one may be requested. The TLS ALPN extension is meant to
13865 replace the TLS NPN extension. See also "ssl_fc_npn".
13866
Willy Tarreau74ca5042013-06-11 23:12:07 +020013867ssl_fc_cipher : string
13868 Returns the name of the used cipher when the incoming connection was made
13869 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020013870
Willy Tarreau74ca5042013-06-11 23:12:07 +020013871ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020013872 Returns true if a client certificate is present in an incoming connection over
13873 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010013874 Note: on SSL session resumption with Session ID or TLS ticket, client
13875 certificate is not present in the current connection but may be retrieved
13876 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
13877 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020013878
Willy Tarreau74ca5042013-06-11 23:12:07 +020013879ssl_fc_has_sni : boolean
13880 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020013881 in an incoming connection was made over an SSL/TLS transport layer. Returns
13882 true when the incoming connection presents a TLS SNI field. This requires
13883 that the SSL library is build with support for TLS extensions enabled (check
13884 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020013885
Nenad Merdanovic1516fe32016-05-17 03:31:21 +020013886ssl_fc_is_resumed : boolean
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020013887 Returns true if the SSL/TLS session has been resumed through the use of
13888 SSL session cache or TLS tickets.
13889
Willy Tarreau74ca5042013-06-11 23:12:07 +020013890ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013891 This extracts the Next Protocol Negotiation field from an incoming connection
Willy Tarreau74ca5042013-06-11 23:12:07 +020013892 made via a TLS transport layer and locally deciphered by haproxy. The result
13893 is a string containing the protocol name advertised by the client. The SSL
13894 library must have been built with support for TLS extensions enabled (check
13895 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
13896 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
13897 forces the client to pick a protocol from this list, any other one may be
13898 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020013899
Willy Tarreau74ca5042013-06-11 23:12:07 +020013900ssl_fc_protocol : string
13901 Returns the name of the used protocol when the incoming connection was made
13902 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020013903
Emeric Brunb73a9b02014-04-30 18:49:19 +020013904ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040013905 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020013906 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
13907 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040013908
Willy Tarreau74ca5042013-06-11 23:12:07 +020013909ssl_fc_session_id : binary
13910 Returns the SSL ID of the front connection when the incoming connection was
13911 made over an SSL/TLS transport layer. It is useful to stick a given client to
13912 a server. It is important to note that some browsers refresh their session ID
13913 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020013914
Willy Tarreau74ca5042013-06-11 23:12:07 +020013915ssl_fc_sni : string
13916 This extracts the Server Name Indication TLS extension (SNI) field from an
13917 incoming connection made via an SSL/TLS transport layer and locally
13918 deciphered by haproxy. The result (when present) typically is a string
13919 matching the HTTPS host name (253 chars or less). The SSL library must have
13920 been built with support for TLS extensions enabled (check haproxy -vv).
13921
13922 This fetch is different from "req_ssl_sni" above in that it applies to the
13923 connection being deciphered by haproxy and not to SSL contents being blindly
13924 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020013925 requires that the SSL library is build with support for TLS extensions
13926 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020013927
Willy Tarreau74ca5042013-06-11 23:12:07 +020013928 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020013929 ssl_fc_sni_end : suffix match
13930 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020013931
Willy Tarreau74ca5042013-06-11 23:12:07 +020013932ssl_fc_use_keysize : integer
13933 Returns the symmetric cipher key size used in bits when the incoming
13934 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020013935
Willy Tarreaub6fb4202008-07-20 11:18:28 +020013936
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200139377.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020013938------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020013939
Willy Tarreau74ca5042013-06-11 23:12:07 +020013940Fetching samples from buffer contents is a bit different from the previous
13941sample fetches above because the sampled data are ephemeral. These data can
13942only be used when they're available and will be lost when they're forwarded.
13943For this reason, samples fetched from buffer contents during a request cannot
13944be used in a response for example. Even while the data are being fetched, they
13945can change. Sometimes it is necessary to set some delays or combine multiple
13946sample fetch methods to ensure that the expected data are complete and usable,
13947for example through TCP request content inspection. Please see the "tcp-request
13948content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020013949
Willy Tarreau74ca5042013-06-11 23:12:07 +020013950payload(<offset>,<length>) : binary (deprecated)
13951 This is an alias for "req.payload" when used in the context of a request (eg:
13952 "stick on", "stick match"), and for "res.payload" when used in the context of
13953 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010013954
Willy Tarreau74ca5042013-06-11 23:12:07 +020013955payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
13956 This is an alias for "req.payload_lv" when used in the context of a request
13957 (eg: "stick on", "stick match"), and for "res.payload_lv" when used in the
13958 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010013959
Willy Tarreau74ca5042013-06-11 23:12:07 +020013960req.len : integer
13961req_len : integer (deprecated)
13962 Returns an integer value corresponding to the number of bytes present in the
13963 request buffer. This is mostly used in ACL. It is important to understand
13964 that this test does not return false as long as the buffer is changing. This
13965 means that a check with equality to zero will almost always immediately match
13966 at the beginning of the session, while a test for more data will wait for
13967 that data to come in and return false only when haproxy is certain that no
13968 more data will come in. This test was designed to be used with TCP request
13969 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013970
Willy Tarreau74ca5042013-06-11 23:12:07 +020013971req.payload(<offset>,<length>) : binary
13972 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020013973 in the request buffer. As a special case, if the <length> argument is zero,
13974 the the whole buffer from <offset> to the end is extracted. This can be used
13975 with ACLs in order to check for the presence of some content in a buffer at
13976 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013977
Willy Tarreau74ca5042013-06-11 23:12:07 +020013978 ACL alternatives :
13979 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013980
Willy Tarreau74ca5042013-06-11 23:12:07 +020013981req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
13982 This extracts a binary block whose size is specified at <offset1> for <length>
13983 bytes, and which starts at <offset2> if specified or just after the length in
13984 the request buffer. The <offset2> parameter also supports relative offsets if
13985 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013986
Willy Tarreau74ca5042013-06-11 23:12:07 +020013987 ACL alternatives :
13988 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013989
Willy Tarreau74ca5042013-06-11 23:12:07 +020013990 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020013991
Willy Tarreau74ca5042013-06-11 23:12:07 +020013992req.proto_http : boolean
13993req_proto_http : boolean (deprecated)
13994 Returns true when data in the request buffer look like HTTP and correctly
13995 parses as such. It is the same parser as the common HTTP request parser which
13996 is used so there should be no surprises. The test does not match until the
13997 request is complete, failed or timed out. This test may be used to report the
13998 protocol in TCP logs, but the biggest use is to block TCP request analysis
13999 until a complete HTTP request is present in the buffer, for example to track
14000 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020014001
Willy Tarreau74ca5042013-06-11 23:12:07 +020014002 Example:
14003 # track request counts per "base" (concatenation of Host+URL)
14004 tcp-request inspect-delay 10s
14005 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014006 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020014007
Willy Tarreau74ca5042013-06-11 23:12:07 +020014008req.rdp_cookie([<name>]) : string
14009rdp_cookie([<name>]) : string (deprecated)
14010 When the request buffer looks like the RDP protocol, extracts the RDP cookie
14011 <name>, or any cookie if unspecified. The parser only checks for the first
14012 cookie, as illustrated in the RDP protocol specification. The cookie name is
14013 case insensitive. Generally the "MSTS" cookie name will be used, as it can
14014 contain the user name of the client connecting to the server if properly
14015 configured on the client. The "MSTSHASH" cookie is often used as well for
14016 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014017
Willy Tarreau74ca5042013-06-11 23:12:07 +020014018 This differs from "balance rdp-cookie" in that any balancing algorithm may be
14019 used and thus the distribution of clients to backend servers is not linked to
14020 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
14021 such as "balance roundrobin" or "balance leastconn" will lead to a more even
14022 distribution of clients to backend servers than the hash used by "balance
14023 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014024
Willy Tarreau74ca5042013-06-11 23:12:07 +020014025 ACL derivatives :
14026 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014027
Willy Tarreau74ca5042013-06-11 23:12:07 +020014028 Example :
14029 listen tse-farm
14030 bind 0.0.0.0:3389
14031 # wait up to 5s for an RDP cookie in the request
14032 tcp-request inspect-delay 5s
14033 tcp-request content accept if RDP_COOKIE
14034 # apply RDP cookie persistence
14035 persist rdp-cookie
14036 # Persist based on the mstshash cookie
14037 # This is only useful makes sense if
14038 # balance rdp-cookie is not used
14039 stick-table type string size 204800
14040 stick on req.rdp_cookie(mstshash)
14041 server srv1 1.1.1.1:3389
14042 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014043
Willy Tarreau74ca5042013-06-11 23:12:07 +020014044 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
14045 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014046
Willy Tarreau74ca5042013-06-11 23:12:07 +020014047req.rdp_cookie_cnt([name]) : integer
14048rdp_cookie_cnt([name]) : integer (deprecated)
14049 Tries to parse the request buffer as RDP protocol, then returns an integer
14050 corresponding to the number of RDP cookies found. If an optional cookie name
14051 is passed, only cookies matching this name are considered. This is mostly
14052 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014053
Willy Tarreau74ca5042013-06-11 23:12:07 +020014054 ACL derivatives :
14055 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014056
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020014057req.ssl_ec_ext : boolean
14058 Returns a boolean identifying if client sent the Supported Elliptic Curves
14059 Extension as defined in RFC4492, section 5.1. within the SSL ClientHello
Cyril Bonté307ee1e2015-09-28 23:16:06 +020014060 message. This can be used to present ECC compatible clients with EC
14061 certificate and to use RSA for all others, on the same IP address. Note that
14062 this only applies to raw contents found in the request buffer and not to
14063 contents deciphered via an SSL data layer, so this will not work with "bind"
14064 lines having the "ssl" option.
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020014065
Willy Tarreau74ca5042013-06-11 23:12:07 +020014066req.ssl_hello_type : integer
14067req_ssl_hello_type : integer (deprecated)
14068 Returns an integer value containing the type of the SSL hello message found
14069 in the request buffer if the buffer contains data that parse as a complete
14070 SSL (v3 or superior) client hello message. Note that this only applies to raw
14071 contents found in the request buffer and not to contents deciphered via an
14072 SSL data layer, so this will not work with "bind" lines having the "ssl"
14073 option. This is mostly used in ACL to detect presence of an SSL hello message
14074 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014075
Willy Tarreau74ca5042013-06-11 23:12:07 +020014076req.ssl_sni : string
14077req_ssl_sni : string (deprecated)
14078 Returns a string containing the value of the Server Name TLS extension sent
14079 by a client in a TLS stream passing through the request buffer if the buffer
14080 contains data that parse as a complete SSL (v3 or superior) client hello
14081 message. Note that this only applies to raw contents found in the request
14082 buffer and not to contents deciphered via an SSL data layer, so this will not
14083 work with "bind" lines having the "ssl" option. SNI normally contains the
14084 name of the host the client tries to connect to (for recent browsers). SNI is
14085 useful for allowing or denying access to certain hosts when SSL/TLS is used
14086 by the client. This test was designed to be used with TCP request content
14087 inspection. If content switching is needed, it is recommended to first wait
14088 for a complete client hello (type 1), like in the example below. See also
14089 "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014090
Willy Tarreau74ca5042013-06-11 23:12:07 +020014091 ACL derivatives :
14092 req_ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014093
Willy Tarreau74ca5042013-06-11 23:12:07 +020014094 Examples :
14095 # Wait for a client hello for at most 5 seconds
14096 tcp-request inspect-delay 5s
14097 tcp-request content accept if { req_ssl_hello_type 1 }
14098 use_backend bk_allow if { req_ssl_sni -f allowed_sites }
14099 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020014100
Pradeep Jindalbb2acf52015-09-29 10:12:57 +053014101req.ssl_st_ext : integer
14102 Returns 0 if the client didn't send a SessionTicket TLS Extension (RFC5077)
14103 Returns 1 if the client sent SessionTicket TLS Extension
14104 Returns 2 if the client also sent non-zero length TLS SessionTicket
14105 Note that this only applies to raw contents found in the request buffer and
14106 not to contents deciphered via an SSL data layer, so this will not work with
14107 "bind" lines having the "ssl" option. This can for example be used to detect
14108 whether the client sent a SessionTicket or not and stick it accordingly, if
14109 no SessionTicket then stick on SessionID or don't stick as there's no server
14110 side state is there when SessionTickets are in use.
14111
Willy Tarreau74ca5042013-06-11 23:12:07 +020014112req.ssl_ver : integer
14113req_ssl_ver : integer (deprecated)
14114 Returns an integer value containing the version of the SSL/TLS protocol of a
14115 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
14116 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
14117 composed of the major version multiplied by 65536, added to the minor
14118 version. Note that this only applies to raw contents found in the request
14119 buffer and not to contents deciphered via an SSL data layer, so this will not
14120 work with "bind" lines having the "ssl" option. The ACL version of the test
14121 matches against a decimal notation in the form MAJOR.MINOR (eg: 3.1). This
14122 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014123
Willy Tarreau74ca5042013-06-11 23:12:07 +020014124 ACL derivatives :
14125 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010014126
Willy Tarreau47e8eba2013-09-11 23:28:46 +020014127res.len : integer
14128 Returns an integer value corresponding to the number of bytes present in the
14129 response buffer. This is mostly used in ACL. It is important to understand
14130 that this test does not return false as long as the buffer is changing. This
14131 means that a check with equality to zero will almost always immediately match
14132 at the beginning of the session, while a test for more data will wait for
14133 that data to come in and return false only when haproxy is certain that no
14134 more data will come in. This test was designed to be used with TCP response
14135 content inspection.
14136
Willy Tarreau74ca5042013-06-11 23:12:07 +020014137res.payload(<offset>,<length>) : binary
14138 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020014139 in the response buffer. As a special case, if the <length> argument is zero,
14140 the the whole buffer from <offset> to the end is extracted. This can be used
14141 with ACLs in order to check for the presence of some content in a buffer at
14142 any location.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014143
Willy Tarreau74ca5042013-06-11 23:12:07 +020014144res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
14145 This extracts a binary block whose size is specified at <offset1> for <length>
14146 bytes, and which starts at <offset2> if specified or just after the length in
14147 the response buffer. The <offset2> parameter also supports relative offsets
14148 if prepended with a '+' or '-' sign.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014149
Willy Tarreau74ca5042013-06-11 23:12:07 +020014150 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014151
Willy Tarreau971f7b62015-09-29 14:06:59 +020014152res.ssl_hello_type : integer
14153rep_ssl_hello_type : integer (deprecated)
14154 Returns an integer value containing the type of the SSL hello message found
14155 in the response buffer if the buffer contains data that parses as a complete
14156 SSL (v3 or superior) hello message. Note that this only applies to raw
14157 contents found in the response buffer and not to contents deciphered via an
14158 SSL data layer, so this will not work with "server" lines having the "ssl"
14159 option. This is mostly used in ACL to detect presence of an SSL hello message
14160 that is supposed to contain an SSL session ID usable for stickiness.
14161
Willy Tarreau74ca5042013-06-11 23:12:07 +020014162wait_end : boolean
14163 This fetch either returns true when the inspection period is over, or does
14164 not fetch. It is only used in ACLs, in conjunction with content analysis to
14165 avoid returning a wrong verdict early. It may also be used to delay some
14166 actions, such as a delayed reject for some special addresses. Since it either
14167 stops the rules evaluation or immediately returns true, it is recommended to
14168 use this acl as the last one in a rule. Please note that the default ACL
14169 "WAIT_END" is always usable without prior declaration. This test was designed
14170 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014171
Willy Tarreau74ca5042013-06-11 23:12:07 +020014172 Examples :
14173 # delay every incoming request by 2 seconds
14174 tcp-request inspect-delay 2s
14175 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010014176
Willy Tarreau74ca5042013-06-11 23:12:07 +020014177 # don't immediately tell bad guys they are rejected
14178 tcp-request inspect-delay 10s
14179 acl goodguys src 10.0.0.0/24
14180 acl badguys src 10.0.1.0/24
14181 tcp-request content accept if goodguys
14182 tcp-request content reject if badguys WAIT_END
14183 tcp-request content reject
14184
14185
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200141867.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020014187--------------------------------------
14188
14189It is possible to fetch samples from HTTP contents, requests and responses.
14190This application layer is also called layer 7. It is only possible to fetch the
14191data in this section when a full HTTP request or response has been parsed from
14192its respective request or response buffer. This is always the case with all
14193HTTP specific rules and for sections running with "mode http". When using TCP
14194content inspection, it may be necessary to support an inspection delay in order
14195to let the request or response come in first. These fetches may require a bit
14196more CPU resources than the layer 4 ones, but not much since the request and
14197response are indexed.
14198
14199base : string
14200 This returns the concatenation of the first Host header and the path part of
14201 the request, which starts at the first slash and ends before the question
14202 mark. It can be useful in virtual hosted environments to detect URL abuses as
14203 well as to improve shared caches efficiency. Using this with a limited size
14204 stick table also allows one to collect statistics about most commonly
14205 requested objects by host/path. With ACLs it can allow simple content
14206 switching rules involving the host and the path at the same time, such as
14207 "www.example.com/favicon.ico". See also "path" and "uri".
14208
14209 ACL derivatives :
14210 base : exact string match
14211 base_beg : prefix match
14212 base_dir : subdir match
14213 base_dom : domain match
14214 base_end : suffix match
14215 base_len : length match
14216 base_reg : regex match
14217 base_sub : substring match
14218
14219base32 : integer
14220 This returns a 32-bit hash of the value returned by the "base" fetch method
14221 above. This is useful to track per-URL activity on high traffic sites without
14222 having to store all URLs. Instead a shorter hash is stored, saving a lot of
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020014223 memory. The output type is an unsigned integer. The hash function used is
14224 SDBM with full avalanche on the output. Technically, base32 is exactly equal
14225 to "base,sdbm(1)".
Willy Tarreau74ca5042013-06-11 23:12:07 +020014226
14227base32+src : binary
14228 This returns the concatenation of the base32 fetch above and the src fetch
14229 below. The resulting type is of type binary, with a size of 8 or 20 bytes
14230 depending on the source address family. This can be used to track per-IP,
14231 per-URL counters.
14232
William Lallemand65ad6e12014-01-31 15:08:02 +010014233capture.req.hdr(<idx>) : string
14234 This extracts the content of the header captured by the "capture request
14235 header", idx is the position of the capture keyword in the configuration.
14236 The first entry is an index of 0. See also: "capture request header".
14237
14238capture.req.method : string
14239 This extracts the METHOD of an HTTP request. It can be used in both request
14240 and response. Unlike "method", it can be used in both request and response
14241 because it's allocated.
14242
14243capture.req.uri : string
14244 This extracts the request's URI, which starts at the first slash and ends
14245 before the first space in the request (without the host part). Unlike "path"
14246 and "url", it can be used in both request and response because it's
14247 allocated.
14248
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020014249capture.req.ver : string
14250 This extracts the request's HTTP version and returns either "HTTP/1.0" or
14251 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
14252 logs because it relies on a persistent flag.
14253
William Lallemand65ad6e12014-01-31 15:08:02 +010014254capture.res.hdr(<idx>) : string
14255 This extracts the content of the header captured by the "capture response
14256 header", idx is the position of the capture keyword in the configuration.
14257 The first entry is an index of 0.
14258 See also: "capture response header"
14259
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020014260capture.res.ver : string
14261 This extracts the response's HTTP version and returns either "HTTP/1.0" or
14262 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
14263 persistent flag.
14264
Willy Tarreaua5910cc2015-05-02 00:46:08 +020014265req.body : binary
14266 This returns the HTTP request's available body as a block of data. It
14267 requires that the request body has been buffered made available using
14268 "option http-buffer-request". In case of chunked-encoded body, currently only
14269 the first chunk is analyzed.
14270
Thierry FOURNIER9826c772015-05-20 15:50:54 +020014271req.body_param([<name>) : string
14272 This fetch assumes that the body of the POST request is url-encoded. The user
14273 can check if the "content-type" contains the value
14274 "application/x-www-form-urlencoded". This extracts the first occurrence of the
14275 parameter <name> in the body, which ends before '&'. The parameter name is
14276 case-sensitive. If no name is given, any parameter will match, and the first
14277 one will be returned. The result is a string corresponding to the value of the
14278 parameter <name> as presented in the request body (no URL decoding is
14279 performed). Note that the ACL version of this fetch iterates over multiple
14280 parameters and will iteratively report all parameters values if no name is
14281 given.
14282
Willy Tarreaua5910cc2015-05-02 00:46:08 +020014283req.body_len : integer
14284 This returns the length of the HTTP request's available body in bytes. It may
14285 be lower than the advertised length if the body is larger than the buffer. It
14286 requires that the request body has been buffered made available using
14287 "option http-buffer-request".
14288
14289req.body_size : integer
14290 This returns the advertised length of the HTTP request's body in bytes. It
14291 will represent the advertised Content-Length header, or the size of the first
14292 chunk in case of chunked encoding. In order to parse the chunks, it requires
14293 that the request body has been buffered made available using
14294 "option http-buffer-request".
14295
Willy Tarreau74ca5042013-06-11 23:12:07 +020014296req.cook([<name>]) : string
14297cook([<name>]) : string (deprecated)
14298 This extracts the last occurrence of the cookie name <name> on a "Cookie"
14299 header line from the request, and returns its value as string. If no name is
14300 specified, the first cookie value is returned. When used with ACLs, all
14301 matching cookies are evaluated. Spaces around the name and the value are
14302 ignored as requested by the Cookie header specification (RFC6265). The cookie
14303 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
14304 well return an empty value if it is present. Use the "found" match to detect
14305 presence. Use the res.cook() variant for response cookies sent by the server.
14306
14307 ACL derivatives :
14308 cook([<name>]) : exact string match
14309 cook_beg([<name>]) : prefix match
14310 cook_dir([<name>]) : subdir match
14311 cook_dom([<name>]) : domain match
14312 cook_end([<name>]) : suffix match
14313 cook_len([<name>]) : length match
14314 cook_reg([<name>]) : regex match
14315 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010014316
Willy Tarreau74ca5042013-06-11 23:12:07 +020014317req.cook_cnt([<name>]) : integer
14318cook_cnt([<name>]) : integer (deprecated)
14319 Returns an integer value representing the number of occurrences of the cookie
14320 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014321
Willy Tarreau74ca5042013-06-11 23:12:07 +020014322req.cook_val([<name>]) : integer
14323cook_val([<name>]) : integer (deprecated)
14324 This extracts the last occurrence of the cookie name <name> on a "Cookie"
14325 header line from the request, and converts its value to an integer which is
14326 returned. If no name is specified, the first cookie value is returned. When
14327 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020014328
Willy Tarreau74ca5042013-06-11 23:12:07 +020014329cookie([<name>]) : string (deprecated)
14330 This extracts the last occurrence of the cookie name <name> on a "Cookie"
14331 header line from the request, or a "Set-Cookie" header from the response, and
14332 returns its value as a string. A typical use is to get multiple clients
14333 sharing a same profile use the same server. This can be similar to what
Willy Tarreau294d0f02015-08-10 19:40:12 +020014334 "appsession" did with the "request-learn" statement, but with support for
Willy Tarreau74ca5042013-06-11 23:12:07 +020014335 multi-peer synchronization and state keeping across restarts. If no name is
14336 specified, the first cookie value is returned. This fetch should not be used
14337 anymore and should be replaced by req.cook() or res.cook() instead as it
14338 ambiguously uses the direction based on the context where it is used.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014339
Willy Tarreau74ca5042013-06-11 23:12:07 +020014340hdr([<name>[,<occ>]]) : string
14341 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
14342 used on responses. Please refer to these respective fetches for more details.
14343 In case of doubt about the fetch direction, please use the explicit ones.
14344 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014345 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014346
Willy Tarreau74ca5042013-06-11 23:12:07 +020014347req.fhdr(<name>[,<occ>]) : string
14348 This extracts the last occurrence of header <name> in an HTTP request. When
14349 used from an ACL, all occurrences are iterated over until a match is found.
14350 Optionally, a specific occurrence might be specified as a position number.
14351 Positive values indicate a position from the first occurrence, with 1 being
14352 the first one. Negative values indicate positions relative to the last one,
14353 with -1 being the last one. It differs from req.hdr() in that any commas
14354 present in the value are returned and are not used as delimiters. This is
14355 sometimes useful with headers such as User-Agent.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014356
Willy Tarreau74ca5042013-06-11 23:12:07 +020014357req.fhdr_cnt([<name>]) : integer
14358 Returns an integer value representing the number of occurrences of request
14359 header field name <name>, or the total number of header fields if <name> is
14360 not specified. Contrary to its req.hdr_cnt() cousin, this function returns
14361 the number of full line headers and does not stop on commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014362
Willy Tarreau74ca5042013-06-11 23:12:07 +020014363req.hdr([<name>[,<occ>]]) : string
14364 This extracts the last occurrence of header <name> in an HTTP request. When
14365 used from an ACL, all occurrences are iterated over until a match is found.
14366 Optionally, a specific occurrence might be specified as a position number.
14367 Positive values indicate a position from the first occurrence, with 1 being
14368 the first one. Negative values indicate positions relative to the last one,
14369 with -1 being the last one. A typical use is with the X-Forwarded-For header
14370 once converted to IP, associated with an IP stick-table. The function
14371 considers any comma as a delimiter for distinct values. If full-line headers
14372 are desired instead, use req.fhdr(). Please carefully check RFC2616 to know
14373 how certain headers are supposed to be parsed. Also, some of them are case
14374 insensitive (eg: Connection).
Willy Tarreaud63335a2010-02-26 12:56:52 +010014375
Willy Tarreau74ca5042013-06-11 23:12:07 +020014376 ACL derivatives :
14377 hdr([<name>[,<occ>]]) : exact string match
14378 hdr_beg([<name>[,<occ>]]) : prefix match
14379 hdr_dir([<name>[,<occ>]]) : subdir match
14380 hdr_dom([<name>[,<occ>]]) : domain match
14381 hdr_end([<name>[,<occ>]]) : suffix match
14382 hdr_len([<name>[,<occ>]]) : length match
14383 hdr_reg([<name>[,<occ>]]) : regex match
14384 hdr_sub([<name>[,<occ>]]) : substring match
14385
14386req.hdr_cnt([<name>]) : integer
14387hdr_cnt([<header>]) : integer (deprecated)
14388 Returns an integer value representing the number of occurrences of request
14389 header field name <name>, or the total number of header field values if
14390 <name> is not specified. It is important to remember that one header line may
14391 count as several headers if it has several values. The function considers any
14392 comma as a delimiter for distinct values. If full-line headers are desired
14393 instead, req.fhdr_cnt() should be used instead. With ACLs, it can be used to
14394 detect presence, absence or abuse of a specific header, as well as to block
14395 request smuggling attacks by rejecting requests which contain more than one
14396 of certain headers. See "req.hdr" for more information on header matching.
14397
14398req.hdr_ip([<name>[,<occ>]]) : ip
14399hdr_ip([<name>[,<occ>]]) : ip (deprecated)
14400 This extracts the last occurrence of header <name> in an HTTP request,
14401 converts it to an IPv4 or IPv6 address and returns this address. When used
14402 with ACLs, all occurrences are checked, and if <name> is omitted, every value
14403 of every header is checked. Optionally, a specific occurrence might be
14404 specified as a position number. Positive values indicate a position from the
14405 first occurrence, with 1 being the first one. Negative values indicate
14406 positions relative to the last one, with -1 being the last one. A typical use
14407 is with the X-Forwarded-For and X-Client-IP headers.
14408
14409req.hdr_val([<name>[,<occ>]]) : integer
14410hdr_val([<name>[,<occ>]]) : integer (deprecated)
14411 This extracts the last occurrence of header <name> in an HTTP request, and
14412 converts it to an integer value. When used with ACLs, all occurrences are
14413 checked, and if <name> is omitted, every value of every header is checked.
14414 Optionally, a specific occurrence might be specified as a position number.
14415 Positive values indicate a position from the first occurrence, with 1 being
14416 the first one. Negative values indicate positions relative to the last one,
14417 with -1 being the last one. A typical use is with the X-Forwarded-For header.
14418
14419http_auth(<userlist>) : boolean
14420 Returns a boolean indicating whether the authentication data received from
14421 the client match a username & password stored in the specified userlist. This
14422 fetch function is not really useful outside of ACLs. Currently only http
14423 basic auth is supported.
14424
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010014425http_auth_group(<userlist>) : string
14426 Returns a string corresponding to the user name found in the authentication
14427 data received from the client if both the user name and password are valid
14428 according to the specified userlist. The main purpose is to use it in ACLs
14429 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014430 This fetch function is not really useful outside of ACLs. Currently only http
14431 basic auth is supported.
14432
14433 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010014434 http_auth_group(<userlist>) : group ...
14435 Returns true when the user extracted from the request and whose password is
14436 valid according to the specified userlist belongs to at least one of the
14437 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014438
14439http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020014440 Returns true when the request being processed is the first one of the
14441 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020014442 from some requests when a request is not the first one, or to help grouping
14443 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020014444
Willy Tarreau74ca5042013-06-11 23:12:07 +020014445method : integer + string
14446 Returns an integer value corresponding to the method in the HTTP request. For
14447 example, "GET" equals 1 (check sources to establish the matching). Value 9
14448 means "other method" and may be converted to a string extracted from the
14449 stream. This should not be used directly as a sample, this is only meant to
14450 be used from ACLs, which transparently convert methods from patterns to these
14451 integer + string values. Some predefined ACL already check for most common
14452 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014453
Willy Tarreau74ca5042013-06-11 23:12:07 +020014454 ACL derivatives :
14455 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020014456
Willy Tarreau74ca5042013-06-11 23:12:07 +020014457 Example :
14458 # only accept GET and HEAD requests
14459 acl valid_method method GET HEAD
14460 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020014461
Willy Tarreau74ca5042013-06-11 23:12:07 +020014462path : string
14463 This extracts the request's URL path, which starts at the first slash and
14464 ends before the question mark (without the host part). A typical use is with
14465 prefetch-capable caches, and with portals which need to aggregate multiple
14466 information from databases and keep them in caches. Note that with outgoing
14467 caches, it would be wiser to use "url" instead. With ACLs, it's typically
14468 used to match exact file names (eg: "/login.php"), or directory parts using
14469 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014470
Willy Tarreau74ca5042013-06-11 23:12:07 +020014471 ACL derivatives :
14472 path : exact string match
14473 path_beg : prefix match
14474 path_dir : subdir match
14475 path_dom : domain match
14476 path_end : suffix match
14477 path_len : length match
14478 path_reg : regex match
14479 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020014480
Willy Tarreau49ad95c2015-01-19 15:06:26 +010014481query : string
14482 This extracts the request's query string, which starts after the first
14483 question mark. If no question mark is present, this fetch returns nothing. If
14484 a question mark is present but nothing follows, it returns an empty string.
14485 This means it's possible to easily know whether a query string is present
Tim Düsterhus4896c442016-11-29 02:15:19 +010014486 using the "found" matching method. This fetch is the complement of "path"
Willy Tarreau49ad95c2015-01-19 15:06:26 +010014487 which stops before the question mark.
14488
Willy Tarreaueb27ec72015-02-20 13:55:29 +010014489req.hdr_names([<delim>]) : string
14490 This builds a string made from the concatenation of all header names as they
14491 appear in the request when the rule is evaluated. The default delimiter is
14492 the comma (',') but it may be overridden as an optional argument <delim>. In
14493 this case, only the first character of <delim> is considered.
14494
Willy Tarreau74ca5042013-06-11 23:12:07 +020014495req.ver : string
14496req_ver : string (deprecated)
14497 Returns the version string from the HTTP request, for example "1.1". This can
14498 be useful for logs, but is mostly there for ACL. Some predefined ACL already
14499 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014500
Willy Tarreau74ca5042013-06-11 23:12:07 +020014501 ACL derivatives :
14502 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020014503
Willy Tarreau74ca5042013-06-11 23:12:07 +020014504res.comp : boolean
14505 Returns the boolean "true" value if the response has been compressed by
14506 HAProxy, otherwise returns boolean "false". This may be used to add
14507 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014508
Willy Tarreau74ca5042013-06-11 23:12:07 +020014509res.comp_algo : string
14510 Returns a string containing the name of the algorithm used if the response
14511 was compressed by HAProxy, for example : "deflate". This may be used to add
14512 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014513
Willy Tarreau74ca5042013-06-11 23:12:07 +020014514res.cook([<name>]) : string
14515scook([<name>]) : string (deprecated)
14516 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
14517 header line from the response, and returns its value as string. If no name is
14518 specified, the first cookie value is returned.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020014519
Willy Tarreau74ca5042013-06-11 23:12:07 +020014520 ACL derivatives :
14521 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020014522
Willy Tarreau74ca5042013-06-11 23:12:07 +020014523res.cook_cnt([<name>]) : integer
14524scook_cnt([<name>]) : integer (deprecated)
14525 Returns an integer value representing the number of occurrences of the cookie
14526 <name> in the response, or all cookies if <name> is not specified. This is
14527 mostly useful when combined with ACLs to detect suspicious responses.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014528
Willy Tarreau74ca5042013-06-11 23:12:07 +020014529res.cook_val([<name>]) : integer
14530scook_val([<name>]) : integer (deprecated)
14531 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
14532 header line from the response, and converts its value to an integer which is
14533 returned. If no name is specified, the first cookie value is returned.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014534
Willy Tarreau74ca5042013-06-11 23:12:07 +020014535res.fhdr([<name>[,<occ>]]) : string
14536 This extracts the last occurrence of header <name> in an HTTP response, or of
14537 the last header if no <name> is specified. Optionally, a specific occurrence
14538 might be specified as a position number. Positive values indicate a position
14539 from the first occurrence, with 1 being the first one. Negative values
14540 indicate positions relative to the last one, with -1 being the last one. It
14541 differs from res.hdr() in that any commas present in the value are returned
14542 and are not used as delimiters. If this is not desired, the res.hdr() fetch
14543 should be used instead. This is sometimes useful with headers such as Date or
14544 Expires.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014545
Willy Tarreau74ca5042013-06-11 23:12:07 +020014546res.fhdr_cnt([<name>]) : integer
14547 Returns an integer value representing the number of occurrences of response
14548 header field name <name>, or the total number of header fields if <name> is
14549 not specified. Contrary to its res.hdr_cnt() cousin, this function returns
14550 the number of full line headers and does not stop on commas. If this is not
14551 desired, the res.hdr_cnt() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014552
Willy Tarreau74ca5042013-06-11 23:12:07 +020014553res.hdr([<name>[,<occ>]]) : string
14554shdr([<name>[,<occ>]]) : string (deprecated)
14555 This extracts the last occurrence of header <name> in an HTTP response, or of
14556 the last header if no <name> is specified. Optionally, a specific occurrence
14557 might be specified as a position number. Positive values indicate a position
14558 from the first occurrence, with 1 being the first one. Negative values
14559 indicate positions relative to the last one, with -1 being the last one. This
14560 can be useful to learn some data into a stick-table. The function considers
14561 any comma as a delimiter for distinct values. If this is not desired, the
14562 res.fhdr() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014563
Willy Tarreau74ca5042013-06-11 23:12:07 +020014564 ACL derivatives :
14565 shdr([<name>[,<occ>]]) : exact string match
14566 shdr_beg([<name>[,<occ>]]) : prefix match
14567 shdr_dir([<name>[,<occ>]]) : subdir match
14568 shdr_dom([<name>[,<occ>]]) : domain match
14569 shdr_end([<name>[,<occ>]]) : suffix match
14570 shdr_len([<name>[,<occ>]]) : length match
14571 shdr_reg([<name>[,<occ>]]) : regex match
14572 shdr_sub([<name>[,<occ>]]) : substring match
14573
14574res.hdr_cnt([<name>]) : integer
14575shdr_cnt([<name>]) : integer (deprecated)
14576 Returns an integer value representing the number of occurrences of response
14577 header field name <name>, or the total number of header fields if <name> is
14578 not specified. The function considers any comma as a delimiter for distinct
14579 values. If this is not desired, the res.fhdr_cnt() fetch should be used
14580 instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014581
Willy Tarreau74ca5042013-06-11 23:12:07 +020014582res.hdr_ip([<name>[,<occ>]]) : ip
14583shdr_ip([<name>[,<occ>]]) : ip (deprecated)
14584 This extracts the last occurrence of header <name> in an HTTP response,
14585 convert it to an IPv4 or IPv6 address and returns this address. Optionally, a
14586 specific occurrence might be specified as a position number. Positive values
14587 indicate a position from the first occurrence, with 1 being the first one.
14588 Negative values indicate positions relative to the last one, with -1 being
14589 the last one. This can be useful to learn some data into a stick table.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014590
Willy Tarreaueb27ec72015-02-20 13:55:29 +010014591res.hdr_names([<delim>]) : string
14592 This builds a string made from the concatenation of all header names as they
14593 appear in the response when the rule is evaluated. The default delimiter is
14594 the comma (',') but it may be overridden as an optional argument <delim>. In
14595 this case, only the first character of <delim> is considered.
14596
Willy Tarreau74ca5042013-06-11 23:12:07 +020014597res.hdr_val([<name>[,<occ>]]) : integer
14598shdr_val([<name>[,<occ>]]) : integer (deprecated)
14599 This extracts the last occurrence of header <name> in an HTTP response, and
14600 converts it to an integer value. Optionally, a specific occurrence might be
14601 specified as a position number. Positive values indicate a position from the
14602 first occurrence, with 1 being the first one. Negative values indicate
14603 positions relative to the last one, with -1 being the last one. This can be
14604 useful to learn some data into a stick table.
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010014605
Willy Tarreau74ca5042013-06-11 23:12:07 +020014606res.ver : string
14607resp_ver : string (deprecated)
14608 Returns the version string from the HTTP response, for example "1.1". This
14609 can be useful for logs, but is mostly there for ACL.
Willy Tarreau0e698542011-09-16 08:32:32 +020014610
Willy Tarreau74ca5042013-06-11 23:12:07 +020014611 ACL derivatives :
14612 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010014613
Willy Tarreau74ca5042013-06-11 23:12:07 +020014614set-cookie([<name>]) : string (deprecated)
14615 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
14616 header line from the response and uses the corresponding value to match. This
Willy Tarreau294d0f02015-08-10 19:40:12 +020014617 can be comparable to what "appsession" did with default options, but with
Willy Tarreau74ca5042013-06-11 23:12:07 +020014618 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010014619
Willy Tarreau74ca5042013-06-11 23:12:07 +020014620 This fetch function is deprecated and has been superseded by the "res.cook"
14621 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010014622
Willy Tarreau74ca5042013-06-11 23:12:07 +020014623status : integer
14624 Returns an integer containing the HTTP status code in the HTTP response, for
14625 example, 302. It is mostly used within ACLs and integer ranges, for example,
14626 to remove any Location header if the response is not a 3xx.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014627
Thierry Fournier0e00dca2016-04-07 15:47:40 +020014628unique-id : string
14629 Returns the unique-id attached to the request. The directive
14630 "unique-id-format" must be set. If it is not set, the unique-id sample fetch
14631 fails. Note that the unique-id is usually used with HTTP requests, however this
14632 sample fetch can be used with other protocols. Obviously, if it is used with
14633 other protocols than HTTP, the unique-id-format directive must not contain
14634 HTTP parts. See: unique-id-format and unique-id-header
14635
Willy Tarreau74ca5042013-06-11 23:12:07 +020014636url : string
14637 This extracts the request's URL as presented in the request. A typical use is
14638 with prefetch-capable caches, and with portals which need to aggregate
14639 multiple information from databases and keep them in caches. With ACLs, using
14640 "path" is preferred over using "url", because clients may send a full URL as
14641 is normally done with proxies. The only real use is to match "*" which does
14642 not match in "path", and for which there is already a predefined ACL. See
14643 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014644
Willy Tarreau74ca5042013-06-11 23:12:07 +020014645 ACL derivatives :
14646 url : exact string match
14647 url_beg : prefix match
14648 url_dir : subdir match
14649 url_dom : domain match
14650 url_end : suffix match
14651 url_len : length match
14652 url_reg : regex match
14653 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014654
Willy Tarreau74ca5042013-06-11 23:12:07 +020014655url_ip : ip
14656 This extracts the IP address from the request's URL when the host part is
14657 presented as an IP address. Its use is very limited. For instance, a
14658 monitoring system might use this field as an alternative for the source IP in
14659 order to test what path a given source address would follow, or to force an
14660 entry in a table for a given source address. With ACLs it can be used to
14661 restrict access to certain systems through a proxy, for example when combined
14662 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014663
Willy Tarreau74ca5042013-06-11 23:12:07 +020014664url_port : integer
14665 This extracts the port part from the request's URL. Note that if the port is
14666 not specified in the request, port 80 is assumed. With ACLs it can be used to
14667 restrict access to certain systems through a proxy, for example when combined
14668 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014669
Willy Tarreau1ede1da2015-05-07 16:06:18 +020014670urlp([<name>[,<delim>]]) : string
14671url_param([<name>[,<delim>]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020014672 This extracts the first occurrence of the parameter <name> in the query
14673 string, which begins after either '?' or <delim>, and which ends before '&',
Willy Tarreau1ede1da2015-05-07 16:06:18 +020014674 ';' or <delim>. The parameter name is case-sensitive. If no name is given,
14675 any parameter will match, and the first one will be returned. The result is
14676 a string corresponding to the value of the parameter <name> as presented in
14677 the request (no URL decoding is performed). This can be used for session
Willy Tarreau74ca5042013-06-11 23:12:07 +020014678 stickiness based on a client ID, to extract an application cookie passed as a
14679 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
Willy Tarreau1ede1da2015-05-07 16:06:18 +020014680 this fetch iterates over multiple parameters and will iteratively report all
14681 parameters values if no name is given
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014682
Willy Tarreau74ca5042013-06-11 23:12:07 +020014683 ACL derivatives :
14684 urlp(<name>[,<delim>]) : exact string match
14685 urlp_beg(<name>[,<delim>]) : prefix match
14686 urlp_dir(<name>[,<delim>]) : subdir match
14687 urlp_dom(<name>[,<delim>]) : domain match
14688 urlp_end(<name>[,<delim>]) : suffix match
14689 urlp_len(<name>[,<delim>]) : length match
14690 urlp_reg(<name>[,<delim>]) : regex match
14691 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014692
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014693
Willy Tarreau74ca5042013-06-11 23:12:07 +020014694 Example :
14695 # match http://example.com/foo?PHPSESSIONID=some_id
14696 stick on urlp(PHPSESSIONID)
14697 # match http://example.com/foo;JSESSIONID=some_id
14698 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020014699
Willy Tarreau1ede1da2015-05-07 16:06:18 +020014700urlp_val([<name>[,<delim>])] : integer
Willy Tarreau74ca5042013-06-11 23:12:07 +020014701 See "urlp" above. This one extracts the URL parameter <name> in the request
14702 and converts it to an integer value. This can be used for session stickiness
14703 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020014704
Dragan Dosen0070cd52016-06-16 12:19:49 +020014705url32 : integer
14706 This returns a 32-bit hash of the value obtained by concatenating the first
14707 Host header and the whole URL including parameters (not only the path part of
14708 the request, as in the "base32" fetch above). This is useful to track per-URL
14709 activity. A shorter hash is stored, saving a lot of memory. The output type
14710 is an unsigned integer.
14711
14712url32+src : binary
14713 This returns the concatenation of the "url32" fetch and the "src" fetch. The
14714 resulting type is of type binary, with a size of 8 or 20 bytes depending on
14715 the source address family. This can be used to track per-IP, per-URL counters.
14716
Willy Tarreau198a7442008-01-17 12:05:32 +010014717
Willy Tarreau74ca5042013-06-11 23:12:07 +0200147187.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014719---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010014720
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014721Some predefined ACLs are hard-coded so that they do not have to be declared in
14722every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020014723order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010014724
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014725ACL name Equivalent to Usage
14726---------------+-----------------------------+---------------------------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014727FALSE always_false never match
Willy Tarreau2492d5b2009-07-11 00:06:00 +020014728HTTP req_proto_http match if protocol is valid HTTP
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014729HTTP_1.0 req_ver 1.0 match HTTP version 1.0
14730HTTP_1.1 req_ver 1.1 match HTTP version 1.1
Willy Tarreaud63335a2010-02-26 12:56:52 +010014731HTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length
14732HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
14733HTTP_URL_SLASH url_beg / match URL beginning with "/"
14734HTTP_URL_STAR url * match URL equal to "*"
14735LOCALHOST src 127.0.0.1/8 match connection from local host
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014736METH_CONNECT method CONNECT match HTTP CONNECT method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020014737METH_DELETE method DELETE match HTTP DELETE method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014738METH_GET method GET HEAD match HTTP GET or HEAD method
14739METH_HEAD method HEAD match HTTP HEAD method
14740METH_OPTIONS method OPTIONS match HTTP OPTIONS method
14741METH_POST method POST match HTTP POST method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020014742METH_PUT method PUT match HTTP PUT method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014743METH_TRACE method TRACE match HTTP TRACE method
Emeric Brunbede3d02009-06-30 17:54:00 +020014744RDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014745REQ_CONTENT req_len gt 0 match data in the request buffer
Willy Tarreaud63335a2010-02-26 12:56:52 +010014746TRUE always_true always match
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014747WAIT_END wait_end wait for end of content analysis
14748---------------+-----------------------------+---------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010014749
Willy Tarreaub937b7e2010-01-12 15:27:54 +010014750
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200147518. Logging
14752----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010014753
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014754One of HAProxy's strong points certainly lies is its precise logs. It probably
14755provides the finest level of information available for such a product, which is
14756very important for troubleshooting complex environments. Standard information
14757provided in logs include client ports, TCP/HTTP state timers, precise session
14758state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010014759to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014760headers.
14761
14762In order to improve administrators reactivity, it offers a great transparency
14763about encountered problems, both internal and external, and it is possible to
14764send logs to different sources at the same time with different level filters :
14765
14766 - global process-level logs (system errors, start/stop, etc..)
14767 - per-instance system and internal errors (lack of resource, bugs, ...)
14768 - per-instance external troubles (servers up/down, max connections)
14769 - per-instance activity (client connections), either at the establishment or
14770 at the termination.
Jim Freeman9e8714b2015-05-26 09:16:34 -060014771 - per-request control of log-level, eg:
14772 http-request set-log-level silent if sensitive_request
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014773
14774The ability to distribute different levels of logs to different log servers
14775allow several production teams to interact and to fix their problems as soon
14776as possible. For example, the system team might monitor system-wide errors,
14777while the application team might be monitoring the up/down for their servers in
14778real time, and the security team might analyze the activity logs with one hour
14779delay.
14780
14781
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200147828.1. Log levels
14783---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014784
Simon Hormandf791f52011-05-29 15:01:10 +090014785TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014786source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090014787HTTP request, HTTP return code, number of bytes transmitted, conditions
14788in which the session ended, and even exchanged cookies values. For example
14789track a particular user's problems. All messages may be sent to up to two
14790syslog servers. Check the "log" keyword in section 4.2 for more information
14791about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014792
14793
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200147948.2. Log formats
14795----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014796
William Lallemand48940402012-01-30 16:47:22 +010014797HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090014798and will be detailed in the following sections. A few of them may vary
14799slightly with the configuration, due to indicators specific to certain
14800options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014801
14802 - the default format, which is very basic and very rarely used. It only
14803 provides very basic information about the incoming connection at the moment
14804 it is accepted : source IP:port, destination IP:port, and frontend-name.
14805 This mode will eventually disappear so it will not be described to great
14806 extents.
14807
14808 - the TCP format, which is more advanced. This format is enabled when "option
14809 tcplog" is set on the frontend. HAProxy will then usually wait for the
14810 connection to terminate before logging. This format provides much richer
14811 information, such as timers, connection counts, queue size, etc... This
14812 format is recommended for pure TCP proxies.
14813
14814 - the HTTP format, which is the most advanced for HTTP proxying. This format
14815 is enabled when "option httplog" is set on the frontend. It provides the
14816 same information as the TCP format with some HTTP-specific fields such as
14817 the request, the status code, and captures of headers and cookies. This
14818 format is recommended for HTTP proxies.
14819
Emeric Brun3a058f32009-06-30 18:26:00 +020014820 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
14821 fields arranged in the same order as the CLF format. In this mode, all
14822 timers, captures, flags, etc... appear one per field after the end of the
14823 common fields, in the same order they appear in the standard HTTP format.
14824
William Lallemand48940402012-01-30 16:47:22 +010014825 - the custom log format, allows you to make your own log line.
14826
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014827Next sections will go deeper into details for each of these formats. Format
14828specification will be performed on a "field" basis. Unless stated otherwise, a
14829field is a portion of text delimited by any number of spaces. Since syslog
14830servers are susceptible of inserting fields at the beginning of a line, it is
14831always assumed that the first field is the one containing the process name and
14832identifier.
14833
14834Note : Since log lines may be quite long, the log examples in sections below
14835 might be broken into multiple lines. The example log lines will be
14836 prefixed with 3 closing angle brackets ('>>>') and each time a log is
14837 broken into multiple lines, each non-final line will end with a
14838 backslash ('\') and the next line will start indented by two characters.
14839
14840
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200148418.2.1. Default log format
14842-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014843
14844This format is used when no specific option is set. The log is emitted as soon
14845as the connection is accepted. One should note that this currently is the only
14846format which logs the request's destination IP and ports.
14847
14848 Example :
14849 listen www
14850 mode http
14851 log global
14852 server srv1 127.0.0.1:8000
14853
14854 >>> Feb 6 12:12:09 localhost \
14855 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
14856 (www/HTTP)
14857
14858 Field Format Extract from the example above
14859 1 process_name '[' pid ']:' haproxy[14385]:
14860 2 'Connect from' Connect from
14861 3 source_ip ':' source_port 10.0.1.2:33312
14862 4 'to' to
14863 5 destination_ip ':' destination_port 10.0.3.31:8012
14864 6 '(' frontend_name '/' mode ')' (www/HTTP)
14865
14866Detailed fields description :
14867 - "source_ip" is the IP address of the client which initiated the connection.
14868 - "source_port" is the TCP port of the client which initiated the connection.
14869 - "destination_ip" is the IP address the client connected to.
14870 - "destination_port" is the TCP port the client connected to.
14871 - "frontend_name" is the name of the frontend (or listener) which received
14872 and processed the connection.
14873 - "mode is the mode the frontend is operating (TCP or HTTP).
14874
Willy Tarreauceb24bc2010-11-09 12:46:41 +010014875In case of a UNIX socket, the source and destination addresses are marked as
14876"unix:" and the ports reflect the internal ID of the socket which accepted the
14877connection (the same ID as reported in the stats).
14878
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014879It is advised not to use this deprecated format for newer installations as it
14880will eventually disappear.
14881
14882
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200148838.2.2. TCP log format
14884---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014885
14886The TCP format is used when "option tcplog" is specified in the frontend, and
14887is the recommended format for pure TCP proxies. It provides a lot of precious
14888information for troubleshooting. Since this format includes timers and byte
14889counts, the log is normally emitted at the end of the session. It can be
14890emitted earlier if "option logasap" is specified, which makes sense in most
14891environments with long sessions such as remote terminals. Sessions which match
14892the "monitor" rules are never logged. It is also possible not to emit logs for
14893sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020014894specifying "option dontlognull" in the frontend. Successful connections will
14895not be logged if "option dontlog-normal" is specified in the frontend. A few
14896fields may slightly vary depending on some configuration options, those are
14897marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014898
14899 Example :
14900 frontend fnt
14901 mode tcp
14902 option tcplog
14903 log global
14904 default_backend bck
14905
14906 backend bck
14907 server srv1 127.0.0.1:8000
14908
14909 >>> Feb 6 12:12:56 localhost \
14910 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
14911 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
14912
14913 Field Format Extract from the example above
14914 1 process_name '[' pid ']:' haproxy[14387]:
14915 2 client_ip ':' client_port 10.0.1.2:33313
14916 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
14917 4 frontend_name fnt
14918 5 backend_name '/' server_name bck/srv1
14919 6 Tw '/' Tc '/' Tt* 0/0/5007
14920 7 bytes_read* 212
14921 8 termination_state --
14922 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
14923 10 srv_queue '/' backend_queue 0/0
14924
14925Detailed fields description :
14926 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010014927 connection to haproxy. If the connection was accepted on a UNIX socket
14928 instead, the IP address would be replaced with the word "unix". Note that
14929 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010014930 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
14931 and the NetScaler Client IP insetion protocol is correctly used, then the
14932 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014933
14934 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010014935 If the connection was accepted on a UNIX socket instead, the port would be
14936 replaced with the ID of the accepting socket, which is also reported in the
14937 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014938
14939 - "accept_date" is the exact date when the connection was received by haproxy
14940 (which might be very slightly different from the date observed on the
14941 network if there was some queuing in the system's backlog). This is usually
14942 the same date which may appear in any upstream firewall's log.
14943
14944 - "frontend_name" is the name of the frontend (or listener) which received
14945 and processed the connection.
14946
14947 - "backend_name" is the name of the backend (or listener) which was selected
14948 to manage the connection to the server. This will be the same as the
14949 frontend if no switching rule has been applied, which is common for TCP
14950 applications.
14951
14952 - "server_name" is the name of the last server to which the connection was
14953 sent, which might differ from the first one if there were connection errors
14954 and a redispatch occurred. Note that this server belongs to the backend
14955 which processed the request. If the connection was aborted before reaching
14956 a server, "<NOSRV>" is indicated instead of a server name.
14957
14958 - "Tw" is the total time in milliseconds spent waiting in the various queues.
14959 It can be "-1" if the connection was aborted before reaching the queue.
14960 See "Timers" below for more details.
14961
14962 - "Tc" is the total time in milliseconds spent waiting for the connection to
14963 establish to the final server, including retries. It can be "-1" if the
14964 connection was aborted before a connection could be established. See
14965 "Timers" below for more details.
14966
14967 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014968 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014969 "option logasap" was specified, then the time counting stops at the moment
14970 the log is emitted. In this case, a '+' sign is prepended before the value,
14971 indicating that the final one will be larger. See "Timers" below for more
14972 details.
14973
14974 - "bytes_read" is the total number of bytes transmitted from the server to
14975 the client when the log is emitted. If "option logasap" is specified, the
14976 this value will be prefixed with a '+' sign indicating that the final one
14977 may be larger. Please note that this value is a 64-bit counter, so log
14978 analysis tools must be able to handle it without overflowing.
14979
14980 - "termination_state" is the condition the session was in when the session
14981 ended. This indicates the session state, which side caused the end of
14982 session to happen, and for what reason (timeout, error, ...). The normal
14983 flags should be "--", indicating the session was closed by either end with
14984 no data remaining in buffers. See below "Session state at disconnection"
14985 for more details.
14986
14987 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040014988 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014989 limits have been reached. For instance, if actconn is close to 512 when
14990 multiple connection errors occur, chances are high that the system limits
14991 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014992 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010014993
14994 - "feconn" is the total number of concurrent connections on the frontend when
14995 the session was logged. It is useful to estimate the amount of resource
14996 required to sustain high loads, and to detect when the frontend's "maxconn"
14997 has been reached. Most often when this value increases by huge jumps, it is
14998 because there is congestion on the backend servers, but sometimes it can be
14999 caused by a denial of service attack.
15000
15001 - "beconn" is the total number of concurrent connections handled by the
15002 backend when the session was logged. It includes the total number of
15003 concurrent connections active on servers as well as the number of
15004 connections pending in queues. It is useful to estimate the amount of
15005 additional servers needed to support high loads for a given application.
15006 Most often when this value increases by huge jumps, it is because there is
15007 congestion on the backend servers, but sometimes it can be caused by a
15008 denial of service attack.
15009
15010 - "srv_conn" is the total number of concurrent connections still active on
15011 the server when the session was logged. It can never exceed the server's
15012 configured "maxconn" parameter. If this value is very often close or equal
15013 to the server's "maxconn", it means that traffic regulation is involved a
15014 lot, meaning that either the server's maxconn value is too low, or that
15015 there aren't enough servers to process the load with an optimal response
15016 time. When only one of the server's "srv_conn" is high, it usually means
15017 that this server has some trouble causing the connections to take longer to
15018 be processed than on other servers.
15019
15020 - "retries" is the number of connection retries experienced by this session
15021 when trying to connect to the server. It must normally be zero, unless a
15022 server is being stopped at the same moment the connection was attempted.
15023 Frequent retries generally indicate either a network problem between
15024 haproxy and the server, or a misconfigured system backlog on the server
15025 preventing new connections from being queued. This field may optionally be
15026 prefixed with a '+' sign, indicating that the session has experienced a
15027 redispatch after the maximal retry count has been reached on the initial
15028 server. In this case, the server name appearing in the log is the one the
15029 connection was redispatched to, and not the first one, though both may
15030 sometimes be the same in case of hashing for instance. So as a general rule
15031 of thumb, when a '+' is present in front of the retry count, this count
15032 should not be attributed to the logged server.
15033
15034 - "srv_queue" is the total number of requests which were processed before
15035 this one in the server queue. It is zero when the request has not gone
15036 through the server queue. It makes it possible to estimate the approximate
15037 server's response time by dividing the time spent in queue by the number of
15038 requests in the queue. It is worth noting that if a session experiences a
15039 redispatch and passes through two server queues, their positions will be
15040 cumulated. A request should not pass through both the server queue and the
15041 backend queue unless a redispatch occurs.
15042
15043 - "backend_queue" is the total number of requests which were processed before
15044 this one in the backend's global queue. It is zero when the request has not
15045 gone through the global queue. It makes it possible to estimate the average
15046 queue length, which easily translates into a number of missing servers when
15047 divided by a server's "maxconn" parameter. It is worth noting that if a
15048 session experiences a redispatch, it may pass twice in the backend's queue,
15049 and then both positions will be cumulated. A request should not pass
15050 through both the server queue and the backend queue unless a redispatch
15051 occurs.
15052
15053
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200150548.2.3. HTTP log format
15055----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015056
15057The HTTP format is the most complete and the best suited for HTTP proxies. It
15058is enabled by when "option httplog" is specified in the frontend. It provides
15059the same level of information as the TCP format with additional features which
15060are specific to the HTTP protocol. Just like the TCP format, the log is usually
15061emitted at the end of the session, unless "option logasap" is specified, which
15062generally only makes sense for download sites. A session which matches the
15063"monitor" rules will never logged. It is also possible not to log sessions for
15064which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020015065frontend. Successful connections will not be logged if "option dontlog-normal"
15066is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015067
15068Most fields are shared with the TCP log, some being different. A few fields may
15069slightly vary depending on some configuration options. Those ones are marked
15070with a star ('*') after the field name below.
15071
15072 Example :
15073 frontend http-in
15074 mode http
15075 option httplog
15076 log global
15077 default_backend bck
15078
15079 backend static
15080 server srv1 127.0.0.1:8000
15081
15082 >>> Feb 6 12:14:14 localhost \
15083 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
15084 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 +010015085 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015086
15087 Field Format Extract from the example above
15088 1 process_name '[' pid ']:' haproxy[14389]:
15089 2 client_ip ':' client_port 10.0.1.2:33317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015090 3 '[' request_date ']' [06/Feb/2009:12:14:14.655]
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015091 4 frontend_name http-in
15092 5 backend_name '/' server_name static/srv1
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015093 6 TR '/' Tw '/' Tc '/' Tr '/' Ta* 10/0/30/69/109
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015094 7 status_code 200
15095 8 bytes_read* 2750
15096 9 captured_request_cookie -
15097 10 captured_response_cookie -
15098 11 termination_state ----
15099 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
15100 13 srv_queue '/' backend_queue 0/0
15101 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
15102 15 '{' captured_response_headers* '}' {}
15103 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010015104
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015105Detailed fields description :
15106 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010015107 connection to haproxy. If the connection was accepted on a UNIX socket
15108 instead, the IP address would be replaced with the word "unix". Note that
15109 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010015110 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
15111 and the NetScaler Client IP insetion protocol is correctly used, then the
15112 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015113
15114 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010015115 If the connection was accepted on a UNIX socket instead, the port would be
15116 replaced with the ID of the accepting socket, which is also reported in the
15117 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015118
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015119 - "request_date" is the exact date when the first byte of the HTTP request
15120 was received by haproxy (log field %tr).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015121
15122 - "frontend_name" is the name of the frontend (or listener) which received
15123 and processed the connection.
15124
15125 - "backend_name" is the name of the backend (or listener) which was selected
15126 to manage the connection to the server. This will be the same as the
15127 frontend if no switching rule has been applied.
15128
15129 - "server_name" is the name of the last server to which the connection was
15130 sent, which might differ from the first one if there were connection errors
15131 and a redispatch occurred. Note that this server belongs to the backend
15132 which processed the request. If the request was aborted before reaching a
15133 server, "<NOSRV>" is indicated instead of a server name. If the request was
15134 intercepted by the stats subsystem, "<STATS>" is indicated instead.
15135
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015136 - "TR" is the total time in milliseconds spent waiting for a full HTTP
15137 request from the client (not counting body) after the first byte was
15138 received. It can be "-1" if the connection was aborted before a complete
15139 request could be received or the a bad request was received. It should
15140 always be very small because a request generally fits in one single packet.
15141 Large times here generally indicate network issues between the client and
15142 haproxy or requests being typed by hand. See "Timers" below for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015143
15144 - "Tw" is the total time in milliseconds spent waiting in the various queues.
15145 It can be "-1" if the connection was aborted before reaching the queue.
15146 See "Timers" below for more details.
15147
15148 - "Tc" is the total time in milliseconds spent waiting for the connection to
15149 establish to the final server, including retries. It can be "-1" if the
15150 request was aborted before a connection could be established. See "Timers"
15151 below for more details.
15152
15153 - "Tr" is the total time in milliseconds spent waiting for the server to send
15154 a full HTTP response, not counting data. It can be "-1" if the request was
15155 aborted before a complete response could be received. It generally matches
15156 the server's processing time for the request, though it may be altered by
15157 the amount of data sent by the client to the server. Large times here on
15158 "GET" requests generally indicate an overloaded server. See "Timers" below
15159 for more details.
15160
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015161 - "Ta" is the time the request remained active in haproxy, which is the total
15162 time in milliseconds elapsed between the first byte of the request was
15163 received and the last byte of response was sent. It covers all possible
15164 processing except the handshake (see Th) and idle time (see Ti). There is
15165 one exception, if "option logasap" was specified, then the time counting
15166 stops at the moment the log is emitted. In this case, a '+' sign is
15167 prepended before the value, indicating that the final one will be larger.
15168 See "Timers" below for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015169
15170 - "status_code" is the HTTP status code returned to the client. This status
15171 is generally set by the server, but it might also be set by haproxy when
15172 the server cannot be reached or when its response is blocked by haproxy.
15173
15174 - "bytes_read" is the total number of bytes transmitted to the client when
15175 the log is emitted. This does include HTTP headers. If "option logasap" is
15176 specified, the this value will be prefixed with a '+' sign indicating that
15177 the final one may be larger. Please note that this value is a 64-bit
15178 counter, so log analysis tools must be able to handle it without
15179 overflowing.
15180
15181 - "captured_request_cookie" is an optional "name=value" entry indicating that
15182 the client had this cookie in the request. The cookie name and its maximum
15183 length are defined by the "capture cookie" statement in the frontend
15184 configuration. The field is a single dash ('-') when the option is not
15185 set. Only one cookie may be captured, it is generally used to track session
15186 ID exchanges between a client and a server to detect session crossing
15187 between clients due to application bugs. For more details, please consult
15188 the section "Capturing HTTP headers and cookies" below.
15189
15190 - "captured_response_cookie" is an optional "name=value" entry indicating
15191 that the server has returned a cookie with its response. The cookie name
15192 and its maximum length are defined by the "capture cookie" statement in the
15193 frontend configuration. The field is a single dash ('-') when the option is
15194 not set. Only one cookie may be captured, it is generally used to track
15195 session ID exchanges between a client and a server to detect session
15196 crossing between clients due to application bugs. For more details, please
15197 consult the section "Capturing HTTP headers and cookies" below.
15198
15199 - "termination_state" is the condition the session was in when the session
15200 ended. This indicates the session state, which side caused the end of
15201 session to happen, for what reason (timeout, error, ...), just like in TCP
15202 logs, and information about persistence operations on cookies in the last
15203 two characters. The normal flags should begin with "--", indicating the
15204 session was closed by either end with no data remaining in buffers. See
15205 below "Session state at disconnection" for more details.
15206
15207 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040015208 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015209 limits have been reached. For instance, if actconn is close to 512 or 1024
15210 when multiple connection errors occur, chances are high that the system
15211 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015212 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015213 system.
15214
15215 - "feconn" is the total number of concurrent connections on the frontend when
15216 the session was logged. It is useful to estimate the amount of resource
15217 required to sustain high loads, and to detect when the frontend's "maxconn"
15218 has been reached. Most often when this value increases by huge jumps, it is
15219 because there is congestion on the backend servers, but sometimes it can be
15220 caused by a denial of service attack.
15221
15222 - "beconn" is the total number of concurrent connections handled by the
15223 backend when the session was logged. It includes the total number of
15224 concurrent connections active on servers as well as the number of
15225 connections pending in queues. It is useful to estimate the amount of
15226 additional servers needed to support high loads for a given application.
15227 Most often when this value increases by huge jumps, it is because there is
15228 congestion on the backend servers, but sometimes it can be caused by a
15229 denial of service attack.
15230
15231 - "srv_conn" is the total number of concurrent connections still active on
15232 the server when the session was logged. It can never exceed the server's
15233 configured "maxconn" parameter. If this value is very often close or equal
15234 to the server's "maxconn", it means that traffic regulation is involved a
15235 lot, meaning that either the server's maxconn value is too low, or that
15236 there aren't enough servers to process the load with an optimal response
15237 time. When only one of the server's "srv_conn" is high, it usually means
15238 that this server has some trouble causing the requests to take longer to be
15239 processed than on other servers.
15240
15241 - "retries" is the number of connection retries experienced by this session
15242 when trying to connect to the server. It must normally be zero, unless a
15243 server is being stopped at the same moment the connection was attempted.
15244 Frequent retries generally indicate either a network problem between
15245 haproxy and the server, or a misconfigured system backlog on the server
15246 preventing new connections from being queued. This field may optionally be
15247 prefixed with a '+' sign, indicating that the session has experienced a
15248 redispatch after the maximal retry count has been reached on the initial
15249 server. In this case, the server name appearing in the log is the one the
15250 connection was redispatched to, and not the first one, though both may
15251 sometimes be the same in case of hashing for instance. So as a general rule
15252 of thumb, when a '+' is present in front of the retry count, this count
15253 should not be attributed to the logged server.
15254
15255 - "srv_queue" is the total number of requests which were processed before
15256 this one in the server queue. It is zero when the request has not gone
15257 through the server queue. It makes it possible to estimate the approximate
15258 server's response time by dividing the time spent in queue by the number of
15259 requests in the queue. It is worth noting that if a session experiences a
15260 redispatch and passes through two server queues, their positions will be
15261 cumulated. A request should not pass through both the server queue and the
15262 backend queue unless a redispatch occurs.
15263
15264 - "backend_queue" is the total number of requests which were processed before
15265 this one in the backend's global queue. It is zero when the request has not
15266 gone through the global queue. It makes it possible to estimate the average
15267 queue length, which easily translates into a number of missing servers when
15268 divided by a server's "maxconn" parameter. It is worth noting that if a
15269 session experiences a redispatch, it may pass twice in the backend's queue,
15270 and then both positions will be cumulated. A request should not pass
15271 through both the server queue and the backend queue unless a redispatch
15272 occurs.
15273
15274 - "captured_request_headers" is a list of headers captured in the request due
15275 to the presence of the "capture request header" statement in the frontend.
15276 Multiple headers can be captured, they will be delimited by a vertical bar
15277 ('|'). When no capture is enabled, the braces do not appear, causing a
15278 shift of remaining fields. It is important to note that this field may
15279 contain spaces, and that using it requires a smarter log parser than when
15280 it's not used. Please consult the section "Capturing HTTP headers and
15281 cookies" below for more details.
15282
15283 - "captured_response_headers" is a list of headers captured in the response
15284 due to the presence of the "capture response header" statement in the
15285 frontend. Multiple headers can be captured, they will be delimited by a
15286 vertical bar ('|'). When no capture is enabled, the braces do not appear,
15287 causing a shift of remaining fields. It is important to note that this
15288 field may contain spaces, and that using it requires a smarter log parser
15289 than when it's not used. Please consult the section "Capturing HTTP headers
15290 and cookies" below for more details.
15291
15292 - "http_request" is the complete HTTP request line, including the method,
15293 request and HTTP version string. Non-printable characters are encoded (see
15294 below the section "Non-printable characters"). This is always the last
15295 field, and it is always delimited by quotes and is the only one which can
15296 contain quotes. If new fields are added to the log format, they will be
15297 added before this field. This field might be truncated if the request is
15298 huge and does not fit in the standard syslog buffer (1024 characters). This
15299 is the reason why this field must always remain the last one.
15300
15301
Cyril Bontédc4d9032012-04-08 21:57:39 +0200153028.2.4. Custom log format
15303------------------------
William Lallemand48940402012-01-30 16:47:22 +010015304
Willy Tarreau2beef582012-12-20 17:22:52 +010015305The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010015306mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010015307
15308HAproxy understands some log format variables. % precedes log format variables.
15309Variables can take arguments using braces ('{}'), and multiple arguments are
15310separated by commas within the braces. Flags may be added or removed by
15311prefixing them with a '+' or '-' sign.
15312
15313Special variable "%o" may be used to propagate its flags to all other
15314variables on the same format string. This is particularly handy with quoted
Dragan Dosen835b9212016-02-12 13:23:03 +010015315("Q") and escaped ("E") string formats.
William Lallemand48940402012-01-30 16:47:22 +010015316
Willy Tarreauc8368452012-12-21 00:09:23 +010015317If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020015318as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010015319less common information such as the client's SSL certificate's DN, or to log
15320the key that would be used to store an entry into a stick table.
15321
William Lallemand48940402012-01-30 16:47:22 +010015322Note: spaces must be escaped. A space character is considered as a separator.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015323In order to emit a verbatim '%', it must be preceded by another '%' resulting
Willy Tarreau06d97f92013-12-02 17:45:48 +010015324in '%%'. HAProxy will automatically merge consecutive separators.
William Lallemand48940402012-01-30 16:47:22 +010015325
Dragan Dosen835b9212016-02-12 13:23:03 +010015326Note: when using the RFC5424 syslog message format, the characters '"',
15327'\' and ']' inside PARAM-VALUE should be escaped with '\' as prefix (see
15328https://tools.ietf.org/html/rfc5424#section-6.3.3 for more details). In
15329such cases, the use of the flag "E" should be considered.
15330
William Lallemand48940402012-01-30 16:47:22 +010015331Flags are :
15332 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040015333 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
Dragan Dosen835b9212016-02-12 13:23:03 +010015334 * E: escape characters '"', '\' and ']' in a string with '\' as prefix
15335 (intended purpose is for the RFC5424 structured-data log formats)
William Lallemand48940402012-01-30 16:47:22 +010015336
15337 Example:
15338
15339 log-format %T\ %t\ Some\ Text
15340 log-format %{+Q}o\ %t\ %s\ %{-Q}r
15341
Dragan Dosen835b9212016-02-12 13:23:03 +010015342 log-format-sd %{+Q,+E}o\ [exampleSDID@1234\ header=%[capture.req.hdr(0)]]
15343
William Lallemand48940402012-01-30 16:47:22 +010015344At the moment, the default HTTP format is defined this way :
15345
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015346 log-format "%ci:%cp [%tr] %ft %b/%s %TR/%Tw/%Tc/%Tr/%Ta %ST %B %CC \
15347 %CS %tsc %ac/%fc/%bc/%sc/%rc %sq/%bq %hr %hs %{+Q}r"
William Lallemand48940402012-01-30 16:47:22 +010015348
William Lallemandbddd4fd2012-02-27 11:23:10 +010015349the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010015350
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015351 log-format "%{+Q}o %{-Q}ci - - [%trg] %r %ST %B \"\" \"\" %cp \
15352 %ms %ft %b %s %TR %Tw %Tc %Tr %Ta %tsc %ac %fc \
15353 %bc %sc %rc %sq %bq %CC %CS %hrl %hsl"
William Lallemand48940402012-01-30 16:47:22 +010015354
William Lallemandbddd4fd2012-02-27 11:23:10 +010015355and the default TCP format is defined this way :
15356
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015357 log-format "%ci:%cp [%t] %ft %b/%s %Tw/%Tc/%Tt %B %ts \
15358 %ac/%fc/%bc/%sc/%rc %sq/%bq"
William Lallemandbddd4fd2012-02-27 11:23:10 +010015359
William Lallemand48940402012-01-30 16:47:22 +010015360Please refer to the table below for currently defined variables :
15361
William Lallemandbddd4fd2012-02-27 11:23:10 +010015362 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020015363 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015364 +---+------+-----------------------------------------------+-------------+
15365 | | %o | special variable, apply flags on all next var | |
15366 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010015367 | | %B | bytes_read (from server to client) | numeric |
15368 | H | %CC | captured_request_cookie | string |
15369 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020015370 | | %H | hostname | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000015371 | H | %HM | HTTP method (ex: POST) | string |
15372 | H | %HP | HTTP request URI without query string (path) | string |
Andrew Hayworthe63ac872015-07-31 16:14:16 +000015373 | H | %HQ | HTTP request URI query string (ex: ?bar=baz) | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000015374 | H | %HU | HTTP request URI (ex: /foo?bar=baz) | string |
15375 | H | %HV | HTTP version (ex: HTTP/1.0) | string |
William Lallemanda73203e2012-03-12 12:48:57 +010015376 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020015377 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020015378 | | %T | gmt_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015379 | | %Ta | Active time of the request (from TR to end) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015380 | | %Tc | Tc | numeric |
Willy Tarreau27b639d2016-05-17 17:55:27 +020015381 | | %Td | Td = Tt - (Tq + Tw + Tc + Tr) | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080015382 | | %Tl | local_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015383 | | %Th | connection handshake time (SSL, PROXY proto) | numeric |
15384 | H | %Ti | idle time before the HTTP request | numeric |
15385 | H | %Tq | Th + Ti + TR | numeric |
15386 | H | %TR | time to receive the full request from 1st byte| numeric |
15387 | H | %Tr | Tr (response time) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020015388 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015389 | | %Tt | Tt | numeric |
15390 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010015391 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015392 | | %ac | actconn | numeric |
15393 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010015394 | | %bc | beconn (backend concurrent connections) | numeric |
15395 | | %bi | backend_source_ip (connecting address) | IP |
15396 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015397 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010015398 | | %ci | client_ip (accepted address) | IP |
15399 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015400 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010015401 | | %fc | feconn (frontend concurrent connections) | numeric |
15402 | | %fi | frontend_ip (accepting address) | IP |
15403 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020015404 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreau7346acb2014-08-28 15:03:15 +020015405 | | %lc | frontend_log_counter | numeric |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020015406 | | %hr | captured_request_headers default style | string |
15407 | | %hrl | captured_request_headers CLF style | string list |
15408 | | %hs | captured_response_headers default style | string |
15409 | | %hsl | captured_response_headers CLF style | string list |
Willy Tarreau812c88e2015-08-09 10:56:35 +020015410 | | %ms | accept date milliseconds (left-padded with 0) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020015411 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020015412 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015413 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010015414 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015415 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010015416 | | %sc | srv_conn (server concurrent connections) | numeric |
15417 | | %si | server_IP (target address) | IP |
15418 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015419 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020015420 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
15421 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010015422 | | %t | date_time (with millisecond resolution) | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015423 | H | %tr | date_time of HTTP request | date |
15424 | H | %trg | gmt_date_time of start of HTTP request | date |
15425 | H | %trl | locla_date_time of start of HTTP request | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015426 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020015427 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010015428 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010015429
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020015430 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010015431
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010015432
154338.2.5. Error log format
15434-----------------------
15435
15436When an incoming connection fails due to an SSL handshake or an invalid PROXY
15437protocol header, haproxy will log the event using a shorter, fixed line format.
15438By default, logs are emitted at the LOG_INFO level, unless the option
15439"log-separate-errors" is set in the backend, in which case the LOG_ERR level
15440will be used. Connections on which no data are exchanged (eg: probes) are not
15441logged if the "dontlognull" option is set.
15442
15443The format looks like this :
15444
15445 >>> Dec 3 18:27:14 localhost \
15446 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
15447 Connection error during SSL handshake
15448
15449 Field Format Extract from the example above
15450 1 process_name '[' pid ']:' haproxy[6103]:
15451 2 client_ip ':' client_port 127.0.0.1:56059
15452 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
15453 4 frontend_name "/" bind_name ":" frt/f1:
15454 5 message Connection error during SSL handshake
15455
15456These fields just provide minimal information to help debugging connection
15457failures.
15458
15459
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200154608.3. Advanced logging options
15461-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015462
15463Some advanced logging options are often looked for but are not easy to find out
15464just by looking at the various options. Here is an entry point for the few
15465options which can enable better logging. Please refer to the keywords reference
15466for more information about their usage.
15467
15468
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200154698.3.1. Disabling logging of external tests
15470------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015471
15472It is quite common to have some monitoring tools perform health checks on
15473haproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
15474commercial load-balancer, and sometimes it will simply be a more complete
15475monitoring system such as Nagios. When the tests are very frequent, users often
15476ask how to disable logging for those checks. There are three possibilities :
15477
15478 - if connections come from everywhere and are just TCP probes, it is often
15479 desired to simply disable logging of connections without data exchange, by
15480 setting "option dontlognull" in the frontend. It also disables logging of
15481 port scans, which may or may not be desired.
15482
15483 - if the connection come from a known source network, use "monitor-net" to
15484 declare this network as monitoring only. Any host in this network will then
15485 only be able to perform health checks, and their requests will not be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015486 logged. This is generally appropriate to designate a list of equipment
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015487 such as other load-balancers.
15488
15489 - if the tests are performed on a known URI, use "monitor-uri" to declare
15490 this URI as dedicated to monitoring. Any host sending this request will
15491 only get the result of a health-check, and the request will not be logged.
15492
15493
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200154948.3.2. Logging before waiting for the session to terminate
15495----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015496
15497The problem with logging at end of connection is that you have no clue about
15498what is happening during very long sessions, such as remote terminal sessions
15499or large file downloads. This problem can be worked around by specifying
15500"option logasap" in the frontend. Haproxy will then log as soon as possible,
15501just before data transfer begins. This means that in case of TCP, it will still
15502log the connection status to the server, and in case of HTTP, it will log just
15503after processing the server headers. In this case, the number of bytes reported
15504is the number of header bytes sent to the client. In order to avoid confusion
15505with normal logs, the total time field and the number of bytes are prefixed
15506with a '+' sign which means that real numbers are certainly larger.
15507
15508
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200155098.3.3. Raising log level upon errors
15510------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020015511
15512Sometimes it is more convenient to separate normal traffic from errors logs,
15513for instance in order to ease error monitoring from log files. When the option
15514"log-separate-errors" is used, connections which experience errors, timeouts,
15515retries, redispatches or HTTP status codes 5xx will see their syslog level
15516raised from "info" to "err". This will help a syslog daemon store the log in
15517a separate file. It is very important to keep the errors in the normal traffic
15518file too, so that log ordering is not altered. You should also be careful if
15519you already have configured your syslog daemon to store all logs higher than
15520"notice" in an "admin" file, because the "err" level is higher than "notice".
15521
15522
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200155238.3.4. Disabling logging of successful connections
15524--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020015525
15526Although this may sound strange at first, some large sites have to deal with
15527multiple thousands of logs per second and are experiencing difficulties keeping
15528them intact for a long time or detecting errors within them. If the option
15529"dontlog-normal" is set on the frontend, all normal connections will not be
15530logged. In this regard, a normal connection is defined as one without any
15531error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
15532and a response with a status 5xx is not considered normal and will be logged
15533too. Of course, doing is is really discouraged as it will remove most of the
15534useful information from the logs. Do this only if you have no other
15535alternative.
15536
15537
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200155388.4. Timing events
15539------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015540
15541Timers provide a great help in troubleshooting network problems. All values are
15542reported in milliseconds (ms). These timers should be used in conjunction with
15543the session termination flags. In TCP mode with "option tcplog" set on the
15544frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015545mode, 5 control points are reported under the form "TR/Tw/Tc/Tr/Ta". In
15546addition, three other measures are provided, "Th", "Ti", and "Tq".
15547
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010015548Timings events in HTTP mode:
15549
15550 first request 2nd request
15551 |<-------------------------------->|<-------------- ...
15552 t tr t tr ...
15553 ---|----|----|----|----|----|----|----|----|--
15554 : Th Ti TR Tw Tc Tr Td : Ti ...
15555 :<---- Tq ---->: :
15556 :<-------------- Tt -------------->:
15557 :<--------- Ta --------->:
15558
15559Timings events in TCP mode:
15560
15561 TCP session
15562 |<----------------->|
15563 t t
15564 ---|----|----|----|----|---
15565 | Th Tw Tc Td |
15566 |<------ Tt ------->|
15567
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015568 - Th: total time to accept tcp connection and execute handshakes for low level
15569 protocols. Currently, these protocoles are proxy-protocol and SSL. This may
15570 only happen once during the whole connection's lifetime. A large time here
15571 may indicate that the client only pre-established the connection without
15572 speaking, that it is experiencing network issues preventing it from
15573 completing a handshake in a reasonable time (eg: MTU issues), or that an
15574 SSL handshake was very expensive to compute.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015575
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015576 - Ti: is the idle time before the HTTP request (HTTP mode only). This timer
15577 counts between the end of the handshakes and the first byte of the HTTP
15578 request. When dealing with a second request in keep-alive mode, it starts
15579 to count after the end of the transmission the previous response. Some
15580 browsers pre-establish connections to a server in order to reduce the
15581 latency of a future request, and keep them pending until they need it. This
15582 delay will be reported as the idle time. A value of -1 indicates that
15583 nothing was received on the connection.
15584
15585 - TR: total time to get the client request (HTTP mode only). It's the time
15586 elapsed between the first bytes received and the moment the proxy received
15587 the empty line marking the end of the HTTP headers. The value "-1"
15588 indicates that the end of headers has never been seen. This happens when
15589 the client closes prematurely or times out. This time is usually very short
15590 since most requests fit in a single packet. A large time may indicate a
15591 request typed by hand during a test.
15592
15593 - Tq: total time to get the client request from the accept date or since the
15594 emission of the last byte of the previous response (HTTP mode only). It's
15595 exactly equalt to Th + Ti + TR unless any of them is -1, in which case it
15596 returns -1 as well. This timer used to be very useful before the arrival of
15597 HTTP keep-alive and browsers' pre-connect feature. It's recommended to drop
15598 it in favor of TR nowadays, as the idle time adds a lot of noise to the
15599 reports.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015600
15601 - Tw: total time spent in the queues waiting for a connection slot. It
15602 accounts for backend queue as well as the server queues, and depends on the
15603 queue size, and the time needed for the server to complete previous
15604 requests. The value "-1" means that the request was killed before reaching
15605 the queue, which is generally what happens with invalid or denied requests.
15606
15607 - Tc: total time to establish the TCP connection to the server. It's the time
15608 elapsed between the moment the proxy sent the connection request, and the
15609 moment it was acknowledged by the server, or between the TCP SYN packet and
15610 the matching SYN/ACK packet in return. The value "-1" means that the
15611 connection never established.
15612
15613 - Tr: server response time (HTTP mode only). It's the time elapsed between
15614 the moment the TCP connection was established to the server and the moment
15615 the server sent its complete response headers. It purely shows its request
15616 processing time, without the network overhead due to the data transmission.
15617 It is worth noting that when the client has data to send to the server, for
15618 instance during a POST request, the time already runs, and this can distort
15619 apparent response time. For this reason, it's generally wise not to trust
15620 too much this field for POST requests initiated from clients behind an
15621 untrusted network. A value of "-1" here means that the last the response
15622 header (empty line) was never seen, most likely because the server timeout
15623 stroke before the server managed to process the request.
15624
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015625 - Ta: total active time for the HTTP request, between the moment the proxy
15626 received the first byte of the request header and the emission of the last
15627 byte of the response body. The exception is when the "logasap" option is
15628 specified. In this case, it only equals (TR+Tw+Tc+Tr), and is prefixed with
15629 a '+' sign. From this field, we can deduce "Td", the data transmission time,
15630 by subtracting other timers when valid :
15631
15632 Td = Ta - (TR + Tw + Tc + Tr)
15633
15634 Timers with "-1" values have to be excluded from this equation. Note that
15635 "Ta" can never be negative.
15636
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015637 - Tt: total session duration time, between the moment the proxy accepted it
15638 and the moment both ends were closed. The exception is when the "logasap"
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015639 option is specified. In this case, it only equals (Th+Ti+TR+Tw+Tc+Tr), and
15640 is prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015641 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015642
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015643 Td = Tt - (Th + Ti + TR + Tw + Tc + Tr)
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015644
15645 Timers with "-1" values have to be excluded from this equation. In TCP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015646 mode, "Ti", "Tq" and "Tr" have to be excluded too. Note that "Tt" can never
15647 be negative and that for HTTP, Tt is simply equal to (Th+Ti+Ta).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015648
15649These timers provide precious indications on trouble causes. Since the TCP
15650protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
15651that timers close to multiples of 3s are nearly always related to lost packets
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015652due to network problems (wires, negotiation, congestion). Moreover, if "Ta" or
15653"Tt" is close to a timeout value specified in the configuration, it often means
15654that a session has been aborted on timeout.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015655
15656Most common cases :
15657
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015658 - If "Th" or "Ti" are close to 3000, a packet has probably been lost between
15659 the client and the proxy. This is very rare on local networks but might
15660 happen when clients are on far remote networks and send large requests. It
15661 may happen that values larger than usual appear here without any network
15662 cause. Sometimes, during an attack or just after a resource starvation has
15663 ended, haproxy may accept thousands of connections in a few milliseconds.
15664 The time spent accepting these connections will inevitably slightly delay
15665 processing of other connections, and it can happen that request times in the
15666 order of a few tens of milliseconds are measured after a few thousands of
15667 new connections have been accepted at once. Using one of the keep-alive
15668 modes may display larger idle times since "Ti" measures the time spent
Patrick Mezard105faca2010-06-12 17:02:46 +020015669 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015670
15671 - If "Tc" is close to 3000, a packet has probably been lost between the
15672 server and the proxy during the server connection phase. This value should
15673 always be very low, such as 1 ms on local networks and less than a few tens
15674 of ms on remote networks.
15675
Willy Tarreau55165fe2009-05-10 12:02:55 +020015676 - If "Tr" is nearly always lower than 3000 except some rare values which seem
15677 to be the average majored by 3000, there are probably some packets lost
15678 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015679
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015680 - If "Ta" is large even for small byte counts, it generally is because
15681 neither the client nor the server decides to close the connection while
15682 haproxy is running in tunnel mode and both have agreed on a keep-alive
15683 connection mode. In order to solve this issue, it will be needed to specify
15684 one of the HTTP options to manipulate keep-alive or close options on either
15685 the frontend or the backend. Having the smallest possible 'Ta' or 'Tt' is
15686 important when connection regulation is used with the "maxconn" option on
15687 the servers, since no new connection will be sent to the server until
15688 another one is released.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015689
15690Other noticeable HTTP log cases ('xx' means any value to be ignored) :
15691
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015692 TR/Tw/Tc/Tr/+Ta The "option logasap" is present on the frontend and the log
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015693 was emitted before the data phase. All the timers are valid
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015694 except "Ta" which is shorter than reality.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015695
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015696 -1/xx/xx/xx/Ta The client was not able to send a complete request in time
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015697 or it aborted too early. Check the session termination flags
15698 then "timeout http-request" and "timeout client" settings.
15699
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015700 TR/-1/xx/xx/Ta It was not possible to process the request, maybe because
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015701 servers were out of order, because the request was invalid
15702 or forbidden by ACL rules. Check the session termination
15703 flags.
15704
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015705 TR/Tw/-1/xx/Ta The connection could not establish on the server. Either it
15706 actively refused it or it timed out after Ta-(TR+Tw) ms.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015707 Check the session termination flags, then check the
15708 "timeout connect" setting. Note that the tarpit action might
15709 return similar-looking patterns, with "Tw" equal to the time
15710 the client connection was maintained open.
15711
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015712 TR/Tw/Tc/-1/Ta The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015713 a complete response in time, or it closed its connection
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020015714 unexpectedly after Ta-(TR+Tw+Tc) ms. Check the session
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015715 termination flags, then check the "timeout server" setting.
15716
15717
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200157188.5. Session state at disconnection
15719-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015720
15721TCP and HTTP logs provide a session termination indicator in the
15722"termination_state" field, just before the number of active connections. It is
157232-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
15724each of which has a special meaning :
15725
15726 - On the first character, a code reporting the first event which caused the
15727 session to terminate :
15728
15729 C : the TCP session was unexpectedly aborted by the client.
15730
15731 S : the TCP session was unexpectedly aborted by the server, or the
15732 server explicitly refused it.
15733
15734 P : the session was prematurely aborted by the proxy, because of a
15735 connection limit enforcement, because a DENY filter was matched,
15736 because of a security check which detected and blocked a dangerous
15737 error in server response which might have caused information leak
Willy Tarreau570f2212013-06-10 16:42:09 +020015738 (eg: cacheable cookie).
15739
15740 L : the session was locally processed by haproxy and was not passed to
15741 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015742
15743 R : a resource on the proxy has been exhausted (memory, sockets, source
15744 ports, ...). Usually, this appears during the connection phase, and
15745 system logs should contain a copy of the precise error. If this
15746 happens, it must be considered as a very serious anomaly which
15747 should be fixed as soon as possible by any means.
15748
15749 I : an internal error was identified by the proxy during a self-check.
15750 This should NEVER happen, and you are encouraged to report any log
15751 containing this, because this would almost certainly be a bug. It
15752 would be wise to preventively restart the process after such an
15753 event too, in case it would be caused by memory corruption.
15754
Simon Horman752dc4a2011-06-21 14:34:59 +090015755 D : the session was killed by haproxy because the server was detected
15756 as down and was configured to kill all connections when going down.
15757
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070015758 U : the session was killed by haproxy on this backup server because an
15759 active server was detected as up and was configured to kill all
15760 backup connections when going up.
15761
Willy Tarreaua2a64e92011-09-07 23:01:56 +020015762 K : the session was actively killed by an admin operating on haproxy.
15763
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015764 c : the client-side timeout expired while waiting for the client to
15765 send or receive data.
15766
15767 s : the server-side timeout expired while waiting for the server to
15768 send or receive data.
15769
15770 - : normal session completion, both the client and the server closed
15771 with nothing left in the buffers.
15772
15773 - on the second character, the TCP or HTTP session state when it was closed :
15774
Willy Tarreauf7b30a92010-12-06 22:59:17 +010015775 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015776 (HTTP mode only). Nothing was sent to any server.
15777
15778 Q : the proxy was waiting in the QUEUE for a connection slot. This can
15779 only happen when servers have a 'maxconn' parameter set. It can
15780 also happen in the global queue after a redispatch consecutive to
15781 a failed attempt to connect to a dying server. If no redispatch is
15782 reported, then no connection attempt was made to any server.
15783
15784 C : the proxy was waiting for the CONNECTION to establish on the
15785 server. The server might at most have noticed a connection attempt.
15786
15787 H : the proxy was waiting for complete, valid response HEADERS from the
15788 server (HTTP only).
15789
15790 D : the session was in the DATA phase.
15791
15792 L : the proxy was still transmitting LAST data to the client while the
15793 server had already finished. This one is very rare as it can only
15794 happen when the client dies while receiving the last packets.
15795
15796 T : the request was tarpitted. It has been held open with the client
15797 during the whole "timeout tarpit" duration or until the client
15798 closed, both of which will be reported in the "Tw" timer.
15799
15800 - : normal session completion after end of data transfer.
15801
15802 - the third character tells whether the persistence cookie was provided by
15803 the client (only in HTTP mode) :
15804
15805 N : the client provided NO cookie. This is usually the case for new
15806 visitors, so counting the number of occurrences of this flag in the
15807 logs generally indicate a valid trend for the site frequentation.
15808
15809 I : the client provided an INVALID cookie matching no known server.
15810 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020015811 cookies between HTTP/HTTPS sites, persistence conditionally
15812 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015813
15814 D : the client provided a cookie designating a server which was DOWN,
15815 so either "option persist" was used and the client was sent to
15816 this server, or it was not set and the client was redispatched to
15817 another server.
15818
Willy Tarreau996a92c2010-10-13 19:30:47 +020015819 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015820 server.
15821
Willy Tarreau996a92c2010-10-13 19:30:47 +020015822 E : the client provided a valid cookie, but with a last date which was
15823 older than what is allowed by the "maxidle" cookie parameter, so
15824 the cookie is consider EXPIRED and is ignored. The request will be
15825 redispatched just as if there was no cookie.
15826
15827 O : the client provided a valid cookie, but with a first date which was
15828 older than what is allowed by the "maxlife" cookie parameter, so
15829 the cookie is consider too OLD and is ignored. The request will be
15830 redispatched just as if there was no cookie.
15831
Willy Tarreauc89ccb62012-04-05 21:18:22 +020015832 U : a cookie was present but was not used to select the server because
15833 some other server selection mechanism was used instead (typically a
15834 "use-server" rule).
15835
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015836 - : does not apply (no cookie set in configuration).
15837
15838 - the last character reports what operations were performed on the persistence
15839 cookie returned by the server (only in HTTP mode) :
15840
15841 N : NO cookie was provided by the server, and none was inserted either.
15842
15843 I : no cookie was provided by the server, and the proxy INSERTED one.
15844 Note that in "cookie insert" mode, if the server provides a cookie,
15845 it will still be overwritten and reported as "I" here.
15846
Willy Tarreau996a92c2010-10-13 19:30:47 +020015847 U : the proxy UPDATED the last date in the cookie that was presented by
15848 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015849 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020015850 date indicated in the cookie. If any other change happens, such as
15851 a redispatch, then the cookie will be marked as inserted instead.
15852
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015853 P : a cookie was PROVIDED by the server and transmitted as-is.
15854
15855 R : the cookie provided by the server was REWRITTEN by the proxy, which
15856 happens in "cookie rewrite" or "cookie prefix" modes.
15857
15858 D : the cookie provided by the server was DELETED by the proxy.
15859
15860 - : does not apply (no cookie set in configuration).
15861
Willy Tarreau996a92c2010-10-13 19:30:47 +020015862The combination of the two first flags gives a lot of information about what
15863was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015864helpful to detect server saturation, network troubles, local system resource
15865starvation, attacks, etc...
15866
15867The most common termination flags combinations are indicated below. They are
15868alphabetically sorted, with the lowercase set just after the upper case for
15869easier finding and understanding.
15870
15871 Flags Reason
15872
15873 -- Normal termination.
15874
15875 CC The client aborted before the connection could be established to the
15876 server. This can happen when haproxy tries to connect to a recently
15877 dead (or unchecked) server, and the client aborts while haproxy is
15878 waiting for the server to respond or for "timeout connect" to expire.
15879
15880 CD The client unexpectedly aborted during data transfer. This can be
15881 caused by a browser crash, by an intermediate equipment between the
15882 client and haproxy which decided to actively break the connection,
15883 by network routing issues between the client and haproxy, or by a
15884 keep-alive session between the server and the client terminated first
15885 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010015886
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015887 cD The client did not send nor acknowledge any data for as long as the
15888 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020015889 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015890
15891 CH The client aborted while waiting for the server to start responding.
15892 It might be the server taking too long to respond or the client
15893 clicking the 'Stop' button too fast.
15894
15895 cH The "timeout client" stroke while waiting for client data during a
15896 POST request. This is sometimes caused by too large TCP MSS values
15897 for PPPoE networks which cannot transport full-sized packets. It can
15898 also happen when client timeout is smaller than server timeout and
15899 the server takes too long to respond.
15900
15901 CQ The client aborted while its session was queued, waiting for a server
15902 with enough empty slots to accept it. It might be that either all the
15903 servers were saturated or that the assigned server was taking too
15904 long a time to respond.
15905
15906 CR The client aborted before sending a full HTTP request. Most likely
15907 the request was typed by hand using a telnet client, and aborted
15908 too early. The HTTP status code is likely a 400 here. Sometimes this
15909 might also be caused by an IDS killing the connection between haproxy
Willy Tarreau0f228a02015-05-01 15:37:53 +020015910 and the client. "option http-ignore-probes" can be used to ignore
15911 connections without any data transfer.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015912
15913 cR The "timeout http-request" stroke before the client sent a full HTTP
15914 request. This is sometimes caused by too large TCP MSS values on the
15915 client side for PPPoE networks which cannot transport full-sized
15916 packets, or by clients sending requests by hand and not typing fast
15917 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020015918 request. The HTTP status code is likely a 408 here. Note: recently,
Willy Tarreau0f228a02015-05-01 15:37:53 +020015919 some browsers started to implement a "pre-connect" feature consisting
15920 in speculatively connecting to some recently visited web sites just
15921 in case the user would like to visit them. This results in many
15922 connections being established to web sites, which end up in 408
15923 Request Timeout if the timeout strikes first, or 400 Bad Request when
15924 the browser decides to close them first. These ones pollute the log
15925 and feed the error counters. Some versions of some browsers have even
15926 been reported to display the error code. It is possible to work
15927 around the undesirable effects of this behaviour by adding "option
15928 http-ignore-probes" in the frontend, resulting in connections with
15929 zero data transfer to be totally ignored. This will definitely hide
15930 the errors of people experiencing connectivity issues though.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015931
15932 CT The client aborted while its session was tarpitted. It is important to
15933 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020015934 wrong tarpit rules have been written. If a lot of them happen, it
15935 might make sense to lower the "timeout tarpit" value to something
15936 closer to the average reported "Tw" timer, in order not to consume
15937 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015938
Willy Tarreau570f2212013-06-10 16:42:09 +020015939 LR The request was intercepted and locally handled by haproxy. Generally
15940 it means that this was a redirect or a stats request.
15941
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010015942 SC The server or an equipment between it and haproxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015943 the TCP connection (the proxy received a TCP RST or an ICMP message
15944 in return). Under some circumstances, it can also be the network
15945 stack telling the proxy that the server is unreachable (eg: no route,
15946 or no ARP response on local network). When this happens in HTTP mode,
15947 the status code is likely a 502 or 503 here.
15948
15949 sC The "timeout connect" stroke before a connection to the server could
15950 complete. When this happens in HTTP mode, the status code is likely a
15951 503 or 504 here.
15952
15953 SD The connection to the server died with an error during the data
15954 transfer. This usually means that haproxy has received an RST from
15955 the server or an ICMP message from an intermediate equipment while
15956 exchanging data with the server. This can be caused by a server crash
15957 or by a network issue on an intermediate equipment.
15958
15959 sD The server did not send nor acknowledge any data for as long as the
15960 "timeout server" setting during the data phase. This is often caused
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010015961 by too short timeouts on L4 equipments before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015962 load-balancers, ...), as well as keep-alive sessions maintained
15963 between the client and the server expiring first on haproxy.
15964
15965 SH The server aborted before sending its full HTTP response headers, or
15966 it crashed while processing the request. Since a server aborting at
15967 this moment is very rare, it would be wise to inspect its logs to
15968 control whether it crashed and why. The logged request may indicate a
15969 small set of faulty requests, demonstrating bugs in the application.
15970 Sometimes this might also be caused by an IDS killing the connection
15971 between haproxy and the server.
15972
15973 sH The "timeout server" stroke before the server could return its
15974 response headers. This is the most common anomaly, indicating too
15975 long transactions, probably caused by server or database saturation.
15976 The immediate workaround consists in increasing the "timeout server"
15977 setting, but it is important to keep in mind that the user experience
15978 will suffer from these long response times. The only long term
15979 solution is to fix the application.
15980
15981 sQ The session spent too much time in queue and has been expired. See
15982 the "timeout queue" and "timeout connect" settings to find out how to
15983 fix this if it happens too often. If it often happens massively in
15984 short periods, it may indicate general problems on the affected
15985 servers due to I/O or database congestion, or saturation caused by
15986 external attacks.
15987
15988 PC The proxy refused to establish a connection to the server because the
15989 process' socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020015990 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010015991 so that it does not happen anymore. This status is very rare and
15992 might happen when the global "ulimit-n" parameter is forced by hand.
15993
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010015994 PD The proxy blocked an incorrectly formatted chunked encoded message in
15995 a request or a response, after the server has emitted its headers. In
15996 most cases, this will indicate an invalid message from the server to
Willy Tarreauf3a3e132013-08-31 08:16:26 +020015997 the client. Haproxy supports chunk sizes of up to 2GB - 1 (2147483647
15998 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010015999
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016000 PH The proxy blocked the server's response, because it was invalid,
16001 incomplete, dangerous (cache control), or matched a security filter.
16002 In any case, an HTTP 502 error is sent to the client. One possible
16003 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010016004 containing unauthorized characters. It is also possible but quite
16005 rare, that the proxy blocked a chunked-encoding request from the
16006 client due to an invalid syntax, before the server responded. In this
16007 case, an HTTP 400 error is sent to the client and reported in the
16008 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016009
16010 PR The proxy blocked the client's HTTP request, either because of an
16011 invalid HTTP syntax, in which case it returned an HTTP 400 error to
16012 the client, or because a deny filter matched, in which case it
16013 returned an HTTP 403 error.
16014
16015 PT The proxy blocked the client's request and has tarpitted its
16016 connection before returning it a 500 server error. Nothing was sent
16017 to the server. The connection was maintained open for as long as
16018 reported by the "Tw" timer field.
16019
16020 RC A local resource has been exhausted (memory, sockets, source ports)
16021 preventing the connection to the server from establishing. The error
16022 logs will tell precisely what was missing. This is very rare and can
16023 only be solved by proper system tuning.
16024
Willy Tarreau996a92c2010-10-13 19:30:47 +020016025The combination of the two last flags gives a lot of information about how
16026persistence was handled by the client, the server and by haproxy. This is very
16027important to troubleshoot disconnections, when users complain they have to
16028re-authenticate. The commonly encountered flags are :
16029
16030 -- Persistence cookie is not enabled.
16031
16032 NN No cookie was provided by the client, none was inserted in the
16033 response. For instance, this can be in insert mode with "postonly"
16034 set on a GET request.
16035
16036 II A cookie designating an invalid server was provided by the client,
16037 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040016038 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020016039 value can be presented by a client when no other server knows it.
16040
16041 NI No cookie was provided by the client, one was inserted in the
16042 response. This typically happens for first requests from every user
16043 in "insert" mode, which makes it an easy way to count real users.
16044
16045 VN A cookie was provided by the client, none was inserted in the
16046 response. This happens for most responses for which the client has
16047 already got a cookie.
16048
16049 VU A cookie was provided by the client, with a last visit date which is
16050 not completely up-to-date, so an updated cookie was provided in
16051 response. This can also happen if there was no date at all, or if
16052 there was a date but the "maxidle" parameter was not set, so that the
16053 cookie can be switched to unlimited time.
16054
16055 EI A cookie was provided by the client, with a last visit date which is
16056 too old for the "maxidle" parameter, so the cookie was ignored and a
16057 new cookie was inserted in the response.
16058
16059 OI A cookie was provided by the client, with a first visit date which is
16060 too old for the "maxlife" parameter, so the cookie was ignored and a
16061 new cookie was inserted in the response.
16062
16063 DI The server designated by the cookie was down, a new server was
16064 selected and a new cookie was emitted in the response.
16065
16066 VI The server designated by the cookie was not marked dead but could not
16067 be reached. A redispatch happened and selected another one, which was
16068 then advertised in the response.
16069
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016070
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200160718.6. Non-printable characters
16072-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016073
16074In order not to cause trouble to log analysis tools or terminals during log
16075consulting, non-printable characters are not sent as-is into log files, but are
16076converted to the two-digits hexadecimal representation of their ASCII code,
16077prefixed by the character '#'. The only characters that can be logged without
16078being escaped are comprised between 32 and 126 (inclusive). Obviously, the
16079escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
16080is the same for the character '"' which becomes "#22", as well as '{', '|' and
16081'}' when logging headers.
16082
16083Note that the space character (' ') is not encoded in headers, which can cause
16084issues for tools relying on space count to locate fields. A typical header
16085containing spaces is "User-Agent".
16086
16087Last, it has been observed that some syslog daemons such as syslog-ng escape
16088the quote ('"') with a backslash ('\'). The reverse operation can safely be
16089performed since no quote may appear anywhere else in the logs.
16090
16091
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200160928.7. Capturing HTTP cookies
16093---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016094
16095Cookie capture simplifies the tracking a complete user session. This can be
16096achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016097section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016098cookie will simultaneously be checked in the request ("Cookie:" header) and in
16099the response ("Set-Cookie:" header). The respective values will be reported in
16100the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016101locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016102not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
16103user switches to a new session for example, because the server will reassign it
16104a new cookie. It is also possible to detect if a server unexpectedly sets a
16105wrong cookie to a client, leading to session crossing.
16106
16107 Examples :
16108 # capture the first cookie whose name starts with "ASPSESSION"
16109 capture cookie ASPSESSION len 32
16110
16111 # capture the first cookie whose name is exactly "vgnvisitor"
16112 capture cookie vgnvisitor= len 32
16113
16114
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200161158.8. Capturing HTTP headers
16116---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016117
16118Header captures are useful to track unique request identifiers set by an upper
16119proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
16120the response, one can search for information about the response length, how the
16121server asked the cache to behave, or an object location during a redirection.
16122
16123Header captures are performed using the "capture request header" and "capture
16124response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016125section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016126
16127It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010016128time. Non-existent headers are logged as empty strings, and if one header
16129appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016130are grouped within braces '{' and '}' in the same order as they were declared,
16131and delimited with a vertical bar '|' without any space. Response headers
16132follow the same representation, but are displayed after a space following the
16133request headers block. These blocks are displayed just before the HTTP request
16134in the logs.
16135
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020016136As a special case, it is possible to specify an HTTP header capture in a TCP
16137frontend. The purpose is to enable logging of headers which will be parsed in
16138an HTTP backend if the request is then switched to this HTTP backend.
16139
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016140 Example :
16141 # This instance chains to the outgoing proxy
16142 listen proxy-out
16143 mode http
16144 option httplog
16145 option logasap
16146 log global
16147 server cache1 192.168.1.1:3128
16148
16149 # log the name of the virtual server
16150 capture request header Host len 20
16151
16152 # log the amount of data uploaded during a POST
16153 capture request header Content-Length len 10
16154
16155 # log the beginning of the referrer
16156 capture request header Referer len 20
16157
16158 # server name (useful for outgoing proxies only)
16159 capture response header Server len 20
16160
16161 # logging the content-length is useful with "option logasap"
16162 capture response header Content-Length len 10
16163
16164 # log the expected cache behaviour on the response
16165 capture response header Cache-Control len 8
16166
16167 # the Via header will report the next proxy's name
16168 capture response header Via len 20
16169
16170 # log the URL location during a redirection
16171 capture response header Location len 20
16172
16173 >>> Aug 9 20:26:09 localhost \
16174 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
16175 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
16176 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
16177 "GET http://fr.adserver.yahoo.com/"
16178
16179 >>> Aug 9 20:30:46 localhost \
16180 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
16181 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
16182 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010016183 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016184
16185 >>> Aug 9 20:30:46 localhost \
16186 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
16187 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
16188 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
16189 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010016190 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016191
16192
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200161938.9. Examples of logs
16194---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016195
16196These are real-world examples of logs accompanied with an explanation. Some of
16197them have been made up by hand. The syslog part has been removed for better
16198reading. Their sole purpose is to explain how to decipher them.
16199
16200 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
16201 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
16202 "HEAD / HTTP/1.0"
16203
16204 => long request (6.5s) entered by hand through 'telnet'. The server replied
16205 in 147 ms, and the session ended normally ('----')
16206
16207 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
16208 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
16209 0/9 "HEAD / HTTP/1.0"
16210
16211 => Idem, but the request was queued in the global queue behind 9 other
16212 requests, and waited there for 1230 ms.
16213
16214 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
16215 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
16216 "GET /image.iso HTTP/1.0"
16217
16218 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010016219 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016220 14 ms, 243 bytes of headers were sent to the client, and total time from
16221 accept to first data byte is 30 ms.
16222
16223 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
16224 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
16225 "GET /cgi-bin/bug.cgi? HTTP/1.0"
16226
16227 => the proxy blocked a server response either because of an "rspdeny" or
16228 "rspideny" filter, or because the response was improperly formatted and
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020016229 not HTTP-compliant, or because it blocked sensitive information which
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016230 risked being cached. In this case, the response is replaced with a "502
16231 bad gateway". The flags ("PH--") tell us that it was haproxy who decided
16232 to return the 502 and not the server.
16233
16234 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010016235 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 +010016236
16237 => the client never completed its request and aborted itself ("C---") after
16238 8.5s, while the proxy was waiting for the request headers ("-R--").
16239 Nothing was sent to any server.
16240
16241 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
16242 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
16243
16244 => The client never completed its request, which was aborted by the
16245 time-out ("c---") after 50s, while the proxy was waiting for the request
16246 headers ("-R--"). Nothing was sent to any server, but the proxy could
16247 send a 408 return code to the client.
16248
16249 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
16250 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
16251
16252 => This log was produced with "option tcplog". The client timed out after
16253 5 seconds ("c----").
16254
16255 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
16256 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010016257 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016258
16259 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016260 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016261 (config says 'retries 3'), and no redispatch (otherwise we would have
16262 seen "/+3"). Status code 503 was returned to the client. There were 115
16263 connections on this server, 202 connections on this proxy, and 205 on
16264 the global process. It is possible that the server refused the
16265 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010016266
Willy Tarreau52b2d222011-09-07 23:48:48 +020016267
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200162689. Supported filters
16269--------------------
16270
16271Here are listed officially supported filters with the list of parameters they
16272accept. Depending on compile options, some of these filters might be
16273unavailable. The list of available filters is reported in haproxy -vv.
16274
16275See also : "filter"
16276
162779.1. Trace
16278----------
16279
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010016280filter trace [name <name>] [random-parsing] [random-forwarding] [hexdump]
Christopher Fauletc3fe5332016-04-07 15:30:10 +020016281
16282 Arguments:
16283 <name> is an arbitrary name that will be reported in
16284 messages. If no name is provided, "TRACE" is used.
16285
16286 <random-parsing> enables the random parsing of data exchanged between
16287 the client and the server. By default, this filter
16288 parses all available data. With this parameter, it
16289 only parses a random amount of the available data.
16290
16291 <random-forwarding> enables the random forwading of parsed data. By
16292 default, this filter forwards all previously parsed
16293 data. With this parameter, it only forwards a random
16294 amount of the parsed data.
16295
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010016296 <hexump> dumps all forwarded data to the server and the client.
16297
Christopher Fauletc3fe5332016-04-07 15:30:10 +020016298This filter can be used as a base to develop new filters. It defines all
16299callbacks and print a message on the standard error stream (stderr) with useful
16300information for all of them. It may be useful to debug the activity of other
16301filters or, quite simply, HAProxy's activity.
16302
16303Using <random-parsing> and/or <random-forwarding> parameters is a good way to
16304tests the behavior of a filter that parses data exchanged between a client and
16305a server by adding some latencies in the processing.
16306
16307
163089.2. HTTP compression
16309---------------------
16310
16311filter compression
16312
16313The HTTP compression has been moved in a filter in HAProxy 1.7. "compression"
16314keyword must still be used to enable and configure the HTTP compression. And
16315when no other filter is used, it is enough. But it is mandatory to explicitly
16316use a filter line to enable the HTTP compression when two or more filters are
16317used for the same listener/frontend/backend. This is important to know the
16318filters evaluation order.
16319
16320See also : "compression"
16321
16322
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +0200163239.3. Stream Processing Offload Engine (SPOE)
16324--------------------------------------------
16325
16326filter spoe [engine <name>] config <file>
16327
16328 Arguments :
16329
16330 <name> is the engine name that will be used to find the right scope in
16331 the configuration file. If not provided, all the file will be
16332 parsed.
16333
16334 <file> is the path of the engine configuration file. This file can
16335 contain configuration of several engines. In this case, each
16336 part must be placed in its own scope.
16337
16338The Stream Processing Offload Engine (SPOE) is a filter communicating with
16339external components. It allows the offload of some specifics processing on the
16340streams in tierce applications. These external components and information
16341exchanged with them are configured in dedicated files, for the main part. It
16342also requires dedicated backends, defined in HAProxy configuration.
16343
16344SPOE communicates with external components using an in-house binary protocol,
16345the Stream Processing Offload Protocol (SPOP).
16346
Tim Düsterhus4896c442016-11-29 02:15:19 +010016347For all information about the SPOE configuration and the SPOP specification, see
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020016348"doc/SPOE.txt".
16349
16350Important note:
16351 The SPOE filter is highly experimental for now and was not heavily
16352 tested. It is really not production ready. So use it carefully.
16353
Willy Tarreau0ba27502007-12-24 16:55:16 +010016354/*
16355 * Local variables:
16356 * fill-column: 79
16357 * End:
16358 */