2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the documentation for the Squid configuration file.
34 This documentation can also be found online at:
35 http://www.squid-cache.org/Doc/config/
37 You may wish to look at the Squid home page and wiki for the
38 FAQ and other documentation:
39 http://www.squid-cache.org/
40 http://wiki.squid-cache.org/SquidFaq
41 http://wiki.squid-cache.org/ConfigExamples
43 This documentation shows what the defaults for various directives
44 happen to be. If you don't need to change the default, you should
45 leave the line out of your squid.conf in most cases.
47 In some cases "none" refers to no default setting at all,
48 while in other cases it refers to the value of the option
49 - the comments for that keyword indicate if this is the case.
54 Configuration options can be included using the "include" directive.
55 Include takes a list of files to include. Quoting and wildcards are
60 include /path/to/included/file/squid.acl.config
62 Includes can be nested up to a hard-coded depth of 16 levels.
63 This arbitrary restriction is to prevent recursive include references
64 from causing Squid entering an infinite loop whilst trying to load
67 Values with byte units
69 Squid accepts size units on some size related directives. All
70 such directives are documented with a default value displaying
73 Units accepted by Squid are:
75 KB - Kilobyte (1024 bytes)
79 Values with spaces, quotes, and other special characters
81 Squid supports directive parameters with spaces, quotes, and other
82 special characters. Surround such parameters with "double quotes". Use
83 the configuration_includes_quoted_values directive to enable or
86 Squid supports reading configuration option parameters from external
87 files using the syntax:
88 parameters("/path/filename")
90 acl whitelist dstdomain parameters("/etc/squid/whitelist.txt")
92 Conditional configuration
94 If-statements can be used to make configuration directives
98 ... regular configuration directives ...
100 ... regular configuration directives ...]
103 The else part is optional. The keywords "if", "else", and "endif"
104 must be typed on their own lines, as if they were regular
105 configuration directives.
107 NOTE: An else-if condition is not supported.
109 These individual conditions types are supported:
112 Always evaluates to true.
114 Always evaluates to false.
115 <integer> = <integer>
116 Equality comparison of two integer numbers.
121 The following SMP-related preprocessor macros can be used.
123 ${process_name} expands to the current Squid process "name"
124 (e.g., squid1, squid2, or cache1).
126 ${process_number} expands to the current Squid process
127 identifier, which is an integer number (e.g., 1, 2, 3) unique
128 across all Squid processes of the current service instance.
130 ${service_name} expands into the current Squid service instance
131 name identifier which is provided by -n on the command line.
135 # options still not yet ported from 2.7 to 3.x
136 NAME: broken_vary_encoding
139 This option is not yet supported by Squid-3.
145 This option is not yet supported by Squid-3.
151 This option is not yet supported by Squid-3.
154 NAME: external_refresh_check
157 This option is not yet supported by Squid-3.
160 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
163 This option is not yet supported by Squid-3.
166 NAME: refresh_stale_hit
169 This option is not yet supported by Squid-3.
172 # Options Removed in 3.3
173 NAME: ignore_ims_on_miss
176 Remove this line. The HTTP/1.1 feature is now fully supported by default.
179 # Options Removed in 3.2
180 NAME: ignore_expect_100
183 Remove this line. The HTTP/1.1 feature is now fully supported by default.
186 NAME: dns_v4_fallback
189 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
195 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
198 NAME: maximum_single_addr_tries
201 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
207 Remove this line. The feature is supported by default in storage types where update is implemented.
210 NAME: url_rewrite_concurrency
213 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
216 # Options Removed in 3.1
220 Remove this line. DNS is no longer tested on startup.
223 NAME: extension_methods
226 Remove this line. All valid methods for HTTP are accepted by default.
229 # 2.7 Options Removed/Replaced in 3.2
234 # 2.7 Options Removed/Replaced in 3.1
242 Remove this line. HTTP/1.1 is supported by default.
245 NAME: upgrade_http0.9
248 Remove this line. ICY/1.0 streaming protocol is supported by default.
251 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
254 Alter these entries. Use the qos_flows directive instead.
257 # Options Removed in 3.0
261 Since squid-3.0 replace with request_header_access or reply_header_access
262 depending on whether you wish to match client requests or server replies.
265 NAME: httpd_accel_no_pmtu_disc
268 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
271 NAME: wais_relay_host
274 Replace this line with 'cache_peer' configuration.
277 NAME: wais_relay_port
280 Replace this line with 'cache_peer' configuration.
284 OPTIONS FOR AUTHENTICATION
285 -----------------------------------------------------------------------------
294 This is used to define parameters for the various authentication
295 schemes supported by Squid.
297 format: auth_param scheme parameter [setting]
299 The order in which authentication schemes are presented to the client is
300 dependent on the order the scheme first appears in config file. IE
301 has a bug (it's not RFC 2617 compliant) in that it will use the basic
302 scheme if basic is the first entry presented, even if more secure
303 schemes are presented. For now use the order in the recommended
304 settings section below. If other browsers have difficulties (don't
305 recognize the schemes offered even if you are using basic) either
306 put basic first, or disable the other schemes (by commenting out their
309 Once an authentication scheme is fully configured, it can only be
310 shutdown by shutting squid down and restarting. Changes can be made on
311 the fly and activated with a reconfigure. I.E. You can change to a
312 different helper, but not unconfigure the helper completely.
314 Please note that while this directive defines how Squid processes
315 authentication it does not automatically activate authentication.
316 To use authentication you must in addition make use of ACLs based
317 on login name in http_access (proxy_auth, proxy_auth_regex or
318 external with %LOGIN used in the format tag). The browser will be
319 challenged for authentication on the first such acl encountered
320 in http_access processing and will also be re-challenged for new
321 login credentials if the request is being denied by a proxy_auth
324 WARNING: authentication can't be used in a transparently intercepting
325 proxy as the client then thinks it is talking to an origin server and
326 not the proxy. This is a limitation of bending the TCP/IP protocol to
327 transparently intercepting port 80, not a limitation in Squid.
328 Ports flagged 'transparent', 'intercept', or 'tproxy' have
329 authentication disabled.
331 === Parameters common to all schemes. ===
334 Specifies the command for the external authenticator. Such a program
335 runs a loop that, on every iteration, reads a request line from
336 the standard and responds with a scheme-specific answer. The loop
337 stops when all input is exchausted (EOF). See scheme-specific
338 "program" descriptions below for details.
341 Specifies a string to be append to request line format for the
342 authentication helper. "Quoted" format values may contain spaces and
343 logformat %macros. In theory, any logformat %macro can be used.
344 In practice, a %macro expands as a dash (-) if the helper request is
345 sent before the required macro information is available to Squid.
346 By default, Squid uses request formats provided in scheme-specific
347 examples below (search for %credentials).
348 The expanded key_extras value is added to the Squid credentials
349 cache and, hence, will affect authentication. It can be used to
350 autenticate different users with identical user names (e.g., when user
351 authentication depends on http_port).
352 Avoid adding frequently changing information to key_extras. For
353 example, if you add user source IP, and it changes frequently
354 in your environment, then max_user_ip ACL is going to treat every
355 user+IP combination as a unique "user", breaking the ACL and
356 wasting a lot of memory on those user records. It will also force
357 users to authenticate from scratch whenever their IP changes.
359 === Parameters for the basic scheme follow. ===
362 Specify the command for the external authenticator. Such a program
363 reads a request line ("username password" by default) and replies
364 with one of three results:
370 the user does not exist.
373 An internal error occurred in the helper, preventing
374 a result being identified.
376 "ERR" and "BH" results may optionally be followed by message="..."
377 containing a description available as %m in the returned error page.
379 If you use an authenticator, make sure you have 1 acl of type
382 By default, the basic authentication scheme is not used unless a
383 program is specified.
385 If you want to use the traditional NCSA proxy authentication, set
386 this line to something like
388 auth_param basic program @DEFAULT_PREFIX@/libexec/basic_ncsa_auth @DEFAULT_PREFIX@/etc/passwd
391 HTTP uses iso-latin-1 as character set, while some authentication
392 backends such as LDAP expects UTF-8. If this is set to on Squid will
393 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
394 username & password to the helper.
396 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
397 The maximum number of authenticator processes to spawn. If you start too few
398 Squid will have to wait for them to process a backlog of credential
399 verifications, slowing it down. When password verifications are
400 done via a (slow) network you are likely to need lots of
401 authenticator processes.
403 The startup= and idle= options permit some skew in the exact amount
404 run. A minimum of startup=N will begin during startup and reconfigure.
405 Squid will start more in groups of up to idle=N in an attempt to meet
406 traffic needs and to keep idle=N free above those traffic needs up to
409 The concurrency= option sets the number of concurrent requests the
410 helper can process. The default of 0 is used for helpers who only
411 supports one request at a time. Setting this to a number greater than
412 0 changes the protocol used to include a channel number first on the
413 request/response line, allowing multiple requests to be sent to the
414 same helper in parallel without waiting for the response.
415 Must not be set unless it's known the helper supports this.
417 auth_param basic children 20 startup=0 idle=1
420 Specifies the realm name which is to be reported to the
421 client for the basic proxy authentication scheme (part of
422 the text the user will see when prompted their username and
423 password). There is no default.
424 auth_param basic realm Squid proxy-caching web server
426 "credentialsttl" timetolive
427 Specifies how long squid assumes an externally validated
428 username:password pair is valid for - in other words how
429 often the helper program is called for that user. Set this
430 low to force revalidation with short lived passwords. Note
431 setting this high does not impact your susceptibility
432 to replay attacks unless you are using an one-time password
433 system (such as SecureID). If you are using such a system,
434 you will be vulnerable to replay attacks unless you also
435 use the max_user_ip ACL in an http_access rule.
437 "casesensitive" on|off
438 Specifies if usernames are case sensitive. Most user databases are
439 case insensitive allowing the same username to be spelled using both
440 lower and upper case letters, but some are case sensitive. This
441 makes a big difference for user_max_ip ACL processing and similar.
442 auth_param basic casesensitive off
444 === Parameters for the digest scheme follow ===
447 Specify the command for the external authenticator. Such a program
448 reads a request_format line ("username":"realm" by default) and
449 replies with one of three results:
452 the user exists. The ha1= key is mandatory and
453 contains the appropriate H(A1) value, hex encoded.
454 See rfc 2616 for the definition of H(A1).
457 the user does not exist.
460 An internal error occurred in the helper, preventing
461 a result being identified.
463 "ERR" and "BH" results may optionally be followed by message="..."
464 containing a description available as %m in the returned error page.
466 By default, the digest authentication scheme is not used unless a
467 program is specified.
469 If you want to use a digest authenticator, set this line to
472 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
475 HTTP uses iso-latin-1 as character set, while some authentication
476 backends such as LDAP expects UTF-8. If this is set to on Squid will
477 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
478 username & password to the helper.
480 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
481 The maximum number of authenticator processes to spawn (default 5).
482 If you start too few Squid will have to wait for them to
483 process a backlog of H(A1) calculations, slowing it down.
484 When the H(A1) calculations are done via a (slow) network
485 you are likely to need lots of authenticator processes.
487 The startup= and idle= options permit some skew in the exact amount
488 run. A minimum of startup=N will begin during startup and reconfigure.
489 Squid will start more in groups of up to idle=N in an attempt to meet
490 traffic needs and to keep idle=N free above those traffic needs up to
493 The concurrency= option sets the number of concurrent requests the
494 helper can process. The default of 0 is used for helpers who only
495 supports one request at a time. Setting this to a number greater than
496 0 changes the protocol used to include a channel number first on the
497 request/response line, allowing multiple requests to be sent to the
498 same helper in parallel without waiting for the response.
499 Must not be set unless it's known the helper supports this.
501 auth_param digest children 20 startup=0 idle=1
504 Specifies the realm name which is to be reported to the
505 client for the digest proxy authentication scheme (part of
506 the text the user will see when prompted their username and
507 password). There is no default.
508 auth_param digest realm Squid proxy-caching web server
510 "nonce_garbage_interval" timeinterval
511 Specifies the interval that nonces that have been issued
512 to client_agent's are checked for validity.
514 "nonce_max_duration" timeinterval
515 Specifies the maximum length of time a given nonce will be
518 "nonce_max_count" number
519 Specifies the maximum number of times a given nonce can be
522 "nonce_strictness" on|off
523 Determines if squid requires strict increment-by-1 behavior
524 for nonce counts, or just incrementing (off - for use when
525 user agents generate nonce counts that occasionally miss 1
526 (ie, 1,2,4,6)). Default off.
528 "check_nonce_count" on|off
529 This directive if set to off can disable the nonce count check
530 completely to work around buggy digest qop implementations in
531 certain mainstream browser versions. Default on to check the
532 nonce count to protect from authentication replay attacks.
534 "post_workaround" on|off
535 This is a workaround to certain buggy browsers who sends
536 an incorrect request digest in POST requests when reusing
537 the same nonce as acquired earlier on a GET request.
539 === NTLM scheme options follow ===
542 Specify the command for the external NTLM authenticator.
543 Such a program reads exchanged NTLMSSP packets with
544 the browser via Squid until authentication is completed.
545 If you use an NTLM authenticator, make sure you have 1 acl
546 of type proxy_auth. By default, the NTLM authenticator program
549 auth_param ntlm program /usr/bin/ntlm_auth
551 "children" numberofchildren [startup=N] [idle=N]
552 The maximum number of authenticator processes to spawn (default 5).
553 If you start too few Squid will have to wait for them to
554 process a backlog of credential verifications, slowing it
555 down. When credential verifications are done via a (slow)
556 network you are likely to need lots of authenticator
559 The startup= and idle= options permit some skew in the exact amount
560 run. A minimum of startup=N will begin during startup and reconfigure.
561 Squid will start more in groups of up to idle=N in an attempt to meet
562 traffic needs and to keep idle=N free above those traffic needs up to
565 auth_param ntlm children 20 startup=0 idle=1
568 If you experience problems with PUT/POST requests when using the
569 Negotiate authentication scheme then you can try setting this to
570 off. This will cause Squid to forcibly close the connection on
571 the initial requests where the browser asks which schemes are
572 supported by the proxy.
574 auth_param ntlm keep_alive on
576 === Options for configuring the NEGOTIATE auth-scheme follow ===
579 Specify the command for the external Negotiate authenticator.
580 This protocol is used in Microsoft Active-Directory enabled setups with
581 the Microsoft Internet Explorer or Mozilla Firefox browsers.
582 Its main purpose is to exchange credentials with the Squid proxy
583 using the Kerberos mechanisms.
584 If you use a Negotiate authenticator, make sure you have at least
585 one acl of type proxy_auth active. By default, the negotiate
586 authenticator program is not used.
587 The only supported program for this role is the ntlm_auth
588 program distributed as part of Samba, version 4 or later.
590 auth_param negotiate program /usr/bin/ntlm_auth --helper-protocol=gss-spnego
592 "children" numberofchildren [startup=N] [idle=N]
593 The maximum number of authenticator processes to spawn (default 5).
594 If you start too few Squid will have to wait for them to
595 process a backlog of credential verifications, slowing it
596 down. When credential verifications are done via a (slow)
597 network you are likely to need lots of authenticator
600 The startup= and idle= options permit some skew in the exact amount
601 run. A minimum of startup=N will begin during startup and reconfigure.
602 Squid will start more in groups of up to idle=N in an attempt to meet
603 traffic needs and to keep idle=N free above those traffic needs up to
606 auth_param negotiate children 20 startup=0 idle=1
609 If you experience problems with PUT/POST requests when using the
610 Negotiate authentication scheme then you can try setting this to
611 off. This will cause Squid to forcibly close the connection on
612 the initial requests where the browser asks which schemes are
613 supported by the proxy.
615 auth_param negotiate keep_alive on
619 #Recommended minimum configuration per scheme:
620 #auth_param negotiate program <uncomment and complete this line to activate>
621 #auth_param negotiate children 20 startup=0 idle=1
622 #auth_param negotiate keep_alive on
624 #auth_param ntlm program <uncomment and complete this line to activate>
625 #auth_param ntlm children 20 startup=0 idle=1
626 #auth_param ntlm keep_alive on
628 #auth_param digest program <uncomment and complete this line>
629 #auth_param digest children 20 startup=0 idle=1
630 #auth_param digest realm Squid proxy-caching web server
631 #auth_param digest nonce_garbage_interval 5 minutes
632 #auth_param digest nonce_max_duration 30 minutes
633 #auth_param digest nonce_max_count 50
635 #auth_param basic program <uncomment and complete this line>
636 #auth_param basic children 5 startup=5 idle=1
637 #auth_param basic realm Squid proxy-caching web server
638 #auth_param basic credentialsttl 2 hours
641 NAME: authenticate_cache_garbage_interval
644 LOC: Config.authenticateGCInterval
646 The time period between garbage collection across the username cache.
647 This is a trade-off between memory utilization (long intervals - say
648 2 days) and CPU (short intervals - say 1 minute). Only change if you
652 NAME: authenticate_ttl
655 LOC: Config.authenticateTTL
657 The time a user & their credentials stay in the logged in
658 user cache since their last request. When the garbage
659 interval passes, all user credentials that have passed their
660 TTL are removed from memory.
663 NAME: authenticate_ip_ttl
665 LOC: Config.authenticateIpTTL
668 If you use proxy authentication and the 'max_user_ip' ACL,
669 this directive controls how long Squid remembers the IP
670 addresses associated with each user. Use a small value
671 (e.g., 60 seconds) if your users might change addresses
672 quickly, as is the case with dialup. You might be safe
673 using a larger value (e.g., 2 hours) in a corporate LAN
674 environment with relatively static address assignments.
679 -----------------------------------------------------------------------------
682 NAME: external_acl_type
683 TYPE: externalAclHelper
684 LOC: Config.externalAclHelperList
687 This option defines external acl classes using a helper program
688 to look up the status
690 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
694 ttl=n TTL in seconds for cached results (defaults to 3600
697 TTL for cached negative lookups (default same
700 Maximum number of acl helper processes spawned to service
701 external acl lookups of this type. (default 20)
703 Minimum number of acl helper processes to spawn during
704 startup and reconfigure to service external acl lookups
705 of this type. (default 0)
707 Number of acl helper processes to keep ahead of traffic
708 loads. Squid will spawn this many at once whenever load
709 rises above the capabilities of existing processes.
710 Up to the value of children-max. (default 1)
711 concurrency=n concurrency level per process. Only used with helpers
712 capable of processing more than one query at a time.
713 cache=n limit the result cache size, default is unbounded.
714 grace=n Percentage remaining of TTL where a refresh of a
715 cached entry should be initiated without needing to
716 wait for a new reply. (default is for no grace period)
717 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
718 ipv4 / ipv6 IP protocol used to communicate with this helper.
719 The default is to auto-detect IPv6 and use it when available.
721 FORMAT specifications
723 %LOGIN Authenticated user login name
724 %EXT_USER Username from previous external acl
725 %EXT_LOG Log details from previous external acl
726 %EXT_TAG Tag from previous external acl
727 %IDENT Ident user name
729 %SRCPORT Client source port
732 %PROTO Requested URL scheme
734 %PATH Requested URL path
735 %METHOD Request method
736 %MYADDR Squid interface address
737 %MYPORT Squid http_port number
738 %PATH Requested URL-path (including query-string if any)
739 %USER_CERT SSL User certificate in PEM format
740 %USER_CERTCHAIN SSL User certificate chain in PEM format
741 %USER_CERT_xx SSL User certificate subject attribute xx
742 %USER_CA_xx SSL User certificate issuer attribute xx
744 %>{Header} HTTP request header "Header"
746 HTTP request header "Hdr" list member "member"
748 HTTP request header list member using ; as
749 list separator. ; can be any non-alphanumeric
752 %<{Header} HTTP reply header "Header"
754 HTTP reply header "Hdr" list member "member"
756 HTTP reply header list member using ; as
757 list separator. ; can be any non-alphanumeric
760 %ACL The name of the ACL being tested.
761 %DATA The ACL arguments. If not used then any arguments
762 is automatically added at the end of the line
764 NOTE: this will encode the arguments as one token,
765 whereas the default will pass each separately.
767 %% The percent sign. Useful for helpers which need
768 an unchanging input format.
771 General request syntax:
773 [channel-ID] FORMAT-values [acl-values ...]
776 FORMAT-values consists of transaction details expanded with
777 whitespace separation per the config file FORMAT specification
778 using the FORMAT macros listed above.
780 acl-values consists of any string specified in the referencing
781 config 'acl ... external' line. see the "acl external" directive.
783 Request values sent to the helper are URL escaped to protect
784 each value in requests against whitespaces.
786 If using protocol=2.5 then the request sent to the helper is not
787 URL escaped to protect against whitespace.
789 NOTE: protocol=3.0 is deprecated as no longer necessary.
791 When using the concurrency= option the protocol is changed by
792 introducing a query channel tag in front of the request/response.
793 The query channel tag is a number between 0 and concurrency-1.
794 This value must be echoed back unchanged to Squid as the first part
795 of the response relating to its request.
798 The helper receives lines expanded per the above format specification
799 and for each input line returns 1 line starting with OK/ERR/BH result
800 code and optionally followed by additional keywords with more details.
803 General result syntax:
805 [channel-ID] result keyword=value ...
807 Result consists of one of the codes:
810 the ACL test produced a match.
813 the ACL test does not produce a match.
816 An internal error occurred in the helper, preventing
817 a result being identified.
819 The meaning of 'a match' is determined by your squid.conf
820 access control configuration. See the Squid wiki for details.
824 user= The users name (login)
826 password= The users password (for login= cache_peer option)
828 message= Message describing the reason for this response.
829 Available as %o in error pages.
830 Useful on (ERR and BH results).
832 tag= Apply a tag to a request. Only sets a tag once,
833 does not alter existing tags.
835 log= String to be logged in access.log. Available as
836 %ea in logformat specifications.
838 Any keywords may be sent on any response whether OK, ERR or BH.
840 All response keyword values need to be a single token with URL
841 escaping, or enclosed in double quotes (") and escaped using \ on
842 any double quotes or \ characters within the value. The wrapping
843 double quotes are removed before the value is interpreted by Squid.
844 \r and \n are also replace by CR and LF.
846 Some example key values:
850 user="J. \"Bob\" Smith"
857 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
858 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
859 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
860 DEFAULT: ssl::certUntrusted ssl_error X509_V_ERR_INVALID_CA X509_V_ERR_SELF_SIGNED_CERT_IN_CHAIN X509_V_ERR_UNABLE_TO_VERIFY_LEAF_SIGNATURE X509_V_ERR_UNABLE_TO_GET_ISSUER_CERT X509_V_ERR_UNABLE_TO_GET_ISSUER_CERT_LOCALLY X509_V_ERR_CERT_UNTRUSTED
861 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
864 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
865 DEFAULT: localhost src 127.0.0.1/32 ::1
866 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
867 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
869 Defining an Access List
871 Every access list definition must begin with an aclname and acltype,
872 followed by either type-specific arguments or a quoted filename that
875 acl aclname acltype argument ...
876 acl aclname acltype "file" ...
878 When using "file", the file should contain one item per line.
880 Some acl types supports options which changes their default behaviour.
881 The available options are:
883 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
884 case-insensitive, use the -i option. To return case-sensitive
885 use the +i option between patterns, or make a new ACL line
888 -n Disable lookups and address type conversions. If lookup or
889 conversion is required because the parameter type (IP or
890 domain name) does not match the message address type (domain
891 name or IP), then the ACL would immediately declare a mismatch
892 without any warnings or lookups.
894 -- Used to stop processing all options, in the case the first acl
895 value has '-' character as first character (for example the '-'
896 is a valid domain name)
898 Some acl types require suspending the current request in order
899 to access some external data source.
900 Those which do are marked with the tag [slow], those which
901 don't are marked as [fast].
902 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
903 for further information
905 ***** ACL TYPES AVAILABLE *****
907 acl aclname src ip-address/mask ... # clients IP address [fast]
908 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
909 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
910 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
912 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
913 # The arp ACL requires the special configure option --enable-arp-acl.
914 # Furthermore, the ARP ACL code is not portable to all operating systems.
915 # It works on Linux, Solaris, Windows, FreeBSD, and some
916 # other *BSD variants.
919 # NOTE: Squid can only determine the MAC address for clients that are on
920 # the same subnet. If the client is on a different subnet,
921 # then Squid cannot find out its MAC address.
923 acl aclname srcdomain .foo.com ...
924 # reverse lookup, from client IP [slow]
925 acl aclname dstdomain [-n] .foo.com ...
926 # Destination server from URL [fast]
927 acl aclname srcdom_regex [-i] \.foo\.com ...
928 # regex matching client name [slow]
929 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
930 # regex matching server [fast]
932 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
933 # based URL is used and no match is found. The name "none" is used
934 # if the reverse lookup fails.
936 acl aclname src_as number ...
937 acl aclname dst_as number ...
939 # Except for access control, AS numbers can be used for
940 # routing of requests to specific caches. Here's an
941 # example for routing all requests for AS#1241 and only
942 # those to mycache.mydomain.net:
943 # acl asexample dst_as 1241
944 # cache_peer_access mycache.mydomain.net allow asexample
945 # cache_peer_access mycache_mydomain.net deny all
947 acl aclname peername myPeer ...
949 # match against a named cache_peer entry
950 # set unique name= on cache_peer lines for reliable use.
952 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
962 # h1:m1 must be less than h2:m2
964 acl aclname url_regex [-i] ^http:// ...
965 # regex matching on whole URL [fast]
966 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
967 # regex matching on URL login field
968 acl aclname urlpath_regex [-i] \.gif$ ...
969 # regex matching on URL path [fast]
971 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
973 acl aclname localport 3128 ... # TCP port the client connected to [fast]
974 # NP: for interception mode this is usually '80'
976 acl aclname myportname 3128 ... # http(s)_port name [fast]
978 acl aclname proto HTTP FTP ... # request protocol [fast]
980 acl aclname method GET POST ... # HTTP request method [fast]
982 acl aclname http_status 200 301 500- 400-403 ...
983 # status code in reply [fast]
985 acl aclname browser [-i] regexp ...
986 # pattern match on User-Agent header (see also req_header below) [fast]
988 acl aclname referer_regex [-i] regexp ...
989 # pattern match on Referer header [fast]
990 # Referer is highly unreliable, so use with care
992 acl aclname ident username ...
993 acl aclname ident_regex [-i] pattern ...
994 # string match on ident output [slow]
995 # use REQUIRED to accept any non-null ident.
997 acl aclname proxy_auth [-i] username ...
998 acl aclname proxy_auth_regex [-i] pattern ...
999 # perform http authentication challenge to the client and match against
1000 # supplied credentials [slow]
1002 # takes a list of allowed usernames.
1003 # use REQUIRED to accept any valid username.
1005 # Will use proxy authentication in forward-proxy scenarios, and plain
1006 # http authenticaiton in reverse-proxy scenarios
1008 # NOTE: when a Proxy-Authentication header is sent but it is not
1009 # needed during ACL checking the username is NOT logged
1012 # NOTE: proxy_auth requires a EXTERNAL authentication program
1013 # to check username/password combinations (see
1014 # auth_param directive).
1016 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1017 # as the browser needs to be configured for using a proxy in order
1018 # to respond to proxy authentication.
1020 acl aclname snmp_community string ...
1021 # A community string to limit access to your SNMP Agent [fast]
1024 # acl snmppublic snmp_community public
1026 acl aclname maxconn number
1027 # This will be matched when the client's IP address has
1028 # more than <number> TCP connections established. [fast]
1029 # NOTE: This only measures direct TCP links so X-Forwarded-For
1030 # indirect clients are not counted.
1032 acl aclname max_user_ip [-s] number
1033 # This will be matched when the user attempts to log in from more
1034 # than <number> different ip addresses. The authenticate_ip_ttl
1035 # parameter controls the timeout on the ip entries. [fast]
1036 # If -s is specified the limit is strict, denying browsing
1037 # from any further IP addresses until the ttl has expired. Without
1038 # -s Squid will just annoy the user by "randomly" denying requests.
1039 # (the counter is reset each time the limit is reached and a
1040 # request is denied)
1041 # NOTE: in acceleration mode or where there is mesh of child proxies,
1042 # clients may appear to come from multiple addresses if they are
1043 # going through proxy farms, so a limit of 1 may cause user problems.
1045 acl aclname random probability
1046 # Pseudo-randomly match requests. Based on the probability given.
1047 # Probability may be written as a decimal (0.333), fraction (1/3)
1048 # or ratio of matches:non-matches (3:5).
1050 acl aclname req_mime_type [-i] mime-type ...
1051 # regex match against the mime type of the request generated
1052 # by the client. Can be used to detect file upload or some
1053 # types HTTP tunneling requests [fast]
1054 # NOTE: This does NOT match the reply. You cannot use this
1055 # to match the returned file type.
1057 acl aclname req_header header-name [-i] any\.regex\.here
1058 # regex match against any of the known request headers. May be
1059 # thought of as a superset of "browser", "referer" and "mime-type"
1062 acl aclname rep_mime_type [-i] mime-type ...
1063 # regex match against the mime type of the reply received by
1064 # squid. Can be used to detect file download or some
1065 # types HTTP tunneling requests. [fast]
1066 # NOTE: This has no effect in http_access rules. It only has
1067 # effect in rules that affect the reply data stream such as
1068 # http_reply_access.
1070 acl aclname rep_header header-name [-i] any\.regex\.here
1071 # regex match against any of the known reply headers. May be
1072 # thought of as a superset of "browser", "referer" and "mime-type"
1075 acl aclname external class_name [arguments...]
1076 # external ACL lookup via a helper class defined by the
1077 # external_acl_type directive [slow]
1079 acl aclname user_cert attribute values...
1080 # match against attributes in a user SSL certificate
1081 # attribute is one of DN/C/O/CN/L/ST [fast]
1083 acl aclname ca_cert attribute values...
1084 # match against attributes a users issuing CA SSL certificate
1085 # attribute is one of DN/C/O/CN/L/ST [fast]
1087 acl aclname ext_user username ...
1088 acl aclname ext_user_regex [-i] pattern ...
1089 # string match on username returned by external acl helper [slow]
1090 # use REQUIRED to accept any non-null user name.
1092 acl aclname tag tagvalue ...
1093 # string match on tag returned by external acl helper [slow]
1095 acl aclname hier_code codename ...
1096 # string match against squid hierarchy code(s); [fast]
1097 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1099 # NOTE: This has no effect in http_access rules. It only has
1100 # effect in rules that affect the reply data stream such as
1101 # http_reply_access.
1103 acl aclname note name [value ...]
1104 # match transaction annotation [fast]
1105 # Without values, matches any annotation with a given name.
1106 # With value(s), matches any annotation with a given name that
1107 # also has one of the given values.
1108 # Names and values are compared using a string equality test.
1109 # Annotation sources include note and adaptation_meta directives
1110 # as well as helper and eCAP responses.
1112 acl aclname adaptation_service service ...
1113 # Matches the name of any icap_service, ecap_service,
1114 # adaptation_service_set, or adaptation_service_chain that Squid
1115 # has used (or attempted to use) for the master transaction.
1116 # This ACL must be defined after the corresponding adaptation
1117 # service is named in squid.conf. This ACL is usable with
1118 # adaptation_meta because it starts matching immediately after
1119 # the service has been selected for adaptation.
1122 acl aclname ssl_error errorname
1123 # match against SSL certificate validation error [fast]
1125 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1128 # The following can be used as shortcuts for certificate properties:
1129 # [ssl::]certHasExpired: the "not after" field is in the past
1130 # [ssl::]certNotYetValid: the "not before" field is in the future
1131 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1132 # [ssl::]certSelfSigned: The certificate is self signed.
1133 # [ssl::]certDomainMismatch: The certificate CN domain does not
1134 # match the name the name of the host we are connecting to.
1136 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1137 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1138 # predefined ACLs, just like the 'all' ACL.
1140 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1141 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1143 acl aclname server_cert_fingerprint [-sha1] fingerprint
1144 # match against server SSL certificate fingerprint [fast]
1146 # The fingerprint is the digest of the DER encoded version
1147 # of the whole certificate. The user should use the form: XX:XX:...
1148 # Optional argument specifies the digest algorithm to use.
1149 # The SHA1 digest algorithm is the default and is currently
1150 # the only algorithm supported (-sha1).
1152 acl aclname atstep step
1153 # match against the current step during ssl_bump evaluation [fast]
1154 # Never matches and should not be used outside the ssl_bump context.
1156 # At each SslBump step, Squid evaluates ssl_bump directives to find
1157 # the next bumping action (e.g., peek or splice). Valid SslBump step
1158 # values and the corresponding ssl_bump evaluation moments are:
1159 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1160 # SslBump2: After getting SSL Client Hello info.
1161 # SslBump3: After getting SSL Server Hello info.
1163 acl aclname any-of acl1 acl2 ...
1164 # match any one of the acls [fast or slow]
1165 # The first matching ACL stops further ACL evaluation.
1167 # ACLs from multiple any-of lines with the same name are ORed.
1168 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1169 # acl A any-of a1 a2
1170 # acl A any-of a3 a4
1172 # This group ACL is fast if all evaluated ACLs in the group are fast
1173 # and slow otherwise.
1175 acl aclname all-of acl1 acl2 ...
1176 # match all of the acls [fast or slow]
1177 # The first mismatching ACL stops further ACL evaluation.
1179 # ACLs from multiple all-of lines with the same name are ORed.
1180 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1181 # acl B all-of b1 b2
1182 # acl B all-of b3 b4
1184 # This group ACL is fast if all evaluated ACLs in the group are fast
1185 # and slow otherwise.
1188 acl macaddress arp 09:00:2b:23:45:67
1189 acl myexample dst_as 1241
1190 acl password proxy_auth REQUIRED
1191 acl fileupload req_mime_type -i ^multipart/form-data$
1192 acl javascript rep_mime_type -i ^application/x-javascript$
1196 # Recommended minimum configuration:
1199 # Example rule allowing access from your local networks.
1200 # Adapt to list your (internal) IP networks from where browsing
1202 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1203 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1204 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1205 acl localnet src fc00::/7 # RFC 4193 local private network range
1206 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1208 acl SSL_ports port 443
1209 acl Safe_ports port 80 # http
1210 acl Safe_ports port 21 # ftp
1211 acl Safe_ports port 443 # https
1212 acl Safe_ports port 70 # gopher
1213 acl Safe_ports port 210 # wais
1214 acl Safe_ports port 1025-65535 # unregistered ports
1215 acl Safe_ports port 280 # http-mgmt
1216 acl Safe_ports port 488 # gss-http
1217 acl Safe_ports port 591 # filemaker
1218 acl Safe_ports port 777 # multiling http
1219 acl CONNECT method CONNECT
1223 NAME: follow_x_forwarded_for
1225 IFDEF: FOLLOW_X_FORWARDED_FOR
1226 LOC: Config.accessList.followXFF
1227 DEFAULT_IF_NONE: deny all
1228 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1230 Allowing or Denying the X-Forwarded-For header to be followed to
1231 find the original source of a request.
1233 Requests may pass through a chain of several other proxies
1234 before reaching us. The X-Forwarded-For header will contain a
1235 comma-separated list of the IP addresses in the chain, with the
1236 rightmost address being the most recent.
1238 If a request reaches us from a source that is allowed by this
1239 configuration item, then we consult the X-Forwarded-For header
1240 to see where that host received the request from. If the
1241 X-Forwarded-For header contains multiple addresses, we continue
1242 backtracking until we reach an address for which we are not allowed
1243 to follow the X-Forwarded-For header, or until we reach the first
1244 address in the list. For the purpose of ACL used in the
1245 follow_x_forwarded_for directive the src ACL type always matches
1246 the address we are testing and srcdomain matches its rDNS.
1248 The end result of this process is an IP address that we will
1249 refer to as the indirect client address. This address may
1250 be treated as the client address for access control, ICAP, delay
1251 pools and logging, depending on the acl_uses_indirect_client,
1252 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1253 log_uses_indirect_client and tproxy_uses_indirect_client options.
1255 This clause only supports fast acl types.
1256 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1258 SECURITY CONSIDERATIONS:
1260 Any host for which we follow the X-Forwarded-For header
1261 can place incorrect information in the header, and Squid
1262 will use the incorrect information as if it were the
1263 source address of the request. This may enable remote
1264 hosts to bypass any access control restrictions that are
1265 based on the client's source addresses.
1269 acl localhost src 127.0.0.1
1270 acl my_other_proxy srcdomain .proxy.example.com
1271 follow_x_forwarded_for allow localhost
1272 follow_x_forwarded_for allow my_other_proxy
1275 NAME: acl_uses_indirect_client
1278 IFDEF: FOLLOW_X_FORWARDED_FOR
1280 LOC: Config.onoff.acl_uses_indirect_client
1282 Controls whether the indirect client address
1283 (see follow_x_forwarded_for) is used instead of the
1284 direct client address in acl matching.
1286 NOTE: maxconn ACL considers direct TCP links and indirect
1287 clients will always have zero. So no match.
1290 NAME: delay_pool_uses_indirect_client
1293 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1295 LOC: Config.onoff.delay_pool_uses_indirect_client
1297 Controls whether the indirect client address
1298 (see follow_x_forwarded_for) is used instead of the
1299 direct client address in delay pools.
1302 NAME: log_uses_indirect_client
1305 IFDEF: FOLLOW_X_FORWARDED_FOR
1307 LOC: Config.onoff.log_uses_indirect_client
1309 Controls whether the indirect client address
1310 (see follow_x_forwarded_for) is used instead of the
1311 direct client address in the access log.
1314 NAME: tproxy_uses_indirect_client
1317 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1319 LOC: Config.onoff.tproxy_uses_indirect_client
1321 Controls whether the indirect client address
1322 (see follow_x_forwarded_for) is used instead of the
1323 direct client address when spoofing the outgoing client.
1325 This has no effect on requests arriving in non-tproxy
1328 SECURITY WARNING: Usage of this option is dangerous
1329 and should not be used trivially. Correct configuration
1330 of follow_x_forewarded_for with a limited set of trusted
1331 sources is required to prevent abuse of your proxy.
1334 NAME: spoof_client_ip
1336 LOC: Config.accessList.spoof_client_ip
1338 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1340 Control client IP address spoofing of TPROXY traffic based on
1341 defined access lists.
1343 spoof_client_ip allow|deny [!]aclname ...
1345 If there are no "spoof_client_ip" lines present, the default
1346 is to "allow" spoofing of any suitable request.
1348 Note that the cache_peer "no-tproxy" option overrides this ACL.
1350 This clause supports fast acl types.
1351 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1356 LOC: Config.accessList.http
1357 DEFAULT_IF_NONE: deny all
1358 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1360 Allowing or Denying access based on defined access lists
1362 Access to the HTTP port:
1363 http_access allow|deny [!]aclname ...
1365 NOTE on default values:
1367 If there are no "access" lines present, the default is to deny
1370 If none of the "access" lines cause a match, the default is the
1371 opposite of the last line in the list. If the last line was
1372 deny, the default is allow. Conversely, if the last line
1373 is allow, the default will be deny. For these reasons, it is a
1374 good idea to have an "deny all" entry at the end of your access
1375 lists to avoid potential confusion.
1377 This clause supports both fast and slow acl types.
1378 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1383 # Recommended minimum Access Permission configuration:
1385 # Deny requests to certain unsafe ports
1386 http_access deny !Safe_ports
1388 # Deny CONNECT to other than secure SSL ports
1389 http_access deny CONNECT !SSL_ports
1391 # Only allow cachemgr access from localhost
1392 http_access allow localhost manager
1393 http_access deny manager
1395 # We strongly recommend the following be uncommented to protect innocent
1396 # web applications running on the proxy server who think the only
1397 # one who can access services on "localhost" is a local user
1398 #http_access deny to_localhost
1401 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1404 # Example rule allowing access from your local networks.
1405 # Adapt localnet in the ACL section to list your (internal) IP networks
1406 # from where browsing should be allowed
1407 http_access allow localnet
1408 http_access allow localhost
1410 # And finally deny all other access to this proxy
1411 http_access deny all
1415 NAME: adapted_http_access http_access2
1417 LOC: Config.accessList.adapted_http
1419 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1421 Allowing or Denying access based on defined access lists
1423 Essentially identical to http_access, but runs after redirectors
1424 and ICAP/eCAP adaptation. Allowing access control based on their
1427 If not set then only http_access is used.
1430 NAME: http_reply_access
1432 LOC: Config.accessList.reply
1434 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1436 Allow replies to client requests. This is complementary to http_access.
1438 http_reply_access allow|deny [!] aclname ...
1440 NOTE: if there are no access lines present, the default is to allow
1443 If none of the access lines cause a match the opposite of the
1444 last line will apply. Thus it is good practice to end the rules
1445 with an "allow all" or "deny all" entry.
1447 This clause supports both fast and slow acl types.
1448 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1453 LOC: Config.accessList.icp
1455 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1457 Allowing or Denying access to the ICP port based on defined
1460 icp_access allow|deny [!]aclname ...
1462 NOTE: The default if no icp_access lines are present is to
1463 deny all traffic. This default may cause problems with peers
1466 This clause only supports fast acl types.
1467 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1469 # Allow ICP queries from local networks only
1470 #icp_access allow localnet
1471 #icp_access deny all
1477 LOC: Config.accessList.htcp
1479 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1481 Allowing or Denying access to the HTCP port based on defined
1484 htcp_access allow|deny [!]aclname ...
1486 See also htcp_clr_access for details on access control for
1487 cache purge (CLR) HTCP messages.
1489 NOTE: The default if no htcp_access lines are present is to
1490 deny all traffic. This default may cause problems with peers
1491 using the htcp option.
1493 This clause only supports fast acl types.
1494 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1496 # Allow HTCP queries from local networks only
1497 #htcp_access allow localnet
1498 #htcp_access deny all
1501 NAME: htcp_clr_access
1504 LOC: Config.accessList.htcp_clr
1506 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1508 Allowing or Denying access to purge content using HTCP based
1509 on defined access lists.
1510 See htcp_access for details on general HTCP access control.
1512 htcp_clr_access allow|deny [!]aclname ...
1514 This clause only supports fast acl types.
1515 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1517 # Allow HTCP CLR requests from trusted peers
1518 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1519 htcp_clr_access allow htcp_clr_peer
1520 htcp_clr_access deny all
1525 LOC: Config.accessList.miss
1527 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1529 Determins whether network access is permitted when satisfying a request.
1532 to force your neighbors to use you as a sibling instead of
1535 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1536 miss_access deny !localclients
1537 miss_access allow all
1539 This means only your local clients are allowed to fetch relayed/MISS
1540 replies from the network and all other clients can only fetch cached
1543 The default for this setting allows all clients who passed the
1544 http_access rules to relay via this proxy.
1546 This clause only supports fast acl types.
1547 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1550 NAME: ident_lookup_access
1554 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1555 LOC: Ident::TheConfig.identLookup
1557 A list of ACL elements which, if matched, cause an ident
1558 (RFC 931) lookup to be performed for this request. For
1559 example, you might choose to always perform ident lookups
1560 for your main multi-user Unix boxes, but not for your Macs
1561 and PCs. By default, ident lookups are not performed for
1564 To enable ident lookups for specific client addresses, you
1565 can follow this example:
1567 acl ident_aware_hosts src 198.168.1.0/24
1568 ident_lookup_access allow ident_aware_hosts
1569 ident_lookup_access deny all
1571 Only src type ACL checks are fully supported. A srcdomain
1572 ACL might work at times, but it will not always provide
1575 This clause only supports fast acl types.
1576 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1579 NAME: reply_body_max_size
1580 COMMENT: size [acl acl...]
1583 DEFAULT_DOC: No limit is applied.
1584 LOC: Config.ReplyBodySize
1586 This option specifies the maximum size of a reply body. It can be
1587 used to prevent users from downloading very large files, such as
1588 MP3's and movies. When the reply headers are received, the
1589 reply_body_max_size lines are processed, and the first line where
1590 all (if any) listed ACLs are true is used as the maximum body size
1593 This size is checked twice. First when we get the reply headers,
1594 we check the content-length value. If the content length value exists
1595 and is larger than the allowed size, the request is denied and the
1596 user receives an error message that says "the request or reply
1597 is too large." If there is no content-length, and the reply
1598 size exceeds this limit, the client's connection is just closed
1599 and they will receive a partial reply.
1601 WARNING: downstream caches probably can not detect a partial reply
1602 if there is no content-length header, so they will cache
1603 partial responses and give them out as hits. You should NOT
1604 use this option if you have downstream caches.
1606 WARNING: A maximum size smaller than the size of squid's error messages
1607 will cause an infinite loop and crash squid. Ensure that the smallest
1608 non-zero value you use is greater that the maximum header size plus
1609 the size of your largest error page.
1611 If you set this parameter none (the default), there will be
1614 Configuration Format is:
1615 reply_body_max_size SIZE UNITS [acl ...]
1617 reply_body_max_size 10 MB
1623 -----------------------------------------------------------------------------
1626 NAME: http_port ascii_port
1629 LOC: Config.Sockaddr.http
1631 Usage: port [mode] [options]
1632 hostname:port [mode] [options]
1633 1.2.3.4:port [mode] [options]
1635 The socket addresses where Squid will listen for HTTP client
1636 requests. You may specify multiple socket addresses.
1637 There are three forms: port alone, hostname with port, and
1638 IP address with port. If you specify a hostname or IP
1639 address, Squid binds the socket to that specific
1640 address. Most likely, you do not need to bind to a specific
1641 address, so you can use the port number alone.
1643 If you are running Squid in accelerator mode, you
1644 probably want to listen on port 80 also, or instead.
1646 The -a command line option may be used to specify additional
1647 port(s) where Squid listens for proxy request. Such ports will
1648 be plain proxy ports with no options.
1650 You may specify multiple socket addresses on multiple lines.
1654 intercept Support for IP-Layer interception of
1655 outgoing requests without browser settings.
1656 NP: disables authentication and IPv6 on the port.
1658 tproxy Support Linux TPROXY for spoofing outgoing
1659 connections using the client IP address.
1660 NP: disables authentication and maybe IPv6 on the port.
1662 accel Accelerator / reverse proxy mode
1664 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1665 establish secure connection with the client and with
1666 the server, decrypt HTTPS messages as they pass through
1667 Squid, and treat them as unencrypted HTTP messages,
1668 becoming the man-in-the-middle.
1670 The ssl_bump option is required to fully enable
1671 bumping of CONNECT requests.
1673 Omitting the mode flag causes default forward proxy mode to be used.
1676 Accelerator Mode Options:
1678 defaultsite=domainname
1679 What to use for the Host: header if it is not present
1680 in a request. Determines what site (not origin server)
1681 accelerators should consider the default.
1683 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1685 protocol= Protocol to reconstruct accelerated and intercepted
1686 requests with. Defaults to HTTP/1.1 for http_port and
1687 HTTPS/1.1 for https_port.
1688 When an unsupported value is configured Squid will
1689 produce a FATAL error.
1690 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1692 vport Virtual host port support. Using the http_port number
1693 instead of the port passed on Host: headers.
1695 vport=NN Virtual host port support. Using the specified port
1696 number instead of the port passed on Host: headers.
1699 Act as if this Squid is the origin server.
1700 This currently means generate new Date: and Expires:
1701 headers on HIT instead of adding Age:.
1703 ignore-cc Ignore request Cache-Control headers.
1705 WARNING: This option violates HTTP specifications if
1706 used in non-accelerator setups.
1708 allow-direct Allow direct forwarding in accelerator mode. Normally
1709 accelerated requests are denied direct forwarding as if
1710 never_direct was used.
1712 WARNING: this option opens accelerator mode to security
1713 vulnerabilities usually only affecting in interception
1714 mode. Make sure to protect forwarding with suitable
1715 http_access rules when using this.
1718 SSL Bump Mode Options:
1719 In addition to these options ssl-bump requires TLS/SSL options.
1721 generate-host-certificates[=<on|off>]
1722 Dynamically create SSL server certificates for the
1723 destination hosts of bumped CONNECT requests.When
1724 enabled, the cert and key options are used to sign
1725 generated certificates. Otherwise generated
1726 certificate will be selfsigned.
1727 If there is a CA certificate lifetime of the generated
1728 certificate equals lifetime of the CA certificate. If
1729 generated certificate is selfsigned lifetime is three
1731 This option is enabled by default when ssl-bump is used.
1732 See the ssl-bump option above for more information.
1734 dynamic_cert_mem_cache_size=SIZE
1735 Approximate total RAM size spent on cached generated
1736 certificates. If set to zero, caching is disabled. The
1737 default value is 4MB.
1741 cert= Path to SSL certificate (PEM format).
1743 key= Path to SSL private key file (PEM format)
1744 if not specified, the certificate file is
1745 assumed to be a combined certificate and
1748 version= The version of SSL/TLS supported
1749 1 automatic (default)
1756 cipher= Colon separated list of supported ciphers.
1757 NOTE: some ciphers such as EDH ciphers depend on
1758 additional settings. If those settings are
1759 omitted the ciphers may be silently ignored
1760 by the OpenSSL library.
1762 options= Various SSL implementation options. The most important
1764 NO_SSLv2 Disallow the use of SSLv2
1765 NO_SSLv3 Disallow the use of SSLv3
1766 NO_TLSv1 Disallow the use of TLSv1.0
1767 NO_TLSv1_1 Disallow the use of TLSv1.1
1768 NO_TLSv1_2 Disallow the use of TLSv1.2
1769 SINGLE_DH_USE Always create a new key when using
1770 temporary/ephemeral DH key exchanges
1771 ALL Enable various bug workarounds
1772 suggested as "harmless" by OpenSSL
1773 Be warned that this reduces SSL/TLS
1774 strength to some attacks.
1775 See OpenSSL SSL_CTX_set_options documentation for a
1776 complete list of options.
1778 clientca= File containing the list of CAs to use when
1779 requesting a client certificate.
1781 cafile= File containing additional CA certificates to
1782 use when verifying client certificates. If unset
1783 clientca will be used.
1785 capath= Directory containing additional CA certificates
1786 and CRL lists to use when verifying client certificates.
1788 crlfile= File of additional CRL lists to use when verifying
1789 the client certificate, in addition to CRLs stored in
1790 the capath. Implies VERIFY_CRL flag below.
1792 dhparams= File containing DH parameters for temporary/ephemeral
1793 DH key exchanges. See OpenSSL documentation for details
1794 on how to create this file.
1795 WARNING: EDH ciphers will be silently disabled if this
1798 sslflags= Various flags modifying the use of SSL:
1800 Don't request client certificates
1801 immediately, but wait until acl processing
1802 requires a certificate (not yet implemented).
1804 Don't use the default CA lists built in
1807 Don't allow for session reuse. Each connection
1808 will result in a new SSL session.
1810 Verify CRL lists when accepting client
1813 Verify CRL lists for all certificates in the
1814 client certificate chain.
1816 sslcontext= SSL session ID context identifier.
1820 connection-auth[=on|off]
1821 use connection-auth=off to tell Squid to prevent
1822 forwarding Microsoft connection oriented authentication
1823 (NTLM, Negotiate and Kerberos)
1825 disable-pmtu-discovery=
1826 Control Path-MTU discovery usage:
1827 off lets OS decide on what to do (default).
1828 transparent disable PMTU discovery when transparent
1830 always disable always PMTU discovery.
1832 In many setups of transparently intercepting proxies
1833 Path-MTU discovery can not work on traffic towards the
1834 clients. This is the case when the intercepting device
1835 does not fully track connections and fails to forward
1836 ICMP must fragment messages to the cache server. If you
1837 have such setup and experience that certain clients
1838 sporadically hang or never complete requests set
1839 disable-pmtu-discovery option to 'transparent'.
1841 name= Specifies a internal name for the port. Defaults to
1842 the port specification (port or addr:port)
1844 tcpkeepalive[=idle,interval,timeout]
1845 Enable TCP keepalive probes of idle connections.
1846 In seconds; idle is the initial time before TCP starts
1847 probing the connection, interval how often to probe, and
1848 timeout the time before giving up.
1850 If you run Squid on a dual-homed machine with an internal
1851 and an external interface we recommend you to specify the
1852 internal address:port in http_port. This way Squid will only be
1853 visible on the internal address.
1857 # Squid normally listens to port 3128
1858 http_port @DEFAULT_HTTP_PORT@
1866 LOC: Config.Sockaddr.https
1868 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1870 The socket address where Squid will listen for client requests made
1871 over TLS or SSL connections. Commonly referred to as HTTPS.
1873 This is most useful for situations where you are running squid in
1874 accelerator mode and you want to do the SSL work at the accelerator level.
1876 You may specify multiple socket addresses on multiple lines,
1877 each with their own SSL certificate and/or options.
1881 accel Accelerator / reverse proxy mode
1883 intercept Support for IP-Layer interception of
1884 outgoing requests without browser settings.
1885 NP: disables authentication and IPv6 on the port.
1887 tproxy Support Linux TPROXY for spoofing outgoing
1888 connections using the client IP address.
1889 NP: disables authentication and maybe IPv6 on the port.
1891 ssl-bump For each intercepted connection allowed by ssl_bump
1892 ACLs, establish a secure connection with the client and with
1893 the server, decrypt HTTPS messages as they pass through
1894 Squid, and treat them as unencrypted HTTP messages,
1895 becoming the man-in-the-middle.
1897 An "ssl_bump server-first" match is required to
1898 fully enable bumping of intercepted SSL connections.
1900 Requires tproxy or intercept.
1902 Omitting the mode flag causes default forward proxy mode to be used.
1905 See http_port for a list of generic options
1910 cert= Path to SSL certificate (PEM format).
1912 key= Path to SSL private key file (PEM format)
1913 if not specified, the certificate file is
1914 assumed to be a combined certificate and
1917 version= The version of SSL/TLS supported
1918 1 automatic (default)
1923 cipher= Colon separated list of supported ciphers.
1925 options= Various SSL engine options. The most important
1927 NO_SSLv2 Disallow the use of SSLv2
1928 NO_SSLv3 Disallow the use of SSLv3
1929 NO_TLSv1 Disallow the use of TLSv1
1930 SINGLE_DH_USE Always create a new key when using
1931 temporary/ephemeral DH key exchanges
1932 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1933 documentation for a complete list of options.
1935 clientca= File containing the list of CAs to use when
1936 requesting a client certificate.
1938 cafile= File containing additional CA certificates to
1939 use when verifying client certificates. If unset
1940 clientca will be used.
1942 capath= Directory containing additional CA certificates
1943 and CRL lists to use when verifying client certificates.
1945 crlfile= File of additional CRL lists to use when verifying
1946 the client certificate, in addition to CRLs stored in
1947 the capath. Implies VERIFY_CRL flag below.
1949 dhparams= File containing DH parameters for temporary/ephemeral
1952 sslflags= Various flags modifying the use of SSL:
1954 Don't request client certificates
1955 immediately, but wait until acl processing
1956 requires a certificate (not yet implemented).
1958 Don't use the default CA lists built in
1961 Don't allow for session reuse. Each connection
1962 will result in a new SSL session.
1964 Verify CRL lists when accepting client
1967 Verify CRL lists for all certificates in the
1968 client certificate chain.
1970 sslcontext= SSL session ID context identifier.
1972 generate-host-certificates[=<on|off>]
1973 Dynamically create SSL server certificates for the
1974 destination hosts of bumped SSL requests.When
1975 enabled, the cert and key options are used to sign
1976 generated certificates. Otherwise generated
1977 certificate will be selfsigned.
1978 If there is CA certificate life time of generated
1979 certificate equals lifetime of CA certificate. If
1980 generated certificate is selfsigned lifetime is three
1982 This option is enabled by default when SslBump is used.
1983 See the sslBump option above for more information.
1985 dynamic_cert_mem_cache_size=SIZE
1986 Approximate total RAM size spent on cached generated
1987 certificates. If set to zero, caching is disabled. The
1988 default value is 4MB.
1990 See http_port for a list of available options.
1993 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1996 LOC: Ip::Qos::TheConfig.tosToServer
1998 Allows you to select a TOS/Diffserv value for packets outgoing
1999 on the server side, based on an ACL.
2001 tcp_outgoing_tos ds-field [!]aclname ...
2003 Example where normal_service_net uses the TOS value 0x00
2004 and good_service_net uses 0x20
2006 acl normal_service_net src 10.0.0.0/24
2007 acl good_service_net src 10.0.1.0/24
2008 tcp_outgoing_tos 0x00 normal_service_net
2009 tcp_outgoing_tos 0x20 good_service_net
2011 TOS/DSCP values really only have local significance - so you should
2012 know what you're specifying. For more information, see RFC2474,
2013 RFC2475, and RFC3260.
2015 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2016 "default" to use whatever default your host has. Note that in
2017 practice often only multiples of 4 is usable as the two rightmost bits
2018 have been redefined for use by ECN (RFC 3168 section 23.1).
2020 Processing proceeds in the order specified, and stops at first fully
2024 NAME: clientside_tos
2027 LOC: Ip::Qos::TheConfig.tosToClient
2029 Allows you to select a TOS/Diffserv value for packets being transmitted
2030 on the client-side, based on an ACL.
2032 clientside_tos ds-field [!]aclname ...
2034 Example where normal_service_net uses the TOS value 0x00
2035 and good_service_net uses 0x20
2037 acl normal_service_net src 10.0.0.0/24
2038 acl good_service_net src 10.0.1.0/24
2039 clientside_tos 0x00 normal_service_net
2040 clientside_tos 0x20 good_service_net
2042 Note: This feature is incompatible with qos_flows. Any TOS values set here
2043 will be overwritten by TOS values in qos_flows.
2046 NAME: tcp_outgoing_mark
2048 IFDEF: SO_MARK&&USE_LIBCAP
2050 LOC: Ip::Qos::TheConfig.nfmarkToServer
2052 Allows you to apply a Netfilter mark value to outgoing packets
2053 on the server side, based on an ACL.
2055 tcp_outgoing_mark mark-value [!]aclname ...
2057 Example where normal_service_net uses the mark value 0x00
2058 and good_service_net uses 0x20
2060 acl normal_service_net src 10.0.0.0/24
2061 acl good_service_net src 10.0.1.0/24
2062 tcp_outgoing_mark 0x00 normal_service_net
2063 tcp_outgoing_mark 0x20 good_service_net
2066 NAME: clientside_mark
2068 IFDEF: SO_MARK&&USE_LIBCAP
2070 LOC: Ip::Qos::TheConfig.nfmarkToClient
2072 Allows you to apply a Netfilter mark value to packets being transmitted
2073 on the client-side, based on an ACL.
2075 clientside_mark mark-value [!]aclname ...
2077 Example where normal_service_net uses the mark value 0x00
2078 and good_service_net uses 0x20
2080 acl normal_service_net src 10.0.0.0/24
2081 acl good_service_net src 10.0.1.0/24
2082 clientside_mark 0x00 normal_service_net
2083 clientside_mark 0x20 good_service_net
2085 Note: This feature is incompatible with qos_flows. Any mark values set here
2086 will be overwritten by mark values in qos_flows.
2093 LOC: Ip::Qos::TheConfig
2095 Allows you to select a TOS/DSCP value to mark outgoing
2096 connections to the client, based on where the reply was sourced.
2097 For platforms using netfilter, allows you to set a netfilter mark
2098 value instead of, or in addition to, a TOS value.
2100 By default this functionality is disabled. To enable it with the default
2101 settings simply use "qos_flows mark" or "qos_flows tos". Default
2102 settings will result in the netfilter mark or TOS value being copied
2103 from the upstream connection to the client. Note that it is the connection
2104 CONNMARK value not the packet MARK value that is copied.
2106 It is not currently possible to copy the mark or TOS value from the
2107 client to the upstream connection request.
2109 TOS values really only have local significance - so you should
2110 know what you're specifying. For more information, see RFC2474,
2111 RFC2475, and RFC3260.
2113 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
2114 in practice often only multiples of 4 is usable as the two rightmost bits
2115 have been redefined for use by ECN (RFC 3168 section 23.1).
2117 Mark values can be any unsigned 32-bit integer value.
2119 This setting is configured by setting the following values:
2121 tos|mark Whether to set TOS or netfilter mark values
2123 local-hit=0xFF Value to mark local cache hits.
2125 sibling-hit=0xFF Value to mark hits from sibling peers.
2127 parent-hit=0xFF Value to mark hits from parent peers.
2129 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2130 over the preserve-miss feature (see below), unless
2131 mask is specified, in which case only the bits
2132 specified in the mask are written.
2134 The TOS variant of the following features are only possible on Linux
2135 and require your kernel to be patched with the TOS preserving ZPH
2136 patch, available from http://zph.bratcheda.org
2137 No patch is needed to preserve the netfilter mark, which will work
2138 with all variants of netfilter.
2140 disable-preserve-miss
2141 This option disables the preservation of the TOS or netfilter
2142 mark. By default, the existing TOS or netfilter mark value of
2143 the response coming from the remote server will be retained
2144 and masked with miss-mark.
2145 NOTE: in the case of a netfilter mark, the mark must be set on
2146 the connection (using the CONNMARK target) not on the packet
2150 Allows you to mask certain bits in the TOS or mark value
2151 received from the remote server, before copying the value to
2152 the TOS sent towards clients.
2153 Default for tos: 0xFF (TOS from server is not changed).
2154 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2156 All of these features require the --enable-zph-qos compilation flag
2157 (enabled by default). Netfilter marking also requires the
2158 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2159 libcap 2.09+ (--with-libcap).
2163 NAME: tcp_outgoing_address
2166 DEFAULT_DOC: Address selection is performed by the operating system.
2167 LOC: Config.accessList.outgoing_address
2169 Allows you to map requests to different outgoing IP addresses
2170 based on the username or source address of the user making
2173 tcp_outgoing_address ipaddr [[!]aclname] ...
2176 Forwarding clients with dedicated IPs for certain subnets.
2178 acl normal_service_net src 10.0.0.0/24
2179 acl good_service_net src 10.0.2.0/24
2181 tcp_outgoing_address 2001:db8::c001 good_service_net
2182 tcp_outgoing_address 10.1.0.2 good_service_net
2184 tcp_outgoing_address 2001:db8::beef normal_service_net
2185 tcp_outgoing_address 10.1.0.1 normal_service_net
2187 tcp_outgoing_address 2001:db8::1
2188 tcp_outgoing_address 10.1.0.3
2190 Processing proceeds in the order specified, and stops at first fully
2193 Squid will add an implicit IP version test to each line.
2194 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2195 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2198 NOTE: The use of this directive using client dependent ACLs is
2199 incompatible with the use of server side persistent connections. To
2200 ensure correct results it is best to set server_persistent_connections
2201 to off when using this directive in such configurations.
2203 NOTE: The use of this directive to set a local IP on outgoing TCP links
2204 is incompatible with using TPROXY to set client IP out outbound TCP links.
2205 When needing to contact peers use the no-tproxy cache_peer option and the
2206 client_dst_passthru directive re-enable normal forwarding such as this.
2210 NAME: host_verify_strict
2213 LOC: Config.onoff.hostStrictVerify
2215 Regardless of this option setting, when dealing with intercepted
2216 traffic, Squid always verifies that the destination IP address matches
2217 the Host header domain or IP (called 'authority form URL').
2219 This enforcement is performed to satisfy a MUST-level requirement in
2220 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2221 authority of the origin server or gateway given by the original URL".
2224 Squid always responds with an HTTP 409 (Conflict) error
2225 page and logs a security warning if there is no match.
2227 Squid verifies that the destination IP address matches
2228 the Host header for forward-proxy and reverse-proxy traffic
2229 as well. For those traffic types, Squid also enables the
2230 following checks, comparing the corresponding Host header
2231 and Request-URI components:
2233 * The host names (domain or IP) must be identical,
2234 but valueless or missing Host header disables all checks.
2235 For the two host names to match, both must be either IP
2238 * Port numbers must be identical, but if a port is missing
2239 the scheme-default port is assumed.
2242 When set to OFF (the default):
2243 Squid allows suspicious requests to continue but logs a
2244 security warning and blocks caching of the response.
2246 * Forward-proxy traffic is not checked at all.
2248 * Reverse-proxy traffic is not checked at all.
2250 * Intercepted traffic which passes verification is handled
2251 according to client_dst_passthru.
2253 * Intercepted requests which fail verification are sent
2254 to the client original destination instead of DIRECT.
2255 This overrides 'client_dst_passthru off'.
2257 For now suspicious intercepted CONNECT requests are always
2258 responded to with an HTTP 409 (Conflict) error page.
2263 As described in CVE-2009-0801 when the Host: header alone is used
2264 to determine the destination of a request it becomes trivial for
2265 malicious scripts on remote websites to bypass browser same-origin
2266 security policy and sandboxing protections.
2268 The cause of this is that such applets are allowed to perform their
2269 own HTTP stack, in which case the same-origin policy of the browser
2270 sandbox only verifies that the applet tries to contact the same IP
2271 as from where it was loaded at the IP level. The Host: header may
2272 be different from the connected IP and approved origin.
2276 NAME: client_dst_passthru
2279 LOC: Config.onoff.client_dst_passthru
2281 With NAT or TPROXY intercepted traffic Squid may pass the request
2282 directly to the original client destination IP or seek a faster
2283 source using the HTTP Host header.
2285 Using Host to locate alternative servers can provide faster
2286 connectivity with a range of failure recovery options.
2287 But can also lead to connectivity trouble when the client and
2288 server are attempting stateful interactions unaware of the proxy.
2290 This option (on by default) prevents alternative DNS entries being
2291 located to send intercepted traffic DIRECT to an origin server.
2292 The clients original destination IP and port will be used instead.
2294 Regardless of this option setting, when dealing with intercepted
2295 traffic Squid will verify the Host: header and any traffic which
2296 fails Host verification will be treated as if this option were ON.
2298 see host_verify_strict for details on the verification process.
2303 -----------------------------------------------------------------------------
2306 NAME: ssl_unclean_shutdown
2310 LOC: Config.SSL.unclean_shutdown
2312 Some browsers (especially MSIE) bugs out on SSL shutdown
2319 LOC: Config.SSL.ssl_engine
2322 The OpenSSL engine to use. You will need to set this if you
2323 would like to use hardware SSL acceleration for example.
2326 NAME: sslproxy_client_certificate
2329 LOC: Config.ssl_client.cert
2332 Client SSL Certificate to use when proxying https:// URLs
2335 NAME: sslproxy_client_key
2338 LOC: Config.ssl_client.key
2341 Client SSL Key to use when proxying https:// URLs
2344 NAME: sslproxy_version
2347 DEFAULT_DOC: automatic SSL/TLS version negotiation
2348 LOC: Config.ssl_client.version
2351 SSL version level to use when proxying https:// URLs
2353 The versions of SSL/TLS supported:
2355 1 automatic (default)
2363 NAME: sslproxy_options
2366 LOC: Config.ssl_client.options
2369 SSL implementation options to use when proxying https:// URLs
2371 The most important being:
2373 NO_SSLv2 Disallow the use of SSLv2
2374 NO_SSLv3 Disallow the use of SSLv3
2375 NO_TLSv1 Disallow the use of TLSv1.0
2376 NO_TLSv1_1 Disallow the use of TLSv1.1
2377 NO_TLSv1_2 Disallow the use of TLSv1.2
2379 Always create a new key when using temporary/ephemeral
2382 Disable use of RFC5077 session tickets. Some servers
2383 may have problems understanding the TLS extension due
2384 to ambiguous specification in RFC4507.
2385 ALL Enable various bug workarounds suggested as "harmless"
2386 by OpenSSL. Be warned that this may reduce SSL/TLS
2387 strength to some attacks.
2389 See the OpenSSL SSL_CTX_set_options documentation for a
2390 complete list of possible options.
2393 NAME: sslproxy_cipher
2396 LOC: Config.ssl_client.cipher
2399 SSL cipher list to use when proxying https:// URLs
2401 Colon separated list of supported ciphers.
2404 NAME: sslproxy_cafile
2407 LOC: Config.ssl_client.cafile
2410 file containing CA certificates to use when verifying server
2411 certificates while proxying https:// URLs
2414 NAME: sslproxy_capath
2417 LOC: Config.ssl_client.capath
2420 directory containing CA certificates to use when verifying
2421 server certificates while proxying https:// URLs
2424 NAME: sslproxy_session_ttl
2427 LOC: Config.SSL.session_ttl
2430 Sets the timeout value for SSL sessions
2433 NAME: sslproxy_session_cache_size
2436 LOC: Config.SSL.sessionCacheSize
2439 Sets the cache size to use for ssl session
2444 TYPE: sslproxy_ssl_bump
2445 LOC: Config.accessList.ssl_bump
2446 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
2449 This option is consulted when a CONNECT request is received on
2450 an http_port (or a new connection is intercepted at an
2451 https_port), provided that port was configured with an ssl-bump
2452 flag. The subsequent data on the connection is either treated as
2453 HTTPS and decrypted OR tunneled at TCP level without decryption,
2454 depending on the first matching bumping "action".
2456 ssl_bump <action> [!]acl ...
2458 The following bumping actions are currently supported:
2461 Become a TCP tunnel without decrypting proxied traffic.
2462 This is the default action.
2465 Establish a secure connection with the server and, using a
2466 mimicked server certificate, with the client.
2469 Receive client (step SslBump1) or server (step SslBump2)
2470 certificate while preserving the possibility of splicing the
2471 connection. Peeking at the server certificate (during step 2)
2472 usually precludes bumping of the connection at step 3.
2475 Receive client (step SslBump1) or server (step SslBump2)
2476 certificate while preserving the possibility of bumping the
2477 connection. Staring at the server certificate (during step 2)
2478 usually precludes splicing of the connection at step 3.
2481 Close client and server connections.
2483 Backward compatibility actions available at step SslBump1:
2486 Bump the connection. Establish a secure connection with the
2487 client first, then connect to the server. This old mode does
2488 not allow Squid to mimic server SSL certificate and does not
2489 work with intercepted SSL connections.
2492 Bump the connection. Establish a secure connection with the
2493 server first, then establish a secure connection with the
2494 client, using a mimicked server certificate. Works with both
2495 CONNECT requests and intercepted SSL connections, but does
2496 not allow to make decisions based on SSL handshake info.
2499 Decide whether to bump or splice the connection based on
2500 client-to-squid and server-to-squid SSL hello messages.
2504 Same as the "splice" action.
2506 All ssl_bump rules are evaluated at each of the supported bumping
2507 steps. Rules with actions that are impossible at the current step are
2508 ignored. The first matching ssl_bump action wins and is applied at the
2509 end of the current step. If no rules match, the splice action is used.
2510 See atstep ACL for a list of the supported SslBump steps.
2512 This clause supports both fast and slow acl types.
2513 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2515 See also: http_port ssl-bump, https_port ssl-bump, and atstep ACL.
2518 # Example: Bump all requests except those originating from
2519 # localhost or those going to example.com.
2521 acl broken_sites dstdomain .example.com
2522 ssl_bump splice localhost
2523 ssl_bump splice broken_sites
2527 NAME: sslproxy_flags
2530 LOC: Config.ssl_client.flags
2533 Various flags modifying the use of SSL while proxying https:// URLs:
2534 DONT_VERIFY_PEER Accept certificates that fail verification.
2535 For refined control, see sslproxy_cert_error.
2536 NO_DEFAULT_CA Don't use the default CA list built in
2540 NAME: sslproxy_cert_error
2543 DEFAULT_DOC: Server certificate errors terminate the transaction.
2544 LOC: Config.ssl_client.cert_error
2547 Use this ACL to bypass server certificate validation errors.
2549 For example, the following lines will bypass all validation errors
2550 when talking to servers for example.com. All other
2551 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2553 acl BrokenButTrustedServers dstdomain example.com
2554 sslproxy_cert_error allow BrokenButTrustedServers
2555 sslproxy_cert_error deny all
2557 This clause only supports fast acl types.
2558 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2559 Using slow acl types may result in server crashes
2561 Without this option, all server certificate validation errors
2562 terminate the transaction to protect Squid and the client.
2564 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2565 but should not happen unless your OpenSSL library is buggy.
2568 Bypassing validation errors is dangerous because an
2569 error usually implies that the server cannot be trusted
2570 and the connection may be insecure.
2572 See also: sslproxy_flags and DONT_VERIFY_PEER.
2575 NAME: sslproxy_cert_sign
2578 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2579 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2580 POSTSCRIPTUM: signTrusted all
2581 TYPE: sslproxy_cert_sign
2582 LOC: Config.ssl_client.cert_sign
2585 sslproxy_cert_sign <signing algorithm> acl ...
2587 The following certificate signing algorithms are supported:
2590 Sign using the configured CA certificate which is usually
2591 placed in and trusted by end-user browsers. This is the
2592 default for trusted origin server certificates.
2595 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2596 This is the default for untrusted origin server certificates
2597 that are not self-signed (see ssl::certUntrusted).
2600 Sign using a self-signed certificate with the right CN to
2601 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2602 browser. This is the default for self-signed origin server
2603 certificates (see ssl::certSelfSigned).
2605 This clause only supports fast acl types.
2607 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2608 signing algorithm to generate the certificate and ignores all
2609 subsequent sslproxy_cert_sign options (the first match wins). If no
2610 acl(s) match, the default signing algorithm is determined by errors
2611 detected when obtaining and validating the origin server certificate.
2613 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2614 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2615 CONNECT request that carries a domain name. In all other cases (CONNECT
2616 to an IP address or an intercepted SSL connection), Squid cannot detect
2617 the domain mismatch at certificate generation time when
2618 bump-server-first is used.
2621 NAME: sslproxy_cert_adapt
2624 TYPE: sslproxy_cert_adapt
2625 LOC: Config.ssl_client.cert_adapt
2628 sslproxy_cert_adapt <adaptation algorithm> acl ...
2630 The following certificate adaptation algorithms are supported:
2633 Sets the "Not After" property to the "Not After" property of
2634 the CA certificate used to sign generated certificates.
2637 Sets the "Not Before" property to the "Not Before" property of
2638 the CA certificate used to sign generated certificates.
2640 setCommonName or setCommonName{CN}
2641 Sets Subject.CN property to the host name specified as a
2642 CN parameter or, if no explicit CN parameter was specified,
2643 extracted from the CONNECT request. It is a misconfiguration
2644 to use setCommonName without an explicit parameter for
2645 intercepted or tproxied SSL connections.
2647 This clause only supports fast acl types.
2649 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2650 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2651 corresponding adaptation algorithm to generate the certificate and
2652 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2653 group (i.e., the first match wins within each algorithm group). If no
2654 acl(s) match, the default mimicking action takes place.
2656 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2657 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2658 CONNECT request that carries a domain name. In all other cases (CONNECT
2659 to an IP address or an intercepted SSL connection), Squid cannot detect
2660 the domain mismatch at certificate generation time when
2661 bump-server-first is used.
2664 NAME: sslpassword_program
2667 LOC: Config.Program.ssl_password
2670 Specify a program used for entering SSL key passphrases
2671 when using encrypted SSL certificate keys. If not specified
2672 keys must either be unencrypted, or Squid started with the -N
2673 option to allow it to query interactively for the passphrase.
2675 The key file name is given as argument to the program allowing
2676 selection of the right password if you have multiple encrypted
2681 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2682 -----------------------------------------------------------------------------
2685 NAME: sslcrtd_program
2688 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2689 LOC: Ssl::TheConfig.ssl_crtd
2691 Specify the location and options of the executable for ssl_crtd process.
2692 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2693 For more information use:
2694 @DEFAULT_SSL_CRTD@ -h
2697 NAME: sslcrtd_children
2698 TYPE: HelperChildConfig
2700 DEFAULT: 32 startup=5 idle=1
2701 LOC: Ssl::TheConfig.ssl_crtdChildren
2703 The maximum number of processes spawn to service ssl server.
2704 The maximum this may be safely set to is 32.
2706 The startup= and idle= options allow some measure of skew in your
2711 Sets the minimum number of processes to spawn when Squid
2712 starts or reconfigures. When set to zero the first request will
2713 cause spawning of the first child process to handle it.
2715 Starting too few children temporary slows Squid under load while it
2716 tries to spawn enough additional processes to cope with traffic.
2720 Sets a minimum of how many processes Squid is to try and keep available
2721 at all times. When traffic begins to rise above what the existing
2722 processes can handle this many more will be spawned up to the maximum
2723 configured. A minimum setting of 1 is required.
2725 You must have at least one ssl_crtd process.
2728 NAME: sslcrtvalidator_program
2732 LOC: Ssl::TheConfig.ssl_crt_validator
2734 Specify the location and options of the executable for ssl_crt_validator
2737 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2740 ttl=n TTL in seconds for cached results. The default is 60 secs
2741 cache=n limit the result cache size. The default value is 2048
2744 NAME: sslcrtvalidator_children
2745 TYPE: HelperChildConfig
2747 DEFAULT: 32 startup=5 idle=1 concurrency=1
2748 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2750 The maximum number of processes spawn to service SSL server.
2751 The maximum this may be safely set to is 32.
2753 The startup= and idle= options allow some measure of skew in your
2758 Sets the minimum number of processes to spawn when Squid
2759 starts or reconfigures. When set to zero the first request will
2760 cause spawning of the first child process to handle it.
2762 Starting too few children temporary slows Squid under load while it
2763 tries to spawn enough additional processes to cope with traffic.
2767 Sets a minimum of how many processes Squid is to try and keep available
2768 at all times. When traffic begins to rise above what the existing
2769 processes can handle this many more will be spawned up to the maximum
2770 configured. A minimum setting of 1 is required.
2774 The number of requests each certificate validator helper can handle in
2775 parallel. A value of 0 indicates the certficate validator does not
2776 support concurrency. Defaults to 1.
2778 When this directive is set to a value >= 1 then the protocol
2779 used to communicate with the helper is modified to include
2780 a request ID in front of the request/response. The request
2781 ID from the request must be echoed back with the response
2784 You must have at least one ssl_crt_validator process.
2788 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2789 -----------------------------------------------------------------------------
2797 To specify other caches in a hierarchy, use the format:
2799 cache_peer hostname type http-port icp-port [options]
2804 # hostname type port port options
2805 # -------------------- -------- ----- ----- -----------
2806 cache_peer parent.foo.net parent 3128 3130 default
2807 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2808 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2809 cache_peer example.com parent 80 0 default
2810 cache_peer cdn.example.com sibling 3128 0
2812 type: either 'parent', 'sibling', or 'multicast'.
2814 proxy-port: The port number where the peer accept HTTP requests.
2815 For other Squid proxies this is usually 3128
2816 For web servers this is usually 80
2818 icp-port: Used for querying neighbor caches about objects.
2819 Set to 0 if the peer does not support ICP or HTCP.
2820 See ICP and HTCP options below for additional details.
2823 ==== ICP OPTIONS ====
2825 You MUST also set icp_port and icp_access explicitly when using these options.
2826 The defaults will prevent peer traffic using ICP.
2829 no-query Disable ICP queries to this neighbor.
2832 Indicates the named peer is a member of a multicast group.
2833 ICP queries will not be sent directly to the peer, but ICP
2834 replies will be accepted from it.
2836 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2837 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2840 To only send ICP queries to this neighbor infrequently.
2841 This is used to keep the neighbor round trip time updated
2842 and is usually used in conjunction with weighted-round-robin.
2845 ==== HTCP OPTIONS ====
2847 You MUST also set htcp_port and htcp_access explicitly when using these options.
2848 The defaults will prevent peer traffic using HTCP.
2851 htcp Send HTCP, instead of ICP, queries to the neighbor.
2852 You probably also want to set the "icp-port" to 4827
2853 instead of 3130. This directive accepts a comma separated
2854 list of options described below.
2856 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2858 htcp=no-clr Send HTCP to the neighbor but without
2859 sending any CLR requests. This cannot be used with
2862 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2863 This cannot be used with no-clr.
2866 Send HTCP to the neighbor including CLRs but only when
2867 they do not result from PURGE requests.
2870 Forward any HTCP CLR requests this proxy receives to the peer.
2873 ==== PEER SELECTION METHODS ====
2875 The default peer selection method is ICP, with the first responding peer
2876 being used as source. These options can be used for better load balancing.
2879 default This is a parent cache which can be used as a "last-resort"
2880 if a peer cannot be located by any of the peer-selection methods.
2881 If specified more than once, only the first is used.
2883 round-robin Load-Balance parents which should be used in a round-robin
2884 fashion in the absence of any ICP queries.
2885 weight=N can be used to add bias.
2887 weighted-round-robin
2888 Load-Balance parents which should be used in a round-robin
2889 fashion with the frequency of each parent being based on the
2890 round trip time. Closer parents are used more often.
2891 Usually used for background-ping parents.
2892 weight=N can be used to add bias.
2894 carp Load-Balance parents which should be used as a CARP array.
2895 The requests will be distributed among the parents based on the
2896 CARP load balancing hash function based on their weight.
2898 userhash Load-balance parents based on the client proxy_auth or ident username.
2900 sourcehash Load-balance parents based on the client source IP.
2903 To be used only for cache peers of type "multicast".
2904 ALL members of this multicast group have "sibling"
2905 relationship with it, not "parent". This is to a multicast
2906 group when the requested object would be fetched only from
2907 a "parent" cache, anyway. It's useful, e.g., when
2908 configuring a pool of redundant Squid proxies, being
2909 members of the same multicast group.
2912 ==== PEER SELECTION OPTIONS ====
2914 weight=N use to affect the selection of a peer during any weighted
2915 peer-selection mechanisms.
2916 The weight must be an integer; default is 1,
2917 larger weights are favored more.
2918 This option does not affect parent selection if a peering
2919 protocol is not in use.
2921 basetime=N Specify a base amount to be subtracted from round trip
2923 It is subtracted before division by weight in calculating
2924 which parent to fectch from. If the rtt is less than the
2925 base time the rtt is set to a minimal value.
2927 ttl=N Specify a TTL to use when sending multicast ICP queries
2929 Only useful when sending to a multicast group.
2930 Because we don't accept ICP replies from random
2931 hosts, you must configure other group members as
2932 peers with the 'multicast-responder' option.
2934 no-delay To prevent access to this neighbor from influencing the
2937 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2938 enabled) for this host from the specified URL rather
2939 than the Squid default location.
2942 ==== CARP OPTIONS ====
2944 carp-key=key-specification
2945 use a different key than the full URL to hash against the peer.
2946 the key-specification is a comma-separated list of the keywords
2947 scheme, host, port, path, params
2948 Order is not important.
2950 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2952 originserver Causes this parent to be contacted as an origin server.
2953 Meant to be used in accelerator setups when the peer
2957 Set the Host header of requests forwarded to this peer.
2958 Useful in accelerator setups where the server (peer)
2959 expects a certain domain name but clients may request
2960 others. ie example.com or www.example.com
2962 no-digest Disable request of cache digests.
2965 Disables requesting ICMP RTT database (NetDB).
2968 ==== AUTHENTICATION OPTIONS ====
2971 If this is a personal/workgroup proxy and your parent
2972 requires proxy authentication.
2974 Note: The string can include URL escapes (i.e. %20 for
2975 spaces). This also means % must be written as %%.
2978 Send login details received from client to this peer.
2979 Both Proxy- and WWW-Authorization headers are passed
2980 without alteration to the peer.
2981 Authentication is not required by Squid for this to work.
2983 Note: This will pass any form of authentication but
2984 only Basic auth will work through a proxy unless the
2985 connection-auth options are also used.
2987 login=PASS Send login details received from client to this peer.
2988 Authentication is not required by this option.
2990 If there are no client-provided authentication headers
2991 to pass on, but username and password are available
2992 from an external ACL user= and password= result tags
2993 they may be sent instead.
2995 Note: To combine this with proxy_auth both proxies must
2996 share the same user database as HTTP only allows for
2997 a single login (one for proxy, one for origin server).
2998 Also be warned this will expose your users proxy
2999 password to the peer. USE WITH CAUTION
3002 Send the username to the upstream cache, but with a
3003 fixed password. This is meant to be used when the peer
3004 is in another administrative domain, but it is still
3005 needed to identify each user.
3006 The star can optionally be followed by some extra
3007 information which is added to the username. This can
3008 be used to identify this proxy to the peer, similar to
3009 the login=username:password option above.
3012 If this is a personal/workgroup proxy and your parent
3013 requires a secure proxy authentication.
3014 The first principal from the default keytab or defined by
3015 the environment variable KRB5_KTNAME will be used.
3017 WARNING: The connection may transmit requests from multiple
3018 clients. Negotiate often assumes end-to-end authentication
3019 and a single-client. Which is not strictly true here.
3021 login=NEGOTIATE:principal_name
3022 If this is a personal/workgroup proxy and your parent
3023 requires a secure proxy authentication.
3024 The principal principal_name from the default keytab or
3025 defined by the environment variable KRB5_KTNAME will be
3028 WARNING: The connection may transmit requests from multiple
3029 clients. Negotiate often assumes end-to-end authentication
3030 and a single-client. Which is not strictly true here.
3032 connection-auth=on|off
3033 Tell Squid that this peer does or not support Microsoft
3034 connection oriented authentication, and any such
3035 challenges received from there should be ignored.
3036 Default is auto to automatically determine the status
3040 ==== SSL / HTTPS / TLS OPTIONS ====
3042 ssl Encrypt connections to this peer with SSL/TLS.
3044 sslcert=/path/to/ssl/certificate
3045 A client SSL certificate to use when connecting to
3048 sslkey=/path/to/ssl/key
3049 The private SSL key corresponding to sslcert above.
3050 If 'sslkey' is not specified 'sslcert' is assumed to
3051 reference a combined file containing both the
3052 certificate and the key.
3054 sslversion=1|2|3|4|5|6
3055 The SSL version to use when connecting to this peer
3056 1 = automatic (default)
3063 sslcipher=... The list of valid SSL ciphers to use when connecting
3066 ssloptions=... Specify various SSL implementation options:
3068 NO_SSLv2 Disallow the use of SSLv2
3069 NO_SSLv3 Disallow the use of SSLv3
3070 NO_TLSv1 Disallow the use of TLSv1.0
3071 NO_TLSv1_1 Disallow the use of TLSv1.1
3072 NO_TLSv1_2 Disallow the use of TLSv1.2
3074 Always create a new key when using
3075 temporary/ephemeral DH key exchanges
3076 ALL Enable various bug workarounds
3077 suggested as "harmless" by OpenSSL
3078 Be warned that this reduces SSL/TLS
3079 strength to some attacks.
3081 See the OpenSSL SSL_CTX_set_options documentation for a
3084 sslcafile=... A file containing additional CA certificates to use
3085 when verifying the peer certificate.
3087 sslcapath=... A directory containing additional CA certificates to
3088 use when verifying the peer certificate.
3090 sslcrlfile=... A certificate revocation list file to use when
3091 verifying the peer certificate.
3093 sslflags=... Specify various flags modifying the SSL implementation:
3096 Accept certificates even if they fail to
3099 Don't use the default CA list built in
3102 Don't verify the peer certificate
3103 matches the server name
3105 ssldomain= The peer name as advertised in it's certificate.
3106 Used for verifying the correctness of the received peer
3107 certificate. If not specified the peer hostname will be
3111 Enable the "Front-End-Https: On" header needed when
3112 using Squid as a SSL frontend in front of Microsoft OWA.
3113 See MS KB document Q307347 for details on this header.
3114 If set to auto the header will only be added if the
3115 request is forwarded as a https:// URL.
3118 ==== GENERAL OPTIONS ====
3121 A peer-specific connect timeout.
3122 Also see the peer_connect_timeout directive.
3124 connect-fail-limit=N
3125 How many times connecting to a peer must fail before
3126 it is marked as down. Standby connection failures
3127 count towards this limit. Default is 10.
3129 allow-miss Disable Squid's use of only-if-cached when forwarding
3130 requests to siblings. This is primarily useful when
3131 icp_hit_stale is used by the sibling. To extensive use
3132 of this option may result in forwarding loops, and you
3133 should avoid having two-way peerings with this option.
3134 For example to deny peer usage on requests from peer
3135 by denying cache_peer_access if the source is a peer.
3137 max-conn=N Limit the number of concurrent connections the Squid
3138 may open to this peer, including already opened idle
3139 and standby connections. There is no peer-specific
3140 connection limit by default.
3142 A peer exceeding the limit is not used for new
3143 requests unless a standby connection is available.
3145 max-conn currently works poorly with idle persistent
3146 connections: When a peer reaches its max-conn limit,
3147 and there are idle persistent connections to the peer,
3148 the peer may not be selected because the limiting code
3149 does not know whether Squid can reuse those idle
3152 standby=N Maintain a pool of N "hot standby" connections to an
3153 UP peer, available for requests when no idle
3154 persistent connection is available (or safe) to use.
3155 By default and with zero N, no such pool is maintained.
3156 N must not exceed the max-conn limit (if any).
3158 At start or after reconfiguration, Squid opens new TCP
3159 standby connections until there are N connections
3160 available and then replenishes the standby pool as
3161 opened connections are used up for requests. A used
3162 connection never goes back to the standby pool, but
3163 may go to the regular idle persistent connection pool
3164 shared by all peers and origin servers.
3166 Squid never opens multiple new standby connections
3167 concurrently. This one-at-a-time approach minimizes
3168 flooding-like effect on peers. Furthermore, just a few
3169 standby connections should be sufficient in most cases
3170 to supply most new requests with a ready-to-use
3173 Standby connections obey server_idle_pconn_timeout.
3174 For the feature to work as intended, the peer must be
3175 configured to accept and keep them open longer than
3176 the idle timeout at the connecting Squid, to minimize
3177 race conditions typical to idle used persistent
3178 connections. Default request_timeout and
3179 server_idle_pconn_timeout values ensure such a
3182 name=xxx Unique name for the peer.
3183 Required if you have multiple peers on the same host
3184 but different ports.
3185 This name can be used in cache_peer_access and similar
3186 directives to dentify the peer.
3187 Can be used by outgoing access controls through the
3190 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3191 requests to this peer. Use normal address selection instead.
3192 This overrides the spoof_client_ip ACL.
3194 proxy-only objects fetched from the peer will not be stored locally.
3198 NAME: cache_peer_domain cache_host_domain
3203 Use to limit the domains for which a neighbor cache will be
3207 cache_peer_domain cache-host domain [domain ...]
3208 cache_peer_domain cache-host !domain
3210 For example, specifying
3212 cache_peer_domain parent.foo.net .edu
3214 has the effect such that UDP query packets are sent to
3215 'bigserver' only when the requested object exists on a
3216 server in the .edu domain. Prefixing the domainname
3217 with '!' means the cache will be queried for objects
3220 NOTE: * Any number of domains may be given for a cache-host,
3221 either on the same or separate lines.
3222 * When multiple domains are given for a particular
3223 cache-host, the first matched domain is applied.
3224 * Cache hosts with no domain restrictions are queried
3226 * There are no defaults.
3227 * There is also a 'cache_peer_access' tag in the ACL
3231 NAME: cache_peer_access
3236 Similar to 'cache_peer_domain' but provides more flexibility by
3240 cache_peer_access cache-host allow|deny [!]aclname ...
3242 The syntax is identical to 'http_access' and the other lists of
3243 ACL elements. See the comments for 'http_access' below, or
3244 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3247 NAME: neighbor_type_domain
3248 TYPE: hostdomaintype
3250 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3253 Modify the cache_peer neighbor type when passing requests
3254 about specific domains to the peer.
3257 neighbor_type_domain neighbor parent|sibling domain domain ...
3260 cache_peer foo.example.com parent 3128 3130
3261 neighbor_type_domain foo.example.com sibling .au .de
3263 The above configuration treats all requests to foo.example.com as a
3264 parent proxy unless the request is for a .au or .de ccTLD domain name.
3267 NAME: dead_peer_timeout
3271 LOC: Config.Timeout.deadPeer
3273 This controls how long Squid waits to declare a peer cache
3274 as "dead." If there are no ICP replies received in this
3275 amount of time, Squid will declare the peer dead and not
3276 expect to receive any further ICP replies. However, it
3277 continues to send ICP queries, and will mark the peer as
3278 alive upon receipt of the first subsequent ICP reply.
3280 This timeout also affects when Squid expects to receive ICP
3281 replies from peers. If more than 'dead_peer' seconds have
3282 passed since the last ICP reply was received, Squid will not
3283 expect to receive an ICP reply on the next query. Thus, if
3284 your time between requests is greater than this timeout, you
3285 will see a lot of requests sent DIRECT to origin servers
3286 instead of to your parents.
3289 NAME: forward_max_tries
3292 LOC: Config.forward_max_tries
3294 Controls how many different forward paths Squid will try
3295 before giving up. See also forward_timeout.
3297 NOTE: connect_retries (default: none) can make each of these
3298 possible forwarding paths be tried multiple times.
3301 NAME: hierarchy_stoplist
3304 LOC: Config.hierarchy_stoplist
3306 A list of words which, if found in a URL, cause the object to
3307 be handled directly by this cache. In other words, use this
3308 to not query neighbor caches for certain objects. You may
3309 list this option multiple times.
3312 hierarchy_stoplist cgi-bin ?
3314 Note: never_direct overrides this option.
3318 MEMORY CACHE OPTIONS
3319 -----------------------------------------------------------------------------
3326 LOC: Config.memMaxSize
3328 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3329 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3330 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3331 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3333 'cache_mem' specifies the ideal amount of memory to be used
3335 * In-Transit objects
3337 * Negative-Cached objects
3339 Data for these objects are stored in 4 KB blocks. This
3340 parameter specifies the ideal upper limit on the total size of
3341 4 KB blocks allocated. In-Transit objects take the highest
3344 In-transit objects have priority over the others. When
3345 additional space is needed for incoming data, negative-cached
3346 and hot objects will be released. In other words, the
3347 negative-cached and hot objects will fill up any unused space
3348 not needed for in-transit objects.
3350 If circumstances require, this limit will be exceeded.
3351 Specifically, if your incoming request rate requires more than
3352 'cache_mem' of memory to hold in-transit objects, Squid will
3353 exceed this limit to satisfy the new requests. When the load
3354 decreases, blocks will be freed until the high-water mark is
3355 reached. Thereafter, blocks will be used to store hot
3358 If shared memory caching is enabled, Squid does not use the shared
3359 cache space for in-transit objects, but they still consume as much
3360 local memory as they need. For more details about the shared memory
3361 cache, see memory_cache_shared.
3364 NAME: maximum_object_size_in_memory
3368 LOC: Config.Store.maxInMemObjSize
3370 Objects greater than this size will not be attempted to kept in
3371 the memory cache. This should be set high enough to keep objects
3372 accessed frequently in memory to improve performance whilst low
3373 enough to keep larger objects from hoarding cache_mem.
3376 NAME: memory_cache_shared
3379 LOC: Config.memShared
3381 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3383 Controls whether the memory cache is shared among SMP workers.
3385 The shared memory cache is meant to occupy cache_mem bytes and replace
3386 the non-shared memory cache, although some entities may still be
3387 cached locally by workers for now (e.g., internal and in-transit
3388 objects may be served from a local memory cache even if shared memory
3389 caching is enabled).
3391 By default, the memory cache is shared if and only if all of the
3392 following conditions are satisfied: Squid runs in SMP mode with
3393 multiple workers, cache_mem is positive, and Squid environment
3394 supports required IPC primitives (e.g., POSIX shared memory segments
3395 and GCC-style atomic operations).
3397 To avoid blocking locks, shared memory uses opportunistic algorithms
3398 that do not guarantee that every cachable entity that could have been
3399 shared among SMP workers will actually be shared.
3401 Currently, entities exceeding 32KB in size cannot be shared.
3404 NAME: memory_cache_mode
3408 DEFAULT_DOC: Keep the most recently fetched objects in memory
3410 Controls which objects to keep in the memory cache (cache_mem)
3412 always Keep most recently fetched objects in memory (default)
3414 disk Only disk cache hits are kept in memory, which means
3415 an object must first be cached on disk and then hit
3416 a second time before cached in memory.
3418 network Only objects fetched from network is kept in memory
3421 NAME: memory_replacement_policy
3423 LOC: Config.memPolicy
3426 The memory replacement policy parameter determines which
3427 objects are purged from memory when memory space is needed.
3429 See cache_replacement_policy for details on algorithms.
3434 -----------------------------------------------------------------------------
3437 NAME: cache_replacement_policy
3439 LOC: Config.replPolicy
3442 The cache replacement policy parameter determines which
3443 objects are evicted (replaced) when disk space is needed.
3445 lru : Squid's original list based LRU policy
3446 heap GDSF : Greedy-Dual Size Frequency
3447 heap LFUDA: Least Frequently Used with Dynamic Aging
3448 heap LRU : LRU policy implemented using a heap
3450 Applies to any cache_dir lines listed below this directive.
3452 The LRU policies keeps recently referenced objects.
3454 The heap GDSF policy optimizes object hit rate by keeping smaller
3455 popular objects in cache so it has a better chance of getting a
3456 hit. It achieves a lower byte hit rate than LFUDA though since
3457 it evicts larger (possibly popular) objects.
3459 The heap LFUDA policy keeps popular objects in cache regardless of
3460 their size and thus optimizes byte hit rate at the expense of
3461 hit rate since one large, popular object will prevent many
3462 smaller, slightly less popular objects from being cached.
3464 Both policies utilize a dynamic aging mechanism that prevents
3465 cache pollution that can otherwise occur with frequency-based
3466 replacement policies.
3468 NOTE: if using the LFUDA replacement policy you should increase
3469 the value of maximum_object_size above its default of 4 MB to
3470 to maximize the potential byte hit rate improvement of LFUDA.
3472 For more information about the GDSF and LFUDA cache replacement
3473 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3474 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3477 NAME: minimum_object_size
3481 DEFAULT_DOC: no limit
3482 LOC: Config.Store.minObjectSize
3484 Objects smaller than this size will NOT be saved on disk. The
3485 value is specified in bytes, and the default is 0 KB, which
3486 means all responses can be stored.
3489 NAME: maximum_object_size
3493 LOC: Config.Store.maxObjectSize
3495 Set the default value for max-size parameter on any cache_dir.
3496 The value is specified in bytes, and the default is 4 MB.
3498 If you wish to get a high BYTES hit ratio, you should probably
3499 increase this (one 32 MB object hit counts for 3200 10KB
3502 If you wish to increase hit ratio more than you want to
3503 save bandwidth you should leave this low.
3505 NOTE: if using the LFUDA replacement policy you should increase
3506 this value to maximize the byte hit rate improvement of LFUDA!
3507 See cache_replacement_policy for a discussion of this policy.
3513 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3514 LOC: Config.cacheSwap
3517 cache_dir Type Directory-Name Fs-specific-data [options]
3519 You can specify multiple cache_dir lines to spread the
3520 cache among different disk partitions.
3522 Type specifies the kind of storage system to use. Only "ufs"
3523 is built by default. To enable any of the other storage systems
3524 see the --enable-storeio configure option.
3526 'Directory' is a top-level directory where cache swap
3527 files will be stored. If you want to use an entire disk
3528 for caching, this can be the mount-point directory.
3529 The directory must exist and be writable by the Squid
3530 process. Squid will NOT create this directory for you.
3532 In SMP configurations, cache_dir must not precede the workers option
3533 and should use configuration macros or conditionals to give each
3534 worker interested in disk caching a dedicated cache directory.
3537 ==== The ufs store type ====
3539 "ufs" is the old well-known Squid storage format that has always
3543 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3545 'Mbytes' is the amount of disk space (MB) to use under this
3546 directory. The default is 100 MB. Change this to suit your
3547 configuration. Do NOT put the size of your disk drive here.
3548 Instead, if you want Squid to use the entire disk drive,
3549 subtract 20% and use that value.
3551 'L1' is the number of first-level subdirectories which
3552 will be created under the 'Directory'. The default is 16.
3554 'L2' is the number of second-level subdirectories which
3555 will be created under each first-level directory. The default
3559 ==== The aufs store type ====
3561 "aufs" uses the same storage format as "ufs", utilizing
3562 POSIX-threads to avoid blocking the main Squid process on
3563 disk-I/O. This was formerly known in Squid as async-io.
3566 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3568 see argument descriptions under ufs above
3571 ==== The diskd store type ====
3573 "diskd" uses the same storage format as "ufs", utilizing a
3574 separate process to avoid blocking the main Squid process on
3578 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3580 see argument descriptions under ufs above
3582 Q1 specifies the number of unacknowledged I/O requests when Squid
3583 stops opening new files. If this many messages are in the queues,
3584 Squid won't open new files. Default is 64
3586 Q2 specifies the number of unacknowledged messages when Squid
3587 starts blocking. If this many messages are in the queues,
3588 Squid blocks until it receives some replies. Default is 72
3590 When Q1 < Q2 (the default), the cache directory is optimized
3591 for lower response time at the expense of a decrease in hit
3592 ratio. If Q1 > Q2, the cache directory is optimized for
3593 higher hit ratio at the expense of an increase in response
3597 ==== The rock store type ====
3600 cache_dir rock Directory-Name Mbytes [options]
3602 The Rock Store type is a database-style storage. All cached
3603 entries are stored in a "database" file, using fixed-size slots.
3604 A single entry occupies one or more slots.
3606 If possible, Squid using Rock Store creates a dedicated kid
3607 process called "disker" to avoid blocking Squid worker(s) on disk
3608 I/O. One disker kid is created for each rock cache_dir. Diskers
3609 are created only when Squid, running in daemon mode, has support
3610 for the IpcIo disk I/O module.
3612 swap-timeout=msec: Squid will not start writing a miss to or
3613 reading a hit from disk if it estimates that the swap operation
3614 will take more than the specified number of milliseconds. By
3615 default and when set to zero, disables the disk I/O time limit
3616 enforcement. Ignored when using blocking I/O module because
3617 blocking synchronous I/O does not allow Squid to estimate the
3618 expected swap wait time.
3620 max-swap-rate=swaps/sec: Artificially limits disk access using
3621 the specified I/O rate limit. Swap out requests that
3622 would cause the average I/O rate to exceed the limit are
3623 delayed. Individual swap in requests (i.e., hits or reads) are
3624 not delayed, but they do contribute to measured swap rate and
3625 since they are placed in the same FIFO queue as swap out
3626 requests, they may wait longer if max-swap-rate is smaller.
3627 This is necessary on file systems that buffer "too
3628 many" writes and then start blocking Squid and other processes
3629 while committing those writes to disk. Usually used together
3630 with swap-timeout to avoid excessive delays and queue overflows
3631 when disk demand exceeds available disk "bandwidth". By default
3632 and when set to zero, disables the disk I/O rate limit
3633 enforcement. Currently supported by IpcIo module only.
3635 slot-size=bytes: The size of a database "record" used for
3636 storing cached responses. A cached response occupies at least
3637 one slot and all database I/O is done using individual slots so
3638 increasing this parameter leads to more disk space waste while
3639 decreasing it leads to more disk I/O overheads. Should be a
3640 multiple of your operating system I/O page size. Defaults to
3641 16KBytes. A housekeeping header is stored with each slot and
3642 smaller slot-sizes will be rejected. The header is smaller than
3646 ==== COMMON OPTIONS ====
3648 no-store no new objects should be stored to this cache_dir.
3650 min-size=n the minimum object size in bytes this cache_dir
3651 will accept. It's used to restrict a cache_dir
3652 to only store large objects (e.g. AUFS) while
3653 other stores are optimized for smaller objects
3657 max-size=n the maximum object size in bytes this cache_dir
3659 The value in maximum_object_size directive sets
3660 the default unless more specific details are
3661 available (ie a small store capacity).
3663 Note: To make optimal use of the max-size limits you should order
3664 the cache_dir lines with the smallest max-size value first.
3668 # Uncomment and adjust the following to add a disk cache directory.
3669 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3673 NAME: store_dir_select_algorithm
3675 LOC: Config.store_dir_select_algorithm
3678 How Squid selects which cache_dir to use when the response
3679 object will fit into more than one.
3681 Regardless of which algorithm is used the cache_dir min-size
3682 and max-size parameters are obeyed. As such they can affect
3683 the selection algorithm by limiting the set of considered
3690 This algorithm is suited to caches with similar cache_dir
3691 sizes and disk speeds.
3693 The disk with the least I/O pending is selected.
3694 When there are multiple disks with the same I/O load ranking
3695 the cache_dir with most available capacity is selected.
3697 When a mix of cache_dir sizes are configured the faster disks
3698 have a naturally lower I/O loading and larger disks have more
3699 capacity. So space used to store objects and data throughput
3700 may be very unbalanced towards larger disks.
3705 This algorithm is suited to caches with unequal cache_dir
3708 Each cache_dir is selected in a rotation. The next suitable
3711 Available cache_dir capacity is only considered in relation
3712 to whether the object will fit and meets the min-size and
3713 max-size parameters.
3715 Disk I/O loading is only considered to prevent overload on slow
3716 disks. This algorithm does not spread objects by size, so any
3717 I/O loading per-disk may appear very unbalanced and volatile.
3721 NAME: max_open_disk_fds
3723 LOC: Config.max_open_disk_fds
3725 DEFAULT_DOC: no limit
3727 To avoid having disk as the I/O bottleneck Squid can optionally
3728 bypass the on-disk cache if more than this amount of disk file
3729 descriptors are open.
3731 A value of 0 indicates no limit.
3734 NAME: cache_swap_low
3735 COMMENT: (percent, 0-100)
3738 LOC: Config.Swap.lowWaterMark
3740 The low-water mark for cache object replacement.
3741 Replacement begins when the swap (disk) usage is above the
3742 low-water mark and attempts to maintain utilization near the
3743 low-water mark. As swap utilization gets close to high-water
3744 mark object eviction becomes more aggressive. If utilization is
3745 close to the low-water mark less replacement is done each time.
3747 Defaults are 90% and 95%. If you have a large cache, 5% could be
3748 hundreds of MB. If this is the case you may wish to set these
3749 numbers closer together.
3751 See also cache_swap_high
3754 NAME: cache_swap_high
3755 COMMENT: (percent, 0-100)
3758 LOC: Config.Swap.highWaterMark
3760 The high-water mark for cache object replacement.
3761 Replacement begins when the swap (disk) usage is above the
3762 low-water mark and attempts to maintain utilization near the
3763 low-water mark. As swap utilization gets close to high-water
3764 mark object eviction becomes more aggressive. If utilization is
3765 close to the low-water mark less replacement is done each time.
3767 Defaults are 90% and 95%. If you have a large cache, 5% could be
3768 hundreds of MB. If this is the case you may wish to set these
3769 numbers closer together.
3771 See also cache_swap_low
3776 -----------------------------------------------------------------------------
3783 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
3787 logformat <name> <format specification>
3789 Defines an access log format.
3791 The <format specification> is a string with embedded % format codes
3793 % format codes all follow the same basic structure where all but
3794 the formatcode is optional. Output strings are automatically escaped
3795 as required according to their context and the output format
3796 modifiers are usually not needed, but can be specified if an explicit
3797 output format is desired.
3799 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3801 " output in quoted string format
3802 [ output in squid text log format as used by log_mime_hdrs
3803 # output in URL quoted format
3808 width minimum and/or maximum field width:
3809 [width_min][.width_max]
3810 When minimum starts with 0, the field is zero-padded.
3811 String values exceeding maximum width are truncated.
3813 {arg} argument such as header name etc
3817 % a literal % character
3818 sn Unique sequence number per log line entry
3819 err_code The ID of an error response served by Squid or
3820 a similar internal error identifier.
3821 err_detail Additional err_code-dependent error information.
3822 note The annotation specified by the argument. Also
3823 logs the adaptation meta headers set by the
3824 adaptation_meta configuration parameter.
3825 If no argument given all annotations logged.
3826 The argument may include a separator to use with
3829 By default, multiple note values are separated with ","
3830 and multiple notes are separated with "\r\n".
3831 When logging named notes with %{name}note, the
3832 explicitly configured separator is used between note
3833 values. When logging all notes with %note, the
3834 explicitly configured separator is used between
3835 individual notes. There is currently no way to
3836 specify both value and notes separators when logging
3837 all notes with %note.
3839 Connection related format codes:
3841 >a Client source IP address
3843 >p Client source port
3844 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3845 >la Local IP address the client connected to
3846 >lp Local port number the client connected to
3847 >qos Client connection TOS/DSCP value set by Squid
3848 >nfmark Client connection netfilter mark set by Squid
3850 la Local listening IP address the client connection was connected to.
3851 lp Local listening port number the client connection was connected to.
3853 <a Server IP address of the last server or peer connection
3854 <A Server FQDN or peer name
3855 <p Server port number of the last server or peer connection
3856 <la Local IP address of the last server or peer connection
3857 <lp Local port number of the last server or peer connection
3858 <qos Server connection TOS/DSCP value set by Squid
3859 <nfmark Server connection netfilter mark set by Squid
3861 Time related format codes:
3863 ts Seconds since epoch
3864 tu subsecond time (milliseconds)
3865 tl Local time. Optional strftime format argument
3866 default %d/%b/%Y:%H:%M:%S %z
3867 tg GMT time. Optional strftime format argument
3868 default %d/%b/%Y:%H:%M:%S %z
3869 tr Response time (milliseconds)
3870 dt Total time spent making DNS lookups (milliseconds)
3871 tS Approximate master transaction start time in
3872 <full seconds since epoch>.<fractional seconds> format.
3873 Currently, Squid considers the master transaction
3874 started when a complete HTTP request header initiating
3875 the transaction is received from the client. This is
3876 the same value that Squid uses to calculate transaction
3877 response time when logging %tr to access.log. Currently,
3878 Squid uses millisecond resolution for %tS values,
3879 similar to the default access.log "current time" field
3882 Access Control related format codes:
3884 et Tag returned by external acl
3885 ea Log string returned by external acl
3886 un User name (any available)
3887 ul User name from authentication
3888 ue User name from external acl helper
3889 ui User name from ident
3890 us User name from SSL
3891 credentials Client credentials. The exact meaning depends on
3892 the authentication scheme: For Basic authentication,
3893 it is the password; for Digest, the realm sent by the
3894 client; for NTLM and Negotiate, the client challenge
3895 or client credentials prefixed with "YR " or "KK ".
3897 HTTP related format codes:
3901 [http::]rm Request method (GET/POST etc)
3902 [http::]>rm Request method from client
3903 [http::]<rm Request method sent to server or peer
3904 [http::]ru Request URL from client (historic, filtered for logging)
3905 [http::]>ru Request URL from client
3906 [http::]<ru Request URL sent to server or peer
3907 [http::]>rs Request URL scheme from client
3908 [http::]<rs Request URL scheme sent to server or peer
3909 [http::]>rd Request URL domain from client
3910 [http::]>rd Request URL domain sent to server or peer
3911 [http::]>rP Request URL port from client
3912 [http::]<rP Request URL port sent to server or peer
3913 [http::]rp Request URL path excluding hostname
3914 [http::]>rp Request URL path excluding hostname from client
3915 [http::]<rp Request URL path excluding hostname sent to server or peer
3916 [http::]rv Request protocol version
3917 [http::]>rv Request protocol version from client
3918 [http::]<rv Request protocol version sent to server or peer
3920 [http::]>h Original received request header.
3921 Usually differs from the request header sent by
3922 Squid, although most fields are often preserved.
3923 Accepts optional header field name/value filter
3924 argument using name[:[separator]element] format.
3925 [http::]>ha Received request header after adaptation and
3926 redirection (pre-cache REQMOD vectoring point).
3927 Usually differs from the request header sent by
3928 Squid, although most fields are often preserved.
3929 Optional header name argument as for >h
3934 [http::]<Hs HTTP status code received from the next hop
3935 [http::]>Hs HTTP status code sent to the client
3937 [http::]<h Reply header. Optional header name argument
3940 [http::]mt MIME content type
3945 [http::]st Total size of request + reply traffic with client
3946 [http::]>st Total size of request received from client.
3947 Excluding chunked encoding bytes.
3948 [http::]<st Total size of reply sent to client (after adaptation)
3950 [http::]>sh Size of request headers received from client
3951 [http::]<sh Size of reply headers sent to client (after adaptation)
3953 [http::]<sH Reply high offset sent
3954 [http::]<sS Upstream object size
3956 [http::]<bs Number of HTTP-equivalent message body bytes
3957 received from the next hop, excluding chunked
3958 transfer encoding and control messages.
3959 Generated FTP/Gopher listings are treated as
3965 [http::]<pt Peer response time in milliseconds. The timer starts
3966 when the last request byte is sent to the next hop
3967 and stops when the last response byte is received.
3968 [http::]<tt Total server-side time in milliseconds. The timer
3969 starts with the first connect request (or write I/O)
3970 sent to the first selected peer. The timer stops
3971 with the last I/O with the last peer.
3973 Squid handling related format codes:
3975 Ss Squid request status (TCP_MISS etc)
3976 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3978 SSL-related format codes:
3980 ssl::bump_mode SslBump decision for the transaction:
3982 For CONNECT requests that initiated bumping of
3983 a connection and for any request received on
3984 an already bumped connection, Squid logs the
3985 corresponding SslBump mode ("server-first" or
3986 "client-first"). See the ssl_bump option for
3987 more information about these modes.
3989 A "none" token is logged for requests that
3990 triggered "ssl_bump" ACL evaluation matching
3991 either a "none" rule or no rules at all.
3993 In all other cases, a single dash ("-") is
3996 If ICAP is enabled, the following code becomes available (as
3997 well as ICAP log codes documented with the icap_log option):
3999 icap::tt Total ICAP processing time for the HTTP
4000 transaction. The timer ticks when ICAP
4001 ACLs are checked and when ICAP
4002 transaction is in progress.
4004 If adaptation is enabled the following three codes become available:
4006 adapt::<last_h The header of the last ICAP response or
4007 meta-information from the last eCAP
4008 transaction related to the HTTP transaction.
4009 Like <h, accepts an optional header name
4012 adapt::sum_trs Summed adaptation transaction response
4013 times recorded as a comma-separated list in
4014 the order of transaction start time. Each time
4015 value is recorded as an integer number,
4016 representing response time of one or more
4017 adaptation (ICAP or eCAP) transaction in
4018 milliseconds. When a failed transaction is
4019 being retried or repeated, its time is not
4020 logged individually but added to the
4021 replacement (next) transaction. See also:
4024 adapt::all_trs All adaptation transaction response times.
4025 Same as adaptation_strs but response times of
4026 individual transactions are never added
4027 together. Instead, all transaction response
4028 times are recorded individually.
4030 You can prefix adapt::*_trs format codes with adaptation
4031 service name in curly braces to record response time(s) specific
4032 to that service. For example: %{my_service}adapt::sum_trs
4034 If SSL is enabled, the following formating codes become available:
4036 %ssl::>cert_subject The Subject field of the received client
4037 SSL certificate or a dash ('-') if Squid has
4038 received an invalid/malformed certificate or
4039 no certificate at all. Consider encoding the
4040 logged value because Subject often has spaces.
4042 %ssl::>cert_issuer The Issuer field of the received client
4043 SSL certificate or a dash ('-') if Squid has
4044 received an invalid/malformed certificate or
4045 no certificate at all. Consider encoding the
4046 logged value because Issuer often has spaces.
4048 The default formats available (which do not need re-defining) are:
4050 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
4051 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
4052 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
4053 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
4054 logformat useragent %>a [%tl] "%{User-Agent}>h"
4056 NOTE: When the log_mime_hdrs directive is set to ON.
4057 The squid, common and combined formats have a safely encoded copy
4058 of the mime headers appended to each line within a pair of brackets.
4060 NOTE: The common and combined formats are not quite true to the Apache definition.
4061 The logs from Squid contain an extra status and hierarchy code appended.
4065 NAME: access_log cache_access_log
4067 LOC: Config.Log.accesslogs
4068 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
4070 Configures whether and how Squid logs HTTP and ICP transactions.
4071 If access logging is enabled, a single line is logged for every
4072 matching HTTP or ICP request. The recommended directive formats are:
4074 access_log <module>:<place> [option ...] [acl acl ...]
4075 access_log none [acl acl ...]
4077 The following directive format is accepted but may be deprecated:
4078 access_log <module>:<place> [<logformat name> [acl acl ...]]
4080 In most cases, the first ACL name must not contain the '=' character
4081 and should not be equal to an existing logformat name. You can always
4082 start with an 'all' ACL to work around those restrictions.
4084 Will log to the specified module:place using the specified format (which
4085 must be defined in a logformat directive) those entries which match
4086 ALL the acl's specified (which must be defined in acl clauses).
4087 If no acl is specified, all requests will be logged to this destination.
4089 ===== Available options for the recommended directive format =====
4091 logformat=name Names log line format (either built-in or
4092 defined by a logformat directive). Defaults
4095 buffer-size=64KB Defines approximate buffering limit for log
4096 records (see buffered_logs). Squid should not
4097 keep more than the specified size and, hence,
4098 should flush records before the buffer becomes
4099 full to avoid overflows under normal
4100 conditions (the exact flushing algorithm is
4101 module-dependent though). The on-error option
4102 controls overflow handling.
4104 on-error=die|drop Defines action on unrecoverable errors. The
4105 'drop' action ignores (i.e., does not log)
4106 affected log records. The default 'die' action
4107 kills the affected worker. The drop action
4108 support has not been tested for modules other
4111 ===== Modules Currently available =====
4113 none Do not log any requests matching these ACL.
4114 Do not specify Place or logformat name.
4116 stdio Write each log line to disk immediately at the completion of
4118 Place: the filename and path to be written.
4120 daemon Very similar to stdio. But instead of writing to disk the log
4121 line is passed to a daemon helper for asychronous handling instead.
4122 Place: varies depending on the daemon.
4124 log_file_daemon Place: the file name and path to be written.
4126 syslog To log each request via syslog facility.
4127 Place: The syslog facility and priority level for these entries.
4128 Place Format: facility.priority
4130 where facility could be any of:
4131 authpriv, daemon, local0 ... local7 or user.
4133 And priority could be any of:
4134 err, warning, notice, info, debug.
4136 udp To send each log line as text data to a UDP receiver.
4137 Place: The destination host name or IP and port.
4138 Place Format: //host:port
4140 tcp To send each log line as text data to a TCP receiver.
4141 Lines may be accumulated before sending (see buffered_logs).
4142 Place: The destination host name or IP and port.
4143 Place Format: //host:port
4146 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4152 LOC: Config.Log.icaplogs
4155 ICAP log files record ICAP transaction summaries, one line per
4158 The icap_log option format is:
4159 icap_log <filepath> [<logformat name> [acl acl ...]]
4160 icap_log none [acl acl ...]]
4162 Please see access_log option documentation for details. The two
4163 kinds of logs share the overall configuration approach and many
4166 ICAP processing of a single HTTP message or transaction may
4167 require multiple ICAP transactions. In such cases, multiple
4168 ICAP transaction log lines will correspond to a single access
4171 ICAP log uses logformat codes that make sense for an ICAP
4172 transaction. Header-related codes are applied to the HTTP header
4173 embedded in an ICAP server response, with the following caveats:
4174 For REQMOD, there is no HTTP response header unless the ICAP
4175 server performed request satisfaction. For RESPMOD, the HTTP
4176 request header is the header sent to the ICAP server. For
4177 OPTIONS, there are no HTTP headers.
4179 The following format codes are also available for ICAP logs:
4181 icap::<A ICAP server IP address. Similar to <A.
4183 icap::<service_name ICAP service name from the icap_service
4184 option in Squid configuration file.
4186 icap::ru ICAP Request-URI. Similar to ru.
4188 icap::rm ICAP request method (REQMOD, RESPMOD, or
4189 OPTIONS). Similar to existing rm.
4191 icap::>st Bytes sent to the ICAP server (TCP payload
4192 only; i.e., what Squid writes to the socket).
4194 icap::<st Bytes received from the ICAP server (TCP
4195 payload only; i.e., what Squid reads from
4198 icap::<bs Number of message body bytes received from the
4199 ICAP server. ICAP message body, if any, usually
4200 includes encapsulated HTTP message headers and
4201 possibly encapsulated HTTP message body. The
4202 HTTP body part is dechunked before its size is
4205 icap::tr Transaction response time (in
4206 milliseconds). The timer starts when
4207 the ICAP transaction is created and
4208 stops when the transaction is completed.
4211 icap::tio Transaction I/O time (in milliseconds). The
4212 timer starts when the first ICAP request
4213 byte is scheduled for sending. The timers
4214 stops when the last byte of the ICAP response
4217 icap::to Transaction outcome: ICAP_ERR* for all
4218 transaction errors, ICAP_OPT for OPTION
4219 transactions, ICAP_ECHO for 204
4220 responses, ICAP_MOD for message
4221 modification, and ICAP_SAT for request
4222 satisfaction. Similar to Ss.
4224 icap::Hs ICAP response status code. Similar to Hs.
4226 icap::>h ICAP request header(s). Similar to >h.
4228 icap::<h ICAP response header(s). Similar to <h.
4230 The default ICAP log format, which can be used without an explicit
4231 definition, is called icap_squid:
4233 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4235 See also: logformat, log_icap, and %adapt::<last_h
4238 NAME: logfile_daemon
4240 DEFAULT: @DEFAULT_LOGFILED@
4241 LOC: Log::TheConfig.logfile_daemon
4243 Specify the path to the logfile-writing daemon. This daemon is
4244 used to write the access and store logs, if configured.
4246 Squid sends a number of commands to the log daemon:
4247 L<data>\n - logfile data
4252 r<n>\n - set rotate count to <n>
4253 b<n>\n - 1 = buffer output, 0 = don't buffer output
4255 No responses is expected.
4261 Remove this line. Use acls with access_log directives to control access logging
4267 Remove this line. Use acls with icap_log directives to control icap logging
4270 NAME: stats_collection
4272 LOC: Config.accessList.stats_collection
4274 DEFAULT_DOC: Allow logging for all transactions.
4275 COMMENT: allow|deny acl acl...
4277 This options allows you to control which requests gets accounted
4278 in performance counters.
4280 This clause only supports fast acl types.
4281 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4284 NAME: cache_store_log
4287 LOC: Config.Log.store
4289 Logs the activities of the storage manager. Shows which
4290 objects are ejected from the cache, and which objects are
4291 saved and for how long.
4292 There are not really utilities to analyze this data, so you can safely
4293 disable it (the default).
4295 Store log uses modular logging outputs. See access_log for the list
4296 of modules supported.
4299 cache_store_log stdio:@DEFAULT_STORE_LOG@
4300 cache_store_log daemon:@DEFAULT_STORE_LOG@
4303 NAME: cache_swap_state cache_swap_log
4305 LOC: Config.Log.swap
4307 DEFAULT_DOC: Store the journal inside its cache_dir
4309 Location for the cache "swap.state" file. This index file holds
4310 the metadata of objects saved on disk. It is used to rebuild
4311 the cache during startup. Normally this file resides in each
4312 'cache_dir' directory, but you may specify an alternate
4313 pathname here. Note you must give a full filename, not just
4314 a directory. Since this is the index for the whole object
4315 list you CANNOT periodically rotate it!
4317 If %s can be used in the file name it will be replaced with a
4318 a representation of the cache_dir name where each / is replaced
4319 with '.'. This is needed to allow adding/removing cache_dir
4320 lines when cache_swap_log is being used.
4322 If have more than one 'cache_dir', and %s is not used in the name
4323 these swap logs will have names such as:
4329 The numbered extension (which is added automatically)
4330 corresponds to the order of the 'cache_dir' lines in this
4331 configuration file. If you change the order of the 'cache_dir'
4332 lines in this file, these index files will NOT correspond to
4333 the correct 'cache_dir' entry (unless you manually rename
4334 them). We recommend you do NOT use this option. It is
4335 better to keep these index files in each 'cache_dir' directory.
4338 NAME: logfile_rotate
4341 LOC: Config.Log.rotateNumber
4343 Specifies the number of logfile rotations to make when you
4344 type 'squid -k rotate'. The default is 10, which will rotate
4345 with extensions 0 through 9. Setting logfile_rotate to 0 will
4346 disable the file name rotation, but the logfiles are still closed
4347 and re-opened. This will enable you to rename the logfiles
4348 yourself just before sending the rotate signal.
4350 Note, the 'squid -k rotate' command normally sends a USR1
4351 signal to the running squid process. In certain situations
4352 (e.g. on Linux with Async I/O), USR1 is used for other
4353 purposes, so -k rotate uses another signal. It is best to get
4354 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4357 Note, from Squid-3.1 this option is only a default for cache.log,
4358 that log can be rotated separately by using debug_options.
4361 NAME: emulate_httpd_log
4364 Replace this with an access_log directive using the format 'common' or 'combined'.
4367 NAME: log_ip_on_direct
4370 Remove this option from your config. To log server or peer names use %<A in the log format.
4375 DEFAULT: @DEFAULT_MIME_TABLE@
4376 LOC: Config.mimeTablePathname
4378 Path to Squid's icon configuration file.
4380 You shouldn't need to change this, but the default file contains
4381 examples and formatting information if you do.
4387 LOC: Config.onoff.log_mime_hdrs
4390 The Cache can record both the request and the response MIME
4391 headers for each HTTP transaction. The headers are encoded
4392 safely and will appear as two bracketed fields at the end of
4393 the access log (for either the native or httpd-emulated log
4394 formats). To enable this logging set log_mime_hdrs to 'on'.
4400 Replace this with an access_log directive using the format 'useragent'.
4403 NAME: referer_log referrer_log
4406 Replace this with an access_log directive using the format 'referrer'.
4411 DEFAULT: @DEFAULT_PID_FILE@
4412 LOC: Config.pidFilename
4414 A filename to write the process-id to. To disable, enter "none".
4420 Remove this option from your config. To log FQDN use %>A in the log format.
4423 NAME: client_netmask
4425 LOC: Config.Addrs.client_netmask
4427 DEFAULT_DOC: Log full client IP address
4429 A netmask for client addresses in logfiles and cachemgr output.
4430 Change this to protect the privacy of your cache clients.
4431 A netmask of 255.255.255.0 will log all IP's in that range with
4432 the last digit set to '0'.
4438 Use a regular access.log with ACL limiting it to MISS events.
4441 NAME: strip_query_terms
4443 LOC: Config.onoff.strip_query_terms
4446 By default, Squid strips query terms from requested URLs before
4447 logging. This protects your user's privacy and reduces log size.
4449 When investigating HIT/MISS or other caching behaviour you
4450 will need to disable this to see the full URL used by Squid.
4457 LOC: Config.onoff.buffered_logs
4459 Whether to write/send access_log records ASAP or accumulate them and
4460 then write/send them in larger chunks. Buffering may improve
4461 performance because it decreases the number of I/Os. However,
4462 buffering increases the delay before log records become available to
4463 the final recipient (e.g., a disk file or logging daemon) and,
4464 hence, increases the risk of log records loss.
4466 Note that even when buffered_logs are off, Squid may have to buffer
4467 records if it cannot write/send them immediately due to pending I/Os
4468 (e.g., the I/O writing the previous log record) or connectivity loss.
4470 Currently honored by 'daemon' and 'tcp' access_log modules only.
4473 NAME: netdb_filename
4475 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4476 LOC: Config.netdbFilename
4479 Where Squid stores it's netdb journal.
4480 When enabled this journal preserves netdb state between restarts.
4482 To disable, enter "none".
4486 OPTIONS FOR TROUBLESHOOTING
4487 -----------------------------------------------------------------------------
4492 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4493 LOC: Debug::cache_log
4495 Squid administrative logging file.
4497 This is where general information about Squid behavior goes. You can
4498 increase the amount of data logged to this file and how often it is
4499 rotated with "debug_options"
4505 DEFAULT_DOC: Log all critical and important messages.
4506 LOC: Debug::debugOptions
4508 Logging options are set as section,level where each source file
4509 is assigned a unique section. Lower levels result in less
4510 output, Full debugging (level 9) can result in a very large
4511 log file, so be careful.
4513 The magic word "ALL" sets debugging levels for all sections.
4514 The default is to run with "ALL,1" to record important warnings.
4516 The rotate=N option can be used to keep more or less of these logs
4517 than would otherwise be kept by logfile_rotate.
4518 For most uses a single log should be enough to monitor current
4519 events affecting Squid.
4524 LOC: Config.coredump_dir
4525 DEFAULT_IF_NONE: none
4526 DEFAULT_DOC: Use the directory from where Squid was started.
4528 By default Squid leaves core files in the directory from where
4529 it was started. If you set 'coredump_dir' to a directory
4530 that exists, Squid will chdir() to that directory at startup
4531 and coredump files will be left there.
4535 # Leave coredumps in the first cache dir
4536 coredump_dir @DEFAULT_SWAP_DIR@
4542 OPTIONS FOR FTP GATEWAYING
4543 -----------------------------------------------------------------------------
4549 LOC: Config.Ftp.anon_user
4551 If you want the anonymous login password to be more informative
4552 (and enable the use of picky FTP servers), set this to something
4553 reasonable for your domain, like wwwuser@somewhere.net
4555 The reason why this is domainless by default is the
4556 request can be made on the behalf of a user in any domain,
4557 depending on how the cache is used.
4558 Some FTP server also validate the email address is valid
4559 (for example perl.com).
4565 LOC: Config.Ftp.passive
4567 If your firewall does not allow Squid to use passive
4568 connections, turn off this option.
4570 Use of ftp_epsv_all option requires this to be ON.
4576 LOC: Config.Ftp.epsv_all
4578 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4580 NATs may be able to put the connection on a "fast path" through the
4581 translator, as the EPRT command will never be used and therefore,
4582 translation of the data portion of the segments will never be needed.
4584 When a client only expects to do two-way FTP transfers this may be
4586 If squid finds that it must do a three-way FTP transfer after issuing
4587 an EPSV ALL command, the FTP session will fail.
4589 If you have any doubts about this option do not use it.
4590 Squid will nicely attempt all other connection methods.
4592 Requires ftp_passive to be ON (default) for any effect.
4598 LOC: Config.accessList.ftp_epsv
4600 FTP Protocol extensions permit the use of a special "EPSV" command.
4602 NATs may be able to put the connection on a "fast path" through the
4603 translator using EPSV, as the EPRT command will never be used
4604 and therefore, translation of the data portion of the segments
4605 will never be needed.
4607 EPSV is often required to interoperate with FTP servers on IPv6
4608 networks. On the other hand, it may break some IPv4 servers.
4610 By default, EPSV may try EPSV with any FTP server. To fine tune
4611 that decision, you may restrict EPSV to certain clients or servers
4614 ftp_epsv allow|deny al1 acl2 ...
4616 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4618 Only fast ACLs are supported.
4619 Requires ftp_passive to be ON (default) for any effect.
4625 LOC: Config.Ftp.eprt
4627 FTP Protocol extensions permit the use of a special "EPRT" command.
4629 This extension provides a protocol neutral alternative to the
4630 IPv4-only PORT command. When supported it enables active FTP data
4631 channels over IPv6 and efficient NAT handling.
4633 Turning this OFF will prevent EPRT being attempted and will skip
4634 straight to using PORT for IPv4 servers.
4636 Some devices are known to not handle this extension correctly and
4637 may result in crashes. Devices which suport EPRT enough to fail
4638 cleanly will result in Squid attempting PORT anyway. This directive
4639 should only be disabled when EPRT results in device failures.
4641 WARNING: Doing so will convert Squid back to the old behavior with all
4642 the related problems with external NAT devices/layers and IPv4-only FTP.
4645 NAME: ftp_sanitycheck
4648 LOC: Config.Ftp.sanitycheck
4650 For security and data integrity reasons Squid by default performs
4651 sanity checks of the addresses of FTP data connections ensure the
4652 data connection is to the requested server. If you need to allow
4653 FTP connections to servers using another IP address for the data
4654 connection turn this off.
4657 NAME: ftp_telnet_protocol
4660 LOC: Config.Ftp.telnet
4662 The FTP protocol is officially defined to use the telnet protocol
4663 as transport channel for the control connection. However, many
4664 implementations are broken and does not respect this aspect of
4667 If you have trouble accessing files with ASCII code 255 in the
4668 path or similar problems involving this ASCII code you can
4669 try setting this directive to off. If that helps, report to the
4670 operator of the FTP server in question that their FTP server
4671 is broken and does not follow the FTP standard.
4675 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4676 -----------------------------------------------------------------------------
4681 DEFAULT: @DEFAULT_DISKD@
4682 LOC: Config.Program.diskd
4684 Specify the location of the diskd executable.
4685 Note this is only useful if you have compiled in
4686 diskd as one of the store io modules.
4689 NAME: unlinkd_program
4692 DEFAULT: @DEFAULT_UNLINKD@
4693 LOC: Config.Program.unlinkd
4695 Specify the location of the executable for file deletion process.
4698 NAME: pinger_program
4700 DEFAULT: @DEFAULT_PINGER@
4701 LOC: Config.pinger.program
4704 Specify the location of the executable for the pinger process.
4710 LOC: Config.pinger.enable
4713 Control whether the pinger is active at run-time.
4714 Enables turning ICMP pinger on and off with a simple
4715 squid -k reconfigure.
4720 OPTIONS FOR URL REWRITING
4721 -----------------------------------------------------------------------------
4724 NAME: url_rewrite_program redirect_program
4726 LOC: Config.Program.redirect
4729 Specify the location of the executable URL rewriter to use.
4730 Since they can perform almost any function there isn't one included.
4732 For each requested URL, the rewriter will receive on line with the format
4734 [channel-ID <SP>] URL [<SP> extras]<NL>
4737 After processing the request the helper must reply using the following format:
4739 [channel-ID <SP>] result [<SP> kv-pairs]
4741 The result code can be:
4743 OK status=30N url="..."
4744 Redirect the URL to the one supplied in 'url='.
4745 'status=' is optional and contains the status code to send
4746 the client in Squids HTTP response. It must be one of the
4747 HTTP redirect status codes: 301, 302, 303, 307, 308.
4748 When no status is given Squid will use 302.
4750 OK rewrite-url="..."
4751 Rewrite the URL to the one supplied in 'rewrite-url='.
4752 The new URL is fetched directly by Squid and returned to
4753 the client as the response to its request.
4756 When neither of url= and rewrite-url= are sent Squid does
4760 Do not change the URL.
4763 An internal error occurred in the helper, preventing
4764 a result being identified. The 'message=' key name is
4765 reserved for delivering a log message.
4768 In the future, the interface protocol will be extended with
4769 key=value pairs ("kv-pairs" shown above). Helper programs
4770 should be prepared to receive and possibly ignore additional
4771 whitespace-separated tokens on each input line.
4773 When using the concurrency= option the protocol is changed by
4774 introducing a query channel tag in front of the request/response.
4775 The query channel tag is a number between 0 and concurrency-1.
4776 This value must be echoed back unchanged to Squid as the first part
4777 of the response relating to its request.
4779 WARNING: URL re-writing ability should be avoided whenever possible.
4780 Use the URL redirect form of response instead.
4782 Re-write creates a difference in the state held by the client
4783 and server. Possibly causing confusion when the server response
4784 contains snippets of its view state. Embeded URLs, response
4785 and content Location headers, etc. are not re-written by this
4788 By default, a URL rewriter is not used.
4791 NAME: url_rewrite_children redirect_children
4792 TYPE: HelperChildConfig
4793 DEFAULT: 20 startup=0 idle=1 concurrency=0
4794 LOC: Config.redirectChildren
4796 The maximum number of redirector processes to spawn. If you limit
4797 it too few Squid will have to wait for them to process a backlog of
4798 URLs, slowing it down. If you allow too many they will use RAM
4799 and other system resources noticably.
4801 The startup= and idle= options allow some measure of skew in your
4806 Sets a minimum of how many processes are to be spawned when Squid
4807 starts or reconfigures. When set to zero the first request will
4808 cause spawning of the first child process to handle it.
4810 Starting too few will cause an initial slowdown in traffic as Squid
4811 attempts to simultaneously spawn enough processes to cope.
4815 Sets a minimum of how many processes Squid is to try and keep available
4816 at all times. When traffic begins to rise above what the existing
4817 processes can handle this many more will be spawned up to the maximum
4818 configured. A minimum setting of 1 is required.
4822 The number of requests each redirector helper can handle in
4823 parallel. Defaults to 0 which indicates the redirector
4824 is a old-style single threaded redirector.
4826 When this directive is set to a value >= 1 then the protocol
4827 used to communicate with the helper is modified to include
4828 an ID in front of the request/response. The ID from the request
4829 must be echoed back with the response to that request.
4832 NAME: url_rewrite_host_header redirect_rewrites_host_header
4835 LOC: Config.onoff.redir_rewrites_host
4837 To preserve same-origin security policies in browsers and
4838 prevent Host: header forgery by redirectors Squid rewrites
4839 any Host: header in redirected requests.
4841 If you are running an accelerator this may not be a wanted
4842 effect of a redirector. This directive enables you disable
4843 Host: alteration in reverse-proxy traffic.
4845 WARNING: Entries are cached on the result of the URL rewriting
4846 process, so be careful if you have domain-virtual hosts.
4848 WARNING: Squid and other software verifies the URL and Host
4849 are matching, so be careful not to relay through other proxies
4850 or inspecting firewalls with this disabled.
4853 NAME: url_rewrite_access redirector_access
4856 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4857 LOC: Config.accessList.redirector
4859 If defined, this access list specifies which requests are
4860 sent to the redirector processes.
4862 This clause supports both fast and slow acl types.
4863 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4866 NAME: url_rewrite_bypass redirector_bypass
4868 LOC: Config.onoff.redirector_bypass
4871 When this is 'on', a request will not go through the
4872 redirector if all the helpers are busy. If this is 'off'
4873 and the redirector queue grows too large, Squid will exit
4874 with a FATAL error and ask you to increase the number of
4875 redirectors. You should only enable this if the redirectors
4876 are not critical to your caching system. If you use
4877 redirectors for access control, and you enable this option,
4878 users may have access to pages they should not
4879 be allowed to request.
4882 NAME: url_rewrite_extras
4883 TYPE: TokenOrQuotedString
4884 LOC: Config.redirector_extras
4885 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4887 Specifies a string to be append to request line format for the
4888 rewriter helper. "Quoted" format values may contain spaces and
4889 logformat %macros. In theory, any logformat %macro can be used.
4890 In practice, a %macro expands as a dash (-) if the helper request is
4891 sent before the required macro information is available to Squid.
4895 OPTIONS FOR STORE ID
4896 -----------------------------------------------------------------------------
4899 NAME: store_id_program storeurl_rewrite_program
4901 LOC: Config.Program.store_id
4904 Specify the location of the executable StoreID helper to use.
4905 Since they can perform almost any function there isn't one included.
4907 For each requested URL, the helper will receive one line with the format
4909 [channel-ID <SP>] URL [<SP> extras]<NL>
4912 After processing the request the helper must reply using the following format:
4914 [channel-ID <SP>] result [<SP> kv-pairs]
4916 The result code can be:
4919 Use the StoreID supplied in 'store-id='.
4922 The default is to use HTTP request URL as the store ID.
4925 An internal error occured in the helper, preventing
4926 a result being identified.
4929 Helper programs should be prepared to receive and possibly ignore
4930 additional whitespace-separated tokens on each input line.
4932 When using the concurrency= option the protocol is changed by
4933 introducing a query channel tag in front of the request/response.
4934 The query channel tag is a number between 0 and concurrency-1.
4935 This value must be echoed back unchanged to Squid as the first part
4936 of the response relating to its request.
4938 NOTE: when using StoreID refresh_pattern will apply to the StoreID
4939 returned from the helper and not the URL.
4941 WARNING: Wrong StoreID value returned by a careless helper may result
4942 in the wrong cached response returned to the user.
4944 By default, a StoreID helper is not used.
4947 NAME: store_id_extras
4948 TYPE: TokenOrQuotedString
4949 LOC: Config.storeId_extras
4950 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4952 Specifies a string to be append to request line format for the
4953 StoreId helper. "Quoted" format values may contain spaces and
4954 logformat %macros. In theory, any logformat %macro can be used.
4955 In practice, a %macro expands as a dash (-) if the helper request is
4956 sent before the required macro information is available to Squid.
4959 NAME: store_id_children storeurl_rewrite_children
4960 TYPE: HelperChildConfig
4961 DEFAULT: 20 startup=0 idle=1 concurrency=0
4962 LOC: Config.storeIdChildren
4964 The maximum number of StoreID helper processes to spawn. If you limit
4965 it too few Squid will have to wait for them to process a backlog of
4966 requests, slowing it down. If you allow too many they will use RAM
4967 and other system resources noticably.
4969 The startup= and idle= options allow some measure of skew in your
4974 Sets a minimum of how many processes are to be spawned when Squid
4975 starts or reconfigures. When set to zero the first request will
4976 cause spawning of the first child process to handle it.
4978 Starting too few will cause an initial slowdown in traffic as Squid
4979 attempts to simultaneously spawn enough processes to cope.
4983 Sets a minimum of how many processes Squid is to try and keep available
4984 at all times. When traffic begins to rise above what the existing
4985 processes can handle this many more will be spawned up to the maximum
4986 configured. A minimum setting of 1 is required.
4990 The number of requests each storeID helper can handle in
4991 parallel. Defaults to 0 which indicates the helper
4992 is a old-style single threaded program.
4994 When this directive is set to a value >= 1 then the protocol
4995 used to communicate with the helper is modified to include
4996 an ID in front of the request/response. The ID from the request
4997 must be echoed back with the response to that request.
5000 NAME: store_id_access storeurl_rewrite_access
5003 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5004 LOC: Config.accessList.store_id
5006 If defined, this access list specifies which requests are
5007 sent to the StoreID processes. By default all requests
5010 This clause supports both fast and slow acl types.
5011 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5014 NAME: store_id_bypass storeurl_rewrite_bypass
5016 LOC: Config.onoff.store_id_bypass
5019 When this is 'on', a request will not go through the
5020 helper if all helpers are busy. If this is 'off'
5021 and the helper queue grows too large, Squid will exit
5022 with a FATAL error and ask you to increase the number of
5023 helpers. You should only enable this if the helperss
5024 are not critical to your caching system. If you use
5025 helpers for critical caching components, and you enable this
5026 option, users may not get objects from cache.
5030 OPTIONS FOR TUNING THE CACHE
5031 -----------------------------------------------------------------------------
5034 NAME: cache no_cache
5037 DEFAULT_DOC: By default, this directive is unused and has no effect.
5038 LOC: Config.accessList.noCache
5040 Requests denied by this directive will not be served from the cache
5041 and their responses will not be stored in the cache. This directive
5042 has no effect on other transactions and on already cached responses.
5044 This clause supports both fast and slow acl types.
5045 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5047 This and the two other similar caching directives listed below are
5048 checked at different transaction processing stages, have different
5049 access to response information, affect different cache operations,
5050 and differ in slow ACLs support:
5052 * cache: Checked before Squid makes a hit/miss determination.
5053 No access to reply information!
5054 Denies both serving a hit and storing a miss.
5055 Supports both fast and slow ACLs.
5056 * send_hit: Checked after a hit was detected.
5057 Has access to reply (hit) information.
5058 Denies serving a hit only.
5059 Supports fast ACLs only.
5060 * store_miss: Checked before storing a cachable miss.
5061 Has access to reply (miss) information.
5062 Denies storing a miss only.
5063 Supports fast ACLs only.
5065 If you are not sure which of the three directives to use, apply the
5066 following decision logic:
5068 * If your ACL(s) are of slow type _and_ need response info, redesign.
5069 Squid does not support that particular combination at this time.
5071 * If your directive ACL(s) are of slow type, use "cache"; and/or
5072 * if your directive ACL(s) need no response info, use "cache".
5074 * If you do not want the response cached, use store_miss; and/or
5075 * if you do not want a hit on a cached response, use send_hit.
5081 DEFAULT_DOC: By default, this directive is unused and has no effect.
5082 LOC: Config.accessList.sendHit
5084 Responses denied by this directive will not be served from the cache
5085 (but may still be cached, see store_miss). This directive has no
5086 effect on the responses it allows and on the cached objects.
5088 Please see the "cache" directive for a summary of differences among
5089 store_miss, send_hit, and cache directives.
5091 Unlike the "cache" directive, send_hit only supports fast acl
5092 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5096 # apply custom Store ID mapping to some URLs
5097 acl MapMe dstdomain .c.example.com
5098 store_id_program ...
5099 store_id_access allow MapMe
5101 # but prevent caching of special responses
5102 # such as 302 redirects that cause StoreID loops
5103 acl Ordinary http_status 200-299
5104 store_miss deny MapMe !Ordinary
5106 # and do not serve any previously stored special responses
5107 # from the cache (in case they were already cached before
5108 # the above store_miss rule was in effect).
5109 send_hit deny MapMe !Ordinary
5115 DEFAULT_DOC: By default, this directive is unused and has no effect.
5116 LOC: Config.accessList.storeMiss
5118 Responses denied by this directive will not be cached (but may still
5119 be served from the cache, see send_hit). This directive has no
5120 effect on the responses it allows and on the already cached responses.
5122 Please see the "cache" directive for a summary of differences among
5123 store_miss, send_hit, and cache directives. See the
5124 send_hit directive for a usage example.
5126 Unlike the "cache" directive, store_miss only supports fast acl
5127 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5133 LOC: Config.maxStale
5136 This option puts an upper limit on how stale content Squid
5137 will serve from the cache if cache validation fails.
5138 Can be overriden by the refresh_pattern max-stale option.
5141 NAME: refresh_pattern
5142 TYPE: refreshpattern
5146 usage: refresh_pattern [-i] regex min percent max [options]
5148 By default, regular expressions are CASE-SENSITIVE. To make
5149 them case-insensitive, use the -i option.
5151 'Min' is the time (in minutes) an object without an explicit
5152 expiry time should be considered fresh. The recommended
5153 value is 0, any higher values may cause dynamic applications
5154 to be erroneously cached unless the application designer
5155 has taken the appropriate actions.
5157 'Percent' is a percentage of the objects age (time since last
5158 modification age) an object without explicit expiry time
5159 will be considered fresh.
5161 'Max' is an upper limit on how long objects without an explicit
5162 expiry time will be considered fresh.
5164 options: override-expire
5169 ignore-must-revalidate
5176 override-expire enforces min age even if the server
5177 sent an explicit expiry time (e.g., with the
5178 Expires: header or Cache-Control: max-age). Doing this
5179 VIOLATES the HTTP standard. Enabling this feature
5180 could make you liable for problems which it causes.
5182 Note: override-expire does not enforce staleness - it only extends
5183 freshness / min. If the server returns a Expires time which
5184 is longer than your max time, Squid will still consider
5185 the object fresh for that period of time.
5187 override-lastmod enforces min age even on objects
5188 that were modified recently.
5190 reload-into-ims changes a client no-cache or ``reload''
5191 request for a cached entry into a conditional request using
5192 If-Modified-Since and/or If-None-Match headers, provided the
5193 cached entry has a Last-Modified and/or a strong ETag header.
5194 Doing this VIOLATES the HTTP standard. Enabling this feature
5195 could make you liable for problems which it causes.
5197 ignore-reload ignores a client no-cache or ``reload''
5198 header. Doing this VIOLATES the HTTP standard. Enabling
5199 this feature could make you liable for problems which
5202 ignore-no-store ignores any ``Cache-control: no-store''
5203 headers received from a server. Doing this VIOLATES
5204 the HTTP standard. Enabling this feature could make you
5205 liable for problems which it causes.
5207 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
5208 headers received from a server. Doing this VIOLATES
5209 the HTTP standard. Enabling this feature could make you
5210 liable for problems which it causes.
5212 ignore-private ignores any ``Cache-control: private''
5213 headers received from a server. Doing this VIOLATES
5214 the HTTP standard. Enabling this feature could make you
5215 liable for problems which it causes.
5217 ignore-auth caches responses to requests with authorization,
5218 as if the originserver had sent ``Cache-control: public''
5219 in the response header. Doing this VIOLATES the HTTP standard.
5220 Enabling this feature could make you liable for problems which
5223 refresh-ims causes squid to contact the origin server
5224 when a client issues an If-Modified-Since request. This
5225 ensures that the client will receive an updated version
5226 if one is available.
5228 store-stale stores responses even if they don't have explicit
5229 freshness or a validator (i.e., Last-Modified or an ETag)
5230 present, or if they're already stale. By default, Squid will
5231 not cache such responses because they usually can't be
5232 reused. Note that such responses will be stale by default.
5234 max-stale=NN provide a maximum staleness factor. Squid won't
5235 serve objects more stale than this even if it failed to
5236 validate the object. Default: use the max_stale global limit.
5238 Basically a cached object is:
5240 FRESH if expires < now, else STALE
5242 FRESH if lm-factor < percent, else STALE
5246 The refresh_pattern lines are checked in the order listed here.
5247 The first entry which matches is used. If none of the entries
5248 match the default will be used.
5250 Note, you must uncomment all the default lines if you want
5251 to change one. The default setting is only active if none is
5257 # Add any of your own refresh_pattern entries above these.
5259 refresh_pattern ^ftp: 1440 20% 10080
5260 refresh_pattern ^gopher: 1440 0% 1440
5261 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5262 refresh_pattern . 0 20% 4320
5266 NAME: quick_abort_min
5270 LOC: Config.quickAbort.min
5273 NAME: quick_abort_max
5277 LOC: Config.quickAbort.max
5280 NAME: quick_abort_pct
5284 LOC: Config.quickAbort.pct
5286 The cache by default continues downloading aborted requests
5287 which are almost completed (less than 16 KB remaining). This
5288 may be undesirable on slow (e.g. SLIP) links and/or very busy
5289 caches. Impatient users may tie up file descriptors and
5290 bandwidth by repeatedly requesting and immediately aborting
5293 When the user aborts a request, Squid will check the
5294 quick_abort values to the amount of data transfered until
5297 If the transfer has less than 'quick_abort_min' KB remaining,
5298 it will finish the retrieval.
5300 If the transfer has more than 'quick_abort_max' KB remaining,
5301 it will abort the retrieval.
5303 If more than 'quick_abort_pct' of the transfer has completed,
5304 it will finish the retrieval.
5306 If you do not want any retrieval to continue after the client
5307 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5310 If you want retrievals to always continue if they are being
5311 cached set 'quick_abort_min' to '-1 KB'.
5314 NAME: read_ahead_gap
5315 COMMENT: buffer-size
5317 LOC: Config.readAheadGap
5320 The amount of data the cache will buffer ahead of what has been
5321 sent to the client when retrieving an object from another server.
5325 IFDEF: USE_HTTP_VIOLATIONS
5328 LOC: Config.negativeTtl
5331 Set the Default Time-to-Live (TTL) for failed requests.
5332 Certain types of failures (such as "connection refused" and
5333 "404 Not Found") are able to be negatively-cached for a short time.
5334 Modern web servers should provide Expires: header, however if they
5335 do not this can provide a minimum TTL.
5336 The default is not to cache errors with unknown expiry details.
5338 Note that this is different from negative caching of DNS lookups.
5340 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5341 this feature could make you liable for problems which it
5345 NAME: positive_dns_ttl
5348 LOC: Config.positiveDnsTtl
5351 Upper limit on how long Squid will cache positive DNS responses.
5352 Default is 6 hours (360 minutes). This directive must be set
5353 larger than negative_dns_ttl.
5356 NAME: negative_dns_ttl
5359 LOC: Config.negativeDnsTtl
5362 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5363 This also sets the lower cache limit on positive lookups.
5364 Minimum value is 1 second, and it is not recommendable to go
5365 much below 10 seconds.
5368 NAME: range_offset_limit
5369 COMMENT: size [acl acl...]
5371 LOC: Config.rangeOffsetLimit
5374 usage: (size) [units] [[!]aclname]
5376 Sets an upper limit on how far (number of bytes) into the file
5377 a Range request may be to cause Squid to prefetch the whole file.
5378 If beyond this limit, Squid forwards the Range request as it is and
5379 the result is NOT cached.
5381 This is to stop a far ahead range request (lets say start at 17MB)
5382 from making Squid fetch the whole object up to that point before
5383 sending anything to the client.
5385 Multiple range_offset_limit lines may be specified, and they will
5386 be searched from top to bottom on each request until a match is found.
5387 The first match found will be used. If no line matches a request, the
5388 default limit of 0 bytes will be used.
5390 'size' is the limit specified as a number of units.
5392 'units' specifies whether to use bytes, KB, MB, etc.
5393 If no units are specified bytes are assumed.
5395 A size of 0 causes Squid to never fetch more than the
5396 client requested. (default)
5398 A size of 'none' causes Squid to always fetch the object from the
5399 beginning so it may cache the result. (2.0 style)
5401 'aclname' is the name of a defined ACL.
5403 NP: Using 'none' as the byte value here will override any quick_abort settings
5404 that may otherwise apply to the range request. The range request will
5405 be fully fetched from start to finish regardless of the client
5406 actions. This affects bandwidth usage.
5409 NAME: minimum_expiry_time
5412 LOC: Config.minimum_expiry_time
5415 The minimum caching time according to (Expires - Date)
5416 headers Squid honors if the object can't be revalidated.
5417 The default is 60 seconds.
5419 In reverse proxy environments it might be desirable to honor
5420 shorter object lifetimes. It is most likely better to make
5421 your server return a meaningful Last-Modified header however.
5423 In ESI environments where page fragments often have short
5424 lifetimes, this will often be best set to 0.
5427 NAME: store_avg_object_size
5431 LOC: Config.Store.avgObjectSize
5433 Average object size, used to estimate number of objects your
5434 cache can hold. The default is 13 KB.
5436 This is used to pre-seed the cache index memory allocation to
5437 reduce expensive reallocate operations while handling clients
5438 traffic. Too-large values may result in memory allocation during
5439 peak traffic, too-small values will result in wasted memory.
5441 Check the cache manager 'info' report metrics for the real
5442 object sizes seen by your Squid before tuning this.
5445 NAME: store_objects_per_bucket
5448 LOC: Config.Store.objectsPerBucket
5450 Target number of objects per bucket in the store hash table.
5451 Lowering this value increases the total number of buckets and
5452 also the storage maintenance rate. The default is 20.
5457 -----------------------------------------------------------------------------
5460 NAME: request_header_max_size
5464 LOC: Config.maxRequestHeaderSize
5466 This specifies the maximum size for HTTP headers in a request.
5467 Request headers are usually relatively small (about 512 bytes).
5468 Placing a limit on the request header size will catch certain
5469 bugs (for example with persistent connections) and possibly
5470 buffer-overflow or denial-of-service attacks.
5473 NAME: reply_header_max_size
5477 LOC: Config.maxReplyHeaderSize
5479 This specifies the maximum size for HTTP headers in a reply.
5480 Reply headers are usually relatively small (about 512 bytes).
5481 Placing a limit on the reply header size will catch certain
5482 bugs (for example with persistent connections) and possibly
5483 buffer-overflow or denial-of-service attacks.
5486 NAME: request_body_max_size
5490 DEFAULT_DOC: No limit.
5491 LOC: Config.maxRequestBodySize
5493 This specifies the maximum size for an HTTP request body.
5494 In other words, the maximum size of a PUT/POST request.
5495 A user who attempts to send a request with a body larger
5496 than this limit receives an "Invalid Request" error message.
5497 If you set this parameter to a zero (the default), there will
5498 be no limit imposed.
5500 See also client_request_buffer_max_size for an alternative
5501 limitation on client uploads which can be configured.
5504 NAME: client_request_buffer_max_size
5508 LOC: Config.maxRequestBufferSize
5510 This specifies the maximum buffer size of a client request.
5511 It prevents squid eating too much memory when somebody uploads
5515 NAME: chunked_request_body_max_size
5519 LOC: Config.maxChunkedRequestBodySize
5521 A broken or confused HTTP/1.1 client may send a chunked HTTP
5522 request to Squid. Squid does not have full support for that
5523 feature yet. To cope with such requests, Squid buffers the
5524 entire request and then dechunks request body to create a
5525 plain HTTP/1.0 request with a known content length. The plain
5526 request is then used by the rest of Squid code as usual.
5528 The option value specifies the maximum size of the buffer used
5529 to hold the request before the conversion. If the chunked
5530 request size exceeds the specified limit, the conversion
5531 fails, and the client receives an "unsupported request" error,
5532 as if dechunking was disabled.
5534 Dechunking is enabled by default. To disable conversion of
5535 chunked requests, set the maximum to zero.
5537 Request dechunking feature and this option in particular are a
5538 temporary hack. When chunking requests and responses are fully
5539 supported, there will be no need to buffer a chunked request.
5543 IFDEF: USE_HTTP_VIOLATIONS
5546 DEFAULT_DOC: Obey RFC 2616.
5547 LOC: Config.accessList.brokenPosts
5549 A list of ACL elements which, if matched, causes Squid to send
5550 an extra CRLF pair after the body of a PUT/POST request.
5552 Some HTTP servers has broken implementations of PUT/POST,
5553 and rely on an extra CRLF pair sent by some WWW clients.
5555 Quote from RFC2616 section 4.1 on this matter:
5557 Note: certain buggy HTTP/1.0 client implementations generate an
5558 extra CRLF's after a POST request. To restate what is explicitly
5559 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5560 a request with an extra CRLF.
5562 This clause only supports fast acl types.
5563 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5566 acl buggy_server url_regex ^http://....
5567 broken_posts allow buggy_server
5570 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5573 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5575 LOC: Adaptation::Config::use_indirect_client
5577 Controls whether the indirect client IP address (instead of the direct
5578 client IP address) is passed to adaptation services.
5580 See also: follow_x_forwarded_for adaptation_send_client_ip
5584 IFDEF: USE_HTTP_VIOLATIONS
5588 LOC: Config.onoff.via
5590 If set (default), Squid will include a Via header in requests and
5591 replies as required by RFC2616.
5597 LOC: Config.onoff.ie_refresh
5600 Microsoft Internet Explorer up until version 5.5 Service
5601 Pack 1 has an issue with transparent proxies, wherein it
5602 is impossible to force a refresh. Turning this on provides
5603 a partial fix to the problem, by causing all IMS-REFRESH
5604 requests from older IE versions to check the origin server
5605 for fresh content. This reduces hit ratio by some amount
5606 (~10% in my experience), but allows users to actually get
5607 fresh content when they want it. Note because Squid
5608 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5609 of 5.5 is unchanged from old versions of Squid (i.e. a
5610 forced refresh is impossible). Newer versions of IE will,
5611 hopefully, continue to have the new behavior and will be
5612 handled based on that assumption. This option defaults to
5613 the old Squid behavior, which is better for hit ratios but
5614 worse for clients using IE, if they need to be able to
5615 force fresh content.
5618 NAME: vary_ignore_expire
5621 LOC: Config.onoff.vary_ignore_expire
5624 Many HTTP servers supporting Vary gives such objects
5625 immediate expiry time with no cache-control header
5626 when requested by a HTTP/1.0 client. This option
5627 enables Squid to ignore such expiry times until
5628 HTTP/1.1 is fully implemented.
5630 WARNING: If turned on this may eventually cause some
5631 varying objects not intended for caching to get cached.
5634 NAME: request_entities
5636 LOC: Config.onoff.request_entities
5639 Squid defaults to deny GET and HEAD requests with request entities,
5640 as the meaning of such requests are undefined in the HTTP standard
5641 even if not explicitly forbidden.
5643 Set this directive to on if you have clients which insists
5644 on sending request entities in GET or HEAD requests. But be warned
5645 that there is server software (both proxies and web servers) which
5646 can fail to properly process this kind of request which may make you
5647 vulnerable to cache pollution attacks if enabled.
5650 NAME: request_header_access
5651 IFDEF: USE_HTTP_VIOLATIONS
5652 TYPE: http_header_access
5653 LOC: Config.request_header_access
5655 DEFAULT_DOC: No limits.
5657 Usage: request_header_access header_name allow|deny [!]aclname ...
5659 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5660 this feature could make you liable for problems which it
5663 This option replaces the old 'anonymize_headers' and the
5664 older 'http_anonymizer' option with something that is much
5665 more configurable. A list of ACLs for each header name allows
5666 removal of specific header fields under specific conditions.
5668 This option only applies to outgoing HTTP request headers (i.e.,
5669 headers sent by Squid to the next HTTP hop such as a cache peer
5670 or an origin server). The option has no effect during cache hit
5671 detection. The equivalent adaptation vectoring point in ICAP
5672 terminology is post-cache REQMOD.
5674 The option is applied to individual outgoing request header
5675 fields. For each request header field F, Squid uses the first
5676 qualifying sets of request_header_access rules:
5678 1. Rules with header_name equal to F's name.
5679 2. Rules with header_name 'Other', provided F's name is not
5680 on the hard-coded list of commonly used HTTP header names.
5681 3. Rules with header_name 'All'.
5683 Within that qualifying rule set, rule ACLs are checked as usual.
5684 If ACLs of an "allow" rule match, the header field is allowed to
5685 go through as is. If ACLs of a "deny" rule match, the header is
5686 removed and request_header_replace is then checked to identify
5687 if the removed header has a replacement. If no rules within the
5688 set have matching ACLs, the header field is left as is.
5690 For example, to achieve the same behavior as the old
5691 'http_anonymizer standard' option, you should use:
5693 request_header_access From deny all
5694 request_header_access Referer deny all
5695 request_header_access User-Agent deny all
5697 Or, to reproduce the old 'http_anonymizer paranoid' feature
5700 request_header_access Authorization allow all
5701 request_header_access Proxy-Authorization allow all
5702 request_header_access Cache-Control allow all
5703 request_header_access Content-Length allow all
5704 request_header_access Content-Type allow all
5705 request_header_access Date allow all
5706 request_header_access Host allow all
5707 request_header_access If-Modified-Since allow all
5708 request_header_access Pragma allow all
5709 request_header_access Accept allow all
5710 request_header_access Accept-Charset allow all
5711 request_header_access Accept-Encoding allow all
5712 request_header_access Accept-Language allow all
5713 request_header_access Connection allow all
5714 request_header_access All deny all
5716 HTTP reply headers are controlled with the reply_header_access directive.
5718 By default, all headers are allowed (no anonymizing is performed).
5721 NAME: reply_header_access
5722 IFDEF: USE_HTTP_VIOLATIONS
5723 TYPE: http_header_access
5724 LOC: Config.reply_header_access
5726 DEFAULT_DOC: No limits.
5728 Usage: reply_header_access header_name allow|deny [!]aclname ...
5730 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5731 this feature could make you liable for problems which it
5734 This option only applies to reply headers, i.e., from the
5735 server to the client.
5737 This is the same as request_header_access, but in the other
5738 direction. Please see request_header_access for detailed
5741 For example, to achieve the same behavior as the old
5742 'http_anonymizer standard' option, you should use:
5744 reply_header_access Server deny all
5745 reply_header_access WWW-Authenticate deny all
5746 reply_header_access Link deny all
5748 Or, to reproduce the old 'http_anonymizer paranoid' feature
5751 reply_header_access Allow allow all
5752 reply_header_access WWW-Authenticate allow all
5753 reply_header_access Proxy-Authenticate allow all
5754 reply_header_access Cache-Control allow all
5755 reply_header_access Content-Encoding allow all
5756 reply_header_access Content-Length allow all
5757 reply_header_access Content-Type allow all
5758 reply_header_access Date allow all
5759 reply_header_access Expires allow all
5760 reply_header_access Last-Modified allow all
5761 reply_header_access Location allow all
5762 reply_header_access Pragma allow all
5763 reply_header_access Content-Language allow all
5764 reply_header_access Retry-After allow all
5765 reply_header_access Title allow all
5766 reply_header_access Content-Disposition allow all
5767 reply_header_access Connection allow all
5768 reply_header_access All deny all
5770 HTTP request headers are controlled with the request_header_access directive.
5772 By default, all headers are allowed (no anonymizing is
5776 NAME: request_header_replace header_replace
5777 IFDEF: USE_HTTP_VIOLATIONS
5778 TYPE: http_header_replace
5779 LOC: Config.request_header_access
5782 Usage: request_header_replace header_name message
5783 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5785 This option allows you to change the contents of headers
5786 denied with request_header_access above, by replacing them
5787 with some fixed string.
5789 This only applies to request headers, not reply headers.
5791 By default, headers are removed if denied.
5794 NAME: reply_header_replace
5795 IFDEF: USE_HTTP_VIOLATIONS
5796 TYPE: http_header_replace
5797 LOC: Config.reply_header_access
5800 Usage: reply_header_replace header_name message
5801 Example: reply_header_replace Server Foo/1.0
5803 This option allows you to change the contents of headers
5804 denied with reply_header_access above, by replacing them
5805 with some fixed string.
5807 This only applies to reply headers, not request headers.
5809 By default, headers are removed if denied.
5812 NAME: request_header_add
5813 TYPE: HeaderWithAclList
5814 LOC: Config.request_header_add
5817 Usage: request_header_add field-name field-value acl1 [acl2] ...
5818 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5820 This option adds header fields to outgoing HTTP requests (i.e.,
5821 request headers sent by Squid to the next HTTP hop such as a
5822 cache peer or an origin server). The option has no effect during
5823 cache hit detection. The equivalent adaptation vectoring point
5824 in ICAP terminology is post-cache REQMOD.
5826 Field-name is a token specifying an HTTP header name. If a
5827 standard HTTP header name is used, Squid does not check whether
5828 the new header conflicts with any existing headers or violates
5829 HTTP rules. If the request to be modified already contains a
5830 field with the same name, the old field is preserved but the
5831 header field values are not merged.
5833 Field-value is either a token or a quoted string. If quoted
5834 string format is used, then the surrounding quotes are removed
5835 while escape sequences and %macros are processed.
5837 In theory, all of the logformat codes can be used as %macros.
5838 However, unlike logging (which happens at the very end of
5839 transaction lifetime), the transaction may not yet have enough
5840 information to expand a macro when the new header value is needed.
5841 And some information may already be available to Squid but not yet
5842 committed where the macro expansion code can access it (report
5843 such instances!). The macro will be expanded into a single dash
5844 ('-') in such cases. Not all macros have been tested.
5846 One or more Squid ACLs may be specified to restrict header
5847 injection to matching requests. As always in squid.conf, all
5848 ACLs in an option ACL list must be satisfied for the insertion
5849 to happen. The request_header_add option supports fast ACLs
5858 This option used to log custom information about the master
5859 transaction. For example, an admin may configure Squid to log
5860 which "user group" the transaction belongs to, where "user group"
5861 will be determined based on a set of ACLs and not [just]
5862 authentication information.
5863 Values of key/value pairs can be logged using %{key}note macros:
5865 note key value acl ...
5866 logformat myFormat ... %{key}note ...
5869 NAME: relaxed_header_parser
5870 COMMENT: on|off|warn
5872 LOC: Config.onoff.relaxed_header_parser
5875 In the default "on" setting Squid accepts certain forms
5876 of non-compliant HTTP messages where it is unambiguous
5877 what the sending application intended even if the message
5878 is not correctly formatted. The messages is then normalized
5879 to the correct form when forwarded by Squid.
5881 If set to "warn" then a warning will be emitted in cache.log
5882 each time such HTTP error is encountered.
5884 If set to "off" then such HTTP errors will cause the request
5885 or response to be rejected.
5888 NAME: collapsed_forwarding
5891 LOC: Config.onoff.collapsed_forwarding
5894 This option controls whether Squid is allowed to merge multiple
5895 potentially cachable requests for the same URI before Squid knows
5896 whether the response is going to be cachable.
5898 This feature is disabled by default: Enabling collapsed forwarding
5899 needlessly delays forwarding requests that look cachable (when they are
5900 collapsed) but then need to be forwarded individually anyway because
5901 they end up being for uncachable content. However, in some cases, such
5902 as accelleration of highly cachable content with periodic or groupped
5903 expiration times, the gains from collapsing [large volumes of
5904 simultenous refresh requests] outweigh losses from such delays.
5909 -----------------------------------------------------------------------------
5912 NAME: forward_timeout
5915 LOC: Config.Timeout.forward
5918 This parameter specifies how long Squid should at most attempt in
5919 finding a forwarding path for the request before giving up.
5922 NAME: connect_timeout
5925 LOC: Config.Timeout.connect
5928 This parameter specifies how long to wait for the TCP connect to
5929 the requested server or peer to complete before Squid should
5930 attempt to find another path where to forward the request.
5933 NAME: peer_connect_timeout
5936 LOC: Config.Timeout.peer_connect
5939 This parameter specifies how long to wait for a pending TCP
5940 connection to a peer cache. The default is 30 seconds. You
5941 may also set different timeout values for individual neighbors
5942 with the 'connect-timeout' option on a 'cache_peer' line.
5948 LOC: Config.Timeout.read
5951 The read_timeout is applied on server-side connections. After
5952 each successful read(), the timeout will be extended by this
5953 amount. If no data is read again after this amount of time,
5954 the request is aborted and logged with ERR_READ_TIMEOUT. The
5955 default is 15 minutes.
5961 LOC: Config.Timeout.write
5964 This timeout is tracked for all connections that have data
5965 available for writing and are waiting for the socket to become
5966 ready. After each successful write, the timeout is extended by
5967 the configured amount. If Squid has data to write but the
5968 connection is not ready for the configured duration, the
5969 transaction associated with the connection is terminated. The
5970 default is 15 minutes.
5973 NAME: request_timeout
5975 LOC: Config.Timeout.request
5978 How long to wait for complete HTTP request headers after initial
5979 connection establishment.
5982 NAME: client_idle_pconn_timeout persistent_request_timeout
5984 LOC: Config.Timeout.clientIdlePconn
5987 How long to wait for the next HTTP request on a persistent
5988 client connection after the previous request completes.
5991 NAME: client_lifetime
5994 LOC: Config.Timeout.lifetime
5997 The maximum amount of time a client (browser) is allowed to
5998 remain connected to the cache process. This protects the Cache
5999 from having a lot of sockets (and hence file descriptors) tied up
6000 in a CLOSE_WAIT state from remote clients that go away without
6001 properly shutting down (either because of a network failure or
6002 because of a poor client implementation). The default is one
6005 NOTE: The default value is intended to be much larger than any
6006 client would ever need to be connected to your cache. You
6007 should probably change client_lifetime only as a last resort.
6008 If you seem to have many client connections tying up
6009 filedescriptors, we recommend first tuning the read_timeout,
6010 request_timeout, persistent_request_timeout and quick_abort values.
6013 NAME: half_closed_clients
6015 LOC: Config.onoff.half_closed_clients
6018 Some clients may shutdown the sending side of their TCP
6019 connections, while leaving their receiving sides open. Sometimes,
6020 Squid can not tell the difference between a half-closed and a
6021 fully-closed TCP connection.
6023 By default, Squid will immediately close client connections when
6024 read(2) returns "no more data to read."
6026 Change this option to 'on' and Squid will keep open connections
6027 until a read(2) or write(2) on the socket returns an error.
6028 This may show some benefits for reverse proxies. But if not
6029 it is recommended to leave OFF.
6032 NAME: server_idle_pconn_timeout pconn_timeout
6034 LOC: Config.Timeout.serverIdlePconn
6037 Timeout for idle persistent connections to servers and other
6044 LOC: Ident::TheConfig.timeout
6047 Maximum time to wait for IDENT lookups to complete.
6049 If this is too high, and you enabled IDENT lookups from untrusted
6050 users, you might be susceptible to denial-of-service by having
6051 many ident requests going at once.
6054 NAME: shutdown_lifetime
6057 LOC: Config.shutdownLifetime
6060 When SIGTERM or SIGHUP is received, the cache is put into
6061 "shutdown pending" mode until all active sockets are closed.
6062 This value is the lifetime to set for all open descriptors
6063 during shutdown mode. Any active clients after this many
6064 seconds will receive a 'timeout' message.
6068 ADMINISTRATIVE PARAMETERS
6069 -----------------------------------------------------------------------------
6075 LOC: Config.adminEmail
6077 Email-address of local cache manager who will receive
6078 mail if the cache dies. The default is "webmaster".
6084 LOC: Config.EmailFrom
6086 From: email-address for mail sent when the cache dies.
6087 The default is to use 'squid@unique_hostname'.
6089 See also: unique_hostname directive.
6095 LOC: Config.EmailProgram
6097 Email program used to send mail if the cache dies.
6098 The default is "mail". The specified program must comply
6099 with the standard Unix mail syntax:
6100 mail-program recipient < mailfile
6102 Optional command line options can be specified.
6105 NAME: cache_effective_user
6107 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
6108 LOC: Config.effectiveUser
6110 If you start Squid as root, it will change its effective/real
6111 UID/GID to the user specified below. The default is to change
6112 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
6113 see also; cache_effective_group
6116 NAME: cache_effective_group
6119 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
6120 LOC: Config.effectiveGroup
6122 Squid sets the GID to the effective user's default group ID
6123 (taken from the password file) and supplementary group list
6124 from the groups membership.
6126 If you want Squid to run with a specific GID regardless of
6127 the group memberships of the effective user then set this
6128 to the group (or GID) you want Squid to run as. When set
6129 all other group privileges of the effective user are ignored
6130 and only this GID is effective. If Squid is not started as
6131 root the user starting Squid MUST be member of the specified
6134 This option is not recommended by the Squid Team.
6135 Our preference is for administrators to configure a secure
6136 user account for squid with UID/GID matching system policies.
6139 NAME: httpd_suppress_version_string
6143 LOC: Config.onoff.httpd_suppress_version_string
6145 Suppress Squid version string info in HTTP headers and HTML error pages.
6148 NAME: visible_hostname
6150 LOC: Config.visibleHostname
6152 DEFAULT_DOC: Automatically detect the system host name
6154 If you want to present a special hostname in error messages, etc,
6155 define this. Otherwise, the return value of gethostname()
6156 will be used. If you have multiple caches in a cluster and
6157 get errors about IP-forwarding you must set them to have individual
6158 names with this setting.
6161 NAME: unique_hostname
6163 LOC: Config.uniqueHostname
6165 DEFAULT_DOC: Copy the value from visible_hostname
6167 If you want to have multiple machines with the same
6168 'visible_hostname' you must give each machine a different
6169 'unique_hostname' so forwarding loops can be detected.
6172 NAME: hostname_aliases
6174 LOC: Config.hostnameAliases
6177 A list of other DNS names your cache has.
6185 Minimum umask which should be enforced while the proxy
6186 is running, in addition to the umask set at startup.
6188 For a traditional octal representation of umasks, start
6193 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6194 -----------------------------------------------------------------------------
6196 This section contains parameters for the (optional) cache
6197 announcement service. This service is provided to help
6198 cache administrators locate one another in order to join or
6199 create cache hierarchies.
6201 An 'announcement' message is sent (via UDP) to the registration
6202 service by Squid. By default, the announcement message is NOT
6203 SENT unless you enable it with 'announce_period' below.
6205 The announcement message includes your hostname, plus the
6206 following information from this configuration file:
6212 All current information is processed regularly and made
6213 available on the Web at http://www.ircache.net/Cache/Tracker/.
6216 NAME: announce_period
6218 LOC: Config.Announce.period
6220 DEFAULT_DOC: Announcement messages disabled.
6222 This is how frequently to send cache announcements.
6224 To enable announcing your cache, just set an announce period.
6227 announce_period 1 day
6232 DEFAULT: tracker.ircache.net
6233 LOC: Config.Announce.host
6235 Set the hostname where announce registration messages will be sent.
6237 See also announce_port and announce_file
6243 LOC: Config.Announce.file
6245 The contents of this file will be included in the announce
6246 registration messages.
6252 LOC: Config.Announce.port
6254 Set the port where announce registration messages will be sent.
6256 See also announce_host and announce_file
6260 HTTPD-ACCELERATOR OPTIONS
6261 -----------------------------------------------------------------------------
6264 NAME: httpd_accel_surrogate_id
6267 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6268 LOC: Config.Accel.surrogate_id
6270 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6271 need an identification token to allow control targeting. Because
6272 a farm of surrogates may all perform the same tasks, they may share
6273 an identification token.
6276 NAME: http_accel_surrogate_remote
6280 LOC: Config.onoff.surrogate_is_remote
6282 Remote surrogates (such as those in a CDN) honour the header
6283 "Surrogate-Control: no-store-remote".
6285 Set this to on to have squid behave as a remote surrogate.
6289 IFDEF: USE_SQUID_ESI
6290 COMMENT: libxml2|expat|custom
6292 LOC: ESIParser::Type
6295 ESI markup is not strictly XML compatible. The custom ESI parser
6296 will give higher performance, but cannot handle non ASCII character
6301 DELAY POOL PARAMETERS
6302 -----------------------------------------------------------------------------
6306 TYPE: delay_pool_count
6308 IFDEF: USE_DELAY_POOLS
6311 This represents the number of delay pools to be used. For example,
6312 if you have one class 2 delay pool and one class 3 delays pool, you
6313 have a total of 2 delay pools.
6315 See also delay_parameters, delay_class, delay_access for pool
6316 configuration details.
6320 TYPE: delay_pool_class
6322 IFDEF: USE_DELAY_POOLS
6325 This defines the class of each delay pool. There must be exactly one
6326 delay_class line for each delay pool. For example, to define two
6327 delay pools, one of class 2 and one of class 3, the settings above
6331 delay_pools 4 # 4 delay pools
6332 delay_class 1 2 # pool 1 is a class 2 pool
6333 delay_class 2 3 # pool 2 is a class 3 pool
6334 delay_class 3 4 # pool 3 is a class 4 pool
6335 delay_class 4 5 # pool 4 is a class 5 pool
6337 The delay pool classes are:
6339 class 1 Everything is limited by a single aggregate
6342 class 2 Everything is limited by a single aggregate
6343 bucket as well as an "individual" bucket chosen
6344 from bits 25 through 32 of the IPv4 address.
6346 class 3 Everything is limited by a single aggregate
6347 bucket as well as a "network" bucket chosen
6348 from bits 17 through 24 of the IP address and a
6349 "individual" bucket chosen from bits 17 through
6350 32 of the IPv4 address.
6352 class 4 Everything in a class 3 delay pool, with an
6353 additional limit on a per user basis. This
6354 only takes effect if the username is established
6355 in advance - by forcing authentication in your
6358 class 5 Requests are grouped according their tag (see
6359 external_acl's tag= reply).
6362 Each pool also requires a delay_parameters directive to configure the pool size
6363 and speed limits used whenever the pool is applied to a request. Along with
6364 a set of delay_access directives to determine when it is used.
6366 NOTE: If an IP address is a.b.c.d
6367 -> bits 25 through 32 are "d"
6368 -> bits 17 through 24 are "c"
6369 -> bits 17 through 32 are "c * 256 + d"
6371 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6372 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6374 This clause only supports fast acl types.
6375 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6377 See also delay_parameters and delay_access.
6381 TYPE: delay_pool_access
6383 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6384 IFDEF: USE_DELAY_POOLS
6387 This is used to determine which delay pool a request falls into.
6389 delay_access is sorted per pool and the matching starts with pool 1,
6390 then pool 2, ..., and finally pool N. The first delay pool where the
6391 request is allowed is selected for the request. If it does not allow
6392 the request to any pool then the request is not delayed (default).
6394 For example, if you want some_big_clients in delay
6395 pool 1 and lotsa_little_clients in delay pool 2:
6397 delay_access 1 allow some_big_clients
6398 delay_access 1 deny all
6399 delay_access 2 allow lotsa_little_clients
6400 delay_access 2 deny all
6401 delay_access 3 allow authenticated_clients
6403 See also delay_parameters and delay_class.
6407 NAME: delay_parameters
6408 TYPE: delay_pool_rates
6410 IFDEF: USE_DELAY_POOLS
6413 This defines the parameters for a delay pool. Each delay pool has
6414 a number of "buckets" associated with it, as explained in the
6415 description of delay_class.
6417 For a class 1 delay pool, the syntax is:
6419 delay_parameters pool aggregate
6421 For a class 2 delay pool:
6423 delay_parameters pool aggregate individual
6425 For a class 3 delay pool:
6427 delay_parameters pool aggregate network individual
6429 For a class 4 delay pool:
6431 delay_parameters pool aggregate network individual user
6433 For a class 5 delay pool:
6435 delay_parameters pool tagrate
6437 The option variables are:
6439 pool a pool number - ie, a number between 1 and the
6440 number specified in delay_pools as used in
6443 aggregate the speed limit parameters for the aggregate bucket
6446 individual the speed limit parameters for the individual
6447 buckets (class 2, 3).
6449 network the speed limit parameters for the network buckets
6452 user the speed limit parameters for the user buckets
6455 tagrate the speed limit parameters for the tag buckets
6458 A pair of delay parameters is written restore/maximum, where restore is
6459 the number of bytes (not bits - modem and network speeds are usually
6460 quoted in bits) per second placed into the bucket, and maximum is the
6461 maximum number of bytes which can be in the bucket at any time.
6463 There must be one delay_parameters line for each delay pool.
6466 For example, if delay pool number 1 is a class 2 delay pool as in the
6467 above example, and is being used to strictly limit each host to 64Kbit/sec
6468 (plus overheads), with no overall limit, the line is:
6470 delay_parameters 1 -1/-1 8000/8000
6472 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6474 Note that the figure -1 is used to represent "unlimited".
6477 And, if delay pool number 2 is a class 3 delay pool as in the above
6478 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6479 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6480 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6481 to permit a decent web page to be downloaded at a decent speed
6482 (if the network is not being limited due to overuse) but slow down
6483 large downloads more significantly:
6485 delay_parameters 2 32000/32000 8000/8000 600/8000
6487 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6488 8 x 8000 KByte/sec -> 64Kbit/sec.
6489 8 x 600 Byte/sec -> 4800bit/sec.
6492 Finally, for a class 4 delay pool as in the example - each user will
6493 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6495 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6498 See also delay_class and delay_access.
6502 NAME: delay_initial_bucket_level
6503 COMMENT: (percent, 0-100)
6506 IFDEF: USE_DELAY_POOLS
6507 LOC: Config.Delay.initial
6509 The initial bucket percentage is used to determine how much is put
6510 in each bucket when squid starts, is reconfigured, or first notices
6511 a host accessing it (in class 2 and class 3, individual hosts and
6512 networks only have buckets associated with them once they have been
6517 CLIENT DELAY POOL PARAMETERS
6518 -----------------------------------------------------------------------------
6521 NAME: client_delay_pools
6522 TYPE: client_delay_pool_count
6524 IFDEF: USE_DELAY_POOLS
6525 LOC: Config.ClientDelay
6527 This option specifies the number of client delay pools used. It must
6528 preceed other client_delay_* options.
6531 client_delay_pools 2
6533 See also client_delay_parameters and client_delay_access.
6536 NAME: client_delay_initial_bucket_level
6537 COMMENT: (percent, 0-no_limit)
6540 IFDEF: USE_DELAY_POOLS
6541 LOC: Config.ClientDelay.initial
6543 This option determines the initial bucket size as a percentage of
6544 max_bucket_size from client_delay_parameters. Buckets are created
6545 at the time of the "first" connection from the matching IP. Idle
6546 buckets are periodically deleted up.
6548 You can specify more than 100 percent but note that such "oversized"
6549 buckets are not refilled until their size goes down to max_bucket_size
6550 from client_delay_parameters.
6553 client_delay_initial_bucket_level 50
6556 NAME: client_delay_parameters
6557 TYPE: client_delay_pool_rates
6559 IFDEF: USE_DELAY_POOLS
6560 LOC: Config.ClientDelay
6563 This option configures client-side bandwidth limits using the
6566 client_delay_parameters pool speed_limit max_bucket_size
6568 pool is an integer ID used for client_delay_access matching.
6570 speed_limit is bytes added to the bucket per second.
6572 max_bucket_size is the maximum size of a bucket, enforced after any
6573 speed_limit additions.
6575 Please see the delay_parameters option for more information and
6579 client_delay_parameters 1 1024 2048
6580 client_delay_parameters 2 51200 16384
6582 See also client_delay_access.
6586 NAME: client_delay_access
6587 TYPE: client_delay_pool_access
6589 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6590 IFDEF: USE_DELAY_POOLS
6591 LOC: Config.ClientDelay
6593 This option determines the client-side delay pool for the
6596 client_delay_access pool_ID allow|deny acl_name
6598 All client_delay_access options are checked in their pool ID
6599 order, starting with pool 1. The first checked pool with allowed
6600 request is selected for the request. If no ACL matches or there
6601 are no client_delay_access options, the request bandwidth is not
6604 The ACL-selected pool is then used to find the
6605 client_delay_parameters for the request. Client-side pools are
6606 not used to aggregate clients. Clients are always aggregated
6607 based on their source IP addresses (one bucket per source IP).
6609 This clause only supports fast acl types.
6610 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6611 Additionally, only the client TCP connection details are available.
6612 ACLs testing HTTP properties will not work.
6614 Please see delay_access for more examples.
6617 client_delay_access 1 allow low_rate_network
6618 client_delay_access 2 allow vips_network
6621 See also client_delay_parameters and client_delay_pools.
6625 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6626 -----------------------------------------------------------------------------
6631 LOC: Config.Wccp.router
6633 DEFAULT_DOC: WCCP disabled.
6636 Use this option to define your WCCP ``home'' router for
6639 wccp_router supports a single WCCP(v1) router
6641 wccp2_router supports multiple WCCPv2 routers
6643 only one of the two may be used at the same time and defines
6644 which version of WCCP to use.
6648 TYPE: IpAddress_list
6649 LOC: Config.Wccp2.router
6651 DEFAULT_DOC: WCCPv2 disabled.
6654 Use this option to define your WCCP ``home'' router for
6657 wccp_router supports a single WCCP(v1) router
6659 wccp2_router supports multiple WCCPv2 routers
6661 only one of the two may be used at the same time and defines
6662 which version of WCCP to use.
6667 LOC: Config.Wccp.version
6671 This directive is only relevant if you need to set up WCCP(v1)
6672 to some very old and end-of-life Cisco routers. In all other
6673 setups it must be left unset or at the default setting.
6674 It defines an internal version in the WCCP(v1) protocol,
6675 with version 4 being the officially documented protocol.
6677 According to some users, Cisco IOS 11.2 and earlier only
6678 support WCCP version 3. If you're using that or an earlier
6679 version of IOS, you may need to change this value to 3, otherwise
6680 do not specify this parameter.
6683 NAME: wccp2_rebuild_wait
6685 LOC: Config.Wccp2.rebuildwait
6689 If this is enabled Squid will wait for the cache dir rebuild to finish
6690 before sending the first wccp2 HereIAm packet
6693 NAME: wccp2_forwarding_method
6695 LOC: Config.Wccp2.forwarding_method
6699 WCCP2 allows the setting of forwarding methods between the
6700 router/switch and the cache. Valid values are as follows:
6702 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6703 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6705 Currently (as of IOS 12.4) cisco routers only support GRE.
6706 Cisco switches only support the L2 redirect assignment method.
6709 NAME: wccp2_return_method
6711 LOC: Config.Wccp2.return_method
6715 WCCP2 allows the setting of return methods between the
6716 router/switch and the cache for packets that the cache
6717 decides not to handle. Valid values are as follows:
6719 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6720 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6722 Currently (as of IOS 12.4) cisco routers only support GRE.
6723 Cisco switches only support the L2 redirect assignment.
6725 If the "ip wccp redirect exclude in" command has been
6726 enabled on the cache interface, then it is still safe for
6727 the proxy server to use a l2 redirect method even if this
6728 option is set to GRE.
6731 NAME: wccp2_assignment_method
6733 LOC: Config.Wccp2.assignment_method
6737 WCCP2 allows the setting of methods to assign the WCCP hash
6738 Valid values are as follows:
6740 hash - Hash assignment
6741 mask - Mask assignment
6743 As a general rule, cisco routers support the hash assignment method
6744 and cisco switches support the mask assignment method.
6749 LOC: Config.Wccp2.info
6750 DEFAULT_IF_NONE: standard 0
6751 DEFAULT_DOC: Use the 'web-cache' standard service.
6754 WCCP2 allows for multiple traffic services. There are two
6755 types: "standard" and "dynamic". The standard type defines
6756 one service id - http (id 0). The dynamic service ids can be from
6757 51 to 255 inclusive. In order to use a dynamic service id
6758 one must define the type of traffic to be redirected; this is done
6759 using the wccp2_service_info option.
6761 The "standard" type does not require a wccp2_service_info option,
6762 just specifying the service id will suffice.
6764 MD5 service authentication can be enabled by adding
6765 "password=<password>" to the end of this service declaration.
6769 wccp2_service standard 0 # for the 'web-cache' standard service
6770 wccp2_service dynamic 80 # a dynamic service type which will be
6771 # fleshed out with subsequent options.
6772 wccp2_service standard 0 password=foo
6775 NAME: wccp2_service_info
6776 TYPE: wccp2_service_info
6777 LOC: Config.Wccp2.info
6781 Dynamic WCCPv2 services require further information to define the
6782 traffic you wish to have diverted.
6786 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6787 priority=<priority> ports=<port>,<port>..
6789 The relevant WCCPv2 flags:
6790 + src_ip_hash, dst_ip_hash
6791 + source_port_hash, dst_port_hash
6792 + src_ip_alt_hash, dst_ip_alt_hash
6793 + src_port_alt_hash, dst_port_alt_hash
6796 The port list can be one to eight entries.
6800 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6801 priority=240 ports=80
6803 Note: the service id must have been defined by a previous
6804 'wccp2_service dynamic <id>' entry.
6809 LOC: Config.Wccp2.weight
6813 Each cache server gets assigned a set of the destination
6814 hash proportional to their weight.
6819 LOC: Config.Wccp.address
6821 DEFAULT_DOC: Address selected by the operating system.
6824 Use this option if you require WCCPv2 to use a specific
6827 The default behavior is to not bind to any specific address.
6832 LOC: Config.Wccp2.address
6834 DEFAULT_DOC: Address selected by the operating system.
6837 Use this option if you require WCCP to use a specific
6840 The default behavior is to not bind to any specific address.
6844 PERSISTENT CONNECTION HANDLING
6845 -----------------------------------------------------------------------------
6847 Also see "pconn_timeout" in the TIMEOUTS section
6850 NAME: client_persistent_connections
6852 LOC: Config.onoff.client_pconns
6855 Persistent connection support for clients.
6856 Squid uses persistent connections (when allowed). You can use
6857 this option to disable persistent connections with clients.
6860 NAME: server_persistent_connections
6862 LOC: Config.onoff.server_pconns
6865 Persistent connection support for servers.
6866 Squid uses persistent connections (when allowed). You can use
6867 this option to disable persistent connections with servers.
6870 NAME: persistent_connection_after_error
6872 LOC: Config.onoff.error_pconns
6875 With this directive the use of persistent connections after
6876 HTTP errors can be disabled. Useful if you have clients
6877 who fail to handle errors on persistent connections proper.
6880 NAME: detect_broken_pconn
6882 LOC: Config.onoff.detect_broken_server_pconns
6885 Some servers have been found to incorrectly signal the use
6886 of HTTP/1.0 persistent connections even on replies not
6887 compatible, causing significant delays. This server problem
6888 has mostly been seen on redirects.
6890 By enabling this directive Squid attempts to detect such
6891 broken replies and automatically assume the reply is finished
6892 after 10 seconds timeout.
6896 CACHE DIGEST OPTIONS
6897 -----------------------------------------------------------------------------
6900 NAME: digest_generation
6901 IFDEF: USE_CACHE_DIGESTS
6903 LOC: Config.onoff.digest_generation
6906 This controls whether the server will generate a Cache Digest
6907 of its contents. By default, Cache Digest generation is
6908 enabled if Squid is compiled with --enable-cache-digests defined.
6911 NAME: digest_bits_per_entry
6912 IFDEF: USE_CACHE_DIGESTS
6914 LOC: Config.digest.bits_per_entry
6917 This is the number of bits of the server's Cache Digest which
6918 will be associated with the Digest entry for a given HTTP
6919 Method and URL (public key) combination. The default is 5.
6922 NAME: digest_rebuild_period
6923 IFDEF: USE_CACHE_DIGESTS
6926 LOC: Config.digest.rebuild_period
6929 This is the wait time between Cache Digest rebuilds.
6932 NAME: digest_rewrite_period
6934 IFDEF: USE_CACHE_DIGESTS
6936 LOC: Config.digest.rewrite_period
6939 This is the wait time between Cache Digest writes to
6943 NAME: digest_swapout_chunk_size
6946 IFDEF: USE_CACHE_DIGESTS
6947 LOC: Config.digest.swapout_chunk_size
6950 This is the number of bytes of the Cache Digest to write to
6951 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6955 NAME: digest_rebuild_chunk_percentage
6956 COMMENT: (percent, 0-100)
6957 IFDEF: USE_CACHE_DIGESTS
6959 LOC: Config.digest.rebuild_chunk_percentage
6962 This is the percentage of the Cache Digest to be scanned at a
6963 time. By default it is set to 10% of the Cache Digest.
6968 -----------------------------------------------------------------------------
6973 LOC: Config.Port.snmp
6975 DEFAULT_DOC: SNMP disabled.
6978 The port number where Squid listens for SNMP requests. To enable
6979 SNMP support set this to a suitable port number. Port number
6980 3401 is often used for the Squid SNMP agent. By default it's
6981 set to "0" (disabled)
6989 LOC: Config.accessList.snmp
6991 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
6994 Allowing or denying access to the SNMP port.
6996 All access to the agent is denied by default.
6999 snmp_access allow|deny [!]aclname ...
7001 This clause only supports fast acl types.
7002 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7005 snmp_access allow snmppublic localhost
7006 snmp_access deny all
7009 NAME: snmp_incoming_address
7011 LOC: Config.Addrs.snmp_incoming
7013 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
7016 Just like 'udp_incoming_address', but for the SNMP port.
7018 snmp_incoming_address is used for the SNMP socket receiving
7019 messages from SNMP agents.
7021 The default snmp_incoming_address is to listen on all
7022 available network interfaces.
7025 NAME: snmp_outgoing_address
7027 LOC: Config.Addrs.snmp_outgoing
7029 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
7032 Just like 'udp_outgoing_address', but for the SNMP port.
7034 snmp_outgoing_address is used for SNMP packets returned to SNMP
7037 If snmp_outgoing_address is not set it will use the same socket
7038 as snmp_incoming_address. Only change this if you want to have
7039 SNMP replies sent using another address than where this Squid
7040 listens for SNMP queries.
7042 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
7043 the same value since they both use the same port.
7048 -----------------------------------------------------------------------------
7051 NAME: icp_port udp_port
7054 DEFAULT_DOC: ICP disabled.
7055 LOC: Config.Port.icp
7057 The port number where Squid sends and receives ICP queries to
7058 and from neighbor caches. The standard UDP port for ICP is 3130.
7061 icp_port @DEFAULT_ICP_PORT@
7068 DEFAULT_DOC: HTCP disabled.
7069 LOC: Config.Port.htcp
7071 The port number where Squid sends and receives HTCP queries to
7072 and from neighbor caches. To turn it on you want to set it to
7079 NAME: log_icp_queries
7083 LOC: Config.onoff.log_udp
7085 If set, ICP queries are logged to access.log. You may wish
7086 do disable this if your ICP load is VERY high to speed things
7087 up or to simplify log analysis.
7090 NAME: udp_incoming_address
7092 LOC:Config.Addrs.udp_incoming
7094 DEFAULT_DOC: Accept packets from all machine interfaces.
7096 udp_incoming_address is used for UDP packets received from other
7099 The default behavior is to not bind to any specific address.
7101 Only change this if you want to have all UDP queries received on
7102 a specific interface/address.
7104 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
7105 modules. Altering it will affect all of them in the same manner.
7107 see also; udp_outgoing_address
7109 NOTE, udp_incoming_address and udp_outgoing_address can not
7110 have the same value since they both use the same port.
7113 NAME: udp_outgoing_address
7115 LOC: Config.Addrs.udp_outgoing
7117 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
7119 udp_outgoing_address is used for UDP packets sent out to other
7122 The default behavior is to not bind to any specific address.
7124 Instead it will use the same socket as udp_incoming_address.
7125 Only change this if you want to have UDP queries sent using another
7126 address than where this Squid listens for UDP queries from other
7129 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
7130 modules. Altering it will affect all of them in the same manner.
7132 see also; udp_incoming_address
7134 NOTE, udp_incoming_address and udp_outgoing_address can not
7135 have the same value since they both use the same port.
7142 LOC: Config.onoff.icp_hit_stale
7144 If you want to return ICP_HIT for stale cache objects, set this
7145 option to 'on'. If you have sibling relationships with caches
7146 in other administrative domains, this should be 'off'. If you only
7147 have sibling relationships with caches under your control,
7148 it is probably okay to set this to 'on'.
7149 If set to 'on', your siblings should use the option "allow-miss"
7150 on their cache_peer lines for connecting to you.
7153 NAME: minimum_direct_hops
7156 LOC: Config.minDirectHops
7158 If using the ICMP pinging stuff, do direct fetches for sites
7159 which are no more than this many hops away.
7162 NAME: minimum_direct_rtt
7166 LOC: Config.minDirectRtt
7168 If using the ICMP pinging stuff, do direct fetches for sites
7169 which are no more than this many rtt milliseconds away.
7175 LOC: Config.Netdb.low
7177 The low water mark for the ICMP measurement database.
7179 Note: high watermark controlled by netdb_high directive.
7181 These watermarks are counts, not percents. The defaults are
7182 (low) 900 and (high) 1000. When the high water mark is
7183 reached, database entries will be deleted until the low
7190 LOC: Config.Netdb.high
7192 The high water mark for the ICMP measurement database.
7194 Note: low watermark controlled by netdb_low directive.
7196 These watermarks are counts, not percents. The defaults are
7197 (low) 900 and (high) 1000. When the high water mark is
7198 reached, database entries will be deleted until the low
7202 NAME: netdb_ping_period
7204 LOC: Config.Netdb.period
7207 The minimum period for measuring a site. There will be at
7208 least this much delay between successive pings to the same
7209 network. The default is five minutes.
7216 LOC: Config.onoff.query_icmp
7218 If you want to ask your peers to include ICMP data in their ICP
7219 replies, enable this option.
7221 If your peer has configured Squid (during compilation) with
7222 '--enable-icmp' that peer will send ICMP pings to origin server
7223 sites of the URLs it receives. If you enable this option the
7224 ICP replies from that peer will include the ICMP data (if available).
7225 Then, when choosing a parent cache, Squid will choose the parent with
7226 the minimal RTT to the origin server. When this happens, the
7227 hierarchy field of the access.log will be
7228 "CLOSEST_PARENT_MISS". This option is off by default.
7231 NAME: test_reachability
7235 LOC: Config.onoff.test_reachability
7237 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7238 instead of ICP_MISS if the target host is NOT in the ICMP
7239 database, or has a zero RTT.
7242 NAME: icp_query_timeout
7245 DEFAULT_DOC: Dynamic detection.
7247 LOC: Config.Timeout.icp_query
7249 Normally Squid will automatically determine an optimal ICP
7250 query timeout value based on the round-trip-time of recent ICP
7251 queries. If you want to override the value determined by
7252 Squid, set this 'icp_query_timeout' to a non-zero value. This
7253 value is specified in MILLISECONDS, so, to use a 2-second
7254 timeout (the old default), you would write:
7256 icp_query_timeout 2000
7259 NAME: maximum_icp_query_timeout
7263 LOC: Config.Timeout.icp_query_max
7265 Normally the ICP query timeout is determined dynamically. But
7266 sometimes it can lead to very large values (say 5 seconds).
7267 Use this option to put an upper limit on the dynamic timeout
7268 value. Do NOT use this option to always use a fixed (instead
7269 of a dynamic) timeout value. To set a fixed timeout see the
7270 'icp_query_timeout' directive.
7273 NAME: minimum_icp_query_timeout
7277 LOC: Config.Timeout.icp_query_min
7279 Normally the ICP query timeout is determined dynamically. But
7280 sometimes it can lead to very small timeouts, even lower than
7281 the normal latency variance on your link due to traffic.
7282 Use this option to put an lower limit on the dynamic timeout
7283 value. Do NOT use this option to always use a fixed (instead
7284 of a dynamic) timeout value. To set a fixed timeout see the
7285 'icp_query_timeout' directive.
7288 NAME: background_ping_rate
7292 LOC: Config.backgroundPingRate
7294 Controls how often the ICP pings are sent to siblings that
7295 have background-ping set.
7299 MULTICAST ICP OPTIONS
7300 -----------------------------------------------------------------------------
7305 LOC: Config.mcast_group_list
7308 This tag specifies a list of multicast groups which your server
7309 should join to receive multicasted ICP queries.
7311 NOTE! Be very careful what you put here! Be sure you
7312 understand the difference between an ICP _query_ and an ICP
7313 _reply_. This option is to be set only if you want to RECEIVE
7314 multicast queries. Do NOT set this option to SEND multicast
7315 ICP (use cache_peer for that). ICP replies are always sent via
7316 unicast, so this option does not affect whether or not you will
7317 receive replies from multicast group members.
7319 You must be very careful to NOT use a multicast address which
7320 is already in use by another group of caches.
7322 If you are unsure about multicast, please read the Multicast
7323 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7325 Usage: mcast_groups 239.128.16.128 224.0.1.20
7327 By default, Squid doesn't listen on any multicast groups.
7330 NAME: mcast_miss_addr
7331 IFDEF: MULTICAST_MISS_STREAM
7333 LOC: Config.mcast_miss.addr
7335 DEFAULT_DOC: disabled.
7337 If you enable this option, every "cache miss" URL will
7338 be sent out on the specified multicast address.
7340 Do not enable this option unless you are are absolutely
7341 certain you understand what you are doing.
7344 NAME: mcast_miss_ttl
7345 IFDEF: MULTICAST_MISS_STREAM
7347 LOC: Config.mcast_miss.ttl
7350 This is the time-to-live value for packets multicasted
7351 when multicasting off cache miss URLs is enabled. By
7352 default this is set to 'site scope', i.e. 16.
7355 NAME: mcast_miss_port
7356 IFDEF: MULTICAST_MISS_STREAM
7358 LOC: Config.mcast_miss.port
7361 This is the port number to be used in conjunction with
7365 NAME: mcast_miss_encode_key
7366 IFDEF: MULTICAST_MISS_STREAM
7368 LOC: Config.mcast_miss.encode_key
7369 DEFAULT: XXXXXXXXXXXXXXXX
7371 The URLs that are sent in the multicast miss stream are
7372 encrypted. This is the encryption key.
7375 NAME: mcast_icp_query_timeout
7379 LOC: Config.Timeout.mcast_icp_query
7381 For multicast peers, Squid regularly sends out ICP "probes" to
7382 count how many other peers are listening on the given multicast
7383 address. This value specifies how long Squid should wait to
7384 count all the replies. The default is 2000 msec, or 2
7389 INTERNAL ICON OPTIONS
7390 -----------------------------------------------------------------------------
7393 NAME: icon_directory
7395 LOC: Config.icons.directory
7396 DEFAULT: @DEFAULT_ICON_DIR@
7398 Where the icons are stored. These are normally kept in
7402 NAME: global_internal_static
7404 LOC: Config.onoff.global_internal_static
7407 This directive controls is Squid should intercept all requests for
7408 /squid-internal-static/ no matter which host the URL is requesting
7409 (default on setting), or if nothing special should be done for
7410 such URLs (off setting). The purpose of this directive is to make
7411 icons etc work better in complex cache hierarchies where it may
7412 not always be possible for all corners in the cache mesh to reach
7413 the server generating a directory listing.
7416 NAME: short_icon_urls
7418 LOC: Config.icons.use_short_names
7421 If this is enabled Squid will use short URLs for icons.
7422 If disabled it will revert to the old behavior of including
7423 it's own name and port in the URL.
7425 If you run a complex cache hierarchy with a mix of Squid and
7426 other proxies you may need to disable this directive.
7431 -----------------------------------------------------------------------------
7434 NAME: error_directory
7436 LOC: Config.errorDirectory
7438 DEFAULT_DOC: Send error pages in the clients preferred language
7440 If you wish to create your own versions of the default
7441 error files to customize them to suit your company copy
7442 the error/template files to another directory and point
7445 WARNING: This option will disable multi-language support
7446 on error pages if used.
7448 The squid developers are interested in making squid available in
7449 a wide variety of languages. If you are making translations for a
7450 language that Squid does not currently provide please consider
7451 contributing your translation back to the project.
7452 http://wiki.squid-cache.org/Translations
7454 The squid developers working on translations are happy to supply drop-in
7455 translated error files in exchange for any new language contributions.
7458 NAME: error_default_language
7459 IFDEF: USE_ERR_LOCALES
7461 LOC: Config.errorDefaultLanguage
7463 DEFAULT_DOC: Generate English language pages.
7465 Set the default language which squid will send error pages in
7466 if no existing translation matches the clients language
7469 If unset (default) generic English will be used.
7471 The squid developers are interested in making squid available in
7472 a wide variety of languages. If you are interested in making
7473 translations for any language see the squid wiki for details.
7474 http://wiki.squid-cache.org/Translations
7477 NAME: error_log_languages
7478 IFDEF: USE_ERR_LOCALES
7480 LOC: Config.errorLogMissingLanguages
7483 Log to cache.log what languages users are attempting to
7484 auto-negotiate for translations.
7486 Successful negotiations are not logged. Only failures
7487 have meaning to indicate that Squid may need an upgrade
7488 of its error page translations.
7491 NAME: err_page_stylesheet
7493 LOC: Config.errorStylesheet
7494 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7496 CSS Stylesheet to pattern the display of Squid default error pages.
7498 For information on CSS see http://www.w3.org/Style/CSS/
7503 LOC: Config.errHtmlText
7506 HTML text to include in error messages. Make this a "mailto"
7507 URL to your admin address, or maybe just a link to your
7508 organizations Web page.
7510 To include this in your error messages, you must rewrite
7511 the error template files (found in the "errors" directory).
7512 Wherever you want the 'err_html_text' line to appear,
7513 insert a %L tag in the error template file.
7516 NAME: email_err_data
7519 LOC: Config.onoff.emailErrData
7522 If enabled, information about the occurred error will be
7523 included in the mailto links of the ERR pages (if %W is set)
7524 so that the email body contains the data.
7525 Syntax is <A HREF="mailto:%w%W">%w</A>
7530 LOC: Config.denyInfoList
7533 Usage: deny_info err_page_name acl
7534 or deny_info http://... acl
7535 or deny_info TCP_RESET acl
7537 This can be used to return a ERR_ page for requests which
7538 do not pass the 'http_access' rules. Squid remembers the last
7539 acl it evaluated in http_access, and if a 'deny_info' line exists
7540 for that ACL Squid returns a corresponding error page.
7542 The acl is typically the last acl on the http_access deny line which
7543 denied access. The exceptions to this rule are:
7544 - When Squid needs to request authentication credentials. It's then
7545 the first authentication related acl encountered
7546 - When none of the http_access lines matches. It's then the last
7547 acl processed on the last http_access line.
7548 - When the decision to deny access was made by an adaptation service,
7549 the acl name is the corresponding eCAP or ICAP service_name.
7551 NP: If providing your own custom error pages with error_directory
7552 you may also specify them by your custom file name:
7553 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7555 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7556 may be specified by prefixing the file name with the code and a colon.
7557 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7559 Alternatively you can tell Squid to reset the TCP connection
7560 by specifying TCP_RESET.
7562 Or you can specify an error URL or URL pattern. The browsers will
7563 get redirected to the specified URL after formatting tags have
7564 been replaced. Redirect will be done with 302 or 307 according to
7565 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7566 the URL. e.g. 303:http://example.com/
7569 %a - username (if available. Password NOT included)
7572 %E - Error description
7574 %H - Request domain name
7575 %i - Client IP Address
7577 %o - Message result from external ACL helper
7578 %p - Request Port number
7579 %P - Request Protocol name
7580 %R - Request URL path
7581 %T - Timestamp in RFC 1123 format
7582 %U - Full canonical URL from client
7583 (HTTPS URLs terminate with *)
7584 %u - Full canonical URL from client
7585 %w - Admin email from squid.conf
7587 %% - Literal percent (%) code
7592 OPTIONS INFLUENCING REQUEST FORWARDING
7593 -----------------------------------------------------------------------------
7596 NAME: nonhierarchical_direct
7598 LOC: Config.onoff.nonhierarchical_direct
7601 By default, Squid will send any non-hierarchical requests
7602 (matching hierarchy_stoplist or not cacheable request type) direct
7605 When this is set to "off", Squid will prefer to send these
7606 requests to parents.
7608 Note that in most configurations, by turning this off you will only
7609 add latency to these request without any improvement in global hit
7612 This option only sets a preference. If the parent is unavailable a
7613 direct connection to the origin server may still be attempted. To
7614 completely prevent direct connections use never_direct.
7619 LOC: Config.onoff.prefer_direct
7622 Normally Squid tries to use parents for most requests. If you for some
7623 reason like it to first try going direct and only use a parent if
7624 going direct fails set this to on.
7626 By combining nonhierarchical_direct off and prefer_direct on you
7627 can set up Squid to use a parent as a backup path if going direct
7630 Note: If you want Squid to use parents for all requests see
7631 the never_direct directive. prefer_direct only modifies how Squid
7632 acts on cacheable requests.
7635 NAME: cache_miss_revalidate
7639 LOC: Config.onoff.cache_miss_revalidate
7641 Whether Squid on cache MISS will pass client revalidation requests
7642 to the server or tries to fetch new content for caching.
7643 This is useful while the cache is mostly empty to more quickly
7644 have the cache populated.
7646 When set to 'on' (default), Squid will pass all client If-* headers
7649 When set to 'off' and if the request is cacheable, Squid will
7650 remove the clients If-Modified-Since and If-None-Match headers from
7651 the request sent to the server.
7656 LOC: Config.accessList.AlwaysDirect
7658 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7660 Usage: always_direct allow|deny [!]aclname ...
7662 Here you can use ACL elements to specify requests which should
7663 ALWAYS be forwarded by Squid to the origin servers without using
7664 any peers. For example, to always directly forward requests for
7665 local servers ignoring any parents or siblings you may have use
7668 acl local-servers dstdomain my.domain.net
7669 always_direct allow local-servers
7671 To always forward FTP requests directly, use
7674 always_direct allow FTP
7676 NOTE: There is a similar, but opposite option named
7677 'never_direct'. You need to be aware that "always_direct deny
7678 foo" is NOT the same thing as "never_direct allow foo". You
7679 may need to use a deny rule to exclude a more-specific case of
7680 some other rule. Example:
7682 acl local-external dstdomain external.foo.net
7683 acl local-servers dstdomain .foo.net
7684 always_direct deny local-external
7685 always_direct allow local-servers
7687 NOTE: If your goal is to make the client forward the request
7688 directly to the origin server bypassing Squid then this needs
7689 to be done in the client configuration. Squid configuration
7690 can only tell Squid how Squid should fetch the object.
7692 NOTE: This directive is not related to caching. The replies
7693 is cached as usual even if you use always_direct. To not cache
7694 the replies see the 'cache' directive.
7696 This clause supports both fast and slow acl types.
7697 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7702 LOC: Config.accessList.NeverDirect
7704 DEFAULT_DOC: Allow DNS results to be used for this request.
7706 Usage: never_direct allow|deny [!]aclname ...
7708 never_direct is the opposite of always_direct. Please read
7709 the description for always_direct if you have not already.
7711 With 'never_direct' you can use ACL elements to specify
7712 requests which should NEVER be forwarded directly to origin
7713 servers. For example, to force the use of a proxy for all
7714 requests, except those in your local domain use something like:
7716 acl local-servers dstdomain .foo.net
7717 never_direct deny local-servers
7718 never_direct allow all
7720 or if Squid is inside a firewall and there are local intranet
7721 servers inside the firewall use something like:
7723 acl local-intranet dstdomain .foo.net
7724 acl local-external dstdomain external.foo.net
7725 always_direct deny local-external
7726 always_direct allow local-intranet
7727 never_direct allow all
7729 This clause supports both fast and slow acl types.
7730 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7734 ADVANCED NETWORKING OPTIONS
7735 -----------------------------------------------------------------------------
7738 NAME: incoming_udp_average incoming_icp_average
7741 LOC: Config.comm_incoming.udp.average
7743 Heavy voodoo here. I can't even believe you are reading this.
7744 Are you crazy? Don't even think about adjusting these unless
7745 you understand the algorithms in comm_select.c first!
7748 NAME: incoming_tcp_average incoming_http_average
7751 LOC: Config.comm_incoming.tcp.average
7753 Heavy voodoo here. I can't even believe you are reading this.
7754 Are you crazy? Don't even think about adjusting these unless
7755 you understand the algorithms in comm_select.c first!
7758 NAME: incoming_dns_average
7761 LOC: Config.comm_incoming.dns.average
7763 Heavy voodoo here. I can't even believe you are reading this.
7764 Are you crazy? Don't even think about adjusting these unless
7765 you understand the algorithms in comm_select.c first!
7768 NAME: min_udp_poll_cnt min_icp_poll_cnt
7771 LOC: Config.comm_incoming.udp.min_poll
7773 Heavy voodoo here. I can't even believe you are reading this.
7774 Are you crazy? Don't even think about adjusting these unless
7775 you understand the algorithms in comm_select.c first!
7778 NAME: min_dns_poll_cnt
7781 LOC: Config.comm_incoming.dns.min_poll
7783 Heavy voodoo here. I can't even believe you are reading this.
7784 Are you crazy? Don't even think about adjusting these unless
7785 you understand the algorithms in comm_select.c first!
7788 NAME: min_tcp_poll_cnt min_http_poll_cnt
7791 LOC: Config.comm_incoming.tcp.min_poll
7793 Heavy voodoo here. I can't even believe you are reading this.
7794 Are you crazy? Don't even think about adjusting these unless
7795 you understand the algorithms in comm_select.c first!
7801 LOC: Config.accept_filter
7805 The name of an accept(2) filter to install on Squid's
7806 listen socket(s). This feature is perhaps specific to
7807 FreeBSD and requires support in the kernel.
7809 The 'httpready' filter delays delivering new connections
7810 to Squid until a full HTTP request has been received.
7811 See the accf_http(9) man page for details.
7813 The 'dataready' filter delays delivering new connections
7814 to Squid until there is some data to process.
7815 See the accf_dataready(9) man page for details.
7819 The 'data' filter delays delivering of new connections
7820 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
7821 You may optionally specify a number of seconds to wait by
7822 'data=N' where N is the number of seconds. Defaults to 30
7823 if not specified. See the tcp(7) man page for details.
7826 accept_filter httpready
7831 NAME: client_ip_max_connections
7833 LOC: Config.client_ip_max_connections
7835 DEFAULT_DOC: No limit.
7837 Set an absolute limit on the number of connections a single
7838 client IP can use. Any more than this and Squid will begin to drop
7839 new connections from the client until it closes some links.
7841 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
7842 connections from the client. For finer control use the ACL access controls.
7844 Requires client_db to be enabled (the default).
7846 WARNING: This may noticably slow down traffic received via external proxies
7847 or NAT devices and cause them to rebound error messages back to their clients.
7850 NAME: tcp_recv_bufsize
7854 DEFAULT_DOC: Use operating system TCP defaults.
7855 LOC: Config.tcpRcvBufsz
7857 Size of receive buffer to set for TCP sockets. Probably just
7858 as easy to change your kernel's default.
7859 Omit from squid.conf to use the default buffer size.
7864 -----------------------------------------------------------------------------
7871 LOC: Adaptation::Icap::TheConfig.onoff
7874 If you want to enable the ICAP module support, set this to on.
7877 NAME: icap_connect_timeout
7880 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7883 This parameter specifies how long to wait for the TCP connect to
7884 the requested ICAP server to complete before giving up and either
7885 terminating the HTTP transaction or bypassing the failure.
7887 The default for optional services is peer_connect_timeout.
7888 The default for essential services is connect_timeout.
7889 If this option is explicitly set, its value applies to all services.
7892 NAME: icap_io_timeout
7896 DEFAULT_DOC: Use read_timeout.
7897 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7900 This parameter specifies how long to wait for an I/O activity on
7901 an established, active ICAP connection before giving up and
7902 either terminating the HTTP transaction or bypassing the
7906 NAME: icap_service_failure_limit
7907 COMMENT: limit [in memory-depth time-units]
7908 TYPE: icap_service_failure_limit
7910 LOC: Adaptation::Icap::TheConfig
7913 The limit specifies the number of failures that Squid tolerates
7914 when establishing a new TCP connection with an ICAP service. If
7915 the number of failures exceeds the limit, the ICAP service is
7916 not used for new ICAP requests until it is time to refresh its
7919 A negative value disables the limit. Without the limit, an ICAP
7920 service will not be considered down due to connectivity failures
7921 between ICAP OPTIONS requests.
7923 Squid forgets ICAP service failures older than the specified
7924 value of memory-depth. The memory fading algorithm
7925 is approximate because Squid does not remember individual
7926 errors but groups them instead, splitting the option
7927 value into ten time slots of equal length.
7929 When memory-depth is 0 and by default this option has no
7930 effect on service failure expiration.
7932 Squid always forgets failures when updating service settings
7933 using an ICAP OPTIONS transaction, regardless of this option
7937 # suspend service usage after 10 failures in 5 seconds:
7938 icap_service_failure_limit 10 in 5 seconds
7941 NAME: icap_service_revival_delay
7944 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7947 The delay specifies the number of seconds to wait after an ICAP
7948 OPTIONS request failure before requesting the options again. The
7949 failed ICAP service is considered "down" until fresh OPTIONS are
7952 The actual delay cannot be smaller than the hardcoded minimum
7953 delay of 30 seconds.
7956 NAME: icap_preview_enable
7960 LOC: Adaptation::Icap::TheConfig.preview_enable
7963 The ICAP Preview feature allows the ICAP server to handle the
7964 HTTP message by looking only at the beginning of the message body
7965 or even without receiving the body at all. In some environments,
7966 previews greatly speedup ICAP processing.
7968 During an ICAP OPTIONS transaction, the server may tell Squid what
7969 HTTP messages should be previewed and how big the preview should be.
7970 Squid will not use Preview if the server did not request one.
7972 To disable ICAP Preview for all ICAP services, regardless of
7973 individual ICAP server OPTIONS responses, set this option to "off".
7975 icap_preview_enable off
7978 NAME: icap_preview_size
7981 LOC: Adaptation::Icap::TheConfig.preview_size
7983 DEFAULT_DOC: No preview sent.
7985 The default size of preview data to be sent to the ICAP server.
7986 This value might be overwritten on a per server basis by OPTIONS requests.
7989 NAME: icap_206_enable
7993 LOC: Adaptation::Icap::TheConfig.allow206_enable
7996 206 (Partial Content) responses is an ICAP extension that allows the
7997 ICAP agents to optionally combine adapted and original HTTP message
7998 content. The decision to combine is postponed until the end of the
7999 ICAP response. Squid supports Partial Content extension by default.
8001 Activation of the Partial Content extension is negotiated with each
8002 ICAP service during OPTIONS exchange. Most ICAP servers should handle
8003 negotation correctly even if they do not support the extension, but
8004 some might fail. To disable Partial Content support for all ICAP
8005 services and to avoid any negotiation, set this option to "off".
8011 NAME: icap_default_options_ttl
8014 LOC: Adaptation::Icap::TheConfig.default_options_ttl
8017 The default TTL value for ICAP OPTIONS responses that don't have
8018 an Options-TTL header.
8021 NAME: icap_persistent_connections
8025 LOC: Adaptation::Icap::TheConfig.reuse_connections
8028 Whether or not Squid should use persistent connections to
8032 NAME: adaptation_send_client_ip icap_send_client_ip
8034 IFDEF: USE_ADAPTATION
8036 LOC: Adaptation::Config::send_client_ip
8039 If enabled, Squid shares HTTP client IP information with adaptation
8040 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
8041 For eCAP, Squid sets the libecap::metaClientIp transaction option.
8043 See also: adaptation_uses_indirect_client
8046 NAME: adaptation_send_username icap_send_client_username
8048 IFDEF: USE_ADAPTATION
8050 LOC: Adaptation::Config::send_username
8053 This sends authenticated HTTP client username (if available) to
8054 the adaptation service.
8056 For ICAP, the username value is encoded based on the
8057 icap_client_username_encode option and is sent using the header
8058 specified by the icap_client_username_header option.
8061 NAME: icap_client_username_header
8064 LOC: Adaptation::Icap::TheConfig.client_username_header
8065 DEFAULT: X-Client-Username
8067 ICAP request header name to use for adaptation_send_username.
8070 NAME: icap_client_username_encode
8074 LOC: Adaptation::Icap::TheConfig.client_username_encode
8077 Whether to base64 encode the authenticated client username.
8081 TYPE: icap_service_type
8083 LOC: Adaptation::Icap::TheConfig
8086 Defines a single ICAP service using the following format:
8088 icap_service id vectoring_point uri [option ...]
8091 an opaque identifier or name which is used to direct traffic to
8092 this specific service. Must be unique among all adaptation
8093 services in squid.conf.
8095 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8096 This specifies at which point of transaction processing the
8097 ICAP service should be activated. *_postcache vectoring points
8098 are not yet supported.
8100 uri: icap://servername:port/servicepath
8101 ICAP server and service location.
8103 ICAP does not allow a single service to handle both REQMOD and RESPMOD
8104 transactions. Squid does not enforce that requirement. You can specify
8105 services with the same service_url and different vectoring_points. You
8106 can even specify multiple identical services as long as their
8107 service_names differ.
8109 To activate a service, use the adaptation_access directive. To group
8110 services, use adaptation_service_chain and adaptation_service_set.
8112 Service options are separated by white space. ICAP services support
8113 the following name=value options:
8116 If set to 'on' or '1', the ICAP service is treated as
8117 optional. If the service cannot be reached or malfunctions,
8118 Squid will try to ignore any errors and process the message as
8119 if the service was not enabled. No all ICAP errors can be
8120 bypassed. If set to 0, the ICAP service is treated as
8121 essential and all ICAP errors will result in an error page
8122 returned to the HTTP client.
8124 Bypass is off by default: services are treated as essential.
8127 If set to 'on' or '1', the ICAP service is allowed to
8128 dynamically change the current message adaptation plan by
8129 returning a chain of services to be used next. The services
8130 are specified using the X-Next-Services ICAP response header
8131 value, formatted as a comma-separated list of service names.
8132 Each named service should be configured in squid.conf. Other
8133 services are ignored. An empty X-Next-Services value results
8134 in an empty plan which ends the current adaptation.
8136 Dynamic adaptation plan may cross or cover multiple supported
8137 vectoring points in their natural processing order.
8139 Routing is not allowed by default: the ICAP X-Next-Services
8140 response header is ignored.
8143 Only has effect on split-stack systems. The default on those systems
8144 is to use IPv4-only connections. When set to 'on' this option will
8145 make Squid use IPv6-only connections to contact this ICAP service.
8147 on-overload=block|bypass|wait|force
8148 If the service Max-Connections limit has been reached, do
8149 one of the following for each new ICAP transaction:
8150 * block: send an HTTP error response to the client
8151 * bypass: ignore the "over-connected" ICAP service
8152 * wait: wait (in a FIFO queue) for an ICAP connection slot
8153 * force: proceed, ignoring the Max-Connections limit
8155 In SMP mode with N workers, each worker assumes the service
8156 connection limit is Max-Connections/N, even though not all
8157 workers may use a given service.
8159 The default value is "bypass" if service is bypassable,
8160 otherwise it is set to "wait".
8164 Use the given number as the Max-Connections limit, regardless
8165 of the Max-Connections value given by the service, if any.
8167 Older icap_service format without optional named parameters is
8168 deprecated but supported for backward compatibility.
8171 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8172 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
8176 TYPE: icap_class_type
8181 This deprecated option was documented to define an ICAP service
8182 chain, even though it actually defined a set of similar, redundant
8183 services, and the chains were not supported.
8185 To define a set of redundant services, please use the
8186 adaptation_service_set directive. For service chains, use
8187 adaptation_service_chain.
8191 TYPE: icap_access_type
8196 This option is deprecated. Please use adaptation_access, which
8197 has the same ICAP functionality, but comes with better
8198 documentation, and eCAP support.
8203 -----------------------------------------------------------------------------
8210 LOC: Adaptation::Ecap::TheConfig.onoff
8213 Controls whether eCAP support is enabled.
8217 TYPE: ecap_service_type
8219 LOC: Adaptation::Ecap::TheConfig
8222 Defines a single eCAP service
8224 ecap_service id vectoring_point uri [option ...]
8227 an opaque identifier or name which is used to direct traffic to
8228 this specific service. Must be unique among all adaptation
8229 services in squid.conf.
8231 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8232 This specifies at which point of transaction processing the
8233 eCAP service should be activated. *_postcache vectoring points
8234 are not yet supported.
8236 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8237 Squid uses the eCAP service URI to match this configuration
8238 line with one of the dynamically loaded services. Each loaded
8239 eCAP service must have a unique URI. Obtain the right URI from
8240 the service provider.
8242 To activate a service, use the adaptation_access directive. To group
8243 services, use adaptation_service_chain and adaptation_service_set.
8245 Service options are separated by white space. eCAP services support
8246 the following name=value options:
8249 If set to 'on' or '1', the eCAP service is treated as optional.
8250 If the service cannot be reached or malfunctions, Squid will try
8251 to ignore any errors and process the message as if the service
8252 was not enabled. No all eCAP errors can be bypassed.
8253 If set to 'off' or '0', the eCAP service is treated as essential
8254 and all eCAP errors will result in an error page returned to the
8257 Bypass is off by default: services are treated as essential.
8260 If set to 'on' or '1', the eCAP service is allowed to
8261 dynamically change the current message adaptation plan by
8262 returning a chain of services to be used next.
8264 Dynamic adaptation plan may cross or cover multiple supported
8265 vectoring points in their natural processing order.
8267 Routing is not allowed by default.
8269 Older ecap_service format without optional named parameters is
8270 deprecated but supported for backward compatibility.
8274 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8275 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8278 NAME: loadable_modules
8280 IFDEF: USE_LOADABLE_MODULES
8281 LOC: Config.loadable_module_names
8284 Instructs Squid to load the specified dynamic module(s) or activate
8285 preloaded module(s).
8287 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8291 MESSAGE ADAPTATION OPTIONS
8292 -----------------------------------------------------------------------------
8295 NAME: adaptation_service_set
8296 TYPE: adaptation_service_set_type
8297 IFDEF: USE_ADAPTATION
8302 Configures an ordered set of similar, redundant services. This is
8303 useful when hot standby or backup adaptation servers are available.
8305 adaptation_service_set set_name service_name1 service_name2 ...
8307 The named services are used in the set declaration order. The first
8308 applicable adaptation service from the set is used first. The next
8309 applicable service is tried if and only if the transaction with the
8310 previous service fails and the message waiting to be adapted is still
8313 When adaptation starts, broken services are ignored as if they were
8314 not a part of the set. A broken service is a down optional service.
8316 The services in a set must be attached to the same vectoring point
8317 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8319 If all services in a set are optional then adaptation failures are
8320 bypassable. If all services in the set are essential, then a
8321 transaction failure with one service may still be retried using
8322 another service from the set, but when all services fail, the master
8323 transaction fails as well.
8325 A set may contain a mix of optional and essential services, but that
8326 is likely to lead to surprising results because broken services become
8327 ignored (see above), making previously bypassable failures fatal.
8328 Technically, it is the bypassability of the last failed service that
8331 See also: adaptation_access adaptation_service_chain
8334 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8335 adaptation service_set svcLogger loggerLocal loggerRemote
8338 NAME: adaptation_service_chain
8339 TYPE: adaptation_service_chain_type
8340 IFDEF: USE_ADAPTATION
8345 Configures a list of complementary services that will be applied
8346 one-by-one, forming an adaptation chain or pipeline. This is useful
8347 when Squid must perform different adaptations on the same message.
8349 adaptation_service_chain chain_name service_name1 svc_name2 ...
8351 The named services are used in the chain declaration order. The first
8352 applicable adaptation service from the chain is used first. The next
8353 applicable service is applied to the successful adaptation results of
8354 the previous service in the chain.
8356 When adaptation starts, broken services are ignored as if they were
8357 not a part of the chain. A broken service is a down optional service.
8359 Request satisfaction terminates the adaptation chain because Squid
8360 does not currently allow declaration of RESPMOD services at the
8361 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8363 The services in a chain must be attached to the same vectoring point
8364 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8366 A chain may contain a mix of optional and essential services. If an
8367 essential adaptation fails (or the failure cannot be bypassed for
8368 other reasons), the master transaction fails. Otherwise, the failure
8369 is bypassed as if the failed adaptation service was not in the chain.
8371 See also: adaptation_access adaptation_service_set
8374 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8377 NAME: adaptation_access
8378 TYPE: adaptation_access_type
8379 IFDEF: USE_ADAPTATION
8382 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8384 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8386 adaptation_access service_name allow|deny [!]aclname...
8387 adaptation_access set_name allow|deny [!]aclname...
8389 At each supported vectoring point, the adaptation_access
8390 statements are processed in the order they appear in this
8391 configuration file. Statements pointing to the following services
8392 are ignored (i.e., skipped without checking their ACL):
8394 - services serving different vectoring points
8395 - "broken-but-bypassable" services
8396 - "up" services configured to ignore such transactions
8397 (e.g., based on the ICAP Transfer-Ignore header).
8399 When a set_name is used, all services in the set are checked
8400 using the same rules, to find the first applicable one. See
8401 adaptation_service_set for details.
8403 If an access list is checked and there is a match, the
8404 processing stops: For an "allow" rule, the corresponding
8405 adaptation service is used for the transaction. For a "deny"
8406 rule, no adaptation service is activated.
8408 It is currently not possible to apply more than one adaptation
8409 service at the same vectoring point to the same HTTP transaction.
8411 See also: icap_service and ecap_service
8414 adaptation_access service_1 allow all
8417 NAME: adaptation_service_iteration_limit
8419 IFDEF: USE_ADAPTATION
8420 LOC: Adaptation::Config::service_iteration_limit
8423 Limits the number of iterations allowed when applying adaptation
8424 services to a message. If your longest adaptation set or chain
8425 may have more than 16 services, increase the limit beyond its
8426 default value of 16. If detecting infinite iteration loops sooner
8427 is critical, make the iteration limit match the actual number
8428 of services in your longest adaptation set or chain.
8430 Infinite adaptation loops are most likely with routing services.
8432 See also: icap_service routing=1
8435 NAME: adaptation_masterx_shared_names
8437 IFDEF: USE_ADAPTATION
8438 LOC: Adaptation::Config::masterx_shared_name
8441 For each master transaction (i.e., the HTTP request and response
8442 sequence, including all related ICAP and eCAP exchanges), Squid
8443 maintains a table of metadata. The table entries are (name, value)
8444 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8445 with the master transaction.
8447 This option specifies the table entry names that Squid must accept
8448 from and forward to the adaptation transactions.
8450 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8451 shared table by returning an ICAP header field with a name
8452 specified in adaptation_masterx_shared_names.
8454 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8455 shared table by implementing the libecap::visitEachOption() API
8456 to provide an option with a name specified in
8457 adaptation_masterx_shared_names.
8459 Squid will store and forward the set entry to subsequent adaptation
8460 transactions within the same master transaction scope.
8462 Only one shared entry name is supported at this time.
8465 # share authentication information among ICAP services
8466 adaptation_masterx_shared_names X-Subscriber-ID
8469 NAME: adaptation_meta
8471 IFDEF: USE_ADAPTATION
8472 LOC: Adaptation::Config::metaHeaders
8475 This option allows Squid administrator to add custom ICAP request
8476 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8477 Use it to pass custom authentication tokens and other
8478 transaction-state related meta information to an ICAP/eCAP service.
8480 The addition of a meta header is ACL-driven:
8481 adaptation_meta name value [!]aclname ...
8483 Processing for a given header name stops after the first ACL list match.
8484 Thus, it is impossible to add two headers with the same name. If no ACL
8485 lists match for a given header name, no such header is added. For
8488 # do not debug transactions except for those that need debugging
8489 adaptation_meta X-Debug 1 needs_debugging
8491 # log all transactions except for those that must remain secret
8492 adaptation_meta X-Log 1 !keep_secret
8494 # mark transactions from users in the "G 1" group
8495 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8497 The "value" parameter may be a regular squid.conf token or a "double
8498 quoted string". Within the quoted string, use backslash (\) to escape
8499 any character, which is currently only useful for escaping backslashes
8500 and double quotes. For example,
8501 "this string has one backslash (\\) and two \"quotes\""
8503 Used adaptation_meta header values may be logged via %note
8504 logformat code. If multiple adaptation_meta headers with the same name
8505 are used during master transaction lifetime, the header values are
8506 logged in the order they were used and duplicate values are ignored
8507 (only the first repeated value will be logged).
8513 LOC: Adaptation::Icap::TheConfig.repeat
8514 DEFAULT_IF_NONE: deny all
8516 This ACL determines which retriable ICAP transactions are
8517 retried. Transactions that received a complete ICAP response
8518 and did not have to consume or produce HTTP bodies to receive
8519 that response are usually retriable.
8521 icap_retry allow|deny [!]aclname ...
8523 Squid automatically retries some ICAP I/O timeouts and errors
8524 due to persistent connection race conditions.
8526 See also: icap_retry_limit
8529 NAME: icap_retry_limit
8532 LOC: Adaptation::Icap::TheConfig.repeat_limit
8534 DEFAULT_DOC: No retries are allowed.
8536 Limits the number of retries allowed.
8538 Communication errors due to persistent connection race
8539 conditions are unavoidable, automatically retried, and do not
8540 count against this limit.
8542 See also: icap_retry
8548 -----------------------------------------------------------------------------
8551 NAME: check_hostnames
8554 LOC: Config.onoff.check_hostnames
8556 For security and stability reasons Squid can check
8557 hostnames for Internet standard RFC compliance. If you want
8558 Squid to perform these checks turn this directive on.
8561 NAME: allow_underscore
8564 LOC: Config.onoff.allow_underscore
8566 Underscore characters is not strictly allowed in Internet hostnames
8567 but nevertheless used by many sites. Set this to off if you want
8568 Squid to be strict about the standard.
8569 This check is performed only when check_hostnames is set to on.
8572 NAME: dns_retransmit_interval
8575 LOC: Config.Timeout.idns_retransmit
8577 Initial retransmit interval for DNS queries. The interval is
8578 doubled each time all configured DNS servers have been tried.
8584 LOC: Config.Timeout.idns_query
8586 DNS Query timeout. If no response is received to a DNS query
8587 within this time all DNS servers for the queried domain
8588 are assumed to be unavailable.
8591 NAME: dns_packet_max
8593 DEFAULT_DOC: EDNS disabled
8595 LOC: Config.dns.packet_max
8597 Maximum number of bytes packet size to advertise via EDNS.
8598 Set to "none" to disable EDNS large packet support.
8600 For legacy reasons DNS UDP replies will default to 512 bytes which
8601 is too small for many responses. EDNS provides a means for Squid to
8602 negotiate receiving larger responses back immediately without having
8603 to failover with repeat requests. Responses larger than this limit
8604 will retain the old behaviour of failover to TCP DNS.
8606 Squid has no real fixed limit internally, but allowing packet sizes
8607 over 1500 bytes requires network jumbogram support and is usually not
8610 WARNING: The RFC also indicates that some older resolvers will reply
8611 with failure of the whole request if the extension is added. Some
8612 resolvers have already been identified which will reply with mangled
8613 EDNS response on occasion. Usually in response to many-KB jumbogram
8614 sizes being advertised by Squid.
8615 Squid will currently treat these both as an unable-to-resolve domain
8616 even if it would be resolvable without EDNS.
8623 DEFAULT_DOC: Search for single-label domain names is disabled.
8624 LOC: Config.onoff.res_defnames
8626 Normally the RES_DEFNAMES resolver option is disabled
8627 (see res_init(3)). This prevents caches in a hierarchy
8628 from interpreting single-component hostnames locally. To allow
8629 Squid to handle single-component names, enable this option.
8632 NAME: dns_multicast_local
8636 DEFAULT_DOC: Search for .local and .arpa names is disabled.
8637 LOC: Config.onoff.dns_mdns
8639 When set to on, Squid sends multicast DNS lookups on the local
8640 network for domains ending in .local and .arpa.
8641 This enables local servers and devices to be contacted in an
8642 ad-hoc or zero-configuration network environment.
8645 NAME: dns_nameservers
8648 DEFAULT_DOC: Use operating system definitions
8649 LOC: Config.dns_nameservers
8651 Use this if you want to specify a list of DNS name servers
8652 (IP addresses) to use instead of those given in your
8653 /etc/resolv.conf file.
8655 On Windows platforms, if no value is specified here or in
8656 the /etc/resolv.conf file, the list of DNS name servers are
8657 taken from the Windows registry, both static and dynamic DHCP
8658 configurations are supported.
8660 Example: dns_nameservers 10.0.0.1 192.172.0.4
8665 DEFAULT: @DEFAULT_HOSTS@
8666 LOC: Config.etcHostsPath
8668 Location of the host-local IP name-address associations
8669 database. Most Operating Systems have such a file on different
8671 - Un*X & Linux: /etc/hosts
8672 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
8673 (%SystemRoot% value install default is c:\winnt)
8674 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
8675 (%SystemRoot% value install default is c:\windows)
8676 - Windows 9x/Me: %windir%\hosts
8677 (%windir% value is usually c:\windows)
8678 - Cygwin: /etc/hosts
8680 The file contains newline-separated definitions, in the
8681 form ip_address_in_dotted_form name [name ...] names are
8682 whitespace-separated. Lines beginning with an hash (#)
8683 character are comments.
8685 The file is checked at startup and upon configuration.
8686 If set to 'none', it won't be checked.
8687 If append_domain is used, that domain will be added to
8688 domain-local (i.e. not containing any dot character) host
8694 LOC: Config.appendDomain
8696 DEFAULT_DOC: Use operating system definitions
8698 Appends local domain name to hostnames without any dots in
8699 them. append_domain must begin with a period.
8701 Be warned there are now Internet names with no dots in
8702 them using only top-domain names, so setting this may
8703 cause some Internet sites to become unavailable.
8706 append_domain .yourdomain.com
8709 NAME: ignore_unknown_nameservers
8711 LOC: Config.onoff.ignore_unknown_nameservers
8714 By default Squid checks that DNS responses are received
8715 from the same IP addresses they are sent to. If they
8716 don't match, Squid ignores the response and writes a warning
8717 message to cache.log. You can allow responses from unknown
8718 nameservers by setting this option to 'off'.
8724 LOC: Config.dns.v4_first
8726 With the IPv6 Internet being as fast or faster than IPv4 Internet
8727 for most networks Squid prefers to contact websites over IPv6.
8729 This option reverses the order of preference to make Squid contact
8730 dual-stack websites over IPv4 first. Squid will still perform both
8731 IPv6 and IPv4 DNS lookups before connecting.
8734 This option will restrict the situations under which IPv6
8735 connectivity is used (and tested). Hiding network problems
8736 which would otherwise be detected and warned about.
8740 COMMENT: (number of entries)
8743 LOC: Config.ipcache.size
8745 Maximum number of DNS IP cache entries.
8752 LOC: Config.ipcache.low
8759 LOC: Config.ipcache.high
8761 The size, low-, and high-water marks for the IP cache.
8764 NAME: fqdncache_size
8765 COMMENT: (number of entries)
8768 LOC: Config.fqdncache.size
8770 Maximum number of FQDN cache entries.
8775 -----------------------------------------------------------------------------
8778 NAME: configuration_includes_quoted_values
8780 TYPE: configuration_includes_quoted_values
8782 LOC: ConfigParser::RecognizeQuotedValues
8784 If set, Squid will recognize each "quoted string" after a configuration
8785 directive as a single parameter. The quotes are stripped before the
8786 parameter value is interpreted or used.
8787 See "Values with spaces, quotes, and other special characters"
8788 section for more details.
8795 LOC: Config.onoff.mem_pools
8797 If set, Squid will keep pools of allocated (but unused) memory
8798 available for future use. If memory is a premium on your
8799 system and you believe your malloc library outperforms Squid
8800 routines, disable this.
8803 NAME: memory_pools_limit
8807 LOC: Config.MemPools.limit
8809 Used only with memory_pools on:
8810 memory_pools_limit 50 MB
8812 If set to a non-zero value, Squid will keep at most the specified
8813 limit of allocated (but unused) memory in memory pools. All free()
8814 requests that exceed this limit will be handled by your malloc
8815 library. Squid does not pre-allocate any memory, just safe-keeps
8816 objects that otherwise would be free()d. Thus, it is safe to set
8817 memory_pools_limit to a reasonably high value even if your
8818 configuration will use less memory.
8820 If set to none, Squid will keep all memory it can. That is, there
8821 will be no limit on the total amount of memory used for safe-keeping.
8823 To disable memory allocation optimization, do not set
8824 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
8826 An overhead for maintaining memory pools is not taken into account
8827 when the limit is checked. This overhead is close to four bytes per
8828 object kept. However, pools may actually _save_ memory because of
8829 reduced memory thrashing in your malloc library.
8833 COMMENT: on|off|transparent|truncate|delete
8836 LOC: opt_forwarded_for
8838 If set to "on", Squid will append your client's IP address
8839 in the HTTP requests it forwards. By default it looks like:
8841 X-Forwarded-For: 192.1.2.3
8843 If set to "off", it will appear as
8845 X-Forwarded-For: unknown
8847 If set to "transparent", Squid will not alter the
8848 X-Forwarded-For header in any way.
8850 If set to "delete", Squid will delete the entire
8851 X-Forwarded-For header.
8853 If set to "truncate", Squid will remove all existing
8854 X-Forwarded-For entries, and place the client IP as the sole entry.
8857 NAME: cachemgr_passwd
8858 TYPE: cachemgrpasswd
8860 DEFAULT_DOC: No password. Actions which require password are denied.
8861 LOC: Config.passwd_list
8863 Specify passwords for cachemgr operations.
8865 Usage: cachemgr_passwd password action action ...
8867 Some valid actions are (see cache manager menu for a full list):
8907 * Indicates actions which will not be performed without a
8908 valid password, others can be performed if not listed here.
8910 To disable an action, set the password to "disable".
8911 To allow performing an action without a password, set the
8914 Use the keyword "all" to set the same password for all actions.
8917 cachemgr_passwd secret shutdown
8918 cachemgr_passwd lesssssssecret info stats/objects
8919 cachemgr_passwd disable all
8926 LOC: Config.onoff.client_db
8928 If you want to disable collecting per-client statistics,
8929 turn off client_db here.
8932 NAME: refresh_all_ims
8936 LOC: Config.onoff.refresh_all_ims
8938 When you enable this option, squid will always check
8939 the origin server for an update when a client sends an
8940 If-Modified-Since request. Many browsers use IMS
8941 requests when the user requests a reload, and this
8942 ensures those clients receive the latest version.
8944 By default (off), squid may return a Not Modified response
8945 based on the age of the cached version.
8948 NAME: reload_into_ims
8949 IFDEF: USE_HTTP_VIOLATIONS
8953 LOC: Config.onoff.reload_into_ims
8955 When you enable this option, client no-cache or ``reload''
8956 requests will be changed to If-Modified-Since requests.
8957 Doing this VIOLATES the HTTP standard. Enabling this
8958 feature could make you liable for problems which it
8961 see also refresh_pattern for a more selective approach.
8964 NAME: connect_retries
8966 LOC: Config.connect_retries
8968 DEFAULT_DOC: Do not retry failed connections.
8970 This sets the maximum number of connection attempts made for each
8971 TCP connection. The connect_retries attempts must all still
8972 complete within the connection timeout period.
8974 The default is not to re-try if the first connection attempt fails.
8975 The (not recommended) maximum is 10 tries.
8977 A warning message will be generated if it is set to a too-high
8978 value and the configured value will be over-ridden.
8980 Note: These re-tries are in addition to forward_max_tries
8981 which limit how many different addresses may be tried to find
8985 NAME: retry_on_error
8987 LOC: Config.retry.onerror
8990 If set to ON Squid will automatically retry requests when
8991 receiving an error response with status 403 (Forbidden),
8992 500 (Internal Error), 501 or 503 (Service not available).
8993 Status 502 and 504 (Gateway errors) are always retried.
8995 This is mainly useful if you are in a complex cache hierarchy to
8996 work around access control errors.
8998 NOTE: This retry will attempt to find another working destination.
8999 Which is different from the server which just failed.
9002 NAME: as_whois_server
9004 LOC: Config.as_whois_server
9005 DEFAULT: whois.ra.net
9007 WHOIS server to query for AS numbers. NOTE: AS numbers are
9008 queried only when Squid starts up, not for every request.
9013 LOC: Config.onoff.offline
9016 Enable this option and Squid will never try to validate cached
9020 NAME: uri_whitespace
9021 TYPE: uri_whitespace
9022 LOC: Config.uri_whitespace
9025 What to do with requests that have whitespace characters in the
9028 strip: The whitespace characters are stripped out of the URL.
9029 This is the behavior recommended by RFC2396 and RFC3986
9030 for tolerant handling of generic URI.
9031 NOTE: This is one difference between generic URI and HTTP URLs.
9033 deny: The request is denied. The user receives an "Invalid
9035 This is the behaviour recommended by RFC2616 for safe
9036 handling of HTTP request URL.
9038 allow: The request is allowed and the URI is not changed. The
9039 whitespace characters remain in the URI. Note the
9040 whitespace is passed to redirector processes if they
9042 Note this may be considered a violation of RFC2616
9043 request parsing where whitespace is prohibited in the
9046 encode: The request is allowed and the whitespace characters are
9047 encoded according to RFC1738.
9049 chop: The request is allowed and the URI is chopped at the
9053 NOTE the current Squid implementation of encode and chop violates
9054 RFC2616 by not using a 301 redirect after altering the URL.
9059 LOC: Config.chroot_dir
9062 Specifies a directory where Squid should do a chroot() while
9063 initializing. This also causes Squid to fully drop root
9064 privileges after initializing. This means, for example, if you
9065 use a HTTP port less than 1024 and try to reconfigure, you may
9066 get an error saying that Squid can not open the port.
9069 NAME: balance_on_multiple_ip
9071 LOC: Config.onoff.balance_on_multiple_ip
9074 Modern IP resolvers in squid sort lookup results by preferred access.
9075 By default squid will use these IP in order and only rotates to
9076 the next listed when the most preffered fails.
9078 Some load balancing servers based on round robin DNS have been
9079 found not to preserve user session state across requests
9080 to different IP addresses.
9082 Enabling this directive Squid rotates IP's per request.
9085 NAME: pipeline_prefetch
9086 TYPE: pipelinePrefetch
9087 LOC: Config.pipeline_max_prefetch
9089 DEFAULT_DOC: Do not pre-parse pipelined requests.
9091 HTTP clients may send a pipeline of 1+N requests to Squid using a
9092 single connection, without waiting for Squid to respond to the first
9093 of those requests. This option limits the number of concurrent
9094 requests Squid will try to handle in parallel. If set to N, Squid
9095 will try to receive and process up to 1+N requests on the same
9096 connection concurrently.
9098 Defaults to 0 (off) for bandwidth management and access logging
9101 NOTE: pipelining requires persistent connections to clients.
9103 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
9106 NAME: high_response_time_warning
9109 LOC: Config.warnings.high_rptm
9111 DEFAULT_DOC: disabled.
9113 If the one-minute median response time exceeds this value,
9114 Squid prints a WARNING with debug level 0 to get the
9115 administrators attention. The value is in milliseconds.
9118 NAME: high_page_fault_warning
9120 LOC: Config.warnings.high_pf
9122 DEFAULT_DOC: disabled.
9124 If the one-minute average page fault rate exceeds this
9125 value, Squid prints a WARNING with debug level 0 to get
9126 the administrators attention. The value is in page faults
9130 NAME: high_memory_warning
9132 LOC: Config.warnings.high_memory
9133 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
9135 DEFAULT_DOC: disabled.
9137 If the memory usage (as determined by gnumalloc, if available and used)
9138 exceeds this amount, Squid prints a WARNING with debug level 0 to get
9139 the administrators attention.
9141 # TODO: link high_memory_warning to mempools?
9143 NAME: sleep_after_fork
9144 COMMENT: (microseconds)
9146 LOC: Config.sleep_after_fork
9149 When this is set to a non-zero value, the main Squid process
9150 sleeps the specified number of microseconds after a fork()
9151 system call. This sleep may help the situation where your
9152 system reports fork() failures due to lack of (virtual)
9153 memory. Note, however, if you have a lot of child
9154 processes, these sleep delays will add up and your
9155 Squid will not service requests for some amount of time
9156 until all the child processes have been started.
9157 On Windows value less then 1000 (1 milliseconds) are
9161 NAME: windows_ipaddrchangemonitor
9162 IFDEF: _SQUID_WINDOWS_
9166 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9168 On Windows Squid by default will monitor IP address changes and will
9169 reconfigure itself after any detected event. This is very useful for
9170 proxies connected to internet with dial-up interfaces.
9171 In some cases (a Proxy server acting as VPN gateway is one) it could be
9172 desiderable to disable this behaviour setting this to 'off'.
9173 Note: after changing this, Squid service must be restarted.
9178 IFDEF: USE_SQUID_EUI
9180 LOC: Eui::TheConfig.euiLookup
9182 Whether to lookup the EUI or MAC address of a connected client.
9185 NAME: max_filedescriptors max_filedesc
9188 DEFAULT_DOC: Use operating system limits set by ulimit.
9189 LOC: Config.max_filedescriptors
9191 Reduce the maximum number of filedescriptors supported below
9192 the usual operating system defaults.
9194 Remove from squid.conf to inherit the current ulimit setting.
9196 Note: Changing this requires a restart of Squid. Also
9197 not all I/O types supports large values (eg on Windows).
9204 DEFAULT_DOC: SMP support disabled.
9206 Number of main Squid processes or "workers" to fork and maintain.
9207 0: "no daemon" mode, like running "squid -N ..."
9208 1: "no SMP" mode, start one main Squid process daemon (default)
9209 N: start N main Squid process daemons (i.e., SMP mode)
9211 In SMP mode, each worker does nearly all what a single Squid daemon
9212 does (e.g., listen on http_port and forward HTTP requests).
9215 NAME: cpu_affinity_map
9216 TYPE: CpuAffinityMap
9217 LOC: Config.cpuAffinityMap
9219 DEFAULT_DOC: Let operating system decide.
9221 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
9223 Sets 1:1 mapping between Squid processes and CPU cores. For example,
9225 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
9227 affects processes 1 through 4 only and places them on the first
9228 four even cores, starting with core #1.
9230 CPU cores are numbered starting from 1. Requires support for
9231 sched_getaffinity(2) and sched_setaffinity(2) system calls.
9233 Multiple cpu_affinity_map options are merged.