1 ## Copyright (C) 1996-2023 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 https://wiki.squid-cache.org/SquidFaq
20 https://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. For a
592 given transaction, (re)authentication is requested in two primary cases
593 detailed below: initial authentication and re-authentication.
595 A client without credentials is requested to authenticate if one of the
596 following ACLs is evaluated by an http_access or adapted_http_access rule:
599 * proxy_auth_regex ACL
601 * external ACL with %ul logformat %code used in FORMAT parameters
602 * external ACL with %LOGIN macro used in FORMAT parameters;
603 this legacy macro currently behaves the same as %ul logformat %code
605 A client with credentials is requested to re-authenticate if http_access
606 or adapted_http_access denies its request _and_ the last evaluated ACL was
607 either proxy_auth, proxy_auth_regex, ident, ident_regex, or an external
608 ACL with %ul or %LOGIN parameter (regardless of whether that last
609 evaluated ACL matched the denied request). Note that a max_user_ip ACL
610 does not have this effect: Requests denied after evaluating max_user_ip
611 trigger an HTTP 403 (Forbidden) response rather than re-authentication.
613 In both initial authentication and re-authentication cases, client access
614 is denied, typically with an HTTP 407 (Proxy Authentication Required) or
615 an HTTP 401 (Unauthorized) response.
617 WARNING: authentication can't be used in a transparently intercepting
618 proxy as the client then thinks it is talking to an origin server and
619 not the proxy. This is a limitation of bending the TCP/IP protocol to
620 transparently intercepting port 80, not a limitation in Squid.
621 Ports flagged 'transparent', 'intercept', or 'tproxy' have
622 authentication disabled.
624 === Parameters common to all schemes. ===
627 Specifies the command for the external authenticator.
629 By default, each authentication scheme is not used unless a
630 program is specified.
632 See https://wiki.squid-cache.org/Features/AddonHelpers for
633 more details on helper operations and creating your own.
636 Specifies a string to be append to request line format for
637 the authentication helper. "Quoted" format values may contain
638 spaces and logformat %macros. In theory, any logformat %macro
639 can be used. In practice, a %macro expands as a dash (-) if
640 the helper request is sent before the required macro
641 information is available to Squid.
643 By default, Squid uses request formats provided in
644 scheme-specific examples below (search for %credentials).
646 The expanded key_extras value is added to the Squid credentials
647 cache and, hence, will affect authentication. It can be used to
648 authenticate different users with identical user names (e.g.,
649 when user authentication depends on http_port).
651 Avoid adding frequently changing information to key_extras. For
652 example, if you add user source IP, and it changes frequently
653 in your environment, then max_user_ip ACL is going to treat
654 every user+IP combination as a unique "user", breaking the ACL
655 and wasting a lot of memory on those user records. It will also
656 force users to authenticate from scratch whenever their IP
660 Specifies the protection scope (aka realm name) which is to be
661 reported to the client for the authentication scheme. It is
662 commonly part of the text the user will see when prompted for
663 their username and password.
665 For Basic the default is "Squid proxy-caching web server".
666 For Digest there is no default, this parameter is mandatory.
667 For NTLM and Negotiate this parameter is ignored.
669 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
670 [queue-size=N] [on-persistent-overload=action]
671 [reservation-timeout=seconds]
673 The maximum number of authenticator processes to spawn. If
674 you start too few Squid will have to wait for them to process
675 a backlog of credential verifications, slowing it down. When
676 password verifications are done via a (slow) network you are
677 likely to need lots of authenticator processes.
679 The startup= and idle= options permit some skew in the exact
680 amount run. A minimum of startup=N will begin during startup
681 and reconfigure. Squid will start more in groups of up to
682 idle=N in an attempt to meet traffic needs and to keep idle=N
683 free above those traffic needs up to the maximum.
685 The concurrency= option sets the number of concurrent requests
686 the helper can process. The default of 0 is used for helpers
687 who only supports one request at a time. Setting this to a
688 number greater than 0 changes the protocol used to include a
689 channel ID field first on the request/response line, allowing
690 multiple requests to be sent to the same helper in parallel
691 without waiting for the response.
693 Concurrency must not be set unless it's known the helper
694 supports the input format with channel-ID fields.
696 The queue-size option sets the maximum number of queued
697 requests. A request is queued when no existing child can
698 accept it due to concurrency limit and no new child can be
699 started due to numberofchildren limit. The default maximum is
700 2*numberofchildren. Squid is allowed to temporarily exceed the
701 configured maximum, marking the affected helper as
702 "overloaded". If the helper overload lasts more than 3
703 minutes, the action prescribed by the on-persistent-overload
706 The on-persistent-overload=action option specifies Squid
707 reaction to a new helper request arriving when the helper
708 has been overloaded for more that 3 minutes already. The number
709 of queued requests determines whether the helper is overloaded
710 (see the queue-size option).
712 Two actions are supported:
714 die Squid worker quits. This is the default behavior.
716 ERR Squid treats the helper request as if it was
717 immediately submitted, and the helper immediately
718 replied with an ERR response. This action has no effect
719 on the already queued and in-progress helper requests.
721 NOTE: NTLM and Negotiate schemes do not support concurrency
722 in the Squid code module even though some helpers can.
724 The reservation-timeout=seconds option allows NTLM and Negotiate
725 helpers to forget about clients that abandon their in-progress
726 connection authentication without closing the connection. The
727 timeout is measured since the last helper response received by
728 Squid for the client. Fractional seconds are not supported.
730 After the timeout, the helper will be used for other clients if
731 there are no unreserved helpers available. In the latter case,
732 the old client attempt to resume authentication will not be
733 forwarded to the helper (and the client should open a new HTTP
734 connection and retry authentication from scratch).
736 By default, reservations do not expire and clients that keep
737 their connections open without completing authentication may
738 exhaust all NTLM and Negotiate helpers.
741 If you experience problems with PUT/POST requests when using
742 the NTLM or Negotiate schemes then you can try setting this
743 to off. This will cause Squid to forcibly close the connection
744 on the initial request where the browser asks which schemes
745 are supported by the proxy.
747 For Basic and Digest this parameter is ignored.
750 Useful for sending credentials to authentication backends that
751 expect UTF-8 encoding (e.g., LDAP).
753 When this option is enabled, Squid uses HTTP Accept-Language
754 request header to guess the received credentials encoding
755 (ISO-Latin-1, CP1251, or UTF-8) and then converts the first
756 two encodings into UTF-8.
758 When this option is disabled and by default, Squid sends
759 credentials in their original (i.e. received) encoding.
761 This parameter is only honored for Basic and Digest schemes.
762 For Basic, the entire username:password credentials are
763 checked and, if necessary, re-encoded. For Digest -- just the
764 username component. For NTLM and Negotiate schemes, this
765 parameter is ignored.
767 IF HAVE_AUTH_MODULE_BASIC
768 === Basic authentication parameters ===
770 "credentialsttl" timetolive
771 Specifies how long squid assumes an externally validated
772 username:password pair is valid for - in other words how
773 often the helper program is called for that user. Set this
774 low to force revalidation with short lived passwords.
776 NOTE: setting this high does not impact your susceptibility
777 to replay attacks unless you are using an one-time password
778 system (such as SecureID). If you are using such a system,
779 you will be vulnerable to replay attacks unless you also
780 use the max_user_ip ACL in an http_access rule.
782 "casesensitive" on|off
783 Specifies whether upper case letters in client-sent usernames are
784 preserved. By default and when explicitly set to "off", a username
785 extracted from Proxy-Authorization or Authorization request header is
786 forced to lower case before user credentials are checked or stored.
788 Most user databases are case insensitive, allowing the same username to be
789 spelled using both lower and upper case letters. For such databases, any
790 casesenstive setting should work, but forcing usernames to lower case may
791 still make a big difference for Squid internal caches like those used by
792 an external ACL with %un logformat code in FORMAT and a user_max_ip ACL.
794 When working with a case sensitive database, set casesensitive to "on".
796 Squid ACLs like proxy_auth are case-sensitive by default. An ACL using
797 upper case letters in user names (e.g., `acl badGuys proxy_auth Bob`)
798 will not match any user with Basic Authentication credentials unless
799 casesensitive is explicitly turned "on" (to preserve "Bob" username
800 instead of converting it to "bob" before the ACL is checked).
803 IF HAVE_AUTH_MODULE_DIGEST
804 === Digest authentication parameters ===
806 "nonce_garbage_interval" timeinterval
807 Specifies the interval that nonces that have been issued
808 to client_agent's are checked for validity.
810 "nonce_max_duration" timeinterval
811 Specifies the maximum length of time a given nonce will be
814 "nonce_max_count" number
815 Specifies the maximum number of times a given nonce can be
818 "nonce_strictness" on|off
819 Determines if squid requires strict increment-by-1 behavior
820 for nonce counts, or just incrementing (off - for use when
821 user agents generate nonce counts that occasionally miss 1
822 (ie, 1,2,4,6)). Default off.
824 "check_nonce_count" on|off
825 This directive if set to off can disable the nonce count check
826 completely to work around buggy digest qop implementations in
827 certain mainstream browser versions. Default on to check the
828 nonce count to protect from authentication replay attacks.
830 "post_workaround" on|off
831 This is a workaround to certain buggy browsers who send an
832 incorrect request digest in POST requests when reusing the
833 same nonce as acquired earlier on a GET request.
837 === Example Configuration ===
839 This configuration displays the recommended authentication scheme
840 order from most to least secure with recommended minimum configuration
841 settings for each scheme:
843 #auth_param negotiate program <uncomment and complete this line to activate>
844 #auth_param negotiate children 20 startup=0 idle=1
846 #auth_param digest program <uncomment and complete this line to activate>
847 #auth_param digest children 20 startup=0 idle=1
848 #auth_param digest realm Squid proxy-caching web server
849 #auth_param digest nonce_garbage_interval 5 minutes
850 #auth_param digest nonce_max_duration 30 minutes
851 #auth_param digest nonce_max_count 50
853 #auth_param ntlm program <uncomment and complete this line to activate>
854 #auth_param ntlm children 20 startup=0 idle=1
856 #auth_param basic program <uncomment and complete this line>
857 #auth_param basic children 5 startup=5 idle=1
858 #auth_param basic credentialsttl 2 hours
861 NAME: authenticate_cache_garbage_interval
865 LOC: Auth::TheConfig.garbageCollectInterval
867 The time period between garbage collection across the username cache.
868 This is a trade-off between memory utilization (long intervals - say
869 2 days) and CPU (short intervals - say 1 minute). Only change if you
873 NAME: authenticate_ttl
877 LOC: Auth::TheConfig.credentialsTtl
879 The time a user & their credentials stay in the logged in
880 user cache since their last request. When the garbage
881 interval passes, all user credentials that have passed their
882 TTL are removed from memory.
885 NAME: authenticate_ip_ttl
888 LOC: Auth::TheConfig.ipTtl
891 If you use proxy authentication and the 'max_user_ip' ACL,
892 this directive controls how long Squid remembers the IP
893 addresses associated with each user. Use a small value
894 (e.g., 60 seconds) if your users might change addresses
895 quickly, as is the case with dialup. You might be safe
896 using a larger value (e.g., 2 hours) in a corporate LAN
897 environment with relatively static address assignments.
902 -----------------------------------------------------------------------------
905 NAME: external_acl_type
906 TYPE: externalAclHelper
907 LOC: Config.externalAclHelperList
910 This option defines external acl classes using a helper program
911 to look up the status
913 external_acl_type name [options] FORMAT /path/to/helper [helper arguments]
917 ttl=n TTL in seconds for cached results (defaults to 3600
921 TTL for cached negative lookups (default same
924 grace=n Percentage remaining of TTL where a refresh of a
925 cached entry should be initiated without needing to
926 wait for a new reply. (default is for no grace period)
928 cache=n The maximum number of entries in the result cache. The
929 default limit is 262144 entries. Each cache entry usually
930 consumes at least 256 bytes. Squid currently does not remove
931 expired cache entries until the limit is reached, so a proxy
932 will sooner or later reach the limit. The expanded FORMAT
933 value is used as the cache key, so if the details in FORMAT
934 are highly variable, a larger cache may be needed to produce
935 reduction in helper load.
938 Maximum number of acl helper processes spawned to service
939 external acl lookups of this type. (default 5)
942 Minimum number of acl helper processes to spawn during
943 startup and reconfigure to service external acl lookups
944 of this type. (default 0)
947 Number of acl helper processes to keep ahead of traffic
948 loads. Squid will spawn this many at once whenever load
949 rises above the capabilities of existing processes.
950 Up to the value of children-max. (default 1)
952 concurrency=n concurrency level per process. Only used with helpers
953 capable of processing more than one query at a time.
955 queue-size=N The queue-size option sets the maximum number of
956 queued requests. A request is queued when no existing
957 helper can accept it due to concurrency limit and no
958 new helper can be started due to children-max limit.
959 If the queued requests exceed queue size, the acl is
960 ignored. The default value is set to 2*children-max.
962 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
964 ipv4 / ipv6 IP protocol used to communicate with this helper.
965 The default is to auto-detect IPv6 and use it when available.
968 FORMAT is a series of %macro codes. See logformat directive for a full list
969 of the accepted codes. Although note that at the time of any external ACL
970 being tested data may not be available and thus some %macro expand to '-'.
972 In addition to the logformat codes; when processing external ACLs these
973 additional macros are made available:
975 %ACL The name of the ACL being tested.
977 %DATA The ACL arguments specified in the referencing config
978 'acl ... external' line, separated by spaces (an
979 "argument string"). see acl external.
981 If there are no ACL arguments %DATA expands to '-'.
983 If you do not specify a DATA macro inside FORMAT,
984 Squid automatically appends %DATA to your FORMAT.
985 Note that Squid-3.x may expand %DATA to whitespace
986 or nothing in this case.
988 By default, Squid applies URL-encoding to each ACL
989 argument inside the argument string. If an explicit
990 encoding modifier is used (e.g., %#DATA), then Squid
991 encodes the whole argument string as a single token
992 (e.g., with %#DATA, spaces between arguments become
995 If SSL is enabled, the following formatting codes become available:
997 %USER_CERT SSL User certificate in PEM format
998 %USER_CERTCHAIN SSL User certificate chain in PEM format
999 %USER_CERT_xx SSL User certificate subject attribute xx
1000 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
1003 NOTE: all other format codes accepted by older Squid versions
1007 General request syntax:
1009 [channel-ID] FORMAT-values
1012 FORMAT-values consists of transaction details expanded with
1013 whitespace separation per the config file FORMAT specification
1014 using the FORMAT macros listed above.
1016 Request values sent to the helper are URL escaped to protect
1017 each value in requests against whitespaces.
1019 If using protocol=2.5 then the request sent to the helper is not
1020 URL escaped to protect against whitespace.
1022 NOTE: protocol=3.0 is deprecated as no longer necessary.
1024 When using the concurrency= option the protocol is changed by
1025 introducing a query channel tag in front of the request/response.
1026 The query channel tag is a number between 0 and concurrency-1.
1027 This value must be echoed back unchanged to Squid as the first part
1028 of the response relating to its request.
1031 The helper receives lines expanded per the above format specification
1032 and for each input line returns 1 line starting with OK/ERR/BH result
1033 code and optionally followed by additional keywords with more details.
1036 General result syntax:
1038 [channel-ID] result keyword=value ...
1040 Result consists of one of the codes:
1043 the ACL test produced a match.
1046 the ACL test does not produce a match.
1049 An internal error occurred in the helper, preventing
1050 a result being identified.
1052 The meaning of 'a match' is determined by your squid.conf
1053 access control configuration. See the Squid wiki for details.
1057 user= The users name (login)
1059 password= The users password (for login= cache_peer option)
1061 message= Message describing the reason for this response.
1062 Available as %o in error pages.
1063 Useful on (ERR and BH results).
1065 tag= Apply a tag to a request. Only sets a tag once,
1066 does not alter existing tags.
1068 log= String to be logged in access.log. Available as
1069 %ea in logformat specifications.
1071 clt_conn_tag= Associates a TAG with the client TCP connection.
1072 Please see url_rewrite_program related documentation
1075 Any keywords may be sent on any response whether OK, ERR or BH.
1077 All response keyword values need to be a single token with URL
1078 escaping, or enclosed in double quotes (") and escaped using \ on
1079 any double quotes or \ characters within the value. The wrapping
1080 double quotes are removed before the value is interpreted by Squid.
1081 \r and \n are also replace by CR and LF.
1083 Some example key values:
1087 user="J. \"Bob\" Smith"
1094 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
1095 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
1096 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
1097 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
1098 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
1100 DEFAULT: all src all
1101 DEFAULT: manager url_regex +i ^[^:]+://[^/]+/squid-internal-mgr/
1102 DEFAULT: localhost src 127.0.0.1/32 ::1
1103 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1/128 ::/128
1104 DEFAULT: to_linklocal dst 169.254.0.0/16 fe80::/10
1105 DEFAULT: CONNECT method CONNECT
1106 DEFAULT_DOC: ACLs all, manager, localhost, to_localhost, to_linklocal, and CONNECT are predefined.
1108 Defining an Access List
1110 Every access list definition must begin with an aclname and acltype,
1111 followed by either type-specific arguments or a quoted filename that
1114 acl aclname acltype argument ...
1115 acl aclname acltype "file" ...
1117 When using "file", the file should contain one item per line.
1122 Some acl types supports options which changes their default behaviour:
1124 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
1125 case-insensitive, use the -i option. To return case-sensitive
1126 use the +i option between patterns, or make a new ACL line
1129 -n Disable lookups and address type conversions. If lookup or
1130 conversion is required because the parameter type (IP or
1131 domain name) does not match the message address type (domain
1132 name or IP), then the ACL would immediately declare a mismatch
1133 without any warnings or lookups.
1136 Perform a list membership test, interpreting values as
1137 comma-separated token lists and matching against individual
1138 tokens instead of whole values.
1139 The optional "delimiters" parameter specifies one or more
1140 alternative non-alphanumeric delimiter characters.
1141 non-alphanumeric delimiter characters.
1143 -- Used to stop processing all options, in the case the first acl
1144 value has '-' character as first character (for example the '-'
1145 is a valid domain name)
1147 Some acl types require suspending the current request in order
1148 to access some external data source.
1149 Those which do are marked with the tag [slow], those which
1150 don't are marked as [fast].
1151 See https://wiki.squid-cache.org/SquidFaq/SquidAcl
1152 for further information
1154 ***** ACL TYPES AVAILABLE *****
1156 acl aclname src ip-address/mask ... # clients IP address [fast]
1157 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
1158 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
1159 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
1162 acl aclname arp mac-address ...
1163 acl aclname eui64 eui64-address ...
1165 # MAC (EUI-48) and EUI-64 addresses use xx:xx:xx:xx:xx:xx notation.
1167 # The 'arp' ACL code is not portable to all operating systems.
1168 # It works on Linux, Solaris, Windows, FreeBSD, and some other
1171 # The eui_lookup directive is required to be 'on' (the default)
1172 # and Squid built with --enable-eui for MAC/EUI addresses to be
1173 # available for this ACL.
1175 # Squid can only determine the MAC/EUI address for IPv4
1176 # clients that are on the same subnet. If the client is on a
1177 # different subnet, then Squid cannot find out its address.
1179 # IPv6 protocol does not contain ARP. MAC/EUI is either
1180 # encoded directly in the IPv6 address or not available.
1182 acl aclname clientside_mark mark[/mask] ...
1183 # matches CONNMARK of an accepted connection [fast]
1184 # DEPRECATED. Use the 'client_connection_mark' instead.
1186 acl aclname client_connection_mark mark[/mask] ...
1187 # matches CONNMARK of an accepted connection [fast]
1189 # mark and mask are unsigned integers (hex, octal, or decimal).
1190 # If multiple marks are given, then the ACL matches if at least
1193 # Uses netfilter-conntrack library.
1194 # Requires building Squid with --enable-linux-netfilter.
1196 # The client, various intermediaries, and Squid itself may set
1197 # CONNMARK at various times. The last CONNMARK set wins. This ACL
1198 # checks the mark present on an accepted connection or set by
1199 # Squid afterwards, depending on the ACL check timing. This ACL
1200 # effectively ignores any mark set by other agents after Squid has
1201 # accepted the connection.
1203 acl aclname srcdomain .foo.com ...
1204 # reverse lookup, from client IP [slow]
1205 acl aclname dstdomain [-n] .foo.com ...
1206 # Destination server from URL [fast]
1207 acl aclname srcdom_regex [-i] \.foo\.com ...
1208 # regex matching client name [slow]
1209 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
1210 # regex matching server [fast]
1212 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
1213 # based URL is used and no match is found. The name "none" is used
1214 # if the reverse lookup fails.
1216 acl aclname src_as number ...
1217 acl aclname dst_as number ...
1219 # Except for access control, AS numbers can be used for
1220 # routing of requests to specific caches. Here's an
1221 # example for routing all requests for AS#1241 and only
1222 # those to mycache.mydomain.net:
1223 # acl asexample dst_as 1241
1224 # cache_peer_access mycache.mydomain.net allow asexample
1225 # cache_peer_access mycache_mydomain.net deny all
1227 acl aclname peername myPeer ...
1228 acl aclname peername_regex [-i] regex-pattern ...
1230 # match against a named cache_peer entry
1231 # set unique name= on cache_peer lines for reliable use.
1233 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
1243 # h1:m1 must be less than h2:m2
1245 acl aclname url_regex [-i] ^http:// ...
1246 # regex matching on whole URL [fast]
1247 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
1248 # regex matching on URL login field
1249 acl aclname urlpath_regex [-i] \.gif$ ...
1250 # regex matching on URL path [fast]
1252 acl aclname port 80 70 21 0-1024 ...
1253 # destination TCP port (or port range) of the request [fast]
1255 # Port 0 matches requests that have no explicit and no default destination
1256 # ports (e.g., HTTP requests with URN targets)
1258 acl aclname localport 3128 ... # TCP port the client connected to [fast]
1259 # NP: for interception mode this is usually '80'
1261 acl aclname myportname 3128 ... # *_port name [fast]
1263 acl aclname proto HTTP FTP ... # request protocol [fast]
1265 acl aclname method GET POST ... # HTTP request method [fast]
1267 acl aclname http_status 200 301 500- 400-403 ...
1268 # status code in reply [fast]
1270 acl aclname browser [-i] regexp ...
1271 # pattern match on User-Agent header (see also req_header below) [fast]
1273 acl aclname referer_regex [-i] regexp ...
1274 # pattern match on Referer header [fast]
1275 # Referer is highly unreliable, so use with care
1277 acl aclname ident [-i] username ...
1278 acl aclname ident_regex [-i] pattern ...
1279 # string match on ident output [slow]
1280 # use REQUIRED to accept any non-null ident.
1282 acl aclname proxy_auth [-i] username ...
1283 acl aclname proxy_auth_regex [-i] pattern ...
1284 # perform http authentication challenge to the client and match against
1285 # supplied credentials [slow]
1287 # takes a list of allowed usernames.
1288 # use REQUIRED to accept any valid username.
1290 # Will use proxy authentication in forward-proxy scenarios, and plain
1291 # http authentication in reverse-proxy scenarios
1293 # NOTE: when a Proxy-Authentication header is sent but it is not
1294 # needed during ACL checking the username is NOT logged
1297 # NOTE: proxy_auth requires a EXTERNAL authentication program
1298 # to check username/password combinations (see
1299 # auth_param directive).
1301 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1302 # as the browser needs to be configured for using a proxy in order
1303 # to respond to proxy authentication.
1305 acl aclname snmp_community string ...
1306 # A community string to limit access to your SNMP Agent [fast]
1309 # acl snmppublic snmp_community public
1311 acl aclname maxconn number
1312 # This will be matched when the client's IP address has
1313 # more than <number> TCP connections established. [fast]
1314 # NOTE: This only measures direct TCP links so X-Forwarded-For
1315 # indirect clients are not counted.
1317 acl aclname max_user_ip [-s] number
1318 # This will be matched when the user attempts to log in from more
1319 # than <number> different ip addresses. The authenticate_ip_ttl
1320 # parameter controls the timeout on the ip entries. [fast]
1321 # If -s is specified the limit is strict, denying browsing
1322 # from any further IP addresses until the ttl has expired. Without
1323 # -s Squid will just annoy the user by "randomly" denying requests.
1324 # (the counter is reset each time the limit is reached and a
1325 # request is denied)
1326 # NOTE: in acceleration mode or where there is mesh of child proxies,
1327 # clients may appear to come from multiple addresses if they are
1328 # going through proxy farms, so a limit of 1 may cause user problems.
1330 acl aclname random probability
1331 # Pseudo-randomly match requests. Based on the probability given.
1332 # Probability may be written as a decimal (0.333), fraction (1/3)
1333 # or ratio of matches:non-matches (3:5).
1335 acl aclname req_mime_type [-i] mime-type ...
1336 # regex match against the mime type of the request generated
1337 # by the client. Can be used to detect file upload or some
1338 # types HTTP tunneling requests [fast]
1339 # NOTE: This does NOT match the reply. You cannot use this
1340 # to match the returned file type.
1342 acl aclname req_header header-name [-i] any\.regex\.here
1343 # regex match against any of the known request headers. May be
1344 # thought of as a superset of "browser", "referer" and "mime-type"
1347 acl aclname rep_mime_type [-i] mime-type ...
1348 # regex match against the mime type of the reply received by
1349 # squid. Can be used to detect file download or some
1350 # types HTTP tunneling requests. [fast]
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 rep_header header-name [-i] any\.regex\.here
1356 # regex match against any of the known reply headers. May be
1357 # thought of as a superset of "browser", "referer" and "mime-type"
1360 acl aclname external class_name [arguments...]
1361 # external ACL lookup via a helper class defined by the
1362 # external_acl_type directive [slow]
1364 acl aclname user_cert attribute values...
1365 # match against attributes in a user SSL certificate
1366 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1368 acl aclname ca_cert attribute values...
1369 # match against attributes a users issuing CA SSL certificate
1370 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1372 acl aclname ext_user [-i] username ...
1373 acl aclname ext_user_regex [-i] pattern ...
1374 # string match on username returned by external acl helper [slow]
1375 # use REQUIRED to accept any non-null user name.
1377 acl aclname tag tagvalue ...
1378 # string match on tag returned by external acl helper [fast]
1379 # DEPRECATED. Only the first tag will match with this ACL.
1380 # Use the 'note' ACL instead for handling multiple tag values.
1382 acl aclname hier_code codename ...
1383 # string match against squid hierarchy code(s); [fast]
1384 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1386 # NOTE: This has no effect in http_access rules. It only has
1387 # effect in rules that affect the reply data stream such as
1388 # http_reply_access.
1390 acl aclname note [-m[=delimiters]] name [value ...]
1391 # match transaction annotation [fast]
1392 # Without values, matches any annotation with a given name.
1393 # With value(s), matches any annotation with a given name that
1394 # also has one of the given values.
1395 # If the -m flag is used, then the value of the named
1396 # annotation is interpreted as a list of tokens, and the ACL
1397 # matches individual name=token pairs rather than whole
1398 # name=value pairs. See "ACL Options" above for more info.
1399 # Annotation sources include note and adaptation_meta directives
1400 # as well as helper and eCAP responses.
1402 acl aclname annotate_transaction [-m[=delimiters]] key=value ...
1403 acl aclname annotate_transaction [-m[=delimiters]] key+=value ...
1404 # Always matches. [fast]
1405 # Used for its side effect: This ACL immediately adds a
1406 # key=value annotation to the current master transaction.
1407 # The added annotation can then be tested using note ACL and
1408 # logged (or sent to helpers) using %note format code.
1410 # Annotations can be specified using replacement and addition
1411 # formats. The key=value form replaces old same-key annotation
1412 # value(s). The key+=value form appends a new value to the old
1413 # same-key annotation. Both forms create a new key=value
1414 # annotation if no same-key annotation exists already. If
1415 # -m flag is used, then the value is interpreted as a list
1416 # and the annotation will contain key=token pair(s) instead of the
1417 # whole key=value pair.
1419 # This ACL is especially useful for recording complex multi-step
1420 # ACL-driven decisions. For example, the following configuration
1421 # avoids logging transactions accepted after aclX matched:
1423 # # First, mark transactions accepted after aclX matched
1424 # acl markSpecial annotate_transaction special=true
1425 # http_access allow acl001
1427 # http_access deny acl100
1428 # http_access allow aclX markSpecial
1430 # # Second, do not log marked transactions:
1431 # acl markedSpecial note special true
1432 # access_log ... deny markedSpecial
1434 # # Note that the following would not have worked because aclX
1435 # # alone does not determine whether the transaction was allowed:
1436 # access_log ... deny aclX # Wrong!
1438 # Warning: This ACL annotates the transaction even when negated
1439 # and even if subsequent ACLs fail to match. For example, the
1440 # following three rules will have exactly the same effect as far
1441 # as annotations set by the "mark" ACL are concerned:
1443 # some_directive acl1 ... mark # rule matches if mark is reached
1444 # some_directive acl1 ... !mark # rule never matches
1445 # some_directive acl1 ... mark !all # rule never matches
1447 acl aclname annotate_client [-m[=delimiters]] key=value ...
1448 acl aclname annotate_client [-m[=delimiters]] key+=value ...
1450 # Always matches. [fast]
1451 # Used for its side effect: This ACL immediately adds a
1452 # key=value annotation to the current client-to-Squid
1453 # connection. Connection annotations are propagated to the current
1454 # and all future master transactions on the annotated connection.
1455 # See the annotate_transaction ACL for details.
1457 # For example, the following configuration avoids rewriting URLs
1458 # of transactions bumped by SslBump:
1460 # # First, mark bumped connections:
1461 # acl markBumped annotate_client bumped=true
1462 # ssl_bump peek acl1
1463 # ssl_bump stare acl2
1464 # ssl_bump bump acl3 markBumped
1465 # ssl_bump splice all
1467 # # Second, do not send marked transactions to the redirector:
1468 # acl markedBumped note bumped true
1469 # url_rewrite_access deny markedBumped
1471 # # Note that the following would not have worked because acl3 alone
1472 # # does not determine whether the connection is going to be bumped:
1473 # url_rewrite_access deny acl3 # Wrong!
1475 acl aclname adaptation_service service ...
1476 # Matches the name of any icap_service, ecap_service,
1477 # adaptation_service_set, or adaptation_service_chain that Squid
1478 # has used (or attempted to use) for the master transaction.
1479 # This ACL must be defined after the corresponding adaptation
1480 # service is named in squid.conf. This ACL is usable with
1481 # adaptation_meta because it starts matching immediately after
1482 # the service has been selected for adaptation.
1484 acl aclname transaction_initiator initiator ...
1485 # Matches transaction's initiator [fast]
1487 # Supported initiators are:
1488 # esi: matches transactions fetching ESI resources
1489 # certificate-fetching: matches transactions fetching
1490 # a missing intermediate TLS certificate
1491 # cache-digest: matches transactions fetching Cache Digests
1493 # htcp: matches HTCP requests from peers
1494 # icp: matches ICP requests to peers
1495 # icmp: matches ICMP RTT database (NetDB) requests to peers
1496 # asn: matches asns db requests
1497 # internal: matches any of the above
1498 # client: matches transactions containing an HTTP or FTP
1499 # client request received at a Squid *_port
1500 # all: matches any transaction, including internal transactions
1501 # without a configurable initiator and hopefully rare
1502 # transactions without a known-to-Squid initiator
1504 # Multiple initiators are ORed.
1506 acl aclname has component
1507 # matches a transaction "component" [fast]
1509 # Supported transaction components are:
1510 # request: transaction has a request header (at least)
1511 # response: transaction has a response header (at least)
1512 # ALE: transaction has an internally-generated Access Log Entry
1513 # structure; bugs notwithstanding, all transaction have it
1515 # For example, the following configuration helps when dealing with HTTP
1516 # clients that close connections without sending a request header:
1518 # acl hasRequest has request
1519 # acl logMe note important_transaction
1520 # # avoid "logMe ACL is used in context without an HTTP request" warnings
1521 # access_log ... logformat=detailed hasRequest logMe
1522 # # log request-less transactions, instead of ignoring them
1523 # access_log ... logformat=brief !hasRequest
1525 # Multiple components are not supported for one "acl" rule, but
1526 # can be specified (and are ORed) using multiple same-name rules:
1528 # # OK, this strange logging daemon needs request or response,
1529 # # but can work without either a request or a response:
1530 # acl hasWhatMyLoggingDaemonNeeds has request
1531 # acl hasWhatMyLoggingDaemonNeeds has response
1533 acl aclname at_step step
1534 # match against the current request processing step [fast]
1536 # GeneratingCONNECT: Generating HTTP CONNECT request headers
1538 # The following ssl_bump processing steps are recognized:
1539 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1540 # SslBump2: After getting SSL Client Hello info.
1541 # SslBump3: After getting SSL Server Hello info.
1545 acl aclname ssl_error errorname
1546 # match against SSL certificate validation error [fast]
1548 # When used with sslproxy_cert_error, this ACL tests a single
1549 # certificate validation error currently being evaluated by that
1550 # directive. When used with slproxy_cert_sign or sslproxy_cert_adapt,
1551 # the ACL tests all past certificate validation errors associated with
1552 # the current Squid-to-server connection (attempt). This ACL is not yet
1553 # supported for use with other directives.
1555 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1558 # The following can be used as shortcuts for certificate properties:
1559 # [ssl::]certHasExpired: the "not after" field is in the past
1560 # [ssl::]certNotYetValid: the "not before" field is in the future
1561 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1562 # [ssl::]certSelfSigned: The certificate is self signed.
1563 # [ssl::]certDomainMismatch: The certificate CN domain does not
1564 # match the name the name of the host we are connecting to.
1566 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1567 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1568 # predefined ACLs, just like the 'all' ACL.
1570 acl aclname server_cert_fingerprint fingerprint
1571 # match against server SSL certificate fingerprint [fast]
1573 # The fingerprint is the digest of the DER encoded version
1574 # of the whole certificate. The user should use the form: XX:XX:...
1575 # The SHA1 digest algorithm is the default and is currently
1576 # the only algorithm supported.
1578 acl aclname ssl::server_name [option] .foo.com ...
1579 # matches server name obtained from various sources [fast]
1581 # The ACL computes server name(s) using such information sources as
1582 # CONNECT request URI, TLS client SNI, and TLS server certificate
1583 # subject (CN and SubjectAltName). The computed server name(s) usually
1584 # change with each SslBump step, as more info becomes available:
1585 # * SNI is used as the server name instead of the request URI,
1586 # * subject name(s) from the server certificate (CN and
1587 # SubjectAltName) are used as the server names instead of SNI.
1589 # When the ACL computes multiple server names, matching any single
1590 # computed name is sufficient for the ACL to match.
1592 # The "none" name can be used to match transactions where the ACL
1593 # could not compute the server name using any information source
1594 # that was both available and allowed to be used by the ACL options at
1595 # the ACL evaluation time.
1597 # Unlike dstdomain, this ACL does not perform DNS lookups.
1599 # An ACL option below may be used to restrict what information
1600 # sources are used to extract the server names from:
1602 # --client-requested
1603 # The server name is SNI regardless of what the server says.
1605 # The server name(s) are the certificate subject name(s), regardless
1606 # of what the client has requested. If the server certificate is
1607 # unavailable, then the name is "none".
1609 # The server name is either SNI (if SNI matches at least one of the
1610 # certificate subject names) or "none" (otherwise). When the server
1611 # certificate is unavailable, the consensus server name is SNI.
1613 # Combining multiple options in one ACL is a fatal configuration
1616 # For all options: If no SNI is available, then the CONNECT request
1617 # target (a.k.a. URI) is used instead of SNI (for an intercepted
1618 # connection, this target is the destination IP address).
1620 acl aclname ssl::server_name_regex [-i] \.foo\.com ...
1621 # regex matches server name obtained from various sources [fast]
1623 acl aclname connections_encrypted
1624 # matches transactions with all HTTP messages received over TLS
1625 # transport connections. [fast]
1627 # The master transaction deals with HTTP messages received from
1628 # various sources. All sources used by the master transaction in the
1629 # past are considered by the ACL. The following rules define whether
1630 # a given message source taints the entire master transaction,
1631 # resulting in ACL mismatches:
1633 # * The HTTP client transport connection is not TLS.
1634 # * An adaptation service connection-encryption flag is off.
1635 # * The peer or origin server transport connection is not TLS.
1637 # Caching currently does not affect these rules. This cache ignorance
1638 # implies that only the current HTTP client transport and REQMOD
1639 # services status determine whether this ACL matches a from-cache
1640 # transaction. The source of the cached response does not have any
1641 # effect on future transaction that use the cached response without
1642 # revalidation. This may change.
1644 # DNS, ICP, and HTCP exchanges during the master transaction do not
1645 # affect these rules.
1647 acl aclname any-of acl1 acl2 ...
1648 # match any one of the acls [fast or slow]
1649 # The first matching ACL stops further ACL evaluation.
1651 # ACLs from multiple any-of lines with the same name are ORed.
1652 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1653 # acl A any-of a1 a2
1654 # acl A any-of a3 a4
1656 # This group ACL is fast if all evaluated ACLs in the group are fast
1657 # and slow otherwise.
1659 acl aclname all-of acl1 acl2 ...
1660 # match all of the acls [fast or slow]
1661 # The first mismatching ACL stops further ACL evaluation.
1663 # ACLs from multiple all-of lines with the same name are ORed.
1664 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1665 # acl B all-of b1 b2
1666 # acl B all-of b3 b4
1668 # This group ACL is fast if all evaluated ACLs in the group are fast
1669 # and slow otherwise.
1672 acl macaddress arp 09:00:2b:23:45:67
1673 acl myexample dst_as 1241
1674 acl password proxy_auth REQUIRED
1675 acl fileupload req_mime_type -i ^multipart/form-data$
1676 acl javascript rep_mime_type -i ^application/x-javascript$
1680 # Recommended minimum configuration:
1683 # Example rule allowing access from your local networks.
1684 # Adapt to list your (internal) IP networks from where browsing
1686 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1687 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1688 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1689 acl localnet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1690 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1691 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1692 acl localnet src fc00::/7 # RFC 4193 local private network range
1693 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1695 acl SSL_ports port 443
1696 acl Safe_ports port 80 # http
1697 acl Safe_ports port 21 # ftp
1698 acl Safe_ports port 443 # https
1699 acl Safe_ports port 70 # gopher
1700 acl Safe_ports port 210 # wais
1701 acl Safe_ports port 1025-65535 # unregistered ports
1702 acl Safe_ports port 280 # http-mgmt
1703 acl Safe_ports port 488 # gss-http
1704 acl Safe_ports port 591 # filemaker
1705 acl Safe_ports port 777 # multiling http
1709 NAME: proxy_protocol_access
1711 LOC: Config.accessList.proxyProtocol
1713 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1715 Determine which client proxies can be trusted to provide correct
1716 information regarding real client IP address using PROXY protocol.
1718 Requests may pass through a chain of several other proxies
1719 before reaching us. The original source details may by sent in:
1720 * HTTP message Forwarded header, or
1721 * HTTP message X-Forwarded-For header, or
1722 * PROXY protocol connection header.
1724 This directive is solely for validating new PROXY protocol
1725 connections received from a port flagged with require-proxy-header.
1726 It is checked only once after TCP connection setup.
1728 A deny match results in TCP connection closure.
1730 An allow match is required for Squid to permit the corresponding
1731 TCP connection, before Squid even looks for HTTP request headers.
1732 If there is an allow match, Squid starts using PROXY header information
1733 to determine the source address of the connection for all future ACL
1734 checks, logging, etc.
1736 SECURITY CONSIDERATIONS:
1738 Any host from which we accept client IP details can place
1739 incorrect information in the relevant header, and Squid
1740 will use the incorrect information as if it were the
1741 source address of the request. This may enable remote
1742 hosts to bypass any access control restrictions that are
1743 based on the client's source addresses.
1745 This clause only supports fast acl types.
1746 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1749 NAME: follow_x_forwarded_for
1751 IFDEF: FOLLOW_X_FORWARDED_FOR
1752 LOC: Config.accessList.followXFF
1753 DEFAULT_IF_NONE: deny all
1754 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1756 Determine which client proxies can be trusted to provide correct
1757 information regarding real client IP address.
1759 Requests may pass through a chain of several other proxies
1760 before reaching us. The original source details may by sent in:
1761 * HTTP message Forwarded header, or
1762 * HTTP message X-Forwarded-For header, or
1763 * PROXY protocol connection header.
1765 PROXY protocol connections are controlled by the proxy_protocol_access
1766 directive which is checked before this.
1768 If a request reaches us from a source that is allowed by this
1769 directive, then we trust the information it provides regarding
1770 the IP of the client it received from (if any).
1772 For the purpose of ACLs used in this directive the src ACL type always
1773 matches the address we are testing and srcdomain matches its rDNS.
1775 On each HTTP request Squid checks for X-Forwarded-For header fields.
1776 If found the header values are iterated in reverse order and an allow
1777 match is required for Squid to continue on to the next value.
1778 The verification ends when a value receives a deny match, cannot be
1779 tested, or there are no more values to test.
1780 NOTE: Squid does not yet follow the Forwarded HTTP header.
1782 The end result of this process is an IP address that we will
1783 refer to as the indirect client address. This address may
1784 be treated as the client address for access control, ICAP, delay
1785 pools and logging, depending on the acl_uses_indirect_client,
1786 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1787 log_uses_indirect_client and tproxy_uses_indirect_client options.
1789 This clause only supports fast acl types.
1790 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1792 SECURITY CONSIDERATIONS:
1794 Any host from which we accept client IP details can place
1795 incorrect information in the relevant header, and Squid
1796 will use the incorrect information as if it were the
1797 source address of the request. This may enable remote
1798 hosts to bypass any access control restrictions that are
1799 based on the client's source addresses.
1803 acl localhost src 127.0.0.1
1804 acl my_other_proxy srcdomain .proxy.example.com
1805 follow_x_forwarded_for allow localhost
1806 follow_x_forwarded_for allow my_other_proxy
1809 NAME: acl_uses_indirect_client
1812 IFDEF: FOLLOW_X_FORWARDED_FOR
1814 LOC: Config.onoff.acl_uses_indirect_client
1816 Controls whether the indirect client address
1817 (see follow_x_forwarded_for) is used instead of the
1818 direct client address in acl matching.
1820 NOTE: maxconn ACL considers direct TCP links and indirect
1821 clients will always have zero. So no match.
1824 NAME: delay_pool_uses_indirect_client
1827 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1829 LOC: Config.onoff.delay_pool_uses_indirect_client
1831 Controls whether the indirect client address
1832 (see follow_x_forwarded_for) is used instead of the
1833 direct client address in delay pools.
1836 NAME: log_uses_indirect_client
1839 IFDEF: FOLLOW_X_FORWARDED_FOR
1841 LOC: Config.onoff.log_uses_indirect_client
1843 Controls whether the indirect client address
1844 (see follow_x_forwarded_for) is used instead of the
1845 direct client address in the access log.
1848 NAME: tproxy_uses_indirect_client
1851 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1853 LOC: Config.onoff.tproxy_uses_indirect_client
1855 Controls whether the indirect client address
1856 (see follow_x_forwarded_for) is used instead of the
1857 direct client address when spoofing the outgoing client.
1859 This has no effect on requests arriving in non-tproxy
1862 SECURITY WARNING: Usage of this option is dangerous
1863 and should not be used trivially. Correct configuration
1864 of follow_x_forwarded_for with a limited set of trusted
1865 sources is required to prevent abuse of your proxy.
1868 NAME: spoof_client_ip
1870 LOC: Config.accessList.spoof_client_ip
1872 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1874 Control client IP address spoofing of TPROXY traffic based on
1875 defined access lists.
1877 spoof_client_ip allow|deny [!]aclname ...
1879 If there are no "spoof_client_ip" lines present, the default
1880 is to "allow" spoofing of any suitable request.
1882 Note that the cache_peer "no-tproxy" option overrides this ACL.
1884 This clause supports fast acl types.
1885 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1890 LOC: Config.accessList.http
1891 DEFAULT_IF_NONE: deny all
1892 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1894 Allowing or Denying access based on defined access lists
1896 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1897 http_access allow|deny [!]aclname ...
1899 NOTE on default values:
1901 If there are no "access" lines present, the default is to deny
1904 If none of the "access" lines cause a match, the default is the
1905 opposite of the last line in the list. If the last line was
1906 deny, the default is allow. Conversely, if the last line
1907 is allow, the default will be deny. For these reasons, it is a
1908 good idea to have an "deny all" entry at the end of your access
1909 lists to avoid potential confusion.
1911 This clause supports both fast and slow acl types.
1912 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1917 # Recommended minimum Access Permission configuration:
1919 # Deny requests to certain unsafe ports
1920 http_access deny !Safe_ports
1922 # Deny CONNECT to other than secure SSL ports
1923 http_access deny CONNECT !SSL_ports
1925 # Only allow cachemgr access from localhost
1926 http_access allow localhost manager
1927 http_access deny manager
1929 # This default configuration only allows localhost requests because a more
1930 # permissive Squid installation could introduce new attack vectors into the
1931 # network by proxying external TCP connections to unprotected services.
1932 http_access allow localhost
1934 # The two deny rules below are unnecessary in this default configuration
1935 # because they are followed by a "deny all" rule. However, they may become
1936 # critically important when you start allowing external requests below them.
1938 # Protect web applications running on the same server as Squid. They often
1939 # assume that only local users can access them at "localhost" ports.
1940 http_access deny to_localhost
1942 # Protect cloud servers that provide local users with sensitive info about
1943 # their server via certain well-known link-local (a.k.a. APIPA) addresses.
1944 http_access deny to_linklocal
1947 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1950 # For example, to allow access from your local networks, you may uncomment the
1951 # following rule (and/or add rules that match your definition of "local"):
1952 # http_access allow localnet
1954 # And finally deny all other access to this proxy
1955 http_access deny all
1959 NAME: adapted_http_access http_access2
1961 LOC: Config.accessList.adapted_http
1963 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1965 Allowing or Denying access based on defined access lists
1967 Essentially identical to http_access, but runs after redirectors
1968 and ICAP/eCAP adaptation. Allowing access control based on their
1971 If not set then only http_access is used.
1974 NAME: http_reply_access
1976 LOC: Config.accessList.reply
1978 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1980 Allow replies to client requests. This is complementary to http_access.
1982 http_reply_access allow|deny [!] aclname ...
1984 NOTE: if there are no access lines present, the default is to allow
1987 If none of the access lines cause a match the opposite of the
1988 last line will apply. Thus it is good practice to end the rules
1989 with an "allow all" or "deny all" entry.
1991 This clause supports both fast and slow acl types.
1992 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1997 LOC: Config.accessList.icp
1999 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
2001 Allowing or Denying access to the ICP port based on defined
2004 icp_access allow|deny [!]aclname ...
2006 NOTE: The default if no icp_access lines are present is to
2007 deny all traffic. This default may cause problems with peers
2010 This clause only supports fast acl types.
2011 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2013 # Allow ICP queries from local networks only
2014 #icp_access allow localnet
2015 #icp_access deny all
2021 LOC: Config.accessList.htcp
2023 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
2025 Allowing or Denying access to the HTCP port based on defined
2028 htcp_access allow|deny [!]aclname ...
2030 See also htcp_clr_access for details on access control for
2031 cache purge (CLR) HTCP messages.
2033 NOTE: The default if no htcp_access lines are present is to
2034 deny all traffic. This default may cause problems with peers
2035 using the htcp option.
2037 This clause only supports fast acl types.
2038 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2040 # Allow HTCP queries from local networks only
2041 #htcp_access allow localnet
2042 #htcp_access deny all
2045 NAME: htcp_clr_access
2048 LOC: Config.accessList.htcp_clr
2050 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
2052 Allowing or Denying access to purge content using HTCP based
2053 on defined access lists.
2054 See htcp_access for details on general HTCP access control.
2056 htcp_clr_access allow|deny [!]aclname ...
2058 This clause only supports fast acl types.
2059 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2061 # Allow HTCP CLR requests from trusted peers
2062 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
2063 htcp_clr_access allow htcp_clr_peer
2064 htcp_clr_access deny all
2069 LOC: Config.accessList.miss
2071 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
2073 Determines whether network access is permitted when satisfying a request.
2076 to force your neighbors to use you as a sibling instead of
2079 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
2080 miss_access deny !localclients
2081 miss_access allow all
2083 This means only your local clients are allowed to fetch relayed/MISS
2084 replies from the network and all other clients can only fetch cached
2087 The default for this setting allows all clients who passed the
2088 http_access rules to relay via this proxy.
2090 This clause only supports fast acl types.
2091 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2094 NAME: ident_lookup_access
2098 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
2099 LOC: Ident::TheConfig.identLookup
2101 A list of ACL elements which, if matched, cause an ident
2102 (RFC 931) lookup to be performed for this request. For
2103 example, you might choose to always perform ident lookups
2104 for your main multi-user Unix boxes, but not for your Macs
2105 and PCs. By default, ident lookups are not performed for
2108 To enable ident lookups for specific client addresses, you
2109 can follow this example:
2111 acl ident_aware_hosts src 198.168.1.0/24
2112 ident_lookup_access allow ident_aware_hosts
2113 ident_lookup_access deny all
2115 Only src type ACL checks are fully supported. A srcdomain
2116 ACL might work at times, but it will not always provide
2119 This clause only supports fast acl types.
2120 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2123 NAME: reply_body_max_size
2124 COMMENT: size [acl acl...]
2127 DEFAULT_DOC: No limit is applied.
2128 LOC: Config.ReplyBodySize
2130 This option specifies the maximum size of a reply body. It can be
2131 used to prevent users from downloading very large files, such as
2132 MP3's and movies. When the reply headers are received, the
2133 reply_body_max_size lines are processed, and the first line where
2134 all (if any) listed ACLs are true is used as the maximum body size
2137 This size is checked twice. First when we get the reply headers,
2138 we check the content-length value. If the content length value exists
2139 and is larger than the allowed size, the request is denied and the
2140 user receives an error message that says "the request or reply
2141 is too large." If there is no content-length, and the reply
2142 size exceeds this limit, the client's connection is just closed
2143 and they will receive a partial reply.
2145 WARNING: downstream caches probably can not detect a partial reply
2146 if there is no content-length header, so they will cache
2147 partial responses and give them out as hits. You should NOT
2148 use this option if you have downstream caches.
2150 WARNING: A maximum size smaller than the size of squid's error messages
2151 will cause an infinite loop and crash squid. Ensure that the smallest
2152 non-zero value you use is greater that the maximum header size plus
2153 the size of your largest error page.
2155 If you set this parameter none (the default), there will be
2158 Configuration Format is:
2159 reply_body_max_size SIZE UNITS [acl ...]
2161 reply_body_max_size 10 MB
2165 NAME: on_unsupported_protocol
2166 TYPE: on_unsupported_protocol
2167 LOC: Config.accessList.on_unsupported_protocol
2169 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
2171 Determines Squid behavior when encountering strange requests at the
2172 beginning of an accepted TCP connection or the beginning of a bumped
2173 CONNECT tunnel. Controlling Squid reaction to unexpected traffic is
2174 especially useful in interception environments where Squid is likely
2175 to see connections for unsupported protocols that Squid should either
2176 terminate or tunnel at TCP level.
2178 on_unsupported_protocol <action> [!]acl ...
2180 The first matching action wins. Only fast ACLs are supported.
2182 Supported actions are:
2184 tunnel: Establish a TCP connection with the intended server and
2185 blindly shovel TCP packets between the client and server.
2187 respond: Respond with an error message, using the transfer protocol
2188 for the Squid port that received the request (e.g., HTTP
2189 for connections intercepted at the http_port). This is the
2192 Squid expects the following traffic patterns:
2194 http_port: a plain HTTP request
2195 https_port: SSL/TLS handshake followed by an [encrypted] HTTP request
2196 ftp_port: a plain FTP command (no on_unsupported_protocol support yet!)
2197 CONNECT tunnel on http_port: same as https_port
2198 CONNECT tunnel on https_port: same as https_port
2200 Currently, this directive has effect on intercepted connections and
2201 bumped tunnels only. Other cases are not supported because Squid
2202 cannot know the intended destination of other traffic.
2205 # define what Squid errors indicate receiving non-HTTP traffic:
2206 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
2207 # define what Squid errors indicate receiving nothing:
2208 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
2209 # tunnel everything that does not look like HTTP:
2210 on_unsupported_protocol tunnel foreignProtocol
2211 # tunnel if we think the client waits for the server to talk first:
2212 on_unsupported_protocol tunnel serverTalksFirstProtocol
2213 # in all other error cases, just send an HTTP "error page" response:
2214 on_unsupported_protocol respond all
2216 See also: squid_error ACL
2222 LOC: Auth::TheConfig.schemeAccess
2224 DEFAULT_DOC: use all auth_param schemes in their configuration order
2226 Use this directive to customize authentication schemes presence and
2227 order in Squid's Unauthorized and Authentication Required responses.
2229 auth_schemes scheme1,scheme2,... [!]aclname ...
2231 where schemeN is the name of one of the authentication schemes
2232 configured using auth_param directives. At least one scheme name is
2233 required. Multiple scheme names are separated by commas. Either
2234 avoid whitespace or quote the entire schemes list.
2236 A special "ALL" scheme name expands to all auth_param-configured
2237 schemes in their configuration order. This directive cannot be used
2238 to configure Squid to offer no authentication schemes at all.
2240 The first matching auth_schemes rule determines the schemes order
2241 for the current Authentication Required transaction. Note that the
2242 future response is not yet available during auth_schemes evaluation.
2244 If this directive is not used or none of its rules match, then Squid
2245 responds with all configured authentication schemes in the order of
2246 auth_param directives in the configuration file.
2248 This directive does not determine when authentication is used or
2249 how each authentication scheme authenticates clients.
2251 The following example sends basic and negotiate authentication
2252 schemes, in that order, when requesting authentication of HTTP
2253 requests matching the isIE ACL (not shown) while sending all
2254 auth_param schemes in their configuration order to other clients:
2256 auth_schemes basic,negotiate isIE
2257 auth_schemes ALL all # explicit default
2259 This directive supports fast ACLs only.
2261 See also: auth_param.
2266 -----------------------------------------------------------------------------
2269 NAME: http_port ascii_port
2274 Usage: port [mode] [options]
2275 hostname:port [mode] [options]
2276 1.2.3.4:port [mode] [options]
2278 The socket addresses where Squid will listen for HTTP client
2279 requests. You may specify multiple socket addresses.
2280 There are three forms: port alone, hostname with port, and
2281 IP address with port. If you specify a hostname or IP
2282 address, Squid binds the socket to that specific
2283 address. Most likely, you do not need to bind to a specific
2284 address, so you can use the port number alone.
2286 If you are running Squid in accelerator mode, you
2287 probably want to listen on port 80 also, or instead.
2289 The -a command line option may be used to specify additional
2290 port(s) where Squid listens for proxy request. Such ports will
2291 be plain proxy ports with no options.
2293 You may specify multiple socket addresses on multiple lines.
2297 intercept Support for IP-Layer NAT interception delivering
2298 traffic to this Squid port.
2299 NP: disables authentication on the port.
2301 tproxy Support Linux TPROXY (or BSD divert-to) with spoofing
2302 of outgoing connections using the client IP address.
2303 NP: disables authentication on the port.
2305 accel Accelerator / reverse proxy mode
2307 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
2308 establish secure connection with the client and with
2309 the server, decrypt HTTPS messages as they pass through
2310 Squid, and treat them as unencrypted HTTP messages,
2311 becoming the man-in-the-middle.
2313 The ssl_bump option is required to fully enable
2314 bumping of CONNECT requests.
2316 Omitting the mode flag causes default forward proxy mode to be used.
2319 Accelerator Mode Options:
2321 defaultsite=domainname
2322 What to use for the Host: header if it is not present
2323 in a request. Determines what site (not origin server)
2324 accelerators should consider the default.
2326 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
2328 protocol= Protocol to reconstruct accelerated and intercepted
2329 requests with. Defaults to HTTP/1.1 for http_port and
2330 HTTPS/1.1 for https_port.
2331 When an unsupported value is configured Squid will
2332 produce a FATAL error.
2333 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
2335 vport Virtual host port support. Using the http_port number
2336 instead of the port passed on Host: headers.
2338 vport=NN Virtual host port support. Using the specified port
2339 number instead of the port passed on Host: headers.
2342 Act as if this Squid is the origin server.
2343 This currently means generate new Date: and Expires:
2344 headers on HIT instead of adding Age:.
2346 ignore-cc Ignore request Cache-Control headers.
2348 WARNING: This option violates HTTP specifications if
2349 used in non-accelerator setups.
2351 allow-direct Allow direct forwarding in accelerator mode. Normally
2352 accelerated requests are denied direct forwarding as if
2353 never_direct was used.
2355 WARNING: this option opens accelerator mode to security
2356 vulnerabilities usually only affecting in interception
2357 mode. Make sure to protect forwarding with suitable
2358 http_access rules when using this.
2361 SSL Bump Mode Options:
2362 In addition to these options ssl-bump requires TLS/SSL options.
2364 generate-host-certificates[=<on|off>]
2365 Dynamically create SSL server certificates for the
2366 destination hosts of bumped CONNECT requests.When
2367 enabled, the cert and key options are used to sign
2368 generated certificates. Otherwise generated
2369 certificate will be selfsigned.
2370 If there is a CA certificate lifetime of the generated
2371 certificate equals lifetime of the CA certificate. If
2372 generated certificate is selfsigned lifetime is three
2374 This option is enabled by default when ssl-bump is used.
2375 See the ssl-bump option above for more information.
2377 dynamic_cert_mem_cache_size=SIZE
2378 Approximate total RAM size spent on cached generated
2379 certificates. If set to zero, caching is disabled. The
2380 default value is 4MB.
2384 tls-cert= Path to file containing an X.509 certificate (PEM format)
2385 to be used in the TLS handshake ServerHello.
2387 If this certificate is constrained by KeyUsage TLS
2388 feature it must allow HTTP server usage, along with
2389 any additional restrictions imposed by your choice
2390 of options= settings.
2392 When OpenSSL is used this file may also contain a
2393 chain of intermediate CA certificates to send in the
2396 When GnuTLS is used this option (and any paired
2397 tls-key= option) may be repeated to load multiple
2398 certificates for different domains.
2400 Also, when generate-host-certificates=on is configured
2401 the first tls-cert= option must be a CA certificate
2402 capable of signing the automatically generated
2405 tls-key= Path to a file containing private key file (PEM format)
2406 for the previous tls-cert= option.
2408 If tls-key= is not specified tls-cert= is assumed to
2409 reference a PEM file containing both the certificate
2412 cipher= Colon separated list of supported ciphers.
2413 NOTE: some ciphers such as EDH ciphers depend on
2414 additional settings. If those settings are
2415 omitted the ciphers may be silently ignored
2416 by the OpenSSL library.
2418 options= Various SSL implementation options. The most important
2421 NO_SSLv3 Disallow the use of SSLv3
2423 NO_TLSv1 Disallow the use of TLSv1.0
2425 NO_TLSv1_1 Disallow the use of TLSv1.1
2427 NO_TLSv1_2 Disallow the use of TLSv1.2
2430 Always create a new key when using
2431 temporary/ephemeral DH key exchanges
2434 Enable ephemeral ECDH key exchange.
2435 The adopted curve should be specified
2436 using the tls-dh option.
2439 Disable use of RFC5077 session tickets.
2440 Some servers may have problems
2441 understanding the TLS extension due
2442 to ambiguous specification in RFC4507.
2444 ALL Enable various bug workarounds
2445 suggested as "harmless" by OpenSSL
2446 Be warned that this reduces SSL/TLS
2447 strength to some attacks.
2449 See the OpenSSL SSL_CTX_set_options documentation for a
2452 clientca= File containing the list of CAs to use when
2453 requesting a client certificate.
2455 tls-cafile= PEM file containing CA certificates to use when verifying
2456 client certificates. If not configured clientca will be
2457 used. May be repeated to load multiple files.
2459 capath= Directory containing additional CA certificates
2460 and CRL lists to use when verifying client certificates.
2461 Requires OpenSSL or LibreSSL.
2463 crlfile= File of additional CRL lists to use when verifying
2464 the client certificate, in addition to CRLs stored in
2465 the capath. Implies VERIFY_CRL flag below.
2468 File containing DH parameters for temporary/ephemeral DH key
2469 exchanges, optionally prefixed by a curve for ephemeral ECDH
2471 See OpenSSL documentation for details on how to create the
2472 DH parameter file. Supported curves for ECDH can be listed
2473 using the "openssl ecparam -list_curves" command.
2474 WARNING: EDH and EECDH ciphers will be silently disabled if
2475 this option is not set.
2477 sslflags= Various flags modifying the use of SSL:
2479 Don't request client certificates
2480 immediately, but wait until acl processing
2481 requires a certificate (not yet implemented).
2483 Request a client certificate during the TLS
2484 handshake, but ignore certificate absence in
2485 the TLS client Hello. If the client does
2486 supply a certificate, it is validated.
2488 Don't allow for session reuse. Each connection
2489 will result in a new SSL session.
2491 Verify CRL lists when accepting client
2494 Verify CRL lists for all certificates in the
2495 client certificate chain.
2497 tls-default-ca[=off]
2498 Whether to use the system Trusted CAs. Default is OFF.
2500 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
2502 sslcontext= SSL session ID context identifier.
2506 connection-auth[=on|off]
2507 use connection-auth=off to tell Squid to prevent
2508 forwarding Microsoft connection oriented authentication
2509 (NTLM, Negotiate and Kerberos)
2511 disable-pmtu-discovery=
2512 Control Path-MTU discovery usage:
2513 off lets OS decide on what to do (default).
2514 transparent disable PMTU discovery when transparent
2516 always disable always PMTU discovery.
2518 In many setups of transparently intercepting proxies
2519 Path-MTU discovery can not work on traffic towards the
2520 clients. This is the case when the intercepting device
2521 does not fully track connections and fails to forward
2522 ICMP must fragment messages to the cache server. If you
2523 have such setup and experience that certain clients
2524 sporadically hang or never complete requests set
2525 disable-pmtu-discovery option to 'transparent'.
2527 name= Specifies a internal name for the port. Defaults to
2528 the port specification (port or addr:port)
2530 tcpkeepalive[=idle,interval,timeout]
2531 Enable TCP keepalive probes of idle connections.
2532 In seconds; idle is the initial time before TCP starts
2533 probing the connection, interval how often to probe, and
2534 timeout the time before giving up.
2536 require-proxy-header
2537 Require PROXY protocol version 1 or 2 connections.
2538 The proxy_protocol_access is required to permit
2539 downstream proxies which can be trusted.
2542 Ask TCP stack to maintain a dedicated listening queue
2543 for each worker accepting requests at this port.
2544 Requires TCP stack that supports the SO_REUSEPORT socket
2547 SECURITY WARNING: Enabling worker-specific queues
2548 allows any process running as Squid's effective user to
2549 easily accept requests destined to this port.
2551 If you run Squid on a dual-homed machine with an internal
2552 and an external interface we recommend you to specify the
2553 internal address:port in http_port. This way Squid will only be
2554 visible on the internal address.
2558 # Squid normally listens to port 3128
2559 http_port @DEFAULT_HTTP_PORT@
2564 IFDEF: HAVE_LIBGNUTLS||USE_OPENSSL
2569 Usage: [ip:]port [mode] tls-cert=certificate.pem [options]
2571 The socket address where Squid will listen for client requests made
2572 over TLS or SSL connections. Commonly referred to as HTTPS.
2574 This is most useful for situations where you are running squid in
2575 accelerator mode and you want to do the TLS work at the accelerator
2578 You may specify multiple socket addresses on multiple lines,
2579 each with their own certificate and/or options.
2581 The tls-cert= option is mandatory on HTTPS ports.
2583 See http_port for a list of modes and options.
2591 Enables Native FTP proxy by specifying the socket address where Squid
2592 listens for FTP client requests. See http_port directive for various
2593 ways to specify the listening address and mode.
2595 Usage: ftp_port address [mode] [options]
2597 WARNING: This is a new, experimental, complex feature that has seen
2598 limited production exposure. Some Squid modules (e.g., caching) do not
2599 currently work with native FTP proxying, and many features have not
2600 even been tested for compatibility. Test well before deploying!
2602 Native FTP proxying differs substantially from proxying HTTP requests
2603 with ftp:// URIs because Squid works as an FTP server and receives
2604 actual FTP commands (rather than HTTP requests with FTP URLs).
2606 Native FTP commands accepted at ftp_port are internally converted or
2607 wrapped into HTTP-like messages. The same happens to Native FTP
2608 responses received from FTP origin servers. Those HTTP-like messages
2609 are shoveled through regular access control and adaptation layers
2610 between the FTP client and the FTP origin server. This allows Squid to
2611 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2612 mechanisms when shoveling wrapped FTP messages. For example,
2613 http_access and adaptation_access directives are used.
2617 intercept Same as http_port intercept. The FTP origin address is
2618 determined based on the intended destination of the
2619 intercepted connection.
2621 tproxy Support Linux TPROXY for spoofing outgoing
2622 connections using the client IP address.
2623 NP: disables authentication and maybe IPv6 on the port.
2625 By default (i.e., without an explicit mode option), Squid extracts the
2626 FTP origin address from the login@origin parameter of the FTP USER
2627 command. Many popular FTP clients support such native FTP proxying.
2631 name=token Specifies an internal name for the port. Defaults to
2632 the port address. Usable with myportname ACL.
2635 Enables tracking of FTP directories by injecting extra
2636 PWD commands and adjusting Request-URI (in wrapping
2637 HTTP requests) to reflect the current FTP server
2638 directory. Tracking is disabled by default.
2640 protocol=FTP Protocol to reconstruct accelerated and intercepted
2641 requests with. Defaults to FTP. No other accepted
2642 values have been tested with. An unsupported value
2643 results in a FATAL error. Accepted values are FTP,
2644 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2646 Other http_port modes and options that are not specific to HTTP and
2647 HTTPS may also work.
2650 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2653 LOC: Ip::Qos::TheConfig.tosToServer
2655 Allows you to select a TOS/Diffserv value for packets outgoing
2656 on the server side, based on an ACL.
2658 tcp_outgoing_tos ds-field [!]aclname ...
2660 Example where normal_service_net uses the TOS value 0x00
2661 and good_service_net uses 0x20
2663 acl normal_service_net src 10.0.0.0/24
2664 acl good_service_net src 10.0.1.0/24
2665 tcp_outgoing_tos 0x00 normal_service_net
2666 tcp_outgoing_tos 0x20 good_service_net
2668 TOS/DSCP values really only have local significance - so you should
2669 know what you're specifying. For more information, see RFC2474,
2670 RFC2475, and RFC3260.
2672 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2673 "default" to use whatever default your host has.
2674 Note that only multiples of 4 are usable as the two rightmost bits have
2675 been redefined for use by ECN (RFC 3168 section 23.1).
2676 The squid parser will enforce this by masking away the ECN bits.
2678 Processing proceeds in the order specified, and stops at first fully
2681 Only fast ACLs are supported.
2684 NAME: clientside_tos
2687 LOC: Ip::Qos::TheConfig.tosToClient
2689 Allows you to select a TOS/DSCP value for packets being transmitted
2690 on the client-side, based on an ACL.
2692 clientside_tos ds-field [!]aclname ...
2694 Example where normal_service_net uses the TOS value 0x00
2695 and good_service_net uses 0x20
2697 acl normal_service_net src 10.0.0.0/24
2698 acl good_service_net src 10.0.1.0/24
2699 clientside_tos 0x00 normal_service_net
2700 clientside_tos 0x20 good_service_net
2702 Note: This feature is incompatible with qos_flows. Any TOS values set here
2703 will be overwritten by TOS values in qos_flows.
2705 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2706 "default" to use whatever default your host has.
2707 Note that only multiples of 4 are usable as the two rightmost bits have
2708 been redefined for use by ECN (RFC 3168 section 23.1).
2709 The squid parser will enforce this by masking away the ECN bits.
2711 This clause only supports fast acl types.
2712 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2715 NAME: tcp_outgoing_mark
2717 IFDEF: SO_MARK&&USE_LIBCAP
2719 LOC: Ip::Qos::TheConfig.nfmarkToServer
2721 Allows you to apply a Netfilter mark value to outgoing packets
2722 on the server side, based on an ACL.
2724 tcp_outgoing_mark mark-value [!]aclname ...
2726 Example where normal_service_net uses the mark value 0x00
2727 and good_service_net uses 0x20
2729 acl normal_service_net src 10.0.0.0/24
2730 acl good_service_net src 10.0.1.0/24
2731 tcp_outgoing_mark 0x00 normal_service_net
2732 tcp_outgoing_mark 0x20 good_service_net
2734 Only fast ACLs are supported.
2737 NAME: mark_client_packet clientside_mark
2739 IFDEF: SO_MARK&&USE_LIBCAP
2741 LOC: Ip::Qos::TheConfig.nfmarkToClient
2743 Allows you to apply a Netfilter MARK value to packets being transmitted
2744 on the client-side, based on an ACL.
2746 mark_client_packet mark-value [!]aclname ...
2748 Example where normal_service_net uses the MARK value 0x00
2749 and good_service_net uses 0x20
2751 acl normal_service_net src 10.0.0.0/24
2752 acl good_service_net src 10.0.1.0/24
2753 mark_client_packet 0x00 normal_service_net
2754 mark_client_packet 0x20 good_service_net
2756 Note: This feature is incompatible with qos_flows. Any mark values set here
2757 will be overwritten by mark values in qos_flows.
2759 This clause only supports fast acl types.
2760 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2763 NAME: mark_client_connection
2765 IFDEF: SO_MARK&&USE_LIBCAP
2767 LOC: Ip::Qos::TheConfig.nfConnmarkToClient
2769 Allows you to apply a Netfilter CONNMARK value to a connection
2770 on the client-side, based on an ACL.
2772 mark_client_connection mark-value[/mask] [!]aclname ...
2774 The mark-value and mask are unsigned integers (hex, octal, or decimal).
2775 The mask may be used to preserve marking previously set by other agents
2778 A matching rule replaces the CONNMARK value. If a mask is also
2779 specified, then the masked bits of the original value are zeroed, and
2780 the configured mark-value is ORed with that adjusted value.
2781 For example, applying a mark-value 0xAB/0xF to 0x5F CONNMARK, results
2782 in a 0xFB marking (rather than a 0xAB or 0x5B).
2784 This directive semantics is similar to iptables --set-mark rather than
2785 --set-xmark functionality.
2787 The directive does not interfere with qos_flows (which uses packet MARKs,
2790 Example where squid marks intercepted FTP connections:
2792 acl proto_ftp proto FTP
2793 mark_client_connection 0x200/0xff00 proto_ftp
2795 This clause only supports fast acl types.
2796 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2803 LOC: Ip::Qos::TheConfig
2805 Allows you to select a TOS/DSCP value to mark outgoing
2806 connections to the client, based on where the reply was sourced.
2807 For platforms using netfilter, allows you to set a netfilter mark
2808 value instead of, or in addition to, a TOS value.
2810 By default this functionality is disabled. To enable it with the default
2811 settings simply use "qos_flows mark" or "qos_flows tos". Default
2812 settings will result in the netfilter mark or TOS value being copied
2813 from the upstream connection to the client. Note that it is the connection
2814 CONNMARK value not the packet MARK value that is copied.
2816 It is not currently possible to copy the mark or TOS value from the
2817 client to the upstream connection request.
2819 TOS values really only have local significance - so you should
2820 know what you're specifying. For more information, see RFC2474,
2821 RFC2475, and RFC3260.
2823 The TOS/DSCP byte must be exactly that - a octet value 0 - 255.
2824 Note that only multiples of 4 are usable as the two rightmost bits have
2825 been redefined for use by ECN (RFC 3168 section 23.1).
2826 The squid parser will enforce this by masking away the ECN bits.
2828 Mark values can be any unsigned 32-bit integer value.
2830 This setting is configured by setting the following values:
2832 tos|mark Whether to set TOS or netfilter mark values
2834 local-hit=0xFF Value to mark local cache hits.
2836 sibling-hit=0xFF Value to mark hits from sibling peers.
2838 parent-hit=0xFF Value to mark hits from parent peers.
2840 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2841 over the preserve-miss feature (see below), unless
2842 mask is specified, in which case only the bits
2843 specified in the mask are written.
2845 The TOS variant of the following features are only possible on Linux
2846 and require your kernel to be patched with the TOS preserving ZPH
2847 patch, available from http://zph.bratcheda.org
2848 No patch is needed to preserve the netfilter mark, which will work
2849 with all variants of netfilter.
2851 disable-preserve-miss
2852 This option disables the preservation of the TOS or netfilter
2853 mark. By default, the existing TOS or netfilter mark value of
2854 the response coming from the remote server will be retained
2855 and masked with miss-mark.
2856 NOTE: in the case of a netfilter mark, the mark must be set on
2857 the connection (using the CONNMARK target) not on the packet
2861 Allows you to mask certain bits in the TOS or mark value
2862 received from the remote server, before copying the value to
2863 the TOS sent towards clients.
2864 Default for tos: 0xFF (TOS from server is not changed).
2865 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2867 All of these features require the --enable-zph-qos compilation flag
2868 (enabled by default). Netfilter marking also requires the
2869 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2870 libcap 2.09+ (--with-libcap).
2874 NAME: tcp_outgoing_address
2877 DEFAULT_DOC: Address selection is performed by the operating system.
2878 LOC: Config.accessList.outgoing_address
2880 Allows you to map requests to different outgoing IP addresses
2881 based on the username or source address of the user making
2884 tcp_outgoing_address ipaddr [[!]aclname] ...
2887 Forwarding clients with dedicated IPs for certain subnets.
2889 acl normal_service_net src 10.0.0.0/24
2890 acl good_service_net src 10.0.2.0/24
2892 tcp_outgoing_address 2001:db8::c001 good_service_net
2893 tcp_outgoing_address 10.1.0.2 good_service_net
2895 tcp_outgoing_address 2001:db8::beef normal_service_net
2896 tcp_outgoing_address 10.1.0.1 normal_service_net
2898 tcp_outgoing_address 2001:db8::1
2899 tcp_outgoing_address 10.1.0.3
2901 Processing proceeds in the order specified, and stops at first fully
2904 Squid will add an implicit IP version test to each line.
2905 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2906 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2909 NOTE: The use of this directive using client dependent ACLs is
2910 incompatible with the use of server side persistent connections. To
2911 ensure correct results it is best to set server_persistent_connections
2912 to off when using this directive in such configurations.
2914 NOTE: The use of this directive to set a local IP on outgoing TCP links
2915 is incompatible with using TPROXY to set client IP out outbound TCP links.
2916 When needing to contact peers use the no-tproxy cache_peer option and the
2917 client_dst_passthru directive re-enable normal forwarding such as this.
2919 This clause only supports fast acl types.
2920 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2923 NAME: host_verify_strict
2926 LOC: Config.onoff.hostStrictVerify
2928 Regardless of this option setting, when dealing with intercepted
2929 traffic, Squid always verifies that the destination IP address matches
2930 the Host header domain or IP (called 'authority form URL').
2932 This enforcement is performed to satisfy a MUST-level requirement in
2933 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2934 authority of the origin server or gateway given by the original URL".
2937 Squid always responds with an HTTP 409 (Conflict) error
2938 page and logs a security warning if there is no match.
2940 Squid verifies that the destination IP address matches
2941 the Host header for forward-proxy and reverse-proxy traffic
2942 as well. For those traffic types, Squid also enables the
2943 following checks, comparing the corresponding Host header
2944 and Request-URI components:
2946 * The host names (domain or IP) must be identical,
2947 but valueless or missing Host header disables all checks.
2948 For the two host names to match, both must be either IP
2951 * Port numbers must be identical, but if a port is missing
2952 the scheme-default port is assumed.
2955 When set to OFF (the default):
2956 Squid allows suspicious requests to continue but logs a
2957 security warning and blocks caching of the response.
2959 * Forward-proxy traffic is not checked at all.
2961 * Reverse-proxy traffic is not checked at all.
2963 * Intercepted traffic which passes verification is handled
2964 according to client_dst_passthru.
2966 * Intercepted requests which fail verification are sent
2967 to the client original destination instead of DIRECT.
2968 This overrides 'client_dst_passthru off'.
2970 For now suspicious intercepted CONNECT requests are always
2971 responded to with an HTTP 409 (Conflict) error page.
2976 As described in CVE-2009-0801 when the Host: header alone is used
2977 to determine the destination of a request it becomes trivial for
2978 malicious scripts on remote websites to bypass browser same-origin
2979 security policy and sandboxing protections.
2981 The cause of this is that such applets are allowed to perform their
2982 own HTTP stack, in which case the same-origin policy of the browser
2983 sandbox only verifies that the applet tries to contact the same IP
2984 as from where it was loaded at the IP level. The Host: header may
2985 be different from the connected IP and approved origin.
2989 NAME: client_dst_passthru
2992 LOC: Config.onoff.client_dst_passthru
2994 With NAT or TPROXY intercepted traffic Squid may pass the request
2995 directly to the original client destination IP or seek a faster
2996 source using the HTTP Host header.
2998 Using Host to locate alternative servers can provide faster
2999 connectivity with a range of failure recovery options.
3000 But can also lead to connectivity trouble when the client and
3001 server are attempting stateful interactions unaware of the proxy.
3003 This option (on by default) prevents alternative DNS entries being
3004 located to send intercepted traffic DIRECT to an origin server.
3005 The clients original destination IP and port will be used instead.
3007 Regardless of this option setting, when dealing with intercepted
3008 traffic Squid will verify the Host: header and any traffic which
3009 fails Host verification will be treated as if this option were ON.
3011 see host_verify_strict for details on the verification process.
3016 -----------------------------------------------------------------------------
3019 NAME: tls_outgoing_options
3020 IFDEF: HAVE_LIBGNUTLS||USE_OPENSSL
3021 TYPE: securePeerOptions
3022 DEFAULT: min-version=1.0
3023 LOC: Security::ProxyOutgoingConfig
3025 disable Do not support https:// URLs.
3027 cert=/path/to/client/certificate
3028 A client X.509 certificate to use when connecting.
3030 key=/path/to/client/private_key
3031 The private key corresponding to the cert= above.
3033 If key= is not specified cert= is assumed to
3034 reference a PEM file containing both the certificate
3037 cipher=... The list of valid TLS ciphers to use.
3040 The minimum TLS protocol version to permit.
3041 To control SSLv3 use the options= parameter.
3042 Supported Values: 1.0 (default), 1.1, 1.2, 1.3
3044 options=... Specify various TLS/SSL implementation options.
3046 OpenSSL options most important are:
3048 NO_SSLv3 Disallow the use of SSLv3
3051 Always create a new key when using
3052 temporary/ephemeral DH key exchanges
3055 Disable use of RFC5077 session tickets.
3056 Some servers may have problems
3057 understanding the TLS extension due
3058 to ambiguous specification in RFC4507.
3060 ALL Enable various bug workarounds
3061 suggested as "harmless" by OpenSSL
3062 Be warned that this reduces SSL/TLS
3063 strength to some attacks.
3065 See the OpenSSL SSL_CTX_set_options documentation
3066 for a more complete list.
3068 GnuTLS options most important are:
3071 Disable use of RFC5077 session tickets.
3072 Some servers may have problems
3073 understanding the TLS extension due
3074 to ambiguous specification in RFC4507.
3076 See the GnuTLS Priority Strings documentation
3077 for a more complete list.
3078 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
3081 cafile= PEM file containing CA certificates to use when verifying
3082 the peer certificate. May be repeated to load multiple files.
3084 capath= A directory containing additional CA certificates to
3085 use when verifying the peer certificate.
3086 Requires OpenSSL or LibreSSL.
3088 crlfile=... A certificate revocation list file to use when
3089 verifying the peer certificate.
3091 flags=... Specify various flags modifying the TLS implementation:
3094 Accept certificates even if they fail to
3097 Don't verify the peer certificate
3098 matches the server name
3101 Whether to use the system Trusted CAs. Default is ON.
3103 domain= The peer name as advertised in its certificate.
3104 Used for verifying the correctness of the received peer
3105 certificate. If not specified the peer hostname will be
3111 -----------------------------------------------------------------------------
3114 NAME: ssl_unclean_shutdown
3118 LOC: Config.SSL.unclean_shutdown
3120 Some browsers (especially MSIE) bugs out on SSL shutdown
3127 LOC: Config.SSL.ssl_engine
3130 The OpenSSL engine to use. You will need to set this if you
3131 would like to use hardware SSL acceleration for example.
3133 Not supported in builds with OpenSSL 3.0 or newer.
3136 NAME: sslproxy_session_ttl
3139 LOC: Config.SSL.session_ttl
3142 Sets the timeout value for SSL sessions
3145 NAME: sslproxy_session_cache_size
3148 LOC: Config.SSL.sessionCacheSize
3151 Sets the cache size to use for ssl session
3154 NAME: sslproxy_foreign_intermediate_certs
3157 LOC: Config.ssl_client.foreignIntermediateCertsPath
3160 Many origin servers fail to send their full server certificate
3161 chain for verification, assuming the client already has or can
3162 easily locate any missing intermediate certificates.
3164 Squid uses the certificates from the specified file to fill in
3165 these missing chains when trying to validate origin server
3168 The file is expected to contain zero or more PEM-encoded
3169 intermediate certificates. These certificates are not treated
3170 as trusted root certificates, and any self-signed certificate in
3171 this file will be ignored.
3174 NAME: sslproxy_cert_sign_hash
3177 LOC: Config.SSL.certSignHash
3180 Sets the hashing algorithm to use when signing generated certificates.
3181 Valid algorithm names depend on the OpenSSL library used. The following
3182 names are usually available: sha1, sha256, sha512, and md5. Please see
3183 your OpenSSL library manual for the available hashes. By default, Squids
3184 that support this option use sha256 hashes.
3186 Squid does not forcefully purge cached certificates that were generated
3187 with an algorithm other than the currently configured one. They remain
3188 in the cache, subject to the regular cache eviction policy, and become
3189 useful if the algorithm changes again.
3194 TYPE: sslproxy_ssl_bump
3195 LOC: Config.accessList.ssl_bump
3196 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
3199 This option is consulted when a CONNECT request is received on
3200 an http_port (or a new connection is intercepted at an
3201 https_port), provided that port was configured with an ssl-bump
3202 flag. The subsequent data on the connection is either treated as
3203 HTTPS and decrypted OR tunneled at TCP level without decryption,
3204 depending on the first matching bumping "action".
3206 ssl_bump <action> [!]acl ...
3208 The following bumping actions are currently supported:
3211 Become a TCP tunnel without decrypting proxied traffic.
3212 This is the default action.
3215 When used on step SslBump1, establishes a secure connection
3216 with the client first, then connect to the server.
3217 When used on step SslBump2 or SslBump3, establishes a secure
3218 connection with the server and, using a mimicked server
3219 certificate, with the client.
3222 Receive client (step SslBump1) or server (step SslBump2)
3223 certificate while preserving the possibility of splicing the
3224 connection. Peeking at the server certificate (during step 2)
3225 usually precludes bumping of the connection at step 3.
3228 Receive client (step SslBump1) or server (step SslBump2)
3229 certificate while preserving the possibility of bumping the
3230 connection. Staring at the server certificate (during step 2)
3231 usually precludes splicing of the connection at step 3.
3234 Close client and server connections.
3236 Backward compatibility actions available at step SslBump1:
3239 Bump the connection. Establish a secure connection with the
3240 client first, then connect to the server. This old mode does
3241 not allow Squid to mimic server SSL certificate and does not
3242 work with intercepted SSL connections.
3245 Bump the connection. Establish a secure connection with the
3246 server first, then establish a secure connection with the
3247 client, using a mimicked server certificate. Works with both
3248 CONNECT requests and intercepted SSL connections, but does
3249 not allow to make decisions based on SSL handshake info.
3252 Decide whether to bump or splice the connection based on
3253 client-to-squid and server-to-squid SSL hello messages.
3257 Same as the "splice" action.
3259 All ssl_bump rules are evaluated at each of the supported bumping
3260 steps. Rules with actions that are impossible at the current step are
3261 ignored. The first matching ssl_bump action wins and is applied at the
3262 end of the current step. If no rules match, the splice action is used.
3263 See the at_step ACL for a list of the supported SslBump steps.
3265 This clause supports both fast and slow acl types.
3266 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3268 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
3271 # Example: Bump all TLS connections except those originating from
3272 # localhost or those going to example.com.
3274 acl broken_sites ssl::server_name .example.com
3275 ssl_bump splice localhost
3276 ssl_bump splice broken_sites
3280 NAME: sslproxy_cert_error
3283 DEFAULT_DOC: Server certificate errors terminate the transaction.
3284 LOC: Config.ssl_client.cert_error
3287 Use this ACL to bypass server certificate validation errors.
3289 For example, the following lines will bypass all validation errors
3290 when talking to servers for example.com. All other
3291 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
3293 acl BrokenButTrustedServers dstdomain example.com
3294 sslproxy_cert_error allow BrokenButTrustedServers
3295 sslproxy_cert_error deny all
3297 This clause only supports fast acl types.
3298 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3299 Using slow acl types may result in server crashes
3301 Without this option, all server certificate validation errors
3302 terminate the transaction to protect Squid and the client.
3304 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
3305 but should not happen unless your OpenSSL library is buggy.
3308 Bypassing validation errors is dangerous because an
3309 error usually implies that the server cannot be trusted
3310 and the connection may be insecure.
3312 See also: sslproxy_flags and DONT_VERIFY_PEER.
3315 NAME: sslproxy_cert_sign
3318 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
3319 POSTSCRIPTUM: signSelf ssl::certSelfSigned
3320 POSTSCRIPTUM: signTrusted all
3321 TYPE: sslproxy_cert_sign
3322 LOC: Config.ssl_client.cert_sign
3325 sslproxy_cert_sign <signing algorithm> acl ...
3327 The following certificate signing algorithms are supported:
3330 Sign using the configured CA certificate which is usually
3331 placed in and trusted by end-user browsers. This is the
3332 default for trusted origin server certificates.
3335 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
3336 This is the default for untrusted origin server certificates
3337 that are not self-signed (see ssl::certUntrusted).
3340 Sign using a self-signed certificate with the right CN to
3341 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
3342 browser. This is the default for self-signed origin server
3343 certificates (see ssl::certSelfSigned).
3345 This clause only supports fast acl types.
3347 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
3348 signing algorithm to generate the certificate and ignores all
3349 subsequent sslproxy_cert_sign options (the first match wins). If no
3350 acl(s) match, the default signing algorithm is determined by errors
3351 detected when obtaining and validating the origin server certificate.
3353 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3354 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3355 CONNECT request that carries a domain name. In all other cases (CONNECT
3356 to an IP address or an intercepted SSL connection), Squid cannot detect
3357 the domain mismatch at certificate generation time when
3358 bump-server-first is used.
3361 NAME: sslproxy_cert_adapt
3364 TYPE: sslproxy_cert_adapt
3365 LOC: Config.ssl_client.cert_adapt
3368 sslproxy_cert_adapt <adaptation algorithm> acl ...
3370 The following certificate adaptation algorithms are supported:
3373 Sets the "Not After" property to the "Not After" property of
3374 the CA certificate used to sign generated certificates.
3377 Sets the "Not Before" property to the "Not Before" property of
3378 the CA certificate used to sign generated certificates.
3380 setCommonName or setCommonName{CN}
3381 Sets Subject.CN property to the host name specified as a
3382 CN parameter or, if no explicit CN parameter was specified,
3383 extracted from the CONNECT request. It is a misconfiguration
3384 to use setCommonName without an explicit parameter for
3385 intercepted or tproxied SSL connections.
3387 This clause only supports fast acl types.
3389 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
3390 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
3391 corresponding adaptation algorithm to generate the certificate and
3392 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
3393 group (i.e., the first match wins within each algorithm group). If no
3394 acl(s) match, the default mimicking action takes place.
3396 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3397 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3398 CONNECT request that carries a domain name. In all other cases (CONNECT
3399 to an IP address or an intercepted SSL connection), Squid cannot detect
3400 the domain mismatch at certificate generation time when
3401 bump-server-first is used.
3404 NAME: sslpassword_program
3407 LOC: Config.Program.ssl_password
3410 Specify a program used for entering SSL key passphrases
3411 when using encrypted SSL certificate keys. If not specified
3412 keys must either be unencrypted, or Squid started with the -N
3413 option to allow it to query interactively for the passphrase.
3415 The key file name is given as argument to the program allowing
3416 selection of the right password if you have multiple encrypted
3421 OPTIONS RELATING TO EXTERNAL SSL_CRTD
3422 -----------------------------------------------------------------------------
3425 NAME: sslcrtd_program
3428 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
3429 LOC: Ssl::TheConfig.ssl_crtd
3431 Specify the location and options of the executable for certificate
3434 @DEFAULT_SSL_CRTD@ program can use a disk cache to improve response
3435 times on repeated requests. To enable caching, specify -s and -M
3436 parameters. If those parameters are not given, the program generates
3437 a new certificate on every request.
3439 For more information use:
3440 @DEFAULT_SSL_CRTD@ -h
3443 NAME: sslcrtd_children
3444 TYPE: HelperChildConfig
3446 DEFAULT: 32 startup=5 idle=1
3447 LOC: Ssl::TheConfig.ssl_crtdChildren
3449 Specifies the maximum number of certificate generation processes that
3450 Squid may spawn (numberofchildren) and several related options. Using
3451 too few of these helper processes (a.k.a. "helpers") creates request
3452 queues. Using too many helpers wastes your system resources. Squid
3453 does not support spawning more than 32 helpers.
3455 Usage: numberofchildren [option]...
3457 The startup= and idle= options allow some measure of skew in your
3462 Sets the minimum number of processes to spawn when Squid
3463 starts or reconfigures. When set to zero the first request will
3464 cause spawning of the first child process to handle it.
3466 Starting too few children temporary slows Squid under load while it
3467 tries to spawn enough additional processes to cope with traffic.
3471 Sets a minimum of how many processes Squid is to try and keep available
3472 at all times. When traffic begins to rise above what the existing
3473 processes can handle this many more will be spawned up to the maximum
3474 configured. A minimum setting of 1 is required.
3478 Sets the maximum number of queued requests. A request is queued when
3479 no existing child is idle and no new child can be started due to
3480 numberofchildren limit. If the queued requests exceed queue size for
3481 more than 3 minutes squid aborts its operation. The default value is
3482 set to 2*numberofchildren.
3484 You must have at least one ssl_crtd process.
3487 NAME: sslcrtvalidator_program
3491 LOC: Ssl::TheConfig.ssl_crt_validator
3493 Specify the location and options of the executable for ssl_crt_validator
3496 Usage: sslcrtvalidator_program [ttl=...] [cache=n] path ...
3501 Limits how much memory Squid can use for caching validator
3502 responses. The default is 67108864 (i.e. 64 MB).
3503 Reconfiguration purges any excess entries. To disable caching,
3504 use cache=0. Currently, cache entry sizes are seriously
3505 underestimated. Even with that bug, a typical estimate for a
3506 single cache entry size would be at least a few kilobytes (the
3507 size of the PEM certificates sent to the validator).
3509 ttl=<seconds|"infinity">
3510 Approximately how long Squid may reuse the validator results
3511 for. The default is 3600 (i.e. 1 hour). Using ttl=infinity
3512 disables TTL checks. Reconfiguration does not affect TTLs of
3513 the already cached entries. To disable caching, use zero cache
3514 size, not zero TTL -- zero TTL allows reuse for the remainder
3515 of the second when the result was cached.
3518 NAME: sslcrtvalidator_children
3519 TYPE: HelperChildConfig
3521 DEFAULT: 32 startup=5 idle=1 concurrency=1
3522 LOC: Ssl::TheConfig.ssl_crt_validator_Children
3524 Specifies the maximum number of certificate validation processes that
3525 Squid may spawn (numberofchildren) and several related options. Using
3526 too few of these helper processes (a.k.a. "helpers") creates request
3527 queues. Using too many helpers wastes your system resources. Squid
3528 does not support spawning more than 32 helpers.
3530 Usage: numberofchildren [option]...
3532 The startup= and idle= options allow some measure of skew in your
3537 Sets the minimum number of processes to spawn when Squid
3538 starts or reconfigures. When set to zero the first request will
3539 cause spawning of the first child process to handle it.
3541 Starting too few children temporary slows Squid under load while it
3542 tries to spawn enough additional processes to cope with traffic.
3546 Sets a minimum of how many processes Squid is to try and keep available
3547 at all times. When traffic begins to rise above what the existing
3548 processes can handle this many more will be spawned up to the maximum
3549 configured. A minimum setting of 1 is required.
3553 The number of requests each certificate validator helper can handle in
3554 parallel. A value of 0 indicates the certificate validator does not
3555 support concurrency. Defaults to 1.
3557 When this directive is set to a value >= 1 then the protocol
3558 used to communicate with the helper is modified to include
3559 a request ID in front of the request/response. The request
3560 ID from the request must be echoed back with the response
3565 Sets the maximum number of queued requests. A request is queued when
3566 no existing child can accept it due to concurrency limit and no new
3567 child can be started due to numberofchildren limit. If the queued
3568 requests exceed queue size for more than 3 minutes squid aborts its
3569 operation. The default value is set to 2*numberofchildren.
3571 You must have at least one ssl_crt_validator process.
3575 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
3576 -----------------------------------------------------------------------------
3584 To specify other caches in a hierarchy, use the format:
3586 cache_peer hostname type http-port icp-port [options]
3591 # hostname type port port options
3592 # -------------------- -------- ----- ----- -----------
3593 cache_peer parent.foo.net parent 3128 3130 default
3594 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
3595 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
3596 cache_peer example.com parent 80 0 default
3597 cache_peer cdn.example.com sibling 3128 0
3599 type: either 'parent', 'sibling', or 'multicast'.
3601 proxy-port: The port number where the peer accept HTTP requests.
3602 For other Squid proxies this is usually 3128
3603 For web servers this is usually 80
3605 icp-port: Used for querying neighbor caches about objects.
3606 Set to 0 if the peer does not support ICP or HTCP.
3607 See ICP and HTCP options below for additional details.
3610 ==== ICP OPTIONS ====
3612 You MUST also set icp_port and icp_access explicitly when using these options.
3613 The defaults will prevent peer traffic using ICP.
3616 no-query Disable ICP queries to this neighbor.
3619 Indicates the named peer is a member of a multicast group.
3620 ICP queries will not be sent directly to the peer, but ICP
3621 replies will be accepted from it.
3623 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
3624 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
3627 To only send ICP queries to this neighbor infrequently.
3628 This is used to keep the neighbor round trip time updated
3629 and is usually used in conjunction with weighted-round-robin.
3632 ==== HTCP OPTIONS ====
3634 You MUST also set htcp_port and htcp_access explicitly when using these options.
3635 The defaults will prevent peer traffic using HTCP.
3638 htcp Send HTCP, instead of ICP, queries to the neighbor.
3639 You probably also want to set the "icp-port" to 4827
3640 instead of 3130. This directive accepts a comma separated
3641 list of options described below.
3643 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3645 htcp=no-clr Send HTCP to the neighbor but without
3646 sending any CLR requests. This cannot be used with
3649 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3650 This cannot be used with no-clr.
3653 Send HTCP to the neighbor including CLRs but only when
3654 they do not result from PURGE requests.
3657 Forward any HTCP CLR requests this proxy receives to the peer.
3660 ==== PEER SELECTION METHODS ====
3662 The default peer selection method is ICP, with the first responding peer
3663 being used as source. These options can be used for better load balancing.
3666 default This is a parent cache which can be used as a "last-resort"
3667 if a peer cannot be located by any of the peer-selection methods.
3668 If specified more than once, only the first is used.
3670 round-robin Load-Balance parents which should be used in a round-robin
3671 fashion in the absence of any ICP queries.
3672 weight=N can be used to add bias.
3674 weighted-round-robin
3675 Load-Balance parents which should be used in a round-robin
3676 fashion with the frequency of each parent being based on the
3677 round trip time. Closer parents are used more often.
3678 Usually used for background-ping parents.
3679 weight=N can be used to add bias.
3681 carp Load-Balance parents which should be used as a CARP array.
3682 The requests will be distributed among the parents based on the
3683 CARP load balancing hash function based on their weight.
3685 userhash Load-balance parents based on the client proxy_auth or ident username.
3687 sourcehash Load-balance parents based on the client source IP.
3690 To be used only for cache peers of type "multicast".
3691 ALL members of this multicast group have "sibling"
3692 relationship with it, not "parent". This is to a multicast
3693 group when the requested object would be fetched only from
3694 a "parent" cache, anyway. It's useful, e.g., when
3695 configuring a pool of redundant Squid proxies, being
3696 members of the same multicast group.
3699 ==== PEER SELECTION OPTIONS ====
3701 weight=N use to affect the selection of a peer during any weighted
3702 peer-selection mechanisms.
3703 The weight must be an integer; default is 1,
3704 larger weights are favored more.
3705 This option does not affect parent selection if a peering
3706 protocol is not in use.
3708 basetime=N Specify a base amount to be subtracted from round trip
3710 It is subtracted before division by weight in calculating
3711 which parent to fectch from. If the rtt is less than the
3712 base time the rtt is set to a minimal value.
3714 ttl=N Specify a TTL to use when sending multicast ICP queries
3716 Only useful when sending to a multicast group.
3717 Because we don't accept ICP replies from random
3718 hosts, you must configure other group members as
3719 peers with the 'multicast-responder' option.
3721 no-delay To prevent access to this neighbor from influencing the
3724 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3725 enabled) for this host from the specified URL rather
3726 than the Squid default location.
3729 ==== CARP OPTIONS ====
3731 carp-key=key-specification
3732 use a different key than the full URL to hash against the peer.
3733 the key-specification is a comma-separated list of the keywords
3734 scheme, host, port, path, params
3735 Order is not important.
3737 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3739 originserver Causes this parent to be contacted as an origin server.
3740 Meant to be used in accelerator setups when the peer
3744 Set the Host header of requests forwarded to this peer.
3745 Useful in accelerator setups where the server (peer)
3746 expects a certain domain name but clients may request
3747 others. ie example.com or www.example.com
3749 no-digest Disable request of cache digests.
3752 Disables requesting ICMP RTT database (NetDB).
3755 ==== AUTHENTICATION OPTIONS ====
3758 If this is a personal/workgroup proxy and your parent
3759 requires proxy authentication.
3761 Note: The string can include URL escapes (i.e. %20 for
3762 spaces). This also means % must be written as %%.
3765 Send login details received from client to this peer.
3766 Both Proxy- and WWW-Authorization headers are passed
3767 without alteration to the peer.
3768 Authentication is not required by Squid for this to work.
3770 Note: This will pass any form of authentication but
3771 only Basic auth will work through a proxy unless the
3772 connection-auth options are also used.
3774 login=PASS Send login details received from client to this peer.
3775 Authentication is not required by this option.
3777 If there are no client-provided authentication headers
3778 to pass on, but username and password are available
3779 from an external ACL user= and password= result tags
3780 they may be sent instead.
3782 Note: To combine this with proxy_auth both proxies must
3783 share the same user database as HTTP only allows for
3784 a single login (one for proxy, one for origin server).
3785 Also be warned this will expose your users proxy
3786 password to the peer. USE WITH CAUTION
3789 Send the username to the upstream cache, but with a
3790 fixed password. This is meant to be used when the peer
3791 is in another administrative domain, but it is still
3792 needed to identify each user.
3793 The star can optionally be followed by some extra
3794 information which is added to the username. This can
3795 be used to identify this proxy to the peer, similar to
3796 the login=username:password option above.
3799 If this is a personal/workgroup proxy and your parent
3800 requires a secure proxy authentication.
3801 The first principal from the default keytab or defined by
3802 the environment variable KRB5_KTNAME will be used.
3804 WARNING: The connection may transmit requests from multiple
3805 clients. Negotiate often assumes end-to-end authentication
3806 and a single-client. Which is not strictly true here.
3808 login=NEGOTIATE:principal_name
3809 If this is a personal/workgroup proxy and your parent
3810 requires a secure proxy authentication.
3811 The principal principal_name from the default keytab or
3812 defined by the environment variable KRB5_KTNAME will be
3815 WARNING: The connection may transmit requests from multiple
3816 clients. Negotiate often assumes end-to-end authentication
3817 and a single-client. Which is not strictly true here.
3819 connection-auth=on|off
3820 Tell Squid that this peer does or not support Microsoft
3821 connection oriented authentication, and any such
3822 challenges received from there should be ignored.
3823 Default is auto to automatically determine the status
3827 Do not use a keytab to authenticate to a peer when
3828 login=NEGOTIATE is specified. Let the GSSAPI
3829 implementation determine which already existing
3830 credentials cache to use instead.
3833 ==== SSL / HTTPS / TLS OPTIONS ====
3835 tls Encrypt connections to this peer with TLS.
3837 sslcert=/path/to/ssl/certificate
3838 A client X.509 certificate to use when connecting to
3841 sslkey=/path/to/ssl/key
3842 The private key corresponding to sslcert above.
3844 If sslkey= is not specified sslcert= is assumed to
3845 reference a PEM file containing both the certificate
3848 sslcipher=... The list of valid SSL ciphers to use when connecting
3852 The minimum TLS protocol version to permit. To control
3853 SSLv3 use the tls-options= parameter.
3854 Supported Values: 1.0 (default), 1.1, 1.2
3856 tls-options=... Specify various TLS implementation options.
3858 OpenSSL options most important are:
3860 NO_SSLv3 Disallow the use of SSLv3
3863 Always create a new key when using
3864 temporary/ephemeral DH key exchanges
3867 Disable use of RFC5077 session tickets.
3868 Some servers may have problems
3869 understanding the TLS extension due
3870 to ambiguous specification in RFC4507.
3872 ALL Enable various bug workarounds
3873 suggested as "harmless" by OpenSSL
3874 Be warned that this reduces SSL/TLS
3875 strength to some attacks.
3877 See the OpenSSL SSL_CTX_set_options documentation for a
3880 GnuTLS options most important are:
3883 Disable use of RFC5077 session tickets.
3884 Some servers may have problems
3885 understanding the TLS extension due
3886 to ambiguous specification in RFC4507.
3888 See the GnuTLS Priority Strings documentation
3889 for a more complete list.
3890 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
3892 tls-cafile= PEM file containing CA certificates to use when verifying
3893 the peer certificate. May be repeated to load multiple files.
3895 sslcapath=... A directory containing additional CA certificates to
3896 use when verifying the peer certificate.
3897 Requires OpenSSL or LibreSSL.
3899 sslcrlfile=... A certificate revocation list file to use when
3900 verifying the peer certificate.
3902 sslflags=... Specify various flags modifying the SSL implementation:
3905 Accept certificates even if they fail to
3909 Don't verify the peer certificate
3910 matches the server name
3912 ssldomain= The peer name as advertised in it's certificate.
3913 Used for verifying the correctness of the received peer
3914 certificate. If not specified the peer hostname will be
3917 front-end-https[=off|on|auto]
3918 Enable the "Front-End-Https: On" header needed when
3919 using Squid as a SSL frontend in front of Microsoft OWA.
3920 See MS KB document Q307347 for details on this header.
3921 If set to auto the header will only be added if the
3922 request is forwarded as a https:// URL.
3924 tls-default-ca[=off]
3925 Whether to use the system Trusted CAs. Default is ON.
3927 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
3929 ==== GENERAL OPTIONS ====
3932 A peer-specific connect timeout.
3933 Also see the peer_connect_timeout directive.
3935 connect-fail-limit=N
3936 How many times connecting to a peer must fail before
3937 it is marked as down. Standby connection failures
3938 count towards this limit. Default is 10.
3940 allow-miss Disable Squid's use of only-if-cached when forwarding
3941 requests to siblings. This is primarily useful when
3942 icp_hit_stale is used by the sibling. Excessive use
3943 of this option may result in forwarding loops. One way
3944 to prevent peering loops when using this option, is to
3945 deny cache peer usage on requests from a peer:
3947 cache_peer_access peerName deny fromPeer
3949 max-conn=N Limit the number of concurrent connections the Squid
3950 may open to this peer, including already opened idle
3951 and standby connections. There is no peer-specific
3952 connection limit by default.
3954 A peer exceeding the limit is not used for new
3955 requests unless a standby connection is available.
3957 max-conn currently works poorly with idle persistent
3958 connections: When a peer reaches its max-conn limit,
3959 and there are idle persistent connections to the peer,
3960 the peer may not be selected because the limiting code
3961 does not know whether Squid can reuse those idle
3964 standby=N Maintain a pool of N "hot standby" connections to an
3965 UP peer, available for requests when no idle
3966 persistent connection is available (or safe) to use.
3967 By default and with zero N, no such pool is maintained.
3968 N must not exceed the max-conn limit (if any).
3970 At start or after reconfiguration, Squid opens new TCP
3971 standby connections until there are N connections
3972 available and then replenishes the standby pool as
3973 opened connections are used up for requests. A used
3974 connection never goes back to the standby pool, but
3975 may go to the regular idle persistent connection pool
3976 shared by all peers and origin servers.
3978 Squid never opens multiple new standby connections
3979 concurrently. This one-at-a-time approach minimizes
3980 flooding-like effect on peers. Furthermore, just a few
3981 standby connections should be sufficient in most cases
3982 to supply most new requests with a ready-to-use
3985 Standby connections obey server_idle_pconn_timeout.
3986 For the feature to work as intended, the peer must be
3987 configured to accept and keep them open longer than
3988 the idle timeout at the connecting Squid, to minimize
3989 race conditions typical to idle used persistent
3990 connections. Default request_timeout and
3991 server_idle_pconn_timeout values ensure such a
3994 name=xxx Unique name for the peer.
3995 Required if you have multiple cache_peers with the same hostname.
3996 Defaults to cache_peer hostname when not explicitly specified.
3998 Other directives (e.g., cache_peer_access), cache manager reports,
3999 and cache.log messages use this name to refer to this cache_peer.
4001 The cache_peer name value affects hashing-based peer selection
4002 methods (e.g., carp and sourcehash).
4004 Can be used by outgoing access controls through the
4007 The name value preserves configured spelling, but name uniqueness
4008 checks and name-based search are case-insensitive.
4010 no-tproxy Do not use the client-spoof TPROXY support when forwarding
4011 requests to this peer. Use normal address selection instead.
4012 This overrides the spoof_client_ip ACL.
4014 proxy-only objects fetched from the peer will not be stored locally.
4018 NAME: cache_peer_access
4021 DEFAULT_DOC: No peer usage restrictions.
4024 Restricts usage of cache_peer proxies.
4027 cache_peer_access peer-name allow|deny [!]aclname ...
4029 For the required peer-name parameter, use either the value of the
4030 cache_peer name=value parameter or, if name=value is missing, the
4031 cache_peer hostname parameter.
4033 This directive narrows down the selection of peering candidates, but
4034 does not determine the order in which the selected candidates are
4035 contacted. That order is determined by the peer selection algorithms
4036 (see PEER SELECTION sections in the cache_peer documentation).
4038 If a deny rule matches, the corresponding peer will not be contacted
4039 for the current transaction -- Squid will not send ICP queries and
4040 will not forward HTTP requests to that peer. An allow match leaves
4041 the corresponding peer in the selection. The first match for a given
4042 peer wins for that peer.
4044 The relative order of cache_peer_access directives for the same peer
4045 matters. The relative order of any two cache_peer_access directives
4046 for different peers does not matter. To ease interpretation, it is a
4047 good idea to group cache_peer_access directives for the same peer
4050 A single cache_peer_access directive may be evaluated multiple times
4051 for a given transaction because individual peer selection algorithms
4052 may check it independently from each other. These redundant checks
4053 may be optimized away in future Squid versions.
4055 This clause only supports fast acl types.
4056 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4060 NAME: neighbor_type_domain
4061 TYPE: hostdomaintype
4063 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
4066 Modify the cache_peer neighbor type when passing requests
4067 about specific domains to the peer.
4070 neighbor_type_domain peer-name parent|sibling domain...
4072 For the required peer-name parameter, use either the value of the
4073 cache_peer name=value parameter or, if name=value is missing, the
4074 cache_peer hostname parameter.
4077 cache_peer foo.example.com parent 3128 3130
4078 neighbor_type_domain foo.example.com sibling .au .de
4080 The above configuration treats all requests to foo.example.com as a
4081 parent proxy unless the request is for a .au or .de ccTLD domain name.
4084 NAME: dead_peer_timeout
4088 LOC: Config.Timeout.deadPeer
4090 This controls how long Squid waits to declare a peer cache
4091 as "dead." If there are no ICP replies received in this
4092 amount of time, Squid will declare the peer dead and not
4093 expect to receive any further ICP replies. However, it
4094 continues to send ICP queries, and will mark the peer as
4095 alive upon receipt of the first subsequent ICP reply.
4097 This timeout also affects when Squid expects to receive ICP
4098 replies from peers. If more than 'dead_peer' seconds have
4099 passed since the last ICP reply was received, Squid will not
4100 expect to receive an ICP reply on the next query. Thus, if
4101 your time between requests is greater than this timeout, you
4102 will see a lot of requests sent DIRECT to origin servers
4103 instead of to your parents.
4106 NAME: forward_max_tries
4109 LOC: Config.forward_max_tries
4111 Limits the number of attempts to forward the request.
4113 For the purpose of this limit, Squid counts all high-level request
4114 forwarding attempts, including any same-destination retries after
4115 certain persistent connection failures and any attempts to use a
4116 different peer. However, these low-level attempts are not counted:
4117 * connection reopening attempts (enabled using connect_retries)
4118 * unfinished Happy Eyeballs connection attempts (prevented by setting
4119 happy_eyeballs_connect_limit to 0)
4121 See also: forward_timeout, connect_retries, and %request_attempts.
4125 MEMORY CACHE OPTIONS
4126 -----------------------------------------------------------------------------
4133 LOC: Config.memMaxSize
4135 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
4136 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
4137 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
4138 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
4140 'cache_mem' specifies the ideal amount of memory to be used
4142 * In-Transit objects
4144 * Negative-Cached objects
4146 Data for these objects are stored in 4 KB blocks. This
4147 parameter specifies the ideal upper limit on the total size of
4148 4 KB blocks allocated. In-Transit objects take the highest
4151 In-transit objects have priority over the others. When
4152 additional space is needed for incoming data, negative-cached
4153 and hot objects will be released. In other words, the
4154 negative-cached and hot objects will fill up any unused space
4155 not needed for in-transit objects.
4157 If circumstances require, this limit will be exceeded.
4158 Specifically, if your incoming request rate requires more than
4159 'cache_mem' of memory to hold in-transit objects, Squid will
4160 exceed this limit to satisfy the new requests. When the load
4161 decreases, blocks will be freed until the high-water mark is
4162 reached. Thereafter, blocks will be used to store hot
4165 If shared memory caching is enabled, Squid does not use the shared
4166 cache space for in-transit objects, but they still consume as much
4167 local memory as they need. For more details about the shared memory
4168 cache, see memory_cache_shared.
4171 NAME: maximum_object_size_in_memory
4175 LOC: Config.Store.maxInMemObjSize
4177 Objects greater than this size will not be attempted to kept in
4178 the memory cache. This should be set high enough to keep objects
4179 accessed frequently in memory to improve performance whilst low
4180 enough to keep larger objects from hoarding cache_mem.
4183 NAME: memory_cache_shared
4186 LOC: Config.memShared
4188 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
4190 Controls whether the memory cache is shared among SMP workers.
4192 The shared memory cache is meant to occupy cache_mem bytes and replace
4193 the non-shared memory cache, although some entities may still be
4194 cached locally by workers for now (e.g., internal and in-transit
4195 objects may be served from a local memory cache even if shared memory
4196 caching is enabled).
4198 By default, the memory cache is shared if and only if all of the
4199 following conditions are satisfied: Squid runs in SMP mode with
4200 multiple workers, cache_mem is positive, and Squid environment
4201 supports required IPC primitives (e.g., POSIX shared memory segments
4202 and GCC-style atomic operations).
4204 To avoid blocking locks, shared memory uses opportunistic algorithms
4205 that do not guarantee that every cachable entity that could have been
4206 shared among SMP workers will actually be shared.
4209 NAME: memory_cache_mode
4213 DEFAULT_DOC: Keep the most recently fetched objects in memory
4215 Controls which objects to keep in the memory cache (cache_mem)
4217 always Keep most recently fetched objects in memory (default)
4219 disk Only disk cache hits are kept in memory, which means
4220 an object must first be cached on disk and then hit
4221 a second time before cached in memory.
4223 network Only objects fetched from network is kept in memory
4226 NAME: memory_replacement_policy
4228 LOC: Config.memPolicy
4231 The memory replacement policy parameter determines which
4232 objects are purged from memory when memory space is needed.
4234 See cache_replacement_policy for details on algorithms.
4239 -----------------------------------------------------------------------------
4242 NAME: cache_replacement_policy
4244 LOC: Config.replPolicy
4247 The cache replacement policy parameter determines which
4248 objects are evicted (replaced) when disk space is needed.
4250 lru : Squid's original list based LRU policy
4251 heap GDSF : Greedy-Dual Size Frequency
4252 heap LFUDA: Least Frequently Used with Dynamic Aging
4253 heap LRU : LRU policy implemented using a heap
4255 Applies to any cache_dir lines listed below this directive.
4257 The LRU policies keeps recently referenced objects.
4259 The heap GDSF policy optimizes object hit rate by keeping smaller
4260 popular objects in cache so it has a better chance of getting a
4261 hit. It achieves a lower byte hit rate than LFUDA though since
4262 it evicts larger (possibly popular) objects.
4264 The heap LFUDA policy keeps popular objects in cache regardless of
4265 their size and thus optimizes byte hit rate at the expense of
4266 hit rate since one large, popular object will prevent many
4267 smaller, slightly less popular objects from being cached.
4269 Both policies utilize a dynamic aging mechanism that prevents
4270 cache pollution that can otherwise occur with frequency-based
4271 replacement policies.
4273 NOTE: if using the LFUDA replacement policy you should increase
4274 the value of maximum_object_size above its default of 4 MB to
4275 to maximize the potential byte hit rate improvement of LFUDA.
4277 For more information about the GDSF and LFUDA cache replacement
4278 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
4279 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
4282 NAME: minimum_object_size
4286 DEFAULT_DOC: no limit
4287 LOC: Config.Store.minObjectSize
4289 Objects smaller than this size will NOT be saved on disk. The
4290 value is specified in bytes, and the default is 0 KB, which
4291 means all responses can be stored.
4294 NAME: maximum_object_size
4298 LOC: Config.Store.maxObjectSize
4300 Set the default value for max-size parameter on any cache_dir.
4301 The value is specified in bytes, and the default is 4 MB.
4303 If you wish to get a high BYTES hit ratio, you should probably
4304 increase this (one 32 MB object hit counts for 3200 10KB
4307 If you wish to increase hit ratio more than you want to
4308 save bandwidth you should leave this low.
4310 NOTE: if using the LFUDA replacement policy you should increase
4311 this value to maximize the byte hit rate improvement of LFUDA!
4312 See cache_replacement_policy for a discussion of this policy.
4318 DEFAULT_DOC: No disk cache. Store cache objects only in memory.
4319 LOC: Config.cacheSwap
4322 cache_dir Type Directory-Name Fs-specific-data [options]
4324 You can specify multiple cache_dir lines to spread the
4325 cache among different disk partitions.
4327 Type specifies the kind of storage system to use. Only "ufs"
4328 is built by default. To enable any of the other storage systems
4329 see the --enable-storeio configure option.
4331 'Directory' is a top-level directory where cache swap
4332 files will be stored. If you want to use an entire disk
4333 for caching, this can be the mount-point directory.
4334 The directory must exist and be writable by the Squid
4335 process. Squid will NOT create this directory for you.
4337 Rock is currently the only SMP-aware cache_dir type. Using other
4338 store types in configurations with multiple workers is not
4339 supported and may lead to HTTP violations or undefined behavior,
4340 even when each such cache_dir is given a dedicated worker using
4341 configuration conditionals.
4344 ==== The ufs store type ====
4346 "ufs" is the old well-known Squid storage format that has always
4350 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
4352 'Mbytes' is the amount of disk space (MB) to use under this
4353 directory. The default is 100 MB. Change this to suit your
4354 configuration. Do NOT put the size of your disk drive here.
4355 Instead, if you want Squid to use the entire disk drive,
4356 subtract 20% and use that value.
4358 'L1' is the number of first-level subdirectories which
4359 will be created under the 'Directory'. The default is 16.
4361 'L2' is the number of second-level subdirectories which
4362 will be created under each first-level directory. The default
4366 ==== The aufs store type ====
4368 "aufs" uses the same storage format as "ufs", utilizing
4369 POSIX-threads to avoid blocking the main Squid process on
4370 disk-I/O. This was formerly known in Squid as async-io.
4373 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
4375 see argument descriptions under ufs above
4378 ==== The diskd store type ====
4380 "diskd" uses the same storage format as "ufs", utilizing a
4381 separate process to avoid blocking the main Squid process on
4385 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
4387 see argument descriptions under ufs above
4389 Q1 specifies the number of unacknowledged I/O requests when Squid
4390 stops opening new files. If this many messages are in the queues,
4391 Squid won't open new files. Default is 64
4393 Q2 specifies the number of unacknowledged messages when Squid
4394 starts blocking. If this many messages are in the queues,
4395 Squid blocks until it receives some replies. Default is 72
4397 When Q1 < Q2 (the default), the cache directory is optimized
4398 for lower response time at the expense of a decrease in hit
4399 ratio. If Q1 > Q2, the cache directory is optimized for
4400 higher hit ratio at the expense of an increase in response
4404 ==== The rock store type ====
4407 cache_dir rock Directory-Name Mbytes [options]
4409 The Rock Store type is a database-style storage. All cached
4410 entries are stored in a "database" file, using fixed-size slots.
4411 A single entry occupies one or more slots.
4413 If possible, Squid using Rock Store creates a dedicated kid
4414 process called "disker" to avoid blocking Squid worker(s) on disk
4415 I/O. One disker kid is created for each rock cache_dir. Diskers
4416 are created only when Squid, running in daemon mode, has support
4417 for the IpcIo disk I/O module.
4419 swap-timeout=msec: Squid will not start writing a miss to or
4420 reading a hit from disk if it estimates that the swap operation
4421 will take more than the specified number of milliseconds. By
4422 default and when set to zero, disables the disk I/O time limit
4423 enforcement. Ignored when using blocking I/O module because
4424 blocking synchronous I/O does not allow Squid to estimate the
4425 expected swap wait time.
4427 max-swap-rate=swaps/sec: Artificially limits disk access using
4428 the specified I/O rate limit. Swap out requests that
4429 would cause the average I/O rate to exceed the limit are
4430 delayed. Individual swap in requests (i.e., hits or reads) are
4431 not delayed, but they do contribute to measured swap rate and
4432 since they are placed in the same FIFO queue as swap out
4433 requests, they may wait longer if max-swap-rate is smaller.
4434 This is necessary on file systems that buffer "too
4435 many" writes and then start blocking Squid and other processes
4436 while committing those writes to disk. Usually used together
4437 with swap-timeout to avoid excessive delays and queue overflows
4438 when disk demand exceeds available disk "bandwidth". By default
4439 and when set to zero, disables the disk I/O rate limit
4440 enforcement. Currently supported by IpcIo module only.
4442 slot-size=bytes: The size of a database "record" used for
4443 storing cached responses. A cached response occupies at least
4444 one slot and all database I/O is done using individual slots so
4445 increasing this parameter leads to more disk space waste while
4446 decreasing it leads to more disk I/O overheads. Should be a
4447 multiple of your operating system I/O page size. Defaults to
4448 16KBytes. A housekeeping header is stored with each slot and
4449 smaller slot-sizes will be rejected. The header is smaller than
4453 ==== COMMON OPTIONS ====
4455 no-store no new objects should be stored to this cache_dir.
4457 min-size=n the minimum object size in bytes this cache_dir
4458 will accept. It's used to restrict a cache_dir
4459 to only store large objects (e.g. AUFS) while
4460 other stores are optimized for smaller objects
4464 max-size=n the maximum object size in bytes this cache_dir
4466 The value in maximum_object_size directive sets
4467 the default unless more specific details are
4468 available (ie a small store capacity).
4470 Note: To make optimal use of the max-size limits you should order
4471 the cache_dir lines with the smallest max-size value first.
4475 # Uncomment and adjust the following to add a disk cache directory.
4476 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
4480 NAME: store_dir_select_algorithm
4482 LOC: Config.store_dir_select_algorithm
4485 How Squid selects which cache_dir to use when the response
4486 object will fit into more than one.
4488 Regardless of which algorithm is used the cache_dir min-size
4489 and max-size parameters are obeyed. As such they can affect
4490 the selection algorithm by limiting the set of considered
4497 This algorithm is suited to caches with similar cache_dir
4498 sizes and disk speeds.
4500 The disk with the least I/O pending is selected.
4501 When there are multiple disks with the same I/O load ranking
4502 the cache_dir with most available capacity is selected.
4504 When a mix of cache_dir sizes are configured the faster disks
4505 have a naturally lower I/O loading and larger disks have more
4506 capacity. So space used to store objects and data throughput
4507 may be very unbalanced towards larger disks.
4512 This algorithm is suited to caches with unequal cache_dir
4515 Each cache_dir is selected in a rotation. The next suitable
4518 Available cache_dir capacity is only considered in relation
4519 to whether the object will fit and meets the min-size and
4520 max-size parameters.
4522 Disk I/O loading is only considered to prevent overload on slow
4523 disks. This algorithm does not spread objects by size, so any
4524 I/O loading per-disk may appear very unbalanced and volatile.
4526 If several cache_dirs use similar min-size, max-size, or other
4527 limits to to reject certain responses, then do not group such
4528 cache_dir lines together, to avoid round-robin selection bias
4529 towards the first cache_dir after the group. Instead, interleave
4530 cache_dir lines from different groups. For example:
4532 store_dir_select_algorithm round-robin
4533 cache_dir rock /hdd1 ... min-size=100000
4534 cache_dir rock /ssd1 ... max-size=99999
4535 cache_dir rock /hdd2 ... min-size=100000
4536 cache_dir rock /ssd2 ... max-size=99999
4537 cache_dir rock /hdd3 ... min-size=100000
4538 cache_dir rock /ssd3 ... max-size=99999
4541 NAME: paranoid_hit_validation
4542 COMMENT: time-units-small
4543 TYPE: time_nanoseconds
4545 DEFAULT_DOC: validation disabled
4546 LOC: Config.paranoid_hit_validation
4548 Controls whether Squid should perform paranoid validation of cache entry
4549 metadata integrity every time a cache entry is hit. This low-level
4550 validation should always succeed. Each failed validation results in a
4551 cache miss, a BUG line reported to cache.log, and the invalid entry
4552 marked as unusable (and eventually purged from the cache).
4554 Squid can only validate shared cache memory and rock cache_dir entries.
4556 * Zero (default) value means that the validation is disabled.
4558 * Positive values enable validation:
4559 - values less than 1 day approximate the maximum time that Squid is allowed
4560 to spend validating a single cache hit.
4561 - values greater or equal to 1 day are considered as no limitation:
4562 in this case all checks will be performed, regardless of how much time
4565 Hits are usually stored using 16KB slots (for rock, the size is
4566 configurable via cache_dir slot-size). Larger hits require scanning more
4567 slots and, hence, take more time. When validation is enabled, at least one
4568 slot is always validated, regardless of the configured time limit.
4570 A worker process validating an entry cannot do anything else (i.e. the
4571 validation is blocking). The validation overhead is environment dependent,
4572 but developers have observed Squid spending 3-10 microseconds to check each
4573 slot of a Rock or shared memory hit entry. If Squid cuts validation short
4574 because it runs out of configured time, it treats the entry as valid.
4576 When hit validation is enabled, its statistics is included in Cache
4577 Manager mgr:counters, mgr:5min, and mgr:60min reports.
4580 NAME: max_open_disk_fds
4582 LOC: Config.max_open_disk_fds
4584 DEFAULT_DOC: no limit
4586 To avoid having disk as the I/O bottleneck Squid can optionally
4587 bypass the on-disk cache if more than this amount of disk file
4588 descriptors are open.
4590 A value of 0 indicates no limit.
4593 NAME: cache_swap_low
4594 COMMENT: (percent, 0-100)
4597 LOC: Config.Swap.lowWaterMark
4599 The low-water mark for AUFS/UFS/diskd cache object eviction by
4600 the cache_replacement_policy algorithm.
4602 Removal begins when the swap (disk) usage of a cache_dir is
4603 above this low-water mark and attempts to maintain utilization
4604 near the low-water mark.
4606 As swap utilization increases towards the high-water mark set
4607 by cache_swap_high object eviction becomes more aggressive.
4609 The value difference in percentages between low- and high-water
4610 marks represent an eviction rate of 300 objects per second and
4611 the rate continues to scale in aggressiveness by multiples of
4612 this above the high-water mark.
4614 Defaults are 90% and 95%. If you have a large cache, 5% could be
4615 hundreds of MB. If this is the case you may wish to set these
4616 numbers closer together.
4618 See also cache_swap_high and cache_replacement_policy
4621 NAME: cache_swap_high
4622 COMMENT: (percent, 0-100)
4625 LOC: Config.Swap.highWaterMark
4627 The high-water mark for AUFS/UFS/diskd cache object eviction by
4628 the cache_replacement_policy algorithm.
4630 Removal begins when the swap (disk) usage of a cache_dir is
4631 above the low-water mark set by cache_swap_low and attempts to
4632 maintain utilization near the low-water mark.
4634 As swap utilization increases towards this high-water mark object
4635 eviction becomes more aggressive.
4637 The value difference in percentages between low- and high-water
4638 marks represent an eviction rate of 300 objects per second and
4639 the rate continues to scale in aggressiveness by multiples of
4640 this above the high-water mark.
4642 Defaults are 90% and 95%. If you have a large cache, 5% could be
4643 hundreds of MB. If this is the case you may wish to set these
4644 numbers closer together.
4646 See also cache_swap_low and cache_replacement_policy
4651 -----------------------------------------------------------------------------
4658 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
4662 logformat <name> <format specification>
4664 Defines an access log format.
4666 The <format specification> is a string with embedded % format codes
4668 % format codes all follow the same basic structure where all
4669 components but the formatcode are optional and usually unnecessary,
4670 especially when dealing with common codes.
4672 % [encoding] [-] [[0]width] [{arg}] formatcode [{arg}]
4674 encoding escapes or otherwise protects "special" characters:
4676 " Quoted string encoding where quote(") and
4677 backslash(\) characters are \-escaped while
4678 CR, LF, and TAB characters are encoded as \r,
4679 \n, and \t two-character sequences.
4681 [ Custom Squid encoding where percent(%), square
4682 brackets([]), backslash(\) and characters with
4683 codes outside of [32,126] range are %-encoded.
4684 SP is not encoded. Used by log_mime_hdrs.
4686 # URL encoding (a.k.a. percent-encoding) where
4687 all URL unsafe and control characters (per RFC
4688 1738) are %-encoded.
4690 / Shell-like encoding where quote(") and
4691 backslash(\) characters are \-escaped while CR
4692 and LF characters are encoded as \r and \n
4693 two-character sequences. Values containing SP
4694 character(s) are surrounded by quotes(").
4696 ' Raw/as-is encoding with no escaping/quoting.
4698 Default encoding: When no explicit encoding is
4699 specified, each %code determines its own encoding.
4700 Most %codes use raw/as-is encoding, but some codes use
4701 a so called "pass-through URL encoding" where all URL
4702 unsafe and control characters (per RFC 1738) are
4703 %-encoded, but the percent character(%) is left as is.
4707 width minimum and/or maximum field width:
4708 [width_min][.width_max]
4709 When minimum starts with 0, the field is zero-padded.
4710 String values exceeding maximum width are truncated.
4712 {arg} argument such as header name etc. This field may be
4713 placed before or after the token, but not both at once.
4717 % a literal % character
4719 byte{value} Adds a single byte with the given value (e.g., %byte{10}
4720 adds an ASCII LF character a.k.a. "new line" or "\n"). The value
4721 parameter is required and must be a positive decimal integer not
4722 exceeding 255. Zero-valued bytes (i.e. ASCII NUL characters) are
4725 sn Unique sequence number per log line entry
4726 err_code The ID of an error response served by Squid or
4727 a similar internal error identifier.
4729 err_detail Additional err_code-dependent error information. Multiple
4730 details are separated by the plus sign ('+'). Admins should not
4731 rely on a particular detail listing order, the uniqueness of the
4732 entries, or individual detail text stability. All those properties
4733 depend on many unstable factors, including external libraries.
4735 note The annotation specified by the argument. Also
4736 logs the adaptation meta headers set by the
4737 adaptation_meta configuration parameter.
4738 If no argument given all annotations logged.
4739 The argument may include a separator to use with
4742 By default, multiple note values are separated with ","
4743 and multiple notes are separated with "\r\n".
4744 When logging named notes with %{name}note, the
4745 explicitly configured separator is used between note
4746 values. When logging all notes with %note, the
4747 explicitly configured separator is used between
4748 individual notes. There is currently no way to
4749 specify both value and notes separators when logging
4750 all notes with %note.
4751 master_xaction The master transaction identifier is an unsigned
4752 integer. These IDs are guaranteed to monotonically
4753 increase within a single worker process lifetime, with
4754 higher values corresponding to transactions that were
4755 accepted or initiated later. Due to current implementation
4756 deficiencies, some IDs are skipped (i.e. never logged).
4757 Concurrent workers and restarted workers use similar,
4758 overlapping sequences of master transaction IDs.
4760 Connection related format codes:
4762 >a Client source IP address
4764 >p Client source port
4765 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
4766 >la Local IP address the client connected to
4767 >lp Local port number the client connected to
4768 >qos Client connection TOS/DSCP value set by Squid
4769 >nfmark Client connection netfilter packet MARK set by Squid
4771 transport::>connection_id Identifies a transport connection
4772 accepted by Squid (e.g., a connection carrying the
4773 logged HTTP request). Currently, Squid only supports
4774 TCP transport connections.
4776 The logged identifier is an unsigned integer. These
4777 IDs are guaranteed to monotonically increase within a
4778 single worker process lifetime, with higher values
4779 corresponding to connections that were accepted later.
4780 Many IDs are skipped (i.e. never logged). Concurrent
4781 workers and restarted workers use similar, partially
4782 overlapping sequences of IDs.
4784 la Local listening IP address the client connection was connected to.
4785 lp Local listening port number the client connection was connected to.
4787 <a Server IP address of the last server or peer connection
4788 <A Server FQDN or peer name
4789 <p Server port number of the last server or peer connection
4790 <la Local IP address of the last server or peer connection
4791 <lp Local port number of the last server or peer connection
4792 <qos Server connection TOS/DSCP value set by Squid
4793 <nfmark Server connection netfilter packet MARK set by Squid
4795 >handshake Raw client handshake
4796 Initial client bytes received by Squid on a newly
4797 accepted TCP connection or inside a just established
4798 CONNECT tunnel. Squid stops accumulating handshake
4799 bytes as soon as the handshake parser succeeds or
4800 fails (determining whether the client is using the
4803 For HTTP clients, the handshake is the request line.
4804 For TLS clients, the handshake consists of all TLS
4805 records up to and including the TLS record that
4806 contains the last byte of the first ClientHello
4807 message. For clients using an unsupported protocol,
4808 this field contains the bytes received by Squid at the
4809 time of the handshake parsing failure.
4811 See the on_unsupported_protocol directive for more
4812 information on Squid handshake traffic expectations.
4814 Current support is limited to these contexts:
4815 - http_port connections, but only when the
4816 on_unsupported_protocol directive is in use.
4817 - https_port connections (and CONNECT tunnels) that
4818 are subject to the ssl_bump peek or stare action.
4820 To protect binary handshake data, this field is always
4821 base64-encoded (RFC 4648 Section 4). If logformat
4822 field encoding is configured, that encoding is applied
4823 on top of base64. Otherwise, the computed base64 value
4826 Time related format codes:
4828 ts Seconds since epoch
4829 tu subsecond time (milliseconds)
4830 tl Local time. Optional strftime format argument
4831 default %d/%b/%Y:%H:%M:%S %z
4832 tg GMT time. Optional strftime format argument
4833 default %d/%b/%Y:%H:%M:%S %z
4834 tr Response time (milliseconds)
4835 dt Total time spent making DNS lookups (milliseconds)
4836 tS Approximate master transaction start time in
4837 <full seconds since epoch>.<fractional seconds> format.
4838 Currently, Squid considers the master transaction
4839 started when a complete HTTP request header initiating
4840 the transaction is received from the client. This is
4841 the same value that Squid uses to calculate transaction
4842 response time when logging %tr to access.log. Currently,
4843 Squid uses millisecond resolution for %tS values,
4844 similar to the default access.log "current time" field
4847 busy_time Time spent in transaction-related code (nanoseconds)
4848 This cumulative measurement excludes periods of time when the
4849 transaction was waiting (e.g., for a server or helper response)
4850 while Squid worked on other transactions or was engaged in
4851 transaction-unrelated activities (e.g., generating a cache index).
4852 In other words, this measurement represents the total amount of
4853 physical time when Squid was busy working on this transaction.
4855 WARNING: This measurement relies on Squid transaction context
4856 tracking features that currently have known context leak bugs and
4857 coverage gaps. Until those features are fully implemented, logged
4858 values may significantly understate or exaggerate actual times.
4859 Do not use this measurement unless you know it works in your case.
4861 Access Control related format codes:
4863 et Tag returned by external acl
4864 ea Log string returned by external acl
4865 un User name (any available)
4866 ul User name from authentication
4867 ue User name from external acl helper
4868 ui User name from ident
4869 un A user name. Expands to the first available name
4870 from the following list of information sources:
4871 - authenticated user name, like %ul
4872 - user name supplied by an external ACL, like %ue
4873 - SSL client name, like %us
4874 - ident user name, like %ui
4875 credentials Client credentials. The exact meaning depends on
4876 the authentication scheme: For Basic authentication,
4877 it is the password; for Digest, the realm sent by the
4878 client; for NTLM and Negotiate, the client challenge
4879 or client credentials prefixed with "YR " or "KK ".
4881 HTTP related format codes:
4885 [http::]rm Request method (GET/POST etc)
4886 [http::]>rm Request method from client
4887 [http::]<rm Request method sent to server or peer
4889 [http::]ru Request URL received (or computed) and sanitized
4891 Logs request URI received from the client, a
4892 request adaptation service, or a request
4893 redirector (whichever was applied last).
4895 Computed URLs are URIs of internally generated
4896 requests and various "error:..." URIs.
4898 Honors strip_query_terms and uri_whitespace.
4900 This field is not encoded by default. Encoding
4901 this field using variants of %-encoding will
4902 clash with uri_whitespace modifications that
4903 also use %-encoding.
4905 [http::]>ru Request URL received from the client (or computed)
4907 Computed URLs are URIs of internally generated
4908 requests and various "error:..." URIs.
4910 Unlike %ru, this request URI is not affected
4911 by request adaptation, URL rewriting services,
4912 and strip_query_terms.
4914 Honors uri_whitespace.
4916 This field is using pass-through URL encoding
4917 by default. Encoding this field using other
4918 variants of %-encoding will clash with
4919 uri_whitespace modifications that also use
4922 [http::]<ru Request URL sent to server or peer
4923 [http::]>rs Request URL scheme from client
4924 [http::]<rs Request URL scheme sent to server or peer
4925 [http::]>rd Request URL domain from client
4926 [http::]<rd Request URL domain sent to server or peer
4927 [http::]>rP Request URL port from client
4928 [http::]<rP Request URL port sent to server or peer
4929 [http::]rp Request URL path excluding hostname
4930 [http::]>rp Request URL path excluding hostname from client
4931 [http::]<rp Request URL path excluding hostname sent to server or peer
4932 [http::]rv Request protocol version
4933 [http::]>rv Request protocol version from client
4934 [http::]<rv Request protocol version sent to server or peer
4936 [http::]>h Original received request header.
4937 Usually differs from the request header sent by
4938 Squid, although most fields are often preserved.
4939 Accepts optional header field name/value filter
4940 argument using name[:[separator]element] format.
4941 [http::]>ha Received request header after adaptation and
4942 redirection (pre-cache REQMOD vectoring point).
4943 Usually differs from the request header sent by
4944 Squid, although most fields are often preserved.
4945 Optional header name argument as for >h
4949 [http::]<Hs HTTP status code received from the next hop
4950 [http::]>Hs HTTP status code sent to the client
4952 [http::]<h Reply header. Optional header name argument
4955 [http::]mt MIME content type
4960 [http::]st Total size of request + reply traffic with client
4961 [http::]>st Total size of request received from client.
4962 Excluding chunked encoding bytes.
4963 [http::]<st Total size of reply sent to client (after adaptation)
4965 [http::]>sh Size of request headers received from client
4966 [http::]<sh Size of reply headers sent to client (after adaptation)
4968 [http::]<sH Reply high offset sent
4969 [http::]<sS Upstream object size
4971 [http::]<bs Number of HTTP-equivalent message body bytes
4972 received from the next hop, excluding chunked
4973 transfer encoding and control messages.
4974 Generated FTP listings are treated as
4979 [http::]<pt Peer response time in milliseconds. The timer starts
4980 when the last request byte is sent to the next hop
4981 and stops when the last response byte is received.
4982 [http::]<tt Total time in milliseconds. The timer
4983 starts with the first connect request (or write I/O)
4984 sent to the first selected peer. The timer stops
4985 with the last I/O with the last peer.
4987 Squid handling related format codes:
4989 Ss Squid request status (TCP_MISS etc)
4990 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4992 [http::]request_attempts Number of request forwarding attempts
4994 See forward_max_tries documentation that details what Squid counts
4995 as a forwarding attempt. Pure cache hits log zero, but cache hits
4996 that triggered HTTP cache revalidation log the number of attempts
4997 made when sending an internal revalidation request. DNS, ICMP,
4998 ICP, HTCP, ESI, ICAP, eCAP, helper, and other secondary requests
4999 sent by Squid as a part of a master transaction do not increment
5000 the counter logged for the received request.
5002 SSL-related format codes:
5004 ssl::bump_mode SslBump decision for the transaction:
5006 For CONNECT requests that initiated bumping of
5007 a connection and for any request received on
5008 an already bumped connection, Squid logs the
5009 corresponding SslBump mode ("splice", "bump",
5010 "peek", "stare", "terminate", "server-first"
5011 or "client-first"). See the ssl_bump option
5012 for more information about these modes.
5014 A "none" token is logged for requests that
5015 triggered "ssl_bump" ACL evaluation matching
5018 In all other cases, a single dash ("-") is
5021 ssl::>sni SSL client SNI sent to Squid.
5024 The Subject field of the received client
5025 SSL certificate or a dash ('-') if Squid has
5026 received an invalid/malformed certificate or
5027 no certificate at all. Consider encoding the
5028 logged value because Subject often has spaces.
5031 The Issuer field of the received client
5032 SSL certificate or a dash ('-') if Squid has
5033 received an invalid/malformed certificate or
5034 no certificate at all. Consider encoding the
5035 logged value because Issuer often has spaces.
5038 The Subject field of the received server
5039 TLS certificate or a dash ('-') if this is
5040 not available. Consider encoding the logged
5041 value because Subject often has spaces.
5044 The Issuer field of the received server
5045 TLS certificate or a dash ('-') if this is
5046 not available. Consider encoding the logged
5047 value because Issuer often has spaces.
5050 The received server x509 certificate in PEM
5051 format, including BEGIN and END lines (or a
5052 dash ('-') if the certificate is unavailable).
5054 WARNING: Large certificates will exceed the
5055 current 8KB access.log record limit, resulting
5056 in truncated records. Such truncation usually
5057 happens in the middle of a record field. The
5058 limit applies to all access logging modules.
5060 The logged certificate may have failed
5061 validation and may not be trusted by Squid.
5062 This field does not include any intermediate
5063 certificates that may have been received from
5064 the server or fetched during certificate
5067 Currently, Squid only collects server
5068 certificates during step3 of SslBump
5069 processing; connections that were not subject
5070 to ssl_bump rules or that did not match a peek
5071 or stare rule at step2 will not have the
5072 server certificate information.
5074 This field is using pass-through URL encoding
5078 The list of certificate validation errors
5079 detected by Squid (including OpenSSL and
5080 certificate validation helper components). The
5081 errors are listed in the discovery order. By
5082 default, the error codes are separated by ':'.
5083 Accepts an optional separator argument.
5085 %ssl::>negotiated_version The negotiated TLS version of the
5088 %ssl::<negotiated_version The negotiated TLS version of the
5089 last server or peer connection.
5091 %ssl::>received_hello_version The TLS version of the Hello
5092 message received from TLS client.
5094 %ssl::<received_hello_version The TLS version of the Hello
5095 message received from TLS server.
5097 %ssl::>received_supported_version The maximum TLS version
5098 supported by the TLS client.
5100 %ssl::<received_supported_version The maximum TLS version
5101 supported by the TLS server.
5103 %ssl::>negotiated_cipher The negotiated cipher of the
5106 %ssl::<negotiated_cipher The negotiated cipher of the
5107 last server or peer connection.
5109 If ICAP is enabled, the following code becomes available (as
5110 well as ICAP log codes documented with the icap_log option):
5112 icap::tt Total ICAP "blocking" time for the HTTP transaction. The
5113 timer ticks while Squid checks adaptation_access and while
5114 ICAP transaction(s) expect ICAP response headers, including
5115 the embedded adapted HTTP message headers (where applicable).
5116 This measurement is meant to estimate ICAP impact on HTTP
5117 transaction response times, but it does not currently account
5118 for slow ICAP response body delivery blocking HTTP progress.
5120 Once Squid receives the final ICAP response headers (e.g.,
5121 ICAP 200 or 204) and the associated adapted HTTP message
5122 headers (if any) from the ICAP service, the corresponding ICAP
5123 transaction stops affecting this measurement, even though the
5124 transaction itself may continue for a long time (e.g., to
5125 finish sending the ICAP request and/or to finish receiving the
5126 ICAP response body).
5128 When "blocking" sections of multiple concurrent ICAP
5129 transactions overlap in time, the overlapping segment is
5132 To see complete ICAP transaction response times (rather than
5133 the cumulative effect of their blocking sections) use the
5134 %adapt::all_trs logformat code or the icap_log directive.
5136 If adaptation is enabled the following codes become available:
5138 adapt::<last_h The header of the last ICAP response or
5139 meta-information from the last eCAP
5140 transaction related to the HTTP transaction.
5141 Like <h, accepts an optional header name
5144 adapt::sum_trs Summed adaptation transaction response
5145 times recorded as a comma-separated list in
5146 the order of transaction start time. Each time
5147 value is recorded as an integer number,
5148 representing response time of one or more
5149 adaptation (ICAP or eCAP) transaction in
5150 milliseconds. When a failed transaction is
5151 being retried or repeated, its time is not
5152 logged individually but added to the
5153 replacement (next) transaction. Lifetimes of individually
5154 listed adaptation transactions may overlap.
5155 See also: %icap::tt and %adapt::all_trs.
5157 adapt::all_trs All adaptation transaction response times.
5158 Same as %adapt::sum_trs but response times of
5159 individual transactions are never added
5160 together. Instead, all transaction response
5161 times are recorded individually.
5163 You can prefix adapt::*_trs format codes with adaptation
5164 service name in curly braces to record response time(s) specific
5165 to that service. For example: %{my_service}adapt::sum_trs
5167 Format codes related to the PROXY protocol:
5169 proxy_protocol::>h PROXY protocol header, including optional TLVs.
5171 Supports the same field and element reporting/extraction logic
5172 as %http::>h. For configuration and reporting purposes, Squid
5173 maps each PROXY TLV to an HTTP header field: the TLV type
5174 (configured as a decimal integer) is the field name, and the
5175 TLV value is the field value. All TLVs of "LOCAL" connections
5176 (in PROXY protocol terminology) are currently skipped/ignored.
5178 Squid also maps the following standard PROXY protocol header
5179 blocks to pseudo HTTP headers (their names use PROXY
5180 terminology and start with a colon, following HTTP tradition
5181 for pseudo headers): :command, :version, :src_addr, :dst_addr,
5182 :src_port, and :dst_port.
5184 Without optional parameters, this logformat code logs
5185 pseudo headers and TLVs.
5187 This format code uses pass-through URL encoding by default.
5190 # relay custom PROXY TLV #224 to adaptation services
5191 adaptation_meta Client-Foo "%proxy_protocol::>h{224}"
5195 The default formats available (which do not need re-defining) are:
5197 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
5198 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
5199 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
5200 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
5201 logformat useragent %>a [%tl] "%{User-Agent}>h"
5203 NOTE: When the log_mime_hdrs directive is set to ON.
5204 The squid, common and combined formats have a safely encoded copy
5205 of the mime headers appended to each line within a pair of brackets.
5207 NOTE: The common and combined formats are not quite true to the Apache definition.
5208 The logs from Squid contain an extra status and hierarchy code appended.
5212 NAME: access_log cache_access_log
5214 LOC: Config.Log.accesslogs
5215 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
5217 Configures whether and how Squid logs HTTP and ICP transactions.
5218 If access logging is enabled, a single line is logged for every
5219 matching HTTP or ICP request. The recommended directive formats are:
5221 access_log <module>:<place> [option ...] [acl acl ...]
5222 access_log none [acl acl ...]
5224 The following directive format is accepted but may be deprecated:
5225 access_log <module>:<place> [<logformat name> [acl acl ...]]
5227 In most cases, the first ACL name must not contain the '=' character
5228 and should not be equal to an existing logformat name. You can always
5229 start with an 'all' ACL to work around those restrictions.
5231 Will log to the specified module:place using the specified format (which
5232 must be defined in a logformat directive) those entries which match
5233 ALL the acl's specified (which must be defined in acl clauses).
5234 If no acl is specified, all requests will be logged to this destination.
5236 ===== Available options for the recommended directive format =====
5238 logformat=name Names log line format (either built-in or
5239 defined by a logformat directive). Defaults
5242 buffer-size=64KB Defines approximate buffering limit for log
5243 records (see buffered_logs). Squid should not
5244 keep more than the specified size and, hence,
5245 should flush records before the buffer becomes
5246 full to avoid overflows under normal
5247 conditions (the exact flushing algorithm is
5248 module-dependent though). The on-error option
5249 controls overflow handling.
5251 on-error=die|drop Defines action on unrecoverable errors. The
5252 'drop' action ignores (i.e., does not log)
5253 affected log records. The default 'die' action
5254 kills the affected worker. The drop action
5255 support has not been tested for modules other
5258 rotate=N Specifies the number of log file rotations to
5259 make when you run 'squid -k rotate'. The default
5260 is to obey the logfile_rotate directive. Setting
5261 rotate=0 will disable the file name rotation,
5262 but the log files are still closed and re-opened.
5263 This will enable you to rename the logfiles
5264 yourself just before sending the rotate signal.
5265 Only supported by the stdio module.
5267 ===== Modules Currently available =====
5269 none Do not log any requests matching these ACL.
5270 Do not specify Place or logformat name.
5272 stdio Write each log line to disk immediately at the completion of
5274 Place: the filename and path to be written.
5276 daemon Very similar to stdio. But instead of writing to disk the log
5277 line is passed to a daemon helper for asychronous handling instead.
5278 Place: varies depending on the daemon.
5280 log_file_daemon Place: the file name and path to be written.
5282 syslog To log each request via syslog facility.
5283 Place: The syslog facility and priority level for these entries.
5284 Place Format: facility.priority
5286 where facility could be any of:
5287 authpriv, daemon, local0 ... local7 or user.
5289 And priority could be any of:
5290 err, warning, notice, info, debug.
5292 udp To send each log line as text data to a UDP receiver.
5293 Place: The destination host name or IP and port.
5294 Place Format: //host:port
5296 tcp To send each log line as text data to a TCP receiver.
5297 Lines may be accumulated before sending (see buffered_logs).
5298 Place: The destination host name or IP and port.
5299 Place Format: //host:port
5302 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
5308 LOC: Config.Log.icaplogs
5311 ICAP log files record ICAP transaction summaries, one line per
5314 The icap_log option format is:
5315 icap_log <filepath> [<logformat name> [acl acl ...]]
5316 icap_log none [acl acl ...]]
5318 Please see access_log option documentation for details. The two
5319 kinds of logs share the overall configuration approach and many
5322 ICAP processing of a single HTTP message or transaction may
5323 require multiple ICAP transactions. In such cases, multiple
5324 ICAP transaction log lines will correspond to a single access
5327 ICAP log supports many access.log logformat %codes. In ICAP context,
5328 HTTP message-related %codes are applied to the HTTP message embedded
5329 in an ICAP message. Logformat "%http::>..." codes are used for HTTP
5330 messages embedded in ICAP requests while "%http::<..." codes are used
5331 for HTTP messages embedded in ICAP responses. For example:
5333 http::>h To-be-adapted HTTP message headers sent by Squid to
5334 the ICAP service. For REQMOD transactions, these are
5335 HTTP request headers. For RESPMOD, these are HTTP
5336 response headers, but Squid currently cannot log them
5337 (i.e., %http::>h will expand to "-" for RESPMOD).
5339 http::<h Adapted HTTP message headers sent by the ICAP
5340 service to Squid (i.e., HTTP request headers in regular
5341 REQMOD; HTTP response headers in RESPMOD and during
5342 request satisfaction in REQMOD).
5344 ICAP OPTIONS transactions do not embed HTTP messages.
5346 Several logformat codes below deal with ICAP message bodies. An ICAP
5347 message body, if any, typically includes a complete HTTP message
5348 (required HTTP headers plus optional HTTP message body). When
5349 computing HTTP message body size for these logformat codes, Squid
5350 either includes or excludes chunked encoding overheads; see
5351 code-specific documentation for details.
5353 For Secure ICAP services, all size-related information is currently
5354 computed before/after TLS encryption/decryption, as if TLS was not
5357 The following format codes are also available for ICAP logs:
5359 icap::<A ICAP server IP address. Similar to <A.
5361 icap::<service_name ICAP service name from the icap_service
5362 option in Squid configuration file.
5364 icap::ru ICAP Request-URI. Similar to ru.
5366 icap::rm ICAP request method (REQMOD, RESPMOD, or
5367 OPTIONS). Similar to existing rm.
5369 icap::>st The total size of the ICAP request sent to the ICAP
5370 server (ICAP headers + ICAP body), including chunking
5373 icap::<st The total size of the ICAP response received from the
5374 ICAP server (ICAP headers + ICAP body), including
5375 chunking metadata (if any).
5377 icap::<bs The size of the ICAP response body received from the
5378 ICAP server, excluding chunking metadata (if any).
5380 icap::tr Transaction response time (in
5381 milliseconds). The timer starts when
5382 the ICAP transaction is created and
5383 stops when the transaction is completed.
5386 icap::tio Transaction I/O time (in milliseconds). The
5387 timer starts when the first ICAP request
5388 byte is scheduled for sending. The timers
5389 stops when the last byte of the ICAP response
5392 icap::to Transaction outcome: ICAP_ERR* for all
5393 transaction errors, ICAP_OPT for OPTION
5394 transactions, ICAP_ECHO for 204
5395 responses, ICAP_MOD for message
5396 modification, and ICAP_SAT for request
5397 satisfaction. Similar to Ss.
5399 icap::Hs ICAP response status code. Similar to Hs.
5401 icap::>h ICAP request header(s). Similar to >h.
5403 icap::<h ICAP response header(s). Similar to <h.
5405 The default ICAP log format, which can be used without an explicit
5406 definition, is called icap_squid:
5408 logformat icap_squid %ts.%03tu %6icap::tr %>A %icap::to/%03icap::Hs %icap::<st %icap::rm %icap::ru %un -/%icap::<A -
5410 See also: logformat and %adapt::<last_h
5413 NAME: logfile_daemon
5415 DEFAULT: @DEFAULT_LOGFILED@
5416 LOC: Log::TheConfig.logfile_daemon
5418 Specify the path to the logfile-writing daemon. This daemon is
5419 used to write the access and store logs, if configured.
5421 Squid sends a number of commands to the log daemon:
5422 L<data>\n - logfile data
5427 r<n>\n - set rotate count to <n>
5428 b<n>\n - 1 = buffer output, 0 = don't buffer output
5430 No responses is expected.
5433 NAME: stats_collection
5435 LOC: Config.accessList.stats_collection
5437 DEFAULT_DOC: Allow logging for all transactions.
5438 COMMENT: allow|deny acl acl...
5440 This options allows you to control which requests gets accounted
5441 in performance counters.
5443 This clause only supports fast acl types.
5444 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5447 NAME: cache_store_log
5450 LOC: Config.Log.store
5452 Logs the activities of the storage manager. Shows which
5453 objects are ejected from the cache, and which objects are
5454 saved and for how long.
5455 There are not really utilities to analyze this data, so you can safely
5456 disable it (the default).
5458 Store log uses modular logging outputs. See access_log for the list
5459 of modules supported.
5462 cache_store_log stdio:@DEFAULT_STORE_LOG@
5463 cache_store_log daemon:@DEFAULT_STORE_LOG@
5466 NAME: cache_swap_state cache_swap_log
5468 LOC: Config.Log.swap
5470 DEFAULT_DOC: Store the journal inside its cache_dir
5472 Location for the cache "swap.state" file. This index file holds
5473 the metadata of objects saved on disk. It is used to rebuild
5474 the cache during startup. Normally this file resides in each
5475 'cache_dir' directory, but you may specify an alternate
5476 pathname here. Note you must give a full filename, not just
5477 a directory. Since this is the index for the whole object
5478 list you CANNOT periodically rotate it!
5480 If %s can be used in the file name it will be replaced with a
5481 a representation of the cache_dir name where each / is replaced
5482 with '.'. This is needed to allow adding/removing cache_dir
5483 lines when cache_swap_log is being used.
5485 If have more than one 'cache_dir', and %s is not used in the name
5486 these swap logs will have names such as:
5492 The numbered extension (which is added automatically)
5493 corresponds to the order of the 'cache_dir' lines in this
5494 configuration file. If you change the order of the 'cache_dir'
5495 lines in this file, these index files will NOT correspond to
5496 the correct 'cache_dir' entry (unless you manually rename
5497 them). We recommend you do NOT use this option. It is
5498 better to keep these index files in each 'cache_dir' directory.
5501 NAME: logfile_rotate
5504 LOC: Config.Log.rotateNumber
5506 Specifies the default number of logfile rotations to make when you
5507 type 'squid -k rotate'. The default is 10, which will rotate
5508 with extensions 0 through 9. Setting logfile_rotate to 0 will
5509 disable the file name rotation, but the logfiles are still closed
5510 and re-opened. This will enable you to rename the logfiles
5511 yourself just before sending the rotate signal.
5513 Note, from Squid-3.1 this option is only a default for cache.log,
5514 that log can be rotated separately by using debug_options.
5516 Note, from Squid-4 this option is only a default for access.log
5517 recorded by stdio: module. Those logs can be rotated separately by
5518 using the rotate=N option on their access_log directive.
5520 Note, the 'squid -k rotate' command normally sends a USR1
5521 signal to the running squid process. In certain situations
5522 (e.g. on Linux with Async I/O), USR1 is used for other
5523 purposes, so -k rotate uses another signal. It is best to get
5524 in the habit of using 'squid -k rotate' instead of 'kill -USR1
5531 DEFAULT: @DEFAULT_MIME_TABLE@
5532 LOC: Config.mimeTablePathname
5534 Path to Squid's icon configuration file.
5536 You shouldn't need to change this, but the default file contains
5537 examples and formatting information if you do.
5543 LOC: Config.onoff.log_mime_hdrs
5546 The Cache can record both the request and the response MIME
5547 headers for each HTTP transaction. The headers are encoded
5548 safely and will appear as two bracketed fields at the end of
5549 the access log (for either the native or httpd-emulated log
5550 formats). To enable this logging set log_mime_hdrs to 'on'.
5555 DEFAULT: @DEFAULT_PID_FILE@
5556 LOC: Config.pidFilename
5558 A filename to write the process-id to. To disable, enter "none".
5561 NAME: client_netmask
5563 LOC: Config.Addrs.client_netmask
5565 DEFAULT_DOC: Log full client IP address
5567 A netmask for client addresses in logfiles and cachemgr output.
5568 Change this to protect the privacy of your cache clients.
5569 A netmask of 255.255.255.0 will log all IP's in that range with
5570 the last digit set to '0'.
5573 NAME: strip_query_terms
5575 LOC: Config.onoff.strip_query_terms
5578 By default, Squid strips query terms from requested URLs before
5579 logging. This protects your user's privacy and reduces log size.
5581 When investigating HIT/MISS or other caching behaviour you
5582 will need to disable this to see the full URL used by Squid.
5589 LOC: Config.onoff.buffered_logs
5591 Whether to write/send access_log records ASAP or accumulate them and
5592 then write/send them in larger chunks. Buffering may improve
5593 performance because it decreases the number of I/Os. However,
5594 buffering increases the delay before log records become available to
5595 the final recipient (e.g., a disk file or logging daemon) and,
5596 hence, increases the risk of log records loss.
5598 Note that even when buffered_logs are off, Squid may have to buffer
5599 records if it cannot write/send them immediately due to pending I/Os
5600 (e.g., the I/O writing the previous log record) or connectivity loss.
5602 Currently honored by 'daemon', 'tcp' and 'udp' access_log modules only.
5605 NAME: netdb_filename
5607 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
5608 LOC: Config.netdbFilename
5611 Where Squid stores it's netdb journal.
5612 When enabled this journal preserves netdb state between restarts.
5614 To disable, enter "none".
5618 TYPE: Security::KeyLog*
5620 LOC: Config.Log.tlsKeys
5623 Configures whether and where Squid records pre-master secret and
5624 related encryption details for TLS connections accepted or established
5625 by Squid. These connections include connections accepted at
5626 https_port, TLS connections opened to origin servers/cache_peers/ICAP
5627 services, and TLS tunnels bumped by Squid using the SslBump feature.
5628 This log (a.k.a. SSLKEYLOGFILE) is meant for triage with traffic
5629 inspection tools like Wireshark.
5631 tls_key_log <destination> [options] [if [!]<acl>...]
5633 WARNING: This log allows anybody to decrypt the corresponding
5634 encrypted TLS connections, both in-flight and postmortem.
5636 At most one log file is supported at this time. Repeated tls_key_log
5637 directives are treated as fatal configuration errors. By default, no
5638 log is created or updated.
5640 If the log file does not exist, Squid creates it. Otherwise, Squid
5641 appends an existing log file.
5643 The directive is consulted whenever a TLS connection is accepted or
5644 established by Squid. TLS connections that fail the handshake may be
5645 logged if Squid got enough information to form a log record. A record
5646 is logged only if all of the configured ACLs match.
5648 While transport-related ACLs like src and dst should work, Squid may
5649 not have access to higher-level information. For example, when logging
5650 accepted https_port connections, Squid does not yet have access to the
5651 expected HTTPS request. Similarly, an HTTPS response is not available
5652 when logging most TLS connections established by Squid.
5654 The log record format is meant to be compatible with TLS deciphering
5655 features of Wireshark which relies on fields like CLIENT_RANDOM and
5656 RSA Master-Key. A single log record usually spans multiple lines.
5657 Technical documentation for that format is maintained inside the
5658 Wireshark code (e.g., see tls_keylog_process_lines() comments as of
5659 Wireshark commit e3d44136f0f0026c5e893fa249f458073f3b7328). TLS key
5660 log does not support custom record formats.
5662 This clause only supports fast acl types.
5663 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5665 See access_log's <module>:<place> parameter for a list of supported
5666 logging destinations.
5668 TLS key log supports all access_log key=value options with the
5669 exception of logformat=name.
5671 Requires Squid built with OpenSSL support.
5676 OPTIONS FOR TROUBLESHOOTING
5677 -----------------------------------------------------------------------------
5682 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
5683 LOC: Debug::cache_log
5685 Squid administrative logging file.
5687 This is where general information about Squid behavior goes. You can
5688 increase the amount of data logged to this file and how often it is
5689 rotated with "debug_options"
5692 NAME: cache_log_message
5693 TYPE: cache_log_message
5695 DEFAULT_DOC: Use debug_options.
5696 LOC: DebugMessagesConfig
5698 Configures logging of individual cache.log messages.
5700 cache_log_message id=<number> option...
5701 cache_log_message ids=<number>-<number> option...
5703 Most messages have _not_ been instrumented to support this directive
5704 yet. For the list of instrumented messages and their IDs, please see
5705 the doc/debug-messages.txt file.
5707 Message ID corresponds to the message semantics rather than message
5708 text or source code location. The ID is stable across Squid
5709 instances and versions. Substantial changes in message semantics
5710 result in a new ID assignment. To reduce the danger of suppressing
5711 an important log message, the old IDs of removed (or substantially
5712 changed) messages are never reused.
5714 If more than one cache_log_message directive refers to the same
5715 message ID, the last directive wins.
5717 Use ids=min-max syntax to apply the same message configuration to an
5718 inclusive range of message IDs. An ID range with N values has
5719 exactly the same effect as typing N cache_log_message lines.
5721 At least one option is required. Supported options are:
5723 level=<number>: The logging level to use for the message. Squid
5724 command line options (-s and -d) as well as the debug_options
5725 directive control which levels go to syslog, stderr, and/or
5726 cache.log. In most environments, using level=2 or higher stops
5727 Squid from logging the message anywhere. By default, the
5728 hard-coded message-specific level is used.
5730 limit=<number>: After logging the specified number of messages at
5731 the configured (or default) debugging level DL, start using
5732 level 3 (for DL 0 and 1) or 8 (for higher DL values). Usually,
5733 level-3+ messages are not logged anywhere so this option can
5734 often be used to effectively suppress the message. Each SMP
5735 Squid process gets the same limit.
5741 DEFAULT_DOC: Log all critical and important messages.
5742 LOC: Debug::debugOptions
5744 Logging options are set as section,level where each source file
5745 is assigned a unique section. Lower levels result in less
5746 output, Full debugging (level 9) can result in a very large
5747 log file, so be careful.
5749 The magic word "ALL" sets debugging levels for all sections.
5750 The default is to run with "ALL,1" to record important warnings.
5752 The rotate=N option can be used to keep more or less of these logs
5753 than would otherwise be kept by logfile_rotate.
5754 For most uses a single log should be enough to monitor current
5755 events affecting Squid.
5760 LOC: Config.coredump_dir
5761 DEFAULT_IF_NONE: none
5762 DEFAULT_DOC: Use the directory from where Squid was started.
5764 By default Squid leaves core files in the directory from where
5765 it was started. If you set 'coredump_dir' to a directory
5766 that exists, Squid will chdir() to that directory at startup
5767 and coredump files will be left there.
5769 In addition to changing the directory, the process permissions are updated
5770 to enable process tracing and/or coredump file generation. The details are
5771 OS-specific, but look for prctl(2) PR_SET_DUMPABLE and procctl(2)
5772 PROC_TRACE_CTL documentation as guiding examples.
5776 # Leave coredumps in the first cache dir
5777 coredump_dir @DEFAULT_SWAP_DIR@
5783 OPTIONS FOR FTP GATEWAYING
5784 -----------------------------------------------------------------------------
5790 LOC: Config.Ftp.anon_user
5792 If you want the anonymous login password to be more informative
5793 (and enable the use of picky FTP servers), set this to something
5794 reasonable for your domain, like wwwuser@somewhere.net
5796 The reason why this is domainless by default is the
5797 request can be made on the behalf of a user in any domain,
5798 depending on how the cache is used.
5799 Some FTP server also validate the email address is valid
5800 (for example perl.com).
5806 LOC: Config.Ftp.passive
5808 If your firewall does not allow Squid to use passive
5809 connections, turn off this option.
5811 Use of ftp_epsv_all option requires this to be ON.
5817 LOC: Config.Ftp.epsv_all
5819 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
5821 NATs may be able to put the connection on a "fast path" through the
5822 translator, as the EPRT command will never be used and therefore,
5823 translation of the data portion of the segments will never be needed.
5825 When a client only expects to do two-way FTP transfers this may be
5827 If squid finds that it must do a three-way FTP transfer after issuing
5828 an EPSV ALL command, the FTP session will fail.
5830 If you have any doubts about this option do not use it.
5831 Squid will nicely attempt all other connection methods.
5833 Requires ftp_passive to be ON (default) for any effect.
5839 LOC: Config.accessList.ftp_epsv
5841 FTP Protocol extensions permit the use of a special "EPSV" command.
5843 NATs may be able to put the connection on a "fast path" through the
5844 translator using EPSV, as the EPRT command will never be used
5845 and therefore, translation of the data portion of the segments
5846 will never be needed.
5848 EPSV is often required to interoperate with FTP servers on IPv6
5849 networks. On the other hand, it may break some IPv4 servers.
5851 By default, EPSV may try EPSV with any FTP server. To fine tune
5852 that decision, you may restrict EPSV to certain clients or servers
5855 ftp_epsv allow|deny al1 acl2 ...
5857 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
5859 Only fast ACLs are supported.
5860 Requires ftp_passive to be ON (default) for any effect.
5866 LOC: Config.Ftp.eprt
5868 FTP Protocol extensions permit the use of a special "EPRT" command.
5870 This extension provides a protocol neutral alternative to the
5871 IPv4-only PORT command. When supported it enables active FTP data
5872 channels over IPv6 and efficient NAT handling.
5874 Turning this OFF will prevent EPRT being attempted and will skip
5875 straight to using PORT for IPv4 servers.
5877 Some devices are known to not handle this extension correctly and
5878 may result in crashes. Devices which support EPRT enough to fail
5879 cleanly will result in Squid attempting PORT anyway. This directive
5880 should only be disabled when EPRT results in device failures.
5882 WARNING: Doing so will convert Squid back to the old behavior with all
5883 the related problems with external NAT devices/layers and IPv4-only FTP.
5886 NAME: ftp_sanitycheck
5889 LOC: Config.Ftp.sanitycheck
5891 For security and data integrity reasons Squid by default performs
5892 sanity checks of the addresses of FTP data connections ensure the
5893 data connection is to the requested server. If you need to allow
5894 FTP connections to servers using another IP address for the data
5895 connection turn this off.
5898 NAME: ftp_telnet_protocol
5901 LOC: Config.Ftp.telnet
5903 The FTP protocol is officially defined to use the telnet protocol
5904 as transport channel for the control connection. However, many
5905 implementations are broken and does not respect this aspect of
5908 If you have trouble accessing files with ASCII code 255 in the
5909 path or similar problems involving this ASCII code you can
5910 try setting this directive to off. If that helps, report to the
5911 operator of the FTP server in question that their FTP server
5912 is broken and does not follow the FTP standard.
5916 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
5917 -----------------------------------------------------------------------------
5922 DEFAULT: @DEFAULT_DISKD@
5923 LOC: Config.Program.diskd
5925 Specify the location of the diskd executable.
5926 Note this is only useful if you have compiled in
5927 diskd as one of the store io modules.
5930 NAME: unlinkd_program
5933 DEFAULT: @DEFAULT_UNLINKD@
5934 LOC: Config.Program.unlinkd
5936 Specify the location of the executable for file deletion process.
5939 NAME: pinger_program
5942 DEFAULT: @DEFAULT_PINGER@
5945 Specify the location of the executable for the pinger process.
5954 Control whether the pinger is active at run-time.
5955 Enables turning ICMP pinger on and off with a simple
5956 squid -k reconfigure.
5961 OPTIONS FOR URL REWRITING
5962 -----------------------------------------------------------------------------
5965 NAME: url_rewrite_program redirect_program
5967 LOC: Config.Program.redirect
5970 The name and command line parameters of an admin-provided executable
5971 for redirecting clients or adjusting/replacing client request URLs.
5973 This helper is consulted after the received request is cleared by
5974 http_access and adapted using eICAP/ICAP services (if any). If the
5975 helper does not redirect the client, Squid checks adapted_http_access
5976 and may consult the cache or forward the request to the next hop.
5979 For each request, the helper gets one line in the following format:
5981 [channel-ID <SP>] request-URL [<SP> extras] <NL>
5983 Use url_rewrite_extras to configure what Squid sends as 'extras'.
5986 The helper must reply to each query using a single line:
5988 [channel-ID <SP>] result [<SP> kv-pairs] <NL>
5990 The result section must match exactly one of the following outcomes:
5992 OK [status=30N] url="..."
5994 Redirect the client to a URL supplied in the 'url' parameter.
5995 Optional 'status' specifies the status code to send to the
5996 client in Squid's HTTP redirect response. It must be one of
5997 the standard HTTP redirect status codes: 301, 302, 303, 307,
5998 or 308. When no specific status is requested, Squid uses 302.
6000 OK rewrite-url="..."
6002 Replace the current request URL with the one supplied in the
6003 'rewrite-url' parameter. Squid fetches the resource specified
6004 by the new URL and forwards the received response (or its
6005 cached copy) to the client.
6007 WARNING: Avoid rewriting URLs! When possible, redirect the
6008 client using an "OK url=..." helper response instead.
6009 Rewriting URLs may create inconsistent requests and/or break
6010 synchronization between internal client and origin server
6011 states, especially when URLs or other message parts contain
6012 snippets of that state. For example, Squid does not adjust
6013 Location headers and embedded URLs after the helper rewrites
6017 Keep the client request intact.
6020 Keep the client request intact.
6023 A helper problem that should be reported to the Squid admin
6024 via a level-1 cache.log message. The 'message' parameter is
6025 reserved for specifying the log message.
6027 In addition to the kv-pairs mentioned above, Squid also understands
6028 the following optional kv-pairs in URL rewriter responses:
6031 Associates a TAG with the client TCP connection.
6033 The clt_conn_tag=TAG pair is treated as a regular transaction
6034 annotation for the current request and also annotates future
6035 requests on the same client connection. A helper may update
6036 the TAG during subsequent requests by returning a new kv-pair.
6039 Helper messages contain the channel-ID part if and only if the
6040 url_rewrite_children directive specifies positive concurrency. As a
6041 channel-ID value, Squid sends a number between 0 and concurrency-1.
6042 The helper must echo back the received channel-ID in its response.
6044 By default, Squid does not use a URL rewriter.
6047 NAME: url_rewrite_children redirect_children
6048 TYPE: HelperChildConfig
6049 DEFAULT: 20 startup=0 idle=1 concurrency=0
6050 LOC: Config.redirectChildren
6052 Specifies the maximum number of redirector processes that Squid may
6053 spawn (numberofchildren) and several related options. Using too few of
6054 these helper processes (a.k.a. "helpers") creates request queues.
6055 Using too many helpers wastes your system resources.
6057 Usage: numberofchildren [option]...
6059 The startup= and idle= options allow some measure of skew in your
6064 Sets a minimum of how many processes are to be spawned when Squid
6065 starts or reconfigures. When set to zero the first request will
6066 cause spawning of the first child process to handle it.
6068 Starting too few will cause an initial slowdown in traffic as Squid
6069 attempts to simultaneously spawn enough processes to cope.
6073 Sets a minimum of how many processes Squid is to try and keep available
6074 at all times. When traffic begins to rise above what the existing
6075 processes can handle this many more will be spawned up to the maximum
6076 configured. A minimum setting of 1 is required.
6080 The number of requests each redirector helper can handle in
6081 parallel. Defaults to 0 which indicates the redirector
6082 is a old-style single threaded redirector.
6084 When this directive is set to a value >= 1 then the protocol
6085 used to communicate with the helper is modified to include
6086 an ID in front of the request/response. The ID from the request
6087 must be echoed back with the response to that request.
6091 Sets the maximum number of queued requests. A request is queued when
6092 no existing child can accept it due to concurrency limit and no new
6093 child can be started due to numberofchildren limit. The default
6094 maximum is zero if url_rewrite_bypass is enabled and
6095 2*numberofchildren otherwise. If the queued requests exceed queue size
6096 and redirector_bypass configuration option is set, then redirector is
6097 bypassed. Otherwise, Squid is allowed to temporarily exceed the
6098 configured maximum, marking the affected helper as "overloaded". If
6099 the helper overload lasts more than 3 minutes, the action prescribed
6100 by the on-persistent-overload option applies.
6102 on-persistent-overload=action
6104 Specifies Squid reaction to a new helper request arriving when the helper
6105 has been overloaded for more that 3 minutes already. The number of queued
6106 requests determines whether the helper is overloaded (see the queue-size
6109 Two actions are supported:
6111 die Squid worker quits. This is the default behavior.
6113 ERR Squid treats the helper request as if it was
6114 immediately submitted, and the helper immediately
6115 replied with an ERR response. This action has no effect
6116 on the already queued and in-progress helper requests.
6119 NAME: url_rewrite_host_header redirect_rewrites_host_header
6122 LOC: Config.onoff.redir_rewrites_host
6124 To preserve same-origin security policies in browsers and
6125 prevent Host: header forgery by redirectors Squid rewrites
6126 any Host: header in redirected requests.
6128 If you are running an accelerator this may not be a wanted
6129 effect of a redirector. This directive enables you disable
6130 Host: alteration in reverse-proxy traffic.
6132 WARNING: Entries are cached on the result of the URL rewriting
6133 process, so be careful if you have domain-virtual hosts.
6135 WARNING: Squid and other software verifies the URL and Host
6136 are matching, so be careful not to relay through other proxies
6137 or inspecting firewalls with this disabled.
6140 NAME: url_rewrite_access redirector_access
6143 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
6144 LOC: Config.accessList.redirector
6146 If defined, this access list specifies which requests are
6147 sent to the redirector processes.
6149 This clause supports both fast and slow acl types.
6150 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6153 NAME: url_rewrite_bypass redirector_bypass
6155 LOC: Config.onoff.redirector_bypass
6158 When this is 'on', a request will not go through the
6159 redirector if all the helpers are busy. If this is 'off' and the
6160 redirector queue grows too large, the action is prescribed by the
6161 on-persistent-overload option. You should only enable this if the
6162 redirectors are not critical to your caching system. If you use
6163 redirectors for access control, and you enable this option,
6164 users may have access to pages they should not
6165 be allowed to request.
6167 Enabling this option sets the default url_rewrite_children queue-size
6171 NAME: url_rewrite_extras
6172 TYPE: TokenOrQuotedString
6173 LOC: Config.redirector_extras
6174 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
6176 Specifies a string to be append to request line format for the
6177 rewriter helper. "Quoted" format values may contain spaces and
6178 logformat %macros. In theory, any logformat %macro can be used.
6179 In practice, a %macro expands as a dash (-) if the helper request is
6180 sent before the required macro information is available to Squid.
6183 NAME: url_rewrite_timeout
6184 TYPE: UrlHelperTimeout
6185 LOC: Config.onUrlRewriteTimeout
6187 DEFAULT_DOC: Squid waits for the helper response forever
6189 Squid times active requests to redirector. The timeout value and Squid
6190 reaction to a timed out request are configurable using the following
6193 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
6195 supported timeout actions:
6196 fail Squid return a ERR_GATEWAY_FAILURE error page
6198 bypass Do not re-write the URL
6200 retry Send the lookup to the helper again
6202 use_configured_response
6203 Use the <quoted-response> as helper response
6207 OPTIONS FOR STORE ID
6208 -----------------------------------------------------------------------------
6211 NAME: store_id_program storeurl_rewrite_program
6213 LOC: Config.Program.store_id
6216 Specify the location of the executable StoreID helper to use.
6217 Since they can perform almost any function there isn't one included.
6219 For each requested URL, the helper will receive one line with the format
6221 [channel-ID <SP>] URL [<SP> extras]<NL>
6224 After processing the request the helper must reply using the following format:
6226 [channel-ID <SP>] result [<SP> kv-pairs]
6228 The result code can be:
6231 Use the StoreID supplied in 'store-id='.
6234 The default is to use HTTP request URL as the store ID.
6237 An internal error occurred in the helper, preventing
6238 a result being identified.
6240 In addition to the above kv-pairs Squid also understands the following
6241 optional kv-pairs received from URL rewriters:
6243 Associates a TAG with the client TCP connection.
6244 Please see url_rewrite_program related documentation for this
6247 Helper programs should be prepared to receive and possibly ignore
6248 additional whitespace-separated tokens on each input line.
6250 When using the concurrency= option the protocol is changed by
6251 introducing a query channel tag in front of the request/response.
6252 The query channel tag is a number between 0 and concurrency-1.
6253 This value must be echoed back unchanged to Squid as the first part
6254 of the response relating to its request.
6256 NOTE: when using StoreID refresh_pattern will apply to the StoreID
6257 returned from the helper and not the URL.
6259 WARNING: Wrong StoreID value returned by a careless helper may result
6260 in the wrong cached response returned to the user.
6262 By default, a StoreID helper is not used.
6265 NAME: store_id_extras
6266 TYPE: TokenOrQuotedString
6267 LOC: Config.storeId_extras
6268 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
6270 Specifies a string to be append to request line format for the
6271 StoreId helper. "Quoted" format values may contain spaces and
6272 logformat %macros. In theory, any logformat %macro can be used.
6273 In practice, a %macro expands as a dash (-) if the helper request is
6274 sent before the required macro information is available to Squid.
6277 NAME: store_id_children storeurl_rewrite_children
6278 TYPE: HelperChildConfig
6279 DEFAULT: 20 startup=0 idle=1 concurrency=0
6280 LOC: Config.storeIdChildren
6282 Specifies the maximum number of StoreID helper processes that Squid
6283 may spawn (numberofchildren) and several related options. Using
6284 too few of these helper processes (a.k.a. "helpers") creates request
6285 queues. Using too many helpers wastes your system resources.
6287 Usage: numberofchildren [option]...
6289 The startup= and idle= options allow some measure of skew in your
6294 Sets a minimum of how many processes are to be spawned when Squid
6295 starts or reconfigures. When set to zero the first request will
6296 cause spawning of the first child process to handle it.
6298 Starting too few will cause an initial slowdown in traffic as Squid
6299 attempts to simultaneously spawn enough processes to cope.
6303 Sets a minimum of how many processes Squid is to try and keep available
6304 at all times. When traffic begins to rise above what the existing
6305 processes can handle this many more will be spawned up to the maximum
6306 configured. A minimum setting of 1 is required.
6310 The number of requests each storeID helper can handle in
6311 parallel. Defaults to 0 which indicates the helper
6312 is a old-style single threaded program.
6314 When this directive is set to a value >= 1 then the protocol
6315 used to communicate with the helper is modified to include
6316 an ID in front of the request/response. The ID from the request
6317 must be echoed back with the response to that request.
6321 Sets the maximum number of queued requests to N. A request is queued
6322 when no existing child can accept it due to concurrency limit and no
6323 new child can be started due to numberofchildren limit. The default
6324 maximum is 2*numberofchildren. If the queued requests exceed queue
6325 size and redirector_bypass configuration option is set, then
6326 redirector is bypassed. Otherwise, Squid is allowed to temporarily
6327 exceed the configured maximum, marking the affected helper as
6328 "overloaded". If the helper overload lasts more than 3 minutes, the
6329 action prescribed by the on-persistent-overload option applies.
6331 on-persistent-overload=action
6333 Specifies Squid reaction to a new helper request arriving when the helper
6334 has been overloaded for more that 3 minutes already. The number of queued
6335 requests determines whether the helper is overloaded (see the queue-size
6338 Two actions are supported:
6340 die Squid worker quits. This is the default behavior.
6342 ERR Squid treats the helper request as if it was
6343 immediately submitted, and the helper immediately
6344 replied with an ERR response. This action has no effect
6345 on the already queued and in-progress helper requests.
6348 NAME: store_id_access storeurl_rewrite_access
6351 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
6352 LOC: Config.accessList.store_id
6354 If defined, this access list specifies which requests are
6355 sent to the StoreID processes. By default all requests
6358 This clause supports both fast and slow acl types.
6359 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6362 NAME: store_id_bypass storeurl_rewrite_bypass
6364 LOC: Config.onoff.store_id_bypass
6367 When this is 'on', a request will not go through the
6368 helper if all helpers are busy. If this is 'off' and the helper
6369 queue grows too large, the action is prescribed by the
6370 on-persistent-overload option. You should only enable this if the
6371 helpers are not critical to your caching system. If you use
6372 helpers for critical caching components, and you enable this
6373 option, users may not get objects from cache.
6374 This options sets default queue-size option of the store_id_children
6379 OPTIONS FOR TUNING THE CACHE
6380 -----------------------------------------------------------------------------
6383 NAME: cache no_cache
6386 DEFAULT_DOC: By default, this directive is unused and has no effect.
6387 LOC: Config.accessList.noCache
6389 Requests denied by this directive will not be served from the cache
6390 and their responses will not be stored in the cache. This directive
6391 has no effect on other transactions and on already cached responses.
6393 This clause supports both fast and slow acl types.
6394 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6396 This and the two other similar caching directives listed below are
6397 checked at different transaction processing stages, have different
6398 access to response information, affect different cache operations,
6399 and differ in slow ACLs support:
6401 * cache: Checked before Squid makes a hit/miss determination.
6402 No access to reply information!
6403 Denies both serving a hit and storing a miss.
6404 Supports both fast and slow ACLs.
6405 * send_hit: Checked after a hit was detected.
6406 Has access to reply (hit) information.
6407 Denies serving a hit only.
6408 Supports fast ACLs only.
6409 * store_miss: Checked before storing a cachable miss.
6410 Has access to reply (miss) information.
6411 Denies storing a miss only.
6412 Supports fast ACLs only.
6414 If you are not sure which of the three directives to use, apply the
6415 following decision logic:
6417 * If your ACL(s) are of slow type _and_ need response info, redesign.
6418 Squid does not support that particular combination at this time.
6420 * If your directive ACL(s) are of slow type, use "cache"; and/or
6421 * if your directive ACL(s) need no response info, use "cache".
6423 * If you do not want the response cached, use store_miss; and/or
6424 * if you do not want a hit on a cached response, use send_hit.
6430 DEFAULT_DOC: By default, this directive is unused and has no effect.
6431 LOC: Config.accessList.sendHit
6433 Responses denied by this directive will not be served from the cache
6434 (but may still be cached, see store_miss). This directive has no
6435 effect on the responses it allows and on the cached objects.
6437 Please see the "cache" directive for a summary of differences among
6438 store_miss, send_hit, and cache directives.
6440 Unlike the "cache" directive, send_hit only supports fast acl
6441 types. See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6445 # apply custom Store ID mapping to some URLs
6446 acl MapMe dstdomain .c.example.com
6447 store_id_program ...
6448 store_id_access allow MapMe
6450 # but prevent caching of special responses
6451 # such as 302 redirects that cause StoreID loops
6452 acl Ordinary http_status 200-299
6453 store_miss deny MapMe !Ordinary
6455 # and do not serve any previously stored special responses
6456 # from the cache (in case they were already cached before
6457 # the above store_miss rule was in effect).
6458 send_hit deny MapMe !Ordinary
6464 DEFAULT_DOC: By default, this directive is unused and has no effect.
6465 LOC: Config.accessList.storeMiss
6467 Responses denied by this directive will not be cached (but may still
6468 be served from the cache, see send_hit). This directive has no
6469 effect on the responses it allows and on the already cached responses.
6471 Please see the "cache" directive for a summary of differences among
6472 store_miss, send_hit, and cache directives. See the
6473 send_hit directive for a usage example.
6475 Unlike the "cache" directive, store_miss only supports fast acl
6476 types. See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6482 LOC: Config.maxStale
6485 This option puts an upper limit on how stale content Squid
6486 will serve from the cache if cache validation fails.
6487 Can be overridden by the refresh_pattern max-stale option.
6490 NAME: refresh_pattern
6491 TYPE: refreshpattern
6495 usage: refresh_pattern [-i] regex min percent max [options]
6497 By default, regular expressions are CASE-SENSITIVE. To make
6498 them case-insensitive, use the -i option.
6500 'Min' is the time (in minutes) an object without an explicit
6501 expiry time should be considered fresh. The recommended
6502 value is 0, any higher values may cause dynamic applications
6503 to be erroneously cached unless the application designer
6504 has taken the appropriate actions.
6506 'Percent' is used to compute the max-age value for responses
6507 with a Last-Modified header and no Cache-Control:max-age nor Expires.
6508 Cache-Control:max-age = ( Date - Last-Modified ) * percent
6510 'Max' is an upper limit on how long objects without an explicit
6511 expiry time will be considered fresh. The value is also used
6512 to form Cache-Control: max-age header for a request sent from
6513 Squid to origin/parent.
6515 options: override-expire
6525 override-expire enforces min age even if the server
6526 sent an explicit expiry time (e.g., with the
6527 Expires: header or Cache-Control: max-age). Doing this
6528 VIOLATES the HTTP standard. Enabling this feature
6529 could make you liable for problems which it causes.
6531 Note: override-expire does not enforce staleness - it only extends
6532 freshness / min. If the server returns a Expires time which
6533 is longer than your max time, Squid will still consider
6534 the object fresh for that period of time.
6536 override-lastmod enforces min age even on objects
6537 that were modified recently.
6539 reload-into-ims changes a client no-cache or ``reload''
6540 request for a cached entry into a conditional request using
6541 If-Modified-Since and/or If-None-Match headers, provided the
6542 cached entry has a Last-Modified and/or a strong ETag header.
6543 Doing this VIOLATES the HTTP standard. Enabling this feature
6544 could make you liable for problems which it causes.
6546 ignore-reload ignores a client no-cache or ``reload''
6547 header. Doing this VIOLATES the HTTP standard. Enabling
6548 this feature could make you liable for problems which
6551 ignore-no-store ignores any ``Cache-control: no-store''
6552 headers received from a server. Doing this VIOLATES
6553 the HTTP standard. Enabling this feature could make you
6554 liable for problems which it causes.
6556 ignore-private ignores any ``Cache-control: private''
6557 headers received from a server. Doing this VIOLATES
6558 the HTTP standard. Enabling this feature could make you
6559 liable for problems which it causes.
6561 refresh-ims causes squid to contact the origin server
6562 when a client issues an If-Modified-Since request. This
6563 ensures that the client will receive an updated version
6564 if one is available.
6566 store-stale stores responses even if they don't have explicit
6567 freshness or a validator (i.e., Last-Modified or an ETag)
6568 present, or if they're already stale. By default, Squid will
6569 not cache such responses because they usually can't be
6570 reused. Note that such responses will be stale by default.
6572 max-stale=NN provide a maximum staleness factor. Squid won't
6573 serve objects more stale than this even if it failed to
6574 validate the object. Default: use the max_stale global limit.
6576 Basically a cached object is:
6578 FRESH if expire > now, else STALE
6580 FRESH if lm-factor < percent, else STALE
6584 The refresh_pattern lines are checked in the order listed here.
6585 The first entry which matches is used. If none of the entries
6586 match the default will be used.
6588 Note, you must uncomment all the default lines if you want
6589 to change one. The default setting is only active if none is
6595 # Add any of your own refresh_pattern entries above these.
6597 refresh_pattern ^ftp: 1440 20% 10080
6598 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
6599 refresh_pattern . 0 20% 4320
6603 NAME: quick_abort_min
6607 LOC: Config.quickAbort.min
6610 NAME: quick_abort_max
6614 LOC: Config.quickAbort.max
6617 NAME: quick_abort_pct
6621 LOC: Config.quickAbort.pct
6623 The cache by default continues downloading aborted requests
6624 which are almost completed (less than 16 KB remaining). This
6625 may be undesirable on slow (e.g. SLIP) links and/or very busy
6626 caches. Impatient users may tie up file descriptors and
6627 bandwidth by repeatedly requesting and immediately aborting
6630 When the user aborts a request, Squid will check the
6631 quick_abort values to the amount of data transferred until
6634 If the transfer has less than 'quick_abort_min' KB remaining,
6635 it will finish the retrieval.
6637 If the transfer has more than 'quick_abort_max' KB remaining,
6638 it will abort the retrieval.
6640 If more than 'quick_abort_pct' of the transfer has completed,
6641 it will finish the retrieval.
6643 If you do not want any retrieval to continue after the client
6644 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
6647 If you want retrievals to always continue if they are being
6648 cached set 'quick_abort_min' to '-1 KB'.
6651 NAME: read_ahead_gap
6652 COMMENT: buffer-size
6654 LOC: Config.readAheadGap
6657 The amount of data the cache will buffer ahead of what has been
6658 sent to the client when retrieving an object from another server.
6662 IFDEF: USE_HTTP_VIOLATIONS
6665 LOC: Config.negativeTtl
6668 Set the Default Time-to-Live (TTL) for failed requests.
6669 Certain types of failures (such as "connection refused" and
6670 "404 Not Found") are able to be negatively-cached for a short time.
6671 Modern web servers should provide Expires: header, however if they
6672 do not this can provide a minimum TTL.
6673 The default is not to cache errors with unknown expiry details.
6675 Note that this is different from negative caching of DNS lookups.
6677 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6678 this feature could make you liable for problems which it
6682 NAME: positive_dns_ttl
6685 LOC: Config.positiveDnsTtl
6688 Upper limit on how long Squid will cache positive DNS responses.
6689 Default is 6 hours (360 minutes). This directive must be set
6690 larger than negative_dns_ttl.
6693 NAME: negative_dns_ttl
6696 LOC: Config.negativeDnsTtl
6699 Time-to-Live (TTL) for negative caching of failed DNS lookups.
6700 This also sets the lower cache limit on positive lookups.
6701 Minimum value is 1 second, and it is not recommendable to go
6702 much below 10 seconds.
6705 NAME: range_offset_limit
6706 COMMENT: size [acl acl...]
6708 LOC: Config.rangeOffsetLimit
6711 usage: (size) [units] [[!]aclname]
6713 Sets an upper limit on how far (number of bytes) into the file
6714 a Range request may be to cause Squid to prefetch the whole file.
6715 If beyond this limit, Squid forwards the Range request as it is and
6716 the result is NOT cached.
6718 This is to stop a far ahead range request (lets say start at 17MB)
6719 from making Squid fetch the whole object up to that point before
6720 sending anything to the client.
6722 Multiple range_offset_limit lines may be specified, and they will
6723 be searched from top to bottom on each request until a match is found.
6724 The first match found will be used. If no line matches a request, the
6725 default limit of 0 bytes will be used.
6727 'size' is the limit specified as a number of units.
6729 'units' specifies whether to use bytes, KB, MB, etc.
6730 If no units are specified bytes are assumed.
6732 A size of 0 causes Squid to never fetch more than the
6733 client requested. (default)
6735 A size of 'none' causes Squid to always fetch the object from the
6736 beginning so it may cache the result. (2.0 style)
6738 'aclname' is the name of a defined ACL.
6740 NP: Using 'none' as the byte value here will override any quick_abort settings
6741 that may otherwise apply to the range request. The range request will
6742 be fully fetched from start to finish regardless of the client
6743 actions. This affects bandwidth usage.
6746 NAME: minimum_expiry_time
6749 LOC: Config.minimum_expiry_time
6752 The minimum caching time according to (Expires - Date)
6753 headers Squid honors if the object can't be revalidated.
6754 The default is 60 seconds.
6756 In reverse proxy environments it might be desirable to honor
6757 shorter object lifetimes. It is most likely better to make
6758 your server return a meaningful Last-Modified header however.
6760 In ESI environments where page fragments often have short
6761 lifetimes, this will often be best set to 0.
6764 NAME: store_avg_object_size
6768 LOC: Config.Store.avgObjectSize
6770 Average object size, used to estimate number of objects your
6771 cache can hold. The default is 13 KB.
6773 This is used to pre-seed the cache index memory allocation to
6774 reduce expensive reallocate operations while handling clients
6775 traffic. Too-large values may result in memory allocation during
6776 peak traffic, too-small values will result in wasted memory.
6778 Check the cache manager 'info' report metrics for the real
6779 object sizes seen by your Squid before tuning this.
6782 NAME: store_objects_per_bucket
6785 LOC: Config.Store.objectsPerBucket
6787 Target number of objects per bucket in the store hash table.
6788 Lowering this value increases the total number of buckets and
6789 also the storage maintenance rate. The default is 20.
6794 -----------------------------------------------------------------------------
6797 NAME: request_header_max_size
6801 LOC: Config.maxRequestHeaderSize
6803 This directives limits the header size of a received HTTP request
6804 (including request-line). Increasing this limit beyond its 64 KB default
6805 exposes certain old Squid code to various denial-of-service attacks. This
6806 limit also applies to received FTP commands.
6808 This limit has no direct affect on Squid memory consumption.
6810 Squid does not check this limit when sending requests.
6813 NAME: reply_header_max_size
6817 LOC: Config.maxReplyHeaderSize
6819 This directives limits the header size of a received HTTP response
6820 (including status-line). Increasing this limit beyond its 64 KB default
6821 exposes certain old Squid code to various denial-of-service attacks. This
6822 limit also applies to FTP command responses.
6824 Squid also checks this limit when loading hit responses from disk cache.
6826 Squid does not check this limit when sending responses.
6829 NAME: request_body_max_size
6833 DEFAULT_DOC: No limit.
6834 LOC: Config.maxRequestBodySize
6836 This specifies the maximum size for an HTTP request body.
6837 In other words, the maximum size of a PUT/POST request.
6838 A user who attempts to send a request with a body larger
6839 than this limit receives an "Invalid Request" error message.
6840 If you set this parameter to a zero (the default), there will
6841 be no limit imposed.
6843 See also client_request_buffer_max_size for an alternative
6844 limitation on client uploads which can be configured.
6847 NAME: client_request_buffer_max_size
6851 LOC: Config.maxRequestBufferSize
6853 This specifies the maximum buffer size of a client request.
6854 It prevents squid eating too much memory when somebody uploads
6859 IFDEF: USE_HTTP_VIOLATIONS
6862 DEFAULT_DOC: Obey RFC 2616.
6863 LOC: Config.accessList.brokenPosts
6865 A list of ACL elements which, if matched, causes Squid to send
6866 an extra CRLF pair after the body of a PUT/POST request.
6868 Some HTTP servers has broken implementations of PUT/POST,
6869 and rely on an extra CRLF pair sent by some WWW clients.
6871 Quote from RFC2616 section 4.1 on this matter:
6873 Note: certain buggy HTTP/1.0 client implementations generate an
6874 extra CRLF's after a POST request. To restate what is explicitly
6875 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
6876 a request with an extra CRLF.
6878 This clause only supports fast acl types.
6879 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6882 acl buggy_server url_regex ^http://....
6883 broken_posts allow buggy_server
6886 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
6889 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
6891 LOC: Adaptation::Config::use_indirect_client
6893 Controls whether the indirect client IP address (instead of the direct
6894 client IP address) is passed to adaptation services.
6896 See also: follow_x_forwarded_for adaptation_send_client_ip
6900 IFDEF: USE_HTTP_VIOLATIONS
6904 LOC: Config.onoff.via
6906 If set (default), Squid will include a Via header in requests and
6907 replies as required by RFC2616.
6910 NAME: vary_ignore_expire
6913 LOC: Config.onoff.vary_ignore_expire
6916 Many HTTP servers supporting Vary gives such objects
6917 immediate expiry time with no cache-control header
6918 when requested by a HTTP/1.0 client. This option
6919 enables Squid to ignore such expiry times until
6920 HTTP/1.1 is fully implemented.
6922 WARNING: If turned on this may eventually cause some
6923 varying objects not intended for caching to get cached.
6926 NAME: request_header_access
6927 IFDEF: USE_HTTP_VIOLATIONS
6928 TYPE: http_header_access
6929 LOC: Config.request_header_access
6931 DEFAULT_DOC: No limits.
6933 Usage: request_header_access header_name allow|deny [!]aclname ...
6935 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6936 this feature could make you liable for problems which it
6939 This option replaces the old 'anonymize_headers' and the
6940 older 'http_anonymizer' option with something that is much
6941 more configurable. A list of ACLs for each header name allows
6942 removal of specific header fields under specific conditions.
6944 This option only applies to outgoing HTTP request headers (i.e.,
6945 headers sent by Squid to the next HTTP hop such as a cache peer
6946 or an origin server). The option has no effect during cache hit
6947 detection. The equivalent adaptation vectoring point in ICAP
6948 terminology is post-cache REQMOD.
6950 The option is applied to individual outgoing request header
6951 fields. For each request header field F, Squid uses the first
6952 qualifying sets of request_header_access rules:
6954 1. Rules with header_name equal to F's name.
6955 2. Rules with header_name 'Other', provided F's name is not
6956 on the hard-coded list of commonly used HTTP header names.
6957 3. Rules with header_name 'All'.
6959 Within that qualifying rule set, rule ACLs are checked as usual.
6960 If ACLs of an "allow" rule match, the header field is allowed to
6961 go through as is. If ACLs of a "deny" rule match, the header is
6962 removed and request_header_replace is then checked to identify
6963 if the removed header has a replacement. If no rules within the
6964 set have matching ACLs, the header field is left as is.
6966 For example, to achieve the same behavior as the old
6967 'http_anonymizer standard' option, you should use:
6969 request_header_access From deny all
6970 request_header_access Referer deny all
6971 request_header_access User-Agent deny all
6973 Or, to reproduce the old 'http_anonymizer paranoid' feature
6976 request_header_access Authorization allow all
6977 request_header_access Proxy-Authorization allow all
6978 request_header_access Cache-Control allow all
6979 request_header_access Content-Length allow all
6980 request_header_access Content-Type allow all
6981 request_header_access Date allow all
6982 request_header_access Host allow all
6983 request_header_access If-Modified-Since allow all
6984 request_header_access Pragma allow all
6985 request_header_access Accept allow all
6986 request_header_access Accept-Charset allow all
6987 request_header_access Accept-Encoding allow all
6988 request_header_access Accept-Language allow all
6989 request_header_access Connection allow all
6990 request_header_access All deny all
6992 HTTP reply headers are controlled with the reply_header_access directive.
6994 By default, all headers are allowed (no anonymizing is performed).
6997 NAME: reply_header_access
6998 IFDEF: USE_HTTP_VIOLATIONS
6999 TYPE: http_header_access
7000 LOC: Config.reply_header_access
7002 DEFAULT_DOC: No limits.
7004 Usage: reply_header_access header_name allow|deny [!]aclname ...
7006 WARNING: Doing this VIOLATES the HTTP standard. Enabling
7007 this feature could make you liable for problems which it
7010 This option only applies to reply headers, i.e., from the
7011 server to the client.
7013 This is the same as request_header_access, but in the other
7014 direction. Please see request_header_access for detailed
7017 For example, to achieve the same behavior as the old
7018 'http_anonymizer standard' option, you should use:
7020 reply_header_access Server deny all
7021 reply_header_access WWW-Authenticate deny all
7022 reply_header_access Link deny all
7024 Or, to reproduce the old 'http_anonymizer paranoid' feature
7027 reply_header_access Allow allow all
7028 reply_header_access WWW-Authenticate allow all
7029 reply_header_access Proxy-Authenticate allow all
7030 reply_header_access Cache-Control allow all
7031 reply_header_access Content-Encoding allow all
7032 reply_header_access Content-Length allow all
7033 reply_header_access Content-Type allow all
7034 reply_header_access Date allow all
7035 reply_header_access Expires allow all
7036 reply_header_access Last-Modified allow all
7037 reply_header_access Location allow all
7038 reply_header_access Pragma allow all
7039 reply_header_access Content-Language allow all
7040 reply_header_access Retry-After allow all
7041 reply_header_access Title allow all
7042 reply_header_access Content-Disposition allow all
7043 reply_header_access Connection allow all
7044 reply_header_access All deny all
7046 HTTP request headers are controlled with the request_header_access directive.
7048 By default, all headers are allowed (no anonymizing is
7052 NAME: request_header_replace header_replace
7053 IFDEF: USE_HTTP_VIOLATIONS
7054 TYPE: http_header_replace
7055 LOC: Config.request_header_access
7058 Usage: request_header_replace header_name message
7059 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
7061 This option allows you to change the contents of headers
7062 denied with request_header_access above, by replacing them
7063 with some fixed string.
7065 This only applies to request headers, not reply headers.
7067 By default, headers are removed if denied.
7070 NAME: reply_header_replace
7071 IFDEF: USE_HTTP_VIOLATIONS
7072 TYPE: http_header_replace
7073 LOC: Config.reply_header_access
7076 Usage: reply_header_replace header_name message
7077 Example: reply_header_replace Server Foo/1.0
7079 This option allows you to change the contents of headers
7080 denied with reply_header_access above, by replacing them
7081 with some fixed string.
7083 This only applies to reply headers, not request headers.
7085 By default, headers are removed if denied.
7088 NAME: request_header_add
7089 TYPE: HeaderWithAclList
7090 LOC: Config.request_header_add
7093 Usage: request_header_add field-name field-value [ acl ... ]
7094 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
7096 This option adds header fields to outgoing HTTP requests (i.e.,
7097 request headers sent by Squid to the next HTTP hop such as a
7098 cache peer or an origin server). The option has no effect during
7099 cache hit detection. The equivalent adaptation vectoring point
7100 in ICAP terminology is post-cache REQMOD.
7102 Field-name is a token specifying an HTTP header name. If a
7103 standard HTTP header name is used, Squid does not check whether
7104 the new header conflicts with any existing headers or violates
7105 HTTP rules. If the request to be modified already contains a
7106 field with the same name, the old field is preserved but the
7107 header field values are not merged.
7109 Field-value is either a token or a quoted string. If quoted
7110 string format is used, then the surrounding quotes are removed
7111 while escape sequences and %macros are processed.
7113 One or more Squid ACLs may be specified to restrict header
7114 injection to matching requests. As always in squid.conf, all
7115 ACLs in the ACL list must be satisfied for the insertion to
7116 happen. The request_header_add supports fast ACLs only.
7118 See also: reply_header_add.
7121 NAME: reply_header_add
7122 TYPE: HeaderWithAclList
7123 LOC: Config.reply_header_add
7126 Usage: reply_header_add field-name field-value [ acl ... ]
7127 Example: reply_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
7129 This option adds header fields to outgoing HTTP responses (i.e., response
7130 headers delivered by Squid to the client). This option has no effect on
7131 cache hit detection. The equivalent adaptation vectoring point in
7132 ICAP terminology is post-cache RESPMOD. This option does not apply to
7133 successful CONNECT replies.
7135 Field-name is a token specifying an HTTP header name. If a
7136 standard HTTP header name is used, Squid does not check whether
7137 the new header conflicts with any existing headers or violates
7138 HTTP rules. If the response to be modified already contains a
7139 field with the same name, the old field is preserved but the
7140 header field values are not merged.
7142 Field-value is either a token or a quoted string. If quoted
7143 string format is used, then the surrounding quotes are removed
7144 while escape sequences and %macros are processed.
7146 One or more Squid ACLs may be specified to restrict header
7147 injection to matching responses. As always in squid.conf, all
7148 ACLs in the ACL list must be satisfied for the insertion to
7149 happen. The reply_header_add option supports fast ACLs only.
7151 See also: request_header_add.
7159 This option used to log custom information about the master
7160 transaction. For example, an admin may configure Squid to log
7161 which "user group" the transaction belongs to, where "user group"
7162 will be determined based on a set of ACLs and not [just]
7163 authentication information.
7164 Values of key/value pairs can be logged using %{key}note macros:
7166 note key value acl ...
7167 logformat myFormat ... %{key}note ...
7169 This clause only supports fast acl types.
7170 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7173 NAME: relaxed_header_parser
7174 COMMENT: on|off|warn
7176 LOC: Config.onoff.relaxed_header_parser
7179 In the default "on" setting Squid accepts certain forms
7180 of non-compliant HTTP messages where it is unambiguous
7181 what the sending application intended even if the message
7182 is not correctly formatted. The messages is then normalized
7183 to the correct form when forwarded by Squid.
7185 If set to "warn" then a warning will be emitted in cache.log
7186 each time such HTTP error is encountered.
7188 If set to "off" then such HTTP errors will cause the request
7189 or response to be rejected.
7192 NAME: collapsed_forwarding
7195 LOC: Config.onoff.collapsed_forwarding
7198 This option controls whether Squid is allowed to merge multiple
7199 potentially cachable requests for the same URI before Squid knows
7200 whether the response is going to be cachable.
7202 When enabled, instead of forwarding each concurrent request for
7203 the same URL, Squid just sends the first of them. The other, so
7204 called "collapsed" requests, wait for the response to the first
7205 request and, if it happens to be cachable, use that response.
7206 Here, "concurrent requests" means "received after the first
7207 request headers were parsed and before the corresponding response
7208 headers were parsed".
7210 This feature is disabled by default: enabling collapsed
7211 forwarding needlessly delays forwarding requests that look
7212 cachable (when they are collapsed) but then need to be forwarded
7213 individually anyway because they end up being for uncachable
7214 content. However, in some cases, such as acceleration of highly
7215 cachable content with periodic or grouped expiration times, the
7216 gains from collapsing [large volumes of simultaneous refresh
7217 requests] outweigh losses from such delays.
7219 Squid collapses two kinds of requests: regular client requests
7220 received on one of the listening ports and internal "cache
7221 revalidation" requests which are triggered by those regular
7222 requests hitting a stale cached object. Revalidation collapsing
7223 is currently disabled for Squid instances containing SMP-aware
7224 disk or memory caches and for Vary-controlled cached objects.
7227 NAME: collapsed_forwarding_access
7230 DEFAULT_DOC: Requests may be collapsed if collapsed_forwarding is on.
7231 LOC: Config.accessList.collapsedForwardingAccess
7233 Use this directive to restrict collapsed forwarding to a subset of
7234 eligible requests. The directive is checked for regular HTTP
7235 requests, internal revalidation requests, and HTCP/ICP requests.
7237 collapsed_forwarding_access allow|deny [!]aclname ...
7239 This directive cannot force collapsing. It has no effect on
7240 collapsing unless collapsed_forwarding is 'on', and all other
7241 collapsing preconditions are satisfied.
7243 * A denied request will not collapse, and future transactions will
7244 not collapse on it (even if they are allowed to collapse).
7246 * An allowed request may collapse, or future transactions may
7247 collapse on it (provided they are allowed to collapse).
7249 This directive is evaluated before receiving HTTP response headers
7250 and without access to Squid-to-peer connection (if any).
7252 Only fast ACLs are supported.
7254 See also: collapsed_forwarding.
7257 NAME: shared_transient_entries_limit collapsed_forwarding_shared_entries_limit
7258 COMMENT: (number of entries)
7260 LOC: Config.shared_transient_entries_limit
7263 This directive limits the size of a table used for sharing current
7264 transaction information among SMP workers. A table entry stores meta
7265 information about a single cache entry being delivered to Squid
7266 client(s) by one or more SMP workers. A single table entry consumes
7267 less than 128 shared memory bytes.
7269 The limit should be significantly larger than the number of
7270 concurrent non-collapsed cachable responses leaving Squid. For a
7271 cache that handles less than 5000 concurrent requests, the default
7272 setting of 16384 should be plenty.
7274 Using excessively large values wastes shared memory. Limiting the
7275 table size too much results in hash collisions, leading to lower hit
7276 ratio and missed SMP request collapsing opportunities: Transactions
7277 left without a table entry cannot cache their responses and are
7278 invisible to other concurrent requests for the same resource.
7280 A zero limit is allowed but unsupported. A positive small limit
7281 lowers hit ratio, but zero limit disables a lot of essential
7282 synchronization among SMP workers, leading to HTTP violations (e.g.,
7283 stale hit responses). It also disables shared collapsed forwarding:
7284 A worker becomes unable to collapse its requests on transactions in
7285 other workers, resulting in more trips to the origin server and more
7291 -----------------------------------------------------------------------------
7294 NAME: forward_timeout
7297 LOC: Config.Timeout.forward
7300 This parameter specifies how long Squid should at most attempt in
7301 finding a forwarding path for the request before giving up.
7304 NAME: connect_timeout
7307 LOC: Config.Timeout.connect
7310 This parameter specifies how long to wait for the TCP connect to
7311 the requested server or peer to complete before Squid should
7312 attempt to find another path where to forward the request.
7315 NAME: peer_connect_timeout
7318 LOC: Config.Timeout.peer_connect
7321 This parameter specifies how long to wait for a pending TCP
7322 connection to a peer cache. The default is 30 seconds. You
7323 may also set different timeout values for individual neighbors
7324 with the 'connect-timeout' option on a 'cache_peer' line.
7330 LOC: Config.Timeout.read
7333 Applied on peer server connections.
7335 After each successful read(), the timeout will be extended by this
7336 amount. If no data is read again after this amount of time,
7337 the request is aborted and logged with ERR_READ_TIMEOUT.
7339 The default is 15 minutes.
7345 LOC: Config.Timeout.write
7348 This timeout is tracked for all connections that have data
7349 available for writing and are waiting for the socket to become
7350 ready. After each successful write, the timeout is extended by
7351 the configured amount. If Squid has data to write but the
7352 connection is not ready for the configured duration, the
7353 transaction associated with the connection is terminated. The
7354 default is 15 minutes.
7357 NAME: request_timeout
7359 LOC: Config.Timeout.request
7362 How long to wait for complete HTTP request headers after initial
7363 connection establishment.
7366 NAME: request_start_timeout
7368 LOC: Config.Timeout.request_start_timeout
7371 How long to wait for the first request byte after initial
7372 connection establishment.
7375 NAME: client_idle_pconn_timeout persistent_request_timeout
7377 LOC: Config.Timeout.clientIdlePconn
7380 How long to wait for the next HTTP request on a persistent
7381 client connection after the previous request completes.
7384 NAME: ftp_client_idle_timeout
7386 LOC: Config.Timeout.ftpClientIdle
7389 How long to wait for an FTP request on a connection to Squid ftp_port.
7390 Many FTP clients do not deal with idle connection closures well,
7391 necessitating a longer default timeout than client_idle_pconn_timeout
7392 used for incoming HTTP requests.
7395 NAME: client_lifetime
7398 LOC: Config.Timeout.lifetime
7401 The maximum amount of time a client (browser) is allowed to
7402 remain connected to the cache process. This protects the Cache
7403 from having a lot of sockets (and hence file descriptors) tied up
7404 in a CLOSE_WAIT state from remote clients that go away without
7405 properly shutting down (either because of a network failure or
7406 because of a poor client implementation). The default is one
7409 NOTE: The default value is intended to be much larger than any
7410 client would ever need to be connected to your cache. You
7411 should probably change client_lifetime only as a last resort.
7412 If you seem to have many client connections tying up
7413 filedescriptors, we recommend first tuning the read_timeout,
7414 request_timeout, persistent_request_timeout and quick_abort values.
7417 NAME: pconn_lifetime
7420 LOC: Config.Timeout.pconnLifetime
7423 Desired maximum lifetime of a persistent connection.
7424 When set, Squid will close a now-idle persistent connection that
7425 exceeded configured lifetime instead of moving the connection into
7426 the idle connection pool (or equivalent). No effect on ongoing/active
7427 transactions. Connection lifetime is the time period from the
7428 connection acceptance or opening time until "now".
7430 This limit is useful in environments with long-lived connections
7431 where Squid configuration or environmental factors change during a
7432 single connection lifetime. If unrestricted, some connections may
7433 last for hours and even days, ignoring those changes that should
7434 have affected their behavior or their existence.
7436 Currently, a new lifetime value supplied via Squid reconfiguration
7437 has no effect on already idle connections unless they become busy.
7439 When set to '0' this limit is not used.
7442 NAME: half_closed_clients
7444 LOC: Config.onoff.half_closed_clients
7447 Some clients may shutdown the sending side of their TCP
7448 connections, while leaving their receiving sides open. Sometimes,
7449 Squid can not tell the difference between a half-closed and a
7450 fully-closed TCP connection.
7452 By default, Squid will immediately close client connections when
7453 read(2) returns "no more data to read."
7455 Change this option to 'on' and Squid will keep open connections
7456 until a read(2) or write(2) on the socket returns an error.
7457 This may show some benefits for reverse proxies. But if not
7458 it is recommended to leave OFF.
7461 NAME: server_idle_pconn_timeout pconn_timeout
7463 LOC: Config.Timeout.serverIdlePconn
7466 Timeout for idle persistent connections to servers and other
7473 LOC: Ident::TheConfig.timeout
7476 Maximum time to wait for IDENT lookups to complete.
7478 If this is too high, and you enabled IDENT lookups from untrusted
7479 users, you might be susceptible to denial-of-service by having
7480 many ident requests going at once.
7483 NAME: shutdown_lifetime
7486 LOC: Config.shutdownLifetime
7489 When SIGTERM or SIGHUP is received, the cache is put into
7490 "shutdown pending" mode until all active sockets are closed.
7491 This value is the lifetime to set for all open descriptors
7492 during shutdown mode. Any active clients after this many
7493 seconds will receive a 'timeout' message.
7497 ADMINISTRATIVE PARAMETERS
7498 -----------------------------------------------------------------------------
7504 LOC: Config.adminEmail
7506 Email-address of local cache manager who will receive
7507 mail if the cache dies. The default is "webmaster".
7513 LOC: Config.EmailFrom
7515 From: email-address for mail sent when the cache dies.
7516 The default is to use 'squid@unique_hostname'.
7518 See also: unique_hostname directive.
7524 LOC: Config.EmailProgram
7526 Email program used to send mail if the cache dies.
7527 The default is "mail". The specified program must comply
7528 with the standard Unix mail syntax:
7529 mail-program recipient < mailfile
7531 Optional command line options can be specified.
7534 NAME: cache_effective_user
7536 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
7537 LOC: Config.effectiveUser
7539 If you start Squid as root, it will change its effective/real
7540 UID/GID to the user specified below. The default is to change
7541 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
7542 see also; cache_effective_group
7545 NAME: cache_effective_group
7548 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
7549 LOC: Config.effectiveGroup
7551 Squid sets the GID to the effective user's default group ID
7552 (taken from the password file) and supplementary group list
7553 from the groups membership.
7555 If you want Squid to run with a specific GID regardless of
7556 the group memberships of the effective user then set this
7557 to the group (or GID) you want Squid to run as. When set
7558 all other group privileges of the effective user are ignored
7559 and only this GID is effective. If Squid is not started as
7560 root the user starting Squid MUST be member of the specified
7563 This option is not recommended by the Squid Team.
7564 Our preference is for administrators to configure a secure
7565 user account for squid with UID/GID matching system policies.
7568 NAME: httpd_suppress_version_string
7572 LOC: Config.onoff.httpd_suppress_version_string
7574 Do not send Squid version string in HTTP metadata and generated content
7575 such as HTML error pages. Squid version string is still present in certain
7576 SNMP responses, HTTP(S) Server response header field,
7577 various console output, and cache.log.
7580 NAME: visible_hostname
7582 LOC: Config.visibleHostname
7584 DEFAULT_DOC: Automatically detect the system host name
7586 If you want to present a special hostname in error messages, etc,
7587 define this. Otherwise, the return value of gethostname()
7588 will be used. If you have multiple caches in a cluster and
7589 get errors about IP-forwarding you must set them to have individual
7590 names with this setting.
7593 NAME: unique_hostname
7595 LOC: Config.uniqueHostname
7597 DEFAULT_DOC: Copy the value from visible_hostname
7599 If you want to have multiple machines with the same
7600 'visible_hostname' you must give each machine a different
7601 'unique_hostname' so forwarding loops can be detected.
7604 NAME: hostname_aliases
7606 LOC: Config.hostnameAliases
7609 A list of other DNS names your cache has.
7617 Minimum umask which should be enforced while the proxy
7618 is running, in addition to the umask set at startup.
7620 For a traditional octal representation of umasks, start
7625 HTTPD-ACCELERATOR OPTIONS
7626 -----------------------------------------------------------------------------
7629 NAME: httpd_accel_surrogate_id
7632 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
7633 LOC: Config.Accel.surrogate_id
7635 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
7636 need an identification token to allow control targeting. Because
7637 a farm of surrogates may all perform the same tasks, they may share
7638 an identification token.
7640 When the surrogate is a reverse-proxy, this ID is also
7641 used as cdn-id for CDN-Loop detection (RFC 8586).
7644 NAME: http_accel_surrogate_remote
7648 LOC: Config.onoff.surrogate_is_remote
7650 Remote surrogates (such as those in a CDN) honour the header
7651 "Surrogate-Control: no-store-remote".
7653 Set this to on to have squid behave as a remote surrogate.
7657 IFDEF: USE_SQUID_ESI
7658 COMMENT: libxml2|expat
7660 LOC: ESIParser::Type
7662 DEFAULT_DOC: Selects libxml2 if available at ./configure time or libexpat otherwise.
7664 Selects the XML parsing library to use when interpreting responses with
7667 To disable ESI handling completely, ./configure Squid with --disable-esi.
7671 DELAY POOL PARAMETERS
7672 -----------------------------------------------------------------------------
7676 TYPE: delay_pool_count
7678 IFDEF: USE_DELAY_POOLS
7681 This represents the number of delay pools to be used. For example,
7682 if you have one class 2 delay pool and one class 3 delays pool, you
7683 have a total of 2 delay pools.
7685 See also delay_parameters, delay_class, delay_access for pool
7686 configuration details.
7690 TYPE: delay_pool_class
7692 IFDEF: USE_DELAY_POOLS
7695 This defines the class of each delay pool. There must be exactly one
7696 delay_class line for each delay pool. For example, to define two
7697 delay pools, one of class 2 and one of class 3, the settings above
7701 delay_pools 4 # 4 delay pools
7702 delay_class 1 2 # pool 1 is a class 2 pool
7703 delay_class 2 3 # pool 2 is a class 3 pool
7704 delay_class 3 4 # pool 3 is a class 4 pool
7705 delay_class 4 5 # pool 4 is a class 5 pool
7707 The delay pool classes are:
7709 class 1 Everything is limited by a single aggregate
7712 class 2 Everything is limited by a single aggregate
7713 bucket as well as an "individual" bucket chosen
7714 from bits 25 through 32 of the IPv4 address.
7716 class 3 Everything is limited by a single aggregate
7717 bucket as well as a "network" bucket chosen
7718 from bits 17 through 24 of the IP address and a
7719 "individual" bucket chosen from bits 17 through
7720 32 of the IPv4 address.
7722 class 4 Everything in a class 3 delay pool, with an
7723 additional limit on a per user basis. This
7724 only takes effect if the username is established
7725 in advance - by forcing authentication in your
7728 class 5 Requests are grouped according their tag (see
7729 external_acl's tag= reply).
7732 Each pool also requires a delay_parameters directive to configure the pool size
7733 and speed limits used whenever the pool is applied to a request. Along with
7734 a set of delay_access directives to determine when it is used.
7736 NOTE: If an IP address is a.b.c.d
7737 -> bits 25 through 32 are "d"
7738 -> bits 17 through 24 are "c"
7739 -> bits 17 through 32 are "c * 256 + d"
7741 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
7742 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
7744 This clause only supports fast acl types.
7745 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7747 See also delay_parameters and delay_access.
7751 TYPE: delay_pool_access
7753 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
7754 IFDEF: USE_DELAY_POOLS
7757 This is used to determine which delay pool a request falls into.
7759 delay_access is sorted per pool and the matching starts with pool 1,
7760 then pool 2, ..., and finally pool N. The first delay pool where the
7761 request is allowed is selected for the request. If it does not allow
7762 the request to any pool then the request is not delayed (default).
7764 For example, if you want some_big_clients in delay
7765 pool 1 and lotsa_little_clients in delay pool 2:
7767 delay_access 1 allow some_big_clients
7768 delay_access 1 deny all
7769 delay_access 2 allow lotsa_little_clients
7770 delay_access 2 deny all
7771 delay_access 3 allow authenticated_clients
7773 See also delay_parameters and delay_class.
7777 NAME: delay_parameters
7778 TYPE: delay_pool_rates
7780 IFDEF: USE_DELAY_POOLS
7783 This defines the parameters for a delay pool. Each delay pool has
7784 a number of "buckets" associated with it, as explained in the
7785 description of delay_class.
7787 For a class 1 delay pool, the syntax is:
7789 delay_parameters pool aggregate
7791 For a class 2 delay pool:
7793 delay_parameters pool aggregate individual
7795 For a class 3 delay pool:
7797 delay_parameters pool aggregate network individual
7799 For a class 4 delay pool:
7801 delay_parameters pool aggregate network individual user
7803 For a class 5 delay pool:
7805 delay_parameters pool tagrate
7807 The option variables are:
7809 pool a pool number - ie, a number between 1 and the
7810 number specified in delay_pools as used in
7813 aggregate the speed limit parameters for the aggregate bucket
7816 individual the speed limit parameters for the individual
7817 buckets (class 2, 3).
7819 network the speed limit parameters for the network buckets
7822 user the speed limit parameters for the user buckets
7825 tagrate the speed limit parameters for the tag buckets
7828 A pair of delay parameters is written restore/maximum, where restore is
7829 the number of bytes (not bits - modem and network speeds are usually
7830 quoted in bits) per second placed into the bucket, and maximum is the
7831 maximum number of bytes which can be in the bucket at any time.
7833 There must be one delay_parameters line for each delay pool.
7836 For example, if delay pool number 1 is a class 2 delay pool as in the
7837 above example, and is being used to strictly limit each host to 64Kbit/sec
7838 (plus overheads), with no overall limit, the line is:
7840 delay_parameters 1 none 8000/8000
7842 Note that 8 x 8K Byte/sec -> 64K bit/sec.
7844 Note that the word 'none' is used to represent no limit.
7847 And, if delay pool number 2 is a class 3 delay pool as in the above
7848 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
7849 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
7850 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
7851 to permit a decent web page to be downloaded at a decent speed
7852 (if the network is not being limited due to overuse) but slow down
7853 large downloads more significantly:
7855 delay_parameters 2 32000/32000 8000/8000 600/8000
7857 Note that 8 x 32K Byte/sec -> 256K bit/sec.
7858 8 x 8K Byte/sec -> 64K bit/sec.
7859 8 x 600 Byte/sec -> 4800 bit/sec.
7862 Finally, for a class 4 delay pool as in the example - each user will
7863 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
7865 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
7868 See also delay_class and delay_access.
7872 NAME: delay_initial_bucket_level
7873 COMMENT: (percent, 0-100)
7876 IFDEF: USE_DELAY_POOLS
7877 LOC: Config.Delay.initial
7879 The initial bucket percentage is used to determine how much is put
7880 in each bucket when squid starts, is reconfigured, or first notices
7881 a host accessing it (in class 2 and class 3, individual hosts and
7882 networks only have buckets associated with them once they have been
7887 CLIENT DELAY POOL PARAMETERS
7888 -----------------------------------------------------------------------------
7891 NAME: client_delay_pools
7892 TYPE: client_delay_pool_count
7894 IFDEF: USE_DELAY_POOLS
7895 LOC: Config.ClientDelay
7897 This option specifies the number of client delay pools used. It must
7898 preceed other client_delay_* options.
7901 client_delay_pools 2
7903 See also client_delay_parameters and client_delay_access.
7906 NAME: client_delay_initial_bucket_level
7907 COMMENT: (percent, 0-no_limit)
7910 IFDEF: USE_DELAY_POOLS
7911 LOC: Config.ClientDelay.initial
7913 This option determines the initial bucket size as a percentage of
7914 max_bucket_size from client_delay_parameters. Buckets are created
7915 at the time of the "first" connection from the matching IP. Idle
7916 buckets are periodically deleted up.
7918 You can specify more than 100 percent but note that such "oversized"
7919 buckets are not refilled until their size goes down to max_bucket_size
7920 from client_delay_parameters.
7923 client_delay_initial_bucket_level 50
7926 NAME: client_delay_parameters
7927 TYPE: client_delay_pool_rates
7929 IFDEF: USE_DELAY_POOLS
7930 LOC: Config.ClientDelay
7933 This option configures client-side bandwidth limits using the
7936 client_delay_parameters pool speed_limit max_bucket_size
7938 pool is an integer ID used for client_delay_access matching.
7940 speed_limit is bytes added to the bucket per second.
7942 max_bucket_size is the maximum size of a bucket, enforced after any
7943 speed_limit additions.
7945 Please see the delay_parameters option for more information and
7949 client_delay_parameters 1 1024 2048
7950 client_delay_parameters 2 51200 16384
7952 See also client_delay_access.
7956 NAME: client_delay_access
7957 TYPE: client_delay_pool_access
7959 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
7960 IFDEF: USE_DELAY_POOLS
7961 LOC: Config.ClientDelay
7963 This option determines the client-side delay pool for the
7966 client_delay_access pool_ID allow|deny acl_name
7968 All client_delay_access options are checked in their pool ID
7969 order, starting with pool 1. The first checked pool with allowed
7970 request is selected for the request. If no ACL matches or there
7971 are no client_delay_access options, the request bandwidth is not
7974 The ACL-selected pool is then used to find the
7975 client_delay_parameters for the request. Client-side pools are
7976 not used to aggregate clients. Clients are always aggregated
7977 based on their source IP addresses (one bucket per source IP).
7979 This clause only supports fast acl types.
7980 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7981 Additionally, only the client TCP connection details are available.
7982 ACLs testing HTTP properties will not work.
7984 Please see delay_access for more examples.
7987 client_delay_access 1 allow low_rate_network
7988 client_delay_access 2 allow vips_network
7991 See also client_delay_parameters and client_delay_pools.
7994 NAME: response_delay_pool
7995 TYPE: response_delay_pool_parameters
7997 IFDEF: USE_DELAY_POOLS
7998 LOC: Config.MessageDelay
8000 This option configures client response bandwidth limits using the
8003 response_delay_pool name [option=value] ...
8005 name the response delay pool name
8009 individual-restore The speed limit of an individual
8010 bucket(bytes/s). To be used in conjunction
8011 with 'individual-maximum'.
8013 individual-maximum The maximum number of bytes which can
8014 be placed into the individual bucket. To be used
8015 in conjunction with 'individual-restore'.
8017 aggregate-restore The speed limit for the aggregate
8018 bucket(bytes/s). To be used in conjunction with
8019 'aggregate-maximum'.
8021 aggregate-maximum The maximum number of bytes which can
8022 be placed into the aggregate bucket. To be used
8023 in conjunction with 'aggregate-restore'.
8025 initial-bucket-level The initial bucket size as a percentage
8026 of individual-maximum.
8028 Individual and(or) aggregate bucket options may not be specified,
8029 meaning no individual and(or) aggregate speed limitation.
8030 See also response_delay_pool_access and delay_parameters for
8031 terminology details.
8034 NAME: response_delay_pool_access
8035 TYPE: response_delay_pool_access
8037 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
8038 IFDEF: USE_DELAY_POOLS
8039 LOC: Config.MessageDelay
8041 Determines whether a specific named response delay pool is used
8042 for the transaction. The syntax for this directive is:
8044 response_delay_pool_access pool_name allow|deny acl_name
8046 All response_delay_pool_access options are checked in the order
8047 they appear in this configuration file. The first rule with a
8048 matching ACL wins. If (and only if) an "allow" rule won, Squid
8049 assigns the response to the corresponding named delay pool.
8053 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
8054 -----------------------------------------------------------------------------
8059 LOC: Config.Wccp.router
8061 DEFAULT_DOC: WCCP disabled.
8064 Use this option to define your WCCP ``home'' router for
8067 wccp_router supports a single WCCP(v1) router
8069 wccp2_router supports multiple WCCPv2 routers
8071 only one of the two may be used at the same time and defines
8072 which version of WCCP to use.
8076 TYPE: IpAddress_list
8077 LOC: Config.Wccp2.router
8079 DEFAULT_DOC: WCCPv2 disabled.
8082 Use this option to define your WCCP ``home'' router for
8085 wccp_router supports a single WCCP(v1) router
8087 wccp2_router supports multiple WCCPv2 routers
8089 only one of the two may be used at the same time and defines
8090 which version of WCCP to use.
8095 LOC: Config.Wccp.version
8099 This directive is only relevant if you need to set up WCCP(v1)
8100 to some very old and end-of-life Cisco routers. In all other
8101 setups it must be left unset or at the default setting.
8102 It defines an internal version in the WCCP(v1) protocol,
8103 with version 4 being the officially documented protocol.
8105 According to some users, Cisco IOS 11.2 and earlier only
8106 support WCCP version 3. If you're using that or an earlier
8107 version of IOS, you may need to change this value to 3, otherwise
8108 do not specify this parameter.
8111 NAME: wccp2_rebuild_wait
8113 LOC: Config.Wccp2.rebuildwait
8117 If this is enabled Squid will wait for the cache dir rebuild to finish
8118 before sending the first wccp2 HereIAm packet
8121 NAME: wccp2_forwarding_method
8123 LOC: Config.Wccp2.forwarding_method
8127 WCCP2 allows the setting of forwarding methods between the
8128 router/switch and the cache. Valid values are as follows:
8130 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
8131 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
8133 Currently (as of IOS 12.4) cisco routers only support GRE.
8134 Cisco switches only support the L2 redirect assignment method.
8137 NAME: wccp2_return_method
8139 LOC: Config.Wccp2.return_method
8143 WCCP2 allows the setting of return methods between the
8144 router/switch and the cache for packets that the cache
8145 decides not to handle. Valid values are as follows:
8147 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
8148 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
8150 Currently (as of IOS 12.4) cisco routers only support GRE.
8151 Cisco switches only support the L2 redirect assignment.
8153 If the "ip wccp redirect exclude in" command has been
8154 enabled on the cache interface, then it is still safe for
8155 the proxy server to use a l2 redirect method even if this
8156 option is set to GRE.
8159 NAME: wccp2_assignment_method
8161 LOC: Config.Wccp2.assignment_method
8165 WCCP2 allows the setting of methods to assign the WCCP hash
8166 Valid values are as follows:
8168 hash - Hash assignment
8169 mask - Mask assignment
8171 As a general rule, cisco routers support the hash assignment method
8172 and cisco switches support the mask assignment method.
8177 LOC: Config.Wccp2.info
8178 DEFAULT_IF_NONE: standard 0
8179 DEFAULT_DOC: Use the 'web-cache' standard service.
8182 WCCP2 allows for multiple traffic services. There are two
8183 types: "standard" and "dynamic". The standard type defines
8184 one service id - http (id 0). The dynamic service ids can be from
8185 51 to 255 inclusive. In order to use a dynamic service id
8186 one must define the type of traffic to be redirected; this is done
8187 using the wccp2_service_info option.
8189 The "standard" type does not require a wccp2_service_info option,
8190 just specifying the service id will suffice.
8192 MD5 service authentication can be enabled by adding
8193 "password=<password>" to the end of this service declaration.
8197 wccp2_service standard 0 # for the 'web-cache' standard service
8198 wccp2_service dynamic 80 # a dynamic service type which will be
8199 # fleshed out with subsequent options.
8200 wccp2_service standard 0 password=foo
8203 NAME: wccp2_service_info
8204 TYPE: wccp2_service_info
8205 LOC: Config.Wccp2.info
8209 Dynamic WCCPv2 services require further information to define the
8210 traffic you wish to have diverted.
8214 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
8215 priority=<priority> ports=<port>,<port>..
8217 The relevant WCCPv2 flags:
8218 + src_ip_hash, dst_ip_hash
8219 + source_port_hash, dst_port_hash
8220 + src_ip_alt_hash, dst_ip_alt_hash
8221 + src_port_alt_hash, dst_port_alt_hash
8224 The port list can be one to eight entries.
8228 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
8229 priority=240 ports=80
8231 Note: the service id must have been defined by a previous
8232 'wccp2_service dynamic <id>' entry.
8237 LOC: Config.Wccp2.weight
8241 Each cache server gets assigned a set of the destination
8242 hash proportional to their weight.
8247 LOC: Config.Wccp.address
8249 DEFAULT_DOC: Address selected by the operating system.
8252 Use this option if you require WCCP(v1) to use a specific
8255 The default behavior is to not bind to any specific address.
8260 LOC: Config.Wccp2.address
8262 DEFAULT_DOC: Address selected by the operating system.
8265 Use this option if you require WCCPv2 to use a specific
8268 The default behavior is to not bind to any specific address.
8272 PERSISTENT CONNECTION HANDLING
8273 -----------------------------------------------------------------------------
8275 Also see "pconn_timeout" in the TIMEOUTS section
8278 NAME: client_persistent_connections
8280 LOC: Config.onoff.client_pconns
8283 Persistent connection support for clients.
8284 Squid uses persistent connections (when allowed). You can use
8285 this option to disable persistent connections with clients.
8288 NAME: server_persistent_connections
8290 LOC: Config.onoff.server_pconns
8293 Persistent connection support for servers.
8294 Squid uses persistent connections (when allowed). You can use
8295 this option to disable persistent connections with servers.
8298 NAME: persistent_connection_after_error
8300 LOC: Config.onoff.error_pconns
8303 With this directive the use of persistent connections after
8304 HTTP errors can be disabled. Useful if you have clients
8305 who fail to handle errors on persistent connections proper.
8308 NAME: detect_broken_pconn
8310 LOC: Config.onoff.detect_broken_server_pconns
8313 Some servers have been found to incorrectly signal the use
8314 of HTTP/1.0 persistent connections even on replies not
8315 compatible, causing significant delays. This server problem
8316 has mostly been seen on redirects.
8318 By enabling this directive Squid attempts to detect such
8319 broken replies and automatically assume the reply is finished
8320 after 10 seconds timeout.
8324 CACHE DIGEST OPTIONS
8325 -----------------------------------------------------------------------------
8328 NAME: digest_generation
8329 IFDEF: USE_CACHE_DIGESTS
8331 LOC: Config.onoff.digest_generation
8334 This controls whether the server will generate a Cache Digest
8335 of its contents. By default, Cache Digest generation is
8336 enabled if Squid is compiled with --enable-cache-digests defined.
8339 NAME: digest_bits_per_entry
8340 IFDEF: USE_CACHE_DIGESTS
8342 LOC: Config.digest.bits_per_entry
8345 This is the number of bits of the server's Cache Digest which
8346 will be associated with the Digest entry for a given HTTP
8347 Method and URL (public key) combination. The default is 5.
8350 NAME: digest_rebuild_period
8351 IFDEF: USE_CACHE_DIGESTS
8354 LOC: Config.digest.rebuild_period
8357 This is the wait time between Cache Digest rebuilds.
8360 NAME: digest_rewrite_period
8362 IFDEF: USE_CACHE_DIGESTS
8364 LOC: Config.digest.rewrite_period
8367 This is the wait time between Cache Digest writes to
8371 NAME: digest_swapout_chunk_size
8374 IFDEF: USE_CACHE_DIGESTS
8375 LOC: Config.digest.swapout_chunk_size
8378 This is the number of bytes of the Cache Digest to write to
8379 disk at a time. It defaults to 4096 bytes (4KB), the Squid
8383 NAME: digest_rebuild_chunk_percentage
8384 COMMENT: (percent, 0-100)
8385 IFDEF: USE_CACHE_DIGESTS
8387 LOC: Config.digest.rebuild_chunk_percentage
8390 This is the percentage of the Cache Digest to be scanned at a
8391 time. By default it is set to 10% of the Cache Digest.
8396 -----------------------------------------------------------------------------
8401 LOC: Config.Port.snmp
8403 DEFAULT_DOC: SNMP disabled.
8406 The port number where Squid listens for SNMP requests. To enable
8407 SNMP support set this to a suitable port number. Port number
8408 3401 is often used for the Squid SNMP agent. By default it's
8409 set to "0" (disabled)
8417 LOC: Config.accessList.snmp
8419 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
8422 Allowing or denying access to the SNMP port.
8424 All access to the agent is denied by default.
8427 snmp_access allow|deny [!]aclname ...
8429 This clause only supports fast acl types.
8430 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8433 snmp_access allow snmppublic localhost
8434 snmp_access deny all
8437 NAME: snmp_incoming_address
8439 LOC: Config.Addrs.snmp_incoming
8441 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
8444 Just like 'udp_incoming_address', but for the SNMP port.
8446 snmp_incoming_address is used for the SNMP socket receiving
8447 messages from SNMP agents.
8449 The default snmp_incoming_address is to listen on all
8450 available network interfaces.
8453 NAME: snmp_outgoing_address
8455 LOC: Config.Addrs.snmp_outgoing
8457 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
8460 Just like 'udp_outgoing_address', but for the SNMP port.
8462 snmp_outgoing_address is used for SNMP packets returned to SNMP
8465 If snmp_outgoing_address is not set it will use the same socket
8466 as snmp_incoming_address. Only change this if you want to have
8467 SNMP replies sent using another address than where this Squid
8468 listens for SNMP queries.
8470 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
8471 the same value since they both use the same port.
8476 -----------------------------------------------------------------------------
8479 NAME: icp_port udp_port
8482 DEFAULT_DOC: ICP disabled.
8483 LOC: Config.Port.icp
8485 The port number where Squid sends and receives ICP queries to
8486 and from neighbor caches. The standard UDP port for ICP is 3130.
8489 icp_port @DEFAULT_ICP_PORT@
8496 DEFAULT_DOC: HTCP disabled.
8497 LOC: Config.Port.htcp
8499 The port number where Squid sends and receives HTCP queries to
8500 and from neighbor caches. To turn it on you want to set it to
8507 NAME: log_icp_queries
8511 LOC: Config.onoff.log_udp
8513 If set, ICP queries are logged to access.log. You may wish
8514 do disable this if your ICP load is VERY high to speed things
8515 up or to simplify log analysis.
8518 NAME: udp_incoming_address
8520 LOC:Config.Addrs.udp_incoming
8522 DEFAULT_DOC: Accept packets from all machine interfaces.
8524 udp_incoming_address is used for UDP packets received from other
8527 The default behavior is to not bind to any specific address.
8529 Only change this if you want to have all UDP queries received on
8530 a specific interface/address.
8532 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
8533 modules. Altering it will affect all of them in the same manner.
8535 see also; udp_outgoing_address
8537 NOTE, udp_incoming_address and udp_outgoing_address can not
8538 have the same value since they both use the same port.
8541 NAME: udp_outgoing_address
8543 LOC: Config.Addrs.udp_outgoing
8545 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
8547 udp_outgoing_address is used for UDP packets sent out to other
8550 The default behavior is to not bind to any specific address.
8552 Instead it will use the same socket as udp_incoming_address.
8553 Only change this if you want to have UDP queries sent using another
8554 address than where this Squid listens for UDP queries from other
8557 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
8558 modules. Altering it will affect all of them in the same manner.
8560 see also; udp_incoming_address
8562 NOTE, udp_incoming_address and udp_outgoing_address can not
8563 have the same value since they both use the same port.
8570 LOC: Config.onoff.icp_hit_stale
8572 If you want to return ICP_HIT for stale cache objects, set this
8573 option to 'on'. If you have sibling relationships with caches
8574 in other administrative domains, this should be 'off'. If you only
8575 have sibling relationships with caches under your control,
8576 it is probably okay to set this to 'on'.
8577 If set to 'on', your siblings should use the option "allow-miss"
8578 on their cache_peer lines for connecting to you.
8581 NAME: minimum_direct_hops
8584 LOC: Config.minDirectHops
8586 If using the ICMP pinging stuff, do direct fetches for sites
8587 which are no more than this many hops away.
8590 NAME: minimum_direct_rtt
8594 LOC: Config.minDirectRtt
8596 If using the ICMP pinging stuff, do direct fetches for sites
8597 which are no more than this many rtt milliseconds away.
8603 LOC: Config.Netdb.low
8605 The low water mark for the ICMP measurement database.
8607 Note: high watermark controlled by netdb_high directive.
8609 These watermarks are counts, not percents. The defaults are
8610 (low) 900 and (high) 1000. When the high water mark is
8611 reached, database entries will be deleted until the low
8618 LOC: Config.Netdb.high
8620 The high water mark for the ICMP measurement database.
8622 Note: low watermark controlled by netdb_low directive.
8624 These watermarks are counts, not percents. The defaults are
8625 (low) 900 and (high) 1000. When the high water mark is
8626 reached, database entries will be deleted until the low
8630 NAME: netdb_ping_period
8632 LOC: Config.Netdb.period
8635 The minimum period for measuring a site. There will be at
8636 least this much delay between successive pings to the same
8637 network. The default is five minutes.
8644 LOC: Config.onoff.query_icmp
8646 If you want to ask your peers to include ICMP data in their ICP
8647 replies, enable this option.
8649 If your peer has configured Squid (during compilation) with
8650 '--enable-icmp' that peer will send ICMP pings to origin server
8651 sites of the URLs it receives. If you enable this option the
8652 ICP replies from that peer will include the ICMP data (if available).
8653 Then, when choosing a parent cache, Squid will choose the parent with
8654 the minimal RTT to the origin server. When this happens, the
8655 hierarchy field of the access.log will be
8656 "CLOSEST_PARENT_MISS". This option is off by default.
8659 NAME: test_reachability
8663 LOC: Config.onoff.test_reachability
8665 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
8666 instead of ICP_MISS if the target host is NOT in the ICMP
8667 database, or has a zero RTT.
8670 NAME: icp_query_timeout
8673 DEFAULT_DOC: Dynamic detection.
8675 LOC: Config.Timeout.icp_query
8677 Normally Squid will automatically determine an optimal ICP
8678 query timeout value based on the round-trip-time of recent ICP
8679 queries. If you want to override the value determined by
8680 Squid, set this 'icp_query_timeout' to a non-zero value. This
8681 value is specified in MILLISECONDS, so, to use a 2-second
8682 timeout (the old default), you would write:
8684 icp_query_timeout 2000
8687 NAME: maximum_icp_query_timeout
8691 LOC: Config.Timeout.icp_query_max
8693 Normally the ICP query timeout is determined dynamically. But
8694 sometimes it can lead to very large values (say 5 seconds).
8695 Use this option to put an upper limit on the dynamic timeout
8696 value. Do NOT use this option to always use a fixed (instead
8697 of a dynamic) timeout value. To set a fixed timeout see the
8698 'icp_query_timeout' directive.
8701 NAME: minimum_icp_query_timeout
8705 LOC: Config.Timeout.icp_query_min
8707 Normally the ICP query timeout is determined dynamically. But
8708 sometimes it can lead to very small timeouts, even lower than
8709 the normal latency variance on your link due to traffic.
8710 Use this option to put an lower limit on the dynamic timeout
8711 value. Do NOT use this option to always use a fixed (instead
8712 of a dynamic) timeout value. To set a fixed timeout see the
8713 'icp_query_timeout' directive.
8716 NAME: background_ping_rate
8720 LOC: Config.backgroundPingRate
8722 Controls how often the ICP pings are sent to siblings that
8723 have background-ping set.
8727 MULTICAST ICP OPTIONS
8728 -----------------------------------------------------------------------------
8733 LOC: Config.mcast_group_list
8736 This tag specifies a list of multicast groups which your server
8737 should join to receive multicasted ICP queries.
8739 NOTE! Be very careful what you put here! Be sure you
8740 understand the difference between an ICP _query_ and an ICP
8741 _reply_. This option is to be set only if you want to RECEIVE
8742 multicast queries. Do NOT set this option to SEND multicast
8743 ICP (use cache_peer for that). ICP replies are always sent via
8744 unicast, so this option does not affect whether or not you will
8745 receive replies from multicast group members.
8747 You must be very careful to NOT use a multicast address which
8748 is already in use by another group of caches.
8750 If you are unsure about multicast, please read the Multicast
8751 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
8753 Usage: mcast_groups 239.128.16.128 224.0.1.20
8755 By default, Squid doesn't listen on any multicast groups.
8758 NAME: mcast_icp_query_timeout
8762 LOC: Config.Timeout.mcast_icp_query
8764 For multicast peers, Squid regularly sends out ICP "probes" to
8765 count how many other peers are listening on the given multicast
8766 address. This value specifies how long Squid should wait to
8767 count all the replies. The default is 2000 msec, or 2
8772 INTERNAL ICON OPTIONS
8773 -----------------------------------------------------------------------------
8776 NAME: icon_directory
8778 LOC: Config.icons.directory
8779 DEFAULT: @DEFAULT_ICON_DIR@
8781 Where the icons are stored. These are normally kept in
8785 NAME: global_internal_static
8787 LOC: Config.onoff.global_internal_static
8790 This directive controls is Squid should intercept all requests for
8791 /squid-internal-static/ no matter which host the URL is requesting
8792 (default on setting), or if nothing special should be done for
8793 such URLs (off setting). The purpose of this directive is to make
8794 icons etc work better in complex cache hierarchies where it may
8795 not always be possible for all corners in the cache mesh to reach
8796 the server generating a directory listing.
8799 NAME: short_icon_urls
8801 LOC: Config.icons.use_short_names
8804 If this is enabled Squid will use short URLs for icons.
8805 If disabled it will revert to the old behavior of including
8806 it's own name and port in the URL.
8808 If you run a complex cache hierarchy with a mix of Squid and
8809 other proxies you may need to disable this directive.
8814 -----------------------------------------------------------------------------
8817 NAME: error_directory
8819 LOC: Config.errorDirectory
8821 DEFAULT_DOC: Send error pages in the clients preferred language
8823 If you wish to create your own versions of the default
8824 error files to customize them to suit your company copy
8825 the error/template files to another directory and point
8828 WARNING: This option will disable multi-language support
8829 on error pages if used.
8831 The squid developers are interested in making squid available in
8832 a wide variety of languages. If you are making translations for a
8833 language that Squid does not currently provide please consider
8834 contributing your translation back to the project.
8835 https://wiki.squid-cache.org/Translations
8837 The squid developers working on translations are happy to supply drop-in
8838 translated error files in exchange for any new language contributions.
8841 NAME: error_default_language
8842 IFDEF: USE_ERR_LOCALES
8844 LOC: Config.errorDefaultLanguage
8846 DEFAULT_DOC: Generate English language pages.
8848 Set the default language which squid will send error pages in
8849 if no existing translation matches the clients language
8852 If unset (default) generic English will be used.
8854 The squid developers are interested in making squid available in
8855 a wide variety of languages. If you are interested in making
8856 translations for any language see the squid wiki for details.
8857 https://wiki.squid-cache.org/Translations
8860 NAME: error_log_languages
8861 IFDEF: USE_ERR_LOCALES
8863 LOC: Config.errorLogMissingLanguages
8866 Log to cache.log what languages users are attempting to
8867 auto-negotiate for translations.
8869 Successful negotiations are not logged. Only failures
8870 have meaning to indicate that Squid may need an upgrade
8871 of its error page translations.
8874 NAME: err_page_stylesheet
8876 LOC: Config.errorStylesheet
8877 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
8879 CSS Stylesheet to pattern the display of Squid default error pages.
8881 For information on CSS see http://www.w3.org/Style/CSS/
8886 LOC: Config.errHtmlText
8889 HTML text to include in error messages. Make this a "mailto"
8890 URL to your admin address, or maybe just a link to your
8891 organizations Web page.
8893 To include this in your error messages, you must rewrite
8894 the error template files (found in the "errors" directory).
8895 Wherever you want the 'err_html_text' line to appear,
8896 insert a %L tag in the error template file.
8899 NAME: email_err_data
8902 LOC: Config.onoff.emailErrData
8905 If enabled, information about the occurred error will be
8906 included in the mailto links of the ERR pages (if %W is set)
8907 so that the email body contains the data.
8908 Syntax is <A HREF="mailto:%w%W">%w</A>
8913 LOC: Config.denyInfoList
8916 Usage: deny_info err_page_name acl
8917 or deny_info http://... acl
8918 or deny_info TCP_RESET acl
8920 This can be used to return a ERR_ page for requests which
8921 do not pass the 'http_access' rules. Squid remembers the last
8922 acl it evaluated in http_access, and if a 'deny_info' line exists
8923 for that ACL Squid returns a corresponding error page.
8925 The acl is typically the last acl on the http_access deny line which
8926 denied access. The exceptions to this rule are:
8927 - When Squid needs to request authentication credentials. It's then
8928 the first authentication related acl encountered
8929 - When none of the http_access lines matches. It's then the last
8930 acl processed on the last http_access line.
8931 - When the decision to deny access was made by an adaptation service,
8932 the acl name is the corresponding eCAP or ICAP service_name.
8934 NP: If providing your own custom error pages with error_directory
8935 you may also specify them by your custom file name:
8936 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
8938 By default Squid will send "403 Forbidden". A different 4xx or 5xx
8939 may be specified by prefixing the file name with the code and a colon.
8940 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
8942 Alternatively you can tell Squid to reset the TCP connection
8943 by specifying TCP_RESET.
8945 Or you can specify an error URL or URL pattern. The browsers will
8946 get redirected to the specified URL after formatting tags have
8947 been replaced. Redirect will be done with 302 or 307 according to
8948 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
8949 the URL. e.g. 303:http://example.com/
8952 %a - username (if available. Password NOT included)
8953 %A - Local listening IP address the client connection was connected to
8956 %E - Error description
8958 %H - Request domain name
8959 %i - Client IP Address
8961 %O - Unescaped message result from external ACL helper
8962 %o - Message result from external ACL helper
8963 %p - Request Port number
8964 %P - Request Protocol name
8965 %R - Request URL path
8966 %T - Timestamp in RFC 1123 format
8967 %U - Full canonical URL from client
8968 (HTTPS URLs terminate with *)
8969 %u - Full canonical URL from client
8970 %w - Admin email from squid.conf
8972 %% - Literal percent (%) code
8977 OPTIONS INFLUENCING REQUEST FORWARDING
8978 -----------------------------------------------------------------------------
8981 NAME: nonhierarchical_direct
8983 LOC: Config.onoff.nonhierarchical_direct
8986 By default, Squid will send any non-hierarchical requests
8987 (not cacheable request type) direct to origin servers.
8989 When this is set to "off", Squid will prefer to send these
8990 requests to parents.
8992 Note that in most configurations, by turning this off you will only
8993 add latency to these request without any improvement in global hit
8996 This option only sets a preference. If the parent is unavailable a
8997 direct connection to the origin server may still be attempted. To
8998 completely prevent direct connections use never_direct.
9003 LOC: Config.onoff.prefer_direct
9006 Normally Squid tries to use parents for most requests. If you for some
9007 reason like it to first try going direct and only use a parent if
9008 going direct fails set this to on.
9010 By combining nonhierarchical_direct off and prefer_direct on you
9011 can set up Squid to use a parent as a backup path if going direct
9014 Note: If you want Squid to use parents for all requests see
9015 the never_direct directive. prefer_direct only modifies how Squid
9016 acts on cacheable requests.
9019 NAME: cache_miss_revalidate
9023 LOC: Config.onoff.cache_miss_revalidate
9025 RFC 7232 defines a conditional request mechanism to prevent
9026 response objects being unnecessarily transferred over the network.
9027 If that mechanism is used by the client and a cache MISS occurs
9028 it can prevent new cache entries being created.
9030 This option determines whether Squid on cache MISS will pass the
9031 client revalidation request to the server or tries to fetch new
9032 content for caching. It can be useful while the cache is mostly
9033 empty to more quickly have the cache populated by generating
9034 non-conditional GETs.
9036 When set to 'on' (default), Squid will pass all client If-* headers
9037 to the server. This permits server responses without a cacheable
9038 payload to be delivered and on MISS no new cache entry is created.
9040 When set to 'off' and if the request is cacheable, Squid will
9041 remove the clients If-Modified-Since and If-None-Match headers from
9042 the request sent to the server. This requests a 200 status response
9043 from the server to create a new cache entry with.
9048 LOC: Config.accessList.AlwaysDirect
9050 DEFAULT_DOC: Prevent any cache_peer being used for this request.
9052 Usage: always_direct allow|deny [!]aclname ...
9054 Here you can use ACL elements to specify requests which should
9055 ALWAYS be forwarded by Squid to the origin servers without using
9056 any peers. For example, to always directly forward requests for
9057 local servers ignoring any parents or siblings you may have use
9060 acl local-servers dstdomain my.domain.net
9061 always_direct allow local-servers
9063 To always forward FTP requests directly, use
9066 always_direct allow FTP
9068 NOTE: There is a similar, but opposite option named
9069 'never_direct'. You need to be aware that "always_direct deny
9070 foo" is NOT the same thing as "never_direct allow foo". You
9071 may need to use a deny rule to exclude a more-specific case of
9072 some other rule. Example:
9074 acl local-external dstdomain external.foo.net
9075 acl local-servers dstdomain .foo.net
9076 always_direct deny local-external
9077 always_direct allow local-servers
9079 NOTE: If your goal is to make the client forward the request
9080 directly to the origin server bypassing Squid then this needs
9081 to be done in the client configuration. Squid configuration
9082 can only tell Squid how Squid should fetch the object.
9084 NOTE: This directive is not related to caching. The replies
9085 is cached as usual even if you use always_direct. To not cache
9086 the replies see the 'cache' directive.
9088 This clause supports both fast and slow acl types.
9089 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
9094 LOC: Config.accessList.NeverDirect
9096 DEFAULT_DOC: Allow DNS results to be used for this request.
9098 Usage: never_direct allow|deny [!]aclname ...
9100 never_direct is the opposite of always_direct. Please read
9101 the description for always_direct if you have not already.
9103 With 'never_direct' you can use ACL elements to specify
9104 requests which should NEVER be forwarded directly to origin
9105 servers. For example, to force the use of a proxy for all
9106 requests, except those in your local domain use something like:
9108 acl local-servers dstdomain .foo.net
9109 never_direct deny local-servers
9110 never_direct allow all
9112 or if Squid is inside a firewall and there are local intranet
9113 servers inside the firewall use something like:
9115 acl local-intranet dstdomain .foo.net
9116 acl local-external dstdomain external.foo.net
9117 always_direct deny local-external
9118 always_direct allow local-intranet
9119 never_direct allow all
9121 This clause supports both fast and slow acl types.
9122 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
9126 ADVANCED NETWORKING OPTIONS
9127 -----------------------------------------------------------------------------
9130 NAME: incoming_udp_average incoming_icp_average
9133 LOC: Config.comm_incoming.udp.average
9135 Heavy voodoo here. I can't even believe you are reading this.
9136 Are you crazy? Don't even think about adjusting these unless
9137 you understand the algorithms in comm_select.c first!
9140 NAME: incoming_tcp_average incoming_http_average
9143 LOC: Config.comm_incoming.tcp.average
9145 Heavy voodoo here. I can't even believe you are reading this.
9146 Are you crazy? Don't even think about adjusting these unless
9147 you understand the algorithms in comm_select.c first!
9150 NAME: incoming_dns_average
9153 LOC: Config.comm_incoming.dns.average
9155 Heavy voodoo here. I can't even believe you are reading this.
9156 Are you crazy? Don't even think about adjusting these unless
9157 you understand the algorithms in comm_select.c first!
9160 NAME: min_udp_poll_cnt min_icp_poll_cnt
9163 LOC: Config.comm_incoming.udp.min_poll
9165 Heavy voodoo here. I can't even believe you are reading this.
9166 Are you crazy? Don't even think about adjusting these unless
9167 you understand the algorithms in comm_select.c first!
9170 NAME: min_dns_poll_cnt
9173 LOC: Config.comm_incoming.dns.min_poll
9175 Heavy voodoo here. I can't even believe you are reading this.
9176 Are you crazy? Don't even think about adjusting these unless
9177 you understand the algorithms in comm_select.c first!
9180 NAME: min_tcp_poll_cnt min_http_poll_cnt
9183 LOC: Config.comm_incoming.tcp.min_poll
9185 Heavy voodoo here. I can't even believe you are reading this.
9186 Are you crazy? Don't even think about adjusting these unless
9187 you understand the algorithms in comm_select.c first!
9193 LOC: Config.accept_filter
9197 The name of an accept(2) filter to install on Squid's
9198 listen socket(s). This feature is perhaps specific to
9199 FreeBSD and requires support in the kernel.
9201 The 'httpready' filter delays delivering new connections
9202 to Squid until a full HTTP request has been received.
9203 See the accf_http(9) man page for details.
9205 The 'dataready' filter delays delivering new connections
9206 to Squid until there is some data to process.
9207 See the accf_dataready(9) man page for details.
9211 The 'data' filter delays delivering of new connections
9212 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
9213 You may optionally specify a number of seconds to wait by
9214 'data=N' where N is the number of seconds. Defaults to 30
9215 if not specified. See the tcp(7) man page for details.
9218 accept_filter httpready
9223 NAME: client_ip_max_connections
9225 LOC: Config.client_ip_max_connections
9227 DEFAULT_DOC: No limit.
9229 Set an absolute limit on the number of connections a single
9230 client IP can use. Any more than this and Squid will begin to drop
9231 new connections from the client until it closes some links.
9233 Note that this is a global limit. It affects all HTTP, HTCP, and FTP
9234 connections from the client. For finer control use the ACL access controls.
9236 Requires client_db to be enabled (the default).
9238 WARNING: This may noticeably slow down traffic received via external proxies
9239 or NAT devices and cause them to rebound error messages back to their clients.
9242 NAME: tcp_recv_bufsize
9246 DEFAULT_DOC: Use operating system TCP defaults.
9247 LOC: Config.tcpRcvBufsz
9249 Size of receive buffer to set for TCP sockets. Probably just
9250 as easy to change your kernel's default.
9251 Omit from squid.conf to use the default buffer size.
9256 -----------------------------------------------------------------------------
9263 LOC: Adaptation::Icap::TheConfig.onoff
9266 If you want to enable the ICAP module support, set this to on.
9269 NAME: icap_connect_timeout
9272 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
9275 This parameter specifies how long to wait for the TCP connect to
9276 the requested ICAP server to complete before giving up and either
9277 terminating the HTTP transaction or bypassing the failure.
9279 The default for optional services is peer_connect_timeout.
9280 The default for essential services is connect_timeout.
9281 If this option is explicitly set, its value applies to all services.
9284 NAME: icap_io_timeout
9288 DEFAULT_DOC: Use read_timeout.
9289 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
9292 This parameter specifies how long to wait for an I/O activity on
9293 an established, active ICAP connection before giving up and
9294 either terminating the HTTP transaction or bypassing the
9298 NAME: icap_service_failure_limit
9299 COMMENT: limit [in memory-depth time-units]
9300 TYPE: icap_service_failure_limit
9302 LOC: Adaptation::Icap::TheConfig
9305 The limit specifies the number of failures that Squid tolerates
9306 when establishing a new TCP connection with an ICAP service. If
9307 the number of failures exceeds the limit, the ICAP service is
9308 not used for new ICAP requests until it is time to refresh its
9311 A negative value disables the limit. Without the limit, an ICAP
9312 service will not be considered down due to connectivity failures
9313 between ICAP OPTIONS requests.
9315 Squid forgets ICAP service failures older than the specified
9316 value of memory-depth. The memory fading algorithm
9317 is approximate because Squid does not remember individual
9318 errors but groups them instead, splitting the option
9319 value into ten time slots of equal length.
9321 When memory-depth is 0 and by default this option has no
9322 effect on service failure expiration.
9324 Squid always forgets failures when updating service settings
9325 using an ICAP OPTIONS transaction, regardless of this option
9329 # suspend service usage after 10 failures in 5 seconds:
9330 icap_service_failure_limit 10 in 5 seconds
9333 NAME: icap_service_revival_delay
9336 LOC: Adaptation::Icap::TheConfig.service_revival_delay
9339 The delay specifies the number of seconds to wait after an ICAP
9340 OPTIONS request failure before requesting the options again. The
9341 failed ICAP service is considered "down" until fresh OPTIONS are
9344 The actual delay cannot be smaller than the hardcoded minimum
9345 delay of 30 seconds.
9348 NAME: icap_preview_enable
9352 LOC: Adaptation::Icap::TheConfig.preview_enable
9355 The ICAP Preview feature allows the ICAP server to handle the
9356 HTTP message by looking only at the beginning of the message body
9357 or even without receiving the body at all. In some environments,
9358 previews greatly speedup ICAP processing.
9360 During an ICAP OPTIONS transaction, the server may tell Squid what
9361 HTTP messages should be previewed and how big the preview should be.
9362 Squid will not use Preview if the server did not request one.
9364 To disable ICAP Preview for all ICAP services, regardless of
9365 individual ICAP server OPTIONS responses, set this option to "off".
9367 icap_preview_enable off
9370 NAME: icap_preview_size
9373 LOC: Adaptation::Icap::TheConfig.preview_size
9375 DEFAULT_DOC: No preview sent.
9377 The default size of preview data to be sent to the ICAP server.
9378 This value might be overwritten on a per server basis by OPTIONS requests.
9381 NAME: icap_206_enable
9385 LOC: Adaptation::Icap::TheConfig.allow206_enable
9388 206 (Partial Content) responses is an ICAP extension that allows the
9389 ICAP agents to optionally combine adapted and original HTTP message
9390 content. The decision to combine is postponed until the end of the
9391 ICAP response. Squid supports Partial Content extension by default.
9393 Activation of the Partial Content extension is negotiated with each
9394 ICAP service during OPTIONS exchange. Most ICAP servers should handle
9395 negotiation correctly even if they do not support the extension, but
9396 some might fail. To disable Partial Content support for all ICAP
9397 services and to avoid any negotiation, set this option to "off".
9403 NAME: icap_default_options_ttl
9406 LOC: Adaptation::Icap::TheConfig.default_options_ttl
9409 The default TTL value for ICAP OPTIONS responses that don't have
9410 an Options-TTL header.
9413 NAME: icap_persistent_connections
9417 LOC: Adaptation::Icap::TheConfig.reuse_connections
9420 Whether or not Squid should use persistent connections to
9424 NAME: adaptation_send_client_ip icap_send_client_ip
9426 IFDEF: USE_ADAPTATION
9428 LOC: Adaptation::Config::send_client_ip
9431 If enabled, Squid shares HTTP client IP information with adaptation
9432 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
9433 For eCAP, Squid sets the libecap::metaClientIp transaction option.
9435 See also: adaptation_uses_indirect_client
9438 NAME: adaptation_send_username icap_send_client_username
9440 IFDEF: USE_ADAPTATION
9442 LOC: Adaptation::Config::send_username
9445 This sends authenticated HTTP client username (if available) to
9446 the adaptation service.
9448 For ICAP, the username value is encoded based on the
9449 icap_client_username_encode option and is sent using the header
9450 specified by the icap_client_username_header option.
9453 NAME: icap_client_username_header
9456 LOC: Adaptation::Icap::TheConfig.client_username_header
9457 DEFAULT: X-Client-Username
9459 ICAP request header name to use for adaptation_send_username.
9462 NAME: icap_client_username_encode
9466 LOC: Adaptation::Icap::TheConfig.client_username_encode
9469 Whether to base64 encode the authenticated client username.
9473 TYPE: icap_service_type
9475 LOC: Adaptation::Icap::TheConfig
9478 Defines a single ICAP service using the following format:
9480 icap_service id vectoring_point uri [option ...]
9483 an opaque identifier or name which is used to direct traffic to
9484 this specific service. Must be unique among all adaptation
9485 services in squid.conf.
9487 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
9488 This specifies at which point of transaction processing the
9489 ICAP service should be activated. *_postcache vectoring points
9490 are not yet supported.
9492 uri: icap://servername:port/servicepath
9493 ICAP server and service location.
9494 icaps://servername:port/servicepath
9495 The "icap:" URI scheme is used for traditional ICAP server and
9496 service location (default port is 1344, connections are not
9497 encrypted). The "icaps:" URI scheme is for Secure ICAP
9498 services that use SSL/TLS-encrypted ICAP connections (by
9499 default, on port 11344).
9501 ICAP does not allow a single service to handle both REQMOD and RESPMOD
9502 transactions. Squid does not enforce that requirement. You can specify
9503 services with the same service_url and different vectoring_points. You
9504 can even specify multiple identical services as long as their
9505 service_names differ.
9507 To activate a service, use the adaptation_access directive. To group
9508 services, use adaptation_service_chain and adaptation_service_set.
9510 Service options are separated by white space. ICAP services support
9511 the following name=value options:
9514 If set to 'on' or '1', the ICAP service is treated as
9515 optional. If the service cannot be reached or malfunctions,
9516 Squid will try to ignore any errors and process the message as
9517 if the service was not enabled. No all ICAP errors can be
9518 bypassed. If set to 0, the ICAP service is treated as
9519 essential and all ICAP errors will result in an error page
9520 returned to the HTTP client.
9522 Bypass is off by default: services are treated as essential.
9525 If set to 'on' or '1', the ICAP service is allowed to
9526 dynamically change the current message adaptation plan by
9527 returning a chain of services to be used next. The services
9528 are specified using the X-Next-Services ICAP response header
9529 value, formatted as a comma-separated list of service names.
9530 Each named service should be configured in squid.conf. Other
9531 services are ignored. An empty X-Next-Services value results
9532 in an empty plan which ends the current adaptation.
9534 Dynamic adaptation plan may cross or cover multiple supported
9535 vectoring points in their natural processing order.
9537 Routing is not allowed by default: the ICAP X-Next-Services
9538 response header is ignored.
9541 Only has effect on split-stack systems. The default on those systems
9542 is to use IPv4-only connections. When set to 'on' this option will
9543 make Squid use IPv6-only connections to contact this ICAP service.
9545 on-overload=block|bypass|wait|force
9546 If the service Max-Connections limit has been reached, do
9547 one of the following for each new ICAP transaction:
9548 * block: send an HTTP error response to the client
9549 * bypass: ignore the "over-connected" ICAP service
9550 * wait: wait (in a FIFO queue) for an ICAP connection slot
9551 * force: proceed, ignoring the Max-Connections limit
9553 In SMP mode with N workers, each worker assumes the service
9554 connection limit is Max-Connections/N, even though not all
9555 workers may use a given service.
9557 The default value is "bypass" if service is bypassable,
9558 otherwise it is set to "wait".
9562 Use the given number as the Max-Connections limit, regardless
9563 of the Max-Connections value given by the service, if any.
9565 connection-encryption=on|off
9566 Determines the ICAP service effect on the connections_encrypted
9569 The default is "on" for Secure ICAP services (i.e., those
9570 with the icaps:// service URIs scheme) and "off" for plain ICAP
9573 Does not affect ICAP connections (e.g., does not turn Secure
9576 ==== ICAPS / TLS OPTIONS ====
9578 These options are used for Secure ICAP (icaps://....) services only.
9580 tls-cert=/path/to/ssl/certificate
9581 A client X.509 certificate to use when connecting to
9584 tls-key=/path/to/ssl/key
9585 The private key corresponding to the previous
9588 If tls-key= is not specified tls-cert= is assumed to
9589 reference a PEM file containing both the certificate
9592 tls-cipher=... The list of valid TLS/SSL ciphers to use when connecting
9593 to this icap server.
9596 The minimum TLS protocol version to permit. To control
9597 SSLv3 use the tls-options= parameter.
9598 Supported Values: 1.0 (default), 1.1, 1.2
9600 tls-options=... Specify various OpenSSL library options:
9602 NO_SSLv3 Disallow the use of SSLv3
9605 Always create a new key when using
9606 temporary/ephemeral DH key exchanges
9608 ALL Enable various bug workarounds
9609 suggested as "harmless" by OpenSSL
9610 Be warned that this reduces SSL/TLS
9611 strength to some attacks.
9613 See the OpenSSL SSL_CTX_set_options documentation for a
9614 more complete list. Options relevant only to SSLv2 are
9617 tls-cafile= PEM file containing CA certificates to use when verifying
9618 the icap server certificate.
9619 Use to specify intermediate CA certificate(s) if not sent
9620 by the server. Or the full CA chain for the server when
9621 using the tls-default-ca=off flag.
9622 May be repeated to load multiple files.
9624 tls-capath=... A directory containing additional CA certificates to
9625 use when verifying the icap server certificate.
9626 Requires OpenSSL or LibreSSL.
9628 tls-crlfile=... A certificate revocation list file to use when
9629 verifying the icap server certificate.
9631 tls-flags=... Specify various flags modifying the Squid TLS implementation:
9634 Accept certificates even if they fail to
9637 Don't verify the icap server certificate
9638 matches the server name
9640 tls-default-ca[=off]
9641 Whether to use the system Trusted CAs. Default is ON.
9643 tls-domain= The icap server name as advertised in it's certificate.
9644 Used for verifying the correctness of the received icap
9645 server certificate. If not specified the icap server
9646 hostname extracted from ICAP URI will be used.
9648 Older icap_service format without optional named parameters is
9649 deprecated but supported for backward compatibility.
9652 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
9653 icap_service svcLogger reqmod_precache icaps://icap2.mydomain.net:11344/reqmod routing=on
9657 TYPE: icap_class_type
9662 This deprecated option was documented to define an ICAP service
9663 chain, even though it actually defined a set of similar, redundant
9664 services, and the chains were not supported.
9666 To define a set of redundant services, please use the
9667 adaptation_service_set directive. For service chains, use
9668 adaptation_service_chain.
9672 TYPE: icap_access_type
9677 This option is deprecated. Please use adaptation_access, which
9678 has the same ICAP functionality, but comes with better
9679 documentation, and eCAP support.
9684 -----------------------------------------------------------------------------
9691 LOC: Adaptation::Ecap::TheConfig.onoff
9694 Controls whether eCAP support is enabled.
9698 TYPE: ecap_service_type
9700 LOC: Adaptation::Ecap::TheConfig
9703 Defines a single eCAP service
9705 ecap_service id vectoring_point uri [option ...]
9708 an opaque identifier or name which is used to direct traffic to
9709 this specific service. Must be unique among all adaptation
9710 services in squid.conf.
9712 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
9713 This specifies at which point of transaction processing the
9714 eCAP service should be activated. *_postcache vectoring points
9715 are not yet supported.
9717 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
9718 Squid uses the eCAP service URI to match this configuration
9719 line with one of the dynamically loaded services. Each loaded
9720 eCAP service must have a unique URI. Obtain the right URI from
9721 the service provider.
9723 To activate a service, use the adaptation_access directive. To group
9724 services, use adaptation_service_chain and adaptation_service_set.
9726 Service options are separated by white space. eCAP services support
9727 the following name=value options:
9730 If set to 'on' or '1', the eCAP service is treated as optional.
9731 If the service cannot be reached or malfunctions, Squid will try
9732 to ignore any errors and process the message as if the service
9733 was not enabled. No all eCAP errors can be bypassed.
9734 If set to 'off' or '0', the eCAP service is treated as essential
9735 and all eCAP errors will result in an error page returned to the
9738 Bypass is off by default: services are treated as essential.
9741 If set to 'on' or '1', the eCAP service is allowed to
9742 dynamically change the current message adaptation plan by
9743 returning a chain of services to be used next.
9745 Dynamic adaptation plan may cross or cover multiple supported
9746 vectoring points in their natural processing order.
9748 Routing is not allowed by default.
9750 connection-encryption=on|off
9751 Determines the eCAP service effect on the connections_encrypted
9754 Defaults to "on", which does not taint the master transaction
9757 Does not affect eCAP API calls.
9759 Older ecap_service format without optional named parameters is
9760 deprecated but supported for backward compatibility.
9764 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
9765 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
9768 NAME: loadable_modules
9770 IFDEF: USE_LOADABLE_MODULES
9771 LOC: Config.loadable_module_names
9774 Instructs Squid to load the specified dynamic module(s) or activate
9775 preloaded module(s).
9777 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
9781 MESSAGE ADAPTATION OPTIONS
9782 -----------------------------------------------------------------------------
9785 NAME: adaptation_service_set
9786 TYPE: adaptation_service_set_type
9787 IFDEF: USE_ADAPTATION
9792 Configures an ordered set of similar, redundant services. This is
9793 useful when hot standby or backup adaptation servers are available.
9795 adaptation_service_set set_name service_name1 service_name2 ...
9797 The named services are used in the set declaration order. The first
9798 applicable adaptation service from the set is used first. The next
9799 applicable service is tried if and only if the transaction with the
9800 previous service fails and the message waiting to be adapted is still
9803 When adaptation starts, broken services are ignored as if they were
9804 not a part of the set. A broken service is a down optional service.
9806 The services in a set must be attached to the same vectoring point
9807 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9809 If all services in a set are optional then adaptation failures are
9810 bypassable. If all services in the set are essential, then a
9811 transaction failure with one service may still be retried using
9812 another service from the set, but when all services fail, the master
9813 transaction fails as well.
9815 A set may contain a mix of optional and essential services, but that
9816 is likely to lead to surprising results because broken services become
9817 ignored (see above), making previously bypassable failures fatal.
9818 Technically, it is the bypassability of the last failed service that
9821 See also: adaptation_access adaptation_service_chain
9824 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
9825 adaptation service_set svcLogger loggerLocal loggerRemote
9828 NAME: adaptation_service_chain
9829 TYPE: adaptation_service_chain_type
9830 IFDEF: USE_ADAPTATION
9835 Configures a list of complementary services that will be applied
9836 one-by-one, forming an adaptation chain or pipeline. This is useful
9837 when Squid must perform different adaptations on the same message.
9839 adaptation_service_chain chain_name service_name1 svc_name2 ...
9841 The named services are used in the chain declaration order. The first
9842 applicable adaptation service from the chain is used first. The next
9843 applicable service is applied to the successful adaptation results of
9844 the previous service in the chain.
9846 When adaptation starts, broken services are ignored as if they were
9847 not a part of the chain. A broken service is a down optional service.
9849 Request satisfaction terminates the adaptation chain because Squid
9850 does not currently allow declaration of RESPMOD services at the
9851 "reqmod_precache" vectoring point (see icap_service or ecap_service).
9853 The services in a chain must be attached to the same vectoring point
9854 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9856 A chain may contain a mix of optional and essential services. If an
9857 essential adaptation fails (or the failure cannot be bypassed for
9858 other reasons), the master transaction fails. Otherwise, the failure
9859 is bypassed as if the failed adaptation service was not in the chain.
9861 See also: adaptation_access adaptation_service_set
9864 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
9867 NAME: adaptation_access
9868 TYPE: adaptation_access_type
9869 IFDEF: USE_ADAPTATION
9872 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
9874 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
9876 adaptation_access service_name allow|deny [!]aclname...
9877 adaptation_access set_name allow|deny [!]aclname...
9879 At each supported vectoring point, the adaptation_access
9880 statements are processed in the order they appear in this
9881 configuration file. Statements pointing to the following services
9882 are ignored (i.e., skipped without checking their ACL):
9884 - services serving different vectoring points
9885 - "broken-but-bypassable" services
9886 - "up" services configured to ignore such transactions
9887 (e.g., based on the ICAP Transfer-Ignore header).
9889 When a set_name is used, all services in the set are checked
9890 using the same rules, to find the first applicable one. See
9891 adaptation_service_set for details.
9893 If an access list is checked and there is a match, the
9894 processing stops: For an "allow" rule, the corresponding
9895 adaptation service is used for the transaction. For a "deny"
9896 rule, no adaptation service is activated.
9898 It is currently not possible to apply more than one adaptation
9899 service at the same vectoring point to the same HTTP transaction.
9901 See also: icap_service and ecap_service
9904 adaptation_access service_1 allow all
9907 NAME: adaptation_service_iteration_limit
9909 IFDEF: USE_ADAPTATION
9910 LOC: Adaptation::Config::service_iteration_limit
9913 Limits the number of iterations allowed when applying adaptation
9914 services to a message. If your longest adaptation set or chain
9915 may have more than 16 services, increase the limit beyond its
9916 default value of 16. If detecting infinite iteration loops sooner
9917 is critical, make the iteration limit match the actual number
9918 of services in your longest adaptation set or chain.
9920 Infinite adaptation loops are most likely with routing services.
9922 See also: icap_service routing=1
9925 NAME: adaptation_masterx_shared_names
9927 IFDEF: USE_ADAPTATION
9928 LOC: Adaptation::Config::masterx_shared_name
9931 For each master transaction (i.e., the HTTP request and response
9932 sequence, including all related ICAP and eCAP exchanges), Squid
9933 maintains a table of metadata. The table entries are (name, value)
9934 pairs shared among eCAP and ICAP exchanges. The table is destroyed
9935 with the master transaction.
9937 This option specifies the table entry names that Squid must accept
9938 from and forward to the adaptation transactions.
9940 An ICAP REQMOD or RESPMOD transaction may set an entry in the
9941 shared table by returning an ICAP header field with a name
9942 specified in adaptation_masterx_shared_names.
9944 An eCAP REQMOD or RESPMOD transaction may set an entry in the
9945 shared table by implementing the libecap::visitEachOption() API
9946 to provide an option with a name specified in
9947 adaptation_masterx_shared_names.
9949 Squid will store and forward the set entry to subsequent adaptation
9950 transactions within the same master transaction scope.
9952 Only one shared entry name is supported at this time.
9955 # share authentication information among ICAP services
9956 adaptation_masterx_shared_names X-Subscriber-ID
9959 NAME: adaptation_meta
9961 IFDEF: USE_ADAPTATION
9962 LOC: Adaptation::Config::metaHeaders
9965 This option allows Squid administrator to add custom ICAP request
9966 headers or eCAP options to Squid ICAP requests or eCAP transactions.
9967 Use it to pass custom authentication tokens and other
9968 transaction-state related meta information to an ICAP/eCAP service.
9970 The addition of a meta header is ACL-driven:
9971 adaptation_meta name value [!]aclname ...
9973 Processing for a given header name stops after the first ACL list match.
9974 Thus, it is impossible to add two headers with the same name. If no ACL
9975 lists match for a given header name, no such header is added. For
9978 # do not debug transactions except for those that need debugging
9979 adaptation_meta X-Debug 1 needs_debugging
9981 # log all transactions except for those that must remain secret
9982 adaptation_meta X-Log 1 !keep_secret
9984 # mark transactions from users in the "G 1" group
9985 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
9987 The "value" parameter may be a regular squid.conf token or a "double
9988 quoted string". Within the quoted string, use backslash (\) to escape
9989 any character, which is currently only useful for escaping backslashes
9990 and double quotes. For example,
9991 "this string has one backslash (\\) and two \"quotes\""
9993 Used adaptation_meta header values may be logged via %note
9994 logformat code. If multiple adaptation_meta headers with the same name
9995 are used during master transaction lifetime, the header values are
9996 logged in the order they were used and duplicate values are ignored
9997 (only the first repeated value will be logged).
10003 LOC: Adaptation::Icap::TheConfig.repeat
10004 DEFAULT_IF_NONE: deny all
10006 This ACL determines which retriable ICAP transactions are
10007 retried. Transactions that received a complete ICAP response
10008 and did not have to consume or produce HTTP bodies to receive
10009 that response are usually retriable.
10011 icap_retry allow|deny [!]aclname ...
10013 Squid automatically retries some ICAP I/O timeouts and errors
10014 due to persistent connection race conditions.
10016 See also: icap_retry_limit
10019 NAME: icap_retry_limit
10022 LOC: Adaptation::Icap::TheConfig.repeat_limit
10024 DEFAULT_DOC: No retries are allowed.
10026 Limits the number of retries allowed.
10028 Communication errors due to persistent connection race
10029 conditions are unavoidable, automatically retried, and do not
10030 count against this limit.
10032 See also: icap_retry
10038 -----------------------------------------------------------------------------
10041 NAME: check_hostnames
10044 LOC: Config.onoff.check_hostnames
10046 For security and stability reasons Squid can check
10047 hostnames for Internet standard RFC compliance. If you want
10048 Squid to perform these checks turn this directive on.
10051 NAME: allow_underscore
10054 LOC: Config.onoff.allow_underscore
10056 Underscore characters is not strictly allowed in Internet hostnames
10057 but nevertheless used by many sites. Set this to off if you want
10058 Squid to be strict about the standard.
10059 This check is performed only when check_hostnames is set to on.
10062 NAME: dns_retransmit_interval
10065 LOC: Config.Timeout.idns_retransmit
10067 Initial retransmit interval for DNS queries. The interval is
10068 doubled each time all configured DNS servers have been tried.
10073 DEFAULT: 30 seconds
10074 LOC: Config.Timeout.idns_query
10076 DNS Query timeout. If no response is received to a DNS query
10077 within this time all DNS servers for the queried domain
10078 are assumed to be unavailable.
10081 NAME: dns_packet_max
10083 DEFAULT_DOC: EDNS disabled
10085 LOC: Config.dns.packet_max
10087 Maximum number of bytes packet size to advertise via EDNS.
10088 Set to "none" to disable EDNS large packet support.
10090 For legacy reasons DNS UDP replies will default to 512 bytes which
10091 is too small for many responses. EDNS provides a means for Squid to
10092 negotiate receiving larger responses back immediately without having
10093 to failover with repeat requests. Responses larger than this limit
10094 will retain the old behaviour of failover to TCP DNS.
10096 Squid has no real fixed limit internally, but allowing packet sizes
10097 over 1500 bytes requires network jumbogram support and is usually not
10100 WARNING: The RFC also indicates that some older resolvers will reply
10101 with failure of the whole request if the extension is added. Some
10102 resolvers have already been identified which will reply with mangled
10103 EDNS response on occasion. Usually in response to many-KB jumbogram
10104 sizes being advertised by Squid.
10105 Squid will currently treat these both as an unable-to-resolve domain
10106 even if it would be resolvable without EDNS.
10113 DEFAULT_DOC: Search for single-label domain names is disabled.
10114 LOC: Config.onoff.res_defnames
10116 Normally the RES_DEFNAMES resolver option is disabled
10117 (see res_init(3)). This prevents caches in a hierarchy
10118 from interpreting single-component hostnames locally. To allow
10119 Squid to handle single-component names, enable this option.
10122 NAME: dns_multicast_local
10126 DEFAULT_DOC: Search for .local and .arpa names is disabled.
10127 LOC: Config.onoff.dns_mdns
10129 When set to on, Squid sends multicast DNS lookups on the local
10130 network for domains ending in .local and .arpa.
10131 This enables local servers and devices to be contacted in an
10132 ad-hoc or zero-configuration network environment.
10135 NAME: dns_nameservers
10138 DEFAULT_DOC: Use operating system definitions
10139 LOC: Config.dns.nameservers
10141 Use this if you want to specify a list of DNS name servers
10142 (IP addresses) to use instead of those given in your
10143 /etc/resolv.conf file.
10145 On Windows platforms, if no value is specified here or in
10146 the /etc/resolv.conf file, the list of DNS name servers are
10147 taken from the Windows registry, both static and dynamic DHCP
10148 configurations are supported.
10150 Example: dns_nameservers 10.0.0.1 192.172.0.4
10155 DEFAULT: @DEFAULT_HOSTS@
10156 LOC: Config.etcHostsPath
10158 Location of the host-local IP name-address associations
10159 database. Most Operating Systems have such a file on different
10161 - Un*X & Linux: /etc/hosts
10162 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
10163 (%SystemRoot% value install default is c:\winnt)
10164 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
10165 (%SystemRoot% value install default is c:\windows)
10166 - Windows 9x/Me: %windir%\hosts
10167 (%windir% value is usually c:\windows)
10168 - Cygwin: /etc/hosts
10170 The file contains newline-separated definitions, in the
10171 form ip_address_in_dotted_form name [name ...] names are
10172 whitespace-separated. Lines beginning with an hash (#)
10173 character are comments.
10175 The file is checked at startup and upon configuration.
10176 If set to 'none', it won't be checked.
10177 If append_domain is used, that domain will be added to
10178 domain-local (i.e. not containing any dot character) host
10182 NAME: append_domain
10184 LOC: Config.appendDomain
10186 DEFAULT_DOC: Use operating system definitions
10188 Appends local domain name to hostnames without any dots in
10189 them. append_domain must begin with a period.
10191 Be warned there are now Internet names with no dots in
10192 them using only top-domain names, so setting this may
10193 cause some Internet sites to become unavailable.
10196 append_domain .yourdomain.com
10199 NAME: ignore_unknown_nameservers
10201 LOC: Config.onoff.ignore_unknown_nameservers
10204 By default Squid checks that DNS responses are received
10205 from the same IP addresses they are sent to. If they
10206 don't match, Squid ignores the response and writes a warning
10207 message to cache.log. You can allow responses from unknown
10208 nameservers by setting this option to 'off'.
10212 COMMENT: (number of entries)
10215 LOC: Config.ipcache.size
10217 Maximum number of DNS IP cache entries.
10224 LOC: Config.ipcache.low
10231 LOC: Config.ipcache.high
10233 The size, low-, and high-water marks for the IP cache.
10236 NAME: fqdncache_size
10237 COMMENT: (number of entries)
10240 LOC: Config.fqdncache.size
10242 Maximum number of FQDN cache entries.
10247 -----------------------------------------------------------------------------
10250 NAME: configuration_includes_quoted_values
10252 TYPE: configuration_includes_quoted_values
10254 LOC: ConfigParser::RecognizeQuotedValues
10256 If set, Squid will recognize each "quoted string" after a configuration
10257 directive as a single parameter. The quotes are stripped before the
10258 parameter value is interpreted or used.
10259 See "Values with spaces, quotes, and other special characters"
10260 section for more details.
10267 LOC: Config.onoff.mem_pools
10269 If set, Squid will keep pools of allocated (but unused) memory
10270 available for future use. If memory is a premium on your
10271 system and you believe your malloc library outperforms Squid
10272 routines, disable this.
10275 NAME: memory_pools_limit
10279 LOC: Config.MemPools.limit
10281 Used only with memory_pools on:
10282 memory_pools_limit 50 MB
10284 If set to a non-zero value, Squid will keep at most the specified
10285 limit of allocated (but unused) memory in memory pools. All free()
10286 requests that exceed this limit will be handled by your malloc
10287 library. Squid does not pre-allocate any memory, just safe-keeps
10288 objects that otherwise would be free()d. Thus, it is safe to set
10289 memory_pools_limit to a reasonably high value even if your
10290 configuration will use less memory.
10292 If set to none, Squid will keep all memory it can. That is, there
10293 will be no limit on the total amount of memory used for safe-keeping.
10295 To disable memory allocation optimization, do not set
10296 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
10298 An overhead for maintaining memory pools is not taken into account
10299 when the limit is checked. This overhead is close to four bytes per
10300 object kept. However, pools may actually _save_ memory because of
10301 reduced memory thrashing in your malloc library.
10304 NAME: forwarded_for
10305 COMMENT: on|off|transparent|truncate|delete
10308 LOC: opt_forwarded_for
10310 If set to "on", Squid will append your client's IP address
10311 in the HTTP requests it forwards. By default it looks like:
10313 X-Forwarded-For: 192.1.2.3
10315 If set to "off", it will appear as
10317 X-Forwarded-For: unknown
10319 If set to "transparent", Squid will not alter the
10320 X-Forwarded-For header in any way.
10322 If set to "delete", Squid will delete the entire
10323 X-Forwarded-For header.
10325 If set to "truncate", Squid will remove all existing
10326 X-Forwarded-For entries, and place the client IP as the sole entry.
10329 NAME: cachemgr_passwd
10330 TYPE: cachemgrpasswd
10332 DEFAULT_DOC: No password. Actions which require password are denied.
10333 LOC: Config.passwd_list
10335 Specify passwords for cachemgr operations.
10337 Usage: cachemgr_passwd password action action ...
10339 Some valid actions are (see cache manager menu for a full list):
10378 * Indicates actions which will not be performed without a
10379 valid password, others can be performed if not listed here.
10381 To disable an action, set the password to "disable".
10382 To allow performing an action without a password, set the
10383 password to "none".
10385 Use the keyword "all" to set the same password for all actions.
10388 cachemgr_passwd secret shutdown
10389 cachemgr_passwd lesssssssecret info stats/objects
10390 cachemgr_passwd disable all
10397 LOC: Config.onoff.client_db
10399 If you want to disable collecting per-client statistics,
10400 turn off client_db here.
10403 NAME: refresh_all_ims
10407 LOC: Config.onoff.refresh_all_ims
10409 When you enable this option, squid will always check
10410 the origin server for an update when a client sends an
10411 If-Modified-Since request. Many browsers use IMS
10412 requests when the user requests a reload, and this
10413 ensures those clients receive the latest version.
10415 By default (off), squid may return a Not Modified response
10416 based on the age of the cached version.
10419 NAME: reload_into_ims
10420 IFDEF: USE_HTTP_VIOLATIONS
10424 LOC: Config.onoff.reload_into_ims
10426 When you enable this option, client no-cache or ``reload''
10427 requests will be changed to If-Modified-Since requests.
10428 Doing this VIOLATES the HTTP standard. Enabling this
10429 feature could make you liable for problems which it
10432 see also refresh_pattern for a more selective approach.
10435 NAME: connect_retries
10437 LOC: Config.connect_retries
10439 DEFAULT_DOC: Do not retry failed connections.
10441 Limits the number of reopening attempts when establishing a single
10442 TCP connection. All these attempts must still complete before the
10443 applicable connection opening timeout expires.
10445 By default and when connect_retries is set to zero, Squid does not
10446 retry failed connection opening attempts.
10448 The (not recommended) maximum is 10 tries. An attempt to configure a
10449 higher value results in the value of 10 being used (with a warning).
10451 Squid may open connections to retry various high-level forwarding
10452 failures. For an outside observer, that activity may look like a
10453 low-level connection reopening attempt, but those high-level retries
10454 are governed by forward_max_tries instead.
10456 See also: connect_timeout, forward_timeout, icap_connect_timeout,
10457 ident_timeout, and forward_max_tries.
10460 NAME: retry_on_error
10462 LOC: Config.retry.onerror
10465 If set to ON Squid will automatically retry requests when
10466 receiving an error response with status 403 (Forbidden),
10467 500 (Internal Error), 501 or 503 (Service not available).
10468 Status 502 and 504 (Gateway errors) are always retried.
10470 This is mainly useful if you are in a complex cache hierarchy to
10471 work around access control errors.
10473 NOTE: This retry will attempt to find another working destination.
10474 Which is different from the server which just failed.
10477 NAME: as_whois_server
10479 LOC: Config.as_whois_server
10480 DEFAULT: whois.ra.net
10482 WHOIS server to query for AS numbers. NOTE: AS numbers are
10483 queried only when Squid starts up, not for every request.
10488 LOC: Config.onoff.offline
10491 Enable this option and Squid will never try to validate cached
10495 NAME: uri_whitespace
10496 TYPE: uri_whitespace
10497 LOC: Config.uri_whitespace
10500 What to do with requests that have whitespace characters in the
10503 strip: The whitespace characters are stripped out of the URL.
10504 This is the behavior recommended by RFC2396 and RFC3986
10505 for tolerant handling of generic URI.
10506 NOTE: This is one difference between generic URI and HTTP URLs.
10508 deny: The request is denied. The user receives an "Invalid
10510 This is the behaviour recommended by RFC2616 for safe
10511 handling of HTTP request URL.
10513 allow: The request is allowed and the URI is not changed. The
10514 whitespace characters remain in the URI. Note the
10515 whitespace is passed to redirector processes if they
10517 Note this may be considered a violation of RFC2616
10518 request parsing where whitespace is prohibited in the
10521 encode: The request is allowed and the whitespace characters are
10522 encoded according to RFC1738.
10524 chop: The request is allowed and the URI is chopped at the
10528 NOTE the current Squid implementation of encode and chop violates
10529 RFC2616 by not using a 301 redirect after altering the URL.
10534 LOC: Config.chroot_dir
10537 Specifies a directory where Squid should do a chroot() while
10538 initializing. This also causes Squid to fully drop root
10539 privileges after initializing. This means, for example, if you
10540 use a HTTP port less than 1024 and try to reconfigure, you may
10541 get an error saying that Squid can not open the port.
10544 NAME: pipeline_prefetch
10545 TYPE: pipelinePrefetch
10546 LOC: Config.pipeline_max_prefetch
10548 DEFAULT_DOC: Do not pre-parse pipelined requests.
10550 HTTP clients may send a pipeline of 1+N requests to Squid using a
10551 single connection, without waiting for Squid to respond to the first
10552 of those requests. This option limits the number of concurrent
10553 requests Squid will try to handle in parallel. If set to N, Squid
10554 will try to receive and process up to 1+N requests on the same
10555 connection concurrently.
10557 Defaults to 0 (off) for bandwidth management and access logging
10560 NOTE: pipelining requires persistent connections to clients.
10562 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
10565 NAME: high_response_time_warning
10568 LOC: Config.warnings.high_rptm
10570 DEFAULT_DOC: disabled.
10572 If the one-minute median response time exceeds this value,
10573 Squid prints a WARNING with debug level 0 to get the
10574 administrators attention. The value is in milliseconds.
10577 NAME: high_page_fault_warning
10579 LOC: Config.warnings.high_pf
10581 DEFAULT_DOC: disabled.
10583 If the one-minute average page fault rate exceeds this
10584 value, Squid prints a WARNING with debug level 0 to get
10585 the administrators attention. The value is in page faults
10589 NAME: high_memory_warning
10591 LOC: Config.warnings.high_memory
10592 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
10594 DEFAULT_DOC: disabled.
10596 If the memory usage (as determined by gnumalloc, if available and used)
10597 exceeds this amount, Squid prints a WARNING with debug level 0 to get
10598 the administrators attention.
10600 # TODO: link high_memory_warning to mempools?
10602 NAME: sleep_after_fork
10603 COMMENT: (microseconds)
10605 LOC: Config.sleep_after_fork
10608 When this is set to a non-zero value, the main Squid process
10609 sleeps the specified number of microseconds after a fork()
10610 system call. This sleep may help the situation where your
10611 system reports fork() failures due to lack of (virtual)
10612 memory. Note, however, if you have a lot of child
10613 processes, these sleep delays will add up and your
10614 Squid will not service requests for some amount of time
10615 until all the child processes have been started.
10618 NAME: windows_ipaddrchangemonitor
10619 IFDEF: _SQUID_WINDOWS_
10623 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
10625 On Windows Squid by default will monitor IP address changes and will
10626 reconfigure itself after any detected event. This is very useful for
10627 proxies connected to internet with dial-up interfaces.
10628 In some cases (a Proxy server acting as VPN gateway is one) it could be
10629 desiderable to disable this behaviour setting this to 'off'.
10630 Note: after changing this, Squid service must be restarted.
10635 IFDEF: USE_SQUID_EUI
10637 LOC: Eui::TheConfig.euiLookup
10639 Whether to lookup the EUI or MAC address of a connected client.
10642 NAME: max_filedescriptors max_filedesc
10645 DEFAULT_DOC: Use operating system soft limit set by ulimit.
10646 LOC: Config.max_filedescriptors
10648 Set the maximum number of filedescriptors, either below the
10649 operating system default or up to the hard limit.
10651 Remove from squid.conf to inherit the current ulimit soft
10654 Note: Changing this requires a restart of Squid. Also
10655 not all I/O types supports large values (eg on Windows).
10658 NAME: force_request_body_continuation
10660 LOC: Config.accessList.forceRequestBodyContinuation
10662 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
10664 This option controls how Squid handles data upload requests from HTTP
10665 and FTP agents that require a "Please Continue" control message response
10666 to actually send the request body to Squid. It is mostly useful in
10667 adaptation environments.
10669 When Squid receives an HTTP request with an "Expect: 100-continue"
10670 header or an FTP upload command (e.g., STOR), Squid normally sends the
10671 request headers or FTP command information to an adaptation service (or
10672 peer) and waits for a response. Most adaptation services (and some
10673 broken peers) may not respond to Squid at that stage because they may
10674 decide to wait for the HTTP request body or FTP data transfer. However,
10675 that request body or data transfer may never come because Squid has not
10676 responded with the HTTP 100 or FTP 150 (Please Continue) control message
10677 to the request sender yet!
10679 An allow match tells Squid to respond with the HTTP 100 or FTP 150
10680 (Please Continue) control message on its own, before forwarding the
10681 request to an adaptation service or peer. Such a response usually forces
10682 the request sender to proceed with sending the body. A deny match tells
10683 Squid to delay that control response until the origin server confirms
10684 that the request body is needed. Delaying is the default behavior.
10687 NAME: http_upgrade_request_protocols
10688 TYPE: http_upgrade_request_protocols
10689 LOC: Config.http_upgrade_request_protocols
10691 DEFAULT_DOC: Upgrade header dropped, effectively blocking an upgrade attempt.
10693 Controls client-initiated and server-confirmed switching from HTTP to
10694 another protocol (or to several protocols) using HTTP Upgrade mechanism
10695 defined in RFC 7230 Section 6.7. Squid itself does not understand the
10696 protocols being upgraded to and participates in the upgraded
10697 communication only as a dumb TCP proxy. Admins should not allow
10698 upgrading to protocols that require a more meaningful proxy
10701 Usage: http_upgrade_request_protocols <protocol> allow|deny [!]acl ...
10703 The required "protocol" parameter is either an all-caps word OTHER or an
10704 explicit protocol name (e.g. "WebSocket") optionally followed by a slash
10705 and a version token (e.g. "HTTP/3"). Explicit protocol names and
10706 versions are case sensitive.
10708 When an HTTP client sends an Upgrade request header, Squid iterates over
10709 the client-offered protocols and, for each protocol P (with an optional
10710 version V), evaluates the first non-empty set of
10711 http_upgrade_request_protocols rules (if any) from the following list:
10713 * All rules with an explicit protocol name equal to P.
10714 * All rules that use OTHER instead of a protocol name.
10716 In other words, rules using OTHER are considered for protocol P if and
10717 only if there are no rules mentioning P by name.
10719 If both of the above sets are empty, then Squid removes protocol P from
10722 If the client sent a versioned protocol offer P/X, then explicit rules
10723 referring to the same-name but different-version protocol P/Y are
10724 declared inapplicable. Inapplicable rules are not evaluated (i.e. are
10725 ignored). However, inapplicable rules still belong to the first set of
10728 Within the applicable rule subset, individual rules are evaluated in
10729 their configuration order. If all ACLs of an applicable "allow" rule
10730 match, then the protocol offered by the client is forwarded to the next
10731 hop as is. If all ACLs of an applicable "deny" rule match, then the
10732 offer is dropped. If no applicable rules have matching ACLs, then the
10733 offer is also dropped. The first matching rule also ends rules
10734 evaluation for the offered protocol.
10736 If all client-offered protocols are removed, then Squid forwards the
10737 client request without the Upgrade header. Squid never sends an empty
10738 Upgrade request header.
10740 An Upgrade request header with a value violating HTTP syntax is dropped
10741 and ignored without an attempt to use extractable individual protocol
10744 Upon receiving an HTTP 101 (Switching Protocols) control message, Squid
10745 checks that the server listed at least one protocol name and sent a
10746 Connection:upgrade response header. Squid does not understand individual
10747 protocol naming and versioning concepts enough to implement stricter
10748 checks, but an admin can restrict HTTP 101 (Switching Protocols)
10749 responses further using http_reply_access. Responses denied by
10750 http_reply_access rules and responses flagged by the internal Upgrade
10751 checks result in HTTP 502 (Bad Gateway) ERR_INVALID_RESP errors and
10752 Squid-to-server connection closures.
10754 If Squid sends an Upgrade request header, and the next hop (e.g., the
10755 origin server) responds with an acceptable HTTP 101 (Switching
10756 Protocols), then Squid forwards that message to the client and becomes
10759 The presence of an Upgrade request header alone does not preclude cache
10760 lookups. In other words, an Upgrade request might be satisfied from the
10761 cache, using regular HTTP caching rules.
10763 This clause only supports fast acl types.
10764 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
10766 Each of the following groups of configuration lines represents a
10767 separate configuration example:
10769 # never upgrade to protocol Foo; all others are OK
10770 http_upgrade_request_protocols Foo deny all
10771 http_upgrade_request_protocols OTHER allow all
10773 # only allow upgrades to protocol Bar (except for its first version)
10774 http_upgrade_request_protocols Bar/1 deny all
10775 http_upgrade_request_protocols Bar allow all
10776 http_upgrade_request_protocols OTHER deny all # this rule is optional
10778 # only allow upgrades to protocol Baz, and only if Baz is the only offer
10779 acl UpgradeHeaderHasMultipleOffers ...
10780 http_upgrade_request_protocols Baz deny UpgradeHeaderHasMultipleOffers
10781 http_upgrade_request_protocols Baz allow all
10784 NAME: server_pconn_for_nonretriable
10787 DEFAULT_DOC: Open new connections for forwarding requests Squid cannot retry safely.
10788 LOC: Config.accessList.serverPconnForNonretriable
10790 This option provides fine-grained control over persistent connection
10791 reuse when forwarding HTTP requests that Squid cannot retry. It is useful
10792 in environments where opening new connections is very expensive
10793 (e.g., all connections are secured with TLS with complex client and server
10794 certificate validation) and race conditions associated with persistent
10795 connections are very rare and/or only cause minor problems.
10797 HTTP prohibits retrying unsafe and non-idempotent requests (e.g., POST).
10798 Squid limitations also prohibit retrying all requests with bodies (e.g., PUT).
10799 By default, when forwarding such "risky" requests, Squid opens a new
10800 connection to the server or cache_peer, even if there is an idle persistent
10801 connection available. When Squid is configured to risk sending a non-retriable
10802 request on a previously used persistent connection, and the server closes
10803 the connection before seeing that risky request, the user gets an error response
10804 from Squid. In most cases, that error response will be HTTP 502 (Bad Gateway)
10805 with ERR_ZERO_SIZE_OBJECT or ERR_WRITE_ERROR (peer connection reset) error detail.
10807 If an allow rule matches, Squid reuses an available idle persistent connection
10808 (if any) for the request that Squid cannot retry. If a deny rule matches, then
10809 Squid opens a new connection for the request that Squid cannot retry.
10811 This option does not affect requests that Squid can retry. They will reuse idle
10812 persistent connections (if any).
10814 This clause only supports fast acl types.
10815 See https://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
10818 acl SpeedIsWorthTheRisk method POST
10819 server_pconn_for_nonretriable allow SpeedIsWorthTheRisk
10822 NAME: happy_eyeballs_connect_timeout
10826 LOC: Config.happyEyeballs.connect_timeout
10828 This Happy Eyeballs (RFC 8305) tuning directive specifies the minimum
10829 delay between opening a primary to-server connection and opening a
10830 spare to-server connection for the same master transaction. This delay
10831 is similar to the Connection Attempt Delay in RFC 8305, but it is only
10832 applied to the first spare connection attempt. Subsequent spare
10833 connection attempts use happy_eyeballs_connect_gap, and primary
10834 connection attempts are not artificially delayed at all.
10836 Terminology: The "primary" and "spare" designations are determined by
10837 the order of DNS answers received by Squid: If Squid DNS AAAA query
10838 was answered first, then primary connections are connections to IPv6
10839 peer addresses (while spare connections use IPv4 addresses).
10840 Similarly, if Squid DNS A query was answered first, then primary
10841 connections are connections to IPv4 peer addresses (while spare
10842 connections use IPv6 addresses).
10844 Shorter happy_eyeballs_connect_timeout values reduce master
10845 transaction response time, potentially improving user-perceived
10846 response times (i.e., making user eyeballs happier). Longer delays
10847 reduce both concurrent connection level and server bombardment with
10848 connection requests, potentially improving overall Squid performance
10849 and reducing the chance of being blocked by servers for opening too
10850 many unused connections.
10852 RFC 8305 prohibits happy_eyeballs_connect_timeout values smaller than
10853 10 (milliseconds) to "avoid congestion collapse in the presence of
10854 high packet-loss rates".
10856 The following Happy Eyeballs directives place additional connection
10857 opening restrictions: happy_eyeballs_connect_gap and
10858 happy_eyeballs_connect_limit.
10861 NAME: happy_eyeballs_connect_gap
10865 DEFAULT_DOC: no artificial delays between spare attempts
10866 LOC: Config.happyEyeballs.connect_gap
10868 This Happy Eyeballs (RFC 8305) tuning directive specifies the
10869 minimum delay between opening spare to-server connections (to any
10870 server; i.e. across all concurrent master transactions in a Squid
10871 instance). Each SMP worker currently multiplies the configured gap
10872 by the total number of workers so that the combined spare connection
10873 opening rate of a Squid instance obeys the configured limit. The
10874 workers do not coordinate connection openings yet; a micro burst
10875 of spare connection openings may violate the configured gap.
10877 This directive has similar trade-offs as
10878 happy_eyeballs_connect_timeout, but its focus is on limiting traffic
10879 amplification effects for Squid as a whole, while
10880 happy_eyeballs_connect_timeout works on an individual master
10883 The following Happy Eyeballs directives place additional connection
10884 opening restrictions: happy_eyeballs_connect_timeout and
10885 happy_eyeballs_connect_limit. See the former for related terminology.
10888 NAME: happy_eyeballs_connect_limit
10891 DEFAULT_DOC: no artificial limit on the number of concurrent spare attempts
10892 LOC: Config.happyEyeballs.connect_limit
10894 This Happy Eyeballs (RFC 8305) tuning directive specifies the
10895 maximum number of spare to-server connections (to any server; i.e.
10896 across all concurrent master transactions in a Squid instance).
10897 Each SMP worker gets an equal share of the total limit. However,
10898 the workers do not share the actual connection counts yet, so one
10899 (busier) worker cannot "borrow" spare connection slots from another
10900 (less loaded) worker.
10902 Setting this limit to zero disables concurrent use of primary and
10903 spare TCP connections: Spare connection attempts are made only after
10904 all primary attempts fail. However, Squid would still use the
10905 DNS-related optimizations of the Happy Eyeballs approach.
10907 This directive has similar trade-offs as happy_eyeballs_connect_gap,
10908 but its focus is on limiting Squid overheads, while
10909 happy_eyeballs_connect_gap focuses on the origin server and peer
10912 The following Happy Eyeballs directives place additional connection
10913 opening restrictions: happy_eyeballs_connect_timeout and
10914 happy_eyeballs_connect_gap. See the former for related terminology.