1 ## Copyright (C) 1996-2022 The Squid Software Foundation and contributors
3 ## Squid software is distributed under GPLv2+ license and includes
4 ## contributions from numerous individuals and organizations.
5 ## Please see the COPYING and CONTRIBUTORS files for details.
10 ----------------------------
12 This is the documentation for the Squid configuration file.
13 This documentation can also be found online at:
14 http://www.squid-cache.org/Doc/config/
16 You may wish to look at the Squid home page and wiki for the
17 FAQ and other documentation:
18 http://www.squid-cache.org/
19 http://wiki.squid-cache.org/SquidFaq
20 http://wiki.squid-cache.org/ConfigExamples
22 This documentation shows what the defaults for various directives
23 happen to be. If you don't need to change the default, you should
24 leave the line out of your squid.conf in most cases.
26 In some cases "none" refers to no default setting at all,
27 while in other cases it refers to the value of the option
28 - the comments for that keyword indicate if this is the case.
33 Configuration options can be included using the "include" directive.
34 Include takes a list of files to include. Quoting and wildcards are
39 include /path/to/included/file/squid.acl.config
41 Includes can be nested up to a hard-coded depth of 16 levels.
42 This arbitrary restriction is to prevent recursive include references
43 from causing Squid entering an infinite loop whilst trying to load
46 Values with byte units
48 Squid accepts size units on some size related directives. All
49 such directives are documented with a default value displaying
52 Units accepted by Squid are:
54 KB - Kilobyte (1024 bytes)
58 Values with time units
60 Time-related directives marked with either "time-units" or
61 "time-units-small" accept a time unit. The supported time units are:
63 nanosecond (time-units-small only)
64 microsecond (time-units-small only)
73 year - 31557790080 milliseconds (just over 365 days)
76 Values with spaces, quotes, and other special characters
78 Squid supports directive parameters with spaces, quotes, and other
79 special characters. Surround such parameters with "double quotes". Use
80 the configuration_includes_quoted_values directive to enable or
83 Squid supports reading configuration option parameters from external
84 files using the syntax:
85 parameters("/path/filename")
87 acl allowlist dstdomain parameters("/etc/squid/allowlist.txt")
89 Conditional configuration
91 If-statements can be used to make configuration directives
95 ... regular configuration directives ...
97 ... regular configuration directives ...]
100 The else part is optional. The keywords "if", "else", and "endif"
101 must be typed on their own lines, as if they were regular
102 configuration directives.
104 NOTE: An else-if condition is not supported.
106 These individual conditions types are supported:
109 Always evaluates to true.
111 Always evaluates to false.
112 <integer> = <integer>
113 Equality comparison of two integer numbers.
118 The following SMP-related preprocessor macros can be used.
120 ${process_name} expands to the current Squid process "name"
121 (e.g., squid1, squid2, or cache1).
123 ${process_number} expands to the current Squid process
124 identifier, which is an integer number (e.g., 1, 2, 3) unique
125 across all Squid processes of the current service instance.
127 ${service_name} expands into the current Squid service instance
128 name identifier which is provided by -n on the command line.
132 Logformat macros can be used in many places outside of the logformat
133 directive. In theory, all of the logformat codes can be used as %macros,
134 where they are supported. In practice, a %macro expands as a dash (-) when
135 the transaction does not yet have enough information and a value is needed.
137 There is no definitive list of what tokens are available at the various
138 stages of the transaction.
140 And some information may already be available to Squid but not yet
141 committed where the macro expansion code can access it (report
142 such instances!). The macro will be expanded into a single dash
143 ('-') in such cases. Not all macros have been tested.
147 # options still not yet ported from 2.7 to 3.x
148 NAME: broken_vary_encoding
151 This option is not yet supported by Squid-3.
157 This option is not yet supported by Squid-3.
163 This option is not yet supported by Squid-3.
166 NAME: external_refresh_check
169 This option is not yet supported by Squid-3.
172 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
175 This option is not yet supported by Squid-3.
178 NAME: refresh_stale_hit
181 This option is not yet supported by Squid-3.
184 # Options removed in 6.x
188 Remove this line. Squid no longer supports this feature.
194 Remove this line. Squid no longer supports this feature.
197 NAME: announce_period
200 Remove this line. Squid no longer supports this feature.
206 Remove this line. Squid no longer supports this feature.
209 NAME: request_entities
212 Remove this line. Squid now accepts HTTP/1.1 requests with bodies.
213 To simplify UI and code, Squid rejects certain HTTP/1.0 requests with bodies.
216 # Options removed in 5.x
220 Remove this line. Squid no longer supports preferential treatment of DNS A records.
223 # Options removed in 4.x
224 NAME: cache_peer_domain cache_host_domain
227 Replace with dstdomain ACLs and cache_peer_access.
233 Remove this line. The behaviour enabled by this is no longer needed.
236 NAME: sslproxy_cafile
239 Remove this line. Use tls_outgoing_options cafile= instead.
242 NAME: sslproxy_capath
245 Remove this line. Use tls_outgoing_options capath= instead.
248 NAME: sslproxy_cipher
251 Remove this line. Use tls_outgoing_options cipher= instead.
254 NAME: sslproxy_client_certificate
257 Remove this line. Use tls_outgoing_options cert= instead.
260 NAME: sslproxy_client_key
263 Remove this line. Use tls_outgoing_options key= instead.
269 Remove this line. Use tls_outgoing_options flags= instead.
272 NAME: sslproxy_options
275 Remove this line. Use tls_outgoing_options options= instead.
278 NAME: sslproxy_version
281 Remove this line. Use tls_outgoing_options options= instead.
284 # Options removed in 3.5
285 NAME: hierarchy_stoplist
288 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
291 # Options removed in 3.4
295 Remove this line. Use acls with access_log directives to control access logging
301 Remove this line. Use acls with icap_log directives to control icap logging
304 # Options Removed in 3.3
305 NAME: ignore_ims_on_miss
308 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
311 # Options Removed in 3.2
312 NAME: balance_on_multiple_ip
315 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, this multiple-IP algorithm is not longer relevant.
318 NAME: chunked_request_body_max_size
321 Remove this line. Squid is now HTTP/1.1 compliant.
324 NAME: dns_v4_fallback
327 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
330 NAME: emulate_httpd_log
333 Replace this with an access_log directive using the format 'common' or 'combined'.
339 Use a regular access.log with ACL limiting it to MISS events.
345 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
348 NAME: ignore_expect_100
351 Remove this line. The HTTP/1.1 feature is now fully supported by default.
357 Remove this option from your config. To log FQDN use %>A in the log format.
360 NAME: log_ip_on_direct
363 Remove this option from your config. To log server or peer names use %<A in the log format.
366 NAME: maximum_single_addr_tries
369 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
372 NAME: referer_log referrer_log
375 Replace this with an access_log directive using the format 'referrer'.
381 Remove this line. The feature is supported by default in storage types where update is implemented.
384 NAME: url_rewrite_concurrency
387 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
393 Replace this with an access_log directive using the format 'useragent'.
396 # Options Removed in 3.1
400 Remove this line. DNS is no longer tested on startup.
403 NAME: extension_methods
406 Remove this line. All valid methods for HTTP are accepted by default.
409 # 2.7 Options Removed/Replaced in 3.2
414 # 2.7 Options Removed/Replaced in 3.1
422 Remove this line. HTTP/1.1 is supported by default.
425 NAME: upgrade_http0.9
428 Remove this line. ICY/1.0 streaming protocol is supported by default.
431 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
434 Alter these entries. Use the qos_flows directive instead.
437 # Options Removed in 3.0
441 Since squid-3.0 replace with request_header_access or reply_header_access
442 depending on whether you wish to match client requests or server replies.
445 NAME: httpd_accel_no_pmtu_disc
448 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
451 NAME: wais_relay_host
454 Replace this line with 'cache_peer' configuration.
457 NAME: wais_relay_port
460 Replace this line with 'cache_peer' configuration.
465 -----------------------------------------------------------------------------
472 DEFAULT_DOC: SMP support disabled.
474 Number of main Squid processes or "workers" to fork and maintain.
475 0: "no daemon" mode, like running "squid -N ..."
476 1: "no SMP" mode, start one main Squid process daemon (default)
477 N: start N main Squid process daemons (i.e., SMP mode)
479 In SMP mode, each worker does nearly all what a single Squid daemon
480 does (e.g., listen on http_port and forward HTTP requests).
483 NAME: cpu_affinity_map
485 LOC: Config.cpuAffinityMap
487 DEFAULT_DOC: Let operating system decide.
489 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
491 Sets 1:1 mapping between Squid processes and CPU cores. For example,
493 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
495 affects processes 1 through 4 only and places them on the first
496 four even cores, starting with core #1.
498 CPU cores are numbered starting from 1. Requires support for
499 sched_getaffinity(2) and sched_setaffinity(2) system calls.
501 Multiple cpu_affinity_map options are merged.
506 NAME: shared_memory_locking
509 LOC: Config.shmLocking
512 Whether to ensure that all required shared memory is available by
513 "locking" that shared memory into RAM when Squid starts. The
514 alternative is faster startup time followed by slightly slower
515 performance and, if not enough RAM is actually available during
516 runtime, mysterious crashes.
518 SMP Squid uses many shared memory segments. These segments are
519 brought into Squid memory space using an mmap(2) system call. During
520 Squid startup, the mmap() call often succeeds regardless of whether
521 the system has enough RAM. In general, Squid cannot tell whether the
522 kernel applies this "optimistic" memory allocation policy (but
523 popular modern kernels usually use it).
525 Later, if Squid attempts to actually access the mapped memory
526 regions beyond what the kernel is willing to allocate, the
527 "optimistic" kernel simply kills Squid kid with a SIGBUS signal.
528 Some of the memory limits enforced by the kernel are currently
529 poorly understood: We do not know how to detect and check them. This
530 option ensures that the mapped memory will be available.
532 This option may have a positive performance side-effect: Locking
533 memory at start avoids runtime paging I/O. Paging slows Squid down.
535 Locking memory may require a large enough RLIMIT_MEMLOCK OS limit,
536 CAP_IPC_LOCK capability, or equivalent.
539 NAME: hopeless_kid_revival_delay
542 LOC: Config.hopelessKidRevivalDelay
545 Normally, when a kid process dies, Squid immediately restarts the
546 kid. A kid experiencing frequent deaths is marked as "hopeless" for
547 the duration specified by this directive. Hopeless kids are not
548 automatically restarted.
550 Currently, zero values are not supported because they result in
551 misconfigured SMP Squid instances running forever, endlessly
552 restarting each dying kid. To effectively disable hopeless kids
553 revival, set the delay to a huge value (e.g., 1 year).
555 Reconfiguration also clears all hopeless kids designations, allowing
556 for manual revival of hopeless kids.
560 OPTIONS FOR AUTHENTICATION
561 -----------------------------------------------------------------------------
567 LOC: Auth::TheConfig.schemes
570 This is used to define parameters for the various authentication
571 schemes supported by Squid.
573 format: auth_param scheme parameter [setting]
575 The order in which authentication schemes are presented to the client is
576 dependent on the order the scheme first appears in config file. IE
577 has a bug (it's not RFC 2617 compliant) in that it will use the basic
578 scheme if basic is the first entry presented, even if more secure
579 schemes are presented. For now use the order in the recommended
580 settings section below. If other browsers have difficulties (don't
581 recognize the schemes offered even if you are using basic) either
582 put basic first, or disable the other schemes (by commenting out their
585 Once an authentication scheme is fully configured, it can only be
586 shutdown by shutting squid down and restarting. Changes can be made on
587 the fly and activated with a reconfigure. I.E. You can change to a
588 different helper, but not unconfigure the helper completely.
590 Please note that while this directive defines how Squid processes
591 authentication it does not automatically activate authentication.
592 To use authentication you must in addition make use of ACLs based
593 on login name in http_access (proxy_auth, proxy_auth_regex or
594 external with %LOGIN used in the format tag). The browser will be
595 challenged for authentication on the first such acl encountered
596 in http_access processing and will also be re-challenged for new
597 login credentials if the request is being denied by a proxy_auth
600 WARNING: authentication can't be used in a transparently intercepting
601 proxy as the client then thinks it is talking to an origin server and
602 not the proxy. This is a limitation of bending the TCP/IP protocol to
603 transparently intercepting port 80, not a limitation in Squid.
604 Ports flagged 'transparent', 'intercept', or 'tproxy' have
605 authentication disabled.
607 === Parameters common to all schemes. ===
610 Specifies the command for the external authenticator.
612 By default, each authentication scheme is not used unless a
613 program is specified.
615 See http://wiki.squid-cache.org/Features/AddonHelpers for
616 more details on helper operations and creating your own.
619 Specifies a string to be append to request line format for
620 the authentication helper. "Quoted" format values may contain
621 spaces and logformat %macros. In theory, any logformat %macro
622 can be used. In practice, a %macro expands as a dash (-) if
623 the helper request is sent before the required macro
624 information is available to Squid.
626 By default, Squid uses request formats provided in
627 scheme-specific examples below (search for %credentials).
629 The expanded key_extras value is added to the Squid credentials
630 cache and, hence, will affect authentication. It can be used to
631 authenticate different users with identical user names (e.g.,
632 when user authentication depends on http_port).
634 Avoid adding frequently changing information to key_extras. For
635 example, if you add user source IP, and it changes frequently
636 in your environment, then max_user_ip ACL is going to treat
637 every user+IP combination as a unique "user", breaking the ACL
638 and wasting a lot of memory on those user records. It will also
639 force users to authenticate from scratch whenever their IP
643 Specifies the protection scope (aka realm name) which is to be
644 reported to the client for the authentication scheme. It is
645 commonly part of the text the user will see when prompted for
646 their username and password.
648 For Basic the default is "Squid proxy-caching web server".
649 For Digest there is no default, this parameter is mandatory.
650 For NTLM and Negotiate this parameter is ignored.
652 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
653 [queue-size=N] [on-persistent-overload=action]
654 [reservation-timeout=seconds]
656 The maximum number of authenticator processes to spawn. If
657 you start too few Squid will have to wait for them to process
658 a backlog of credential verifications, slowing it down. When
659 password verifications are done via a (slow) network you are
660 likely to need lots of authenticator processes.
662 The startup= and idle= options permit some skew in the exact
663 amount run. A minimum of startup=N will begin during startup
664 and reconfigure. Squid will start more in groups of up to
665 idle=N in an attempt to meet traffic needs and to keep idle=N
666 free above those traffic needs up to the maximum.
668 The concurrency= option sets the number of concurrent requests
669 the helper can process. The default of 0 is used for helpers
670 who only supports one request at a time. Setting this to a
671 number greater than 0 changes the protocol used to include a
672 channel ID field first on the request/response line, allowing
673 multiple requests to be sent to the same helper in parallel
674 without waiting for the response.
676 Concurrency must not be set unless it's known the helper
677 supports the input format with channel-ID fields.
679 The queue-size option sets the maximum number of queued
680 requests. A request is queued when no existing child can
681 accept it due to concurrency limit and no new child can be
682 started due to numberofchildren limit. The default maximum is
683 2*numberofchildren. Squid is allowed to temporarily exceed the
684 configured maximum, marking the affected helper as
685 "overloaded". If the helper overload lasts more than 3
686 minutes, the action prescribed by the on-persistent-overload
689 The on-persistent-overload=action option specifies Squid
690 reaction to a new helper request arriving when the helper
691 has been overloaded for more that 3 minutes already. The number
692 of queued requests determines whether the helper is overloaded
693 (see the queue-size option).
695 Two actions are supported:
697 die Squid worker quits. This is the default behavior.
699 ERR Squid treats the helper request as if it was
700 immediately submitted, and the helper immediately
701 replied with an ERR response. This action has no effect
702 on the already queued and in-progress helper requests.
704 NOTE: NTLM and Negotiate schemes do not support concurrency
705 in the Squid code module even though some helpers can.
707 The reservation-timeout=seconds option allows NTLM and Negotiate
708 helpers to forget about clients that abandon their in-progress
709 connection authentication without closing the connection. The
710 timeout is measured since the last helper response received by
711 Squid for the client. Fractional seconds are not supported.
713 After the timeout, the helper will be used for other clients if
714 there are no unreserved helpers available. In the latter case,
715 the old client attempt to resume authentication will not be
716 forwarded to the helper (and the client should open a new HTTP
717 connection and retry authentication from scratch).
719 By default, reservations do not expire and clients that keep
720 their connections open without completing authentication may
721 exhaust all NTLM and Negotiate helpers.
724 If you experience problems with PUT/POST requests when using
725 the NTLM or Negotiate schemes then you can try setting this
726 to off. This will cause Squid to forcibly close the connection
727 on the initial request where the browser asks which schemes
728 are supported by the proxy.
730 For Basic and Digest this parameter is ignored.
733 Useful for sending credentials to authentication backends that
734 expect UTF-8 encoding (e.g., LDAP).
736 When this option is enabled, Squid uses HTTP Accept-Language
737 request header to guess the received credentials encoding
738 (ISO-Latin-1, CP1251, or UTF-8) and then converts the first
739 two encodings into UTF-8.
741 When this option is disabled and by default, Squid sends
742 credentials in their original (i.e. received) encoding.
744 This parameter is only honored for Basic and Digest schemes.
745 For Basic, the entire username:password credentials are
746 checked and, if necessary, re-encoded. For Digest -- just the
747 username component. For NTLM and Negotiate schemes, this
748 parameter is ignored.
750 IF HAVE_AUTH_MODULE_BASIC
751 === Basic authentication parameters ===
753 "credentialsttl" timetolive
754 Specifies how long squid assumes an externally validated
755 username:password pair is valid for - in other words how
756 often the helper program is called for that user. Set this
757 low to force revalidation with short lived passwords.
759 NOTE: setting this high does not impact your susceptibility
760 to replay attacks unless you are using an one-time password
761 system (such as SecureID). If you are using such a system,
762 you will be vulnerable to replay attacks unless you also
763 use the max_user_ip ACL in an http_access rule.
765 "casesensitive" on|off
766 Specifies if usernames are case sensitive. Most user databases
767 are case insensitive allowing the same username to be spelled
768 using both lower and upper case letters, but some are case
769 sensitive. This makes a big difference for user_max_ip ACL
770 processing and similar.
773 IF HAVE_AUTH_MODULE_DIGEST
774 === Digest authentication parameters ===
776 "nonce_garbage_interval" timeinterval
777 Specifies the interval that nonces that have been issued
778 to client_agent's are checked for validity.
780 "nonce_max_duration" timeinterval
781 Specifies the maximum length of time a given nonce will be
784 "nonce_max_count" number
785 Specifies the maximum number of times a given nonce can be
788 "nonce_strictness" on|off
789 Determines if squid requires strict increment-by-1 behavior
790 for nonce counts, or just incrementing (off - for use when
791 user agents generate nonce counts that occasionally miss 1
792 (ie, 1,2,4,6)). Default off.
794 "check_nonce_count" on|off
795 This directive if set to off can disable the nonce count check
796 completely to work around buggy digest qop implementations in
797 certain mainstream browser versions. Default on to check the
798 nonce count to protect from authentication replay attacks.
800 "post_workaround" on|off
801 This is a workaround to certain buggy browsers who send an
802 incorrect request digest in POST requests when reusing the
803 same nonce as acquired earlier on a GET request.
807 === Example Configuration ===
809 This configuration displays the recommended authentication scheme
810 order from most to least secure with recommended minimum configuration
811 settings for each scheme:
813 #auth_param negotiate program <uncomment and complete this line to activate>
814 #auth_param negotiate children 20 startup=0 idle=1
816 #auth_param digest program <uncomment and complete this line to activate>
817 #auth_param digest children 20 startup=0 idle=1
818 #auth_param digest realm Squid proxy-caching web server
819 #auth_param digest nonce_garbage_interval 5 minutes
820 #auth_param digest nonce_max_duration 30 minutes
821 #auth_param digest nonce_max_count 50
823 #auth_param ntlm program <uncomment and complete this line to activate>
824 #auth_param ntlm children 20 startup=0 idle=1
826 #auth_param basic program <uncomment and complete this line>
827 #auth_param basic children 5 startup=5 idle=1
828 #auth_param basic credentialsttl 2 hours
831 NAME: authenticate_cache_garbage_interval
835 LOC: Auth::TheConfig.garbageCollectInterval
837 The time period between garbage collection across the username cache.
838 This is a trade-off between memory utilization (long intervals - say
839 2 days) and CPU (short intervals - say 1 minute). Only change if you
843 NAME: authenticate_ttl
847 LOC: Auth::TheConfig.credentialsTtl
849 The time a user & their credentials stay in the logged in
850 user cache since their last request. When the garbage
851 interval passes, all user credentials that have passed their
852 TTL are removed from memory.
855 NAME: authenticate_ip_ttl
858 LOC: Auth::TheConfig.ipTtl
861 If you use proxy authentication and the 'max_user_ip' ACL,
862 this directive controls how long Squid remembers the IP
863 addresses associated with each user. Use a small value
864 (e.g., 60 seconds) if your users might change addresses
865 quickly, as is the case with dialup. You might be safe
866 using a larger value (e.g., 2 hours) in a corporate LAN
867 environment with relatively static address assignments.
872 -----------------------------------------------------------------------------
875 NAME: external_acl_type
876 TYPE: externalAclHelper
877 LOC: Config.externalAclHelperList
880 This option defines external acl classes using a helper program
881 to look up the status
883 external_acl_type name [options] FORMAT /path/to/helper [helper arguments]
887 ttl=n TTL in seconds for cached results (defaults to 3600
891 TTL for cached negative lookups (default same
894 grace=n Percentage remaining of TTL where a refresh of a
895 cached entry should be initiated without needing to
896 wait for a new reply. (default is for no grace period)
898 cache=n The maximum number of entries in the result cache. The
899 default limit is 262144 entries. Each cache entry usually
900 consumes at least 256 bytes. Squid currently does not remove
901 expired cache entries until the limit is reached, so a proxy
902 will sooner or later reach the limit. The expanded FORMAT
903 value is used as the cache key, so if the details in FORMAT
904 are highly variable, a larger cache may be needed to produce
905 reduction in helper load.
908 Maximum number of acl helper processes spawned to service
909 external acl lookups of this type. (default 5)
912 Minimum number of acl helper processes to spawn during
913 startup and reconfigure to service external acl lookups
914 of this type. (default 0)
917 Number of acl helper processes to keep ahead of traffic
918 loads. Squid will spawn this many at once whenever load
919 rises above the capabilities of existing processes.
920 Up to the value of children-max. (default 1)
922 concurrency=n concurrency level per process. Only used with helpers
923 capable of processing more than one query at a time.
925 queue-size=N The queue-size option sets the maximum number of
926 queued requests. A request is queued when no existing
927 helper can accept it due to concurrency limit and no
928 new helper can be started due to children-max limit.
929 If the queued requests exceed queue size, the acl is
930 ignored. The default value is set to 2*children-max.
932 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
934 ipv4 / ipv6 IP protocol used to communicate with this helper.
935 The default is to auto-detect IPv6 and use it when available.
938 FORMAT is a series of %macro codes. See logformat directive for a full list
939 of the accepted codes. Although note that at the time of any external ACL
940 being tested data may not be available and thus some %macro expand to '-'.
942 In addition to the logformat codes; when processing external ACLs these
943 additional macros are made available:
945 %ACL The name of the ACL being tested.
947 %DATA The ACL arguments specified in the referencing config
948 'acl ... external' line, separated by spaces (an
949 "argument string"). see acl external.
951 If there are no ACL arguments %DATA expands to '-'.
953 If you do not specify a DATA macro inside FORMAT,
954 Squid automatically appends %DATA to your FORMAT.
955 Note that Squid-3.x may expand %DATA to whitespace
956 or nothing in this case.
958 By default, Squid applies URL-encoding to each ACL
959 argument inside the argument string. If an explicit
960 encoding modifier is used (e.g., %#DATA), then Squid
961 encodes the whole argument string as a single token
962 (e.g., with %#DATA, spaces between arguments become
965 If SSL is enabled, the following formatting codes become available:
967 %USER_CERT SSL User certificate in PEM format
968 %USER_CERTCHAIN SSL User certificate chain in PEM format
969 %USER_CERT_xx SSL User certificate subject attribute xx
970 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
973 NOTE: all other format codes accepted by older Squid versions
977 General request syntax:
979 [channel-ID] FORMAT-values
982 FORMAT-values consists of transaction details expanded with
983 whitespace separation per the config file FORMAT specification
984 using the FORMAT macros listed above.
986 Request values sent to the helper are URL escaped to protect
987 each value in requests against whitespaces.
989 If using protocol=2.5 then the request sent to the helper is not
990 URL escaped to protect against whitespace.
992 NOTE: protocol=3.0 is deprecated as no longer necessary.
994 When using the concurrency= option the protocol is changed by
995 introducing a query channel tag in front of the request/response.
996 The query channel tag is a number between 0 and concurrency-1.
997 This value must be echoed back unchanged to Squid as the first part
998 of the response relating to its request.
1001 The helper receives lines expanded per the above format specification
1002 and for each input line returns 1 line starting with OK/ERR/BH result
1003 code and optionally followed by additional keywords with more details.
1006 General result syntax:
1008 [channel-ID] result keyword=value ...
1010 Result consists of one of the codes:
1013 the ACL test produced a match.
1016 the ACL test does not produce a match.
1019 An internal error occurred in the helper, preventing
1020 a result being identified.
1022 The meaning of 'a match' is determined by your squid.conf
1023 access control configuration. See the Squid wiki for details.
1027 user= The users name (login)
1029 password= The users password (for login= cache_peer option)
1031 message= Message describing the reason for this response.
1032 Available as %o in error pages.
1033 Useful on (ERR and BH results).
1035 tag= Apply a tag to a request. Only sets a tag once,
1036 does not alter existing tags.
1038 log= String to be logged in access.log. Available as
1039 %ea in logformat specifications.
1041 clt_conn_tag= Associates a TAG with the client TCP connection.
1042 Please see url_rewrite_program related documentation
1045 Any keywords may be sent on any response whether OK, ERR or BH.
1047 All response keyword values need to be a single token with URL
1048 escaping, or enclosed in double quotes (") and escaped using \ on
1049 any double quotes or \ characters within the value. The wrapping
1050 double quotes are removed before the value is interpreted by Squid.
1051 \r and \n are also replace by CR and LF.
1053 Some example key values:
1057 user="J. \"Bob\" Smith"
1064 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
1065 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
1066 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
1067 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
1068 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
1070 DEFAULT: all src all
1071 DEFAULT: manager url_regex -i ^cache_object:// +i ^[^:]+://[^/]+/squid-internal-mgr/
1072 DEFAULT: localhost src 127.0.0.1/32 ::1
1073 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1/128 ::/128
1074 DEFAULT: CONNECT method CONNECT
1075 DEFAULT_DOC: ACLs all, manager, localhost, to_localhost, and CONNECT are predefined.
1077 Defining an Access List
1079 Every access list definition must begin with an aclname and acltype,
1080 followed by either type-specific arguments or a quoted filename that
1083 acl aclname acltype argument ...
1084 acl aclname acltype "file" ...
1086 When using "file", the file should contain one item per line.
1091 Some acl types supports options which changes their default behaviour:
1093 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
1094 case-insensitive, use the -i option. To return case-sensitive
1095 use the +i option between patterns, or make a new ACL line
1098 -n Disable lookups and address type conversions. If lookup or
1099 conversion is required because the parameter type (IP or
1100 domain name) does not match the message address type (domain
1101 name or IP), then the ACL would immediately declare a mismatch
1102 without any warnings or lookups.
1105 Perform a list membership test, interpreting values as
1106 comma-separated token lists and matching against individual
1107 tokens instead of whole values.
1108 The optional "delimiters" parameter specifies one or more
1109 alternative non-alphanumeric delimiter characters.
1110 non-alphanumeric delimiter characters.
1112 -- Used to stop processing all options, in the case the first acl
1113 value has '-' character as first character (for example the '-'
1114 is a valid domain name)
1116 Some acl types require suspending the current request in order
1117 to access some external data source.
1118 Those which do are marked with the tag [slow], those which
1119 don't are marked as [fast].
1120 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
1121 for further information
1123 ***** ACL TYPES AVAILABLE *****
1125 acl aclname src ip-address/mask ... # clients IP address [fast]
1126 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
1127 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
1128 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
1131 acl aclname arp mac-address ...
1132 acl aclname eui64 eui64-address ...
1134 # MAC (EUI-48) and EUI-64 addresses use xx:xx:xx:xx:xx:xx notation.
1136 # The 'arp' ACL code is not portable to all operating systems.
1137 # It works on Linux, Solaris, Windows, FreeBSD, and some other
1140 # The eui_lookup directive is required to be 'on' (the default)
1141 # and Squid built with --enable-eui for MAC/EUI addresses to be
1142 # available for this ACL.
1144 # Squid can only determine the MAC/EUI address for IPv4
1145 # clients that are on the same subnet. If the client is on a
1146 # different subnet, then Squid cannot find out its address.
1148 # IPv6 protocol does not contain ARP. MAC/EUI is either
1149 # encoded directly in the IPv6 address or not available.
1151 acl aclname clientside_mark mark[/mask] ...
1152 # matches CONNMARK of an accepted connection [fast]
1153 # DEPRECATED. Use the 'client_connection_mark' instead.
1155 acl aclname client_connection_mark mark[/mask] ...
1156 # matches CONNMARK of an accepted connection [fast]
1158 # mark and mask are unsigned integers (hex, octal, or decimal).
1159 # If multiple marks are given, then the ACL matches if at least
1162 # Uses netfilter-conntrack library.
1163 # Requires building Squid with --enable-linux-netfilter.
1165 # The client, various intermediaries, and Squid itself may set
1166 # CONNMARK at various times. The last CONNMARK set wins. This ACL
1167 # checks the mark present on an accepted connection or set by
1168 # Squid afterwards, depending on the ACL check timing. This ACL
1169 # effectively ignores any mark set by other agents after Squid has
1170 # accepted the connection.
1172 acl aclname srcdomain .foo.com ...
1173 # reverse lookup, from client IP [slow]
1174 acl aclname dstdomain [-n] .foo.com ...
1175 # Destination server from URL [fast]
1176 acl aclname srcdom_regex [-i] \.foo\.com ...
1177 # regex matching client name [slow]
1178 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
1179 # regex matching server [fast]
1181 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
1182 # based URL is used and no match is found. The name "none" is used
1183 # if the reverse lookup fails.
1185 acl aclname src_as number ...
1186 acl aclname dst_as number ...
1188 # Except for access control, AS numbers can be used for
1189 # routing of requests to specific caches. Here's an
1190 # example for routing all requests for AS#1241 and only
1191 # those to mycache.mydomain.net:
1192 # acl asexample dst_as 1241
1193 # cache_peer_access mycache.mydomain.net allow asexample
1194 # cache_peer_access mycache_mydomain.net deny all
1196 acl aclname peername myPeer ...
1197 acl aclname peername_regex [-i] regex-pattern ...
1199 # match against a named cache_peer entry
1200 # set unique name= on cache_peer lines for reliable use.
1202 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
1212 # h1:m1 must be less than h2:m2
1214 acl aclname url_regex [-i] ^http:// ...
1215 # regex matching on whole URL [fast]
1216 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
1217 # regex matching on URL login field
1218 acl aclname urlpath_regex [-i] \.gif$ ...
1219 # regex matching on URL path [fast]
1221 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
1222 # ranges are allowed
1223 acl aclname localport 3128 ... # TCP port the client connected to [fast]
1224 # NP: for interception mode this is usually '80'
1226 acl aclname myportname 3128 ... # *_port name [fast]
1228 acl aclname proto HTTP FTP ... # request protocol [fast]
1230 acl aclname method GET POST ... # HTTP request method [fast]
1232 acl aclname http_status 200 301 500- 400-403 ...
1233 # status code in reply [fast]
1235 acl aclname browser [-i] regexp ...
1236 # pattern match on User-Agent header (see also req_header below) [fast]
1238 acl aclname referer_regex [-i] regexp ...
1239 # pattern match on Referer header [fast]
1240 # Referer is highly unreliable, so use with care
1242 acl aclname ident [-i] username ...
1243 acl aclname ident_regex [-i] pattern ...
1244 # string match on ident output [slow]
1245 # use REQUIRED to accept any non-null ident.
1247 acl aclname proxy_auth [-i] username ...
1248 acl aclname proxy_auth_regex [-i] pattern ...
1249 # perform http authentication challenge to the client and match against
1250 # supplied credentials [slow]
1252 # takes a list of allowed usernames.
1253 # use REQUIRED to accept any valid username.
1255 # Will use proxy authentication in forward-proxy scenarios, and plain
1256 # http authentication in reverse-proxy scenarios
1258 # NOTE: when a Proxy-Authentication header is sent but it is not
1259 # needed during ACL checking the username is NOT logged
1262 # NOTE: proxy_auth requires a EXTERNAL authentication program
1263 # to check username/password combinations (see
1264 # auth_param directive).
1266 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1267 # as the browser needs to be configured for using a proxy in order
1268 # to respond to proxy authentication.
1270 acl aclname snmp_community string ...
1271 # A community string to limit access to your SNMP Agent [fast]
1274 # acl snmppublic snmp_community public
1276 acl aclname maxconn number
1277 # This will be matched when the client's IP address has
1278 # more than <number> TCP connections established. [fast]
1279 # NOTE: This only measures direct TCP links so X-Forwarded-For
1280 # indirect clients are not counted.
1282 acl aclname max_user_ip [-s] number
1283 # This will be matched when the user attempts to log in from more
1284 # than <number> different ip addresses. The authenticate_ip_ttl
1285 # parameter controls the timeout on the ip entries. [fast]
1286 # If -s is specified the limit is strict, denying browsing
1287 # from any further IP addresses until the ttl has expired. Without
1288 # -s Squid will just annoy the user by "randomly" denying requests.
1289 # (the counter is reset each time the limit is reached and a
1290 # request is denied)
1291 # NOTE: in acceleration mode or where there is mesh of child proxies,
1292 # clients may appear to come from multiple addresses if they are
1293 # going through proxy farms, so a limit of 1 may cause user problems.
1295 acl aclname random probability
1296 # Pseudo-randomly match requests. Based on the probability given.
1297 # Probability may be written as a decimal (0.333), fraction (1/3)
1298 # or ratio of matches:non-matches (3:5).
1300 acl aclname req_mime_type [-i] mime-type ...
1301 # regex match against the mime type of the request generated
1302 # by the client. Can be used to detect file upload or some
1303 # types HTTP tunneling requests [fast]
1304 # NOTE: This does NOT match the reply. You cannot use this
1305 # to match the returned file type.
1307 acl aclname req_header header-name [-i] any\.regex\.here
1308 # regex match against any of the known request headers. May be
1309 # thought of as a superset of "browser", "referer" and "mime-type"
1312 acl aclname rep_mime_type [-i] mime-type ...
1313 # regex match against the mime type of the reply received by
1314 # squid. Can be used to detect file download or some
1315 # types HTTP tunneling requests. [fast]
1316 # NOTE: This has no effect in http_access rules. It only has
1317 # effect in rules that affect the reply data stream such as
1318 # http_reply_access.
1320 acl aclname rep_header header-name [-i] any\.regex\.here
1321 # regex match against any of the known reply headers. May be
1322 # thought of as a superset of "browser", "referer" and "mime-type"
1325 acl aclname external class_name [arguments...]
1326 # external ACL lookup via a helper class defined by the
1327 # external_acl_type directive [slow]
1329 acl aclname user_cert attribute values...
1330 # match against attributes in a user SSL certificate
1331 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1333 acl aclname ca_cert attribute values...
1334 # match against attributes a users issuing CA SSL certificate
1335 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1337 acl aclname ext_user [-i] username ...
1338 acl aclname ext_user_regex [-i] pattern ...
1339 # string match on username returned by external acl helper [slow]
1340 # use REQUIRED to accept any non-null user name.
1342 acl aclname tag tagvalue ...
1343 # string match on tag returned by external acl helper [fast]
1344 # DEPRECATED. Only the first tag will match with this ACL.
1345 # Use the 'note' ACL instead for handling multiple tag values.
1347 acl aclname hier_code codename ...
1348 # string match against squid hierarchy code(s); [fast]
1349 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1351 # NOTE: This has no effect in http_access rules. It only has
1352 # effect in rules that affect the reply data stream such as
1353 # http_reply_access.
1355 acl aclname note [-m[=delimiters]] name [value ...]
1356 # match transaction annotation [fast]
1357 # Without values, matches any annotation with a given name.
1358 # With value(s), matches any annotation with a given name that
1359 # also has one of the given values.
1360 # If the -m flag is used, then the value of the named
1361 # annotation is interpreted as a list of tokens, and the ACL
1362 # matches individual name=token pairs rather than whole
1363 # name=value pairs. See "ACL Options" above for more info.
1364 # Annotation sources include note and adaptation_meta directives
1365 # as well as helper and eCAP responses.
1367 acl aclname annotate_transaction [-m[=delimiters]] key=value ...
1368 acl aclname annotate_transaction [-m[=delimiters]] key+=value ...
1369 # Always matches. [fast]
1370 # Used for its side effect: This ACL immediately adds a
1371 # key=value annotation to the current master transaction.
1372 # The added annotation can then be tested using note ACL and
1373 # logged (or sent to helpers) using %note format code.
1375 # Annotations can be specified using replacement and addition
1376 # formats. The key=value form replaces old same-key annotation
1377 # value(s). The key+=value form appends a new value to the old
1378 # same-key annotation. Both forms create a new key=value
1379 # annotation if no same-key annotation exists already. If
1380 # -m flag is used, then the value is interpreted as a list
1381 # and the annotation will contain key=token pair(s) instead of the
1382 # whole key=value pair.
1384 # This ACL is especially useful for recording complex multi-step
1385 # ACL-driven decisions. For example, the following configuration
1386 # avoids logging transactions accepted after aclX matched:
1388 # # First, mark transactions accepted after aclX matched
1389 # acl markSpecial annotate_transaction special=true
1390 # http_access allow acl001
1392 # http_access deny acl100
1393 # http_access allow aclX markSpecial
1395 # # Second, do not log marked transactions:
1396 # acl markedSpecial note special true
1397 # access_log ... deny markedSpecial
1399 # # Note that the following would not have worked because aclX
1400 # # alone does not determine whether the transaction was allowed:
1401 # access_log ... deny aclX # Wrong!
1403 # Warning: This ACL annotates the transaction even when negated
1404 # and even if subsequent ACLs fail to match. For example, the
1405 # following three rules will have exactly the same effect as far
1406 # as annotations set by the "mark" ACL are concerned:
1408 # some_directive acl1 ... mark # rule matches if mark is reached
1409 # some_directive acl1 ... !mark # rule never matches
1410 # some_directive acl1 ... mark !all # rule never matches
1412 acl aclname annotate_client [-m[=delimiters]] key=value ...
1413 acl aclname annotate_client [-m[=delimiters]] key+=value ...
1415 # Always matches. [fast]
1416 # Used for its side effect: This ACL immediately adds a
1417 # key=value annotation to the current client-to-Squid
1418 # connection. Connection annotations are propagated to the current
1419 # and all future master transactions on the annotated connection.
1420 # See the annotate_transaction ACL for details.
1422 # For example, the following configuration avoids rewriting URLs
1423 # of transactions bumped by SslBump:
1425 # # First, mark bumped connections:
1426 # acl markBumped annotate_client bumped=true
1427 # ssl_bump peek acl1
1428 # ssl_bump stare acl2
1429 # ssl_bump bump acl3 markBumped
1430 # ssl_bump splice all
1432 # # Second, do not send marked transactions to the redirector:
1433 # acl markedBumped note bumped true
1434 # url_rewrite_access deny markedBumped
1436 # # Note that the following would not have worked because acl3 alone
1437 # # does not determine whether the connection is going to be bumped:
1438 # url_rewrite_access deny acl3 # Wrong!
1440 acl aclname adaptation_service service ...
1441 # Matches the name of any icap_service, ecap_service,
1442 # adaptation_service_set, or adaptation_service_chain that Squid
1443 # has used (or attempted to use) for the master transaction.
1444 # This ACL must be defined after the corresponding adaptation
1445 # service is named in squid.conf. This ACL is usable with
1446 # adaptation_meta because it starts matching immediately after
1447 # the service has been selected for adaptation.
1449 acl aclname transaction_initiator initiator ...
1450 # Matches transaction's initiator [fast]
1452 # Supported initiators are:
1453 # esi: matches transactions fetching ESI resources
1454 # certificate-fetching: matches transactions fetching
1455 # a missing intermediate TLS certificate
1456 # cache-digest: matches transactions fetching Cache Digests
1458 # htcp: matches HTCP requests from peers
1459 # icp: matches ICP requests to peers
1460 # icmp: matches ICMP RTT database (NetDB) requests to peers
1461 # asn: matches asns db requests
1462 # internal: matches any of the above
1463 # client: matches transactions containing an HTTP or FTP
1464 # client request received at a Squid *_port
1465 # all: matches any transaction, including internal transactions
1466 # without a configurable initiator and hopefully rare
1467 # transactions without a known-to-Squid initiator
1469 # Multiple initiators are ORed.
1471 acl aclname has component
1472 # matches a transaction "component" [fast]
1474 # Supported transaction components are:
1475 # request: transaction has a request header (at least)
1476 # response: transaction has a response header (at least)
1477 # ALE: transaction has an internally-generated Access Log Entry
1478 # structure; bugs notwithstanding, all transaction have it
1480 # For example, the following configuration helps when dealing with HTTP
1481 # clients that close connections without sending a request header:
1483 # acl hasRequest has request
1484 # acl logMe note important_transaction
1485 # # avoid "logMe ACL is used in context without an HTTP request" warnings
1486 # access_log ... logformat=detailed hasRequest logMe
1487 # # log request-less transactions, instead of ignoring them
1488 # access_log ... logformat=brief !hasRequest
1490 # Multiple components are not supported for one "acl" rule, but
1491 # can be specified (and are ORed) using multiple same-name rules:
1493 # # OK, this strange logging daemon needs request or response,
1494 # # but can work without either a request or a response:
1495 # acl hasWhatMyLoggingDaemonNeeds has request
1496 # acl hasWhatMyLoggingDaemonNeeds has response
1498 acl aclname at_step step
1499 # match against the current request processing step [fast]
1501 # GeneratingCONNECT: Generating HTTP CONNECT request headers
1503 # The following ssl_bump processing steps are recognized:
1504 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1505 # SslBump2: After getting SSL Client Hello info.
1506 # SslBump3: After getting SSL Server Hello info.
1510 acl aclname ssl_error errorname
1511 # match against SSL certificate validation error [fast]
1513 # When used with sslproxy_cert_error, this ACL tests a single
1514 # certificate validation error currently being evaluated by that
1515 # directive. When used with slproxy_cert_sign or sslproxy_cert_adapt,
1516 # the ACL tests all past certificate validation errors associated with
1517 # the current Squid-to-server connection (attempt). This ACL is not yet
1518 # supported for use with other directives.
1520 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1523 # The following can be used as shortcuts for certificate properties:
1524 # [ssl::]certHasExpired: the "not after" field is in the past
1525 # [ssl::]certNotYetValid: the "not before" field is in the future
1526 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1527 # [ssl::]certSelfSigned: The certificate is self signed.
1528 # [ssl::]certDomainMismatch: The certificate CN domain does not
1529 # match the name the name of the host we are connecting to.
1531 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1532 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1533 # predefined ACLs, just like the 'all' ACL.
1535 acl aclname server_cert_fingerprint [-sha1] fingerprint
1536 # match against server SSL certificate fingerprint [fast]
1538 # The fingerprint is the digest of the DER encoded version
1539 # of the whole certificate. The user should use the form: XX:XX:...
1540 # Optional argument specifies the digest algorithm to use.
1541 # The SHA1 digest algorithm is the default and is currently
1542 # the only algorithm supported (-sha1).
1544 acl aclname ssl::server_name [option] .foo.com ...
1545 # matches server name obtained from various sources [fast]
1547 # The ACL computes server name(s) using such information sources as
1548 # CONNECT request URI, TLS client SNI, and TLS server certificate
1549 # subject (CN and SubjectAltName). The computed server name(s) usually
1550 # change with each SslBump step, as more info becomes available:
1551 # * SNI is used as the server name instead of the request URI,
1552 # * subject name(s) from the server certificate (CN and
1553 # SubjectAltName) are used as the server names instead of SNI.
1555 # When the ACL computes multiple server names, matching any single
1556 # computed name is sufficient for the ACL to match.
1558 # The "none" name can be used to match transactions where the ACL
1559 # could not compute the server name using any information source
1560 # that was both available and allowed to be used by the ACL options at
1561 # the ACL evaluation time.
1563 # Unlike dstdomain, this ACL does not perform DNS lookups.
1565 # An ACL option below may be used to restrict what information
1566 # sources are used to extract the server names from:
1568 # --client-requested
1569 # The server name is SNI regardless of what the server says.
1571 # The server name(s) are the certificate subject name(s), regardless
1572 # of what the client has requested. If the server certificate is
1573 # unavailable, then the name is "none".
1575 # The server name is either SNI (if SNI matches at least one of the
1576 # certificate subject names) or "none" (otherwise). When the server
1577 # certificate is unavailable, the consensus server name is SNI.
1579 # Combining multiple options in one ACL is a fatal configuration
1582 # For all options: If no SNI is available, then the CONNECT request
1583 # target (a.k.a. URI) is used instead of SNI (for an intercepted
1584 # connection, this target is the destination IP address).
1586 acl aclname ssl::server_name_regex [-i] \.foo\.com ...
1587 # regex matches server name obtained from various sources [fast]
1589 acl aclname connections_encrypted
1590 # matches transactions with all HTTP messages received over TLS
1591 # transport connections. [fast]
1593 # The master transaction deals with HTTP messages received from
1594 # various sources. All sources used by the master transaction in the
1595 # past are considered by the ACL. The following rules define whether
1596 # a given message source taints the entire master transaction,
1597 # resulting in ACL mismatches:
1599 # * The HTTP client transport connection is not TLS.
1600 # * An adaptation service connection-encryption flag is off.
1601 # * The peer or origin server transport connection is not TLS.
1603 # Caching currently does not affect these rules. This cache ignorance
1604 # implies that only the current HTTP client transport and REQMOD
1605 # services status determine whether this ACL matches a from-cache
1606 # transaction. The source of the cached response does not have any
1607 # effect on future transaction that use the cached response without
1608 # revalidation. This may change.
1610 # DNS, ICP, and HTCP exchanges during the master transaction do not
1611 # affect these rules.
1613 acl aclname any-of acl1 acl2 ...
1614 # match any one of the acls [fast or slow]
1615 # The first matching ACL stops further ACL evaluation.
1617 # ACLs from multiple any-of lines with the same name are ORed.
1618 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1619 # acl A any-of a1 a2
1620 # acl A any-of a3 a4
1622 # This group ACL is fast if all evaluated ACLs in the group are fast
1623 # and slow otherwise.
1625 acl aclname all-of acl1 acl2 ...
1626 # match all of the acls [fast or slow]
1627 # The first mismatching ACL stops further ACL evaluation.
1629 # ACLs from multiple all-of lines with the same name are ORed.
1630 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1631 # acl B all-of b1 b2
1632 # acl B all-of b3 b4
1634 # This group ACL is fast if all evaluated ACLs in the group are fast
1635 # and slow otherwise.
1638 acl macaddress arp 09:00:2b:23:45:67
1639 acl myexample dst_as 1241
1640 acl password proxy_auth REQUIRED
1641 acl fileupload req_mime_type -i ^multipart/form-data$
1642 acl javascript rep_mime_type -i ^application/x-javascript$
1646 # Recommended minimum configuration:
1649 # Example rule allowing access from your local networks.
1650 # Adapt to list your (internal) IP networks from where browsing
1652 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1653 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1654 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1655 acl localnet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1656 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1657 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1658 acl localnet src fc00::/7 # RFC 4193 local private network range
1659 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1661 acl SSL_ports port 443
1662 acl Safe_ports port 80 # http
1663 acl Safe_ports port 21 # ftp
1664 acl Safe_ports port 443 # https
1665 acl Safe_ports port 70 # gopher
1666 acl Safe_ports port 210 # wais
1667 acl Safe_ports port 1025-65535 # unregistered ports
1668 acl Safe_ports port 280 # http-mgmt
1669 acl Safe_ports port 488 # gss-http
1670 acl Safe_ports port 591 # filemaker
1671 acl Safe_ports port 777 # multiling http
1675 NAME: proxy_protocol_access
1677 LOC: Config.accessList.proxyProtocol
1679 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1681 Determine which client proxies can be trusted to provide correct
1682 information regarding real client IP address using PROXY protocol.
1684 Requests may pass through a chain of several other proxies
1685 before reaching us. The original source details may by sent in:
1686 * HTTP message Forwarded header, or
1687 * HTTP message X-Forwarded-For header, or
1688 * PROXY protocol connection header.
1690 This directive is solely for validating new PROXY protocol
1691 connections received from a port flagged with require-proxy-header.
1692 It is checked only once after TCP connection setup.
1694 A deny match results in TCP connection closure.
1696 An allow match is required for Squid to permit the corresponding
1697 TCP connection, before Squid even looks for HTTP request headers.
1698 If there is an allow match, Squid starts using PROXY header information
1699 to determine the source address of the connection for all future ACL
1700 checks, logging, etc.
1702 SECURITY CONSIDERATIONS:
1704 Any host from which we accept client IP details can place
1705 incorrect information in the relevant header, and Squid
1706 will use the incorrect information as if it were the
1707 source address of the request. This may enable remote
1708 hosts to bypass any access control restrictions that are
1709 based on the client's source addresses.
1711 This clause only supports fast acl types.
1712 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1715 NAME: follow_x_forwarded_for
1717 IFDEF: FOLLOW_X_FORWARDED_FOR
1718 LOC: Config.accessList.followXFF
1719 DEFAULT_IF_NONE: deny all
1720 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1722 Determine which client proxies can be trusted to provide correct
1723 information regarding real client IP address.
1725 Requests may pass through a chain of several other proxies
1726 before reaching us. The original source details may by sent in:
1727 * HTTP message Forwarded header, or
1728 * HTTP message X-Forwarded-For header, or
1729 * PROXY protocol connection header.
1731 PROXY protocol connections are controlled by the proxy_protocol_access
1732 directive which is checked before this.
1734 If a request reaches us from a source that is allowed by this
1735 directive, then we trust the information it provides regarding
1736 the IP of the client it received from (if any).
1738 For the purpose of ACLs used in this directive the src ACL type always
1739 matches the address we are testing and srcdomain matches its rDNS.
1741 On each HTTP request Squid checks for X-Forwarded-For header fields.
1742 If found the header values are iterated in reverse order and an allow
1743 match is required for Squid to continue on to the next value.
1744 The verification ends when a value receives a deny match, cannot be
1745 tested, or there are no more values to test.
1746 NOTE: Squid does not yet follow the Forwarded HTTP header.
1748 The end result of this process is an IP address that we will
1749 refer to as the indirect client address. This address may
1750 be treated as the client address for access control, ICAP, delay
1751 pools and logging, depending on the acl_uses_indirect_client,
1752 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1753 log_uses_indirect_client and tproxy_uses_indirect_client options.
1755 This clause only supports fast acl types.
1756 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1758 SECURITY CONSIDERATIONS:
1760 Any host from which we accept client IP details can place
1761 incorrect information in the relevant header, and Squid
1762 will use the incorrect information as if it were the
1763 source address of the request. This may enable remote
1764 hosts to bypass any access control restrictions that are
1765 based on the client's source addresses.
1769 acl localhost src 127.0.0.1
1770 acl my_other_proxy srcdomain .proxy.example.com
1771 follow_x_forwarded_for allow localhost
1772 follow_x_forwarded_for allow my_other_proxy
1775 NAME: acl_uses_indirect_client
1778 IFDEF: FOLLOW_X_FORWARDED_FOR
1780 LOC: Config.onoff.acl_uses_indirect_client
1782 Controls whether the indirect client address
1783 (see follow_x_forwarded_for) is used instead of the
1784 direct client address in acl matching.
1786 NOTE: maxconn ACL considers direct TCP links and indirect
1787 clients will always have zero. So no match.
1790 NAME: delay_pool_uses_indirect_client
1793 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1795 LOC: Config.onoff.delay_pool_uses_indirect_client
1797 Controls whether the indirect client address
1798 (see follow_x_forwarded_for) is used instead of the
1799 direct client address in delay pools.
1802 NAME: log_uses_indirect_client
1805 IFDEF: FOLLOW_X_FORWARDED_FOR
1807 LOC: Config.onoff.log_uses_indirect_client
1809 Controls whether the indirect client address
1810 (see follow_x_forwarded_for) is used instead of the
1811 direct client address in the access log.
1814 NAME: tproxy_uses_indirect_client
1817 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1819 LOC: Config.onoff.tproxy_uses_indirect_client
1821 Controls whether the indirect client address
1822 (see follow_x_forwarded_for) is used instead of the
1823 direct client address when spoofing the outgoing client.
1825 This has no effect on requests arriving in non-tproxy
1828 SECURITY WARNING: Usage of this option is dangerous
1829 and should not be used trivially. Correct configuration
1830 of follow_x_forwarded_for with a limited set of trusted
1831 sources is required to prevent abuse of your proxy.
1834 NAME: spoof_client_ip
1836 LOC: Config.accessList.spoof_client_ip
1838 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1840 Control client IP address spoofing of TPROXY traffic based on
1841 defined access lists.
1843 spoof_client_ip allow|deny [!]aclname ...
1845 If there are no "spoof_client_ip" lines present, the default
1846 is to "allow" spoofing of any suitable request.
1848 Note that the cache_peer "no-tproxy" option overrides this ACL.
1850 This clause supports fast acl types.
1851 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1856 LOC: Config.accessList.http
1857 DEFAULT_IF_NONE: deny all
1858 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1860 Allowing or Denying access based on defined access lists
1862 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1863 http_access allow|deny [!]aclname ...
1865 NOTE on default values:
1867 If there are no "access" lines present, the default is to deny
1870 If none of the "access" lines cause a match, the default is the
1871 opposite of the last line in the list. If the last line was
1872 deny, the default is allow. Conversely, if the last line
1873 is allow, the default will be deny. For these reasons, it is a
1874 good idea to have an "deny all" entry at the end of your access
1875 lists to avoid potential confusion.
1877 This clause supports both fast and slow acl types.
1878 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1883 # Recommended minimum Access Permission configuration:
1885 # Deny requests to certain unsafe ports
1886 http_access deny !Safe_ports
1888 # Deny CONNECT to other than secure SSL ports
1889 http_access deny CONNECT !SSL_ports
1891 # Only allow cachemgr access from localhost
1892 http_access allow localhost manager
1893 http_access deny manager
1895 # We strongly recommend the following be uncommented to protect innocent
1896 # web applications running on the proxy server who think the only
1897 # one who can access services on "localhost" is a local user
1898 #http_access deny to_localhost
1901 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1904 # Example rule allowing access from your local networks.
1905 # Adapt localnet in the ACL section to list your (internal) IP networks
1906 # from where browsing should be allowed
1907 http_access allow localnet
1908 http_access allow localhost
1910 # And finally deny all other access to this proxy
1911 http_access deny all
1915 NAME: adapted_http_access http_access2
1917 LOC: Config.accessList.adapted_http
1919 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1921 Allowing or Denying access based on defined access lists
1923 Essentially identical to http_access, but runs after redirectors
1924 and ICAP/eCAP adaptation. Allowing access control based on their
1927 If not set then only http_access is used.
1930 NAME: http_reply_access
1932 LOC: Config.accessList.reply
1934 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1936 Allow replies to client requests. This is complementary to http_access.
1938 http_reply_access allow|deny [!] aclname ...
1940 NOTE: if there are no access lines present, the default is to allow
1943 If none of the access lines cause a match the opposite of the
1944 last line will apply. Thus it is good practice to end the rules
1945 with an "allow all" or "deny all" entry.
1947 This clause supports both fast and slow acl types.
1948 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1953 LOC: Config.accessList.icp
1955 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1957 Allowing or Denying access to the ICP port based on defined
1960 icp_access allow|deny [!]aclname ...
1962 NOTE: The default if no icp_access lines are present is to
1963 deny all traffic. This default may cause problems with peers
1966 This clause only supports fast acl types.
1967 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1969 # Allow ICP queries from local networks only
1970 #icp_access allow localnet
1971 #icp_access deny all
1977 LOC: Config.accessList.htcp
1979 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1981 Allowing or Denying access to the HTCP port based on defined
1984 htcp_access allow|deny [!]aclname ...
1986 See also htcp_clr_access for details on access control for
1987 cache purge (CLR) HTCP messages.
1989 NOTE: The default if no htcp_access lines are present is to
1990 deny all traffic. This default may cause problems with peers
1991 using the htcp option.
1993 This clause only supports fast acl types.
1994 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1996 # Allow HTCP queries from local networks only
1997 #htcp_access allow localnet
1998 #htcp_access deny all
2001 NAME: htcp_clr_access
2004 LOC: Config.accessList.htcp_clr
2006 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
2008 Allowing or Denying access to purge content using HTCP based
2009 on defined access lists.
2010 See htcp_access for details on general HTCP access control.
2012 htcp_clr_access allow|deny [!]aclname ...
2014 This clause only supports fast acl types.
2015 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2017 # Allow HTCP CLR requests from trusted peers
2018 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
2019 htcp_clr_access allow htcp_clr_peer
2020 htcp_clr_access deny all
2025 LOC: Config.accessList.miss
2027 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
2029 Determines whether network access is permitted when satisfying a request.
2032 to force your neighbors to use you as a sibling instead of
2035 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
2036 miss_access deny !localclients
2037 miss_access allow all
2039 This means only your local clients are allowed to fetch relayed/MISS
2040 replies from the network and all other clients can only fetch cached
2043 The default for this setting allows all clients who passed the
2044 http_access rules to relay via this proxy.
2046 This clause only supports fast acl types.
2047 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2050 NAME: ident_lookup_access
2054 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
2055 LOC: Ident::TheConfig.identLookup
2057 A list of ACL elements which, if matched, cause an ident
2058 (RFC 931) lookup to be performed for this request. For
2059 example, you might choose to always perform ident lookups
2060 for your main multi-user Unix boxes, but not for your Macs
2061 and PCs. By default, ident lookups are not performed for
2064 To enable ident lookups for specific client addresses, you
2065 can follow this example:
2067 acl ident_aware_hosts src 198.168.1.0/24
2068 ident_lookup_access allow ident_aware_hosts
2069 ident_lookup_access deny all
2071 Only src type ACL checks are fully supported. A srcdomain
2072 ACL might work at times, but it will not always provide
2075 This clause only supports fast acl types.
2076 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2079 NAME: reply_body_max_size
2080 COMMENT: size [acl acl...]
2083 DEFAULT_DOC: No limit is applied.
2084 LOC: Config.ReplyBodySize
2086 This option specifies the maximum size of a reply body. It can be
2087 used to prevent users from downloading very large files, such as
2088 MP3's and movies. When the reply headers are received, the
2089 reply_body_max_size lines are processed, and the first line where
2090 all (if any) listed ACLs are true is used as the maximum body size
2093 This size is checked twice. First when we get the reply headers,
2094 we check the content-length value. If the content length value exists
2095 and is larger than the allowed size, the request is denied and the
2096 user receives an error message that says "the request or reply
2097 is too large." If there is no content-length, and the reply
2098 size exceeds this limit, the client's connection is just closed
2099 and they will receive a partial reply.
2101 WARNING: downstream caches probably can not detect a partial reply
2102 if there is no content-length header, so they will cache
2103 partial responses and give them out as hits. You should NOT
2104 use this option if you have downstream caches.
2106 WARNING: A maximum size smaller than the size of squid's error messages
2107 will cause an infinite loop and crash squid. Ensure that the smallest
2108 non-zero value you use is greater that the maximum header size plus
2109 the size of your largest error page.
2111 If you set this parameter none (the default), there will be
2114 Configuration Format is:
2115 reply_body_max_size SIZE UNITS [acl ...]
2117 reply_body_max_size 10 MB
2121 NAME: on_unsupported_protocol
2122 TYPE: on_unsupported_protocol
2123 LOC: Config.accessList.on_unsupported_protocol
2125 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
2127 Determines Squid behavior when encountering strange requests at the
2128 beginning of an accepted TCP connection or the beginning of a bumped
2129 CONNECT tunnel. Controlling Squid reaction to unexpected traffic is
2130 especially useful in interception environments where Squid is likely
2131 to see connections for unsupported protocols that Squid should either
2132 terminate or tunnel at TCP level.
2134 on_unsupported_protocol <action> [!]acl ...
2136 The first matching action wins. Only fast ACLs are supported.
2138 Supported actions are:
2140 tunnel: Establish a TCP connection with the intended server and
2141 blindly shovel TCP packets between the client and server.
2143 respond: Respond with an error message, using the transfer protocol
2144 for the Squid port that received the request (e.g., HTTP
2145 for connections intercepted at the http_port). This is the
2148 Squid expects the following traffic patterns:
2150 http_port: a plain HTTP request
2151 https_port: SSL/TLS handshake followed by an [encrypted] HTTP request
2152 ftp_port: a plain FTP command (no on_unsupported_protocol support yet!)
2153 CONNECT tunnel on http_port: same as https_port
2154 CONNECT tunnel on https_port: same as https_port
2156 Currently, this directive has effect on intercepted connections and
2157 bumped tunnels only. Other cases are not supported because Squid
2158 cannot know the intended destination of other traffic.
2161 # define what Squid errors indicate receiving non-HTTP traffic:
2162 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
2163 # define what Squid errors indicate receiving nothing:
2164 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
2165 # tunnel everything that does not look like HTTP:
2166 on_unsupported_protocol tunnel foreignProtocol
2167 # tunnel if we think the client waits for the server to talk first:
2168 on_unsupported_protocol tunnel serverTalksFirstProtocol
2169 # in all other error cases, just send an HTTP "error page" response:
2170 on_unsupported_protocol respond all
2172 See also: squid_error ACL
2178 LOC: Auth::TheConfig.schemeAccess
2180 DEFAULT_DOC: use all auth_param schemes in their configuration order
2182 Use this directive to customize authentication schemes presence and
2183 order in Squid's Unauthorized and Authentication Required responses.
2185 auth_schemes scheme1,scheme2,... [!]aclname ...
2187 where schemeN is the name of one of the authentication schemes
2188 configured using auth_param directives. At least one scheme name is
2189 required. Multiple scheme names are separated by commas. Either
2190 avoid whitespace or quote the entire schemes list.
2192 A special "ALL" scheme name expands to all auth_param-configured
2193 schemes in their configuration order. This directive cannot be used
2194 to configure Squid to offer no authentication schemes at all.
2196 The first matching auth_schemes rule determines the schemes order
2197 for the current Authentication Required transaction. Note that the
2198 future response is not yet available during auth_schemes evaluation.
2200 If this directive is not used or none of its rules match, then Squid
2201 responds with all configured authentication schemes in the order of
2202 auth_param directives in the configuration file.
2204 This directive does not determine when authentication is used or
2205 how each authentication scheme authenticates clients.
2207 The following example sends basic and negotiate authentication
2208 schemes, in that order, when requesting authentication of HTTP
2209 requests matching the isIE ACL (not shown) while sending all
2210 auth_param schemes in their configuration order to other clients:
2212 auth_schemes basic,negotiate isIE
2213 auth_schemes ALL all # explicit default
2215 This directive supports fast ACLs only.
2217 See also: auth_param.
2222 -----------------------------------------------------------------------------
2225 NAME: http_port ascii_port
2230 Usage: port [mode] [options]
2231 hostname:port [mode] [options]
2232 1.2.3.4:port [mode] [options]
2234 The socket addresses where Squid will listen for HTTP client
2235 requests. You may specify multiple socket addresses.
2236 There are three forms: port alone, hostname with port, and
2237 IP address with port. If you specify a hostname or IP
2238 address, Squid binds the socket to that specific
2239 address. Most likely, you do not need to bind to a specific
2240 address, so you can use the port number alone.
2242 If you are running Squid in accelerator mode, you
2243 probably want to listen on port 80 also, or instead.
2245 The -a command line option may be used to specify additional
2246 port(s) where Squid listens for proxy request. Such ports will
2247 be plain proxy ports with no options.
2249 You may specify multiple socket addresses on multiple lines.
2253 intercept Support for IP-Layer NAT interception delivering
2254 traffic to this Squid port.
2255 NP: disables authentication on the port.
2257 tproxy Support Linux TPROXY (or BSD divert-to) with spoofing
2258 of outgoing connections using the client IP address.
2259 NP: disables authentication on the port.
2261 accel Accelerator / reverse proxy mode
2263 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
2264 establish secure connection with the client and with
2265 the server, decrypt HTTPS messages as they pass through
2266 Squid, and treat them as unencrypted HTTP messages,
2267 becoming the man-in-the-middle.
2269 The ssl_bump option is required to fully enable
2270 bumping of CONNECT requests.
2272 Omitting the mode flag causes default forward proxy mode to be used.
2275 Accelerator Mode Options:
2277 defaultsite=domainname
2278 What to use for the Host: header if it is not present
2279 in a request. Determines what site (not origin server)
2280 accelerators should consider the default.
2282 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
2284 protocol= Protocol to reconstruct accelerated and intercepted
2285 requests with. Defaults to HTTP/1.1 for http_port and
2286 HTTPS/1.1 for https_port.
2287 When an unsupported value is configured Squid will
2288 produce a FATAL error.
2289 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
2291 vport Virtual host port support. Using the http_port number
2292 instead of the port passed on Host: headers.
2294 vport=NN Virtual host port support. Using the specified port
2295 number instead of the port passed on Host: headers.
2298 Act as if this Squid is the origin server.
2299 This currently means generate new Date: and Expires:
2300 headers on HIT instead of adding Age:.
2302 ignore-cc Ignore request Cache-Control headers.
2304 WARNING: This option violates HTTP specifications if
2305 used in non-accelerator setups.
2307 allow-direct Allow direct forwarding in accelerator mode. Normally
2308 accelerated requests are denied direct forwarding as if
2309 never_direct was used.
2311 WARNING: this option opens accelerator mode to security
2312 vulnerabilities usually only affecting in interception
2313 mode. Make sure to protect forwarding with suitable
2314 http_access rules when using this.
2317 SSL Bump Mode Options:
2318 In addition to these options ssl-bump requires TLS/SSL options.
2320 generate-host-certificates[=<on|off>]
2321 Dynamically create SSL server certificates for the
2322 destination hosts of bumped CONNECT requests.When
2323 enabled, the cert and key options are used to sign
2324 generated certificates. Otherwise generated
2325 certificate will be selfsigned.
2326 If there is a CA certificate lifetime of the generated
2327 certificate equals lifetime of the CA certificate. If
2328 generated certificate is selfsigned lifetime is three
2330 This option is enabled by default when ssl-bump is used.
2331 See the ssl-bump option above for more information.
2333 dynamic_cert_mem_cache_size=SIZE
2334 Approximate total RAM size spent on cached generated
2335 certificates. If set to zero, caching is disabled. The
2336 default value is 4MB.
2340 tls-cert= Path to file containing an X.509 certificate (PEM format)
2341 to be used in the TLS handshake ServerHello.
2343 If this certificate is constrained by KeyUsage TLS
2344 feature it must allow HTTP server usage, along with
2345 any additional restrictions imposed by your choice
2346 of options= settings.
2348 When OpenSSL is used this file may also contain a
2349 chain of intermediate CA certificates to send in the
2352 When GnuTLS is used this option (and any paired
2353 tls-key= option) may be repeated to load multiple
2354 certificates for different domains.
2356 Also, when generate-host-certificates=on is configured
2357 the first tls-cert= option must be a CA certificate
2358 capable of signing the automatically generated
2361 tls-key= Path to a file containing private key file (PEM format)
2362 for the previous tls-cert= option.
2364 If tls-key= is not specified tls-cert= is assumed to
2365 reference a PEM file containing both the certificate
2368 cipher= Colon separated list of supported ciphers.
2369 NOTE: some ciphers such as EDH ciphers depend on
2370 additional settings. If those settings are
2371 omitted the ciphers may be silently ignored
2372 by the OpenSSL library.
2374 options= Various SSL implementation options. The most important
2377 NO_SSLv3 Disallow the use of SSLv3
2379 NO_TLSv1 Disallow the use of TLSv1.0
2381 NO_TLSv1_1 Disallow the use of TLSv1.1
2383 NO_TLSv1_2 Disallow the use of TLSv1.2
2386 Always create a new key when using
2387 temporary/ephemeral DH key exchanges
2390 Enable ephemeral ECDH key exchange.
2391 The adopted curve should be specified
2392 using the tls-dh option.
2395 Disable use of RFC5077 session tickets.
2396 Some servers may have problems
2397 understanding the TLS extension due
2398 to ambiguous specification in RFC4507.
2400 ALL Enable various bug workarounds
2401 suggested as "harmless" by OpenSSL
2402 Be warned that this reduces SSL/TLS
2403 strength to some attacks.
2405 See the OpenSSL SSL_CTX_set_options documentation for a
2408 clientca= File containing the list of CAs to use when
2409 requesting a client certificate.
2411 tls-cafile= PEM file containing CA certificates to use when verifying
2412 client certificates. If not configured clientca will be
2413 used. May be repeated to load multiple files.
2415 capath= Directory containing additional CA certificates
2416 and CRL lists to use when verifying client certificates.
2417 Requires OpenSSL or LibreSSL.
2419 crlfile= File of additional CRL lists to use when verifying
2420 the client certificate, in addition to CRLs stored in
2421 the capath. Implies VERIFY_CRL flag below.
2424 File containing DH parameters for temporary/ephemeral DH key
2425 exchanges, optionally prefixed by a curve for ephemeral ECDH
2427 See OpenSSL documentation for details on how to create the
2428 DH parameter file. Supported curves for ECDH can be listed
2429 using the "openssl ecparam -list_curves" command.
2430 WARNING: EDH and EECDH ciphers will be silently disabled if
2431 this option is not set.
2433 sslflags= Various flags modifying the use of SSL:
2435 Don't request client certificates
2436 immediately, but wait until acl processing
2437 requires a certificate (not yet implemented).
2439 Request a client certificate during the TLS
2440 handshake, but ignore certificate absence in
2441 the TLS client Hello. If the client does
2442 supply a certificate, it is validated.
2444 Don't allow for session reuse. Each connection
2445 will result in a new SSL session.
2447 Verify CRL lists when accepting client
2450 Verify CRL lists for all certificates in the
2451 client certificate chain.
2453 tls-default-ca[=off]
2454 Whether to use the system Trusted CAs. Default is OFF.
2456 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
2458 sslcontext= SSL session ID context identifier.
2462 connection-auth[=on|off]
2463 use connection-auth=off to tell Squid to prevent
2464 forwarding Microsoft connection oriented authentication
2465 (NTLM, Negotiate and Kerberos)
2467 disable-pmtu-discovery=
2468 Control Path-MTU discovery usage:
2469 off lets OS decide on what to do (default).
2470 transparent disable PMTU discovery when transparent
2472 always disable always PMTU discovery.
2474 In many setups of transparently intercepting proxies
2475 Path-MTU discovery can not work on traffic towards the
2476 clients. This is the case when the intercepting device
2477 does not fully track connections and fails to forward
2478 ICMP must fragment messages to the cache server. If you
2479 have such setup and experience that certain clients
2480 sporadically hang or never complete requests set
2481 disable-pmtu-discovery option to 'transparent'.
2483 name= Specifies a internal name for the port. Defaults to
2484 the port specification (port or addr:port)
2486 tcpkeepalive[=idle,interval,timeout]
2487 Enable TCP keepalive probes of idle connections.
2488 In seconds; idle is the initial time before TCP starts
2489 probing the connection, interval how often to probe, and
2490 timeout the time before giving up.
2492 require-proxy-header
2493 Require PROXY protocol version 1 or 2 connections.
2494 The proxy_protocol_access is required to permit
2495 downstream proxies which can be trusted.
2498 Ask TCP stack to maintain a dedicated listening queue
2499 for each worker accepting requests at this port.
2500 Requires TCP stack that supports the SO_REUSEPORT socket
2503 SECURITY WARNING: Enabling worker-specific queues
2504 allows any process running as Squid's effective user to
2505 easily accept requests destined to this port.
2507 If you run Squid on a dual-homed machine with an internal
2508 and an external interface we recommend you to specify the
2509 internal address:port in http_port. This way Squid will only be
2510 visible on the internal address.
2514 # Squid normally listens to port 3128
2515 http_port @DEFAULT_HTTP_PORT@
2520 IFDEF: USE_GNUTLS||USE_OPENSSL
2525 Usage: [ip:]port [mode] tls-cert=certificate.pem [options]
2527 The socket address where Squid will listen for client requests made
2528 over TLS or SSL connections. Commonly referred to as HTTPS.
2530 This is most useful for situations where you are running squid in
2531 accelerator mode and you want to do the TLS work at the accelerator
2534 You may specify multiple socket addresses on multiple lines,
2535 each with their own certificate and/or options.
2537 The tls-cert= option is mandatory on HTTPS ports.
2539 See http_port for a list of modes and options.
2547 Enables Native FTP proxy by specifying the socket address where Squid
2548 listens for FTP client requests. See http_port directive for various
2549 ways to specify the listening address and mode.
2551 Usage: ftp_port address [mode] [options]
2553 WARNING: This is a new, experimental, complex feature that has seen
2554 limited production exposure. Some Squid modules (e.g., caching) do not
2555 currently work with native FTP proxying, and many features have not
2556 even been tested for compatibility. Test well before deploying!
2558 Native FTP proxying differs substantially from proxying HTTP requests
2559 with ftp:// URIs because Squid works as an FTP server and receives
2560 actual FTP commands (rather than HTTP requests with FTP URLs).
2562 Native FTP commands accepted at ftp_port are internally converted or
2563 wrapped into HTTP-like messages. The same happens to Native FTP
2564 responses received from FTP origin servers. Those HTTP-like messages
2565 are shoveled through regular access control and adaptation layers
2566 between the FTP client and the FTP origin server. This allows Squid to
2567 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2568 mechanisms when shoveling wrapped FTP messages. For example,
2569 http_access and adaptation_access directives are used.
2573 intercept Same as http_port intercept. The FTP origin address is
2574 determined based on the intended destination of the
2575 intercepted connection.
2577 tproxy Support Linux TPROXY for spoofing outgoing
2578 connections using the client IP address.
2579 NP: disables authentication and maybe IPv6 on the port.
2581 By default (i.e., without an explicit mode option), Squid extracts the
2582 FTP origin address from the login@origin parameter of the FTP USER
2583 command. Many popular FTP clients support such native FTP proxying.
2587 name=token Specifies an internal name for the port. Defaults to
2588 the port address. Usable with myportname ACL.
2591 Enables tracking of FTP directories by injecting extra
2592 PWD commands and adjusting Request-URI (in wrapping
2593 HTTP requests) to reflect the current FTP server
2594 directory. Tracking is disabled by default.
2596 protocol=FTP Protocol to reconstruct accelerated and intercepted
2597 requests with. Defaults to FTP. No other accepted
2598 values have been tested with. An unsupported value
2599 results in a FATAL error. Accepted values are FTP,
2600 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2602 Other http_port modes and options that are not specific to HTTP and
2603 HTTPS may also work.
2606 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2609 LOC: Ip::Qos::TheConfig.tosToServer
2611 Allows you to select a TOS/Diffserv value for packets outgoing
2612 on the server side, based on an ACL.
2614 tcp_outgoing_tos ds-field [!]aclname ...
2616 Example where normal_service_net uses the TOS value 0x00
2617 and good_service_net uses 0x20
2619 acl normal_service_net src 10.0.0.0/24
2620 acl good_service_net src 10.0.1.0/24
2621 tcp_outgoing_tos 0x00 normal_service_net
2622 tcp_outgoing_tos 0x20 good_service_net
2624 TOS/DSCP values really only have local significance - so you should
2625 know what you're specifying. For more information, see RFC2474,
2626 RFC2475, and RFC3260.
2628 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2629 "default" to use whatever default your host has.
2630 Note that only multiples of 4 are usable as the two rightmost bits have
2631 been redefined for use by ECN (RFC 3168 section 23.1).
2632 The squid parser will enforce this by masking away the ECN bits.
2634 Processing proceeds in the order specified, and stops at first fully
2637 Only fast ACLs are supported.
2640 NAME: clientside_tos
2643 LOC: Ip::Qos::TheConfig.tosToClient
2645 Allows you to select a TOS/DSCP value for packets being transmitted
2646 on the client-side, based on an ACL.
2648 clientside_tos ds-field [!]aclname ...
2650 Example where normal_service_net uses the TOS value 0x00
2651 and good_service_net uses 0x20
2653 acl normal_service_net src 10.0.0.0/24
2654 acl good_service_net src 10.0.1.0/24
2655 clientside_tos 0x00 normal_service_net
2656 clientside_tos 0x20 good_service_net
2658 Note: This feature is incompatible with qos_flows. Any TOS values set here
2659 will be overwritten by TOS values in qos_flows.
2661 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2662 "default" to use whatever default your host has.
2663 Note that only multiples of 4 are usable as the two rightmost bits have
2664 been redefined for use by ECN (RFC 3168 section 23.1).
2665 The squid parser will enforce this by masking away the ECN bits.
2667 This clause only supports fast acl types.
2668 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2671 NAME: tcp_outgoing_mark
2673 IFDEF: SO_MARK&&USE_LIBCAP
2675 LOC: Ip::Qos::TheConfig.nfmarkToServer
2677 Allows you to apply a Netfilter mark value to outgoing packets
2678 on the server side, based on an ACL.
2680 tcp_outgoing_mark mark-value [!]aclname ...
2682 Example where normal_service_net uses the mark value 0x00
2683 and good_service_net uses 0x20
2685 acl normal_service_net src 10.0.0.0/24
2686 acl good_service_net src 10.0.1.0/24
2687 tcp_outgoing_mark 0x00 normal_service_net
2688 tcp_outgoing_mark 0x20 good_service_net
2690 Only fast ACLs are supported.
2693 NAME: mark_client_packet clientside_mark
2695 IFDEF: SO_MARK&&USE_LIBCAP
2697 LOC: Ip::Qos::TheConfig.nfmarkToClient
2699 Allows you to apply a Netfilter MARK value to packets being transmitted
2700 on the client-side, based on an ACL.
2702 mark_client_packet mark-value [!]aclname ...
2704 Example where normal_service_net uses the MARK value 0x00
2705 and good_service_net uses 0x20
2707 acl normal_service_net src 10.0.0.0/24
2708 acl good_service_net src 10.0.1.0/24
2709 mark_client_packet 0x00 normal_service_net
2710 mark_client_packet 0x20 good_service_net
2712 Note: This feature is incompatible with qos_flows. Any mark values set here
2713 will be overwritten by mark values in qos_flows.
2715 This clause only supports fast acl types.
2716 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2719 NAME: mark_client_connection
2721 IFDEF: SO_MARK&&USE_LIBCAP
2723 LOC: Ip::Qos::TheConfig.nfConnmarkToClient
2725 Allows you to apply a Netfilter CONNMARK value to a connection
2726 on the client-side, based on an ACL.
2728 mark_client_connection mark-value[/mask] [!]aclname ...
2730 The mark-value and mask are unsigned integers (hex, octal, or decimal).
2731 The mask may be used to preserve marking previously set by other agents
2734 A matching rule replaces the CONNMARK value. If a mask is also
2735 specified, then the masked bits of the original value are zeroed, and
2736 the configured mark-value is ORed with that adjusted value.
2737 For example, applying a mark-value 0xAB/0xF to 0x5F CONNMARK, results
2738 in a 0xFB marking (rather than a 0xAB or 0x5B).
2740 This directive semantics is similar to iptables --set-mark rather than
2741 --set-xmark functionality.
2743 The directive does not interfere with qos_flows (which uses packet MARKs,
2746 Example where squid marks intercepted FTP connections:
2748 acl proto_ftp proto FTP
2749 mark_client_connection 0x200/0xff00 proto_ftp
2751 This clause only supports fast acl types.
2752 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2759 LOC: Ip::Qos::TheConfig
2761 Allows you to select a TOS/DSCP value to mark outgoing
2762 connections to the client, based on where the reply was sourced.
2763 For platforms using netfilter, allows you to set a netfilter mark
2764 value instead of, or in addition to, a TOS value.
2766 By default this functionality is disabled. To enable it with the default
2767 settings simply use "qos_flows mark" or "qos_flows tos". Default
2768 settings will result in the netfilter mark or TOS value being copied
2769 from the upstream connection to the client. Note that it is the connection
2770 CONNMARK value not the packet MARK value that is copied.
2772 It is not currently possible to copy the mark or TOS value from the
2773 client to the upstream connection request.
2775 TOS values really only have local significance - so you should
2776 know what you're specifying. For more information, see RFC2474,
2777 RFC2475, and RFC3260.
2779 The TOS/DSCP byte must be exactly that - a octet value 0 - 255.
2780 Note that only multiples of 4 are usable as the two rightmost bits have
2781 been redefined for use by ECN (RFC 3168 section 23.1).
2782 The squid parser will enforce this by masking away the ECN bits.
2784 Mark values can be any unsigned 32-bit integer value.
2786 This setting is configured by setting the following values:
2788 tos|mark Whether to set TOS or netfilter mark values
2790 local-hit=0xFF Value to mark local cache hits.
2792 sibling-hit=0xFF Value to mark hits from sibling peers.
2794 parent-hit=0xFF Value to mark hits from parent peers.
2796 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2797 over the preserve-miss feature (see below), unless
2798 mask is specified, in which case only the bits
2799 specified in the mask are written.
2801 The TOS variant of the following features are only possible on Linux
2802 and require your kernel to be patched with the TOS preserving ZPH
2803 patch, available from http://zph.bratcheda.org
2804 No patch is needed to preserve the netfilter mark, which will work
2805 with all variants of netfilter.
2807 disable-preserve-miss
2808 This option disables the preservation of the TOS or netfilter
2809 mark. By default, the existing TOS or netfilter mark value of
2810 the response coming from the remote server will be retained
2811 and masked with miss-mark.
2812 NOTE: in the case of a netfilter mark, the mark must be set on
2813 the connection (using the CONNMARK target) not on the packet
2817 Allows you to mask certain bits in the TOS or mark value
2818 received from the remote server, before copying the value to
2819 the TOS sent towards clients.
2820 Default for tos: 0xFF (TOS from server is not changed).
2821 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2823 All of these features require the --enable-zph-qos compilation flag
2824 (enabled by default). Netfilter marking also requires the
2825 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2826 libcap 2.09+ (--with-libcap).
2830 NAME: tcp_outgoing_address
2833 DEFAULT_DOC: Address selection is performed by the operating system.
2834 LOC: Config.accessList.outgoing_address
2836 Allows you to map requests to different outgoing IP addresses
2837 based on the username or source address of the user making
2840 tcp_outgoing_address ipaddr [[!]aclname] ...
2843 Forwarding clients with dedicated IPs for certain subnets.
2845 acl normal_service_net src 10.0.0.0/24
2846 acl good_service_net src 10.0.2.0/24
2848 tcp_outgoing_address 2001:db8::c001 good_service_net
2849 tcp_outgoing_address 10.1.0.2 good_service_net
2851 tcp_outgoing_address 2001:db8::beef normal_service_net
2852 tcp_outgoing_address 10.1.0.1 normal_service_net
2854 tcp_outgoing_address 2001:db8::1
2855 tcp_outgoing_address 10.1.0.3
2857 Processing proceeds in the order specified, and stops at first fully
2860 Squid will add an implicit IP version test to each line.
2861 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2862 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2865 NOTE: The use of this directive using client dependent ACLs is
2866 incompatible with the use of server side persistent connections. To
2867 ensure correct results it is best to set server_persistent_connections
2868 to off when using this directive in such configurations.
2870 NOTE: The use of this directive to set a local IP on outgoing TCP links
2871 is incompatible with using TPROXY to set client IP out outbound TCP links.
2872 When needing to contact peers use the no-tproxy cache_peer option and the
2873 client_dst_passthru directive re-enable normal forwarding such as this.
2875 This clause only supports fast acl types.
2876 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2879 NAME: host_verify_strict
2882 LOC: Config.onoff.hostStrictVerify
2884 Regardless of this option setting, when dealing with intercepted
2885 traffic, Squid always verifies that the destination IP address matches
2886 the Host header domain or IP (called 'authority form URL').
2888 This enforcement is performed to satisfy a MUST-level requirement in
2889 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2890 authority of the origin server or gateway given by the original URL".
2893 Squid always responds with an HTTP 409 (Conflict) error
2894 page and logs a security warning if there is no match.
2896 Squid verifies that the destination IP address matches
2897 the Host header for forward-proxy and reverse-proxy traffic
2898 as well. For those traffic types, Squid also enables the
2899 following checks, comparing the corresponding Host header
2900 and Request-URI components:
2902 * The host names (domain or IP) must be identical,
2903 but valueless or missing Host header disables all checks.
2904 For the two host names to match, both must be either IP
2907 * Port numbers must be identical, but if a port is missing
2908 the scheme-default port is assumed.
2911 When set to OFF (the default):
2912 Squid allows suspicious requests to continue but logs a
2913 security warning and blocks caching of the response.
2915 * Forward-proxy traffic is not checked at all.
2917 * Reverse-proxy traffic is not checked at all.
2919 * Intercepted traffic which passes verification is handled
2920 according to client_dst_passthru.
2922 * Intercepted requests which fail verification are sent
2923 to the client original destination instead of DIRECT.
2924 This overrides 'client_dst_passthru off'.
2926 For now suspicious intercepted CONNECT requests are always
2927 responded to with an HTTP 409 (Conflict) error page.
2932 As described in CVE-2009-0801 when the Host: header alone is used
2933 to determine the destination of a request it becomes trivial for
2934 malicious scripts on remote websites to bypass browser same-origin
2935 security policy and sandboxing protections.
2937 The cause of this is that such applets are allowed to perform their
2938 own HTTP stack, in which case the same-origin policy of the browser
2939 sandbox only verifies that the applet tries to contact the same IP
2940 as from where it was loaded at the IP level. The Host: header may
2941 be different from the connected IP and approved origin.
2945 NAME: client_dst_passthru
2948 LOC: Config.onoff.client_dst_passthru
2950 With NAT or TPROXY intercepted traffic Squid may pass the request
2951 directly to the original client destination IP or seek a faster
2952 source using the HTTP Host header.
2954 Using Host to locate alternative servers can provide faster
2955 connectivity with a range of failure recovery options.
2956 But can also lead to connectivity trouble when the client and
2957 server are attempting stateful interactions unaware of the proxy.
2959 This option (on by default) prevents alternative DNS entries being
2960 located to send intercepted traffic DIRECT to an origin server.
2961 The clients original destination IP and port will be used instead.
2963 Regardless of this option setting, when dealing with intercepted
2964 traffic Squid will verify the Host: header and any traffic which
2965 fails Host verification will be treated as if this option were ON.
2967 see host_verify_strict for details on the verification process.
2972 -----------------------------------------------------------------------------
2975 NAME: tls_outgoing_options
2976 IFDEF: USE_GNUTLS||USE_OPENSSL
2977 TYPE: securePeerOptions
2978 DEFAULT: min-version=1.0
2979 LOC: Security::ProxyOutgoingConfig
2981 disable Do not support https:// URLs.
2983 cert=/path/to/client/certificate
2984 A client X.509 certificate to use when connecting.
2986 key=/path/to/client/private_key
2987 The private key corresponding to the cert= above.
2989 If key= is not specified cert= is assumed to
2990 reference a PEM file containing both the certificate
2993 cipher=... The list of valid TLS ciphers to use.
2996 The minimum TLS protocol version to permit.
2997 To control SSLv3 use the options= parameter.
2998 Supported Values: 1.0 (default), 1.1, 1.2, 1.3
3000 options=... Specify various TLS/SSL implementation options.
3002 OpenSSL options most important are:
3004 NO_SSLv3 Disallow the use of SSLv3
3007 Always create a new key when using
3008 temporary/ephemeral DH key exchanges
3011 Disable use of RFC5077 session tickets.
3012 Some servers may have problems
3013 understanding the TLS extension due
3014 to ambiguous specification in RFC4507.
3016 ALL Enable various bug workarounds
3017 suggested as "harmless" by OpenSSL
3018 Be warned that this reduces SSL/TLS
3019 strength to some attacks.
3021 See the OpenSSL SSL_CTX_set_options documentation
3022 for a more complete list.
3024 GnuTLS options most important are:
3027 Disable use of RFC5077 session tickets.
3028 Some servers may have problems
3029 understanding the TLS extension due
3030 to ambiguous specification in RFC4507.
3032 See the GnuTLS Priority Strings documentation
3033 for a more complete list.
3034 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
3037 cafile= PEM file containing CA certificates to use when verifying
3038 the peer certificate. May be repeated to load multiple files.
3040 capath= A directory containing additional CA certificates to
3041 use when verifying the peer certificate.
3042 Requires OpenSSL or LibreSSL.
3044 crlfile=... A certificate revocation list file to use when
3045 verifying the peer certificate.
3047 flags=... Specify various flags modifying the TLS implementation:
3050 Accept certificates even if they fail to
3053 Don't verify the peer certificate
3054 matches the server name
3057 Whether to use the system Trusted CAs. Default is ON.
3059 domain= The peer name as advertised in its certificate.
3060 Used for verifying the correctness of the received peer
3061 certificate. If not specified the peer hostname will be
3067 -----------------------------------------------------------------------------
3070 NAME: ssl_unclean_shutdown
3074 LOC: Config.SSL.unclean_shutdown
3076 Some browsers (especially MSIE) bugs out on SSL shutdown
3083 LOC: Config.SSL.ssl_engine
3086 The OpenSSL engine to use. You will need to set this if you
3087 would like to use hardware SSL acceleration for example.
3089 Not supported in builds with OpenSSL v3 or newer.
3092 NAME: sslproxy_session_ttl
3095 LOC: Config.SSL.session_ttl
3098 Sets the timeout value for SSL sessions
3101 NAME: sslproxy_session_cache_size
3104 LOC: Config.SSL.sessionCacheSize
3107 Sets the cache size to use for ssl session
3110 NAME: sslproxy_foreign_intermediate_certs
3113 LOC: Config.ssl_client.foreignIntermediateCertsPath
3116 Many origin servers fail to send their full server certificate
3117 chain for verification, assuming the client already has or can
3118 easily locate any missing intermediate certificates.
3120 Squid uses the certificates from the specified file to fill in
3121 these missing chains when trying to validate origin server
3124 The file is expected to contain zero or more PEM-encoded
3125 intermediate certificates. These certificates are not treated
3126 as trusted root certificates, and any self-signed certificate in
3127 this file will be ignored.
3130 NAME: sslproxy_cert_sign_hash
3133 LOC: Config.SSL.certSignHash
3136 Sets the hashing algorithm to use when signing generated certificates.
3137 Valid algorithm names depend on the OpenSSL library used. The following
3138 names are usually available: sha1, sha256, sha512, and md5. Please see
3139 your OpenSSL library manual for the available hashes. By default, Squids
3140 that support this option use sha256 hashes.
3142 Squid does not forcefully purge cached certificates that were generated
3143 with an algorithm other than the currently configured one. They remain
3144 in the cache, subject to the regular cache eviction policy, and become
3145 useful if the algorithm changes again.
3150 TYPE: sslproxy_ssl_bump
3151 LOC: Config.accessList.ssl_bump
3152 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
3155 This option is consulted when a CONNECT request is received on
3156 an http_port (or a new connection is intercepted at an
3157 https_port), provided that port was configured with an ssl-bump
3158 flag. The subsequent data on the connection is either treated as
3159 HTTPS and decrypted OR tunneled at TCP level without decryption,
3160 depending on the first matching bumping "action".
3162 ssl_bump <action> [!]acl ...
3164 The following bumping actions are currently supported:
3167 Become a TCP tunnel without decrypting proxied traffic.
3168 This is the default action.
3171 When used on step SslBump1, establishes a secure connection
3172 with the client first, then connect to the server.
3173 When used on step SslBump2 or SslBump3, establishes a secure
3174 connection with the server and, using a mimicked server
3175 certificate, with the client.
3178 Receive client (step SslBump1) or server (step SslBump2)
3179 certificate while preserving the possibility of splicing the
3180 connection. Peeking at the server certificate (during step 2)
3181 usually precludes bumping of the connection at step 3.
3184 Receive client (step SslBump1) or server (step SslBump2)
3185 certificate while preserving the possibility of bumping the
3186 connection. Staring at the server certificate (during step 2)
3187 usually precludes splicing of the connection at step 3.
3190 Close client and server connections.
3192 Backward compatibility actions available at step SslBump1:
3195 Bump the connection. Establish a secure connection with the
3196 client first, then connect to the server. This old mode does
3197 not allow Squid to mimic server SSL certificate and does not
3198 work with intercepted SSL connections.
3201 Bump the connection. Establish a secure connection with the
3202 server first, then establish a secure connection with the
3203 client, using a mimicked server certificate. Works with both
3204 CONNECT requests and intercepted SSL connections, but does
3205 not allow to make decisions based on SSL handshake info.
3208 Decide whether to bump or splice the connection based on
3209 client-to-squid and server-to-squid SSL hello messages.
3213 Same as the "splice" action.
3215 All ssl_bump rules are evaluated at each of the supported bumping
3216 steps. Rules with actions that are impossible at the current step are
3217 ignored. The first matching ssl_bump action wins and is applied at the
3218 end of the current step. If no rules match, the splice action is used.
3219 See the at_step ACL for a list of the supported SslBump steps.
3221 This clause supports both fast and slow acl types.
3222 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3224 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
3227 # Example: Bump all TLS connections except those originating from
3228 # localhost or those going to example.com.
3230 acl broken_sites ssl::server_name .example.com
3231 ssl_bump splice localhost
3232 ssl_bump splice broken_sites
3236 NAME: sslproxy_cert_error
3239 DEFAULT_DOC: Server certificate errors terminate the transaction.
3240 LOC: Config.ssl_client.cert_error
3243 Use this ACL to bypass server certificate validation errors.
3245 For example, the following lines will bypass all validation errors
3246 when talking to servers for example.com. All other
3247 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
3249 acl BrokenButTrustedServers dstdomain example.com
3250 sslproxy_cert_error allow BrokenButTrustedServers
3251 sslproxy_cert_error deny all
3253 This clause only supports fast acl types.
3254 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3255 Using slow acl types may result in server crashes
3257 Without this option, all server certificate validation errors
3258 terminate the transaction to protect Squid and the client.
3260 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
3261 but should not happen unless your OpenSSL library is buggy.
3264 Bypassing validation errors is dangerous because an
3265 error usually implies that the server cannot be trusted
3266 and the connection may be insecure.
3268 See also: sslproxy_flags and DONT_VERIFY_PEER.
3271 NAME: sslproxy_cert_sign
3274 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
3275 POSTSCRIPTUM: signSelf ssl::certSelfSigned
3276 POSTSCRIPTUM: signTrusted all
3277 TYPE: sslproxy_cert_sign
3278 LOC: Config.ssl_client.cert_sign
3281 sslproxy_cert_sign <signing algorithm> acl ...
3283 The following certificate signing algorithms are supported:
3286 Sign using the configured CA certificate which is usually
3287 placed in and trusted by end-user browsers. This is the
3288 default for trusted origin server certificates.
3291 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
3292 This is the default for untrusted origin server certificates
3293 that are not self-signed (see ssl::certUntrusted).
3296 Sign using a self-signed certificate with the right CN to
3297 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
3298 browser. This is the default for self-signed origin server
3299 certificates (see ssl::certSelfSigned).
3301 This clause only supports fast acl types.
3303 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
3304 signing algorithm to generate the certificate and ignores all
3305 subsequent sslproxy_cert_sign options (the first match wins). If no
3306 acl(s) match, the default signing algorithm is determined by errors
3307 detected when obtaining and validating the origin server certificate.
3309 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3310 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3311 CONNECT request that carries a domain name. In all other cases (CONNECT
3312 to an IP address or an intercepted SSL connection), Squid cannot detect
3313 the domain mismatch at certificate generation time when
3314 bump-server-first is used.
3317 NAME: sslproxy_cert_adapt
3320 TYPE: sslproxy_cert_adapt
3321 LOC: Config.ssl_client.cert_adapt
3324 sslproxy_cert_adapt <adaptation algorithm> acl ...
3326 The following certificate adaptation algorithms are supported:
3329 Sets the "Not After" property to the "Not After" property of
3330 the CA certificate used to sign generated certificates.
3333 Sets the "Not Before" property to the "Not Before" property of
3334 the CA certificate used to sign generated certificates.
3336 setCommonName or setCommonName{CN}
3337 Sets Subject.CN property to the host name specified as a
3338 CN parameter or, if no explicit CN parameter was specified,
3339 extracted from the CONNECT request. It is a misconfiguration
3340 to use setCommonName without an explicit parameter for
3341 intercepted or tproxied SSL connections.
3343 This clause only supports fast acl types.
3345 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
3346 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
3347 corresponding adaptation algorithm to generate the certificate and
3348 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
3349 group (i.e., the first match wins within each algorithm group). If no
3350 acl(s) match, the default mimicking action takes place.
3352 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3353 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3354 CONNECT request that carries a domain name. In all other cases (CONNECT
3355 to an IP address or an intercepted SSL connection), Squid cannot detect
3356 the domain mismatch at certificate generation time when
3357 bump-server-first is used.
3360 NAME: sslpassword_program
3363 LOC: Config.Program.ssl_password
3366 Specify a program used for entering SSL key passphrases
3367 when using encrypted SSL certificate keys. If not specified
3368 keys must either be unencrypted, or Squid started with the -N
3369 option to allow it to query interactively for the passphrase.
3371 The key file name is given as argument to the program allowing
3372 selection of the right password if you have multiple encrypted
3377 OPTIONS RELATING TO EXTERNAL SSL_CRTD
3378 -----------------------------------------------------------------------------
3381 NAME: sslcrtd_program
3384 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
3385 LOC: Ssl::TheConfig.ssl_crtd
3387 Specify the location and options of the executable for certificate
3390 @DEFAULT_SSL_CRTD@ program can use a disk cache to improve response
3391 times on repeated requests. To enable caching, specify -s and -M
3392 parameters. If those parameters are not given, the program generates
3393 a new certificate on every request.
3395 For more information use:
3396 @DEFAULT_SSL_CRTD@ -h
3399 NAME: sslcrtd_children
3400 TYPE: HelperChildConfig
3402 DEFAULT: 32 startup=5 idle=1
3403 LOC: Ssl::TheConfig.ssl_crtdChildren
3405 Specifies the maximum number of certificate generation processes that
3406 Squid may spawn (numberofchildren) and several related options. Using
3407 too few of these helper processes (a.k.a. "helpers") creates request
3408 queues. Using too many helpers wastes your system resources. Squid
3409 does not support spawning more than 32 helpers.
3411 Usage: numberofchildren [option]...
3413 The startup= and idle= options allow some measure of skew in your
3418 Sets the minimum number of processes to spawn when Squid
3419 starts or reconfigures. When set to zero the first request will
3420 cause spawning of the first child process to handle it.
3422 Starting too few children temporary slows Squid under load while it
3423 tries to spawn enough additional processes to cope with traffic.
3427 Sets a minimum of how many processes Squid is to try and keep available
3428 at all times. When traffic begins to rise above what the existing
3429 processes can handle this many more will be spawned up to the maximum
3430 configured. A minimum setting of 1 is required.
3434 Sets the maximum number of queued requests. A request is queued when
3435 no existing child is idle and no new child can be started due to
3436 numberofchildren limit. If the queued requests exceed queue size for
3437 more than 3 minutes squid aborts its operation. The default value is
3438 set to 2*numberofchildren.
3440 You must have at least one ssl_crtd process.
3443 NAME: sslcrtvalidator_program
3447 LOC: Ssl::TheConfig.ssl_crt_validator
3449 Specify the location and options of the executable for ssl_crt_validator
3452 Usage: sslcrtvalidator_program [ttl=...] [cache=n] path ...
3457 Limits how much memory Squid can use for caching validator
3458 responses. The default is 67108864 (i.e. 64 MB).
3459 Reconfiguration purges any excess entries. To disable caching,
3460 use cache=0. Currently, cache entry sizes are seriously
3461 underestimated. Even with that bug, a typical estimate for a
3462 single cache entry size would be at least a few kilobytes (the
3463 size of the PEM certificates sent to the validator).
3465 ttl=<seconds|"infinity">
3466 Approximately how long Squid may reuse the validator results
3467 for. The default is 3600 (i.e. 1 hour). Using ttl=infinity
3468 disables TTL checks. Reconfiguration does not affect TTLs of
3469 the already cached entries. To disable caching, use zero cache
3470 size, not zero TTL -- zero TTL allows reuse for the remainder
3471 of the second when the result was cached.
3474 NAME: sslcrtvalidator_children
3475 TYPE: HelperChildConfig
3477 DEFAULT: 32 startup=5 idle=1 concurrency=1
3478 LOC: Ssl::TheConfig.ssl_crt_validator_Children
3480 Specifies the maximum number of certificate validation processes that
3481 Squid may spawn (numberofchildren) and several related options. Using
3482 too few of these helper processes (a.k.a. "helpers") creates request
3483 queues. Using too many helpers wastes your system resources. Squid
3484 does not support spawning more than 32 helpers.
3486 Usage: numberofchildren [option]...
3488 The startup= and idle= options allow some measure of skew in your
3493 Sets the minimum number of processes to spawn when Squid
3494 starts or reconfigures. When set to zero the first request will
3495 cause spawning of the first child process to handle it.
3497 Starting too few children temporary slows Squid under load while it
3498 tries to spawn enough additional processes to cope with traffic.
3502 Sets a minimum of how many processes Squid is to try and keep available
3503 at all times. When traffic begins to rise above what the existing
3504 processes can handle this many more will be spawned up to the maximum
3505 configured. A minimum setting of 1 is required.
3509 The number of requests each certificate validator helper can handle in
3510 parallel. A value of 0 indicates the certificate validator does not
3511 support concurrency. Defaults to 1.
3513 When this directive is set to a value >= 1 then the protocol
3514 used to communicate with the helper is modified to include
3515 a request ID in front of the request/response. The request
3516 ID from the request must be echoed back with the response
3521 Sets the maximum number of queued requests. A request is queued when
3522 no existing child can accept it due to concurrency limit and no new
3523 child can be started due to numberofchildren limit. If the queued
3524 requests exceed queue size for more than 3 minutes squid aborts its
3525 operation. The default value is set to 2*numberofchildren.
3527 You must have at least one ssl_crt_validator process.
3531 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
3532 -----------------------------------------------------------------------------
3540 To specify other caches in a hierarchy, use the format:
3542 cache_peer hostname type http-port icp-port [options]
3547 # hostname type port port options
3548 # -------------------- -------- ----- ----- -----------
3549 cache_peer parent.foo.net parent 3128 3130 default
3550 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
3551 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
3552 cache_peer example.com parent 80 0 default
3553 cache_peer cdn.example.com sibling 3128 0
3555 type: either 'parent', 'sibling', or 'multicast'.
3557 proxy-port: The port number where the peer accept HTTP requests.
3558 For other Squid proxies this is usually 3128
3559 For web servers this is usually 80
3561 icp-port: Used for querying neighbor caches about objects.
3562 Set to 0 if the peer does not support ICP or HTCP.
3563 See ICP and HTCP options below for additional details.
3566 ==== ICP OPTIONS ====
3568 You MUST also set icp_port and icp_access explicitly when using these options.
3569 The defaults will prevent peer traffic using ICP.
3572 no-query Disable ICP queries to this neighbor.
3575 Indicates the named peer is a member of a multicast group.
3576 ICP queries will not be sent directly to the peer, but ICP
3577 replies will be accepted from it.
3579 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
3580 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
3583 To only send ICP queries to this neighbor infrequently.
3584 This is used to keep the neighbor round trip time updated
3585 and is usually used in conjunction with weighted-round-robin.
3588 ==== HTCP OPTIONS ====
3590 You MUST also set htcp_port and htcp_access explicitly when using these options.
3591 The defaults will prevent peer traffic using HTCP.
3594 htcp Send HTCP, instead of ICP, queries to the neighbor.
3595 You probably also want to set the "icp-port" to 4827
3596 instead of 3130. This directive accepts a comma separated
3597 list of options described below.
3599 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3601 htcp=no-clr Send HTCP to the neighbor but without
3602 sending any CLR requests. This cannot be used with
3605 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3606 This cannot be used with no-clr.
3609 Send HTCP to the neighbor including CLRs but only when
3610 they do not result from PURGE requests.
3613 Forward any HTCP CLR requests this proxy receives to the peer.
3616 ==== PEER SELECTION METHODS ====
3618 The default peer selection method is ICP, with the first responding peer
3619 being used as source. These options can be used for better load balancing.
3622 default This is a parent cache which can be used as a "last-resort"
3623 if a peer cannot be located by any of the peer-selection methods.
3624 If specified more than once, only the first is used.
3626 round-robin Load-Balance parents which should be used in a round-robin
3627 fashion in the absence of any ICP queries.
3628 weight=N can be used to add bias.
3630 weighted-round-robin
3631 Load-Balance parents which should be used in a round-robin
3632 fashion with the frequency of each parent being based on the
3633 round trip time. Closer parents are used more often.
3634 Usually used for background-ping parents.
3635 weight=N can be used to add bias.
3637 carp Load-Balance parents which should be used as a CARP array.
3638 The requests will be distributed among the parents based on the
3639 CARP load balancing hash function based on their weight.
3641 userhash Load-balance parents based on the client proxy_auth or ident username.
3643 sourcehash Load-balance parents based on the client source IP.
3646 To be used only for cache peers of type "multicast".
3647 ALL members of this multicast group have "sibling"
3648 relationship with it, not "parent". This is to a multicast
3649 group when the requested object would be fetched only from
3650 a "parent" cache, anyway. It's useful, e.g., when
3651 configuring a pool of redundant Squid proxies, being
3652 members of the same multicast group.
3655 ==== PEER SELECTION OPTIONS ====
3657 weight=N use to affect the selection of a peer during any weighted
3658 peer-selection mechanisms.
3659 The weight must be an integer; default is 1,
3660 larger weights are favored more.
3661 This option does not affect parent selection if a peering
3662 protocol is not in use.
3664 basetime=N Specify a base amount to be subtracted from round trip
3666 It is subtracted before division by weight in calculating
3667 which parent to fectch from. If the rtt is less than the
3668 base time the rtt is set to a minimal value.
3670 ttl=N Specify a TTL to use when sending multicast ICP queries
3672 Only useful when sending to a multicast group.
3673 Because we don't accept ICP replies from random
3674 hosts, you must configure other group members as
3675 peers with the 'multicast-responder' option.
3677 no-delay To prevent access to this neighbor from influencing the
3680 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3681 enabled) for this host from the specified URL rather
3682 than the Squid default location.
3685 ==== CARP OPTIONS ====
3687 carp-key=key-specification
3688 use a different key than the full URL to hash against the peer.
3689 the key-specification is a comma-separated list of the keywords
3690 scheme, host, port, path, params
3691 Order is not important.
3693 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3695 originserver Causes this parent to be contacted as an origin server.
3696 Meant to be used in accelerator setups when the peer
3700 Set the Host header of requests forwarded to this peer.
3701 Useful in accelerator setups where the server (peer)
3702 expects a certain domain name but clients may request
3703 others. ie example.com or www.example.com
3705 no-digest Disable request of cache digests.
3708 Disables requesting ICMP RTT database (NetDB).
3711 ==== AUTHENTICATION OPTIONS ====
3714 If this is a personal/workgroup proxy and your parent
3715 requires proxy authentication.
3717 Note: The string can include URL escapes (i.e. %20 for
3718 spaces). This also means % must be written as %%.
3721 Send login details received from client to this peer.
3722 Both Proxy- and WWW-Authorization headers are passed
3723 without alteration to the peer.
3724 Authentication is not required by Squid for this to work.
3726 Note: This will pass any form of authentication but
3727 only Basic auth will work through a proxy unless the
3728 connection-auth options are also used.
3730 login=PASS Send login details received from client to this peer.
3731 Authentication is not required by this option.
3733 If there are no client-provided authentication headers
3734 to pass on, but username and password are available
3735 from an external ACL user= and password= result tags
3736 they may be sent instead.
3738 Note: To combine this with proxy_auth both proxies must
3739 share the same user database as HTTP only allows for
3740 a single login (one for proxy, one for origin server).
3741 Also be warned this will expose your users proxy
3742 password to the peer. USE WITH CAUTION
3745 Send the username to the upstream cache, but with a
3746 fixed password. This is meant to be used when the peer
3747 is in another administrative domain, but it is still
3748 needed to identify each user.
3749 The star can optionally be followed by some extra
3750 information which is added to the username. This can
3751 be used to identify this proxy to the peer, similar to
3752 the login=username:password option above.
3755 If this is a personal/workgroup proxy and your parent
3756 requires a secure proxy authentication.
3757 The first principal from the default keytab or defined by
3758 the environment variable KRB5_KTNAME will be used.
3760 WARNING: The connection may transmit requests from multiple
3761 clients. Negotiate often assumes end-to-end authentication
3762 and a single-client. Which is not strictly true here.
3764 login=NEGOTIATE:principal_name
3765 If this is a personal/workgroup proxy and your parent
3766 requires a secure proxy authentication.
3767 The principal principal_name from the default keytab or
3768 defined by the environment variable KRB5_KTNAME will be
3771 WARNING: The connection may transmit requests from multiple
3772 clients. Negotiate often assumes end-to-end authentication
3773 and a single-client. Which is not strictly true here.
3775 connection-auth=on|off
3776 Tell Squid that this peer does or not support Microsoft
3777 connection oriented authentication, and any such
3778 challenges received from there should be ignored.
3779 Default is auto to automatically determine the status
3783 Do not use a keytab to authenticate to a peer when
3784 login=NEGOTIATE is specified. Let the GSSAPI
3785 implementation determine which already existing
3786 credentials cache to use instead.
3789 ==== SSL / HTTPS / TLS OPTIONS ====
3791 tls Encrypt connections to this peer with TLS.
3793 sslcert=/path/to/ssl/certificate
3794 A client X.509 certificate to use when connecting to
3797 sslkey=/path/to/ssl/key
3798 The private key corresponding to sslcert above.
3800 If sslkey= is not specified sslcert= is assumed to
3801 reference a PEM file containing both the certificate
3804 sslcipher=... The list of valid SSL ciphers to use when connecting
3808 The minimum TLS protocol version to permit. To control
3809 SSLv3 use the tls-options= parameter.
3810 Supported Values: 1.0 (default), 1.1, 1.2
3812 tls-options=... Specify various TLS implementation options.
3814 OpenSSL options most important are:
3816 NO_SSLv3 Disallow the use of SSLv3
3819 Always create a new key when using
3820 temporary/ephemeral DH key exchanges
3823 Disable use of RFC5077 session tickets.
3824 Some servers may have problems
3825 understanding the TLS extension due
3826 to ambiguous specification in RFC4507.
3828 ALL Enable various bug workarounds
3829 suggested as "harmless" by OpenSSL
3830 Be warned that this reduces SSL/TLS
3831 strength to some attacks.
3833 See the OpenSSL SSL_CTX_set_options documentation for a
3836 GnuTLS options most important are:
3839 Disable use of RFC5077 session tickets.
3840 Some servers may have problems
3841 understanding the TLS extension due
3842 to ambiguous specification in RFC4507.
3844 See the GnuTLS Priority Strings documentation
3845 for a more complete list.
3846 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
3848 tls-cafile= PEM file containing CA certificates to use when verifying
3849 the peer certificate. May be repeated to load multiple files.
3851 sslcapath=... A directory containing additional CA certificates to
3852 use when verifying the peer certificate.
3853 Requires OpenSSL or LibreSSL.
3855 sslcrlfile=... A certificate revocation list file to use when
3856 verifying the peer certificate.
3858 sslflags=... Specify various flags modifying the SSL implementation:
3861 Accept certificates even if they fail to
3865 Don't verify the peer certificate
3866 matches the server name
3868 ssldomain= The peer name as advertised in it's certificate.
3869 Used for verifying the correctness of the received peer
3870 certificate. If not specified the peer hostname will be
3873 front-end-https[=off|on|auto]
3874 Enable the "Front-End-Https: On" header needed when
3875 using Squid as a SSL frontend in front of Microsoft OWA.
3876 See MS KB document Q307347 for details on this header.
3877 If set to auto the header will only be added if the
3878 request is forwarded as a https:// URL.
3880 tls-default-ca[=off]
3881 Whether to use the system Trusted CAs. Default is ON.
3883 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
3885 ==== GENERAL OPTIONS ====
3888 A peer-specific connect timeout.
3889 Also see the peer_connect_timeout directive.
3891 connect-fail-limit=N
3892 How many times connecting to a peer must fail before
3893 it is marked as down. Standby connection failures
3894 count towards this limit. Default is 10.
3896 allow-miss Disable Squid's use of only-if-cached when forwarding
3897 requests to siblings. This is primarily useful when
3898 icp_hit_stale is used by the sibling. Excessive use
3899 of this option may result in forwarding loops. One way
3900 to prevent peering loops when using this option, is to
3901 deny cache peer usage on requests from a peer:
3903 cache_peer_access peerName deny fromPeer
3905 max-conn=N Limit the number of concurrent connections the Squid
3906 may open to this peer, including already opened idle
3907 and standby connections. There is no peer-specific
3908 connection limit by default.
3910 A peer exceeding the limit is not used for new
3911 requests unless a standby connection is available.
3913 max-conn currently works poorly with idle persistent
3914 connections: When a peer reaches its max-conn limit,
3915 and there are idle persistent connections to the peer,
3916 the peer may not be selected because the limiting code
3917 does not know whether Squid can reuse those idle
3920 standby=N Maintain a pool of N "hot standby" connections to an
3921 UP peer, available for requests when no idle
3922 persistent connection is available (or safe) to use.
3923 By default and with zero N, no such pool is maintained.
3924 N must not exceed the max-conn limit (if any).
3926 At start or after reconfiguration, Squid opens new TCP
3927 standby connections until there are N connections
3928 available and then replenishes the standby pool as
3929 opened connections are used up for requests. A used
3930 connection never goes back to the standby pool, but
3931 may go to the regular idle persistent connection pool
3932 shared by all peers and origin servers.
3934 Squid never opens multiple new standby connections
3935 concurrently. This one-at-a-time approach minimizes
3936 flooding-like effect on peers. Furthermore, just a few
3937 standby connections should be sufficient in most cases
3938 to supply most new requests with a ready-to-use
3941 Standby connections obey server_idle_pconn_timeout.
3942 For the feature to work as intended, the peer must be
3943 configured to accept and keep them open longer than
3944 the idle timeout at the connecting Squid, to minimize
3945 race conditions typical to idle used persistent
3946 connections. Default request_timeout and
3947 server_idle_pconn_timeout values ensure such a
3950 name=xxx Unique name for the peer.
3951 Required if you have multiple peers on the same host
3952 but different ports.
3953 This name can be used in cache_peer_access and similar
3954 directives to identify the peer.
3955 Can be used by outgoing access controls through the
3958 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3959 requests to this peer. Use normal address selection instead.
3960 This overrides the spoof_client_ip ACL.
3962 proxy-only objects fetched from the peer will not be stored locally.
3966 NAME: cache_peer_access
3969 DEFAULT_DOC: No peer usage restrictions.
3972 Restricts usage of cache_peer proxies.
3975 cache_peer_access peer-name allow|deny [!]aclname ...
3977 For the required peer-name parameter, use either the value of the
3978 cache_peer name=value parameter or, if name=value is missing, the
3979 cache_peer hostname parameter.
3981 This directive narrows down the selection of peering candidates, but
3982 does not determine the order in which the selected candidates are
3983 contacted. That order is determined by the peer selection algorithms
3984 (see PEER SELECTION sections in the cache_peer documentation).
3986 If a deny rule matches, the corresponding peer will not be contacted
3987 for the current transaction -- Squid will not send ICP queries and
3988 will not forward HTTP requests to that peer. An allow match leaves
3989 the corresponding peer in the selection. The first match for a given
3990 peer wins for that peer.
3992 The relative order of cache_peer_access directives for the same peer
3993 matters. The relative order of any two cache_peer_access directives
3994 for different peers does not matter. To ease interpretation, it is a
3995 good idea to group cache_peer_access directives for the same peer
3998 A single cache_peer_access directive may be evaluated multiple times
3999 for a given transaction because individual peer selection algorithms
4000 may check it independently from each other. These redundant checks
4001 may be optimized away in future Squid versions.
4003 This clause only supports fast acl types.
4004 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4008 NAME: neighbor_type_domain
4009 TYPE: hostdomaintype
4011 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
4014 Modify the cache_peer neighbor type when passing requests
4015 about specific domains to the peer.
4018 neighbor_type_domain neighbor parent|sibling domain domain ...
4021 cache_peer foo.example.com parent 3128 3130
4022 neighbor_type_domain foo.example.com sibling .au .de
4024 The above configuration treats all requests to foo.example.com as a
4025 parent proxy unless the request is for a .au or .de ccTLD domain name.
4028 NAME: dead_peer_timeout
4032 LOC: Config.Timeout.deadPeer
4034 This controls how long Squid waits to declare a peer cache
4035 as "dead." If there are no ICP replies received in this
4036 amount of time, Squid will declare the peer dead and not
4037 expect to receive any further ICP replies. However, it
4038 continues to send ICP queries, and will mark the peer as
4039 alive upon receipt of the first subsequent ICP reply.
4041 This timeout also affects when Squid expects to receive ICP
4042 replies from peers. If more than 'dead_peer' seconds have
4043 passed since the last ICP reply was received, Squid will not
4044 expect to receive an ICP reply on the next query. Thus, if
4045 your time between requests is greater than this timeout, you
4046 will see a lot of requests sent DIRECT to origin servers
4047 instead of to your parents.
4050 NAME: forward_max_tries
4053 LOC: Config.forward_max_tries
4055 Limits the number of attempts to forward the request.
4057 For the purpose of this limit, Squid counts all high-level request
4058 forwarding attempts, including any same-destination retries after
4059 certain persistent connection failures and any attempts to use a
4060 different peer. However, these low-level attempts are not counted:
4061 * connection reopening attempts (enabled using connect_retries)
4062 * unfinished Happy Eyeballs connection attempts (prevented by setting
4063 happy_eyeballs_connect_limit to 0)
4065 See also: forward_timeout and connect_retries.
4069 MEMORY CACHE OPTIONS
4070 -----------------------------------------------------------------------------
4077 LOC: Config.memMaxSize
4079 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
4080 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
4081 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
4082 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
4084 'cache_mem' specifies the ideal amount of memory to be used
4086 * In-Transit objects
4088 * Negative-Cached objects
4090 Data for these objects are stored in 4 KB blocks. This
4091 parameter specifies the ideal upper limit on the total size of
4092 4 KB blocks allocated. In-Transit objects take the highest
4095 In-transit objects have priority over the others. When
4096 additional space is needed for incoming data, negative-cached
4097 and hot objects will be released. In other words, the
4098 negative-cached and hot objects will fill up any unused space
4099 not needed for in-transit objects.
4101 If circumstances require, this limit will be exceeded.
4102 Specifically, if your incoming request rate requires more than
4103 'cache_mem' of memory to hold in-transit objects, Squid will
4104 exceed this limit to satisfy the new requests. When the load
4105 decreases, blocks will be freed until the high-water mark is
4106 reached. Thereafter, blocks will be used to store hot
4109 If shared memory caching is enabled, Squid does not use the shared
4110 cache space for in-transit objects, but they still consume as much
4111 local memory as they need. For more details about the shared memory
4112 cache, see memory_cache_shared.
4115 NAME: maximum_object_size_in_memory
4119 LOC: Config.Store.maxInMemObjSize
4121 Objects greater than this size will not be attempted to kept in
4122 the memory cache. This should be set high enough to keep objects
4123 accessed frequently in memory to improve performance whilst low
4124 enough to keep larger objects from hoarding cache_mem.
4127 NAME: memory_cache_shared
4130 LOC: Config.memShared
4132 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
4134 Controls whether the memory cache is shared among SMP workers.
4136 The shared memory cache is meant to occupy cache_mem bytes and replace
4137 the non-shared memory cache, although some entities may still be
4138 cached locally by workers for now (e.g., internal and in-transit
4139 objects may be served from a local memory cache even if shared memory
4140 caching is enabled).
4142 By default, the memory cache is shared if and only if all of the
4143 following conditions are satisfied: Squid runs in SMP mode with
4144 multiple workers, cache_mem is positive, and Squid environment
4145 supports required IPC primitives (e.g., POSIX shared memory segments
4146 and GCC-style atomic operations).
4148 To avoid blocking locks, shared memory uses opportunistic algorithms
4149 that do not guarantee that every cachable entity that could have been
4150 shared among SMP workers will actually be shared.
4153 NAME: memory_cache_mode
4157 DEFAULT_DOC: Keep the most recently fetched objects in memory
4159 Controls which objects to keep in the memory cache (cache_mem)
4161 always Keep most recently fetched objects in memory (default)
4163 disk Only disk cache hits are kept in memory, which means
4164 an object must first be cached on disk and then hit
4165 a second time before cached in memory.
4167 network Only objects fetched from network is kept in memory
4170 NAME: memory_replacement_policy
4172 LOC: Config.memPolicy
4175 The memory replacement policy parameter determines which
4176 objects are purged from memory when memory space is needed.
4178 See cache_replacement_policy for details on algorithms.
4183 -----------------------------------------------------------------------------
4186 NAME: cache_replacement_policy
4188 LOC: Config.replPolicy
4191 The cache replacement policy parameter determines which
4192 objects are evicted (replaced) when disk space is needed.
4194 lru : Squid's original list based LRU policy
4195 heap GDSF : Greedy-Dual Size Frequency
4196 heap LFUDA: Least Frequently Used with Dynamic Aging
4197 heap LRU : LRU policy implemented using a heap
4199 Applies to any cache_dir lines listed below this directive.
4201 The LRU policies keeps recently referenced objects.
4203 The heap GDSF policy optimizes object hit rate by keeping smaller
4204 popular objects in cache so it has a better chance of getting a
4205 hit. It achieves a lower byte hit rate than LFUDA though since
4206 it evicts larger (possibly popular) objects.
4208 The heap LFUDA policy keeps popular objects in cache regardless of
4209 their size and thus optimizes byte hit rate at the expense of
4210 hit rate since one large, popular object will prevent many
4211 smaller, slightly less popular objects from being cached.
4213 Both policies utilize a dynamic aging mechanism that prevents
4214 cache pollution that can otherwise occur with frequency-based
4215 replacement policies.
4217 NOTE: if using the LFUDA replacement policy you should increase
4218 the value of maximum_object_size above its default of 4 MB to
4219 to maximize the potential byte hit rate improvement of LFUDA.
4221 For more information about the GDSF and LFUDA cache replacement
4222 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
4223 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
4226 NAME: minimum_object_size
4230 DEFAULT_DOC: no limit
4231 LOC: Config.Store.minObjectSize
4233 Objects smaller than this size will NOT be saved on disk. The
4234 value is specified in bytes, and the default is 0 KB, which
4235 means all responses can be stored.
4238 NAME: maximum_object_size
4242 LOC: Config.Store.maxObjectSize
4244 Set the default value for max-size parameter on any cache_dir.
4245 The value is specified in bytes, and the default is 4 MB.
4247 If you wish to get a high BYTES hit ratio, you should probably
4248 increase this (one 32 MB object hit counts for 3200 10KB
4251 If you wish to increase hit ratio more than you want to
4252 save bandwidth you should leave this low.
4254 NOTE: if using the LFUDA replacement policy you should increase
4255 this value to maximize the byte hit rate improvement of LFUDA!
4256 See cache_replacement_policy for a discussion of this policy.
4262 DEFAULT_DOC: No disk cache. Store cache objects only in memory.
4263 LOC: Config.cacheSwap
4266 cache_dir Type Directory-Name Fs-specific-data [options]
4268 You can specify multiple cache_dir lines to spread the
4269 cache among different disk partitions.
4271 Type specifies the kind of storage system to use. Only "ufs"
4272 is built by default. To enable any of the other storage systems
4273 see the --enable-storeio configure option.
4275 'Directory' is a top-level directory where cache swap
4276 files will be stored. If you want to use an entire disk
4277 for caching, this can be the mount-point directory.
4278 The directory must exist and be writable by the Squid
4279 process. Squid will NOT create this directory for you.
4281 In SMP configurations, cache_dir must not precede the workers option
4282 and should use configuration macros or conditionals to give each
4283 worker interested in disk caching a dedicated cache directory.
4286 ==== The ufs store type ====
4288 "ufs" is the old well-known Squid storage format that has always
4292 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
4294 'Mbytes' is the amount of disk space (MB) to use under this
4295 directory. The default is 100 MB. Change this to suit your
4296 configuration. Do NOT put the size of your disk drive here.
4297 Instead, if you want Squid to use the entire disk drive,
4298 subtract 20% and use that value.
4300 'L1' is the number of first-level subdirectories which
4301 will be created under the 'Directory'. The default is 16.
4303 'L2' is the number of second-level subdirectories which
4304 will be created under each first-level directory. The default
4308 ==== The aufs store type ====
4310 "aufs" uses the same storage format as "ufs", utilizing
4311 POSIX-threads to avoid blocking the main Squid process on
4312 disk-I/O. This was formerly known in Squid as async-io.
4315 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
4317 see argument descriptions under ufs above
4320 ==== The diskd store type ====
4322 "diskd" uses the same storage format as "ufs", utilizing a
4323 separate process to avoid blocking the main Squid process on
4327 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
4329 see argument descriptions under ufs above
4331 Q1 specifies the number of unacknowledged I/O requests when Squid
4332 stops opening new files. If this many messages are in the queues,
4333 Squid won't open new files. Default is 64
4335 Q2 specifies the number of unacknowledged messages when Squid
4336 starts blocking. If this many messages are in the queues,
4337 Squid blocks until it receives some replies. Default is 72
4339 When Q1 < Q2 (the default), the cache directory is optimized
4340 for lower response time at the expense of a decrease in hit
4341 ratio. If Q1 > Q2, the cache directory is optimized for
4342 higher hit ratio at the expense of an increase in response
4346 ==== The rock store type ====
4349 cache_dir rock Directory-Name Mbytes [options]
4351 The Rock Store type is a database-style storage. All cached
4352 entries are stored in a "database" file, using fixed-size slots.
4353 A single entry occupies one or more slots.
4355 If possible, Squid using Rock Store creates a dedicated kid
4356 process called "disker" to avoid blocking Squid worker(s) on disk
4357 I/O. One disker kid is created for each rock cache_dir. Diskers
4358 are created only when Squid, running in daemon mode, has support
4359 for the IpcIo disk I/O module.
4361 swap-timeout=msec: Squid will not start writing a miss to or
4362 reading a hit from disk if it estimates that the swap operation
4363 will take more than the specified number of milliseconds. By
4364 default and when set to zero, disables the disk I/O time limit
4365 enforcement. Ignored when using blocking I/O module because
4366 blocking synchronous I/O does not allow Squid to estimate the
4367 expected swap wait time.
4369 max-swap-rate=swaps/sec: Artificially limits disk access using
4370 the specified I/O rate limit. Swap out requests that
4371 would cause the average I/O rate to exceed the limit are
4372 delayed. Individual swap in requests (i.e., hits or reads) are
4373 not delayed, but they do contribute to measured swap rate and
4374 since they are placed in the same FIFO queue as swap out
4375 requests, they may wait longer if max-swap-rate is smaller.
4376 This is necessary on file systems that buffer "too
4377 many" writes and then start blocking Squid and other processes
4378 while committing those writes to disk. Usually used together
4379 with swap-timeout to avoid excessive delays and queue overflows
4380 when disk demand exceeds available disk "bandwidth". By default
4381 and when set to zero, disables the disk I/O rate limit
4382 enforcement. Currently supported by IpcIo module only.
4384 slot-size=bytes: The size of a database "record" used for
4385 storing cached responses. A cached response occupies at least
4386 one slot and all database I/O is done using individual slots so
4387 increasing this parameter leads to more disk space waste while
4388 decreasing it leads to more disk I/O overheads. Should be a
4389 multiple of your operating system I/O page size. Defaults to
4390 16KBytes. A housekeeping header is stored with each slot and
4391 smaller slot-sizes will be rejected. The header is smaller than
4395 ==== COMMON OPTIONS ====
4397 no-store no new objects should be stored to this cache_dir.
4399 min-size=n the minimum object size in bytes this cache_dir
4400 will accept. It's used to restrict a cache_dir
4401 to only store large objects (e.g. AUFS) while
4402 other stores are optimized for smaller objects
4406 max-size=n the maximum object size in bytes this cache_dir
4408 The value in maximum_object_size directive sets
4409 the default unless more specific details are
4410 available (ie a small store capacity).
4412 Note: To make optimal use of the max-size limits you should order
4413 the cache_dir lines with the smallest max-size value first.
4417 # Uncomment and adjust the following to add a disk cache directory.
4418 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
4422 NAME: store_dir_select_algorithm
4424 LOC: Config.store_dir_select_algorithm
4427 How Squid selects which cache_dir to use when the response
4428 object will fit into more than one.
4430 Regardless of which algorithm is used the cache_dir min-size
4431 and max-size parameters are obeyed. As such they can affect
4432 the selection algorithm by limiting the set of considered
4439 This algorithm is suited to caches with similar cache_dir
4440 sizes and disk speeds.
4442 The disk with the least I/O pending is selected.
4443 When there are multiple disks with the same I/O load ranking
4444 the cache_dir with most available capacity is selected.
4446 When a mix of cache_dir sizes are configured the faster disks
4447 have a naturally lower I/O loading and larger disks have more
4448 capacity. So space used to store objects and data throughput
4449 may be very unbalanced towards larger disks.
4454 This algorithm is suited to caches with unequal cache_dir
4457 Each cache_dir is selected in a rotation. The next suitable
4460 Available cache_dir capacity is only considered in relation
4461 to whether the object will fit and meets the min-size and
4462 max-size parameters.
4464 Disk I/O loading is only considered to prevent overload on slow
4465 disks. This algorithm does not spread objects by size, so any
4466 I/O loading per-disk may appear very unbalanced and volatile.
4468 If several cache_dirs use similar min-size, max-size, or other
4469 limits to to reject certain responses, then do not group such
4470 cache_dir lines together, to avoid round-robin selection bias
4471 towards the first cache_dir after the group. Instead, interleave
4472 cache_dir lines from different groups. For example:
4474 store_dir_select_algorithm round-robin
4475 cache_dir rock /hdd1 ... min-size=100000
4476 cache_dir rock /ssd1 ... max-size=99999
4477 cache_dir rock /hdd2 ... min-size=100000
4478 cache_dir rock /ssd2 ... max-size=99999
4479 cache_dir rock /hdd3 ... min-size=100000
4480 cache_dir rock /ssd3 ... max-size=99999
4483 NAME: paranoid_hit_validation
4484 COMMENT: time-units-small
4485 TYPE: time_nanoseconds
4487 DEFAULT_DOC: validation disabled
4488 LOC: Config.paranoid_hit_validation
4490 Controls whether Squid should perform paranoid validation of cache entry
4491 metadata integrity every time a cache entry is hit. This low-level
4492 validation should always succeed. Each failed validation results in a
4493 cache miss, a BUG line reported to cache.log, and the invalid entry
4494 marked as unusable (and eventually purged from the cache).
4496 Squid can only validate shared cache memory and rock cache_dir entries.
4498 * Zero (default) value means that the validation is disabled.
4500 * Positive values enable validation:
4501 - values less than 1 day approximate the maximum time that Squid is allowed
4502 to spend validating a single cache hit.
4503 - values greater or equal to 1 day are considered as no limitation:
4504 in this case all checks will be performed, regardless of how much time
4507 Hits are usually stored using 16KB slots (for rock, the size is
4508 configurable via cache_dir slot-size). Larger hits require scanning more
4509 slots and, hence, take more time. When validation is enabled, at least one
4510 slot is always validated, regardless of the configured time limit.
4512 A worker process validating an entry cannot do anything else (i.e. the
4513 validation is blocking). The validation overhead is environment dependent,
4514 but developers have observed Squid spending 3-10 microseconds to check each
4515 slot of a Rock or shared memory hit entry. If Squid cuts validation short
4516 because it runs out of configured time, it treats the entry as valid.
4518 When hit validation is enabled, its statistics is included in Cache
4519 Manager mgr:counters, mgr:5min, and mgr:60min reports.
4522 NAME: max_open_disk_fds
4524 LOC: Config.max_open_disk_fds
4526 DEFAULT_DOC: no limit
4528 To avoid having disk as the I/O bottleneck Squid can optionally
4529 bypass the on-disk cache if more than this amount of disk file
4530 descriptors are open.
4532 A value of 0 indicates no limit.
4535 NAME: cache_swap_low
4536 COMMENT: (percent, 0-100)
4539 LOC: Config.Swap.lowWaterMark
4541 The low-water mark for AUFS/UFS/diskd cache object eviction by
4542 the cache_replacement_policy algorithm.
4544 Removal begins when the swap (disk) usage of a cache_dir is
4545 above this low-water mark and attempts to maintain utilization
4546 near the low-water mark.
4548 As swap utilization increases towards the high-water mark set
4549 by cache_swap_high object eviction becomes more aggressive.
4551 The value difference in percentages between low- and high-water
4552 marks represent an eviction rate of 300 objects per second and
4553 the rate continues to scale in aggressiveness by multiples of
4554 this above the high-water mark.
4556 Defaults are 90% and 95%. If you have a large cache, 5% could be
4557 hundreds of MB. If this is the case you may wish to set these
4558 numbers closer together.
4560 See also cache_swap_high and cache_replacement_policy
4563 NAME: cache_swap_high
4564 COMMENT: (percent, 0-100)
4567 LOC: Config.Swap.highWaterMark
4569 The high-water mark for AUFS/UFS/diskd cache object eviction by
4570 the cache_replacement_policy algorithm.
4572 Removal begins when the swap (disk) usage of a cache_dir is
4573 above the low-water mark set by cache_swap_low and attempts to
4574 maintain utilization near the low-water mark.
4576 As swap utilization increases towards this high-water mark object
4577 eviction becomes more aggressive.
4579 The value difference in percentages between low- and high-water
4580 marks represent an eviction rate of 300 objects per second and
4581 the rate continues to scale in aggressiveness by multiples of
4582 this above the high-water mark.
4584 Defaults are 90% and 95%. If you have a large cache, 5% could be
4585 hundreds of MB. If this is the case you may wish to set these
4586 numbers closer together.
4588 See also cache_swap_low and cache_replacement_policy
4593 -----------------------------------------------------------------------------
4600 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
4604 logformat <name> <format specification>
4606 Defines an access log format.
4608 The <format specification> is a string with embedded % format codes
4610 % format codes all follow the same basic structure where all
4611 components but the formatcode are optional and usually unnecessary,
4612 especially when dealing with common codes.
4614 % [encoding] [-] [[0]width] [{arg}] formatcode [{arg}]
4616 encoding escapes or otherwise protects "special" characters:
4618 " Quoted string encoding where quote(") and
4619 backslash(\) characters are \-escaped while
4620 CR, LF, and TAB characters are encoded as \r,
4621 \n, and \t two-character sequences.
4623 [ Custom Squid encoding where percent(%), square
4624 brackets([]), backslash(\) and characters with
4625 codes outside of [32,126] range are %-encoded.
4626 SP is not encoded. Used by log_mime_hdrs.
4628 # URL encoding (a.k.a. percent-encoding) where
4629 all URL unsafe and control characters (per RFC
4630 1738) are %-encoded.
4632 / Shell-like encoding where quote(") and
4633 backslash(\) characters are \-escaped while CR
4634 and LF characters are encoded as \r and \n
4635 two-character sequences. Values containing SP
4636 character(s) are surrounded by quotes(").
4638 ' Raw/as-is encoding with no escaping/quoting.
4640 Default encoding: When no explicit encoding is
4641 specified, each %code determines its own encoding.
4642 Most %codes use raw/as-is encoding, but some codes use
4643 a so called "pass-through URL encoding" where all URL
4644 unsafe and control characters (per RFC 1738) are
4645 %-encoded, but the percent character(%) is left as is.
4649 width minimum and/or maximum field width:
4650 [width_min][.width_max]
4651 When minimum starts with 0, the field is zero-padded.
4652 String values exceeding maximum width are truncated.
4654 {arg} argument such as header name etc. This field may be
4655 placed before or after the token, but not both at once.
4659 % a literal % character
4660 sn Unique sequence number per log line entry
4661 err_code The ID of an error response served by Squid or
4662 a similar internal error identifier.
4663 err_detail Additional err_code-dependent error information.
4664 note The annotation specified by the argument. Also
4665 logs the adaptation meta headers set by the
4666 adaptation_meta configuration parameter.
4667 If no argument given all annotations logged.
4668 The argument may include a separator to use with
4671 By default, multiple note values are separated with ","
4672 and multiple notes are separated with "\r\n".
4673 When logging named notes with %{name}note, the
4674 explicitly configured separator is used between note
4675 values. When logging all notes with %note, the
4676 explicitly configured separator is used between
4677 individual notes. There is currently no way to
4678 specify both value and notes separators when logging
4679 all notes with %note.
4680 master_xaction The master transaction identifier is an unsigned
4681 integer. These IDs are guaranteed to monotonically
4682 increase within a single worker process lifetime, with
4683 higher values corresponding to transactions that were
4684 accepted or initiated later. Due to current implementation
4685 deficiencies, some IDs are skipped (i.e. never logged).
4686 Concurrent workers and restarted workers use similar,
4687 overlapping sequences of master transaction IDs.
4689 Connection related format codes:
4691 >a Client source IP address
4693 >p Client source port
4694 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
4695 >la Local IP address the client connected to
4696 >lp Local port number the client connected to
4697 >qos Client connection TOS/DSCP value set by Squid
4698 >nfmark Client connection netfilter packet MARK set by Squid
4700 transport::>connection_id Identifies a transport connection
4701 accepted by Squid (e.g., a connection carrying the
4702 logged HTTP request). Currently, Squid only supports
4703 TCP transport connections.
4705 The logged identifier is an unsigned integer. These
4706 IDs are guaranteed to monotonically increase within a
4707 single worker process lifetime, with higher values
4708 corresponding to connections that were accepted later.
4709 Many IDs are skipped (i.e. never logged). Concurrent
4710 workers and restarted workers use similar, partially
4711 overlapping sequences of IDs.
4713 la Local listening IP address the client connection was connected to.
4714 lp Local listening port number the client connection was connected to.
4716 <a Server IP address of the last server or peer connection
4717 <A Server FQDN or peer name
4718 <p Server port number of the last server or peer connection
4719 <la Local IP address of the last server or peer connection
4720 <lp Local port number of the last server or peer connection
4721 <qos Server connection TOS/DSCP value set by Squid
4722 <nfmark Server connection netfilter packet MARK set by Squid
4724 >handshake Raw client handshake
4725 Initial client bytes received by Squid on a newly
4726 accepted TCP connection or inside a just established
4727 CONNECT tunnel. Squid stops accumulating handshake
4728 bytes as soon as the handshake parser succeeds or
4729 fails (determining whether the client is using the
4732 For HTTP clients, the handshake is the request line.
4733 For TLS clients, the handshake consists of all TLS
4734 records up to and including the TLS record that
4735 contains the last byte of the first ClientHello
4736 message. For clients using an unsupported protocol,
4737 this field contains the bytes received by Squid at the
4738 time of the handshake parsing failure.
4740 See the on_unsupported_protocol directive for more
4741 information on Squid handshake traffic expectations.
4743 Current support is limited to these contexts:
4744 - http_port connections, but only when the
4745 on_unsupported_protocol directive is in use.
4746 - https_port connections (and CONNECT tunnels) that
4747 are subject to the ssl_bump peek or stare action.
4749 To protect binary handshake data, this field is always
4750 base64-encoded (RFC 4648 Section 4). If logformat
4751 field encoding is configured, that encoding is applied
4752 on top of base64. Otherwise, the computed base64 value
4755 Time related format codes:
4757 ts Seconds since epoch
4758 tu subsecond time (milliseconds)
4759 tl Local time. Optional strftime format argument
4760 default %d/%b/%Y:%H:%M:%S %z
4761 tg GMT time. Optional strftime format argument
4762 default %d/%b/%Y:%H:%M:%S %z
4763 tr Response time (milliseconds)
4764 dt Total time spent making DNS lookups (milliseconds)
4765 tS Approximate master transaction start time in
4766 <full seconds since epoch>.<fractional seconds> format.
4767 Currently, Squid considers the master transaction
4768 started when a complete HTTP request header initiating
4769 the transaction is received from the client. This is
4770 the same value that Squid uses to calculate transaction
4771 response time when logging %tr to access.log. Currently,
4772 Squid uses millisecond resolution for %tS values,
4773 similar to the default access.log "current time" field
4776 Access Control related format codes:
4778 et Tag returned by external acl
4779 ea Log string returned by external acl
4780 un User name (any available)
4781 ul User name from authentication
4782 ue User name from external acl helper
4783 ui User name from ident
4784 un A user name. Expands to the first available name
4785 from the following list of information sources:
4786 - authenticated user name, like %ul
4787 - user name supplied by an external ACL, like %ue
4788 - SSL client name, like %us
4789 - ident user name, like %ui
4790 credentials Client credentials. The exact meaning depends on
4791 the authentication scheme: For Basic authentication,
4792 it is the password; for Digest, the realm sent by the
4793 client; for NTLM and Negotiate, the client challenge
4794 or client credentials prefixed with "YR " or "KK ".
4796 HTTP related format codes:
4800 [http::]rm Request method (GET/POST etc)
4801 [http::]>rm Request method from client
4802 [http::]<rm Request method sent to server or peer
4804 [http::]ru Request URL received (or computed) and sanitized
4806 Logs request URI received from the client, a
4807 request adaptation service, or a request
4808 redirector (whichever was applied last).
4810 Computed URLs are URIs of internally generated
4811 requests and various "error:..." URIs.
4813 Honors strip_query_terms and uri_whitespace.
4815 This field is not encoded by default. Encoding
4816 this field using variants of %-encoding will
4817 clash with uri_whitespace modifications that
4818 also use %-encoding.
4820 [http::]>ru Request URL received from the client (or computed)
4822 Computed URLs are URIs of internally generated
4823 requests and various "error:..." URIs.
4825 Unlike %ru, this request URI is not affected
4826 by request adaptation, URL rewriting services,
4827 and strip_query_terms.
4829 Honors uri_whitespace.
4831 This field is using pass-through URL encoding
4832 by default. Encoding this field using other
4833 variants of %-encoding will clash with
4834 uri_whitespace modifications that also use
4837 [http::]<ru Request URL sent to server or peer
4838 [http::]>rs Request URL scheme from client
4839 [http::]<rs Request URL scheme sent to server or peer
4840 [http::]>rd Request URL domain from client
4841 [http::]<rd Request URL domain sent to server or peer
4842 [http::]>rP Request URL port from client
4843 [http::]<rP Request URL port sent to server or peer
4844 [http::]rp Request URL path excluding hostname
4845 [http::]>rp Request URL path excluding hostname from client
4846 [http::]<rp Request URL path excluding hostname sent to server or peer
4847 [http::]rv Request protocol version
4848 [http::]>rv Request protocol version from client
4849 [http::]<rv Request protocol version sent to server or peer
4851 [http::]>h Original received request header.
4852 Usually differs from the request header sent by
4853 Squid, although most fields are often preserved.
4854 Accepts optional header field name/value filter
4855 argument using name[:[separator]element] format.
4856 [http::]>ha Received request header after adaptation and
4857 redirection (pre-cache REQMOD vectoring point).
4858 Usually differs from the request header sent by
4859 Squid, although most fields are often preserved.
4860 Optional header name argument as for >h
4864 [http::]<Hs HTTP status code received from the next hop
4865 [http::]>Hs HTTP status code sent to the client
4867 [http::]<h Reply header. Optional header name argument
4870 [http::]mt MIME content type
4875 [http::]st Total size of request + reply traffic with client
4876 [http::]>st Total size of request received from client.
4877 Excluding chunked encoding bytes.
4878 [http::]<st Total size of reply sent to client (after adaptation)
4880 [http::]>sh Size of request headers received from client
4881 [http::]<sh Size of reply headers sent to client (after adaptation)
4883 [http::]<sH Reply high offset sent
4884 [http::]<sS Upstream object size
4886 [http::]<bs Number of HTTP-equivalent message body bytes
4887 received from the next hop, excluding chunked
4888 transfer encoding and control messages.
4889 Generated FTP listings are treated as
4894 [http::]<pt Peer response time in milliseconds. The timer starts
4895 when the last request byte is sent to the next hop
4896 and stops when the last response byte is received.
4897 [http::]<tt Total time in milliseconds. The timer
4898 starts with the first connect request (or write I/O)
4899 sent to the first selected peer. The timer stops
4900 with the last I/O with the last peer.
4902 Squid handling related format codes:
4904 Ss Squid request status (TCP_MISS etc)
4905 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4907 SSL-related format codes:
4909 ssl::bump_mode SslBump decision for the transaction:
4911 For CONNECT requests that initiated bumping of
4912 a connection and for any request received on
4913 an already bumped connection, Squid logs the
4914 corresponding SslBump mode ("splice", "bump",
4915 "peek", "stare", "terminate", "server-first"
4916 or "client-first"). See the ssl_bump option
4917 for more information about these modes.
4919 A "none" token is logged for requests that
4920 triggered "ssl_bump" ACL evaluation matching
4923 In all other cases, a single dash ("-") is
4926 ssl::>sni SSL client SNI sent to Squid.
4929 The Subject field of the received client
4930 SSL certificate or a dash ('-') if Squid has
4931 received an invalid/malformed certificate or
4932 no certificate at all. Consider encoding the
4933 logged value because Subject often has spaces.
4936 The Issuer field of the received client
4937 SSL certificate or a dash ('-') if Squid has
4938 received an invalid/malformed certificate or
4939 no certificate at all. Consider encoding the
4940 logged value because Issuer often has spaces.
4943 The Subject field of the received server
4944 TLS certificate or a dash ('-') if this is
4945 not available. Consider encoding the logged
4946 value because Subject often has spaces.
4949 The Issuer field of the received server
4950 TLS certificate or a dash ('-') if this is
4951 not available. Consider encoding the logged
4952 value because Issuer often has spaces.
4955 The received server x509 certificate in PEM
4956 format, including BEGIN and END lines (or a
4957 dash ('-') if the certificate is unavailable).
4959 WARNING: Large certificates will exceed the
4960 current 8KB access.log record limit, resulting
4961 in truncated records. Such truncation usually
4962 happens in the middle of a record field. The
4963 limit applies to all access logging modules.
4965 The logged certificate may have failed
4966 validation and may not be trusted by Squid.
4967 This field does not include any intermediate
4968 certificates that may have been received from
4969 the server or fetched during certificate
4972 Currently, Squid only collects server
4973 certificates during step3 of SslBump
4974 processing; connections that were not subject
4975 to ssl_bump rules or that did not match a peek
4976 or stare rule at step2 will not have the
4977 server certificate information.
4979 This field is using pass-through URL encoding
4983 The list of certificate validation errors
4984 detected by Squid (including OpenSSL and
4985 certificate validation helper components). The
4986 errors are listed in the discovery order. By
4987 default, the error codes are separated by ':'.
4988 Accepts an optional separator argument.
4990 %ssl::>negotiated_version The negotiated TLS version of the
4993 %ssl::<negotiated_version The negotiated TLS version of the
4994 last server or peer connection.
4996 %ssl::>received_hello_version The TLS version of the Hello
4997 message received from TLS client.
4999 %ssl::<received_hello_version The TLS version of the Hello
5000 message received from TLS server.
5002 %ssl::>received_supported_version The maximum TLS version
5003 supported by the TLS client.
5005 %ssl::<received_supported_version The maximum TLS version
5006 supported by the TLS server.
5008 %ssl::>negotiated_cipher The negotiated cipher of the
5011 %ssl::<negotiated_cipher The negotiated cipher of the
5012 last server or peer connection.
5014 If ICAP is enabled, the following code becomes available (as
5015 well as ICAP log codes documented with the icap_log option):
5017 icap::tt Total ICAP "blocking" time for the HTTP transaction. The
5018 timer ticks while Squid checks adaptation_access and while
5019 ICAP transaction(s) expect ICAP response headers, including
5020 the embedded adapted HTTP message headers (where applicable).
5021 This measurement is meant to estimate ICAP impact on HTTP
5022 transaction response times, but it does not currently account
5023 for slow ICAP response body delivery blocking HTTP progress.
5025 Once Squid receives the final ICAP response headers (e.g.,
5026 ICAP 200 or 204) and the associated adapted HTTP message
5027 headers (if any) from the ICAP service, the corresponding ICAP
5028 transaction stops affecting this measurement, even though the
5029 transaction itself may continue for a long time (e.g., to
5030 finish sending the ICAP request and/or to finish receiving the
5031 ICAP response body).
5033 When "blocking" sections of multiple concurrent ICAP
5034 transactions overlap in time, the overlapping segment is
5037 To see complete ICAP transaction response times (rather than
5038 the cumulative effect of their blocking sections) use the
5039 %adapt::all_trs logformat code or the icap_log directive.
5041 If adaptation is enabled the following codes become available:
5043 adapt::<last_h The header of the last ICAP response or
5044 meta-information from the last eCAP
5045 transaction related to the HTTP transaction.
5046 Like <h, accepts an optional header name
5049 adapt::sum_trs Summed adaptation transaction response
5050 times recorded as a comma-separated list in
5051 the order of transaction start time. Each time
5052 value is recorded as an integer number,
5053 representing response time of one or more
5054 adaptation (ICAP or eCAP) transaction in
5055 milliseconds. When a failed transaction is
5056 being retried or repeated, its time is not
5057 logged individually but added to the
5058 replacement (next) transaction. Lifetimes of individually
5059 listed adaptation transactions may overlap.
5060 See also: %icap::tt and %adapt::all_trs.
5062 adapt::all_trs All adaptation transaction response times.
5063 Same as %adapt::sum_trs but response times of
5064 individual transactions are never added
5065 together. Instead, all transaction response
5066 times are recorded individually.
5068 You can prefix adapt::*_trs format codes with adaptation
5069 service name in curly braces to record response time(s) specific
5070 to that service. For example: %{my_service}adapt::sum_trs
5072 Format codes related to the PROXY protocol:
5074 proxy_protocol::>h PROXY protocol header, including optional TLVs.
5076 Supports the same field and element reporting/extraction logic
5077 as %http::>h. For configuration and reporting purposes, Squid
5078 maps each PROXY TLV to an HTTP header field: the TLV type
5079 (configured as a decimal integer) is the field name, and the
5080 TLV value is the field value. All TLVs of "LOCAL" connections
5081 (in PROXY protocol terminology) are currently skipped/ignored.
5083 Squid also maps the following standard PROXY protocol header
5084 blocks to pseudo HTTP headers (their names use PROXY
5085 terminology and start with a colon, following HTTP tradition
5086 for pseudo headers): :command, :version, :src_addr, :dst_addr,
5087 :src_port, and :dst_port.
5089 Without optional parameters, this logformat code logs
5090 pseudo headers and TLVs.
5092 This format code uses pass-through URL encoding by default.
5095 # relay custom PROXY TLV #224 to adaptation services
5096 adaptation_meta Client-Foo "%proxy_protocol::>h{224}"
5100 The default formats available (which do not need re-defining) are:
5102 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
5103 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
5104 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
5105 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
5106 logformat useragent %>a [%tl] "%{User-Agent}>h"
5108 NOTE: When the log_mime_hdrs directive is set to ON.
5109 The squid, common and combined formats have a safely encoded copy
5110 of the mime headers appended to each line within a pair of brackets.
5112 NOTE: The common and combined formats are not quite true to the Apache definition.
5113 The logs from Squid contain an extra status and hierarchy code appended.
5117 NAME: access_log cache_access_log
5119 LOC: Config.Log.accesslogs
5120 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
5122 Configures whether and how Squid logs HTTP and ICP transactions.
5123 If access logging is enabled, a single line is logged for every
5124 matching HTTP or ICP request. The recommended directive formats are:
5126 access_log <module>:<place> [option ...] [acl acl ...]
5127 access_log none [acl acl ...]
5129 The following directive format is accepted but may be deprecated:
5130 access_log <module>:<place> [<logformat name> [acl acl ...]]
5132 In most cases, the first ACL name must not contain the '=' character
5133 and should not be equal to an existing logformat name. You can always
5134 start with an 'all' ACL to work around those restrictions.
5136 Will log to the specified module:place using the specified format (which
5137 must be defined in a logformat directive) those entries which match
5138 ALL the acl's specified (which must be defined in acl clauses).
5139 If no acl is specified, all requests will be logged to this destination.
5141 ===== Available options for the recommended directive format =====
5143 logformat=name Names log line format (either built-in or
5144 defined by a logformat directive). Defaults
5147 buffer-size=64KB Defines approximate buffering limit for log
5148 records (see buffered_logs). Squid should not
5149 keep more than the specified size and, hence,
5150 should flush records before the buffer becomes
5151 full to avoid overflows under normal
5152 conditions (the exact flushing algorithm is
5153 module-dependent though). The on-error option
5154 controls overflow handling.
5156 on-error=die|drop Defines action on unrecoverable errors. The
5157 'drop' action ignores (i.e., does not log)
5158 affected log records. The default 'die' action
5159 kills the affected worker. The drop action
5160 support has not been tested for modules other
5163 rotate=N Specifies the number of log file rotations to
5164 make when you run 'squid -k rotate'. The default
5165 is to obey the logfile_rotate directive. Setting
5166 rotate=0 will disable the file name rotation,
5167 but the log files are still closed and re-opened.
5168 This will enable you to rename the logfiles
5169 yourself just before sending the rotate signal.
5170 Only supported by the stdio module.
5172 ===== Modules Currently available =====
5174 none Do not log any requests matching these ACL.
5175 Do not specify Place or logformat name.
5177 stdio Write each log line to disk immediately at the completion of
5179 Place: the filename and path to be written.
5181 daemon Very similar to stdio. But instead of writing to disk the log
5182 line is passed to a daemon helper for asychronous handling instead.
5183 Place: varies depending on the daemon.
5185 log_file_daemon Place: the file name and path to be written.
5187 syslog To log each request via syslog facility.
5188 Place: The syslog facility and priority level for these entries.
5189 Place Format: facility.priority
5191 where facility could be any of:
5192 authpriv, daemon, local0 ... local7 or user.
5194 And priority could be any of:
5195 err, warning, notice, info, debug.
5197 udp To send each log line as text data to a UDP receiver.
5198 Place: The destination host name or IP and port.
5199 Place Format: //host:port
5201 tcp To send each log line as text data to a TCP receiver.
5202 Lines may be accumulated before sending (see buffered_logs).
5203 Place: The destination host name or IP and port.
5204 Place Format: //host:port
5207 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
5213 LOC: Config.Log.icaplogs
5216 ICAP log files record ICAP transaction summaries, one line per
5219 The icap_log option format is:
5220 icap_log <filepath> [<logformat name> [acl acl ...]]
5221 icap_log none [acl acl ...]]
5223 Please see access_log option documentation for details. The two
5224 kinds of logs share the overall configuration approach and many
5227 ICAP processing of a single HTTP message or transaction may
5228 require multiple ICAP transactions. In such cases, multiple
5229 ICAP transaction log lines will correspond to a single access
5232 ICAP log supports many access.log logformat %codes. In ICAP context,
5233 HTTP message-related %codes are applied to the HTTP message embedded
5234 in an ICAP message. Logformat "%http::>..." codes are used for HTTP
5235 messages embedded in ICAP requests while "%http::<..." codes are used
5236 for HTTP messages embedded in ICAP responses. For example:
5238 http::>h To-be-adapted HTTP message headers sent by Squid to
5239 the ICAP service. For REQMOD transactions, these are
5240 HTTP request headers. For RESPMOD, these are HTTP
5241 response headers, but Squid currently cannot log them
5242 (i.e., %http::>h will expand to "-" for RESPMOD).
5244 http::<h Adapted HTTP message headers sent by the ICAP
5245 service to Squid (i.e., HTTP request headers in regular
5246 REQMOD; HTTP response headers in RESPMOD and during
5247 request satisfaction in REQMOD).
5249 ICAP OPTIONS transactions do not embed HTTP messages.
5251 Several logformat codes below deal with ICAP message bodies. An ICAP
5252 message body, if any, typically includes a complete HTTP message
5253 (required HTTP headers plus optional HTTP message body). When
5254 computing HTTP message body size for these logformat codes, Squid
5255 either includes or excludes chunked encoding overheads; see
5256 code-specific documentation for details.
5258 For Secure ICAP services, all size-related information is currently
5259 computed before/after TLS encryption/decryption, as if TLS was not
5262 The following format codes are also available for ICAP logs:
5264 icap::<A ICAP server IP address. Similar to <A.
5266 icap::<service_name ICAP service name from the icap_service
5267 option in Squid configuration file.
5269 icap::ru ICAP Request-URI. Similar to ru.
5271 icap::rm ICAP request method (REQMOD, RESPMOD, or
5272 OPTIONS). Similar to existing rm.
5274 icap::>st The total size of the ICAP request sent to the ICAP
5275 server (ICAP headers + ICAP body), including chunking
5278 icap::<st The total size of the ICAP response received from the
5279 ICAP server (ICAP headers + ICAP body), including
5280 chunking metadata (if any).
5282 icap::<bs The size of the ICAP response body received from the
5283 ICAP server, excluding chunking metadata (if any).
5285 icap::tr Transaction response time (in
5286 milliseconds). The timer starts when
5287 the ICAP transaction is created and
5288 stops when the transaction is completed.
5291 icap::tio Transaction I/O time (in milliseconds). The
5292 timer starts when the first ICAP request
5293 byte is scheduled for sending. The timers
5294 stops when the last byte of the ICAP response
5297 icap::to Transaction outcome: ICAP_ERR* for all
5298 transaction errors, ICAP_OPT for OPTION
5299 transactions, ICAP_ECHO for 204
5300 responses, ICAP_MOD for message
5301 modification, and ICAP_SAT for request
5302 satisfaction. Similar to Ss.
5304 icap::Hs ICAP response status code. Similar to Hs.
5306 icap::>h ICAP request header(s). Similar to >h.
5308 icap::<h ICAP response header(s). Similar to <h.
5310 The default ICAP log format, which can be used without an explicit
5311 definition, is called icap_squid:
5313 logformat icap_squid %ts.%03tu %6icap::tr %>A %icap::to/%03icap::Hs %icap::<st %icap::rm %icap::ru %un -/%icap::<A -
5315 See also: logformat and %adapt::<last_h
5318 NAME: logfile_daemon
5320 DEFAULT: @DEFAULT_LOGFILED@
5321 LOC: Log::TheConfig.logfile_daemon
5323 Specify the path to the logfile-writing daemon. This daemon is
5324 used to write the access and store logs, if configured.
5326 Squid sends a number of commands to the log daemon:
5327 L<data>\n - logfile data
5332 r<n>\n - set rotate count to <n>
5333 b<n>\n - 1 = buffer output, 0 = don't buffer output
5335 No responses is expected.
5338 NAME: stats_collection
5340 LOC: Config.accessList.stats_collection
5342 DEFAULT_DOC: Allow logging for all transactions.
5343 COMMENT: allow|deny acl acl...
5345 This options allows you to control which requests gets accounted
5346 in performance counters.
5348 This clause only supports fast acl types.
5349 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5352 NAME: cache_store_log
5355 LOC: Config.Log.store
5357 Logs the activities of the storage manager. Shows which
5358 objects are ejected from the cache, and which objects are
5359 saved and for how long.
5360 There are not really utilities to analyze this data, so you can safely
5361 disable it (the default).
5363 Store log uses modular logging outputs. See access_log for the list
5364 of modules supported.
5367 cache_store_log stdio:@DEFAULT_STORE_LOG@
5368 cache_store_log daemon:@DEFAULT_STORE_LOG@
5371 NAME: cache_swap_state cache_swap_log
5373 LOC: Config.Log.swap
5375 DEFAULT_DOC: Store the journal inside its cache_dir
5377 Location for the cache "swap.state" file. This index file holds
5378 the metadata of objects saved on disk. It is used to rebuild
5379 the cache during startup. Normally this file resides in each
5380 'cache_dir' directory, but you may specify an alternate
5381 pathname here. Note you must give a full filename, not just
5382 a directory. Since this is the index for the whole object
5383 list you CANNOT periodically rotate it!
5385 If %s can be used in the file name it will be replaced with a
5386 a representation of the cache_dir name where each / is replaced
5387 with '.'. This is needed to allow adding/removing cache_dir
5388 lines when cache_swap_log is being used.
5390 If have more than one 'cache_dir', and %s is not used in the name
5391 these swap logs will have names such as:
5397 The numbered extension (which is added automatically)
5398 corresponds to the order of the 'cache_dir' lines in this
5399 configuration file. If you change the order of the 'cache_dir'
5400 lines in this file, these index files will NOT correspond to
5401 the correct 'cache_dir' entry (unless you manually rename
5402 them). We recommend you do NOT use this option. It is
5403 better to keep these index files in each 'cache_dir' directory.
5406 NAME: logfile_rotate
5409 LOC: Config.Log.rotateNumber
5411 Specifies the default number of logfile rotations to make when you
5412 type 'squid -k rotate'. The default is 10, which will rotate
5413 with extensions 0 through 9. Setting logfile_rotate to 0 will
5414 disable the file name rotation, but the logfiles are still closed
5415 and re-opened. This will enable you to rename the logfiles
5416 yourself just before sending the rotate signal.
5418 Note, from Squid-3.1 this option is only a default for cache.log,
5419 that log can be rotated separately by using debug_options.
5421 Note, from Squid-4 this option is only a default for access.log
5422 recorded by stdio: module. Those logs can be rotated separately by
5423 using the rotate=N option on their access_log directive.
5425 Note, the 'squid -k rotate' command normally sends a USR1
5426 signal to the running squid process. In certain situations
5427 (e.g. on Linux with Async I/O), USR1 is used for other
5428 purposes, so -k rotate uses another signal. It is best to get
5429 in the habit of using 'squid -k rotate' instead of 'kill -USR1
5436 DEFAULT: @DEFAULT_MIME_TABLE@
5437 LOC: Config.mimeTablePathname
5439 Path to Squid's icon configuration file.
5441 You shouldn't need to change this, but the default file contains
5442 examples and formatting information if you do.
5448 LOC: Config.onoff.log_mime_hdrs
5451 The Cache can record both the request and the response MIME
5452 headers for each HTTP transaction. The headers are encoded
5453 safely and will appear as two bracketed fields at the end of
5454 the access log (for either the native or httpd-emulated log
5455 formats). To enable this logging set log_mime_hdrs to 'on'.
5460 DEFAULT: @DEFAULT_PID_FILE@
5461 LOC: Config.pidFilename
5463 A filename to write the process-id to. To disable, enter "none".
5466 NAME: client_netmask
5468 LOC: Config.Addrs.client_netmask
5470 DEFAULT_DOC: Log full client IP address
5472 A netmask for client addresses in logfiles and cachemgr output.
5473 Change this to protect the privacy of your cache clients.
5474 A netmask of 255.255.255.0 will log all IP's in that range with
5475 the last digit set to '0'.
5478 NAME: strip_query_terms
5480 LOC: Config.onoff.strip_query_terms
5483 By default, Squid strips query terms from requested URLs before
5484 logging. This protects your user's privacy and reduces log size.
5486 When investigating HIT/MISS or other caching behaviour you
5487 will need to disable this to see the full URL used by Squid.
5494 LOC: Config.onoff.buffered_logs
5496 Whether to write/send access_log records ASAP or accumulate them and
5497 then write/send them in larger chunks. Buffering may improve
5498 performance because it decreases the number of I/Os. However,
5499 buffering increases the delay before log records become available to
5500 the final recipient (e.g., a disk file or logging daemon) and,
5501 hence, increases the risk of log records loss.
5503 Note that even when buffered_logs are off, Squid may have to buffer
5504 records if it cannot write/send them immediately due to pending I/Os
5505 (e.g., the I/O writing the previous log record) or connectivity loss.
5507 Currently honored by 'daemon' and 'tcp' access_log modules only.
5510 NAME: netdb_filename
5512 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
5513 LOC: Config.netdbFilename
5516 Where Squid stores it's netdb journal.
5517 When enabled this journal preserves netdb state between restarts.
5519 To disable, enter "none".
5523 TYPE: Security::KeyLog*
5525 LOC: Config.Log.tlsKeys
5528 Configures whether and where Squid records pre-master secret and
5529 related encryption details for TLS connections accepted or established
5530 by Squid. These connections include connections accepted at
5531 https_port, TLS connections opened to origin servers/cache_peers/ICAP
5532 services, and TLS tunnels bumped by Squid using the SslBump feature.
5533 This log (a.k.a. SSLKEYLOGFILE) is meant for triage with traffic
5534 inspection tools like Wireshark.
5536 tls_key_log <destination> [options] [if [!]<acl>...]
5538 WARNING: This log allows anybody to decrypt the corresponding
5539 encrypted TLS connections, both in-flight and postmortem.
5541 At most one log file is supported at this time. Repeated tls_key_log
5542 directives are treated as fatal configuration errors. By default, no
5543 log is created or updated.
5545 If the log file does not exist, Squid creates it. Otherwise, Squid
5546 appends an existing log file.
5548 The directive is consulted whenever a TLS connection is accepted or
5549 established by Squid. TLS connections that fail the handshake may be
5550 logged if Squid got enough information to form a log record. A record
5551 is logged only if all of the configured ACLs match.
5553 While transport-related ACLs like src and dst should work, Squid may
5554 not have access to higher-level information. For example, when logging
5555 accepted https_port connections, Squid does not yet have access to the
5556 expected HTTPS request. Similarly, an HTTPS response is not available
5557 when logging most TLS connections established by Squid.
5559 The log record format is meant to be compatible with TLS deciphering
5560 features of Wireshark which relies on fields like CLIENT_RANDOM and
5561 RSA Master-Key. A single log record usually spans multiple lines.
5562 Technical documentation for that format is maintained inside the
5563 Wireshark code (e.g., see tls_keylog_process_lines() comments as of
5564 Wireshark commit e3d44136f0f0026c5e893fa249f458073f3b7328). TLS key
5565 log does not support custom record formats.
5567 This clause only supports fast acl types.
5568 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5570 See access_log's <module>:<place> parameter for a list of supported
5571 logging destinations.
5573 TLS key log supports all access_log key=value options with the
5574 exception of logformat=name.
5576 Requires Squid built with OpenSSL support.
5581 OPTIONS FOR TROUBLESHOOTING
5582 -----------------------------------------------------------------------------
5587 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
5588 LOC: Debug::cache_log
5590 Squid administrative logging file.
5592 This is where general information about Squid behavior goes. You can
5593 increase the amount of data logged to this file and how often it is
5594 rotated with "debug_options"
5597 NAME: cache_log_message
5598 TYPE: cache_log_message
5600 DEFAULT_DOC: Use debug_options.
5601 LOC: Config.debugMessages
5603 Configures logging of individual cache.log messages.
5605 cache_log_message id=<number> option...
5606 cache_log_message ids=<number>-<number> option...
5608 Most messages have _not_ been instrumented to support this directive
5609 yet. For the list of instrumented messages and their IDs, please see
5610 the doc/debug-messages.txt file.
5612 Message ID corresponds to the message semantics rather than message
5613 text or source code location. The ID is stable across Squid
5614 instances and versions. Substantial changes in message semantics
5615 result in a new ID assignment. To reduce the danger of suppressing
5616 an important log message, the old IDs of removed (or substantially
5617 changed) messages are never reused.
5619 If more than one cache_log_message directive refers to the same
5620 message ID, the last directive wins.
5622 Use ids=min-max syntax to apply the same message configuration to an
5623 inclusive range of message IDs. An ID range with N values has
5624 exactly the same effect as typing N cache_log_message lines.
5626 At least one option is required. Supported options are:
5628 level=<number>: The logging level to use for the message. Squid
5629 command line options (-s and -d) as well as the debug_options
5630 directive control which levels go to syslog, stderr, and/or
5631 cache.log. In most environments, using level=2 or higher stops
5632 Squid from logging the message anywhere. By default, the
5633 hard-coded message-specific level is used.
5635 limit=<number>: After logging the specified number of messages at
5636 the configured (or default) debugging level DL, start using
5637 level 3 (for DL 0 and 1) or 8 (for higher DL values). Usually,
5638 level-3+ messages are not logged anywhere so this option can
5639 often be used to effectively suppress the message. Each SMP
5640 Squid process gets the same limit.
5646 DEFAULT_DOC: Log all critical and important messages.
5647 LOC: Debug::debugOptions
5649 Logging options are set as section,level where each source file
5650 is assigned a unique section. Lower levels result in less
5651 output, Full debugging (level 9) can result in a very large
5652 log file, so be careful.
5654 The magic word "ALL" sets debugging levels for all sections.
5655 The default is to run with "ALL,1" to record important warnings.
5657 The rotate=N option can be used to keep more or less of these logs
5658 than would otherwise be kept by logfile_rotate.
5659 For most uses a single log should be enough to monitor current
5660 events affecting Squid.
5665 LOC: Config.coredump_dir
5666 DEFAULT_IF_NONE: none
5667 DEFAULT_DOC: Use the directory from where Squid was started.
5669 By default Squid leaves core files in the directory from where
5670 it was started. If you set 'coredump_dir' to a directory
5671 that exists, Squid will chdir() to that directory at startup
5672 and coredump files will be left there.
5674 In addition to changing the directory, the process permissions are updated
5675 to enable process tracing and/or coredump file generation. The details are
5676 OS-specific, but look for prctl(2) PR_SET_DUMPABLE and procctl(2)
5677 PROC_TRACE_CTL documentation as guiding examples.
5681 # Leave coredumps in the first cache dir
5682 coredump_dir @DEFAULT_SWAP_DIR@
5688 OPTIONS FOR FTP GATEWAYING
5689 -----------------------------------------------------------------------------
5695 LOC: Config.Ftp.anon_user
5697 If you want the anonymous login password to be more informative
5698 (and enable the use of picky FTP servers), set this to something
5699 reasonable for your domain, like wwwuser@somewhere.net
5701 The reason why this is domainless by default is the
5702 request can be made on the behalf of a user in any domain,
5703 depending on how the cache is used.
5704 Some FTP server also validate the email address is valid
5705 (for example perl.com).
5711 LOC: Config.Ftp.passive
5713 If your firewall does not allow Squid to use passive
5714 connections, turn off this option.
5716 Use of ftp_epsv_all option requires this to be ON.
5722 LOC: Config.Ftp.epsv_all
5724 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
5726 NATs may be able to put the connection on a "fast path" through the
5727 translator, as the EPRT command will never be used and therefore,
5728 translation of the data portion of the segments will never be needed.
5730 When a client only expects to do two-way FTP transfers this may be
5732 If squid finds that it must do a three-way FTP transfer after issuing
5733 an EPSV ALL command, the FTP session will fail.
5735 If you have any doubts about this option do not use it.
5736 Squid will nicely attempt all other connection methods.
5738 Requires ftp_passive to be ON (default) for any effect.
5744 LOC: Config.accessList.ftp_epsv
5746 FTP Protocol extensions permit the use of a special "EPSV" command.
5748 NATs may be able to put the connection on a "fast path" through the
5749 translator using EPSV, as the EPRT command will never be used
5750 and therefore, translation of the data portion of the segments
5751 will never be needed.
5753 EPSV is often required to interoperate with FTP servers on IPv6
5754 networks. On the other hand, it may break some IPv4 servers.
5756 By default, EPSV may try EPSV with any FTP server. To fine tune
5757 that decision, you may restrict EPSV to certain clients or servers
5760 ftp_epsv allow|deny al1 acl2 ...
5762 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
5764 Only fast ACLs are supported.
5765 Requires ftp_passive to be ON (default) for any effect.
5771 LOC: Config.Ftp.eprt
5773 FTP Protocol extensions permit the use of a special "EPRT" command.
5775 This extension provides a protocol neutral alternative to the
5776 IPv4-only PORT command. When supported it enables active FTP data
5777 channels over IPv6 and efficient NAT handling.
5779 Turning this OFF will prevent EPRT being attempted and will skip
5780 straight to using PORT for IPv4 servers.
5782 Some devices are known to not handle this extension correctly and
5783 may result in crashes. Devices which support EPRT enough to fail
5784 cleanly will result in Squid attempting PORT anyway. This directive
5785 should only be disabled when EPRT results in device failures.
5787 WARNING: Doing so will convert Squid back to the old behavior with all
5788 the related problems with external NAT devices/layers and IPv4-only FTP.
5791 NAME: ftp_sanitycheck
5794 LOC: Config.Ftp.sanitycheck
5796 For security and data integrity reasons Squid by default performs
5797 sanity checks of the addresses of FTP data connections ensure the
5798 data connection is to the requested server. If you need to allow
5799 FTP connections to servers using another IP address for the data
5800 connection turn this off.
5803 NAME: ftp_telnet_protocol
5806 LOC: Config.Ftp.telnet
5808 The FTP protocol is officially defined to use the telnet protocol
5809 as transport channel for the control connection. However, many
5810 implementations are broken and does not respect this aspect of
5813 If you have trouble accessing files with ASCII code 255 in the
5814 path or similar problems involving this ASCII code you can
5815 try setting this directive to off. If that helps, report to the
5816 operator of the FTP server in question that their FTP server
5817 is broken and does not follow the FTP standard.
5821 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
5822 -----------------------------------------------------------------------------
5827 DEFAULT: @DEFAULT_DISKD@
5828 LOC: Config.Program.diskd
5830 Specify the location of the diskd executable.
5831 Note this is only useful if you have compiled in
5832 diskd as one of the store io modules.
5835 NAME: unlinkd_program
5838 DEFAULT: @DEFAULT_UNLINKD@
5839 LOC: Config.Program.unlinkd
5841 Specify the location of the executable for file deletion process.
5844 NAME: pinger_program
5847 DEFAULT: @DEFAULT_PINGER@
5850 Specify the location of the executable for the pinger process.
5859 Control whether the pinger is active at run-time.
5860 Enables turning ICMP pinger on and off with a simple
5861 squid -k reconfigure.
5866 OPTIONS FOR URL REWRITING
5867 -----------------------------------------------------------------------------
5870 NAME: url_rewrite_program redirect_program
5872 LOC: Config.Program.redirect
5875 The name and command line parameters of an admin-provided executable
5876 for redirecting clients or adjusting/replacing client request URLs.
5878 This helper is consulted after the received request is cleared by
5879 http_access and adapted using eICAP/ICAP services (if any). If the
5880 helper does not redirect the client, Squid checks adapted_http_access
5881 and may consult the cache or forward the request to the next hop.
5884 For each request, the helper gets one line in the following format:
5886 [channel-ID <SP>] request-URL [<SP> extras] <NL>
5888 Use url_rewrite_extras to configure what Squid sends as 'extras'.
5891 The helper must reply to each query using a single line:
5893 [channel-ID <SP>] result [<SP> kv-pairs] <NL>
5895 The result section must match exactly one of the following outcomes:
5897 OK [status=30N] url="..."
5899 Redirect the client to a URL supplied in the 'url' parameter.
5900 Optional 'status' specifies the status code to send to the
5901 client in Squid's HTTP redirect response. It must be one of
5902 the standard HTTP redirect status codes: 301, 302, 303, 307,
5903 or 308. When no specific status is requested, Squid uses 302.
5905 OK rewrite-url="..."
5907 Replace the current request URL with the one supplied in the
5908 'rewrite-url' parameter. Squid fetches the resource specified
5909 by the new URL and forwards the received response (or its
5910 cached copy) to the client.
5912 WARNING: Avoid rewriting URLs! When possible, redirect the
5913 client using an "OK url=..." helper response instead.
5914 Rewriting URLs may create inconsistent requests and/or break
5915 synchronization between internal client and origin server
5916 states, especially when URLs or other message parts contain
5917 snippets of that state. For example, Squid does not adjust
5918 Location headers and embedded URLs after the helper rewrites
5922 Keep the client request intact.
5925 Keep the client request intact.
5928 A helper problem that should be reported to the Squid admin
5929 via a level-1 cache.log message. The 'message' parameter is
5930 reserved for specifying the log message.
5932 In addition to the kv-pairs mentioned above, Squid also understands
5933 the following optional kv-pairs in URL rewriter responses:
5936 Associates a TAG with the client TCP connection.
5938 The clt_conn_tag=TAG pair is treated as a regular transaction
5939 annotation for the current request and also annotates future
5940 requests on the same client connection. A helper may update
5941 the TAG during subsequent requests by returning a new kv-pair.
5944 Helper messages contain the channel-ID part if and only if the
5945 url_rewrite_children directive specifies positive concurrency. As a
5946 channel-ID value, Squid sends a number between 0 and concurrency-1.
5947 The helper must echo back the received channel-ID in its response.
5949 By default, Squid does not use a URL rewriter.
5952 NAME: url_rewrite_children redirect_children
5953 TYPE: HelperChildConfig
5954 DEFAULT: 20 startup=0 idle=1 concurrency=0
5955 LOC: Config.redirectChildren
5957 Specifies the maximum number of redirector processes that Squid may
5958 spawn (numberofchildren) and several related options. Using too few of
5959 these helper processes (a.k.a. "helpers") creates request queues.
5960 Using too many helpers wastes your system resources.
5962 Usage: numberofchildren [option]...
5964 The startup= and idle= options allow some measure of skew in your
5969 Sets a minimum of how many processes are to be spawned when Squid
5970 starts or reconfigures. When set to zero the first request will
5971 cause spawning of the first child process to handle it.
5973 Starting too few will cause an initial slowdown in traffic as Squid
5974 attempts to simultaneously spawn enough processes to cope.
5978 Sets a minimum of how many processes Squid is to try and keep available
5979 at all times. When traffic begins to rise above what the existing
5980 processes can handle this many more will be spawned up to the maximum
5981 configured. A minimum setting of 1 is required.
5985 The number of requests each redirector helper can handle in
5986 parallel. Defaults to 0 which indicates the redirector
5987 is a old-style single threaded redirector.
5989 When this directive is set to a value >= 1 then the protocol
5990 used to communicate with the helper is modified to include
5991 an ID in front of the request/response. The ID from the request
5992 must be echoed back with the response to that request.
5996 Sets the maximum number of queued requests. A request is queued when
5997 no existing child can accept it due to concurrency limit and no new
5998 child can be started due to numberofchildren limit. The default
5999 maximum is zero if url_rewrite_bypass is enabled and
6000 2*numberofchildren otherwise. If the queued requests exceed queue size
6001 and redirector_bypass configuration option is set, then redirector is
6002 bypassed. Otherwise, Squid is allowed to temporarily exceed the
6003 configured maximum, marking the affected helper as "overloaded". If
6004 the helper overload lasts more than 3 minutes, the action prescribed
6005 by the on-persistent-overload option applies.
6007 on-persistent-overload=action
6009 Specifies Squid reaction to a new helper request arriving when the helper
6010 has been overloaded for more that 3 minutes already. The number of queued
6011 requests determines whether the helper is overloaded (see the queue-size
6014 Two actions are supported:
6016 die Squid worker quits. This is the default behavior.
6018 ERR Squid treats the helper request as if it was
6019 immediately submitted, and the helper immediately
6020 replied with an ERR response. This action has no effect
6021 on the already queued and in-progress helper requests.
6024 NAME: url_rewrite_host_header redirect_rewrites_host_header
6027 LOC: Config.onoff.redir_rewrites_host
6029 To preserve same-origin security policies in browsers and
6030 prevent Host: header forgery by redirectors Squid rewrites
6031 any Host: header in redirected requests.
6033 If you are running an accelerator this may not be a wanted
6034 effect of a redirector. This directive enables you disable
6035 Host: alteration in reverse-proxy traffic.
6037 WARNING: Entries are cached on the result of the URL rewriting
6038 process, so be careful if you have domain-virtual hosts.
6040 WARNING: Squid and other software verifies the URL and Host
6041 are matching, so be careful not to relay through other proxies
6042 or inspecting firewalls with this disabled.
6045 NAME: url_rewrite_access redirector_access
6048 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
6049 LOC: Config.accessList.redirector
6051 If defined, this access list specifies which requests are
6052 sent to the redirector processes.
6054 This clause supports both fast and slow acl types.
6055 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6058 NAME: url_rewrite_bypass redirector_bypass
6060 LOC: Config.onoff.redirector_bypass
6063 When this is 'on', a request will not go through the
6064 redirector if all the helpers are busy. If this is 'off' and the
6065 redirector queue grows too large, the action is prescribed by the
6066 on-persistent-overload option. You should only enable this if the
6067 redirectors are not critical to your caching system. If you use
6068 redirectors for access control, and you enable this option,
6069 users may have access to pages they should not
6070 be allowed to request.
6072 Enabling this option sets the default url_rewrite_children queue-size
6076 NAME: url_rewrite_extras
6077 TYPE: TokenOrQuotedString
6078 LOC: Config.redirector_extras
6079 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
6081 Specifies a string to be append to request line format for the
6082 rewriter helper. "Quoted" format values may contain spaces and
6083 logformat %macros. In theory, any logformat %macro can be used.
6084 In practice, a %macro expands as a dash (-) if the helper request is
6085 sent before the required macro information is available to Squid.
6088 NAME: url_rewrite_timeout
6089 TYPE: UrlHelperTimeout
6090 LOC: Config.onUrlRewriteTimeout
6092 DEFAULT_DOC: Squid waits for the helper response forever
6094 Squid times active requests to redirector. The timeout value and Squid
6095 reaction to a timed out request are configurable using the following
6098 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
6100 supported timeout actions:
6101 fail Squid return a ERR_GATEWAY_FAILURE error page
6103 bypass Do not re-write the URL
6105 retry Send the lookup to the helper again
6107 use_configured_response
6108 Use the <quoted-response> as helper response
6112 OPTIONS FOR STORE ID
6113 -----------------------------------------------------------------------------
6116 NAME: store_id_program storeurl_rewrite_program
6118 LOC: Config.Program.store_id
6121 Specify the location of the executable StoreID helper to use.
6122 Since they can perform almost any function there isn't one included.
6124 For each requested URL, the helper will receive one line with the format
6126 [channel-ID <SP>] URL [<SP> extras]<NL>
6129 After processing the request the helper must reply using the following format:
6131 [channel-ID <SP>] result [<SP> kv-pairs]
6133 The result code can be:
6136 Use the StoreID supplied in 'store-id='.
6139 The default is to use HTTP request URL as the store ID.
6142 An internal error occurred in the helper, preventing
6143 a result being identified.
6145 In addition to the above kv-pairs Squid also understands the following
6146 optional kv-pairs received from URL rewriters:
6148 Associates a TAG with the client TCP connection.
6149 Please see url_rewrite_program related documentation for this
6152 Helper programs should be prepared to receive and possibly ignore
6153 additional whitespace-separated tokens on each input line.
6155 When using the concurrency= option the protocol is changed by
6156 introducing a query channel tag in front of the request/response.
6157 The query channel tag is a number between 0 and concurrency-1.
6158 This value must be echoed back unchanged to Squid as the first part
6159 of the response relating to its request.
6161 NOTE: when using StoreID refresh_pattern will apply to the StoreID
6162 returned from the helper and not the URL.
6164 WARNING: Wrong StoreID value returned by a careless helper may result
6165 in the wrong cached response returned to the user.
6167 By default, a StoreID helper is not used.
6170 NAME: store_id_extras
6171 TYPE: TokenOrQuotedString
6172 LOC: Config.storeId_extras
6173 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
6175 Specifies a string to be append to request line format for the
6176 StoreId helper. "Quoted" format values may contain spaces and
6177 logformat %macros. In theory, any logformat %macro can be used.
6178 In practice, a %macro expands as a dash (-) if the helper request is
6179 sent before the required macro information is available to Squid.
6182 NAME: store_id_children storeurl_rewrite_children
6183 TYPE: HelperChildConfig
6184 DEFAULT: 20 startup=0 idle=1 concurrency=0
6185 LOC: Config.storeIdChildren
6187 Specifies the maximum number of StoreID helper processes that Squid
6188 may spawn (numberofchildren) and several related options. Using
6189 too few of these helper processes (a.k.a. "helpers") creates request
6190 queues. Using too many helpers wastes your system resources.
6192 Usage: numberofchildren [option]...
6194 The startup= and idle= options allow some measure of skew in your
6199 Sets a minimum of how many processes are to be spawned when Squid
6200 starts or reconfigures. When set to zero the first request will
6201 cause spawning of the first child process to handle it.
6203 Starting too few will cause an initial slowdown in traffic as Squid
6204 attempts to simultaneously spawn enough processes to cope.
6208 Sets a minimum of how many processes Squid is to try and keep available
6209 at all times. When traffic begins to rise above what the existing
6210 processes can handle this many more will be spawned up to the maximum
6211 configured. A minimum setting of 1 is required.
6215 The number of requests each storeID helper can handle in
6216 parallel. Defaults to 0 which indicates the helper
6217 is a old-style single threaded program.
6219 When this directive is set to a value >= 1 then the protocol
6220 used to communicate with the helper is modified to include
6221 an ID in front of the request/response. The ID from the request
6222 must be echoed back with the response to that request.
6226 Sets the maximum number of queued requests to N. A request is queued
6227 when no existing child can accept it due to concurrency limit and no
6228 new child can be started due to numberofchildren limit. The default
6229 maximum is 2*numberofchildren. If the queued requests exceed queue
6230 size and redirector_bypass configuration option is set, then
6231 redirector is bypassed. Otherwise, Squid is allowed to temporarily
6232 exceed the configured maximum, marking the affected helper as
6233 "overloaded". If the helper overload lasts more than 3 minutes, the
6234 action prescribed by the on-persistent-overload option applies.
6236 on-persistent-overload=action
6238 Specifies Squid reaction to a new helper request arriving when the helper
6239 has been overloaded for more that 3 minutes already. The number of queued
6240 requests determines whether the helper is overloaded (see the queue-size
6243 Two actions are supported:
6245 die Squid worker quits. This is the default behavior.
6247 ERR Squid treats the helper request as if it was
6248 immediately submitted, and the helper immediately
6249 replied with an ERR response. This action has no effect
6250 on the already queued and in-progress helper requests.
6253 NAME: store_id_access storeurl_rewrite_access
6256 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
6257 LOC: Config.accessList.store_id
6259 If defined, this access list specifies which requests are
6260 sent to the StoreID processes. By default all requests
6263 This clause supports both fast and slow acl types.
6264 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6267 NAME: store_id_bypass storeurl_rewrite_bypass
6269 LOC: Config.onoff.store_id_bypass
6272 When this is 'on', a request will not go through the
6273 helper if all helpers are busy. If this is 'off' and the helper
6274 queue grows too large, the action is prescribed by the
6275 on-persistent-overload option. You should only enable this if the
6276 helpers are not critical to your caching system. If you use
6277 helpers for critical caching components, and you enable this
6278 option, users may not get objects from cache.
6279 This options sets default queue-size option of the store_id_children
6284 OPTIONS FOR TUNING THE CACHE
6285 -----------------------------------------------------------------------------
6288 NAME: cache no_cache
6291 DEFAULT_DOC: By default, this directive is unused and has no effect.
6292 LOC: Config.accessList.noCache
6294 Requests denied by this directive will not be served from the cache
6295 and their responses will not be stored in the cache. This directive
6296 has no effect on other transactions and on already cached responses.
6298 This clause supports both fast and slow acl types.
6299 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6301 This and the two other similar caching directives listed below are
6302 checked at different transaction processing stages, have different
6303 access to response information, affect different cache operations,
6304 and differ in slow ACLs support:
6306 * cache: Checked before Squid makes a hit/miss determination.
6307 No access to reply information!
6308 Denies both serving a hit and storing a miss.
6309 Supports both fast and slow ACLs.
6310 * send_hit: Checked after a hit was detected.
6311 Has access to reply (hit) information.
6312 Denies serving a hit only.
6313 Supports fast ACLs only.
6314 * store_miss: Checked before storing a cachable miss.
6315 Has access to reply (miss) information.
6316 Denies storing a miss only.
6317 Supports fast ACLs only.
6319 If you are not sure which of the three directives to use, apply the
6320 following decision logic:
6322 * If your ACL(s) are of slow type _and_ need response info, redesign.
6323 Squid does not support that particular combination at this time.
6325 * If your directive ACL(s) are of slow type, use "cache"; and/or
6326 * if your directive ACL(s) need no response info, use "cache".
6328 * If you do not want the response cached, use store_miss; and/or
6329 * if you do not want a hit on a cached response, use send_hit.
6335 DEFAULT_DOC: By default, this directive is unused and has no effect.
6336 LOC: Config.accessList.sendHit
6338 Responses denied by this directive will not be served from the cache
6339 (but may still be cached, see store_miss). This directive has no
6340 effect on the responses it allows and on the cached objects.
6342 Please see the "cache" directive for a summary of differences among
6343 store_miss, send_hit, and cache directives.
6345 Unlike the "cache" directive, send_hit only supports fast acl
6346 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6350 # apply custom Store ID mapping to some URLs
6351 acl MapMe dstdomain .c.example.com
6352 store_id_program ...
6353 store_id_access allow MapMe
6355 # but prevent caching of special responses
6356 # such as 302 redirects that cause StoreID loops
6357 acl Ordinary http_status 200-299
6358 store_miss deny MapMe !Ordinary
6360 # and do not serve any previously stored special responses
6361 # from the cache (in case they were already cached before
6362 # the above store_miss rule was in effect).
6363 send_hit deny MapMe !Ordinary
6369 DEFAULT_DOC: By default, this directive is unused and has no effect.
6370 LOC: Config.accessList.storeMiss
6372 Responses denied by this directive will not be cached (but may still
6373 be served from the cache, see send_hit). This directive has no
6374 effect on the responses it allows and on the already cached responses.
6376 Please see the "cache" directive for a summary of differences among
6377 store_miss, send_hit, and cache directives. See the
6378 send_hit directive for a usage example.
6380 Unlike the "cache" directive, store_miss only supports fast acl
6381 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6387 LOC: Config.maxStale
6390 This option puts an upper limit on how stale content Squid
6391 will serve from the cache if cache validation fails.
6392 Can be overridden by the refresh_pattern max-stale option.
6395 NAME: refresh_pattern
6396 TYPE: refreshpattern
6400 usage: refresh_pattern [-i] regex min percent max [options]
6402 By default, regular expressions are CASE-SENSITIVE. To make
6403 them case-insensitive, use the -i option.
6405 'Min' is the time (in minutes) an object without an explicit
6406 expiry time should be considered fresh. The recommended
6407 value is 0, any higher values may cause dynamic applications
6408 to be erroneously cached unless the application designer
6409 has taken the appropriate actions.
6411 'Percent' is a percentage of the objects age (time since last
6412 modification age) an object without explicit expiry time
6413 will be considered fresh.
6415 'Max' is an upper limit on how long objects without an explicit
6416 expiry time will be considered fresh. The value is also used
6417 to form Cache-Control: max-age header for a request sent from
6418 Squid to origin/parent.
6420 options: override-expire
6430 override-expire enforces min age even if the server
6431 sent an explicit expiry time (e.g., with the
6432 Expires: header or Cache-Control: max-age). Doing this
6433 VIOLATES the HTTP standard. Enabling this feature
6434 could make you liable for problems which it causes.
6436 Note: override-expire does not enforce staleness - it only extends
6437 freshness / min. If the server returns a Expires time which
6438 is longer than your max time, Squid will still consider
6439 the object fresh for that period of time.
6441 override-lastmod enforces min age even on objects
6442 that were modified recently.
6444 reload-into-ims changes a client no-cache or ``reload''
6445 request for a cached entry into a conditional request using
6446 If-Modified-Since and/or If-None-Match headers, provided the
6447 cached entry has a Last-Modified and/or a strong ETag header.
6448 Doing this VIOLATES the HTTP standard. Enabling this feature
6449 could make you liable for problems which it causes.
6451 ignore-reload ignores a client no-cache or ``reload''
6452 header. Doing this VIOLATES the HTTP standard. Enabling
6453 this feature could make you liable for problems which
6456 ignore-no-store ignores any ``Cache-control: no-store''
6457 headers received from a server. Doing this VIOLATES
6458 the HTTP standard. Enabling this feature could make you
6459 liable for problems which it causes.
6461 ignore-private ignores any ``Cache-control: private''
6462 headers received from a server. Doing this VIOLATES
6463 the HTTP standard. Enabling this feature could make you
6464 liable for problems which it causes.
6466 refresh-ims causes squid to contact the origin server
6467 when a client issues an If-Modified-Since request. This
6468 ensures that the client will receive an updated version
6469 if one is available.
6471 store-stale stores responses even if they don't have explicit
6472 freshness or a validator (i.e., Last-Modified or an ETag)
6473 present, or if they're already stale. By default, Squid will
6474 not cache such responses because they usually can't be
6475 reused. Note that such responses will be stale by default.
6477 max-stale=NN provide a maximum staleness factor. Squid won't
6478 serve objects more stale than this even if it failed to
6479 validate the object. Default: use the max_stale global limit.
6481 Basically a cached object is:
6483 FRESH if expire > now, else STALE
6485 FRESH if lm-factor < percent, else STALE
6489 The refresh_pattern lines are checked in the order listed here.
6490 The first entry which matches is used. If none of the entries
6491 match the default will be used.
6493 Note, you must uncomment all the default lines if you want
6494 to change one. The default setting is only active if none is
6500 # Add any of your own refresh_pattern entries above these.
6502 refresh_pattern ^ftp: 1440 20% 10080
6503 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
6504 refresh_pattern . 0 20% 4320
6508 NAME: quick_abort_min
6512 LOC: Config.quickAbort.min
6515 NAME: quick_abort_max
6519 LOC: Config.quickAbort.max
6522 NAME: quick_abort_pct
6526 LOC: Config.quickAbort.pct
6528 The cache by default continues downloading aborted requests
6529 which are almost completed (less than 16 KB remaining). This
6530 may be undesirable on slow (e.g. SLIP) links and/or very busy
6531 caches. Impatient users may tie up file descriptors and
6532 bandwidth by repeatedly requesting and immediately aborting
6535 When the user aborts a request, Squid will check the
6536 quick_abort values to the amount of data transferred until
6539 If the transfer has less than 'quick_abort_min' KB remaining,
6540 it will finish the retrieval.
6542 If the transfer has more than 'quick_abort_max' KB remaining,
6543 it will abort the retrieval.
6545 If more than 'quick_abort_pct' of the transfer has completed,
6546 it will finish the retrieval.
6548 If you do not want any retrieval to continue after the client
6549 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
6552 If you want retrievals to always continue if they are being
6553 cached set 'quick_abort_min' to '-1 KB'.
6556 NAME: read_ahead_gap
6557 COMMENT: buffer-size
6559 LOC: Config.readAheadGap
6562 The amount of data the cache will buffer ahead of what has been
6563 sent to the client when retrieving an object from another server.
6567 IFDEF: USE_HTTP_VIOLATIONS
6570 LOC: Config.negativeTtl
6573 Set the Default Time-to-Live (TTL) for failed requests.
6574 Certain types of failures (such as "connection refused" and
6575 "404 Not Found") are able to be negatively-cached for a short time.
6576 Modern web servers should provide Expires: header, however if they
6577 do not this can provide a minimum TTL.
6578 The default is not to cache errors with unknown expiry details.
6580 Note that this is different from negative caching of DNS lookups.
6582 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6583 this feature could make you liable for problems which it
6587 NAME: positive_dns_ttl
6590 LOC: Config.positiveDnsTtl
6593 Upper limit on how long Squid will cache positive DNS responses.
6594 Default is 6 hours (360 minutes). This directive must be set
6595 larger than negative_dns_ttl.
6598 NAME: negative_dns_ttl
6601 LOC: Config.negativeDnsTtl
6604 Time-to-Live (TTL) for negative caching of failed DNS lookups.
6605 This also sets the lower cache limit on positive lookups.
6606 Minimum value is 1 second, and it is not recommendable to go
6607 much below 10 seconds.
6610 NAME: range_offset_limit
6611 COMMENT: size [acl acl...]
6613 LOC: Config.rangeOffsetLimit
6616 usage: (size) [units] [[!]aclname]
6618 Sets an upper limit on how far (number of bytes) into the file
6619 a Range request may be to cause Squid to prefetch the whole file.
6620 If beyond this limit, Squid forwards the Range request as it is and
6621 the result is NOT cached.
6623 This is to stop a far ahead range request (lets say start at 17MB)
6624 from making Squid fetch the whole object up to that point before
6625 sending anything to the client.
6627 Multiple range_offset_limit lines may be specified, and they will
6628 be searched from top to bottom on each request until a match is found.
6629 The first match found will be used. If no line matches a request, the
6630 default limit of 0 bytes will be used.
6632 'size' is the limit specified as a number of units.
6634 'units' specifies whether to use bytes, KB, MB, etc.
6635 If no units are specified bytes are assumed.
6637 A size of 0 causes Squid to never fetch more than the
6638 client requested. (default)
6640 A size of 'none' causes Squid to always fetch the object from the
6641 beginning so it may cache the result. (2.0 style)
6643 'aclname' is the name of a defined ACL.
6645 NP: Using 'none' as the byte value here will override any quick_abort settings
6646 that may otherwise apply to the range request. The range request will
6647 be fully fetched from start to finish regardless of the client
6648 actions. This affects bandwidth usage.
6651 NAME: minimum_expiry_time
6654 LOC: Config.minimum_expiry_time
6657 The minimum caching time according to (Expires - Date)
6658 headers Squid honors if the object can't be revalidated.
6659 The default is 60 seconds.
6661 In reverse proxy environments it might be desirable to honor
6662 shorter object lifetimes. It is most likely better to make
6663 your server return a meaningful Last-Modified header however.
6665 In ESI environments where page fragments often have short
6666 lifetimes, this will often be best set to 0.
6669 NAME: store_avg_object_size
6673 LOC: Config.Store.avgObjectSize
6675 Average object size, used to estimate number of objects your
6676 cache can hold. The default is 13 KB.
6678 This is used to pre-seed the cache index memory allocation to
6679 reduce expensive reallocate operations while handling clients
6680 traffic. Too-large values may result in memory allocation during
6681 peak traffic, too-small values will result in wasted memory.
6683 Check the cache manager 'info' report metrics for the real
6684 object sizes seen by your Squid before tuning this.
6687 NAME: store_objects_per_bucket
6690 LOC: Config.Store.objectsPerBucket
6692 Target number of objects per bucket in the store hash table.
6693 Lowering this value increases the total number of buckets and
6694 also the storage maintenance rate. The default is 20.
6699 -----------------------------------------------------------------------------
6702 NAME: request_header_max_size
6706 LOC: Config.maxRequestHeaderSize
6708 This specifies the maximum size for HTTP headers in a request.
6709 Request headers are usually relatively small (about 512 bytes).
6710 Placing a limit on the request header size will catch certain
6711 bugs (for example with persistent connections) and possibly
6712 buffer-overflow or denial-of-service attacks.
6715 NAME: reply_header_max_size
6719 LOC: Config.maxReplyHeaderSize
6721 This specifies the maximum size for HTTP headers in a reply.
6722 Reply headers are usually relatively small (about 512 bytes).
6723 Placing a limit on the reply header size will catch certain
6724 bugs (for example with persistent connections) and possibly
6725 buffer-overflow or denial-of-service attacks.
6728 NAME: request_body_max_size
6732 DEFAULT_DOC: No limit.
6733 LOC: Config.maxRequestBodySize
6735 This specifies the maximum size for an HTTP request body.
6736 In other words, the maximum size of a PUT/POST request.
6737 A user who attempts to send a request with a body larger
6738 than this limit receives an "Invalid Request" error message.
6739 If you set this parameter to a zero (the default), there will
6740 be no limit imposed.
6742 See also client_request_buffer_max_size for an alternative
6743 limitation on client uploads which can be configured.
6746 NAME: client_request_buffer_max_size
6750 LOC: Config.maxRequestBufferSize
6752 This specifies the maximum buffer size of a client request.
6753 It prevents squid eating too much memory when somebody uploads
6758 IFDEF: USE_HTTP_VIOLATIONS
6761 DEFAULT_DOC: Obey RFC 2616.
6762 LOC: Config.accessList.brokenPosts
6764 A list of ACL elements which, if matched, causes Squid to send
6765 an extra CRLF pair after the body of a PUT/POST request.
6767 Some HTTP servers has broken implementations of PUT/POST,
6768 and rely on an extra CRLF pair sent by some WWW clients.
6770 Quote from RFC2616 section 4.1 on this matter:
6772 Note: certain buggy HTTP/1.0 client implementations generate an
6773 extra CRLF's after a POST request. To restate what is explicitly
6774 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
6775 a request with an extra CRLF.
6777 This clause only supports fast acl types.
6778 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6781 acl buggy_server url_regex ^http://....
6782 broken_posts allow buggy_server
6785 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
6788 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
6790 LOC: Adaptation::Config::use_indirect_client
6792 Controls whether the indirect client IP address (instead of the direct
6793 client IP address) is passed to adaptation services.
6795 See also: follow_x_forwarded_for adaptation_send_client_ip
6799 IFDEF: USE_HTTP_VIOLATIONS
6803 LOC: Config.onoff.via
6805 If set (default), Squid will include a Via header in requests and
6806 replies as required by RFC2616.
6809 NAME: vary_ignore_expire
6812 LOC: Config.onoff.vary_ignore_expire
6815 Many HTTP servers supporting Vary gives such objects
6816 immediate expiry time with no cache-control header
6817 when requested by a HTTP/1.0 client. This option
6818 enables Squid to ignore such expiry times until
6819 HTTP/1.1 is fully implemented.
6821 WARNING: If turned on this may eventually cause some
6822 varying objects not intended for caching to get cached.
6825 NAME: request_header_access
6826 IFDEF: USE_HTTP_VIOLATIONS
6827 TYPE: http_header_access
6828 LOC: Config.request_header_access
6830 DEFAULT_DOC: No limits.
6832 Usage: request_header_access header_name allow|deny [!]aclname ...
6834 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6835 this feature could make you liable for problems which it
6838 This option replaces the old 'anonymize_headers' and the
6839 older 'http_anonymizer' option with something that is much
6840 more configurable. A list of ACLs for each header name allows
6841 removal of specific header fields under specific conditions.
6843 This option only applies to outgoing HTTP request headers (i.e.,
6844 headers sent by Squid to the next HTTP hop such as a cache peer
6845 or an origin server). The option has no effect during cache hit
6846 detection. The equivalent adaptation vectoring point in ICAP
6847 terminology is post-cache REQMOD.
6849 The option is applied to individual outgoing request header
6850 fields. For each request header field F, Squid uses the first
6851 qualifying sets of request_header_access rules:
6853 1. Rules with header_name equal to F's name.
6854 2. Rules with header_name 'Other', provided F's name is not
6855 on the hard-coded list of commonly used HTTP header names.
6856 3. Rules with header_name 'All'.
6858 Within that qualifying rule set, rule ACLs are checked as usual.
6859 If ACLs of an "allow" rule match, the header field is allowed to
6860 go through as is. If ACLs of a "deny" rule match, the header is
6861 removed and request_header_replace is then checked to identify
6862 if the removed header has a replacement. If no rules within the
6863 set have matching ACLs, the header field is left as is.
6865 For example, to achieve the same behavior as the old
6866 'http_anonymizer standard' option, you should use:
6868 request_header_access From deny all
6869 request_header_access Referer deny all
6870 request_header_access User-Agent deny all
6872 Or, to reproduce the old 'http_anonymizer paranoid' feature
6875 request_header_access Authorization allow all
6876 request_header_access Proxy-Authorization allow all
6877 request_header_access Cache-Control allow all
6878 request_header_access Content-Length allow all
6879 request_header_access Content-Type allow all
6880 request_header_access Date allow all
6881 request_header_access Host allow all
6882 request_header_access If-Modified-Since allow all
6883 request_header_access Pragma allow all
6884 request_header_access Accept allow all
6885 request_header_access Accept-Charset allow all
6886 request_header_access Accept-Encoding allow all
6887 request_header_access Accept-Language allow all
6888 request_header_access Connection allow all
6889 request_header_access All deny all
6891 HTTP reply headers are controlled with the reply_header_access directive.
6893 By default, all headers are allowed (no anonymizing is performed).
6896 NAME: reply_header_access
6897 IFDEF: USE_HTTP_VIOLATIONS
6898 TYPE: http_header_access
6899 LOC: Config.reply_header_access
6901 DEFAULT_DOC: No limits.
6903 Usage: reply_header_access header_name allow|deny [!]aclname ...
6905 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6906 this feature could make you liable for problems which it
6909 This option only applies to reply headers, i.e., from the
6910 server to the client.
6912 This is the same as request_header_access, but in the other
6913 direction. Please see request_header_access for detailed
6916 For example, to achieve the same behavior as the old
6917 'http_anonymizer standard' option, you should use:
6919 reply_header_access Server deny all
6920 reply_header_access WWW-Authenticate deny all
6921 reply_header_access Link deny all
6923 Or, to reproduce the old 'http_anonymizer paranoid' feature
6926 reply_header_access Allow allow all
6927 reply_header_access WWW-Authenticate allow all
6928 reply_header_access Proxy-Authenticate allow all
6929 reply_header_access Cache-Control allow all
6930 reply_header_access Content-Encoding allow all
6931 reply_header_access Content-Length allow all
6932 reply_header_access Content-Type allow all
6933 reply_header_access Date allow all
6934 reply_header_access Expires allow all
6935 reply_header_access Last-Modified allow all
6936 reply_header_access Location allow all
6937 reply_header_access Pragma allow all
6938 reply_header_access Content-Language allow all
6939 reply_header_access Retry-After allow all
6940 reply_header_access Title allow all
6941 reply_header_access Content-Disposition allow all
6942 reply_header_access Connection allow all
6943 reply_header_access All deny all
6945 HTTP request headers are controlled with the request_header_access directive.
6947 By default, all headers are allowed (no anonymizing is
6951 NAME: request_header_replace header_replace
6952 IFDEF: USE_HTTP_VIOLATIONS
6953 TYPE: http_header_replace
6954 LOC: Config.request_header_access
6957 Usage: request_header_replace header_name message
6958 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
6960 This option allows you to change the contents of headers
6961 denied with request_header_access above, by replacing them
6962 with some fixed string.
6964 This only applies to request headers, not reply headers.
6966 By default, headers are removed if denied.
6969 NAME: reply_header_replace
6970 IFDEF: USE_HTTP_VIOLATIONS
6971 TYPE: http_header_replace
6972 LOC: Config.reply_header_access
6975 Usage: reply_header_replace header_name message
6976 Example: reply_header_replace Server Foo/1.0
6978 This option allows you to change the contents of headers
6979 denied with reply_header_access above, by replacing them
6980 with some fixed string.
6982 This only applies to reply headers, not request headers.
6984 By default, headers are removed if denied.
6987 NAME: request_header_add
6988 TYPE: HeaderWithAclList
6989 LOC: Config.request_header_add
6992 Usage: request_header_add field-name field-value [ acl ... ]
6993 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6995 This option adds header fields to outgoing HTTP requests (i.e.,
6996 request headers sent by Squid to the next HTTP hop such as a
6997 cache peer or an origin server). The option has no effect during
6998 cache hit detection. The equivalent adaptation vectoring point
6999 in ICAP terminology is post-cache REQMOD.
7001 Field-name is a token specifying an HTTP header name. If a
7002 standard HTTP header name is used, Squid does not check whether
7003 the new header conflicts with any existing headers or violates
7004 HTTP rules. If the request to be modified already contains a
7005 field with the same name, the old field is preserved but the
7006 header field values are not merged.
7008 Field-value is either a token or a quoted string. If quoted
7009 string format is used, then the surrounding quotes are removed
7010 while escape sequences and %macros are processed.
7012 One or more Squid ACLs may be specified to restrict header
7013 injection to matching requests. As always in squid.conf, all
7014 ACLs in the ACL list must be satisfied for the insertion to
7015 happen. The request_header_add supports fast ACLs only.
7017 See also: reply_header_add.
7020 NAME: reply_header_add
7021 TYPE: HeaderWithAclList
7022 LOC: Config.reply_header_add
7025 Usage: reply_header_add field-name field-value [ acl ... ]
7026 Example: reply_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
7028 This option adds header fields to outgoing HTTP responses (i.e., response
7029 headers delivered by Squid to the client). This option has no effect on
7030 cache hit detection. The equivalent adaptation vectoring point in
7031 ICAP terminology is post-cache RESPMOD. This option does not apply to
7032 successful CONNECT replies.
7034 Field-name is a token specifying an HTTP header name. If a
7035 standard HTTP header name is used, Squid does not check whether
7036 the new header conflicts with any existing headers or violates
7037 HTTP rules. If the response to be modified already contains a
7038 field with the same name, the old field is preserved but the
7039 header field values are not merged.
7041 Field-value is either a token or a quoted string. If quoted
7042 string format is used, then the surrounding quotes are removed
7043 while escape sequences and %macros are processed.
7045 One or more Squid ACLs may be specified to restrict header
7046 injection to matching responses. As always in squid.conf, all
7047 ACLs in the ACL list must be satisfied for the insertion to
7048 happen. The reply_header_add option supports fast ACLs only.
7050 See also: request_header_add.
7058 This option used to log custom information about the master
7059 transaction. For example, an admin may configure Squid to log
7060 which "user group" the transaction belongs to, where "user group"
7061 will be determined based on a set of ACLs and not [just]
7062 authentication information.
7063 Values of key/value pairs can be logged using %{key}note macros:
7065 note key value acl ...
7066 logformat myFormat ... %{key}note ...
7068 This clause only supports fast acl types.
7069 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7072 NAME: relaxed_header_parser
7073 COMMENT: on|off|warn
7075 LOC: Config.onoff.relaxed_header_parser
7078 In the default "on" setting Squid accepts certain forms
7079 of non-compliant HTTP messages where it is unambiguous
7080 what the sending application intended even if the message
7081 is not correctly formatted. The messages is then normalized
7082 to the correct form when forwarded by Squid.
7084 If set to "warn" then a warning will be emitted in cache.log
7085 each time such HTTP error is encountered.
7087 If set to "off" then such HTTP errors will cause the request
7088 or response to be rejected.
7091 NAME: collapsed_forwarding
7094 LOC: Config.onoff.collapsed_forwarding
7097 This option controls whether Squid is allowed to merge multiple
7098 potentially cachable requests for the same URI before Squid knows
7099 whether the response is going to be cachable.
7101 When enabled, instead of forwarding each concurrent request for
7102 the same URL, Squid just sends the first of them. The other, so
7103 called "collapsed" requests, wait for the response to the first
7104 request and, if it happens to be cachable, use that response.
7105 Here, "concurrent requests" means "received after the first
7106 request headers were parsed and before the corresponding response
7107 headers were parsed".
7109 This feature is disabled by default: enabling collapsed
7110 forwarding needlessly delays forwarding requests that look
7111 cachable (when they are collapsed) but then need to be forwarded
7112 individually anyway because they end up being for uncachable
7113 content. However, in some cases, such as acceleration of highly
7114 cachable content with periodic or grouped expiration times, the
7115 gains from collapsing [large volumes of simultaneous refresh
7116 requests] outweigh losses from such delays.
7118 Squid collapses two kinds of requests: regular client requests
7119 received on one of the listening ports and internal "cache
7120 revalidation" requests which are triggered by those regular
7121 requests hitting a stale cached object. Revalidation collapsing
7122 is currently disabled for Squid instances containing SMP-aware
7123 disk or memory caches and for Vary-controlled cached objects.
7126 NAME: collapsed_forwarding_access
7129 DEFAULT_DOC: Requests may be collapsed if collapsed_forwarding is on.
7130 LOC: Config.accessList.collapsedForwardingAccess
7132 Use this directive to restrict collapsed forwarding to a subset of
7133 eligible requests. The directive is checked for regular HTTP
7134 requests, internal revalidation requests, and HTCP/ICP requests.
7136 collapsed_forwarding_access allow|deny [!]aclname ...
7138 This directive cannot force collapsing. It has no effect on
7139 collapsing unless collapsed_forwarding is 'on', and all other
7140 collapsing preconditions are satisfied.
7142 * A denied request will not collapse, and future transactions will
7143 not collapse on it (even if they are allowed to collapse).
7145 * An allowed request may collapse, or future transactions may
7146 collapse on it (provided they are allowed to collapse).
7148 This directive is evaluated before receiving HTTP response headers
7149 and without access to Squid-to-peer connection (if any).
7151 Only fast ACLs are supported.
7153 See also: collapsed_forwarding.
7156 NAME: shared_transient_entries_limit collapsed_forwarding_shared_entries_limit
7157 COMMENT: (number of entries)
7159 LOC: Config.shared_transient_entries_limit
7162 This directive limits the size of a table used for sharing current
7163 transaction information among SMP workers. A table entry stores meta
7164 information about a single cache entry being delivered to Squid
7165 client(s) by one or more SMP workers. A single table entry consumes
7166 less than 128 shared memory bytes.
7168 The limit should be significantly larger than the number of
7169 concurrent non-collapsed cachable responses leaving Squid. For a
7170 cache that handles less than 5000 concurrent requests, the default
7171 setting of 16384 should be plenty.
7173 Using excessively large values wastes shared memory. Limiting the
7174 table size too much results in hash collisions, leading to lower hit
7175 ratio and missed SMP request collapsing opportunities: Transactions
7176 left without a table entry cannot cache their responses and are
7177 invisible to other concurrent requests for the same resource.
7179 A zero limit is allowed but unsupported. A positive small limit
7180 lowers hit ratio, but zero limit disables a lot of essential
7181 synchronization among SMP workers, leading to HTTP violations (e.g.,
7182 stale hit responses). It also disables shared collapsed forwarding:
7183 A worker becomes unable to collapse its requests on transactions in
7184 other workers, resulting in more trips to the origin server and more
7190 -----------------------------------------------------------------------------
7193 NAME: forward_timeout
7196 LOC: Config.Timeout.forward
7199 This parameter specifies how long Squid should at most attempt in
7200 finding a forwarding path for the request before giving up.
7203 NAME: connect_timeout
7206 LOC: Config.Timeout.connect
7209 This parameter specifies how long to wait for the TCP connect to
7210 the requested server or peer to complete before Squid should
7211 attempt to find another path where to forward the request.
7214 NAME: peer_connect_timeout
7217 LOC: Config.Timeout.peer_connect
7220 This parameter specifies how long to wait for a pending TCP
7221 connection to a peer cache. The default is 30 seconds. You
7222 may also set different timeout values for individual neighbors
7223 with the 'connect-timeout' option on a 'cache_peer' line.
7229 LOC: Config.Timeout.read
7232 Applied on peer server connections.
7234 After each successful read(), the timeout will be extended by this
7235 amount. If no data is read again after this amount of time,
7236 the request is aborted and logged with ERR_READ_TIMEOUT.
7238 The default is 15 minutes.
7244 LOC: Config.Timeout.write
7247 This timeout is tracked for all connections that have data
7248 available for writing and are waiting for the socket to become
7249 ready. After each successful write, the timeout is extended by
7250 the configured amount. If Squid has data to write but the
7251 connection is not ready for the configured duration, the
7252 transaction associated with the connection is terminated. The
7253 default is 15 minutes.
7256 NAME: request_timeout
7258 LOC: Config.Timeout.request
7261 How long to wait for complete HTTP request headers after initial
7262 connection establishment.
7265 NAME: request_start_timeout
7267 LOC: Config.Timeout.request_start_timeout
7270 How long to wait for the first request byte after initial
7271 connection establishment.
7274 NAME: client_idle_pconn_timeout persistent_request_timeout
7276 LOC: Config.Timeout.clientIdlePconn
7279 How long to wait for the next HTTP request on a persistent
7280 client connection after the previous request completes.
7283 NAME: ftp_client_idle_timeout
7285 LOC: Config.Timeout.ftpClientIdle
7288 How long to wait for an FTP request on a connection to Squid ftp_port.
7289 Many FTP clients do not deal with idle connection closures well,
7290 necessitating a longer default timeout than client_idle_pconn_timeout
7291 used for incoming HTTP requests.
7294 NAME: client_lifetime
7297 LOC: Config.Timeout.lifetime
7300 The maximum amount of time a client (browser) is allowed to
7301 remain connected to the cache process. This protects the Cache
7302 from having a lot of sockets (and hence file descriptors) tied up
7303 in a CLOSE_WAIT state from remote clients that go away without
7304 properly shutting down (either because of a network failure or
7305 because of a poor client implementation). The default is one
7308 NOTE: The default value is intended to be much larger than any
7309 client would ever need to be connected to your cache. You
7310 should probably change client_lifetime only as a last resort.
7311 If you seem to have many client connections tying up
7312 filedescriptors, we recommend first tuning the read_timeout,
7313 request_timeout, persistent_request_timeout and quick_abort values.
7316 NAME: pconn_lifetime
7319 LOC: Config.Timeout.pconnLifetime
7322 Desired maximum lifetime of a persistent connection.
7323 When set, Squid will close a now-idle persistent connection that
7324 exceeded configured lifetime instead of moving the connection into
7325 the idle connection pool (or equivalent). No effect on ongoing/active
7326 transactions. Connection lifetime is the time period from the
7327 connection acceptance or opening time until "now".
7329 This limit is useful in environments with long-lived connections
7330 where Squid configuration or environmental factors change during a
7331 single connection lifetime. If unrestricted, some connections may
7332 last for hours and even days, ignoring those changes that should
7333 have affected their behavior or their existence.
7335 Currently, a new lifetime value supplied via Squid reconfiguration
7336 has no effect on already idle connections unless they become busy.
7338 When set to '0' this limit is not used.
7341 NAME: half_closed_clients
7343 LOC: Config.onoff.half_closed_clients
7346 Some clients may shutdown the sending side of their TCP
7347 connections, while leaving their receiving sides open. Sometimes,
7348 Squid can not tell the difference between a half-closed and a
7349 fully-closed TCP connection.
7351 By default, Squid will immediately close client connections when
7352 read(2) returns "no more data to read."
7354 Change this option to 'on' and Squid will keep open connections
7355 until a read(2) or write(2) on the socket returns an error.
7356 This may show some benefits for reverse proxies. But if not
7357 it is recommended to leave OFF.
7360 NAME: server_idle_pconn_timeout pconn_timeout
7362 LOC: Config.Timeout.serverIdlePconn
7365 Timeout for idle persistent connections to servers and other
7372 LOC: Ident::TheConfig.timeout
7375 Maximum time to wait for IDENT lookups to complete.
7377 If this is too high, and you enabled IDENT lookups from untrusted
7378 users, you might be susceptible to denial-of-service by having
7379 many ident requests going at once.
7382 NAME: shutdown_lifetime
7385 LOC: Config.shutdownLifetime
7388 When SIGTERM or SIGHUP is received, the cache is put into
7389 "shutdown pending" mode until all active sockets are closed.
7390 This value is the lifetime to set for all open descriptors
7391 during shutdown mode. Any active clients after this many
7392 seconds will receive a 'timeout' message.
7396 ADMINISTRATIVE PARAMETERS
7397 -----------------------------------------------------------------------------
7403 LOC: Config.adminEmail
7405 Email-address of local cache manager who will receive
7406 mail if the cache dies. The default is "webmaster".
7412 LOC: Config.EmailFrom
7414 From: email-address for mail sent when the cache dies.
7415 The default is to use 'squid@unique_hostname'.
7417 See also: unique_hostname directive.
7423 LOC: Config.EmailProgram
7425 Email program used to send mail if the cache dies.
7426 The default is "mail". The specified program must comply
7427 with the standard Unix mail syntax:
7428 mail-program recipient < mailfile
7430 Optional command line options can be specified.
7433 NAME: cache_effective_user
7435 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
7436 LOC: Config.effectiveUser
7438 If you start Squid as root, it will change its effective/real
7439 UID/GID to the user specified below. The default is to change
7440 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
7441 see also; cache_effective_group
7444 NAME: cache_effective_group
7447 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
7448 LOC: Config.effectiveGroup
7450 Squid sets the GID to the effective user's default group ID
7451 (taken from the password file) and supplementary group list
7452 from the groups membership.
7454 If you want Squid to run with a specific GID regardless of
7455 the group memberships of the effective user then set this
7456 to the group (or GID) you want Squid to run as. When set
7457 all other group privileges of the effective user are ignored
7458 and only this GID is effective. If Squid is not started as
7459 root the user starting Squid MUST be member of the specified
7462 This option is not recommended by the Squid Team.
7463 Our preference is for administrators to configure a secure
7464 user account for squid with UID/GID matching system policies.
7467 NAME: httpd_suppress_version_string
7471 LOC: Config.onoff.httpd_suppress_version_string
7473 Do not send Squid version string in HTTP metadata and generated content
7474 such as HTML error pages. Squid version string is still present in certain
7475 SNMP responses, cachemgr.cgi output, squidclient User-Agent request header
7476 field, various console output, and cache.log.
7479 NAME: visible_hostname
7481 LOC: Config.visibleHostname
7483 DEFAULT_DOC: Automatically detect the system host name
7485 If you want to present a special hostname in error messages, etc,
7486 define this. Otherwise, the return value of gethostname()
7487 will be used. If you have multiple caches in a cluster and
7488 get errors about IP-forwarding you must set them to have individual
7489 names with this setting.
7492 NAME: unique_hostname
7494 LOC: Config.uniqueHostname
7496 DEFAULT_DOC: Copy the value from visible_hostname
7498 If you want to have multiple machines with the same
7499 'visible_hostname' you must give each machine a different
7500 'unique_hostname' so forwarding loops can be detected.
7503 NAME: hostname_aliases
7505 LOC: Config.hostnameAliases
7508 A list of other DNS names your cache has.
7516 Minimum umask which should be enforced while the proxy
7517 is running, in addition to the umask set at startup.
7519 For a traditional octal representation of umasks, start
7524 HTTPD-ACCELERATOR OPTIONS
7525 -----------------------------------------------------------------------------
7528 NAME: httpd_accel_surrogate_id
7531 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
7532 LOC: Config.Accel.surrogate_id
7534 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
7535 need an identification token to allow control targeting. Because
7536 a farm of surrogates may all perform the same tasks, they may share
7537 an identification token.
7539 When the surrogate is a reverse-proxy, this ID is also
7540 used as cdn-id for CDN-Loop detection (RFC 8586).
7543 NAME: http_accel_surrogate_remote
7547 LOC: Config.onoff.surrogate_is_remote
7549 Remote surrogates (such as those in a CDN) honour the header
7550 "Surrogate-Control: no-store-remote".
7552 Set this to on to have squid behave as a remote surrogate.
7556 IFDEF: USE_SQUID_ESI
7557 COMMENT: libxml2|expat
7559 LOC: ESIParser::Type
7561 DEFAULT_DOC: Selects libxml2 if available at ./configure time or libexpat otherwise.
7563 Selects the XML parsing library to use when interpreting responses with
7566 To disable ESI handling completely, ./configure Squid with --disable-esi.
7570 DELAY POOL PARAMETERS
7571 -----------------------------------------------------------------------------
7575 TYPE: delay_pool_count
7577 IFDEF: USE_DELAY_POOLS
7580 This represents the number of delay pools to be used. For example,
7581 if you have one class 2 delay pool and one class 3 delays pool, you
7582 have a total of 2 delay pools.
7584 See also delay_parameters, delay_class, delay_access for pool
7585 configuration details.
7589 TYPE: delay_pool_class
7591 IFDEF: USE_DELAY_POOLS
7594 This defines the class of each delay pool. There must be exactly one
7595 delay_class line for each delay pool. For example, to define two
7596 delay pools, one of class 2 and one of class 3, the settings above
7600 delay_pools 4 # 4 delay pools
7601 delay_class 1 2 # pool 1 is a class 2 pool
7602 delay_class 2 3 # pool 2 is a class 3 pool
7603 delay_class 3 4 # pool 3 is a class 4 pool
7604 delay_class 4 5 # pool 4 is a class 5 pool
7606 The delay pool classes are:
7608 class 1 Everything is limited by a single aggregate
7611 class 2 Everything is limited by a single aggregate
7612 bucket as well as an "individual" bucket chosen
7613 from bits 25 through 32 of the IPv4 address.
7615 class 3 Everything is limited by a single aggregate
7616 bucket as well as a "network" bucket chosen
7617 from bits 17 through 24 of the IP address and a
7618 "individual" bucket chosen from bits 17 through
7619 32 of the IPv4 address.
7621 class 4 Everything in a class 3 delay pool, with an
7622 additional limit on a per user basis. This
7623 only takes effect if the username is established
7624 in advance - by forcing authentication in your
7627 class 5 Requests are grouped according their tag (see
7628 external_acl's tag= reply).
7631 Each pool also requires a delay_parameters directive to configure the pool size
7632 and speed limits used whenever the pool is applied to a request. Along with
7633 a set of delay_access directives to determine when it is used.
7635 NOTE: If an IP address is a.b.c.d
7636 -> bits 25 through 32 are "d"
7637 -> bits 17 through 24 are "c"
7638 -> bits 17 through 32 are "c * 256 + d"
7640 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
7641 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
7643 This clause only supports fast acl types.
7644 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7646 See also delay_parameters and delay_access.
7650 TYPE: delay_pool_access
7652 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
7653 IFDEF: USE_DELAY_POOLS
7656 This is used to determine which delay pool a request falls into.
7658 delay_access is sorted per pool and the matching starts with pool 1,
7659 then pool 2, ..., and finally pool N. The first delay pool where the
7660 request is allowed is selected for the request. If it does not allow
7661 the request to any pool then the request is not delayed (default).
7663 For example, if you want some_big_clients in delay
7664 pool 1 and lotsa_little_clients in delay pool 2:
7666 delay_access 1 allow some_big_clients
7667 delay_access 1 deny all
7668 delay_access 2 allow lotsa_little_clients
7669 delay_access 2 deny all
7670 delay_access 3 allow authenticated_clients
7672 See also delay_parameters and delay_class.
7676 NAME: delay_parameters
7677 TYPE: delay_pool_rates
7679 IFDEF: USE_DELAY_POOLS
7682 This defines the parameters for a delay pool. Each delay pool has
7683 a number of "buckets" associated with it, as explained in the
7684 description of delay_class.
7686 For a class 1 delay pool, the syntax is:
7688 delay_parameters pool aggregate
7690 For a class 2 delay pool:
7692 delay_parameters pool aggregate individual
7694 For a class 3 delay pool:
7696 delay_parameters pool aggregate network individual
7698 For a class 4 delay pool:
7700 delay_parameters pool aggregate network individual user
7702 For a class 5 delay pool:
7704 delay_parameters pool tagrate
7706 The option variables are:
7708 pool a pool number - ie, a number between 1 and the
7709 number specified in delay_pools as used in
7712 aggregate the speed limit parameters for the aggregate bucket
7715 individual the speed limit parameters for the individual
7716 buckets (class 2, 3).
7718 network the speed limit parameters for the network buckets
7721 user the speed limit parameters for the user buckets
7724 tagrate the speed limit parameters for the tag buckets
7727 A pair of delay parameters is written restore/maximum, where restore is
7728 the number of bytes (not bits - modem and network speeds are usually
7729 quoted in bits) per second placed into the bucket, and maximum is the
7730 maximum number of bytes which can be in the bucket at any time.
7732 There must be one delay_parameters line for each delay pool.
7735 For example, if delay pool number 1 is a class 2 delay pool as in the
7736 above example, and is being used to strictly limit each host to 64Kbit/sec
7737 (plus overheads), with no overall limit, the line is:
7739 delay_parameters 1 none 8000/8000
7741 Note that 8 x 8K Byte/sec -> 64K bit/sec.
7743 Note that the word 'none' is used to represent no limit.
7746 And, if delay pool number 2 is a class 3 delay pool as in the above
7747 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
7748 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
7749 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
7750 to permit a decent web page to be downloaded at a decent speed
7751 (if the network is not being limited due to overuse) but slow down
7752 large downloads more significantly:
7754 delay_parameters 2 32000/32000 8000/8000 600/8000
7756 Note that 8 x 32K Byte/sec -> 256K bit/sec.
7757 8 x 8K Byte/sec -> 64K bit/sec.
7758 8 x 600 Byte/sec -> 4800 bit/sec.
7761 Finally, for a class 4 delay pool as in the example - each user will
7762 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
7764 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
7767 See also delay_class and delay_access.
7771 NAME: delay_initial_bucket_level
7772 COMMENT: (percent, 0-100)
7775 IFDEF: USE_DELAY_POOLS
7776 LOC: Config.Delay.initial
7778 The initial bucket percentage is used to determine how much is put
7779 in each bucket when squid starts, is reconfigured, or first notices
7780 a host accessing it (in class 2 and class 3, individual hosts and
7781 networks only have buckets associated with them once they have been
7786 CLIENT DELAY POOL PARAMETERS
7787 -----------------------------------------------------------------------------
7790 NAME: client_delay_pools
7791 TYPE: client_delay_pool_count
7793 IFDEF: USE_DELAY_POOLS
7794 LOC: Config.ClientDelay
7796 This option specifies the number of client delay pools used. It must
7797 preceed other client_delay_* options.
7800 client_delay_pools 2
7802 See also client_delay_parameters and client_delay_access.
7805 NAME: client_delay_initial_bucket_level
7806 COMMENT: (percent, 0-no_limit)
7809 IFDEF: USE_DELAY_POOLS
7810 LOC: Config.ClientDelay.initial
7812 This option determines the initial bucket size as a percentage of
7813 max_bucket_size from client_delay_parameters. Buckets are created
7814 at the time of the "first" connection from the matching IP. Idle
7815 buckets are periodically deleted up.
7817 You can specify more than 100 percent but note that such "oversized"
7818 buckets are not refilled until their size goes down to max_bucket_size
7819 from client_delay_parameters.
7822 client_delay_initial_bucket_level 50
7825 NAME: client_delay_parameters
7826 TYPE: client_delay_pool_rates
7828 IFDEF: USE_DELAY_POOLS
7829 LOC: Config.ClientDelay
7832 This option configures client-side bandwidth limits using the
7835 client_delay_parameters pool speed_limit max_bucket_size
7837 pool is an integer ID used for client_delay_access matching.
7839 speed_limit is bytes added to the bucket per second.
7841 max_bucket_size is the maximum size of a bucket, enforced after any
7842 speed_limit additions.
7844 Please see the delay_parameters option for more information and
7848 client_delay_parameters 1 1024 2048
7849 client_delay_parameters 2 51200 16384
7851 See also client_delay_access.
7855 NAME: client_delay_access
7856 TYPE: client_delay_pool_access
7858 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
7859 IFDEF: USE_DELAY_POOLS
7860 LOC: Config.ClientDelay
7862 This option determines the client-side delay pool for the
7865 client_delay_access pool_ID allow|deny acl_name
7867 All client_delay_access options are checked in their pool ID
7868 order, starting with pool 1. The first checked pool with allowed
7869 request is selected for the request. If no ACL matches or there
7870 are no client_delay_access options, the request bandwidth is not
7873 The ACL-selected pool is then used to find the
7874 client_delay_parameters for the request. Client-side pools are
7875 not used to aggregate clients. Clients are always aggregated
7876 based on their source IP addresses (one bucket per source IP).
7878 This clause only supports fast acl types.
7879 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7880 Additionally, only the client TCP connection details are available.
7881 ACLs testing HTTP properties will not work.
7883 Please see delay_access for more examples.
7886 client_delay_access 1 allow low_rate_network
7887 client_delay_access 2 allow vips_network
7890 See also client_delay_parameters and client_delay_pools.
7893 NAME: response_delay_pool
7894 TYPE: response_delay_pool_parameters
7896 IFDEF: USE_DELAY_POOLS
7897 LOC: Config.MessageDelay
7899 This option configures client response bandwidth limits using the
7902 response_delay_pool name [option=value] ...
7904 name the response delay pool name
7908 individual-restore The speed limit of an individual
7909 bucket(bytes/s). To be used in conjunction
7910 with 'individual-maximum'.
7912 individual-maximum The maximum number of bytes which can
7913 be placed into the individual bucket. To be used
7914 in conjunction with 'individual-restore'.
7916 aggregate-restore The speed limit for the aggregate
7917 bucket(bytes/s). To be used in conjunction with
7918 'aggregate-maximum'.
7920 aggregate-maximum The maximum number of bytes which can
7921 be placed into the aggregate bucket. To be used
7922 in conjunction with 'aggregate-restore'.
7924 initial-bucket-level The initial bucket size as a percentage
7925 of individual-maximum.
7927 Individual and(or) aggregate bucket options may not be specified,
7928 meaning no individual and(or) aggregate speed limitation.
7929 See also response_delay_pool_access and delay_parameters for
7930 terminology details.
7933 NAME: response_delay_pool_access
7934 TYPE: response_delay_pool_access
7936 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
7937 IFDEF: USE_DELAY_POOLS
7938 LOC: Config.MessageDelay
7940 Determines whether a specific named response delay pool is used
7941 for the transaction. The syntax for this directive is:
7943 response_delay_pool_access pool_name allow|deny acl_name
7945 All response_delay_pool_access options are checked in the order
7946 they appear in this configuration file. The first rule with a
7947 matching ACL wins. If (and only if) an "allow" rule won, Squid
7948 assigns the response to the corresponding named delay pool.
7952 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
7953 -----------------------------------------------------------------------------
7958 LOC: Config.Wccp.router
7960 DEFAULT_DOC: WCCP disabled.
7963 Use this option to define your WCCP ``home'' router for
7966 wccp_router supports a single WCCP(v1) router
7968 wccp2_router supports multiple WCCPv2 routers
7970 only one of the two may be used at the same time and defines
7971 which version of WCCP to use.
7975 TYPE: IpAddress_list
7976 LOC: Config.Wccp2.router
7978 DEFAULT_DOC: WCCPv2 disabled.
7981 Use this option to define your WCCP ``home'' router for
7984 wccp_router supports a single WCCP(v1) router
7986 wccp2_router supports multiple WCCPv2 routers
7988 only one of the two may be used at the same time and defines
7989 which version of WCCP to use.
7994 LOC: Config.Wccp.version
7998 This directive is only relevant if you need to set up WCCP(v1)
7999 to some very old and end-of-life Cisco routers. In all other
8000 setups it must be left unset or at the default setting.
8001 It defines an internal version in the WCCP(v1) protocol,
8002 with version 4 being the officially documented protocol.
8004 According to some users, Cisco IOS 11.2 and earlier only
8005 support WCCP version 3. If you're using that or an earlier
8006 version of IOS, you may need to change this value to 3, otherwise
8007 do not specify this parameter.
8010 NAME: wccp2_rebuild_wait
8012 LOC: Config.Wccp2.rebuildwait
8016 If this is enabled Squid will wait for the cache dir rebuild to finish
8017 before sending the first wccp2 HereIAm packet
8020 NAME: wccp2_forwarding_method
8022 LOC: Config.Wccp2.forwarding_method
8026 WCCP2 allows the setting of forwarding methods between the
8027 router/switch and the cache. Valid values are as follows:
8029 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
8030 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
8032 Currently (as of IOS 12.4) cisco routers only support GRE.
8033 Cisco switches only support the L2 redirect assignment method.
8036 NAME: wccp2_return_method
8038 LOC: Config.Wccp2.return_method
8042 WCCP2 allows the setting of return methods between the
8043 router/switch and the cache for packets that the cache
8044 decides not to handle. Valid values are as follows:
8046 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
8047 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
8049 Currently (as of IOS 12.4) cisco routers only support GRE.
8050 Cisco switches only support the L2 redirect assignment.
8052 If the "ip wccp redirect exclude in" command has been
8053 enabled on the cache interface, then it is still safe for
8054 the proxy server to use a l2 redirect method even if this
8055 option is set to GRE.
8058 NAME: wccp2_assignment_method
8060 LOC: Config.Wccp2.assignment_method
8064 WCCP2 allows the setting of methods to assign the WCCP hash
8065 Valid values are as follows:
8067 hash - Hash assignment
8068 mask - Mask assignment
8070 As a general rule, cisco routers support the hash assignment method
8071 and cisco switches support the mask assignment method.
8076 LOC: Config.Wccp2.info
8077 DEFAULT_IF_NONE: standard 0
8078 DEFAULT_DOC: Use the 'web-cache' standard service.
8081 WCCP2 allows for multiple traffic services. There are two
8082 types: "standard" and "dynamic". The standard type defines
8083 one service id - http (id 0). The dynamic service ids can be from
8084 51 to 255 inclusive. In order to use a dynamic service id
8085 one must define the type of traffic to be redirected; this is done
8086 using the wccp2_service_info option.
8088 The "standard" type does not require a wccp2_service_info option,
8089 just specifying the service id will suffice.
8091 MD5 service authentication can be enabled by adding
8092 "password=<password>" to the end of this service declaration.
8096 wccp2_service standard 0 # for the 'web-cache' standard service
8097 wccp2_service dynamic 80 # a dynamic service type which will be
8098 # fleshed out with subsequent options.
8099 wccp2_service standard 0 password=foo
8102 NAME: wccp2_service_info
8103 TYPE: wccp2_service_info
8104 LOC: Config.Wccp2.info
8108 Dynamic WCCPv2 services require further information to define the
8109 traffic you wish to have diverted.
8113 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
8114 priority=<priority> ports=<port>,<port>..
8116 The relevant WCCPv2 flags:
8117 + src_ip_hash, dst_ip_hash
8118 + source_port_hash, dst_port_hash
8119 + src_ip_alt_hash, dst_ip_alt_hash
8120 + src_port_alt_hash, dst_port_alt_hash
8123 The port list can be one to eight entries.
8127 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
8128 priority=240 ports=80
8130 Note: the service id must have been defined by a previous
8131 'wccp2_service dynamic <id>' entry.
8136 LOC: Config.Wccp2.weight
8140 Each cache server gets assigned a set of the destination
8141 hash proportional to their weight.
8146 LOC: Config.Wccp.address
8148 DEFAULT_DOC: Address selected by the operating system.
8151 Use this option if you require WCCP(v1) to use a specific
8154 The default behavior is to not bind to any specific address.
8159 LOC: Config.Wccp2.address
8161 DEFAULT_DOC: Address selected by the operating system.
8164 Use this option if you require WCCPv2 to use a specific
8167 The default behavior is to not bind to any specific address.
8171 PERSISTENT CONNECTION HANDLING
8172 -----------------------------------------------------------------------------
8174 Also see "pconn_timeout" in the TIMEOUTS section
8177 NAME: client_persistent_connections
8179 LOC: Config.onoff.client_pconns
8182 Persistent connection support for clients.
8183 Squid uses persistent connections (when allowed). You can use
8184 this option to disable persistent connections with clients.
8187 NAME: server_persistent_connections
8189 LOC: Config.onoff.server_pconns
8192 Persistent connection support for servers.
8193 Squid uses persistent connections (when allowed). You can use
8194 this option to disable persistent connections with servers.
8197 NAME: persistent_connection_after_error
8199 LOC: Config.onoff.error_pconns
8202 With this directive the use of persistent connections after
8203 HTTP errors can be disabled. Useful if you have clients
8204 who fail to handle errors on persistent connections proper.
8207 NAME: detect_broken_pconn
8209 LOC: Config.onoff.detect_broken_server_pconns
8212 Some servers have been found to incorrectly signal the use
8213 of HTTP/1.0 persistent connections even on replies not
8214 compatible, causing significant delays. This server problem
8215 has mostly been seen on redirects.
8217 By enabling this directive Squid attempts to detect such
8218 broken replies and automatically assume the reply is finished
8219 after 10 seconds timeout.
8223 CACHE DIGEST OPTIONS
8224 -----------------------------------------------------------------------------
8227 NAME: digest_generation
8228 IFDEF: USE_CACHE_DIGESTS
8230 LOC: Config.onoff.digest_generation
8233 This controls whether the server will generate a Cache Digest
8234 of its contents. By default, Cache Digest generation is
8235 enabled if Squid is compiled with --enable-cache-digests defined.
8238 NAME: digest_bits_per_entry
8239 IFDEF: USE_CACHE_DIGESTS
8241 LOC: Config.digest.bits_per_entry
8244 This is the number of bits of the server's Cache Digest which
8245 will be associated with the Digest entry for a given HTTP
8246 Method and URL (public key) combination. The default is 5.
8249 NAME: digest_rebuild_period
8250 IFDEF: USE_CACHE_DIGESTS
8253 LOC: Config.digest.rebuild_period
8256 This is the wait time between Cache Digest rebuilds.
8259 NAME: digest_rewrite_period
8261 IFDEF: USE_CACHE_DIGESTS
8263 LOC: Config.digest.rewrite_period
8266 This is the wait time between Cache Digest writes to
8270 NAME: digest_swapout_chunk_size
8273 IFDEF: USE_CACHE_DIGESTS
8274 LOC: Config.digest.swapout_chunk_size
8277 This is the number of bytes of the Cache Digest to write to
8278 disk at a time. It defaults to 4096 bytes (4KB), the Squid
8282 NAME: digest_rebuild_chunk_percentage
8283 COMMENT: (percent, 0-100)
8284 IFDEF: USE_CACHE_DIGESTS
8286 LOC: Config.digest.rebuild_chunk_percentage
8289 This is the percentage of the Cache Digest to be scanned at a
8290 time. By default it is set to 10% of the Cache Digest.
8295 -----------------------------------------------------------------------------
8300 LOC: Config.Port.snmp
8302 DEFAULT_DOC: SNMP disabled.
8305 The port number where Squid listens for SNMP requests. To enable
8306 SNMP support set this to a suitable port number. Port number
8307 3401 is often used for the Squid SNMP agent. By default it's
8308 set to "0" (disabled)
8316 LOC: Config.accessList.snmp
8318 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
8321 Allowing or denying access to the SNMP port.
8323 All access to the agent is denied by default.
8326 snmp_access allow|deny [!]aclname ...
8328 This clause only supports fast acl types.
8329 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8332 snmp_access allow snmppublic localhost
8333 snmp_access deny all
8336 NAME: snmp_incoming_address
8338 LOC: Config.Addrs.snmp_incoming
8340 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
8343 Just like 'udp_incoming_address', but for the SNMP port.
8345 snmp_incoming_address is used for the SNMP socket receiving
8346 messages from SNMP agents.
8348 The default snmp_incoming_address is to listen on all
8349 available network interfaces.
8352 NAME: snmp_outgoing_address
8354 LOC: Config.Addrs.snmp_outgoing
8356 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
8359 Just like 'udp_outgoing_address', but for the SNMP port.
8361 snmp_outgoing_address is used for SNMP packets returned to SNMP
8364 If snmp_outgoing_address is not set it will use the same socket
8365 as snmp_incoming_address. Only change this if you want to have
8366 SNMP replies sent using another address than where this Squid
8367 listens for SNMP queries.
8369 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
8370 the same value since they both use the same port.
8375 -----------------------------------------------------------------------------
8378 NAME: icp_port udp_port
8381 DEFAULT_DOC: ICP disabled.
8382 LOC: Config.Port.icp
8384 The port number where Squid sends and receives ICP queries to
8385 and from neighbor caches. The standard UDP port for ICP is 3130.
8388 icp_port @DEFAULT_ICP_PORT@
8395 DEFAULT_DOC: HTCP disabled.
8396 LOC: Config.Port.htcp
8398 The port number where Squid sends and receives HTCP queries to
8399 and from neighbor caches. To turn it on you want to set it to
8406 NAME: log_icp_queries
8410 LOC: Config.onoff.log_udp
8412 If set, ICP queries are logged to access.log. You may wish
8413 do disable this if your ICP load is VERY high to speed things
8414 up or to simplify log analysis.
8417 NAME: udp_incoming_address
8419 LOC:Config.Addrs.udp_incoming
8421 DEFAULT_DOC: Accept packets from all machine interfaces.
8423 udp_incoming_address is used for UDP packets received from other
8426 The default behavior is to not bind to any specific address.
8428 Only change this if you want to have all UDP queries received on
8429 a specific interface/address.
8431 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
8432 modules. Altering it will affect all of them in the same manner.
8434 see also; udp_outgoing_address
8436 NOTE, udp_incoming_address and udp_outgoing_address can not
8437 have the same value since they both use the same port.
8440 NAME: udp_outgoing_address
8442 LOC: Config.Addrs.udp_outgoing
8444 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
8446 udp_outgoing_address is used for UDP packets sent out to other
8449 The default behavior is to not bind to any specific address.
8451 Instead it will use the same socket as udp_incoming_address.
8452 Only change this if you want to have UDP queries sent using another
8453 address than where this Squid listens for UDP queries from other
8456 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
8457 modules. Altering it will affect all of them in the same manner.
8459 see also; udp_incoming_address
8461 NOTE, udp_incoming_address and udp_outgoing_address can not
8462 have the same value since they both use the same port.
8469 LOC: Config.onoff.icp_hit_stale
8471 If you want to return ICP_HIT for stale cache objects, set this
8472 option to 'on'. If you have sibling relationships with caches
8473 in other administrative domains, this should be 'off'. If you only
8474 have sibling relationships with caches under your control,
8475 it is probably okay to set this to 'on'.
8476 If set to 'on', your siblings should use the option "allow-miss"
8477 on their cache_peer lines for connecting to you.
8480 NAME: minimum_direct_hops
8483 LOC: Config.minDirectHops
8485 If using the ICMP pinging stuff, do direct fetches for sites
8486 which are no more than this many hops away.
8489 NAME: minimum_direct_rtt
8493 LOC: Config.minDirectRtt
8495 If using the ICMP pinging stuff, do direct fetches for sites
8496 which are no more than this many rtt milliseconds away.
8502 LOC: Config.Netdb.low
8504 The low water mark for the ICMP measurement database.
8506 Note: high watermark controlled by netdb_high directive.
8508 These watermarks are counts, not percents. The defaults are
8509 (low) 900 and (high) 1000. When the high water mark is
8510 reached, database entries will be deleted until the low
8517 LOC: Config.Netdb.high
8519 The high water mark for the ICMP measurement database.
8521 Note: low watermark controlled by netdb_low directive.
8523 These watermarks are counts, not percents. The defaults are
8524 (low) 900 and (high) 1000. When the high water mark is
8525 reached, database entries will be deleted until the low
8529 NAME: netdb_ping_period
8531 LOC: Config.Netdb.period
8534 The minimum period for measuring a site. There will be at
8535 least this much delay between successive pings to the same
8536 network. The default is five minutes.
8543 LOC: Config.onoff.query_icmp
8545 If you want to ask your peers to include ICMP data in their ICP
8546 replies, enable this option.
8548 If your peer has configured Squid (during compilation) with
8549 '--enable-icmp' that peer will send ICMP pings to origin server
8550 sites of the URLs it receives. If you enable this option the
8551 ICP replies from that peer will include the ICMP data (if available).
8552 Then, when choosing a parent cache, Squid will choose the parent with
8553 the minimal RTT to the origin server. When this happens, the
8554 hierarchy field of the access.log will be
8555 "CLOSEST_PARENT_MISS". This option is off by default.
8558 NAME: test_reachability
8562 LOC: Config.onoff.test_reachability
8564 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
8565 instead of ICP_MISS if the target host is NOT in the ICMP
8566 database, or has a zero RTT.
8569 NAME: icp_query_timeout
8572 DEFAULT_DOC: Dynamic detection.
8574 LOC: Config.Timeout.icp_query
8576 Normally Squid will automatically determine an optimal ICP
8577 query timeout value based on the round-trip-time of recent ICP
8578 queries. If you want to override the value determined by
8579 Squid, set this 'icp_query_timeout' to a non-zero value. This
8580 value is specified in MILLISECONDS, so, to use a 2-second
8581 timeout (the old default), you would write:
8583 icp_query_timeout 2000
8586 NAME: maximum_icp_query_timeout
8590 LOC: Config.Timeout.icp_query_max
8592 Normally the ICP query timeout is determined dynamically. But
8593 sometimes it can lead to very large values (say 5 seconds).
8594 Use this option to put an upper limit on the dynamic timeout
8595 value. Do NOT use this option to always use a fixed (instead
8596 of a dynamic) timeout value. To set a fixed timeout see the
8597 'icp_query_timeout' directive.
8600 NAME: minimum_icp_query_timeout
8604 LOC: Config.Timeout.icp_query_min
8606 Normally the ICP query timeout is determined dynamically. But
8607 sometimes it can lead to very small timeouts, even lower than
8608 the normal latency variance on your link due to traffic.
8609 Use this option to put an lower limit on the dynamic timeout
8610 value. Do NOT use this option to always use a fixed (instead
8611 of a dynamic) timeout value. To set a fixed timeout see the
8612 'icp_query_timeout' directive.
8615 NAME: background_ping_rate
8619 LOC: Config.backgroundPingRate
8621 Controls how often the ICP pings are sent to siblings that
8622 have background-ping set.
8626 MULTICAST ICP OPTIONS
8627 -----------------------------------------------------------------------------
8632 LOC: Config.mcast_group_list
8635 This tag specifies a list of multicast groups which your server
8636 should join to receive multicasted ICP queries.
8638 NOTE! Be very careful what you put here! Be sure you
8639 understand the difference between an ICP _query_ and an ICP
8640 _reply_. This option is to be set only if you want to RECEIVE
8641 multicast queries. Do NOT set this option to SEND multicast
8642 ICP (use cache_peer for that). ICP replies are always sent via
8643 unicast, so this option does not affect whether or not you will
8644 receive replies from multicast group members.
8646 You must be very careful to NOT use a multicast address which
8647 is already in use by another group of caches.
8649 If you are unsure about multicast, please read the Multicast
8650 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
8652 Usage: mcast_groups 239.128.16.128 224.0.1.20
8654 By default, Squid doesn't listen on any multicast groups.
8657 NAME: mcast_miss_addr
8658 IFDEF: MULTICAST_MISS_STREAM
8660 LOC: Config.mcast_miss.addr
8662 DEFAULT_DOC: disabled.
8664 If you enable this option, every "cache miss" URL will
8665 be sent out on the specified multicast address.
8667 Do not enable this option unless you are are absolutely
8668 certain you understand what you are doing.
8671 NAME: mcast_miss_ttl
8672 IFDEF: MULTICAST_MISS_STREAM
8674 LOC: Config.mcast_miss.ttl
8677 This is the time-to-live value for packets multicasted
8678 when multicasting off cache miss URLs is enabled. By
8679 default this is set to 'site scope', i.e. 16.
8682 NAME: mcast_miss_port
8683 IFDEF: MULTICAST_MISS_STREAM
8685 LOC: Config.mcast_miss.port
8688 This is the port number to be used in conjunction with
8692 NAME: mcast_miss_encode_key
8693 IFDEF: MULTICAST_MISS_STREAM
8695 LOC: Config.mcast_miss.encode_key
8696 DEFAULT: XXXXXXXXXXXXXXXX
8698 The URLs that are sent in the multicast miss stream are
8699 encrypted. This is the encryption key.
8702 NAME: mcast_icp_query_timeout
8706 LOC: Config.Timeout.mcast_icp_query
8708 For multicast peers, Squid regularly sends out ICP "probes" to
8709 count how many other peers are listening on the given multicast
8710 address. This value specifies how long Squid should wait to
8711 count all the replies. The default is 2000 msec, or 2
8716 INTERNAL ICON OPTIONS
8717 -----------------------------------------------------------------------------
8720 NAME: icon_directory
8722 LOC: Config.icons.directory
8723 DEFAULT: @DEFAULT_ICON_DIR@
8725 Where the icons are stored. These are normally kept in
8729 NAME: global_internal_static
8731 LOC: Config.onoff.global_internal_static
8734 This directive controls is Squid should intercept all requests for
8735 /squid-internal-static/ no matter which host the URL is requesting
8736 (default on setting), or if nothing special should be done for
8737 such URLs (off setting). The purpose of this directive is to make
8738 icons etc work better in complex cache hierarchies where it may
8739 not always be possible for all corners in the cache mesh to reach
8740 the server generating a directory listing.
8743 NAME: short_icon_urls
8745 LOC: Config.icons.use_short_names
8748 If this is enabled Squid will use short URLs for icons.
8749 If disabled it will revert to the old behavior of including
8750 it's own name and port in the URL.
8752 If you run a complex cache hierarchy with a mix of Squid and
8753 other proxies you may need to disable this directive.
8758 -----------------------------------------------------------------------------
8761 NAME: error_directory
8763 LOC: Config.errorDirectory
8765 DEFAULT_DOC: Send error pages in the clients preferred language
8767 If you wish to create your own versions of the default
8768 error files to customize them to suit your company copy
8769 the error/template files to another directory and point
8772 WARNING: This option will disable multi-language support
8773 on error pages if used.
8775 The squid developers are interested in making squid available in
8776 a wide variety of languages. If you are making translations for a
8777 language that Squid does not currently provide please consider
8778 contributing your translation back to the project.
8779 http://wiki.squid-cache.org/Translations
8781 The squid developers working on translations are happy to supply drop-in
8782 translated error files in exchange for any new language contributions.
8785 NAME: error_default_language
8786 IFDEF: USE_ERR_LOCALES
8788 LOC: Config.errorDefaultLanguage
8790 DEFAULT_DOC: Generate English language pages.
8792 Set the default language which squid will send error pages in
8793 if no existing translation matches the clients language
8796 If unset (default) generic English will be used.
8798 The squid developers are interested in making squid available in
8799 a wide variety of languages. If you are interested in making
8800 translations for any language see the squid wiki for details.
8801 http://wiki.squid-cache.org/Translations
8804 NAME: error_log_languages
8805 IFDEF: USE_ERR_LOCALES
8807 LOC: Config.errorLogMissingLanguages
8810 Log to cache.log what languages users are attempting to
8811 auto-negotiate for translations.
8813 Successful negotiations are not logged. Only failures
8814 have meaning to indicate that Squid may need an upgrade
8815 of its error page translations.
8818 NAME: err_page_stylesheet
8820 LOC: Config.errorStylesheet
8821 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
8823 CSS Stylesheet to pattern the display of Squid default error pages.
8825 For information on CSS see http://www.w3.org/Style/CSS/
8830 LOC: Config.errHtmlText
8833 HTML text to include in error messages. Make this a "mailto"
8834 URL to your admin address, or maybe just a link to your
8835 organizations Web page.
8837 To include this in your error messages, you must rewrite
8838 the error template files (found in the "errors" directory).
8839 Wherever you want the 'err_html_text' line to appear,
8840 insert a %L tag in the error template file.
8843 NAME: email_err_data
8846 LOC: Config.onoff.emailErrData
8849 If enabled, information about the occurred error will be
8850 included in the mailto links of the ERR pages (if %W is set)
8851 so that the email body contains the data.
8852 Syntax is <A HREF="mailto:%w%W">%w</A>
8857 LOC: Config.denyInfoList
8860 Usage: deny_info err_page_name acl
8861 or deny_info http://... acl
8862 or deny_info TCP_RESET acl
8864 This can be used to return a ERR_ page for requests which
8865 do not pass the 'http_access' rules. Squid remembers the last
8866 acl it evaluated in http_access, and if a 'deny_info' line exists
8867 for that ACL Squid returns a corresponding error page.
8869 The acl is typically the last acl on the http_access deny line which
8870 denied access. The exceptions to this rule are:
8871 - When Squid needs to request authentication credentials. It's then
8872 the first authentication related acl encountered
8873 - When none of the http_access lines matches. It's then the last
8874 acl processed on the last http_access line.
8875 - When the decision to deny access was made by an adaptation service,
8876 the acl name is the corresponding eCAP or ICAP service_name.
8878 NP: If providing your own custom error pages with error_directory
8879 you may also specify them by your custom file name:
8880 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
8882 By default Squid will send "403 Forbidden". A different 4xx or 5xx
8883 may be specified by prefixing the file name with the code and a colon.
8884 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
8886 Alternatively you can tell Squid to reset the TCP connection
8887 by specifying TCP_RESET.
8889 Or you can specify an error URL or URL pattern. The browsers will
8890 get redirected to the specified URL after formatting tags have
8891 been replaced. Redirect will be done with 302 or 307 according to
8892 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
8893 the URL. e.g. 303:http://example.com/
8896 %a - username (if available. Password NOT included)
8897 %A - Local listening IP address the client connection was connected to
8900 %E - Error description
8902 %H - Request domain name
8903 %i - Client IP Address
8905 %O - Unescaped message result from external ACL helper
8906 %o - Message result from external ACL helper
8907 %p - Request Port number
8908 %P - Request Protocol name
8909 %R - Request URL path
8910 %T - Timestamp in RFC 1123 format
8911 %U - Full canonical URL from client
8912 (HTTPS URLs terminate with *)
8913 %u - Full canonical URL from client
8914 %w - Admin email from squid.conf
8916 %% - Literal percent (%) code
8921 OPTIONS INFLUENCING REQUEST FORWARDING
8922 -----------------------------------------------------------------------------
8925 NAME: nonhierarchical_direct
8927 LOC: Config.onoff.nonhierarchical_direct
8930 By default, Squid will send any non-hierarchical requests
8931 (not cacheable request type) direct to origin servers.
8933 When this is set to "off", Squid will prefer to send these
8934 requests to parents.
8936 Note that in most configurations, by turning this off you will only
8937 add latency to these request without any improvement in global hit
8940 This option only sets a preference. If the parent is unavailable a
8941 direct connection to the origin server may still be attempted. To
8942 completely prevent direct connections use never_direct.
8947 LOC: Config.onoff.prefer_direct
8950 Normally Squid tries to use parents for most requests. If you for some
8951 reason like it to first try going direct and only use a parent if
8952 going direct fails set this to on.
8954 By combining nonhierarchical_direct off and prefer_direct on you
8955 can set up Squid to use a parent as a backup path if going direct
8958 Note: If you want Squid to use parents for all requests see
8959 the never_direct directive. prefer_direct only modifies how Squid
8960 acts on cacheable requests.
8963 NAME: cache_miss_revalidate
8967 LOC: Config.onoff.cache_miss_revalidate
8969 RFC 7232 defines a conditional request mechanism to prevent
8970 response objects being unnecessarily transferred over the network.
8971 If that mechanism is used by the client and a cache MISS occurs
8972 it can prevent new cache entries being created.
8974 This option determines whether Squid on cache MISS will pass the
8975 client revalidation request to the server or tries to fetch new
8976 content for caching. It can be useful while the cache is mostly
8977 empty to more quickly have the cache populated by generating
8978 non-conditional GETs.
8980 When set to 'on' (default), Squid will pass all client If-* headers
8981 to the server. This permits server responses without a cacheable
8982 payload to be delivered and on MISS no new cache entry is created.
8984 When set to 'off' and if the request is cacheable, Squid will
8985 remove the clients If-Modified-Since and If-None-Match headers from
8986 the request sent to the server. This requests a 200 status response
8987 from the server to create a new cache entry with.
8992 LOC: Config.accessList.AlwaysDirect
8994 DEFAULT_DOC: Prevent any cache_peer being used for this request.
8996 Usage: always_direct allow|deny [!]aclname ...
8998 Here you can use ACL elements to specify requests which should
8999 ALWAYS be forwarded by Squid to the origin servers without using
9000 any peers. For example, to always directly forward requests for
9001 local servers ignoring any parents or siblings you may have use
9004 acl local-servers dstdomain my.domain.net
9005 always_direct allow local-servers
9007 To always forward FTP requests directly, use
9010 always_direct allow FTP
9012 NOTE: There is a similar, but opposite option named
9013 'never_direct'. You need to be aware that "always_direct deny
9014 foo" is NOT the same thing as "never_direct allow foo". You
9015 may need to use a deny rule to exclude a more-specific case of
9016 some other rule. Example:
9018 acl local-external dstdomain external.foo.net
9019 acl local-servers dstdomain .foo.net
9020 always_direct deny local-external
9021 always_direct allow local-servers
9023 NOTE: If your goal is to make the client forward the request
9024 directly to the origin server bypassing Squid then this needs
9025 to be done in the client configuration. Squid configuration
9026 can only tell Squid how Squid should fetch the object.
9028 NOTE: This directive is not related to caching. The replies
9029 is cached as usual even if you use always_direct. To not cache
9030 the replies see the 'cache' directive.
9032 This clause supports both fast and slow acl types.
9033 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
9038 LOC: Config.accessList.NeverDirect
9040 DEFAULT_DOC: Allow DNS results to be used for this request.
9042 Usage: never_direct allow|deny [!]aclname ...
9044 never_direct is the opposite of always_direct. Please read
9045 the description for always_direct if you have not already.
9047 With 'never_direct' you can use ACL elements to specify
9048 requests which should NEVER be forwarded directly to origin
9049 servers. For example, to force the use of a proxy for all
9050 requests, except those in your local domain use something like:
9052 acl local-servers dstdomain .foo.net
9053 never_direct deny local-servers
9054 never_direct allow all
9056 or if Squid is inside a firewall and there are local intranet
9057 servers inside the firewall use something like:
9059 acl local-intranet dstdomain .foo.net
9060 acl local-external dstdomain external.foo.net
9061 always_direct deny local-external
9062 always_direct allow local-intranet
9063 never_direct allow all
9065 This clause supports both fast and slow acl types.
9066 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
9070 ADVANCED NETWORKING OPTIONS
9071 -----------------------------------------------------------------------------
9074 NAME: incoming_udp_average incoming_icp_average
9077 LOC: Config.comm_incoming.udp.average
9079 Heavy voodoo here. I can't even believe you are reading this.
9080 Are you crazy? Don't even think about adjusting these unless
9081 you understand the algorithms in comm_select.c first!
9084 NAME: incoming_tcp_average incoming_http_average
9087 LOC: Config.comm_incoming.tcp.average
9089 Heavy voodoo here. I can't even believe you are reading this.
9090 Are you crazy? Don't even think about adjusting these unless
9091 you understand the algorithms in comm_select.c first!
9094 NAME: incoming_dns_average
9097 LOC: Config.comm_incoming.dns.average
9099 Heavy voodoo here. I can't even believe you are reading this.
9100 Are you crazy? Don't even think about adjusting these unless
9101 you understand the algorithms in comm_select.c first!
9104 NAME: min_udp_poll_cnt min_icp_poll_cnt
9107 LOC: Config.comm_incoming.udp.min_poll
9109 Heavy voodoo here. I can't even believe you are reading this.
9110 Are you crazy? Don't even think about adjusting these unless
9111 you understand the algorithms in comm_select.c first!
9114 NAME: min_dns_poll_cnt
9117 LOC: Config.comm_incoming.dns.min_poll
9119 Heavy voodoo here. I can't even believe you are reading this.
9120 Are you crazy? Don't even think about adjusting these unless
9121 you understand the algorithms in comm_select.c first!
9124 NAME: min_tcp_poll_cnt min_http_poll_cnt
9127 LOC: Config.comm_incoming.tcp.min_poll
9129 Heavy voodoo here. I can't even believe you are reading this.
9130 Are you crazy? Don't even think about adjusting these unless
9131 you understand the algorithms in comm_select.c first!
9137 LOC: Config.accept_filter
9141 The name of an accept(2) filter to install on Squid's
9142 listen socket(s). This feature is perhaps specific to
9143 FreeBSD and requires support in the kernel.
9145 The 'httpready' filter delays delivering new connections
9146 to Squid until a full HTTP request has been received.
9147 See the accf_http(9) man page for details.
9149 The 'dataready' filter delays delivering new connections
9150 to Squid until there is some data to process.
9151 See the accf_dataready(9) man page for details.
9155 The 'data' filter delays delivering of new connections
9156 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
9157 You may optionally specify a number of seconds to wait by
9158 'data=N' where N is the number of seconds. Defaults to 30
9159 if not specified. See the tcp(7) man page for details.
9162 accept_filter httpready
9167 NAME: client_ip_max_connections
9169 LOC: Config.client_ip_max_connections
9171 DEFAULT_DOC: No limit.
9173 Set an absolute limit on the number of connections a single
9174 client IP can use. Any more than this and Squid will begin to drop
9175 new connections from the client until it closes some links.
9177 Note that this is a global limit. It affects all HTTP, HTCP, and FTP
9178 connections from the client. For finer control use the ACL access controls.
9180 Requires client_db to be enabled (the default).
9182 WARNING: This may noticeably slow down traffic received via external proxies
9183 or NAT devices and cause them to rebound error messages back to their clients.
9186 NAME: tcp_recv_bufsize
9190 DEFAULT_DOC: Use operating system TCP defaults.
9191 LOC: Config.tcpRcvBufsz
9193 Size of receive buffer to set for TCP sockets. Probably just
9194 as easy to change your kernel's default.
9195 Omit from squid.conf to use the default buffer size.
9200 -----------------------------------------------------------------------------
9207 LOC: Adaptation::Icap::TheConfig.onoff
9210 If you want to enable the ICAP module support, set this to on.
9213 NAME: icap_connect_timeout
9216 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
9219 This parameter specifies how long to wait for the TCP connect to
9220 the requested ICAP server to complete before giving up and either
9221 terminating the HTTP transaction or bypassing the failure.
9223 The default for optional services is peer_connect_timeout.
9224 The default for essential services is connect_timeout.
9225 If this option is explicitly set, its value applies to all services.
9228 NAME: icap_io_timeout
9232 DEFAULT_DOC: Use read_timeout.
9233 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
9236 This parameter specifies how long to wait for an I/O activity on
9237 an established, active ICAP connection before giving up and
9238 either terminating the HTTP transaction or bypassing the
9242 NAME: icap_service_failure_limit
9243 COMMENT: limit [in memory-depth time-units]
9244 TYPE: icap_service_failure_limit
9246 LOC: Adaptation::Icap::TheConfig
9249 The limit specifies the number of failures that Squid tolerates
9250 when establishing a new TCP connection with an ICAP service. If
9251 the number of failures exceeds the limit, the ICAP service is
9252 not used for new ICAP requests until it is time to refresh its
9255 A negative value disables the limit. Without the limit, an ICAP
9256 service will not be considered down due to connectivity failures
9257 between ICAP OPTIONS requests.
9259 Squid forgets ICAP service failures older than the specified
9260 value of memory-depth. The memory fading algorithm
9261 is approximate because Squid does not remember individual
9262 errors but groups them instead, splitting the option
9263 value into ten time slots of equal length.
9265 When memory-depth is 0 and by default this option has no
9266 effect on service failure expiration.
9268 Squid always forgets failures when updating service settings
9269 using an ICAP OPTIONS transaction, regardless of this option
9273 # suspend service usage after 10 failures in 5 seconds:
9274 icap_service_failure_limit 10 in 5 seconds
9277 NAME: icap_service_revival_delay
9280 LOC: Adaptation::Icap::TheConfig.service_revival_delay
9283 The delay specifies the number of seconds to wait after an ICAP
9284 OPTIONS request failure before requesting the options again. The
9285 failed ICAP service is considered "down" until fresh OPTIONS are
9288 The actual delay cannot be smaller than the hardcoded minimum
9289 delay of 30 seconds.
9292 NAME: icap_preview_enable
9296 LOC: Adaptation::Icap::TheConfig.preview_enable
9299 The ICAP Preview feature allows the ICAP server to handle the
9300 HTTP message by looking only at the beginning of the message body
9301 or even without receiving the body at all. In some environments,
9302 previews greatly speedup ICAP processing.
9304 During an ICAP OPTIONS transaction, the server may tell Squid what
9305 HTTP messages should be previewed and how big the preview should be.
9306 Squid will not use Preview if the server did not request one.
9308 To disable ICAP Preview for all ICAP services, regardless of
9309 individual ICAP server OPTIONS responses, set this option to "off".
9311 icap_preview_enable off
9314 NAME: icap_preview_size
9317 LOC: Adaptation::Icap::TheConfig.preview_size
9319 DEFAULT_DOC: No preview sent.
9321 The default size of preview data to be sent to the ICAP server.
9322 This value might be overwritten on a per server basis by OPTIONS requests.
9325 NAME: icap_206_enable
9329 LOC: Adaptation::Icap::TheConfig.allow206_enable
9332 206 (Partial Content) responses is an ICAP extension that allows the
9333 ICAP agents to optionally combine adapted and original HTTP message
9334 content. The decision to combine is postponed until the end of the
9335 ICAP response. Squid supports Partial Content extension by default.
9337 Activation of the Partial Content extension is negotiated with each
9338 ICAP service during OPTIONS exchange. Most ICAP servers should handle
9339 negotiation correctly even if they do not support the extension, but
9340 some might fail. To disable Partial Content support for all ICAP
9341 services and to avoid any negotiation, set this option to "off".
9347 NAME: icap_default_options_ttl
9350 LOC: Adaptation::Icap::TheConfig.default_options_ttl
9353 The default TTL value for ICAP OPTIONS responses that don't have
9354 an Options-TTL header.
9357 NAME: icap_persistent_connections
9361 LOC: Adaptation::Icap::TheConfig.reuse_connections
9364 Whether or not Squid should use persistent connections to
9368 NAME: adaptation_send_client_ip icap_send_client_ip
9370 IFDEF: USE_ADAPTATION
9372 LOC: Adaptation::Config::send_client_ip
9375 If enabled, Squid shares HTTP client IP information with adaptation
9376 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
9377 For eCAP, Squid sets the libecap::metaClientIp transaction option.
9379 See also: adaptation_uses_indirect_client
9382 NAME: adaptation_send_username icap_send_client_username
9384 IFDEF: USE_ADAPTATION
9386 LOC: Adaptation::Config::send_username
9389 This sends authenticated HTTP client username (if available) to
9390 the adaptation service.
9392 For ICAP, the username value is encoded based on the
9393 icap_client_username_encode option and is sent using the header
9394 specified by the icap_client_username_header option.
9397 NAME: icap_client_username_header
9400 LOC: Adaptation::Icap::TheConfig.client_username_header
9401 DEFAULT: X-Client-Username
9403 ICAP request header name to use for adaptation_send_username.
9406 NAME: icap_client_username_encode
9410 LOC: Adaptation::Icap::TheConfig.client_username_encode
9413 Whether to base64 encode the authenticated client username.
9417 TYPE: icap_service_type
9419 LOC: Adaptation::Icap::TheConfig
9422 Defines a single ICAP service using the following format:
9424 icap_service id vectoring_point uri [option ...]
9427 an opaque identifier or name which is used to direct traffic to
9428 this specific service. Must be unique among all adaptation
9429 services in squid.conf.
9431 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
9432 This specifies at which point of transaction processing the
9433 ICAP service should be activated. *_postcache vectoring points
9434 are not yet supported.
9436 uri: icap://servername:port/servicepath
9437 ICAP server and service location.
9438 icaps://servername:port/servicepath
9439 The "icap:" URI scheme is used for traditional ICAP server and
9440 service location (default port is 1344, connections are not
9441 encrypted). The "icaps:" URI scheme is for Secure ICAP
9442 services that use SSL/TLS-encrypted ICAP connections (by
9443 default, on port 11344).
9445 ICAP does not allow a single service to handle both REQMOD and RESPMOD
9446 transactions. Squid does not enforce that requirement. You can specify
9447 services with the same service_url and different vectoring_points. You
9448 can even specify multiple identical services as long as their
9449 service_names differ.
9451 To activate a service, use the adaptation_access directive. To group
9452 services, use adaptation_service_chain and adaptation_service_set.
9454 Service options are separated by white space. ICAP services support
9455 the following name=value options:
9458 If set to 'on' or '1', the ICAP service is treated as
9459 optional. If the service cannot be reached or malfunctions,
9460 Squid will try to ignore any errors and process the message as
9461 if the service was not enabled. No all ICAP errors can be
9462 bypassed. If set to 0, the ICAP service is treated as
9463 essential and all ICAP errors will result in an error page
9464 returned to the HTTP client.
9466 Bypass is off by default: services are treated as essential.
9469 If set to 'on' or '1', the ICAP service is allowed to
9470 dynamically change the current message adaptation plan by
9471 returning a chain of services to be used next. The services
9472 are specified using the X-Next-Services ICAP response header
9473 value, formatted as a comma-separated list of service names.
9474 Each named service should be configured in squid.conf. Other
9475 services are ignored. An empty X-Next-Services value results
9476 in an empty plan which ends the current adaptation.
9478 Dynamic adaptation plan may cross or cover multiple supported
9479 vectoring points in their natural processing order.
9481 Routing is not allowed by default: the ICAP X-Next-Services
9482 response header is ignored.
9485 Only has effect on split-stack systems. The default on those systems
9486 is to use IPv4-only connections. When set to 'on' this option will
9487 make Squid use IPv6-only connections to contact this ICAP service.
9489 on-overload=block|bypass|wait|force
9490 If the service Max-Connections limit has been reached, do
9491 one of the following for each new ICAP transaction:
9492 * block: send an HTTP error response to the client
9493 * bypass: ignore the "over-connected" ICAP service
9494 * wait: wait (in a FIFO queue) for an ICAP connection slot
9495 * force: proceed, ignoring the Max-Connections limit
9497 In SMP mode with N workers, each worker assumes the service
9498 connection limit is Max-Connections/N, even though not all
9499 workers may use a given service.
9501 The default value is "bypass" if service is bypassable,
9502 otherwise it is set to "wait".
9506 Use the given number as the Max-Connections limit, regardless
9507 of the Max-Connections value given by the service, if any.
9509 connection-encryption=on|off
9510 Determines the ICAP service effect on the connections_encrypted
9513 The default is "on" for Secure ICAP services (i.e., those
9514 with the icaps:// service URIs scheme) and "off" for plain ICAP
9517 Does not affect ICAP connections (e.g., does not turn Secure
9520 ==== ICAPS / TLS OPTIONS ====
9522 These options are used for Secure ICAP (icaps://....) services only.
9524 tls-cert=/path/to/ssl/certificate
9525 A client X.509 certificate to use when connecting to
9528 tls-key=/path/to/ssl/key
9529 The private key corresponding to the previous
9532 If tls-key= is not specified tls-cert= is assumed to
9533 reference a PEM file containing both the certificate
9536 tls-cipher=... The list of valid TLS/SSL ciphers to use when connecting
9537 to this icap server.
9540 The minimum TLS protocol version to permit. To control
9541 SSLv3 use the tls-options= parameter.
9542 Supported Values: 1.0 (default), 1.1, 1.2
9544 tls-options=... Specify various OpenSSL library options:
9546 NO_SSLv3 Disallow the use of SSLv3
9549 Always create a new key when using
9550 temporary/ephemeral DH key exchanges
9552 ALL Enable various bug workarounds
9553 suggested as "harmless" by OpenSSL
9554 Be warned that this reduces SSL/TLS
9555 strength to some attacks.
9557 See the OpenSSL SSL_CTX_set_options documentation for a
9558 more complete list. Options relevant only to SSLv2 are
9561 tls-cafile= PEM file containing CA certificates to use when verifying
9562 the icap server certificate.
9563 Use to specify intermediate CA certificate(s) if not sent
9564 by the server. Or the full CA chain for the server when
9565 using the tls-default-ca=off flag.
9566 May be repeated to load multiple files.
9568 tls-capath=... A directory containing additional CA certificates to
9569 use when verifying the icap server certificate.
9570 Requires OpenSSL or LibreSSL.
9572 tls-crlfile=... A certificate revocation list file to use when
9573 verifying the icap server certificate.
9575 tls-flags=... Specify various flags modifying the Squid TLS implementation:
9578 Accept certificates even if they fail to
9581 Don't verify the icap server certificate
9582 matches the server name
9584 tls-default-ca[=off]
9585 Whether to use the system Trusted CAs. Default is ON.
9587 tls-domain= The icap server name as advertised in it's certificate.
9588 Used for verifying the correctness of the received icap
9589 server certificate. If not specified the icap server
9590 hostname extracted from ICAP URI will be used.
9592 Older icap_service format without optional named parameters is
9593 deprecated but supported for backward compatibility.
9596 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
9597 icap_service svcLogger reqmod_precache icaps://icap2.mydomain.net:11344/reqmod routing=on
9601 TYPE: icap_class_type
9606 This deprecated option was documented to define an ICAP service
9607 chain, even though it actually defined a set of similar, redundant
9608 services, and the chains were not supported.
9610 To define a set of redundant services, please use the
9611 adaptation_service_set directive. For service chains, use
9612 adaptation_service_chain.
9616 TYPE: icap_access_type
9621 This option is deprecated. Please use adaptation_access, which
9622 has the same ICAP functionality, but comes with better
9623 documentation, and eCAP support.
9628 -----------------------------------------------------------------------------
9635 LOC: Adaptation::Ecap::TheConfig.onoff
9638 Controls whether eCAP support is enabled.
9642 TYPE: ecap_service_type
9644 LOC: Adaptation::Ecap::TheConfig
9647 Defines a single eCAP service
9649 ecap_service id vectoring_point uri [option ...]
9652 an opaque identifier or name which is used to direct traffic to
9653 this specific service. Must be unique among all adaptation
9654 services in squid.conf.
9656 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
9657 This specifies at which point of transaction processing the
9658 eCAP service should be activated. *_postcache vectoring points
9659 are not yet supported.
9661 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
9662 Squid uses the eCAP service URI to match this configuration
9663 line with one of the dynamically loaded services. Each loaded
9664 eCAP service must have a unique URI. Obtain the right URI from
9665 the service provider.
9667 To activate a service, use the adaptation_access directive. To group
9668 services, use adaptation_service_chain and adaptation_service_set.
9670 Service options are separated by white space. eCAP services support
9671 the following name=value options:
9674 If set to 'on' or '1', the eCAP service is treated as optional.
9675 If the service cannot be reached or malfunctions, Squid will try
9676 to ignore any errors and process the message as if the service
9677 was not enabled. No all eCAP errors can be bypassed.
9678 If set to 'off' or '0', the eCAP service is treated as essential
9679 and all eCAP errors will result in an error page returned to the
9682 Bypass is off by default: services are treated as essential.
9685 If set to 'on' or '1', the eCAP service is allowed to
9686 dynamically change the current message adaptation plan by
9687 returning a chain of services to be used next.
9689 Dynamic adaptation plan may cross or cover multiple supported
9690 vectoring points in their natural processing order.
9692 Routing is not allowed by default.
9694 connection-encryption=on|off
9695 Determines the eCAP service effect on the connections_encrypted
9698 Defaults to "on", which does not taint the master transaction
9701 Does not affect eCAP API calls.
9703 Older ecap_service format without optional named parameters is
9704 deprecated but supported for backward compatibility.
9708 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
9709 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
9712 NAME: loadable_modules
9714 IFDEF: USE_LOADABLE_MODULES
9715 LOC: Config.loadable_module_names
9718 Instructs Squid to load the specified dynamic module(s) or activate
9719 preloaded module(s).
9721 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
9725 MESSAGE ADAPTATION OPTIONS
9726 -----------------------------------------------------------------------------
9729 NAME: adaptation_service_set
9730 TYPE: adaptation_service_set_type
9731 IFDEF: USE_ADAPTATION
9736 Configures an ordered set of similar, redundant services. This is
9737 useful when hot standby or backup adaptation servers are available.
9739 adaptation_service_set set_name service_name1 service_name2 ...
9741 The named services are used in the set declaration order. The first
9742 applicable adaptation service from the set is used first. The next
9743 applicable service is tried if and only if the transaction with the
9744 previous service fails and the message waiting to be adapted is still
9747 When adaptation starts, broken services are ignored as if they were
9748 not a part of the set. A broken service is a down optional service.
9750 The services in a set must be attached to the same vectoring point
9751 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9753 If all services in a set are optional then adaptation failures are
9754 bypassable. If all services in the set are essential, then a
9755 transaction failure with one service may still be retried using
9756 another service from the set, but when all services fail, the master
9757 transaction fails as well.
9759 A set may contain a mix of optional and essential services, but that
9760 is likely to lead to surprising results because broken services become
9761 ignored (see above), making previously bypassable failures fatal.
9762 Technically, it is the bypassability of the last failed service that
9765 See also: adaptation_access adaptation_service_chain
9768 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
9769 adaptation service_set svcLogger loggerLocal loggerRemote
9772 NAME: adaptation_service_chain
9773 TYPE: adaptation_service_chain_type
9774 IFDEF: USE_ADAPTATION
9779 Configures a list of complementary services that will be applied
9780 one-by-one, forming an adaptation chain or pipeline. This is useful
9781 when Squid must perform different adaptations on the same message.
9783 adaptation_service_chain chain_name service_name1 svc_name2 ...
9785 The named services are used in the chain declaration order. The first
9786 applicable adaptation service from the chain is used first. The next
9787 applicable service is applied to the successful adaptation results of
9788 the previous service in the chain.
9790 When adaptation starts, broken services are ignored as if they were
9791 not a part of the chain. A broken service is a down optional service.
9793 Request satisfaction terminates the adaptation chain because Squid
9794 does not currently allow declaration of RESPMOD services at the
9795 "reqmod_precache" vectoring point (see icap_service or ecap_service).
9797 The services in a chain must be attached to the same vectoring point
9798 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9800 A chain may contain a mix of optional and essential services. If an
9801 essential adaptation fails (or the failure cannot be bypassed for
9802 other reasons), the master transaction fails. Otherwise, the failure
9803 is bypassed as if the failed adaptation service was not in the chain.
9805 See also: adaptation_access adaptation_service_set
9808 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
9811 NAME: adaptation_access
9812 TYPE: adaptation_access_type
9813 IFDEF: USE_ADAPTATION
9816 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
9818 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
9820 adaptation_access service_name allow|deny [!]aclname...
9821 adaptation_access set_name allow|deny [!]aclname...
9823 At each supported vectoring point, the adaptation_access
9824 statements are processed in the order they appear in this
9825 configuration file. Statements pointing to the following services
9826 are ignored (i.e., skipped without checking their ACL):
9828 - services serving different vectoring points
9829 - "broken-but-bypassable" services
9830 - "up" services configured to ignore such transactions
9831 (e.g., based on the ICAP Transfer-Ignore header).
9833 When a set_name is used, all services in the set are checked
9834 using the same rules, to find the first applicable one. See
9835 adaptation_service_set for details.
9837 If an access list is checked and there is a match, the
9838 processing stops: For an "allow" rule, the corresponding
9839 adaptation service is used for the transaction. For a "deny"
9840 rule, no adaptation service is activated.
9842 It is currently not possible to apply more than one adaptation
9843 service at the same vectoring point to the same HTTP transaction.
9845 See also: icap_service and ecap_service
9848 adaptation_access service_1 allow all
9851 NAME: adaptation_service_iteration_limit
9853 IFDEF: USE_ADAPTATION
9854 LOC: Adaptation::Config::service_iteration_limit
9857 Limits the number of iterations allowed when applying adaptation
9858 services to a message. If your longest adaptation set or chain
9859 may have more than 16 services, increase the limit beyond its
9860 default value of 16. If detecting infinite iteration loops sooner
9861 is critical, make the iteration limit match the actual number
9862 of services in your longest adaptation set or chain.
9864 Infinite adaptation loops are most likely with routing services.
9866 See also: icap_service routing=1
9869 NAME: adaptation_masterx_shared_names
9871 IFDEF: USE_ADAPTATION
9872 LOC: Adaptation::Config::masterx_shared_name
9875 For each master transaction (i.e., the HTTP request and response
9876 sequence, including all related ICAP and eCAP exchanges), Squid
9877 maintains a table of metadata. The table entries are (name, value)
9878 pairs shared among eCAP and ICAP exchanges. The table is destroyed
9879 with the master transaction.
9881 This option specifies the table entry names that Squid must accept
9882 from and forward to the adaptation transactions.
9884 An ICAP REQMOD or RESPMOD transaction may set an entry in the
9885 shared table by returning an ICAP header field with a name
9886 specified in adaptation_masterx_shared_names.
9888 An eCAP REQMOD or RESPMOD transaction may set an entry in the
9889 shared table by implementing the libecap::visitEachOption() API
9890 to provide an option with a name specified in
9891 adaptation_masterx_shared_names.
9893 Squid will store and forward the set entry to subsequent adaptation
9894 transactions within the same master transaction scope.
9896 Only one shared entry name is supported at this time.
9899 # share authentication information among ICAP services
9900 adaptation_masterx_shared_names X-Subscriber-ID
9903 NAME: adaptation_meta
9905 IFDEF: USE_ADAPTATION
9906 LOC: Adaptation::Config::metaHeaders
9909 This option allows Squid administrator to add custom ICAP request
9910 headers or eCAP options to Squid ICAP requests or eCAP transactions.
9911 Use it to pass custom authentication tokens and other
9912 transaction-state related meta information to an ICAP/eCAP service.
9914 The addition of a meta header is ACL-driven:
9915 adaptation_meta name value [!]aclname ...
9917 Processing for a given header name stops after the first ACL list match.
9918 Thus, it is impossible to add two headers with the same name. If no ACL
9919 lists match for a given header name, no such header is added. For
9922 # do not debug transactions except for those that need debugging
9923 adaptation_meta X-Debug 1 needs_debugging
9925 # log all transactions except for those that must remain secret
9926 adaptation_meta X-Log 1 !keep_secret
9928 # mark transactions from users in the "G 1" group
9929 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
9931 The "value" parameter may be a regular squid.conf token or a "double
9932 quoted string". Within the quoted string, use backslash (\) to escape
9933 any character, which is currently only useful for escaping backslashes
9934 and double quotes. For example,
9935 "this string has one backslash (\\) and two \"quotes\""
9937 Used adaptation_meta header values may be logged via %note
9938 logformat code. If multiple adaptation_meta headers with the same name
9939 are used during master transaction lifetime, the header values are
9940 logged in the order they were used and duplicate values are ignored
9941 (only the first repeated value will be logged).
9947 LOC: Adaptation::Icap::TheConfig.repeat
9948 DEFAULT_IF_NONE: deny all
9950 This ACL determines which retriable ICAP transactions are
9951 retried. Transactions that received a complete ICAP response
9952 and did not have to consume or produce HTTP bodies to receive
9953 that response are usually retriable.
9955 icap_retry allow|deny [!]aclname ...
9957 Squid automatically retries some ICAP I/O timeouts and errors
9958 due to persistent connection race conditions.
9960 See also: icap_retry_limit
9963 NAME: icap_retry_limit
9966 LOC: Adaptation::Icap::TheConfig.repeat_limit
9968 DEFAULT_DOC: No retries are allowed.
9970 Limits the number of retries allowed.
9972 Communication errors due to persistent connection race
9973 conditions are unavoidable, automatically retried, and do not
9974 count against this limit.
9976 See also: icap_retry
9982 -----------------------------------------------------------------------------
9985 NAME: check_hostnames
9988 LOC: Config.onoff.check_hostnames
9990 For security and stability reasons Squid can check
9991 hostnames for Internet standard RFC compliance. If you want
9992 Squid to perform these checks turn this directive on.
9995 NAME: allow_underscore
9998 LOC: Config.onoff.allow_underscore
10000 Underscore characters is not strictly allowed in Internet hostnames
10001 but nevertheless used by many sites. Set this to off if you want
10002 Squid to be strict about the standard.
10003 This check is performed only when check_hostnames is set to on.
10006 NAME: dns_retransmit_interval
10009 LOC: Config.Timeout.idns_retransmit
10011 Initial retransmit interval for DNS queries. The interval is
10012 doubled each time all configured DNS servers have been tried.
10017 DEFAULT: 30 seconds
10018 LOC: Config.Timeout.idns_query
10020 DNS Query timeout. If no response is received to a DNS query
10021 within this time all DNS servers for the queried domain
10022 are assumed to be unavailable.
10025 NAME: dns_packet_max
10027 DEFAULT_DOC: EDNS disabled
10029 LOC: Config.dns.packet_max
10031 Maximum number of bytes packet size to advertise via EDNS.
10032 Set to "none" to disable EDNS large packet support.
10034 For legacy reasons DNS UDP replies will default to 512 bytes which
10035 is too small for many responses. EDNS provides a means for Squid to
10036 negotiate receiving larger responses back immediately without having
10037 to failover with repeat requests. Responses larger than this limit
10038 will retain the old behaviour of failover to TCP DNS.
10040 Squid has no real fixed limit internally, but allowing packet sizes
10041 over 1500 bytes requires network jumbogram support and is usually not
10044 WARNING: The RFC also indicates that some older resolvers will reply
10045 with failure of the whole request if the extension is added. Some
10046 resolvers have already been identified which will reply with mangled
10047 EDNS response on occasion. Usually in response to many-KB jumbogram
10048 sizes being advertised by Squid.
10049 Squid will currently treat these both as an unable-to-resolve domain
10050 even if it would be resolvable without EDNS.
10057 DEFAULT_DOC: Search for single-label domain names is disabled.
10058 LOC: Config.onoff.res_defnames
10060 Normally the RES_DEFNAMES resolver option is disabled
10061 (see res_init(3)). This prevents caches in a hierarchy
10062 from interpreting single-component hostnames locally. To allow
10063 Squid to handle single-component names, enable this option.
10066 NAME: dns_multicast_local
10070 DEFAULT_DOC: Search for .local and .arpa names is disabled.
10071 LOC: Config.onoff.dns_mdns
10073 When set to on, Squid sends multicast DNS lookups on the local
10074 network for domains ending in .local and .arpa.
10075 This enables local servers and devices to be contacted in an
10076 ad-hoc or zero-configuration network environment.
10079 NAME: dns_nameservers
10082 DEFAULT_DOC: Use operating system definitions
10083 LOC: Config.dns.nameservers
10085 Use this if you want to specify a list of DNS name servers
10086 (IP addresses) to use instead of those given in your
10087 /etc/resolv.conf file.
10089 On Windows platforms, if no value is specified here or in
10090 the /etc/resolv.conf file, the list of DNS name servers are
10091 taken from the Windows registry, both static and dynamic DHCP
10092 configurations are supported.
10094 Example: dns_nameservers 10.0.0.1 192.172.0.4
10099 DEFAULT: @DEFAULT_HOSTS@
10100 LOC: Config.etcHostsPath
10102 Location of the host-local IP name-address associations
10103 database. Most Operating Systems have such a file on different
10105 - Un*X & Linux: /etc/hosts
10106 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
10107 (%SystemRoot% value install default is c:\winnt)
10108 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
10109 (%SystemRoot% value install default is c:\windows)
10110 - Windows 9x/Me: %windir%\hosts
10111 (%windir% value is usually c:\windows)
10112 - Cygwin: /etc/hosts
10114 The file contains newline-separated definitions, in the
10115 form ip_address_in_dotted_form name [name ...] names are
10116 whitespace-separated. Lines beginning with an hash (#)
10117 character are comments.
10119 The file is checked at startup and upon configuration.
10120 If set to 'none', it won't be checked.
10121 If append_domain is used, that domain will be added to
10122 domain-local (i.e. not containing any dot character) host
10126 NAME: append_domain
10128 LOC: Config.appendDomain
10130 DEFAULT_DOC: Use operating system definitions
10132 Appends local domain name to hostnames without any dots in
10133 them. append_domain must begin with a period.
10135 Be warned there are now Internet names with no dots in
10136 them using only top-domain names, so setting this may
10137 cause some Internet sites to become unavailable.
10140 append_domain .yourdomain.com
10143 NAME: ignore_unknown_nameservers
10145 LOC: Config.onoff.ignore_unknown_nameservers
10148 By default Squid checks that DNS responses are received
10149 from the same IP addresses they are sent to. If they
10150 don't match, Squid ignores the response and writes a warning
10151 message to cache.log. You can allow responses from unknown
10152 nameservers by setting this option to 'off'.
10156 COMMENT: (number of entries)
10159 LOC: Config.ipcache.size
10161 Maximum number of DNS IP cache entries.
10168 LOC: Config.ipcache.low
10175 LOC: Config.ipcache.high
10177 The size, low-, and high-water marks for the IP cache.
10180 NAME: fqdncache_size
10181 COMMENT: (number of entries)
10184 LOC: Config.fqdncache.size
10186 Maximum number of FQDN cache entries.
10191 -----------------------------------------------------------------------------
10194 NAME: configuration_includes_quoted_values
10196 TYPE: configuration_includes_quoted_values
10198 LOC: ConfigParser::RecognizeQuotedValues
10200 If set, Squid will recognize each "quoted string" after a configuration
10201 directive as a single parameter. The quotes are stripped before the
10202 parameter value is interpreted or used.
10203 See "Values with spaces, quotes, and other special characters"
10204 section for more details.
10211 LOC: Config.onoff.mem_pools
10213 If set, Squid will keep pools of allocated (but unused) memory
10214 available for future use. If memory is a premium on your
10215 system and you believe your malloc library outperforms Squid
10216 routines, disable this.
10219 NAME: memory_pools_limit
10223 LOC: Config.MemPools.limit
10225 Used only with memory_pools on:
10226 memory_pools_limit 50 MB
10228 If set to a non-zero value, Squid will keep at most the specified
10229 limit of allocated (but unused) memory in memory pools. All free()
10230 requests that exceed this limit will be handled by your malloc
10231 library. Squid does not pre-allocate any memory, just safe-keeps
10232 objects that otherwise would be free()d. Thus, it is safe to set
10233 memory_pools_limit to a reasonably high value even if your
10234 configuration will use less memory.
10236 If set to none, Squid will keep all memory it can. That is, there
10237 will be no limit on the total amount of memory used for safe-keeping.
10239 To disable memory allocation optimization, do not set
10240 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
10242 An overhead for maintaining memory pools is not taken into account
10243 when the limit is checked. This overhead is close to four bytes per
10244 object kept. However, pools may actually _save_ memory because of
10245 reduced memory thrashing in your malloc library.
10248 NAME: forwarded_for
10249 COMMENT: on|off|transparent|truncate|delete
10252 LOC: opt_forwarded_for
10254 If set to "on", Squid will append your client's IP address
10255 in the HTTP requests it forwards. By default it looks like:
10257 X-Forwarded-For: 192.1.2.3
10259 If set to "off", it will appear as
10261 X-Forwarded-For: unknown
10263 If set to "transparent", Squid will not alter the
10264 X-Forwarded-For header in any way.
10266 If set to "delete", Squid will delete the entire
10267 X-Forwarded-For header.
10269 If set to "truncate", Squid will remove all existing
10270 X-Forwarded-For entries, and place the client IP as the sole entry.
10273 NAME: cachemgr_passwd
10274 TYPE: cachemgrpasswd
10276 DEFAULT_DOC: No password. Actions which require password are denied.
10277 LOC: Config.passwd_list
10279 Specify passwords for cachemgr operations.
10281 Usage: cachemgr_passwd password action action ...
10283 Some valid actions are (see cache manager menu for a full list):
10323 * Indicates actions which will not be performed without a
10324 valid password, others can be performed if not listed here.
10326 To disable an action, set the password to "disable".
10327 To allow performing an action without a password, set the
10328 password to "none".
10330 Use the keyword "all" to set the same password for all actions.
10333 cachemgr_passwd secret shutdown
10334 cachemgr_passwd lesssssssecret info stats/objects
10335 cachemgr_passwd disable all
10342 LOC: Config.onoff.client_db
10344 If you want to disable collecting per-client statistics,
10345 turn off client_db here.
10348 NAME: refresh_all_ims
10352 LOC: Config.onoff.refresh_all_ims
10354 When you enable this option, squid will always check
10355 the origin server for an update when a client sends an
10356 If-Modified-Since request. Many browsers use IMS
10357 requests when the user requests a reload, and this
10358 ensures those clients receive the latest version.
10360 By default (off), squid may return a Not Modified response
10361 based on the age of the cached version.
10364 NAME: reload_into_ims
10365 IFDEF: USE_HTTP_VIOLATIONS
10369 LOC: Config.onoff.reload_into_ims
10371 When you enable this option, client no-cache or ``reload''
10372 requests will be changed to If-Modified-Since requests.
10373 Doing this VIOLATES the HTTP standard. Enabling this
10374 feature could make you liable for problems which it
10377 see also refresh_pattern for a more selective approach.
10380 NAME: connect_retries
10382 LOC: Config.connect_retries
10384 DEFAULT_DOC: Do not retry failed connections.
10386 Limits the number of reopening attempts when establishing a single
10387 TCP connection. All these attempts must still complete before the
10388 applicable connection opening timeout expires.
10390 By default and when connect_retries is set to zero, Squid does not
10391 retry failed connection opening attempts.
10393 The (not recommended) maximum is 10 tries. An attempt to configure a
10394 higher value results in the value of 10 being used (with a warning).
10396 Squid may open connections to retry various high-level forwarding
10397 failures. For an outside observer, that activity may look like a
10398 low-level connection reopening attempt, but those high-level retries
10399 are governed by forward_max_tries instead.
10401 See also: connect_timeout, forward_timeout, icap_connect_timeout,
10402 ident_timeout, and forward_max_tries.
10405 NAME: retry_on_error
10407 LOC: Config.retry.onerror
10410 If set to ON Squid will automatically retry requests when
10411 receiving an error response with status 403 (Forbidden),
10412 500 (Internal Error), 501 or 503 (Service not available).
10413 Status 502 and 504 (Gateway errors) are always retried.
10415 This is mainly useful if you are in a complex cache hierarchy to
10416 work around access control errors.
10418 NOTE: This retry will attempt to find another working destination.
10419 Which is different from the server which just failed.
10422 NAME: as_whois_server
10424 LOC: Config.as_whois_server
10425 DEFAULT: whois.ra.net
10427 WHOIS server to query for AS numbers. NOTE: AS numbers are
10428 queried only when Squid starts up, not for every request.
10433 LOC: Config.onoff.offline
10436 Enable this option and Squid will never try to validate cached
10440 NAME: uri_whitespace
10441 TYPE: uri_whitespace
10442 LOC: Config.uri_whitespace
10445 What to do with requests that have whitespace characters in the
10448 strip: The whitespace characters are stripped out of the URL.
10449 This is the behavior recommended by RFC2396 and RFC3986
10450 for tolerant handling of generic URI.
10451 NOTE: This is one difference between generic URI and HTTP URLs.
10453 deny: The request is denied. The user receives an "Invalid
10455 This is the behaviour recommended by RFC2616 for safe
10456 handling of HTTP request URL.
10458 allow: The request is allowed and the URI is not changed. The
10459 whitespace characters remain in the URI. Note the
10460 whitespace is passed to redirector processes if they
10462 Note this may be considered a violation of RFC2616
10463 request parsing where whitespace is prohibited in the
10466 encode: The request is allowed and the whitespace characters are
10467 encoded according to RFC1738.
10469 chop: The request is allowed and the URI is chopped at the
10473 NOTE the current Squid implementation of encode and chop violates
10474 RFC2616 by not using a 301 redirect after altering the URL.
10479 LOC: Config.chroot_dir
10482 Specifies a directory where Squid should do a chroot() while
10483 initializing. This also causes Squid to fully drop root
10484 privileges after initializing. This means, for example, if you
10485 use a HTTP port less than 1024 and try to reconfigure, you may
10486 get an error saying that Squid can not open the port.
10489 NAME: pipeline_prefetch
10490 TYPE: pipelinePrefetch
10491 LOC: Config.pipeline_max_prefetch
10493 DEFAULT_DOC: Do not pre-parse pipelined requests.
10495 HTTP clients may send a pipeline of 1+N requests to Squid using a
10496 single connection, without waiting for Squid to respond to the first
10497 of those requests. This option limits the number of concurrent
10498 requests Squid will try to handle in parallel. If set to N, Squid
10499 will try to receive and process up to 1+N requests on the same
10500 connection concurrently.
10502 Defaults to 0 (off) for bandwidth management and access logging
10505 NOTE: pipelining requires persistent connections to clients.
10507 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
10510 NAME: high_response_time_warning
10513 LOC: Config.warnings.high_rptm
10515 DEFAULT_DOC: disabled.
10517 If the one-minute median response time exceeds this value,
10518 Squid prints a WARNING with debug level 0 to get the
10519 administrators attention. The value is in milliseconds.
10522 NAME: high_page_fault_warning
10524 LOC: Config.warnings.high_pf
10526 DEFAULT_DOC: disabled.
10528 If the one-minute average page fault rate exceeds this
10529 value, Squid prints a WARNING with debug level 0 to get
10530 the administrators attention. The value is in page faults
10534 NAME: high_memory_warning
10536 LOC: Config.warnings.high_memory
10537 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
10539 DEFAULT_DOC: disabled.
10541 If the memory usage (as determined by gnumalloc, if available and used)
10542 exceeds this amount, Squid prints a WARNING with debug level 0 to get
10543 the administrators attention.
10545 # TODO: link high_memory_warning to mempools?
10547 NAME: sleep_after_fork
10548 COMMENT: (microseconds)
10550 LOC: Config.sleep_after_fork
10553 When this is set to a non-zero value, the main Squid process
10554 sleeps the specified number of microseconds after a fork()
10555 system call. This sleep may help the situation where your
10556 system reports fork() failures due to lack of (virtual)
10557 memory. Note, however, if you have a lot of child
10558 processes, these sleep delays will add up and your
10559 Squid will not service requests for some amount of time
10560 until all the child processes have been started.
10561 On Windows value less then 1000 (1 milliseconds) are
10565 NAME: windows_ipaddrchangemonitor
10566 IFDEF: _SQUID_WINDOWS_
10570 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
10572 On Windows Squid by default will monitor IP address changes and will
10573 reconfigure itself after any detected event. This is very useful for
10574 proxies connected to internet with dial-up interfaces.
10575 In some cases (a Proxy server acting as VPN gateway is one) it could be
10576 desiderable to disable this behaviour setting this to 'off'.
10577 Note: after changing this, Squid service must be restarted.
10582 IFDEF: USE_SQUID_EUI
10584 LOC: Eui::TheConfig.euiLookup
10586 Whether to lookup the EUI or MAC address of a connected client.
10589 NAME: max_filedescriptors max_filedesc
10592 DEFAULT_DOC: Use operating system soft limit set by ulimit.
10593 LOC: Config.max_filedescriptors
10595 Set the maximum number of filedescriptors, either below the
10596 operating system default or up to the hard limit.
10598 Remove from squid.conf to inherit the current ulimit soft
10601 Note: Changing this requires a restart of Squid. Also
10602 not all I/O types supports large values (eg on Windows).
10605 NAME: force_request_body_continuation
10607 LOC: Config.accessList.forceRequestBodyContinuation
10609 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
10611 This option controls how Squid handles data upload requests from HTTP
10612 and FTP agents that require a "Please Continue" control message response
10613 to actually send the request body to Squid. It is mostly useful in
10614 adaptation environments.
10616 When Squid receives an HTTP request with an "Expect: 100-continue"
10617 header or an FTP upload command (e.g., STOR), Squid normally sends the
10618 request headers or FTP command information to an adaptation service (or
10619 peer) and waits for a response. Most adaptation services (and some
10620 broken peers) may not respond to Squid at that stage because they may
10621 decide to wait for the HTTP request body or FTP data transfer. However,
10622 that request body or data transfer may never come because Squid has not
10623 responded with the HTTP 100 or FTP 150 (Please Continue) control message
10624 to the request sender yet!
10626 An allow match tells Squid to respond with the HTTP 100 or FTP 150
10627 (Please Continue) control message on its own, before forwarding the
10628 request to an adaptation service or peer. Such a response usually forces
10629 the request sender to proceed with sending the body. A deny match tells
10630 Squid to delay that control response until the origin server confirms
10631 that the request body is needed. Delaying is the default behavior.
10634 NAME: http_upgrade_request_protocols
10635 TYPE: http_upgrade_request_protocols
10636 LOC: Config.http_upgrade_request_protocols
10638 DEFAULT_DOC: Upgrade header dropped, effectively blocking an upgrade attempt.
10640 Controls client-initiated and server-confirmed switching from HTTP to
10641 another protocol (or to several protocols) using HTTP Upgrade mechanism
10642 defined in RFC 7230 Section 6.7. Squid itself does not understand the
10643 protocols being upgraded to and participates in the upgraded
10644 communication only as a dumb TCP proxy. Admins should not allow
10645 upgrading to protocols that require a more meaningful proxy
10648 Usage: http_upgrade_request_protocols <protocol> allow|deny [!]acl ...
10650 The required "protocol" parameter is either an all-caps word OTHER or an
10651 explicit protocol name (e.g. "WebSocket") optionally followed by a slash
10652 and a version token (e.g. "HTTP/3"). Explicit protocol names and
10653 versions are case sensitive.
10655 When an HTTP client sends an Upgrade request header, Squid iterates over
10656 the client-offered protocols and, for each protocol P (with an optional
10657 version V), evaluates the first non-empty set of
10658 http_upgrade_request_protocols rules (if any) from the following list:
10660 * All rules with an explicit protocol name equal to P.
10661 * All rules that use OTHER instead of a protocol name.
10663 In other words, rules using OTHER are considered for protocol P if and
10664 only if there are no rules mentioning P by name.
10666 If both of the above sets are empty, then Squid removes protocol P from
10669 If the client sent a versioned protocol offer P/X, then explicit rules
10670 referring to the same-name but different-version protocol P/Y are
10671 declared inapplicable. Inapplicable rules are not evaluated (i.e. are
10672 ignored). However, inapplicable rules still belong to the first set of
10675 Within the applicable rule subset, individual rules are evaluated in
10676 their configuration order. If all ACLs of an applicable "allow" rule
10677 match, then the protocol offered by the client is forwarded to the next
10678 hop as is. If all ACLs of an applicable "deny" rule match, then the
10679 offer is dropped. If no applicable rules have matching ACLs, then the
10680 offer is also dropped. The first matching rule also ends rules
10681 evaluation for the offered protocol.
10683 If all client-offered protocols are removed, then Squid forwards the
10684 client request without the Upgrade header. Squid never sends an empty
10685 Upgrade request header.
10687 An Upgrade request header with a value violating HTTP syntax is dropped
10688 and ignored without an attempt to use extractable individual protocol
10691 Upon receiving an HTTP 101 (Switching Protocols) control message, Squid
10692 checks that the server listed at least one protocol name and sent a
10693 Connection:upgrade response header. Squid does not understand individual
10694 protocol naming and versioning concepts enough to implement stricter
10695 checks, but an admin can restrict HTTP 101 (Switching Protocols)
10696 responses further using http_reply_access. Responses denied by
10697 http_reply_access rules and responses flagged by the internal Upgrade
10698 checks result in HTTP 502 (Bad Gateway) ERR_INVALID_RESP errors and
10699 Squid-to-server connection closures.
10701 If Squid sends an Upgrade request header, and the next hop (e.g., the
10702 origin server) responds with an acceptable HTTP 101 (Switching
10703 Protocols), then Squid forwards that message to the client and becomes
10706 The presence of an Upgrade request header alone does not preclude cache
10707 lookups. In other words, an Upgrade request might be satisfied from the
10708 cache, using regular HTTP caching rules.
10710 This clause only supports fast acl types.
10711 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
10713 Each of the following groups of configuration lines represents a
10714 separate configuration example:
10716 # never upgrade to protocol Foo; all others are OK
10717 http_upgrade_request_protocols Foo deny all
10718 http_upgrade_request_protocols OTHER allow all
10720 # only allow upgrades to protocol Bar (except for its first version)
10721 http_upgrade_request_protocols Bar/1 deny all
10722 http_upgrade_request_protocols Bar allow all
10723 http_upgrade_request_protocols OTHER deny all # this rule is optional
10725 # only allow upgrades to protocol Baz, and only if Baz is the only offer
10726 acl UpgradeHeaderHasMultipleOffers ...
10727 http_upgrade_request_protocols Baz deny UpgradeHeaderHasMultipleOffers
10728 http_upgrade_request_protocols Baz allow all
10731 NAME: server_pconn_for_nonretriable
10734 DEFAULT_DOC: Open new connections for forwarding requests Squid cannot retry safely.
10735 LOC: Config.accessList.serverPconnForNonretriable
10737 This option provides fine-grained control over persistent connection
10738 reuse when forwarding HTTP requests that Squid cannot retry. It is useful
10739 in environments where opening new connections is very expensive
10740 (e.g., all connections are secured with TLS with complex client and server
10741 certificate validation) and race conditions associated with persistent
10742 connections are very rare and/or only cause minor problems.
10744 HTTP prohibits retrying unsafe and non-idempotent requests (e.g., POST).
10745 Squid limitations also prohibit retrying all requests with bodies (e.g., PUT).
10746 By default, when forwarding such "risky" requests, Squid opens a new
10747 connection to the server or cache_peer, even if there is an idle persistent
10748 connection available. When Squid is configured to risk sending a non-retriable
10749 request on a previously used persistent connection, and the server closes
10750 the connection before seeing that risky request, the user gets an error response
10751 from Squid. In most cases, that error response will be HTTP 502 (Bad Gateway)
10752 with ERR_ZERO_SIZE_OBJECT or ERR_WRITE_ERROR (peer connection reset) error detail.
10754 If an allow rule matches, Squid reuses an available idle persistent connection
10755 (if any) for the request that Squid cannot retry. If a deny rule matches, then
10756 Squid opens a new connection for the request that Squid cannot retry.
10758 This option does not affect requests that Squid can retry. They will reuse idle
10759 persistent connections (if any).
10761 This clause only supports fast acl types.
10762 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
10765 acl SpeedIsWorthTheRisk method POST
10766 server_pconn_for_nonretriable allow SpeedIsWorthTheRisk
10769 NAME: happy_eyeballs_connect_timeout
10773 LOC: Config.happyEyeballs.connect_timeout
10775 This Happy Eyeballs (RFC 8305) tuning directive specifies the minimum
10776 delay between opening a primary to-server connection and opening a
10777 spare to-server connection for the same master transaction. This delay
10778 is similar to the Connection Attempt Delay in RFC 8305, but it is only
10779 applied to the first spare connection attempt. Subsequent spare
10780 connection attempts use happy_eyeballs_connect_gap, and primary
10781 connection attempts are not artificially delayed at all.
10783 Terminology: The "primary" and "spare" designations are determined by
10784 the order of DNS answers received by Squid: If Squid DNS AAAA query
10785 was answered first, then primary connections are connections to IPv6
10786 peer addresses (while spare connections use IPv4 addresses).
10787 Similarly, if Squid DNS A query was answered first, then primary
10788 connections are connections to IPv4 peer addresses (while spare
10789 connections use IPv6 addresses).
10791 Shorter happy_eyeballs_connect_timeout values reduce master
10792 transaction response time, potentially improving user-perceived
10793 response times (i.e., making user eyeballs happier). Longer delays
10794 reduce both concurrent connection level and server bombardment with
10795 connection requests, potentially improving overall Squid performance
10796 and reducing the chance of being blocked by servers for opening too
10797 many unused connections.
10799 RFC 8305 prohibits happy_eyeballs_connect_timeout values smaller than
10800 10 (milliseconds) to "avoid congestion collapse in the presence of
10801 high packet-loss rates".
10803 The following Happy Eyeballs directives place additional connection
10804 opening restrictions: happy_eyeballs_connect_gap and
10805 happy_eyeballs_connect_limit.
10808 NAME: happy_eyeballs_connect_gap
10812 DEFAULT_DOC: no artificial delays between spare attempts
10813 LOC: Config.happyEyeballs.connect_gap
10815 This Happy Eyeballs (RFC 8305) tuning directive specifies the
10816 minimum delay between opening spare to-server connections (to any
10817 server; i.e. across all concurrent master transactions in a Squid
10818 instance). Each SMP worker currently multiplies the configured gap
10819 by the total number of workers so that the combined spare connection
10820 opening rate of a Squid instance obeys the configured limit. The
10821 workers do not coordinate connection openings yet; a micro burst
10822 of spare connection openings may violate the configured gap.
10824 This directive has similar trade-offs as
10825 happy_eyeballs_connect_timeout, but its focus is on limiting traffic
10826 amplification effects for Squid as a whole, while
10827 happy_eyeballs_connect_timeout works on an individual master
10830 The following Happy Eyeballs directives place additional connection
10831 opening restrictions: happy_eyeballs_connect_timeout and
10832 happy_eyeballs_connect_limit. See the former for related terminology.
10835 NAME: happy_eyeballs_connect_limit
10838 DEFAULT_DOC: no artificial limit on the number of concurrent spare attempts
10839 LOC: Config.happyEyeballs.connect_limit
10841 This Happy Eyeballs (RFC 8305) tuning directive specifies the
10842 maximum number of spare to-server connections (to any server; i.e.
10843 across all concurrent master transactions in a Squid instance).
10844 Each SMP worker gets an equal share of the total limit. However,
10845 the workers do not share the actual connection counts yet, so one
10846 (busier) worker cannot "borrow" spare connection slots from another
10847 (less loaded) worker.
10849 Setting this limit to zero disables concurrent use of primary and
10850 spare TCP connections: Spare connection attempts are made only after
10851 all primary attempts fail. However, Squid would still use the
10852 DNS-related optimizations of the Happy Eyeballs approach.
10854 This directive has similar trade-offs as happy_eyeballs_connect_gap,
10855 but its focus is on limiting Squid overheads, while
10856 happy_eyeballs_connect_gap focuses on the origin server and peer
10859 The following Happy Eyeballs directives place additional connection
10860 opening restrictions: happy_eyeballs_connect_timeout and
10861 happy_eyeballs_connect_gap. See the former for related terminology.