1 ## Copyright (C) 1996-2018 The Squid Software Foundation and contributors
3 ## Squid software is distributed under GPLv2+ license and includes
4 ## contributions from numerous individuals and organizations.
5 ## Please see the COPYING and CONTRIBUTORS files for details.
10 ----------------------------
12 This is the documentation for the Squid configuration file.
13 This documentation can also be found online at:
14 http://www.squid-cache.org/Doc/config/
16 You may wish to look at the Squid home page and wiki for the
17 FAQ and other documentation:
18 http://www.squid-cache.org/
19 http://wiki.squid-cache.org/SquidFaq
20 http://wiki.squid-cache.org/ConfigExamples
22 This documentation shows what the defaults for various directives
23 happen to be. If you don't need to change the default, you should
24 leave the line out of your squid.conf in most cases.
26 In some cases "none" refers to no default setting at all,
27 while in other cases it refers to the value of the option
28 - the comments for that keyword indicate if this is the case.
33 Configuration options can be included using the "include" directive.
34 Include takes a list of files to include. Quoting and wildcards are
39 include /path/to/included/file/squid.acl.config
41 Includes can be nested up to a hard-coded depth of 16 levels.
42 This arbitrary restriction is to prevent recursive include references
43 from causing Squid entering an infinite loop whilst trying to load
46 Values with byte units
48 Squid accepts size units on some size related directives. All
49 such directives are documented with a default value displaying
52 Units accepted by Squid are:
54 KB - Kilobyte (1024 bytes)
58 Values with spaces, quotes, and other special characters
60 Squid supports directive parameters with spaces, quotes, and other
61 special characters. Surround such parameters with "double quotes". Use
62 the configuration_includes_quoted_values directive to enable or
65 Squid supports reading configuration option parameters from external
66 files using the syntax:
67 parameters("/path/filename")
69 acl whitelist dstdomain parameters("/etc/squid/whitelist.txt")
71 Conditional configuration
73 If-statements can be used to make configuration directives
77 ... regular configuration directives ...
79 ... regular configuration directives ...]
82 The else part is optional. The keywords "if", "else", and "endif"
83 must be typed on their own lines, as if they were regular
84 configuration directives.
86 NOTE: An else-if condition is not supported.
88 These individual conditions types are supported:
91 Always evaluates to true.
93 Always evaluates to false.
95 Equality comparison of two integer numbers.
100 The following SMP-related preprocessor macros can be used.
102 ${process_name} expands to the current Squid process "name"
103 (e.g., squid1, squid2, or cache1).
105 ${process_number} expands to the current Squid process
106 identifier, which is an integer number (e.g., 1, 2, 3) unique
107 across all Squid processes of the current service instance.
109 ${service_name} expands into the current Squid service instance
110 name identifier which is provided by -n on the command line.
114 Logformat macros can be used in many places outside of the logformat
115 directive. In theory, all of the logformat codes can be used as %macros,
116 where they are supported. In practice, a %macro expands as a dash (-) when
117 the transaction does not yet have enough information and a value is needed.
119 There is no definitive list of what tokens are available at the various
120 stages of the transaction.
122 And some information may already be available to Squid but not yet
123 committed where the macro expansion code can access it (report
124 such instances!). The macro will be expanded into a single dash
125 ('-') in such cases. Not all macros have been tested.
129 # options still not yet ported from 2.7 to 3.x
130 NAME: broken_vary_encoding
133 This option is not yet supported by Squid-3.
139 This option is not yet supported by Squid-3.
145 This option is not yet supported by Squid-3.
148 NAME: external_refresh_check
151 This option is not yet supported by Squid-3.
154 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
157 This option is not yet supported by Squid-3.
160 NAME: refresh_stale_hit
163 This option is not yet supported by Squid-3.
166 # Options removed in 4.x
167 NAME: cache_peer_domain cache_host_domain
170 Replace with dstdomain ACLs and cache_peer_access.
176 Remove this line. The behaviour enabled by this is no longer needed.
179 NAME: sslproxy_cafile
182 Remove this line. Use tls_outgoing_options cafile= instead.
185 NAME: sslproxy_capath
188 Remove this line. Use tls_outgoing_options capath= instead.
191 NAME: sslproxy_cipher
194 Remove this line. Use tls_outgoing_options cipher= instead.
197 NAME: sslproxy_client_certificate
200 Remove this line. Use tls_outgoing_options cert= instead.
203 NAME: sslproxy_client_key
206 Remove this line. Use tls_outgoing_options key= instead.
212 Remove this line. Use tls_outgoing_options flags= instead.
215 NAME: sslproxy_options
218 Remove this line. Use tls_outgoing_options options= instead.
221 NAME: sslproxy_version
224 Remove this line. Use tls_outgoing_options options= instead.
227 # Options removed in 3.5
228 NAME: hierarchy_stoplist
231 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
234 # Options removed in 3.4
238 Remove this line. Use acls with access_log directives to control access logging
244 Remove this line. Use acls with icap_log directives to control icap logging
247 # Options Removed in 3.3
248 NAME: ignore_ims_on_miss
251 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
254 # Options Removed in 3.2
255 NAME: chunked_request_body_max_size
258 Remove this line. Squid is now HTTP/1.1 compliant.
261 NAME: dns_v4_fallback
264 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
267 NAME: emulate_httpd_log
270 Replace this with an access_log directive using the format 'common' or 'combined'.
276 Use a regular access.log with ACL limiting it to MISS events.
282 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
285 NAME: ignore_expect_100
288 Remove this line. The HTTP/1.1 feature is now fully supported by default.
294 Remove this option from your config. To log FQDN use %>A in the log format.
297 NAME: log_ip_on_direct
300 Remove this option from your config. To log server or peer names use %<A in the log format.
303 NAME: maximum_single_addr_tries
306 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
309 NAME: referer_log referrer_log
312 Replace this with an access_log directive using the format 'referrer'.
318 Remove this line. The feature is supported by default in storage types where update is implemented.
321 NAME: url_rewrite_concurrency
324 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
330 Replace this with an access_log directive using the format 'useragent'.
333 # Options Removed in 3.1
337 Remove this line. DNS is no longer tested on startup.
340 NAME: extension_methods
343 Remove this line. All valid methods for HTTP are accepted by default.
346 # 2.7 Options Removed/Replaced in 3.2
351 # 2.7 Options Removed/Replaced in 3.1
359 Remove this line. HTTP/1.1 is supported by default.
362 NAME: upgrade_http0.9
365 Remove this line. ICY/1.0 streaming protocol is supported by default.
368 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
371 Alter these entries. Use the qos_flows directive instead.
374 # Options Removed in 3.0
378 Since squid-3.0 replace with request_header_access or reply_header_access
379 depending on whether you wish to match client requests or server replies.
382 NAME: httpd_accel_no_pmtu_disc
385 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
388 NAME: wais_relay_host
391 Replace this line with 'cache_peer' configuration.
394 NAME: wais_relay_port
397 Replace this line with 'cache_peer' configuration.
402 -----------------------------------------------------------------------------
409 DEFAULT_DOC: SMP support disabled.
411 Number of main Squid processes or "workers" to fork and maintain.
412 0: "no daemon" mode, like running "squid -N ..."
413 1: "no SMP" mode, start one main Squid process daemon (default)
414 N: start N main Squid process daemons (i.e., SMP mode)
416 In SMP mode, each worker does nearly all what a single Squid daemon
417 does (e.g., listen on http_port and forward HTTP requests).
420 NAME: cpu_affinity_map
422 LOC: Config.cpuAffinityMap
424 DEFAULT_DOC: Let operating system decide.
426 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
428 Sets 1:1 mapping between Squid processes and CPU cores. For example,
430 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
432 affects processes 1 through 4 only and places them on the first
433 four even cores, starting with core #1.
435 CPU cores are numbered starting from 1. Requires support for
436 sched_getaffinity(2) and sched_setaffinity(2) system calls.
438 Multiple cpu_affinity_map options are merged.
443 NAME: shared_memory_locking
446 LOC: Config.shmLocking
449 Whether to ensure that all required shared memory is available by
450 "locking" that shared memory into RAM when Squid starts. The
451 alternative is faster startup time followed by slightly slower
452 performance and, if not enough RAM is actually available during
453 runtime, mysterious crashes.
455 SMP Squid uses many shared memory segments. These segments are
456 brought into Squid memory space using an mmap(2) system call. During
457 Squid startup, the mmap() call often succeeds regardless of whether
458 the system has enough RAM. In general, Squid cannot tell whether the
459 kernel applies this "optimistic" memory allocation policy (but
460 popular modern kernels usually use it).
462 Later, if Squid attempts to actually access the mapped memory
463 regions beyond what the kernel is willing to allocate, the
464 "optimistic" kernel simply kills Squid kid with a SIGBUS signal.
465 Some of the memory limits enforced by the kernel are currently
466 poorly understood: We do not know how to detect and check them. This
467 option ensures that the mapped memory will be available.
469 This option may have a positive performance side-effect: Locking
470 memory at start avoids runtime paging I/O. Paging slows Squid down.
472 Locking memory may require a large enough RLIMIT_MEMLOCK OS limit,
473 CAP_IPC_LOCK capability, or equivalent.
476 NAME: hopeless_kid_revival_delay
479 LOC: Config.hopelessKidRevivalDelay
482 Normally, when a kid process dies, Squid immediately restarts the
483 kid. A kid experiencing frequent deaths is marked as "hopeless" for
484 the duration specified by this directive. Hopeless kids are not
485 automatically restarted.
487 Currently, zero values are not supported because they result in
488 misconfigured SMP Squid instances running forever, endlessly
489 restarting each dying kid. To effectively disable hopeless kids
490 revival, set the delay to a huge value (e.g., 1 year).
492 Reconfiguration also clears all hopeless kids designations, allowing
493 for manual revival of hopeless kids.
497 OPTIONS FOR AUTHENTICATION
498 -----------------------------------------------------------------------------
504 LOC: Auth::TheConfig.schemes
507 This is used to define parameters for the various authentication
508 schemes supported by Squid.
510 format: auth_param scheme parameter [setting]
512 The order in which authentication schemes are presented to the client is
513 dependent on the order the scheme first appears in config file. IE
514 has a bug (it's not RFC 2617 compliant) in that it will use the basic
515 scheme if basic is the first entry presented, even if more secure
516 schemes are presented. For now use the order in the recommended
517 settings section below. If other browsers have difficulties (don't
518 recognize the schemes offered even if you are using basic) either
519 put basic first, or disable the other schemes (by commenting out their
522 Once an authentication scheme is fully configured, it can only be
523 shutdown by shutting squid down and restarting. Changes can be made on
524 the fly and activated with a reconfigure. I.E. You can change to a
525 different helper, but not unconfigure the helper completely.
527 Please note that while this directive defines how Squid processes
528 authentication it does not automatically activate authentication.
529 To use authentication you must in addition make use of ACLs based
530 on login name in http_access (proxy_auth, proxy_auth_regex or
531 external with %LOGIN used in the format tag). The browser will be
532 challenged for authentication on the first such acl encountered
533 in http_access processing and will also be re-challenged for new
534 login credentials if the request is being denied by a proxy_auth
537 WARNING: authentication can't be used in a transparently intercepting
538 proxy as the client then thinks it is talking to an origin server and
539 not the proxy. This is a limitation of bending the TCP/IP protocol to
540 transparently intercepting port 80, not a limitation in Squid.
541 Ports flagged 'transparent', 'intercept', or 'tproxy' have
542 authentication disabled.
544 === Parameters common to all schemes. ===
547 Specifies the command for the external authenticator.
549 By default, each authentication scheme is not used unless a
550 program is specified.
552 See http://wiki.squid-cache.org/Features/AddonHelpers for
553 more details on helper operations and creating your own.
556 Specifies a string to be append to request line format for
557 the authentication helper. "Quoted" format values may contain
558 spaces and logformat %macros. In theory, any logformat %macro
559 can be used. In practice, a %macro expands as a dash (-) if
560 the helper request is sent before the required macro
561 information is available to Squid.
563 By default, Squid uses request formats provided in
564 scheme-specific examples below (search for %credentials).
566 The expanded key_extras value is added to the Squid credentials
567 cache and, hence, will affect authentication. It can be used to
568 autenticate different users with identical user names (e.g.,
569 when user authentication depends on http_port).
571 Avoid adding frequently changing information to key_extras. For
572 example, if you add user source IP, and it changes frequently
573 in your environment, then max_user_ip ACL is going to treat
574 every user+IP combination as a unique "user", breaking the ACL
575 and wasting a lot of memory on those user records. It will also
576 force users to authenticate from scratch whenever their IP
580 Specifies the protection scope (aka realm name) which is to be
581 reported to the client for the authentication scheme. It is
582 commonly part of the text the user will see when prompted for
583 their username and password.
585 For Basic the default is "Squid proxy-caching web server".
586 For Digest there is no default, this parameter is mandatory.
587 For NTLM and Negotiate this parameter is ignored.
589 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
590 [queue-size=N] [on-persistent-overload=action]
592 The maximum number of authenticator processes to spawn. If
593 you start too few Squid will have to wait for them to process
594 a backlog of credential verifications, slowing it down. When
595 password verifications are done via a (slow) network you are
596 likely to need lots of authenticator processes.
598 The startup= and idle= options permit some skew in the exact
599 amount run. A minimum of startup=N will begin during startup
600 and reconfigure. Squid will start more in groups of up to
601 idle=N in an attempt to meet traffic needs and to keep idle=N
602 free above those traffic needs up to the maximum.
604 The concurrency= option sets the number of concurrent requests
605 the helper can process. The default of 0 is used for helpers
606 who only supports one request at a time. Setting this to a
607 number greater than 0 changes the protocol used to include a
608 channel ID field first on the request/response line, allowing
609 multiple requests to be sent to the same helper in parallel
610 without waiting for the response.
612 Concurrency must not be set unless it's known the helper
613 supports the input format with channel-ID fields.
615 The queue-size=N option sets the maximum number of queued
616 requests to N. The default maximum is 2*numberofchildren. Squid
617 is allowed to temporarily exceed the configured maximum, marking
618 the affected helper as "overloaded". If the helper overload
619 lasts more than 3 minutes, the action prescribed by the
620 on-persistent-overload option applies.
622 The on-persistent-overload=action option specifies Squid
623 reaction to a new helper request arriving when the helper
624 has been overloaded for more that 3 minutes already. The number
625 of queued requests determines whether the helper is overloaded
626 (see the queue-size option).
628 Two actions are supported:
630 die Squid worker quits. This is the default behavior.
632 ERR Squid treats the helper request as if it was
633 immediately submitted, and the helper immediately
634 replied with an ERR response. This action has no effect
635 on the already queued and in-progress helper requests.
637 NOTE: NTLM and Negotiate schemes do not support concurrency
638 in the Squid code module even though some helpers can.
641 If you experience problems with PUT/POST requests when using
642 the NTLM or Negotiate schemes then you can try setting this
643 to off. This will cause Squid to forcibly close the connection
644 on the initial request where the browser asks which schemes
645 are supported by the proxy.
647 For Basic and Digest this parameter is ignored.
650 HTTP uses iso-latin-1 as character set, while some
651 authentication backends such as LDAP expects UTF-8. If this is
652 set to on Squid will translate the HTTP iso-latin-1 charset to
653 UTF-8 before sending the username and password to the helper.
655 For NTLM and Negotiate this parameter is ignored.
657 IF HAVE_AUTH_MODULE_BASIC
658 === Basic authentication parameters ===
660 "credentialsttl" timetolive
661 Specifies how long squid assumes an externally validated
662 username:password pair is valid for - in other words how
663 often the helper program is called for that user. Set this
664 low to force revalidation with short lived passwords.
666 NOTE: setting this high does not impact your susceptibility
667 to replay attacks unless you are using an one-time password
668 system (such as SecureID). If you are using such a system,
669 you will be vulnerable to replay attacks unless you also
670 use the max_user_ip ACL in an http_access rule.
672 "casesensitive" on|off
673 Specifies if usernames are case sensitive. Most user databases
674 are case insensitive allowing the same username to be spelled
675 using both lower and upper case letters, but some are case
676 sensitive. This makes a big difference for user_max_ip ACL
677 processing and similar.
680 IF HAVE_AUTH_MODULE_DIGEST
681 === Digest authentication parameters ===
683 "nonce_garbage_interval" timeinterval
684 Specifies the interval that nonces that have been issued
685 to client_agent's are checked for validity.
687 "nonce_max_duration" timeinterval
688 Specifies the maximum length of time a given nonce will be
691 "nonce_max_count" number
692 Specifies the maximum number of times a given nonce can be
695 "nonce_strictness" on|off
696 Determines if squid requires strict increment-by-1 behavior
697 for nonce counts, or just incrementing (off - for use when
698 user agents generate nonce counts that occasionally miss 1
699 (ie, 1,2,4,6)). Default off.
701 "check_nonce_count" on|off
702 This directive if set to off can disable the nonce count check
703 completely to work around buggy digest qop implementations in
704 certain mainstream browser versions. Default on to check the
705 nonce count to protect from authentication replay attacks.
707 "post_workaround" on|off
708 This is a workaround to certain buggy browsers who send an
709 incorrect request digest in POST requests when reusing the
710 same nonce as acquired earlier on a GET request.
714 === Example Configuration ===
716 This configuration displays the recommended authentication scheme
717 order from most to least secure with recommended minimum configuration
718 settings for each scheme:
720 #auth_param negotiate program <uncomment and complete this line to activate>
721 #auth_param negotiate children 20 startup=0 idle=1
723 #auth_param digest program <uncomment and complete this line to activate>
724 #auth_param digest children 20 startup=0 idle=1
725 #auth_param digest realm Squid proxy-caching web server
726 #auth_param digest nonce_garbage_interval 5 minutes
727 #auth_param digest nonce_max_duration 30 minutes
728 #auth_param digest nonce_max_count 50
730 #auth_param ntlm program <uncomment and complete this line to activate>
731 #auth_param ntlm children 20 startup=0 idle=1
733 #auth_param basic program <uncomment and complete this line>
734 #auth_param basic children 5 startup=5 idle=1
735 #auth_param basic credentialsttl 2 hours
738 NAME: authenticate_cache_garbage_interval
742 LOC: Auth::TheConfig.garbageCollectInterval
744 The time period between garbage collection across the username cache.
745 This is a trade-off between memory utilization (long intervals - say
746 2 days) and CPU (short intervals - say 1 minute). Only change if you
750 NAME: authenticate_ttl
754 LOC: Auth::TheConfig.credentialsTtl
756 The time a user & their credentials stay in the logged in
757 user cache since their last request. When the garbage
758 interval passes, all user credentials that have passed their
759 TTL are removed from memory.
762 NAME: authenticate_ip_ttl
765 LOC: Auth::TheConfig.ipTtl
768 If you use proxy authentication and the 'max_user_ip' ACL,
769 this directive controls how long Squid remembers the IP
770 addresses associated with each user. Use a small value
771 (e.g., 60 seconds) if your users might change addresses
772 quickly, as is the case with dialup. You might be safe
773 using a larger value (e.g., 2 hours) in a corporate LAN
774 environment with relatively static address assignments.
779 -----------------------------------------------------------------------------
782 NAME: external_acl_type
783 TYPE: externalAclHelper
784 LOC: Config.externalAclHelperList
787 This option defines external acl classes using a helper program
788 to look up the status
790 external_acl_type name [options] FORMAT /path/to/helper [helper arguments]
794 ttl=n TTL in seconds for cached results (defaults to 3600
798 TTL for cached negative lookups (default same
801 grace=n Percentage remaining of TTL where a refresh of a
802 cached entry should be initiated without needing to
803 wait for a new reply. (default is for no grace period)
805 cache=n The maximum number of entries in the result cache. The
806 default limit is 262144 entries. Each cache entry usually
807 consumes at least 256 bytes. Squid currently does not remove
808 expired cache entries until the limit is reached, so a proxy
809 will sooner or later reach the limit. The expanded FORMAT
810 value is used as the cache key, so if the details in FORMAT
811 are highly variable, a larger cache may be needed to produce
812 reduction in helper load.
815 Maximum number of acl helper processes spawned to service
816 external acl lookups of this type. (default 5)
819 Minimum number of acl helper processes to spawn during
820 startup and reconfigure to service external acl lookups
821 of this type. (default 0)
824 Number of acl helper processes to keep ahead of traffic
825 loads. Squid will spawn this many at once whenever load
826 rises above the capabilities of existing processes.
827 Up to the value of children-max. (default 1)
829 concurrency=n concurrency level per process. Only used with helpers
830 capable of processing more than one query at a time.
832 queue-size=N The queue-size= option sets the maximum number of queued
833 requests. If the queued requests exceed queue size
835 The default value is set to 2*children-max.
837 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
839 ipv4 / ipv6 IP protocol used to communicate with this helper.
840 The default is to auto-detect IPv6 and use it when available.
843 FORMAT is a series of %macro codes. See logformat directive for a full list
844 of the accepted codes. Although note that at the time of any external ACL
845 being tested data may not be available and thus some %macro expand to '-'.
847 In addition to the logformat codes; when processing external ACLs these
848 additional macros are made available:
850 %ACL The name of the ACL being tested.
852 %DATA The ACL arguments specified in the referencing config
853 'acl ... external' line, separated by spaces (an
854 "argument string"). see acl external.
856 If there are no ACL arguments %DATA expands to '-'.
858 If you do not specify a DATA macro inside FORMAT,
859 Squid automatically appends %DATA to your FORMAT.
861 By default, Squid applies URL-encoding to each ACL
862 argument inside the argument string. If an explicit
863 encoding modifier is used (e.g., %#DATA), then Squid
864 encodes the whole argument string as a single token
865 (e.g., with %#DATA, spaces between arguments become
868 If SSL is enabled, the following formating codes become available:
870 %USER_CERT SSL User certificate in PEM format
871 %USER_CERTCHAIN SSL User certificate chain in PEM format
872 %USER_CERT_xx SSL User certificate subject attribute xx
873 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
876 NOTE: all other format codes accepted by older Squid versions
880 General request syntax:
882 [channel-ID] FORMAT-values
885 FORMAT-values consists of transaction details expanded with
886 whitespace separation per the config file FORMAT specification
887 using the FORMAT macros listed above.
889 Request values sent to the helper are URL escaped to protect
890 each value in requests against whitespaces.
892 If using protocol=2.5 then the request sent to the helper is not
893 URL escaped to protect against whitespace.
895 NOTE: protocol=3.0 is deprecated as no longer necessary.
897 When using the concurrency= option the protocol is changed by
898 introducing a query channel tag in front of the request/response.
899 The query channel tag is a number between 0 and concurrency-1.
900 This value must be echoed back unchanged to Squid as the first part
901 of the response relating to its request.
904 The helper receives lines expanded per the above format specification
905 and for each input line returns 1 line starting with OK/ERR/BH result
906 code and optionally followed by additional keywords with more details.
909 General result syntax:
911 [channel-ID] result keyword=value ...
913 Result consists of one of the codes:
916 the ACL test produced a match.
919 the ACL test does not produce a match.
922 An internal error occurred in the helper, preventing
923 a result being identified.
925 The meaning of 'a match' is determined by your squid.conf
926 access control configuration. See the Squid wiki for details.
930 user= The users name (login)
932 password= The users password (for login= cache_peer option)
934 message= Message describing the reason for this response.
935 Available as %o in error pages.
936 Useful on (ERR and BH results).
938 tag= Apply a tag to a request. Only sets a tag once,
939 does not alter existing tags.
941 log= String to be logged in access.log. Available as
942 %ea in logformat specifications.
944 clt_conn_tag= Associates a TAG with the client TCP connection.
945 Please see url_rewrite_program related documentation
948 Any keywords may be sent on any response whether OK, ERR or BH.
950 All response keyword values need to be a single token with URL
951 escaping, or enclosed in double quotes (") and escaped using \ on
952 any double quotes or \ characters within the value. The wrapping
953 double quotes are removed before the value is interpreted by Squid.
954 \r and \n are also replace by CR and LF.
956 Some example key values:
960 user="J. \"Bob\" Smith"
967 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
968 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
969 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
970 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
971 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
974 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
975 DEFAULT: localhost src 127.0.0.1/32 ::1
976 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
977 DEFAULT: CONNECT method CONNECT
978 DEFAULT_DOC: ACLs all, manager, localhost, to_localhost, and CONNECT are predefined.
980 Defining an Access List
982 Every access list definition must begin with an aclname and acltype,
983 followed by either type-specific arguments or a quoted filename that
986 acl aclname acltype argument ...
987 acl aclname acltype "file" ...
989 When using "file", the file should contain one item per line.
994 Some acl types supports options which changes their default behaviour:
996 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
997 case-insensitive, use the -i option. To return case-sensitive
998 use the +i option between patterns, or make a new ACL line
1001 -n Disable lookups and address type conversions. If lookup or
1002 conversion is required because the parameter type (IP or
1003 domain name) does not match the message address type (domain
1004 name or IP), then the ACL would immediately declare a mismatch
1005 without any warnings or lookups.
1008 Perform a list membership test, interpreting values as
1009 comma-separated token lists and matching against individual
1010 tokens instead of whole values.
1011 The optional "delimiters" parameter specifies one or more
1012 alternative non-alphanumeric delimiter characters.
1013 non-alphanumeric delimiter characters.
1015 -- Used to stop processing all options, in the case the first acl
1016 value has '-' character as first character (for example the '-'
1017 is a valid domain name)
1019 Some acl types require suspending the current request in order
1020 to access some external data source.
1021 Those which do are marked with the tag [slow], those which
1022 don't are marked as [fast].
1023 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
1024 for further information
1026 ***** ACL TYPES AVAILABLE *****
1028 acl aclname src ip-address/mask ... # clients IP address [fast]
1029 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
1030 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
1031 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
1033 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
1035 # The 'arp' ACL code is not portable to all operating systems.
1036 # It works on Linux, Solaris, Windows, FreeBSD, and some other
1039 # NOTE: Squid can only determine the MAC/EUI address for IPv4
1040 # clients that are on the same subnet. If the client is on a
1041 # different subnet, then Squid cannot find out its address.
1043 # NOTE 2: IPv6 protocol does not contain ARP. MAC/EUI is either
1044 # encoded directly in the IPv6 address or not available.
1046 acl aclname clientside_mark mark[/mask] ...
1047 # matches CONNMARK of an accepted connection [fast]
1049 # mark and mask are unsigned integers (hex, octal, or decimal).
1050 # If multiple marks are given, then the ACL matches if at least
1053 # Uses netfilter-conntrack library.
1054 # Requires building Squid with --enable-linux-netfilter.
1056 # The client, various intermediaries, and Squid itself may set
1057 # CONNMARK at various times. The last CONNMARK set wins. This ACL
1058 # checks the mark present on an accepted connection or set by
1059 # Squid afterwards, depending on the ACL check timing. This ACL
1060 # effectively ignores any mark set by other agents after Squid has
1061 # accepted the connection.
1063 acl aclname srcdomain .foo.com ...
1064 # reverse lookup, from client IP [slow]
1065 acl aclname dstdomain [-n] .foo.com ...
1066 # Destination server from URL [fast]
1067 acl aclname srcdom_regex [-i] \.foo\.com ...
1068 # regex matching client name [slow]
1069 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
1070 # regex matching server [fast]
1072 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
1073 # based URL is used and no match is found. The name "none" is used
1074 # if the reverse lookup fails.
1076 acl aclname src_as number ...
1077 acl aclname dst_as number ...
1079 # Except for access control, AS numbers can be used for
1080 # routing of requests to specific caches. Here's an
1081 # example for routing all requests for AS#1241 and only
1082 # those to mycache.mydomain.net:
1083 # acl asexample dst_as 1241
1084 # cache_peer_access mycache.mydomain.net allow asexample
1085 # cache_peer_access mycache_mydomain.net deny all
1087 acl aclname peername myPeer ...
1089 # match against a named cache_peer entry
1090 # set unique name= on cache_peer lines for reliable use.
1092 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
1102 # h1:m1 must be less than h2:m2
1104 acl aclname url_regex [-i] ^http:// ...
1105 # regex matching on whole URL [fast]
1106 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
1107 # regex matching on URL login field
1108 acl aclname urlpath_regex [-i] \.gif$ ...
1109 # regex matching on URL path [fast]
1111 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
1113 acl aclname localport 3128 ... # TCP port the client connected to [fast]
1114 # NP: for interception mode this is usually '80'
1116 acl aclname myportname 3128 ... # *_port name [fast]
1118 acl aclname proto HTTP FTP ... # request protocol [fast]
1120 acl aclname method GET POST ... # HTTP request method [fast]
1122 acl aclname http_status 200 301 500- 400-403 ...
1123 # status code in reply [fast]
1125 acl aclname browser [-i] regexp ...
1126 # pattern match on User-Agent header (see also req_header below) [fast]
1128 acl aclname referer_regex [-i] regexp ...
1129 # pattern match on Referer header [fast]
1130 # Referer is highly unreliable, so use with care
1132 acl aclname ident username ...
1133 acl aclname ident_regex [-i] pattern ...
1134 # string match on ident output [slow]
1135 # use REQUIRED to accept any non-null ident.
1137 acl aclname proxy_auth [-i] username ...
1138 acl aclname proxy_auth_regex [-i] pattern ...
1139 # perform http authentication challenge to the client and match against
1140 # supplied credentials [slow]
1142 # takes a list of allowed usernames.
1143 # use REQUIRED to accept any valid username.
1145 # Will use proxy authentication in forward-proxy scenarios, and plain
1146 # http authenticaiton in reverse-proxy scenarios
1148 # NOTE: when a Proxy-Authentication header is sent but it is not
1149 # needed during ACL checking the username is NOT logged
1152 # NOTE: proxy_auth requires a EXTERNAL authentication program
1153 # to check username/password combinations (see
1154 # auth_param directive).
1156 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1157 # as the browser needs to be configured for using a proxy in order
1158 # to respond to proxy authentication.
1160 acl aclname snmp_community string ...
1161 # A community string to limit access to your SNMP Agent [fast]
1164 # acl snmppublic snmp_community public
1166 acl aclname maxconn number
1167 # This will be matched when the client's IP address has
1168 # more than <number> TCP connections established. [fast]
1169 # NOTE: This only measures direct TCP links so X-Forwarded-For
1170 # indirect clients are not counted.
1172 acl aclname max_user_ip [-s] number
1173 # This will be matched when the user attempts to log in from more
1174 # than <number> different ip addresses. The authenticate_ip_ttl
1175 # parameter controls the timeout on the ip entries. [fast]
1176 # If -s is specified the limit is strict, denying browsing
1177 # from any further IP addresses until the ttl has expired. Without
1178 # -s Squid will just annoy the user by "randomly" denying requests.
1179 # (the counter is reset each time the limit is reached and a
1180 # request is denied)
1181 # NOTE: in acceleration mode or where there is mesh of child proxies,
1182 # clients may appear to come from multiple addresses if they are
1183 # going through proxy farms, so a limit of 1 may cause user problems.
1185 acl aclname random probability
1186 # Pseudo-randomly match requests. Based on the probability given.
1187 # Probability may be written as a decimal (0.333), fraction (1/3)
1188 # or ratio of matches:non-matches (3:5).
1190 acl aclname req_mime_type [-i] mime-type ...
1191 # regex match against the mime type of the request generated
1192 # by the client. Can be used to detect file upload or some
1193 # types HTTP tunneling requests [fast]
1194 # NOTE: This does NOT match the reply. You cannot use this
1195 # to match the returned file type.
1197 acl aclname req_header header-name [-i] any\.regex\.here
1198 # regex match against any of the known request headers. May be
1199 # thought of as a superset of "browser", "referer" and "mime-type"
1202 acl aclname rep_mime_type [-i] mime-type ...
1203 # regex match against the mime type of the reply received by
1204 # squid. Can be used to detect file download or some
1205 # types HTTP tunneling requests. [fast]
1206 # NOTE: This has no effect in http_access rules. It only has
1207 # effect in rules that affect the reply data stream such as
1208 # http_reply_access.
1210 acl aclname rep_header header-name [-i] any\.regex\.here
1211 # regex match against any of the known reply headers. May be
1212 # thought of as a superset of "browser", "referer" and "mime-type"
1215 acl aclname external class_name [arguments...]
1216 # external ACL lookup via a helper class defined by the
1217 # external_acl_type directive [slow]
1219 acl aclname user_cert attribute values...
1220 # match against attributes in a user SSL certificate
1221 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1223 acl aclname ca_cert attribute values...
1224 # match against attributes a users issuing CA SSL certificate
1225 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1227 acl aclname ext_user username ...
1228 acl aclname ext_user_regex [-i] pattern ...
1229 # string match on username returned by external acl helper [slow]
1230 # use REQUIRED to accept any non-null user name.
1232 acl aclname tag tagvalue ...
1233 # string match on tag returned by external acl helper [fast]
1234 # DEPRECATED. Only the first tag will match with this ACL.
1235 # Use the 'note' ACL instead for handling multiple tag values.
1237 acl aclname hier_code codename ...
1238 # string match against squid hierarchy code(s); [fast]
1239 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1241 # NOTE: This has no effect in http_access rules. It only has
1242 # effect in rules that affect the reply data stream such as
1243 # http_reply_access.
1245 acl aclname note [-m[=delimiters]] name [value ...]
1246 # match transaction annotation [fast]
1247 # Without values, matches any annotation with a given name.
1248 # With value(s), matches any annotation with a given name that
1249 # also has one of the given values.
1250 # If the -m flag is used, then the value of the named
1251 # annotation is interpreted as a list of tokens, and the ACL
1252 # matches individual name=token pairs rather than whole
1253 # name=value pairs. See "ACL Options" above for more info.
1254 # Annotation sources include note and adaptation_meta directives
1255 # as well as helper and eCAP responses.
1257 acl aclname annotate_transaction [-m[=delimiters]] key=value ...
1258 acl aclname annotate_transaction [-m[=delimiters]] key+=value ...
1259 # Always matches. [fast]
1260 # Used for its side effect: This ACL immediately adds a
1261 # key=value annotation to the current master transaction.
1262 # The added annotation can then be tested using note ACL and
1263 # logged (or sent to helpers) using %note format code.
1265 # Annotations can be specified using replacement and addition
1266 # formats. The key=value form replaces old same-key annotation
1267 # value(s). The key+=value form appends a new value to the old
1268 # same-key annotation. Both forms create a new key=value
1269 # annotation if no same-key annotation exists already. If
1270 # -m flag is used, then the value is interpreted as a list
1271 # and the annotation will contain key=token pair(s) instead of the
1272 # whole key=value pair.
1274 # This ACL is especially useful for recording complex multi-step
1275 # ACL-driven decisions. For example, the following configuration
1276 # avoids logging transactions accepted after aclX matched:
1278 # # First, mark transactions accepted after aclX matched
1279 # acl markSpecial annotate_transaction special=true
1280 # http_access allow acl001
1282 # http_access deny acl100
1283 # http_access allow aclX markSpecial
1285 # # Second, do not log marked transactions:
1286 # acl markedSpecial note special true
1287 # access_log ... deny markedSpecial
1289 # # Note that the following would not have worked because aclX
1290 # # alone does not determine whether the transaction was allowed:
1291 # access_log ... deny aclX # Wrong!
1293 # Warning: This ACL annotates the transaction even when negated
1294 # and even if subsequent ACLs fail to match. For example, the
1295 # following three rules will have exactly the same effect as far
1296 # as annotations set by the "mark" ACL are concerned:
1298 # some_directive acl1 ... mark # rule matches if mark is reached
1299 # some_directive acl1 ... !mark # rule never matches
1300 # some_directive acl1 ... mark !all # rule never matches
1302 acl aclname annotate_client [-m[=delimiters]] key=value ...
1303 acl aclname annotate_client [-m[=delimiters]] key+=value ...
1305 # Always matches. [fast]
1306 # Used for its side effect: This ACL immediately adds a
1307 # key=value annotation to the current client-to-Squid
1308 # connection. Connection annotations are propagated to the current
1309 # and all future master transactions on the annotated connection.
1310 # See the annotate_transaction ACL for details.
1312 # For example, the following configuration avoids rewriting URLs
1313 # of transactions bumped by SslBump:
1315 # # First, mark bumped connections:
1316 # acl markBumped annotate_client bumped=true
1317 # ssl_bump peek acl1
1318 # ssl_bump stare acl2
1319 # ssl_bump bump acl3 markBumped
1320 # ssl_bump splice all
1322 # # Second, do not send marked transactions to the redirector:
1323 # acl markedBumped note bumped true
1324 # url_rewrite_access deny markedBumped
1326 # # Note that the following would not have worked because acl3 alone
1327 # # does not determine whether the connection is going to be bumped:
1328 # url_rewrite_access deny acl3 # Wrong!
1330 acl aclname adaptation_service service ...
1331 # Matches the name of any icap_service, ecap_service,
1332 # adaptation_service_set, or adaptation_service_chain that Squid
1333 # has used (or attempted to use) for the master transaction.
1334 # This ACL must be defined after the corresponding adaptation
1335 # service is named in squid.conf. This ACL is usable with
1336 # adaptation_meta because it starts matching immediately after
1337 # the service has been selected for adaptation.
1339 acl aclname transaction_initiator initiator ...
1340 # Matches transaction's initiator [fast]
1342 # Supported initiators are:
1343 # esi: matches transactions fetching ESI resources
1344 # certificate-fetching: matches transactions fetching
1345 # a missing intermediate TLS certificate
1346 # cache-digest: matches transactions fetching Cache Digests
1348 # htcp: matches HTCP requests from peers
1349 # icp: matches ICP requests to peers
1350 # icmp: matches ICMP RTT database (NetDB) requests to peers
1351 # asn: matches asns db requests
1352 # internal: matches any of the above
1353 # client: matches transactions containing an HTTP or FTP
1354 # client request received at a Squid *_port
1355 # all: matches any transaction, including internal transactions
1356 # without a configurable initiator and hopefully rare
1357 # transactions without a known-to-Squid initiator
1359 # Multiple initiators are ORed.
1361 acl aclname has component
1362 # matches a transaction "component" [fast]
1364 # Supported transaction components are:
1365 # request: transaction has a request header (at least)
1366 # response: transaction has a response header (at least)
1367 # ALE: transaction has an internally-generated Access Log Entry
1368 # structure; bugs notwithstanding, all transaction have it
1370 # For example, the following configuration helps when dealing with HTTP
1371 # clients that close connections without sending a request header:
1373 # acl hasRequest has request
1374 # acl logMe note important_transaction
1375 # # avoid "logMe ACL is used in context without an HTTP request" warnings
1376 # access_log ... logformat=detailed hasRequest logMe
1377 # # log request-less transactions, instead of ignoring them
1378 # access_log ... logformat=brief !hasRequest
1380 # Multiple components are not supported for one "acl" rule, but
1381 # can be specified (and are ORed) using multiple same-name rules:
1383 # # OK, this strange logging daemon needs request or response,
1384 # # but can work without either a request or a response:
1385 # acl hasWhatMyLoggingDaemonNeeds has request
1386 # acl hasWhatMyLoggingDaemonNeeds has response
1389 acl aclname ssl_error errorname
1390 # match against SSL certificate validation error [fast]
1392 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1395 # The following can be used as shortcuts for certificate properties:
1396 # [ssl::]certHasExpired: the "not after" field is in the past
1397 # [ssl::]certNotYetValid: the "not before" field is in the future
1398 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1399 # [ssl::]certSelfSigned: The certificate is self signed.
1400 # [ssl::]certDomainMismatch: The certificate CN domain does not
1401 # match the name the name of the host we are connecting to.
1403 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1404 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1405 # predefined ACLs, just like the 'all' ACL.
1407 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1408 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1410 acl aclname server_cert_fingerprint [-sha1] fingerprint
1411 # match against server SSL certificate fingerprint [fast]
1413 # The fingerprint is the digest of the DER encoded version
1414 # of the whole certificate. The user should use the form: XX:XX:...
1415 # Optional argument specifies the digest algorithm to use.
1416 # The SHA1 digest algorithm is the default and is currently
1417 # the only algorithm supported (-sha1).
1419 acl aclname at_step step
1420 # match against the current step during ssl_bump evaluation [fast]
1421 # Never matches and should not be used outside the ssl_bump context.
1423 # At each SslBump step, Squid evaluates ssl_bump directives to find
1424 # the next bumping action (e.g., peek or splice). Valid SslBump step
1425 # values and the corresponding ssl_bump evaluation moments are:
1426 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1427 # SslBump2: After getting SSL Client Hello info.
1428 # SslBump3: After getting SSL Server Hello info.
1430 acl aclname ssl::server_name [option] .foo.com ...
1431 # matches server name obtained from various sources [fast]
1433 # The ACL computes server name(s) using such information sources as
1434 # CONNECT request URI, TLS client SNI, and TLS server certificate
1435 # subject (CN and SubjectAltName). The computed server name(s) usually
1436 # change with each SslBump step, as more info becomes available:
1437 # * SNI is used as the server name instead of the request URI,
1438 # * subject name(s) from the server certificate (CN and
1439 # SubjectAltName) are used as the server names instead of SNI.
1441 # When the ACL computes multiple server names, matching any single
1442 # computed name is sufficient for the ACL to match.
1444 # The "none" name can be used to match transactions where the ACL
1445 # could not compute the server name using any information source
1446 # that was both available and allowed to be used by the ACL options at
1447 # the ACL evaluation time.
1449 # Unlike dstdomain, this ACL does not perform DNS lookups.
1451 # An ACL option below may be used to restrict what information
1452 # sources are used to extract the server names from:
1454 # --client-requested
1455 # The server name is SNI regardless of what the server says.
1457 # The server name(s) are the certificate subject name(s), regardless
1458 # of what the client has requested. If the server certificate is
1459 # unavailable, then the name is "none".
1461 # The server name is either SNI (if SNI matches at least one of the
1462 # certificate subject names) or "none" (otherwise). When the server
1463 # certificate is unavailable, the consensus server name is SNI.
1465 # Combining multiple options in one ACL is a fatal configuration
1468 # For all options: If no SNI is available, then the CONNECT request
1469 # target (a.k.a. URI) is used instead of SNI (for an intercepted
1470 # connection, this target is the destination IP address).
1472 acl aclname ssl::server_name_regex [-i] \.foo\.com ...
1473 # regex matches server name obtained from various sources [fast]
1475 acl aclname connections_encrypted
1476 # matches transactions with all HTTP messages received over TLS
1477 # transport connections. [fast]
1479 # The master transaction deals with HTTP messages received from
1480 # various sources. All sources used by the master transaction in the
1481 # past are considered by the ACL. The following rules define whether
1482 # a given message source taints the entire master transaction,
1483 # resulting in ACL mismatches:
1485 # * The HTTP client transport connection is not TLS.
1486 # * An adaptation service connection-encryption flag is off.
1487 # * The peer or origin server transport connection is not TLS.
1489 # Caching currently does not affect these rules. This cache ignorance
1490 # implies that only the current HTTP client transport and REQMOD
1491 # services status determine whether this ACL matches a from-cache
1492 # transaction. The source of the cached response does not have any
1493 # effect on future transaction that use the cached response without
1494 # revalidation. This may change.
1496 # DNS, ICP, and HTCP exchanges during the master transaction do not
1497 # affect these rules.
1499 acl aclname any-of acl1 acl2 ...
1500 # match any one of the acls [fast or slow]
1501 # The first matching ACL stops further ACL evaluation.
1503 # ACLs from multiple any-of lines with the same name are ORed.
1504 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1505 # acl A any-of a1 a2
1506 # acl A any-of a3 a4
1508 # This group ACL is fast if all evaluated ACLs in the group are fast
1509 # and slow otherwise.
1511 acl aclname all-of acl1 acl2 ...
1512 # match all of the acls [fast or slow]
1513 # The first mismatching ACL stops further ACL evaluation.
1515 # ACLs from multiple all-of lines with the same name are ORed.
1516 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1517 # acl B all-of b1 b2
1518 # acl B all-of b3 b4
1520 # This group ACL is fast if all evaluated ACLs in the group are fast
1521 # and slow otherwise.
1524 acl macaddress arp 09:00:2b:23:45:67
1525 acl myexample dst_as 1241
1526 acl password proxy_auth REQUIRED
1527 acl fileupload req_mime_type -i ^multipart/form-data$
1528 acl javascript rep_mime_type -i ^application/x-javascript$
1532 # Recommended minimum configuration:
1535 # Example rule allowing access from your local networks.
1536 # Adapt to list your (internal) IP networks from where browsing
1538 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1539 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1540 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1541 acl localhet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1542 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1543 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1544 acl localnet src fc00::/7 # RFC 4193 local private network range
1545 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1547 acl SSL_ports port 443
1548 acl Safe_ports port 80 # http
1549 acl Safe_ports port 21 # ftp
1550 acl Safe_ports port 443 # https
1551 acl Safe_ports port 70 # gopher
1552 acl Safe_ports port 210 # wais
1553 acl Safe_ports port 1025-65535 # unregistered ports
1554 acl Safe_ports port 280 # http-mgmt
1555 acl Safe_ports port 488 # gss-http
1556 acl Safe_ports port 591 # filemaker
1557 acl Safe_ports port 777 # multiling http
1561 NAME: proxy_protocol_access
1563 LOC: Config.accessList.proxyProtocol
1565 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1567 Determine which client proxies can be trusted to provide correct
1568 information regarding real client IP address using PROXY protocol.
1570 Requests may pass through a chain of several other proxies
1571 before reaching us. The original source details may by sent in:
1572 * HTTP message Forwarded header, or
1573 * HTTP message X-Forwarded-For header, or
1574 * PROXY protocol connection header.
1576 This directive is solely for validating new PROXY protocol
1577 connections received from a port flagged with require-proxy-header.
1578 It is checked only once after TCP connection setup.
1580 A deny match results in TCP connection closure.
1582 An allow match is required for Squid to permit the corresponding
1583 TCP connection, before Squid even looks for HTTP request headers.
1584 If there is an allow match, Squid starts using PROXY header information
1585 to determine the source address of the connection for all future ACL
1586 checks, logging, etc.
1588 SECURITY CONSIDERATIONS:
1590 Any host from which we accept client IP details can place
1591 incorrect information in the relevant header, and Squid
1592 will use the incorrect information as if it were the
1593 source address of the request. This may enable remote
1594 hosts to bypass any access control restrictions that are
1595 based on the client's source addresses.
1597 This clause only supports fast acl types.
1598 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1601 NAME: follow_x_forwarded_for
1603 IFDEF: FOLLOW_X_FORWARDED_FOR
1604 LOC: Config.accessList.followXFF
1605 DEFAULT_IF_NONE: deny all
1606 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1608 Determine which client proxies can be trusted to provide correct
1609 information regarding real client IP address.
1611 Requests may pass through a chain of several other proxies
1612 before reaching us. The original source details may by sent in:
1613 * HTTP message Forwarded header, or
1614 * HTTP message X-Forwarded-For header, or
1615 * PROXY protocol connection header.
1617 PROXY protocol connections are controlled by the proxy_protocol_access
1618 directive which is checked before this.
1620 If a request reaches us from a source that is allowed by this
1621 directive, then we trust the information it provides regarding
1622 the IP of the client it received from (if any).
1624 For the purpose of ACLs used in this directive the src ACL type always
1625 matches the address we are testing and srcdomain matches its rDNS.
1627 On each HTTP request Squid checks for X-Forwarded-For header fields.
1628 If found the header values are iterated in reverse order and an allow
1629 match is required for Squid to continue on to the next value.
1630 The verification ends when a value receives a deny match, cannot be
1631 tested, or there are no more values to test.
1632 NOTE: Squid does not yet follow the Forwarded HTTP header.
1634 The end result of this process is an IP address that we will
1635 refer to as the indirect client address. This address may
1636 be treated as the client address for access control, ICAP, delay
1637 pools and logging, depending on the acl_uses_indirect_client,
1638 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1639 log_uses_indirect_client and tproxy_uses_indirect_client options.
1641 This clause only supports fast acl types.
1642 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1644 SECURITY CONSIDERATIONS:
1646 Any host from which we accept client IP details can place
1647 incorrect information in the relevant header, and Squid
1648 will use the incorrect information as if it were the
1649 source address of the request. This may enable remote
1650 hosts to bypass any access control restrictions that are
1651 based on the client's source addresses.
1655 acl localhost src 127.0.0.1
1656 acl my_other_proxy srcdomain .proxy.example.com
1657 follow_x_forwarded_for allow localhost
1658 follow_x_forwarded_for allow my_other_proxy
1661 NAME: acl_uses_indirect_client
1664 IFDEF: FOLLOW_X_FORWARDED_FOR
1666 LOC: Config.onoff.acl_uses_indirect_client
1668 Controls whether the indirect client address
1669 (see follow_x_forwarded_for) is used instead of the
1670 direct client address in acl matching.
1672 NOTE: maxconn ACL considers direct TCP links and indirect
1673 clients will always have zero. So no match.
1676 NAME: delay_pool_uses_indirect_client
1679 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1681 LOC: Config.onoff.delay_pool_uses_indirect_client
1683 Controls whether the indirect client address
1684 (see follow_x_forwarded_for) is used instead of the
1685 direct client address in delay pools.
1688 NAME: log_uses_indirect_client
1691 IFDEF: FOLLOW_X_FORWARDED_FOR
1693 LOC: Config.onoff.log_uses_indirect_client
1695 Controls whether the indirect client address
1696 (see follow_x_forwarded_for) is used instead of the
1697 direct client address in the access log.
1700 NAME: tproxy_uses_indirect_client
1703 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1705 LOC: Config.onoff.tproxy_uses_indirect_client
1707 Controls whether the indirect client address
1708 (see follow_x_forwarded_for) is used instead of the
1709 direct client address when spoofing the outgoing client.
1711 This has no effect on requests arriving in non-tproxy
1714 SECURITY WARNING: Usage of this option is dangerous
1715 and should not be used trivially. Correct configuration
1716 of follow_x_forwarded_for with a limited set of trusted
1717 sources is required to prevent abuse of your proxy.
1720 NAME: spoof_client_ip
1722 LOC: Config.accessList.spoof_client_ip
1724 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1726 Control client IP address spoofing of TPROXY traffic based on
1727 defined access lists.
1729 spoof_client_ip allow|deny [!]aclname ...
1731 If there are no "spoof_client_ip" lines present, the default
1732 is to "allow" spoofing of any suitable request.
1734 Note that the cache_peer "no-tproxy" option overrides this ACL.
1736 This clause supports fast acl types.
1737 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1742 LOC: Config.accessList.http
1743 DEFAULT_IF_NONE: deny all
1744 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1746 Allowing or Denying access based on defined access lists
1748 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1749 http_access allow|deny [!]aclname ...
1751 NOTE on default values:
1753 If there are no "access" lines present, the default is to deny
1756 If none of the "access" lines cause a match, the default is the
1757 opposite of the last line in the list. If the last line was
1758 deny, the default is allow. Conversely, if the last line
1759 is allow, the default will be deny. For these reasons, it is a
1760 good idea to have an "deny all" entry at the end of your access
1761 lists to avoid potential confusion.
1763 This clause supports both fast and slow acl types.
1764 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1769 # Recommended minimum Access Permission configuration:
1771 # Deny requests to certain unsafe ports
1772 http_access deny !Safe_ports
1774 # Deny CONNECT to other than secure SSL ports
1775 http_access deny CONNECT !SSL_ports
1777 # Only allow cachemgr access from localhost
1778 http_access allow localhost manager
1779 http_access deny manager
1781 # We strongly recommend the following be uncommented to protect innocent
1782 # web applications running on the proxy server who think the only
1783 # one who can access services on "localhost" is a local user
1784 #http_access deny to_localhost
1787 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1790 # Example rule allowing access from your local networks.
1791 # Adapt localnet in the ACL section to list your (internal) IP networks
1792 # from where browsing should be allowed
1793 http_access allow localnet
1794 http_access allow localhost
1796 # And finally deny all other access to this proxy
1797 http_access deny all
1801 NAME: adapted_http_access http_access2
1803 LOC: Config.accessList.adapted_http
1805 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1807 Allowing or Denying access based on defined access lists
1809 Essentially identical to http_access, but runs after redirectors
1810 and ICAP/eCAP adaptation. Allowing access control based on their
1813 If not set then only http_access is used.
1816 NAME: http_reply_access
1818 LOC: Config.accessList.reply
1820 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1822 Allow replies to client requests. This is complementary to http_access.
1824 http_reply_access allow|deny [!] aclname ...
1826 NOTE: if there are no access lines present, the default is to allow
1829 If none of the access lines cause a match the opposite of the
1830 last line will apply. Thus it is good practice to end the rules
1831 with an "allow all" or "deny all" entry.
1833 This clause supports both fast and slow acl types.
1834 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1839 LOC: Config.accessList.icp
1841 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1843 Allowing or Denying access to the ICP port based on defined
1846 icp_access allow|deny [!]aclname ...
1848 NOTE: The default if no icp_access lines are present is to
1849 deny all traffic. This default may cause problems with peers
1852 This clause only supports fast acl types.
1853 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1855 # Allow ICP queries from local networks only
1856 #icp_access allow localnet
1857 #icp_access deny all
1863 LOC: Config.accessList.htcp
1865 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1867 Allowing or Denying access to the HTCP port based on defined
1870 htcp_access allow|deny [!]aclname ...
1872 See also htcp_clr_access for details on access control for
1873 cache purge (CLR) HTCP messages.
1875 NOTE: The default if no htcp_access lines are present is to
1876 deny all traffic. This default may cause problems with peers
1877 using the htcp option.
1879 This clause only supports fast acl types.
1880 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1882 # Allow HTCP queries from local networks only
1883 #htcp_access allow localnet
1884 #htcp_access deny all
1887 NAME: htcp_clr_access
1890 LOC: Config.accessList.htcp_clr
1892 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1894 Allowing or Denying access to purge content using HTCP based
1895 on defined access lists.
1896 See htcp_access for details on general HTCP access control.
1898 htcp_clr_access allow|deny [!]aclname ...
1900 This clause only supports fast acl types.
1901 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1903 # Allow HTCP CLR requests from trusted peers
1904 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1905 htcp_clr_access allow htcp_clr_peer
1906 htcp_clr_access deny all
1911 LOC: Config.accessList.miss
1913 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1915 Determines whether network access is permitted when satisfying a request.
1918 to force your neighbors to use you as a sibling instead of
1921 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1922 miss_access deny !localclients
1923 miss_access allow all
1925 This means only your local clients are allowed to fetch relayed/MISS
1926 replies from the network and all other clients can only fetch cached
1929 The default for this setting allows all clients who passed the
1930 http_access rules to relay via this proxy.
1932 This clause only supports fast acl types.
1933 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1936 NAME: ident_lookup_access
1940 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1941 LOC: Ident::TheConfig.identLookup
1943 A list of ACL elements which, if matched, cause an ident
1944 (RFC 931) lookup to be performed for this request. For
1945 example, you might choose to always perform ident lookups
1946 for your main multi-user Unix boxes, but not for your Macs
1947 and PCs. By default, ident lookups are not performed for
1950 To enable ident lookups for specific client addresses, you
1951 can follow this example:
1953 acl ident_aware_hosts src 198.168.1.0/24
1954 ident_lookup_access allow ident_aware_hosts
1955 ident_lookup_access deny all
1957 Only src type ACL checks are fully supported. A srcdomain
1958 ACL might work at times, but it will not always provide
1961 This clause only supports fast acl types.
1962 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1965 NAME: reply_body_max_size
1966 COMMENT: size [acl acl...]
1969 DEFAULT_DOC: No limit is applied.
1970 LOC: Config.ReplyBodySize
1972 This option specifies the maximum size of a reply body. It can be
1973 used to prevent users from downloading very large files, such as
1974 MP3's and movies. When the reply headers are received, the
1975 reply_body_max_size lines are processed, and the first line where
1976 all (if any) listed ACLs are true is used as the maximum body size
1979 This size is checked twice. First when we get the reply headers,
1980 we check the content-length value. If the content length value exists
1981 and is larger than the allowed size, the request is denied and the
1982 user receives an error message that says "the request or reply
1983 is too large." If there is no content-length, and the reply
1984 size exceeds this limit, the client's connection is just closed
1985 and they will receive a partial reply.
1987 WARNING: downstream caches probably can not detect a partial reply
1988 if there is no content-length header, so they will cache
1989 partial responses and give them out as hits. You should NOT
1990 use this option if you have downstream caches.
1992 WARNING: A maximum size smaller than the size of squid's error messages
1993 will cause an infinite loop and crash squid. Ensure that the smallest
1994 non-zero value you use is greater that the maximum header size plus
1995 the size of your largest error page.
1997 If you set this parameter none (the default), there will be
2000 Configuration Format is:
2001 reply_body_max_size SIZE UNITS [acl ...]
2003 reply_body_max_size 10 MB
2007 NAME: on_unsupported_protocol
2008 TYPE: on_unsupported_protocol
2009 LOC: Config.accessList.on_unsupported_protocol
2011 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
2013 Determines Squid behavior when encountering strange requests at the
2014 beginning of an accepted TCP connection or the beginning of a bumped
2015 CONNECT tunnel. Controlling Squid reaction to unexpected traffic is
2016 especially useful in interception environments where Squid is likely
2017 to see connections for unsupported protocols that Squid should either
2018 terminate or tunnel at TCP level.
2020 on_unsupported_protocol <action> [!]acl ...
2022 The first matching action wins. Only fast ACLs are supported.
2024 Supported actions are:
2026 tunnel: Establish a TCP connection with the intended server and
2027 blindly shovel TCP packets between the client and server.
2029 respond: Respond with an error message, using the transfer protocol
2030 for the Squid port that received the request (e.g., HTTP
2031 for connections intercepted at the http_port). This is the
2034 Squid expects the following traffic patterns:
2036 http_port: a plain HTTP request
2037 https_port: SSL/TLS handshake followed by an [encrypted] HTTP request
2038 ftp_port: a plain FTP command (no on_unsupported_protocol support yet!)
2039 CONNECT tunnel on http_port: same as https_port
2040 CONNECT tunnel on https_port: same as https_port
2042 Currently, this directive has effect on intercepted connections and
2043 bumped tunnels only. Other cases are not supported because Squid
2044 cannot know the intended destination of other traffic.
2047 # define what Squid errors indicate receiving non-HTTP traffic:
2048 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
2049 # define what Squid errors indicate receiving nothing:
2050 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
2051 # tunnel everything that does not look like HTTP:
2052 on_unsupported_protocol tunnel foreignProtocol
2053 # tunnel if we think the client waits for the server to talk first:
2054 on_unsupported_protocol tunnel serverTalksFirstProtocol
2055 # in all other error cases, just send an HTTP "error page" response:
2056 on_unsupported_protocol respond all
2058 See also: squid_error ACL
2064 LOC: Auth::TheConfig.schemeAccess
2066 DEFAULT_DOC: use all auth_param schemes in their configuration order
2068 Use this directive to customize authentication schemes presence and
2069 order in Squid's Unauthorized and Authentication Required responses.
2071 auth_schemes scheme1,scheme2,... [!]aclname ...
2073 where schemeN is the name of one of the authentication schemes
2074 configured using auth_param directives. At least one scheme name is
2075 required. Multiple scheme names are separated by commas. Either
2076 avoid whitespace or quote the entire schemes list.
2078 A special "ALL" scheme name expands to all auth_param-configured
2079 schemes in their configuration order. This directive cannot be used
2080 to configure Squid to offer no authentication schemes at all.
2082 The first matching auth_schemes rule determines the schemes order
2083 for the current Authentication Required transaction. Note that the
2084 future response is not yet available during auth_schemes evaluation.
2086 If this directive is not used or none of its rules match, then Squid
2087 responds with all configured authentication schemes in the order of
2088 auth_param directives in the configuration file.
2090 This directive does not determine when authentication is used or
2091 how each authentication scheme authenticates clients.
2093 The following example sends basic and negotiate authentication
2094 schemes, in that order, when requesting authentication of HTTP
2095 requests matching the isIE ACL (not shown) while sending all
2096 auth_param schemes in their configuration order to other clients:
2098 auth_schemes basic,negotiate isIE
2099 auth_schemes ALL all # explicit default
2101 This directive supports fast ACLs only.
2103 See also: auth_param.
2108 -----------------------------------------------------------------------------
2111 NAME: http_port ascii_port
2116 Usage: port [mode] [options]
2117 hostname:port [mode] [options]
2118 1.2.3.4:port [mode] [options]
2120 The socket addresses where Squid will listen for HTTP client
2121 requests. You may specify multiple socket addresses.
2122 There are three forms: port alone, hostname with port, and
2123 IP address with port. If you specify a hostname or IP
2124 address, Squid binds the socket to that specific
2125 address. Most likely, you do not need to bind to a specific
2126 address, so you can use the port number alone.
2128 If you are running Squid in accelerator mode, you
2129 probably want to listen on port 80 also, or instead.
2131 The -a command line option may be used to specify additional
2132 port(s) where Squid listens for proxy request. Such ports will
2133 be plain proxy ports with no options.
2135 You may specify multiple socket addresses on multiple lines.
2139 intercept Support for IP-Layer NAT interception delivering
2140 traffic to this Squid port.
2141 NP: disables authentication on the port.
2143 tproxy Support Linux TPROXY (or BSD divert-to) with spoofing
2144 of outgoing connections using the client IP address.
2145 NP: disables authentication on the port.
2147 accel Accelerator / reverse proxy mode
2149 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
2150 establish secure connection with the client and with
2151 the server, decrypt HTTPS messages as they pass through
2152 Squid, and treat them as unencrypted HTTP messages,
2153 becoming the man-in-the-middle.
2155 The ssl_bump option is required to fully enable
2156 bumping of CONNECT requests.
2158 Omitting the mode flag causes default forward proxy mode to be used.
2161 Accelerator Mode Options:
2163 defaultsite=domainname
2164 What to use for the Host: header if it is not present
2165 in a request. Determines what site (not origin server)
2166 accelerators should consider the default.
2168 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
2170 protocol= Protocol to reconstruct accelerated and intercepted
2171 requests with. Defaults to HTTP/1.1 for http_port and
2172 HTTPS/1.1 for https_port.
2173 When an unsupported value is configured Squid will
2174 produce a FATAL error.
2175 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
2177 vport Virtual host port support. Using the http_port number
2178 instead of the port passed on Host: headers.
2180 vport=NN Virtual host port support. Using the specified port
2181 number instead of the port passed on Host: headers.
2184 Act as if this Squid is the origin server.
2185 This currently means generate new Date: and Expires:
2186 headers on HIT instead of adding Age:.
2188 ignore-cc Ignore request Cache-Control headers.
2190 WARNING: This option violates HTTP specifications if
2191 used in non-accelerator setups.
2193 allow-direct Allow direct forwarding in accelerator mode. Normally
2194 accelerated requests are denied direct forwarding as if
2195 never_direct was used.
2197 WARNING: this option opens accelerator mode to security
2198 vulnerabilities usually only affecting in interception
2199 mode. Make sure to protect forwarding with suitable
2200 http_access rules when using this.
2203 SSL Bump Mode Options:
2204 In addition to these options ssl-bump requires TLS/SSL options.
2206 generate-host-certificates[=<on|off>]
2207 Dynamically create SSL server certificates for the
2208 destination hosts of bumped CONNECT requests.When
2209 enabled, the cert and key options are used to sign
2210 generated certificates. Otherwise generated
2211 certificate will be selfsigned.
2212 If there is a CA certificate lifetime of the generated
2213 certificate equals lifetime of the CA certificate. If
2214 generated certificate is selfsigned lifetime is three
2216 This option is enabled by default when ssl-bump is used.
2217 See the ssl-bump option above for more information.
2219 dynamic_cert_mem_cache_size=SIZE
2220 Approximate total RAM size spent on cached generated
2221 certificates. If set to zero, caching is disabled. The
2222 default value is 4MB.
2226 cert= Path to SSL certificate (PEM format).
2228 key= Path to SSL private key file (PEM format)
2229 if not specified, the certificate file is
2230 assumed to be a combined certificate and
2233 cipher= Colon separated list of supported ciphers.
2234 NOTE: some ciphers such as EDH ciphers depend on
2235 additional settings. If those settings are
2236 omitted the ciphers may be silently ignored
2237 by the OpenSSL library.
2239 options= Various SSL implementation options. The most important
2242 NO_SSLv3 Disallow the use of SSLv3
2244 NO_TLSv1 Disallow the use of TLSv1.0
2246 NO_TLSv1_1 Disallow the use of TLSv1.1
2248 NO_TLSv1_2 Disallow the use of TLSv1.2
2251 Always create a new key when using
2252 temporary/ephemeral DH key exchanges
2255 Enable ephemeral ECDH key exchange.
2256 The adopted curve should be specified
2257 using the tls-dh option.
2260 Disable use of RFC5077 session tickets.
2261 Some servers may have problems
2262 understanding the TLS extension due
2263 to ambiguous specification in RFC4507.
2265 ALL Enable various bug workarounds
2266 suggested as "harmless" by OpenSSL
2267 Be warned that this reduces SSL/TLS
2268 strength to some attacks.
2270 See the OpenSSL SSL_CTX_set_options documentation for a
2273 clientca= File containing the list of CAs to use when
2274 requesting a client certificate.
2276 tls-cafile= PEM file containing CA certificates to use when verifying
2277 client certificates. If not configured clientca will be
2278 used. May be repeated to load multiple files.
2280 capath= Directory containing additional CA certificates
2281 and CRL lists to use when verifying client certificates.
2282 Requires OpenSSL or LibreSSL.
2284 crlfile= File of additional CRL lists to use when verifying
2285 the client certificate, in addition to CRLs stored in
2286 the capath. Implies VERIFY_CRL flag below.
2289 File containing DH parameters for temporary/ephemeral DH key
2290 exchanges, optionally prefixed by a curve for ephemeral ECDH
2292 See OpenSSL documentation for details on how to create the
2293 DH parameter file. Supported curves for ECDH can be listed
2294 using the "openssl ecparam -list_curves" command.
2295 WARNING: EDH and EECDH ciphers will be silently disabled if
2296 this option is not set.
2298 sslflags= Various flags modifying the use of SSL:
2300 Don't request client certificates
2301 immediately, but wait until acl processing
2302 requires a certificate (not yet implemented).
2304 Don't allow for session reuse. Each connection
2305 will result in a new SSL session.
2307 Verify CRL lists when accepting client
2310 Verify CRL lists for all certificates in the
2311 client certificate chain.
2313 tls-default-ca[=off]
2314 Whether to use the system Trusted CAs. Default is OFF.
2316 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
2318 sslcontext= SSL session ID context identifier.
2322 connection-auth[=on|off]
2323 use connection-auth=off to tell Squid to prevent
2324 forwarding Microsoft connection oriented authentication
2325 (NTLM, Negotiate and Kerberos)
2327 disable-pmtu-discovery=
2328 Control Path-MTU discovery usage:
2329 off lets OS decide on what to do (default).
2330 transparent disable PMTU discovery when transparent
2332 always disable always PMTU discovery.
2334 In many setups of transparently intercepting proxies
2335 Path-MTU discovery can not work on traffic towards the
2336 clients. This is the case when the intercepting device
2337 does not fully track connections and fails to forward
2338 ICMP must fragment messages to the cache server. If you
2339 have such setup and experience that certain clients
2340 sporadically hang or never complete requests set
2341 disable-pmtu-discovery option to 'transparent'.
2343 name= Specifies a internal name for the port. Defaults to
2344 the port specification (port or addr:port)
2346 tcpkeepalive[=idle,interval,timeout]
2347 Enable TCP keepalive probes of idle connections.
2348 In seconds; idle is the initial time before TCP starts
2349 probing the connection, interval how often to probe, and
2350 timeout the time before giving up.
2352 require-proxy-header
2353 Require PROXY protocol version 1 or 2 connections.
2354 The proxy_protocol_access is required to whitelist
2355 downstream proxies which can be trusted.
2357 If you run Squid on a dual-homed machine with an internal
2358 and an external interface we recommend you to specify the
2359 internal address:port in http_port. This way Squid will only be
2360 visible on the internal address.
2364 # Squid normally listens to port 3128
2365 http_port @DEFAULT_HTTP_PORT@
2370 IFDEF: USE_GNUTLS||USE_OPENSSL
2375 Usage: [ip:]port [mode] cert=certificate.pem [options]
2377 The socket address where Squid will listen for client requests made
2378 over TLS or SSL connections. Commonly referred to as HTTPS.
2380 This is most useful for situations where you are running squid in
2381 accelerator mode and you want to do the TLS work at the accelerator level.
2383 You may specify multiple socket addresses on multiple lines,
2384 each with their own certificate and/or options.
2386 The TLS cert= option is mandatory on HTTPS ports.
2388 See http_port for a list of modes and options.
2396 Enables Native FTP proxy by specifying the socket address where Squid
2397 listens for FTP client requests. See http_port directive for various
2398 ways to specify the listening address and mode.
2400 Usage: ftp_port address [mode] [options]
2402 WARNING: This is a new, experimental, complex feature that has seen
2403 limited production exposure. Some Squid modules (e.g., caching) do not
2404 currently work with native FTP proxying, and many features have not
2405 even been tested for compatibility. Test well before deploying!
2407 Native FTP proxying differs substantially from proxying HTTP requests
2408 with ftp:// URIs because Squid works as an FTP server and receives
2409 actual FTP commands (rather than HTTP requests with FTP URLs).
2411 Native FTP commands accepted at ftp_port are internally converted or
2412 wrapped into HTTP-like messages. The same happens to Native FTP
2413 responses received from FTP origin servers. Those HTTP-like messages
2414 are shoveled through regular access control and adaptation layers
2415 between the FTP client and the FTP origin server. This allows Squid to
2416 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2417 mechanisms when shoveling wrapped FTP messages. For example,
2418 http_access and adaptation_access directives are used.
2422 intercept Same as http_port intercept. The FTP origin address is
2423 determined based on the intended destination of the
2424 intercepted connection.
2426 tproxy Support Linux TPROXY for spoofing outgoing
2427 connections using the client IP address.
2428 NP: disables authentication and maybe IPv6 on the port.
2430 By default (i.e., without an explicit mode option), Squid extracts the
2431 FTP origin address from the login@origin parameter of the FTP USER
2432 command. Many popular FTP clients support such native FTP proxying.
2436 name=token Specifies an internal name for the port. Defaults to
2437 the port address. Usable with myportname ACL.
2440 Enables tracking of FTP directories by injecting extra
2441 PWD commands and adjusting Request-URI (in wrapping
2442 HTTP requests) to reflect the current FTP server
2443 directory. Tracking is disabled by default.
2445 protocol=FTP Protocol to reconstruct accelerated and intercepted
2446 requests with. Defaults to FTP. No other accepted
2447 values have been tested with. An unsupported value
2448 results in a FATAL error. Accepted values are FTP,
2449 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2451 Other http_port modes and options that are not specific to HTTP and
2452 HTTPS may also work.
2455 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2458 LOC: Ip::Qos::TheConfig.tosToServer
2460 Allows you to select a TOS/Diffserv value for packets outgoing
2461 on the server side, based on an ACL.
2463 tcp_outgoing_tos ds-field [!]aclname ...
2465 Example where normal_service_net uses the TOS value 0x00
2466 and good_service_net uses 0x20
2468 acl normal_service_net src 10.0.0.0/24
2469 acl good_service_net src 10.0.1.0/24
2470 tcp_outgoing_tos 0x00 normal_service_net
2471 tcp_outgoing_tos 0x20 good_service_net
2473 TOS/DSCP values really only have local significance - so you should
2474 know what you're specifying. For more information, see RFC2474,
2475 RFC2475, and RFC3260.
2477 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2478 "default" to use whatever default your host has.
2479 Note that only multiples of 4 are usable as the two rightmost bits have
2480 been redefined for use by ECN (RFC 3168 section 23.1).
2481 The squid parser will enforce this by masking away the ECN bits.
2483 Processing proceeds in the order specified, and stops at first fully
2486 Only fast ACLs are supported.
2489 NAME: clientside_tos
2492 LOC: Ip::Qos::TheConfig.tosToClient
2494 Allows you to select a TOS/DSCP value for packets being transmitted
2495 on the client-side, based on an ACL.
2497 clientside_tos ds-field [!]aclname ...
2499 Example where normal_service_net uses the TOS value 0x00
2500 and good_service_net uses 0x20
2502 acl normal_service_net src 10.0.0.0/24
2503 acl good_service_net src 10.0.1.0/24
2504 clientside_tos 0x00 normal_service_net
2505 clientside_tos 0x20 good_service_net
2507 Note: This feature is incompatible with qos_flows. Any TOS values set here
2508 will be overwritten by TOS values in qos_flows.
2510 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2511 "default" to use whatever default your host has.
2512 Note that only multiples of 4 are usable as the two rightmost bits have
2513 been redefined for use by ECN (RFC 3168 section 23.1).
2514 The squid parser will enforce this by masking away the ECN bits.
2516 This clause only supports fast acl types.
2517 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2520 NAME: tcp_outgoing_mark
2522 IFDEF: SO_MARK&&USE_LIBCAP
2524 LOC: Ip::Qos::TheConfig.nfmarkToServer
2526 Allows you to apply a Netfilter mark value to outgoing packets
2527 on the server side, based on an ACL.
2529 tcp_outgoing_mark mark-value [!]aclname ...
2531 Example where normal_service_net uses the mark value 0x00
2532 and good_service_net uses 0x20
2534 acl normal_service_net src 10.0.0.0/24
2535 acl good_service_net src 10.0.1.0/24
2536 tcp_outgoing_mark 0x00 normal_service_net
2537 tcp_outgoing_mark 0x20 good_service_net
2539 Only fast ACLs are supported.
2542 NAME: clientside_mark
2544 IFDEF: SO_MARK&&USE_LIBCAP
2546 LOC: Ip::Qos::TheConfig.nfmarkToClient
2548 Allows you to apply a Netfilter mark value to packets being transmitted
2549 on the client-side, based on an ACL.
2551 clientside_mark mark-value [!]aclname ...
2553 Example where normal_service_net uses the mark value 0x00
2554 and good_service_net uses 0x20
2556 acl normal_service_net src 10.0.0.0/24
2557 acl good_service_net src 10.0.1.0/24
2558 clientside_mark 0x00 normal_service_net
2559 clientside_mark 0x20 good_service_net
2561 Note: This feature is incompatible with qos_flows. Any mark values set here
2562 will be overwritten by mark values in qos_flows.
2564 This clause only supports fast acl types.
2565 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2572 LOC: Ip::Qos::TheConfig
2574 Allows you to select a TOS/DSCP value to mark outgoing
2575 connections to the client, based on where the reply was sourced.
2576 For platforms using netfilter, allows you to set a netfilter mark
2577 value instead of, or in addition to, a TOS value.
2579 By default this functionality is disabled. To enable it with the default
2580 settings simply use "qos_flows mark" or "qos_flows tos". Default
2581 settings will result in the netfilter mark or TOS value being copied
2582 from the upstream connection to the client. Note that it is the connection
2583 CONNMARK value not the packet MARK value that is copied.
2585 It is not currently possible to copy the mark or TOS value from the
2586 client to the upstream connection request.
2588 TOS values really only have local significance - so you should
2589 know what you're specifying. For more information, see RFC2474,
2590 RFC2475, and RFC3260.
2592 The TOS/DSCP byte must be exactly that - a octet value 0 - 255.
2593 Note that only multiples of 4 are usable as the two rightmost bits have
2594 been redefined for use by ECN (RFC 3168 section 23.1).
2595 The squid parser will enforce this by masking away the ECN bits.
2597 Mark values can be any unsigned 32-bit integer value.
2599 This setting is configured by setting the following values:
2601 tos|mark Whether to set TOS or netfilter mark values
2603 local-hit=0xFF Value to mark local cache hits.
2605 sibling-hit=0xFF Value to mark hits from sibling peers.
2607 parent-hit=0xFF Value to mark hits from parent peers.
2609 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2610 over the preserve-miss feature (see below), unless
2611 mask is specified, in which case only the bits
2612 specified in the mask are written.
2614 The TOS variant of the following features are only possible on Linux
2615 and require your kernel to be patched with the TOS preserving ZPH
2616 patch, available from http://zph.bratcheda.org
2617 No patch is needed to preserve the netfilter mark, which will work
2618 with all variants of netfilter.
2620 disable-preserve-miss
2621 This option disables the preservation of the TOS or netfilter
2622 mark. By default, the existing TOS or netfilter mark value of
2623 the response coming from the remote server will be retained
2624 and masked with miss-mark.
2625 NOTE: in the case of a netfilter mark, the mark must be set on
2626 the connection (using the CONNMARK target) not on the packet
2630 Allows you to mask certain bits in the TOS or mark value
2631 received from the remote server, before copying the value to
2632 the TOS sent towards clients.
2633 Default for tos: 0xFF (TOS from server is not changed).
2634 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2636 All of these features require the --enable-zph-qos compilation flag
2637 (enabled by default). Netfilter marking also requires the
2638 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2639 libcap 2.09+ (--with-libcap).
2643 NAME: tcp_outgoing_address
2646 DEFAULT_DOC: Address selection is performed by the operating system.
2647 LOC: Config.accessList.outgoing_address
2649 Allows you to map requests to different outgoing IP addresses
2650 based on the username or source address of the user making
2653 tcp_outgoing_address ipaddr [[!]aclname] ...
2656 Forwarding clients with dedicated IPs for certain subnets.
2658 acl normal_service_net src 10.0.0.0/24
2659 acl good_service_net src 10.0.2.0/24
2661 tcp_outgoing_address 2001:db8::c001 good_service_net
2662 tcp_outgoing_address 10.1.0.2 good_service_net
2664 tcp_outgoing_address 2001:db8::beef normal_service_net
2665 tcp_outgoing_address 10.1.0.1 normal_service_net
2667 tcp_outgoing_address 2001:db8::1
2668 tcp_outgoing_address 10.1.0.3
2670 Processing proceeds in the order specified, and stops at first fully
2673 Squid will add an implicit IP version test to each line.
2674 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2675 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2678 NOTE: The use of this directive using client dependent ACLs is
2679 incompatible with the use of server side persistent connections. To
2680 ensure correct results it is best to set server_persistent_connections
2681 to off when using this directive in such configurations.
2683 NOTE: The use of this directive to set a local IP on outgoing TCP links
2684 is incompatible with using TPROXY to set client IP out outbound TCP links.
2685 When needing to contact peers use the no-tproxy cache_peer option and the
2686 client_dst_passthru directive re-enable normal forwarding such as this.
2688 This clause only supports fast acl types.
2689 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2692 NAME: host_verify_strict
2695 LOC: Config.onoff.hostStrictVerify
2697 Regardless of this option setting, when dealing with intercepted
2698 traffic, Squid always verifies that the destination IP address matches
2699 the Host header domain or IP (called 'authority form URL').
2701 This enforcement is performed to satisfy a MUST-level requirement in
2702 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2703 authority of the origin server or gateway given by the original URL".
2706 Squid always responds with an HTTP 409 (Conflict) error
2707 page and logs a security warning if there is no match.
2709 Squid verifies that the destination IP address matches
2710 the Host header for forward-proxy and reverse-proxy traffic
2711 as well. For those traffic types, Squid also enables the
2712 following checks, comparing the corresponding Host header
2713 and Request-URI components:
2715 * The host names (domain or IP) must be identical,
2716 but valueless or missing Host header disables all checks.
2717 For the two host names to match, both must be either IP
2720 * Port numbers must be identical, but if a port is missing
2721 the scheme-default port is assumed.
2724 When set to OFF (the default):
2725 Squid allows suspicious requests to continue but logs a
2726 security warning and blocks caching of the response.
2728 * Forward-proxy traffic is not checked at all.
2730 * Reverse-proxy traffic is not checked at all.
2732 * Intercepted traffic which passes verification is handled
2733 according to client_dst_passthru.
2735 * Intercepted requests which fail verification are sent
2736 to the client original destination instead of DIRECT.
2737 This overrides 'client_dst_passthru off'.
2739 For now suspicious intercepted CONNECT requests are always
2740 responded to with an HTTP 409 (Conflict) error page.
2745 As described in CVE-2009-0801 when the Host: header alone is used
2746 to determine the destination of a request it becomes trivial for
2747 malicious scripts on remote websites to bypass browser same-origin
2748 security policy and sandboxing protections.
2750 The cause of this is that such applets are allowed to perform their
2751 own HTTP stack, in which case the same-origin policy of the browser
2752 sandbox only verifies that the applet tries to contact the same IP
2753 as from where it was loaded at the IP level. The Host: header may
2754 be different from the connected IP and approved origin.
2758 NAME: client_dst_passthru
2761 LOC: Config.onoff.client_dst_passthru
2763 With NAT or TPROXY intercepted traffic Squid may pass the request
2764 directly to the original client destination IP or seek a faster
2765 source using the HTTP Host header.
2767 Using Host to locate alternative servers can provide faster
2768 connectivity with a range of failure recovery options.
2769 But can also lead to connectivity trouble when the client and
2770 server are attempting stateful interactions unaware of the proxy.
2772 This option (on by default) prevents alternative DNS entries being
2773 located to send intercepted traffic DIRECT to an origin server.
2774 The clients original destination IP and port will be used instead.
2776 Regardless of this option setting, when dealing with intercepted
2777 traffic Squid will verify the Host: header and any traffic which
2778 fails Host verification will be treated as if this option were ON.
2780 see host_verify_strict for details on the verification process.
2785 -----------------------------------------------------------------------------
2788 NAME: tls_outgoing_options
2789 IFDEF: USE_GNUTLS||USE_OPENSSL
2790 TYPE: securePeerOptions
2791 DEFAULT: min-version=1.0
2792 LOC: Security::ProxyOutgoingConfig
2794 disable Do not support https:// URLs.
2796 cert=/path/to/client/certificate
2797 A client TLS certificate to use when connecting.
2799 key=/path/to/client/private_key
2800 The private TLS key corresponding to the cert= above.
2801 If key= is not specified cert= is assumed to reference
2802 a PEM file containing both the certificate and the key.
2804 cipher=... The list of valid TLS ciphers to use.
2807 The minimum TLS protocol version to permit.
2808 To control SSLv3 use the options= parameter.
2809 Supported Values: 1.0 (default), 1.1, 1.2
2811 options=... Specify various TLS/SSL implementation options.
2813 OpenSSL options most important are:
2815 NO_SSLv3 Disallow the use of SSLv3
2818 Always create a new key when using
2819 temporary/ephemeral DH key exchanges
2822 Disable use of RFC5077 session tickets.
2823 Some servers may have problems
2824 understanding the TLS extension due
2825 to ambiguous specification in RFC4507.
2827 ALL Enable various bug workarounds
2828 suggested as "harmless" by OpenSSL
2829 Be warned that this reduces SSL/TLS
2830 strength to some attacks.
2832 See the OpenSSL SSL_CTX_set_options documentation
2833 for a more complete list.
2835 GnuTLS options most important are:
2838 Disable use of RFC5077 session tickets.
2839 Some servers may have problems
2840 understanding the TLS extension due
2841 to ambiguous specification in RFC4507.
2843 See the GnuTLS Priority Strings documentation
2844 for a more complete list.
2845 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
2848 cafile= PEM file containing CA certificates to use when verifying
2849 the peer certificate. May be repeated to load multiple files.
2851 capath= A directory containing additional CA certificates to
2852 use when verifying the peer certificate.
2853 Requires OpenSSL or LibreSSL.
2855 crlfile=... A certificate revocation list file to use when
2856 verifying the peer certificate.
2858 flags=... Specify various flags modifying the TLS implementation:
2861 Accept certificates even if they fail to
2864 Don't verify the peer certificate
2865 matches the server name
2868 Whether to use the system Trusted CAs. Default is ON.
2870 domain= The peer name as advertised in its certificate.
2871 Used for verifying the correctness of the received peer
2872 certificate. If not specified the peer hostname will be
2878 -----------------------------------------------------------------------------
2881 NAME: ssl_unclean_shutdown
2885 LOC: Config.SSL.unclean_shutdown
2887 Some browsers (especially MSIE) bugs out on SSL shutdown
2894 LOC: Config.SSL.ssl_engine
2897 The OpenSSL engine to use. You will need to set this if you
2898 would like to use hardware SSL acceleration for example.
2901 NAME: sslproxy_session_ttl
2904 LOC: Config.SSL.session_ttl
2907 Sets the timeout value for SSL sessions
2910 NAME: sslproxy_session_cache_size
2913 LOC: Config.SSL.sessionCacheSize
2916 Sets the cache size to use for ssl session
2919 NAME: sslproxy_foreign_intermediate_certs
2922 LOC: Config.ssl_client.foreignIntermediateCertsPath
2925 Many origin servers fail to send their full server certificate
2926 chain for verification, assuming the client already has or can
2927 easily locate any missing intermediate certificates.
2929 Squid uses the certificates from the specified file to fill in
2930 these missing chains when trying to validate origin server
2933 The file is expected to contain zero or more PEM-encoded
2934 intermediate certificates. These certificates are not treated
2935 as trusted root certificates, and any self-signed certificate in
2936 this file will be ignored.
2939 NAME: sslproxy_cert_sign_hash
2942 LOC: Config.SSL.certSignHash
2945 Sets the hashing algorithm to use when signing generated certificates.
2946 Valid algorithm names depend on the OpenSSL library used. The following
2947 names are usually available: sha1, sha256, sha512, and md5. Please see
2948 your OpenSSL library manual for the available hashes. By default, Squids
2949 that support this option use sha256 hashes.
2951 Squid does not forcefully purge cached certificates that were generated
2952 with an algorithm other than the currently configured one. They remain
2953 in the cache, subject to the regular cache eviction policy, and become
2954 useful if the algorithm changes again.
2959 TYPE: sslproxy_ssl_bump
2960 LOC: Config.accessList.ssl_bump
2961 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
2964 This option is consulted when a CONNECT request is received on
2965 an http_port (or a new connection is intercepted at an
2966 https_port), provided that port was configured with an ssl-bump
2967 flag. The subsequent data on the connection is either treated as
2968 HTTPS and decrypted OR tunneled at TCP level without decryption,
2969 depending on the first matching bumping "action".
2971 ssl_bump <action> [!]acl ...
2973 The following bumping actions are currently supported:
2976 Become a TCP tunnel without decrypting proxied traffic.
2977 This is the default action.
2980 When used on step SslBump1, establishes a secure connection
2981 with the client first, then connect to the server.
2982 When used on step SslBump2 or SslBump3, establishes a secure
2983 connection with the server and, using a mimicked server
2984 certificate, with the client.
2987 Receive client (step SslBump1) or server (step SslBump2)
2988 certificate while preserving the possibility of splicing the
2989 connection. Peeking at the server certificate (during step 2)
2990 usually precludes bumping of the connection at step 3.
2993 Receive client (step SslBump1) or server (step SslBump2)
2994 certificate while preserving the possibility of bumping the
2995 connection. Staring at the server certificate (during step 2)
2996 usually precludes splicing of the connection at step 3.
2999 Close client and server connections.
3001 Backward compatibility actions available at step SslBump1:
3004 Bump the connection. Establish a secure connection with the
3005 client first, then connect to the server. This old mode does
3006 not allow Squid to mimic server SSL certificate and does not
3007 work with intercepted SSL connections.
3010 Bump the connection. Establish a secure connection with the
3011 server first, then establish a secure connection with the
3012 client, using a mimicked server certificate. Works with both
3013 CONNECT requests and intercepted SSL connections, but does
3014 not allow to make decisions based on SSL handshake info.
3017 Decide whether to bump or splice the connection based on
3018 client-to-squid and server-to-squid SSL hello messages.
3022 Same as the "splice" action.
3024 All ssl_bump rules are evaluated at each of the supported bumping
3025 steps. Rules with actions that are impossible at the current step are
3026 ignored. The first matching ssl_bump action wins and is applied at the
3027 end of the current step. If no rules match, the splice action is used.
3028 See the at_step ACL for a list of the supported SslBump steps.
3030 This clause supports both fast and slow acl types.
3031 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3033 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
3036 # Example: Bump all TLS connections except those originating from
3037 # localhost or those going to example.com.
3039 acl broken_sites ssl::server_name .example.com
3040 ssl_bump splice localhost
3041 ssl_bump splice broken_sites
3045 NAME: sslproxy_cert_error
3048 DEFAULT_DOC: Server certificate errors terminate the transaction.
3049 LOC: Config.ssl_client.cert_error
3052 Use this ACL to bypass server certificate validation errors.
3054 For example, the following lines will bypass all validation errors
3055 when talking to servers for example.com. All other
3056 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
3058 acl BrokenButTrustedServers dstdomain example.com
3059 sslproxy_cert_error allow BrokenButTrustedServers
3060 sslproxy_cert_error deny all
3062 This clause only supports fast acl types.
3063 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3064 Using slow acl types may result in server crashes
3066 Without this option, all server certificate validation errors
3067 terminate the transaction to protect Squid and the client.
3069 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
3070 but should not happen unless your OpenSSL library is buggy.
3073 Bypassing validation errors is dangerous because an
3074 error usually implies that the server cannot be trusted
3075 and the connection may be insecure.
3077 See also: sslproxy_flags and DONT_VERIFY_PEER.
3080 NAME: sslproxy_cert_sign
3083 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
3084 POSTSCRIPTUM: signSelf ssl::certSelfSigned
3085 POSTSCRIPTUM: signTrusted all
3086 TYPE: sslproxy_cert_sign
3087 LOC: Config.ssl_client.cert_sign
3090 sslproxy_cert_sign <signing algorithm> acl ...
3092 The following certificate signing algorithms are supported:
3095 Sign using the configured CA certificate which is usually
3096 placed in and trusted by end-user browsers. This is the
3097 default for trusted origin server certificates.
3100 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
3101 This is the default for untrusted origin server certificates
3102 that are not self-signed (see ssl::certUntrusted).
3105 Sign using a self-signed certificate with the right CN to
3106 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
3107 browser. This is the default for self-signed origin server
3108 certificates (see ssl::certSelfSigned).
3110 This clause only supports fast acl types.
3112 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
3113 signing algorithm to generate the certificate and ignores all
3114 subsequent sslproxy_cert_sign options (the first match wins). If no
3115 acl(s) match, the default signing algorithm is determined by errors
3116 detected when obtaining and validating the origin server certificate.
3118 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3119 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3120 CONNECT request that carries a domain name. In all other cases (CONNECT
3121 to an IP address or an intercepted SSL connection), Squid cannot detect
3122 the domain mismatch at certificate generation time when
3123 bump-server-first is used.
3126 NAME: sslproxy_cert_adapt
3129 TYPE: sslproxy_cert_adapt
3130 LOC: Config.ssl_client.cert_adapt
3133 sslproxy_cert_adapt <adaptation algorithm> acl ...
3135 The following certificate adaptation algorithms are supported:
3138 Sets the "Not After" property to the "Not After" property of
3139 the CA certificate used to sign generated certificates.
3142 Sets the "Not Before" property to the "Not Before" property of
3143 the CA certificate used to sign generated certificates.
3145 setCommonName or setCommonName{CN}
3146 Sets Subject.CN property to the host name specified as a
3147 CN parameter or, if no explicit CN parameter was specified,
3148 extracted from the CONNECT request. It is a misconfiguration
3149 to use setCommonName without an explicit parameter for
3150 intercepted or tproxied SSL connections.
3152 This clause only supports fast acl types.
3154 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
3155 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
3156 corresponding adaptation algorithm to generate the certificate and
3157 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
3158 group (i.e., the first match wins within each algorithm group). If no
3159 acl(s) match, the default mimicking action takes place.
3161 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
3162 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
3163 CONNECT request that carries a domain name. In all other cases (CONNECT
3164 to an IP address or an intercepted SSL connection), Squid cannot detect
3165 the domain mismatch at certificate generation time when
3166 bump-server-first is used.
3169 NAME: sslpassword_program
3172 LOC: Config.Program.ssl_password
3175 Specify a program used for entering SSL key passphrases
3176 when using encrypted SSL certificate keys. If not specified
3177 keys must either be unencrypted, or Squid started with the -N
3178 option to allow it to query interactively for the passphrase.
3180 The key file name is given as argument to the program allowing
3181 selection of the right password if you have multiple encrypted
3186 OPTIONS RELATING TO EXTERNAL SSL_CRTD
3187 -----------------------------------------------------------------------------
3190 NAME: sslcrtd_program
3193 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
3194 LOC: Ssl::TheConfig.ssl_crtd
3196 Specify the location and options of the executable for certificate
3198 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
3199 For more information use:
3200 @DEFAULT_SSL_CRTD@ -h
3203 NAME: sslcrtd_children
3204 TYPE: HelperChildConfig
3206 DEFAULT: 32 startup=5 idle=1
3207 LOC: Ssl::TheConfig.ssl_crtdChildren
3209 The maximum number of processes spawn to service ssl server.
3210 The maximum this may be safely set to is 32.
3212 The startup= and idle= options allow some measure of skew in your
3217 Sets the minimum number of processes to spawn when Squid
3218 starts or reconfigures. When set to zero the first request will
3219 cause spawning of the first child process to handle it.
3221 Starting too few children temporary slows Squid under load while it
3222 tries to spawn enough additional processes to cope with traffic.
3226 Sets a minimum of how many processes Squid is to try and keep available
3227 at all times. When traffic begins to rise above what the existing
3228 processes can handle this many more will be spawned up to the maximum
3229 configured. A minimum setting of 1 is required.
3233 Sets the maximum number of queued requests.
3234 If the queued requests exceed queue size for more than 3 minutes
3235 squid aborts its operation.
3236 The default value is set to 2*numberofchildren.
3238 You must have at least one ssl_crtd process.
3241 NAME: sslcrtvalidator_program
3245 LOC: Ssl::TheConfig.ssl_crt_validator
3247 Specify the location and options of the executable for ssl_crt_validator
3250 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
3253 ttl=n TTL in seconds for cached results. The default is 60 secs
3254 cache=n limit the result cache size. The default value is 2048
3257 NAME: sslcrtvalidator_children
3258 TYPE: HelperChildConfig
3260 DEFAULT: 32 startup=5 idle=1 concurrency=1
3261 LOC: Ssl::TheConfig.ssl_crt_validator_Children
3263 The maximum number of processes spawn to service SSL server.
3264 The maximum this may be safely set to is 32.
3266 The startup= and idle= options allow some measure of skew in your
3271 Sets the minimum number of processes to spawn when Squid
3272 starts or reconfigures. When set to zero the first request will
3273 cause spawning of the first child process to handle it.
3275 Starting too few children temporary slows Squid under load while it
3276 tries to spawn enough additional processes to cope with traffic.
3280 Sets a minimum of how many processes Squid is to try and keep available
3281 at all times. When traffic begins to rise above what the existing
3282 processes can handle this many more will be spawned up to the maximum
3283 configured. A minimum setting of 1 is required.
3287 The number of requests each certificate validator helper can handle in
3288 parallel. A value of 0 indicates the certficate validator does not
3289 support concurrency. Defaults to 1.
3291 When this directive is set to a value >= 1 then the protocol
3292 used to communicate with the helper is modified to include
3293 a request ID in front of the request/response. The request
3294 ID from the request must be echoed back with the response
3299 Sets the maximum number of queued requests.
3300 If the queued requests exceed queue size for more than 3 minutes
3301 squid aborts its operation.
3302 The default value is set to 2*numberofchildren.
3304 You must have at least one ssl_crt_validator process.
3308 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
3309 -----------------------------------------------------------------------------
3317 To specify other caches in a hierarchy, use the format:
3319 cache_peer hostname type http-port icp-port [options]
3324 # hostname type port port options
3325 # -------------------- -------- ----- ----- -----------
3326 cache_peer parent.foo.net parent 3128 3130 default
3327 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
3328 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
3329 cache_peer example.com parent 80 0 default
3330 cache_peer cdn.example.com sibling 3128 0
3332 type: either 'parent', 'sibling', or 'multicast'.
3334 proxy-port: The port number where the peer accept HTTP requests.
3335 For other Squid proxies this is usually 3128
3336 For web servers this is usually 80
3338 icp-port: Used for querying neighbor caches about objects.
3339 Set to 0 if the peer does not support ICP or HTCP.
3340 See ICP and HTCP options below for additional details.
3343 ==== ICP OPTIONS ====
3345 You MUST also set icp_port and icp_access explicitly when using these options.
3346 The defaults will prevent peer traffic using ICP.
3349 no-query Disable ICP queries to this neighbor.
3352 Indicates the named peer is a member of a multicast group.
3353 ICP queries will not be sent directly to the peer, but ICP
3354 replies will be accepted from it.
3356 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
3357 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
3360 To only send ICP queries to this neighbor infrequently.
3361 This is used to keep the neighbor round trip time updated
3362 and is usually used in conjunction with weighted-round-robin.
3365 ==== HTCP OPTIONS ====
3367 You MUST also set htcp_port and htcp_access explicitly when using these options.
3368 The defaults will prevent peer traffic using HTCP.
3371 htcp Send HTCP, instead of ICP, queries to the neighbor.
3372 You probably also want to set the "icp-port" to 4827
3373 instead of 3130. This directive accepts a comma separated
3374 list of options described below.
3376 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3378 htcp=no-clr Send HTCP to the neighbor but without
3379 sending any CLR requests. This cannot be used with
3382 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3383 This cannot be used with no-clr.
3386 Send HTCP to the neighbor including CLRs but only when
3387 they do not result from PURGE requests.
3390 Forward any HTCP CLR requests this proxy receives to the peer.
3393 ==== PEER SELECTION METHODS ====
3395 The default peer selection method is ICP, with the first responding peer
3396 being used as source. These options can be used for better load balancing.
3399 default This is a parent cache which can be used as a "last-resort"
3400 if a peer cannot be located by any of the peer-selection methods.
3401 If specified more than once, only the first is used.
3403 round-robin Load-Balance parents which should be used in a round-robin
3404 fashion in the absence of any ICP queries.
3405 weight=N can be used to add bias.
3407 weighted-round-robin
3408 Load-Balance parents which should be used in a round-robin
3409 fashion with the frequency of each parent being based on the
3410 round trip time. Closer parents are used more often.
3411 Usually used for background-ping parents.
3412 weight=N can be used to add bias.
3414 carp Load-Balance parents which should be used as a CARP array.
3415 The requests will be distributed among the parents based on the
3416 CARP load balancing hash function based on their weight.
3418 userhash Load-balance parents based on the client proxy_auth or ident username.
3420 sourcehash Load-balance parents based on the client source IP.
3423 To be used only for cache peers of type "multicast".
3424 ALL members of this multicast group have "sibling"
3425 relationship with it, not "parent". This is to a multicast
3426 group when the requested object would be fetched only from
3427 a "parent" cache, anyway. It's useful, e.g., when
3428 configuring a pool of redundant Squid proxies, being
3429 members of the same multicast group.
3432 ==== PEER SELECTION OPTIONS ====
3434 weight=N use to affect the selection of a peer during any weighted
3435 peer-selection mechanisms.
3436 The weight must be an integer; default is 1,
3437 larger weights are favored more.
3438 This option does not affect parent selection if a peering
3439 protocol is not in use.
3441 basetime=N Specify a base amount to be subtracted from round trip
3443 It is subtracted before division by weight in calculating
3444 which parent to fectch from. If the rtt is less than the
3445 base time the rtt is set to a minimal value.
3447 ttl=N Specify a TTL to use when sending multicast ICP queries
3449 Only useful when sending to a multicast group.
3450 Because we don't accept ICP replies from random
3451 hosts, you must configure other group members as
3452 peers with the 'multicast-responder' option.
3454 no-delay To prevent access to this neighbor from influencing the
3457 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3458 enabled) for this host from the specified URL rather
3459 than the Squid default location.
3462 ==== CARP OPTIONS ====
3464 carp-key=key-specification
3465 use a different key than the full URL to hash against the peer.
3466 the key-specification is a comma-separated list of the keywords
3467 scheme, host, port, path, params
3468 Order is not important.
3470 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3472 originserver Causes this parent to be contacted as an origin server.
3473 Meant to be used in accelerator setups when the peer
3477 Set the Host header of requests forwarded to this peer.
3478 Useful in accelerator setups where the server (peer)
3479 expects a certain domain name but clients may request
3480 others. ie example.com or www.example.com
3482 no-digest Disable request of cache digests.
3485 Disables requesting ICMP RTT database (NetDB).
3488 ==== AUTHENTICATION OPTIONS ====
3491 If this is a personal/workgroup proxy and your parent
3492 requires proxy authentication.
3494 Note: The string can include URL escapes (i.e. %20 for
3495 spaces). This also means % must be written as %%.
3498 Send login details received from client to this peer.
3499 Both Proxy- and WWW-Authorization headers are passed
3500 without alteration to the peer.
3501 Authentication is not required by Squid for this to work.
3503 Note: This will pass any form of authentication but
3504 only Basic auth will work through a proxy unless the
3505 connection-auth options are also used.
3507 login=PASS Send login details received from client to this peer.
3508 Authentication is not required by this option.
3510 If there are no client-provided authentication headers
3511 to pass on, but username and password are available
3512 from an external ACL user= and password= result tags
3513 they may be sent instead.
3515 Note: To combine this with proxy_auth both proxies must
3516 share the same user database as HTTP only allows for
3517 a single login (one for proxy, one for origin server).
3518 Also be warned this will expose your users proxy
3519 password to the peer. USE WITH CAUTION
3522 Send the username to the upstream cache, but with a
3523 fixed password. This is meant to be used when the peer
3524 is in another administrative domain, but it is still
3525 needed to identify each user.
3526 The star can optionally be followed by some extra
3527 information which is added to the username. This can
3528 be used to identify this proxy to the peer, similar to
3529 the login=username:password option above.
3532 If this is a personal/workgroup proxy and your parent
3533 requires a secure proxy authentication.
3534 The first principal from the default keytab or defined by
3535 the environment variable KRB5_KTNAME will be used.
3537 WARNING: The connection may transmit requests from multiple
3538 clients. Negotiate often assumes end-to-end authentication
3539 and a single-client. Which is not strictly true here.
3541 login=NEGOTIATE:principal_name
3542 If this is a personal/workgroup proxy and your parent
3543 requires a secure proxy authentication.
3544 The principal principal_name from the default keytab or
3545 defined by the environment variable KRB5_KTNAME will be
3548 WARNING: The connection may transmit requests from multiple
3549 clients. Negotiate often assumes end-to-end authentication
3550 and a single-client. Which is not strictly true here.
3552 connection-auth=on|off
3553 Tell Squid that this peer does or not support Microsoft
3554 connection oriented authentication, and any such
3555 challenges received from there should be ignored.
3556 Default is auto to automatically determine the status
3560 Do not use a keytab to authenticate to a peer when
3561 login=NEGOTIATE is specified. Let the GSSAPI
3562 implementation determine which already existing
3563 credentials cache to use instead.
3566 ==== SSL / HTTPS / TLS OPTIONS ====
3568 tls Encrypt connections to this peer with TLS.
3570 sslcert=/path/to/ssl/certificate
3571 A client SSL certificate to use when connecting to
3574 sslkey=/path/to/ssl/key
3575 The private SSL key corresponding to sslcert above.
3576 If 'sslkey' is not specified 'sslcert' is assumed to
3577 reference a combined file containing both the
3578 certificate and the key.
3580 sslcipher=... The list of valid SSL ciphers to use when connecting
3584 The minimum TLS protocol version to permit. To control
3585 SSLv3 use the tls-options= parameter.
3586 Supported Values: 1.0 (default), 1.1, 1.2
3588 tls-options=... Specify various TLS implementation options.
3590 OpenSSL options most important are:
3592 NO_SSLv3 Disallow the use of SSLv3
3595 Always create a new key when using
3596 temporary/ephemeral DH key exchanges
3599 Disable use of RFC5077 session tickets.
3600 Some servers may have problems
3601 understanding the TLS extension due
3602 to ambiguous specification in RFC4507.
3604 ALL Enable various bug workarounds
3605 suggested as "harmless" by OpenSSL
3606 Be warned that this reduces SSL/TLS
3607 strength to some attacks.
3609 See the OpenSSL SSL_CTX_set_options documentation for a
3612 GnuTLS options most important are:
3615 Disable use of RFC5077 session tickets.
3616 Some servers may have problems
3617 understanding the TLS extension due
3618 to ambiguous specification in RFC4507.
3620 See the GnuTLS Priority Strings documentation
3621 for a more complete list.
3622 http://www.gnutls.org/manual/gnutls.html#Priority-Strings
3624 tls-cafile= PEM file containing CA certificates to use when verifying
3625 the peer certificate. May be repeated to load multiple files.
3627 sslcapath=... A directory containing additional CA certificates to
3628 use when verifying the peer certificate.
3629 Requires OpenSSL or LibreSSL.
3631 sslcrlfile=... A certificate revocation list file to use when
3632 verifying the peer certificate.
3634 sslflags=... Specify various flags modifying the SSL implementation:
3637 Accept certificates even if they fail to
3641 Don't verify the peer certificate
3642 matches the server name
3644 ssldomain= The peer name as advertised in it's certificate.
3645 Used for verifying the correctness of the received peer
3646 certificate. If not specified the peer hostname will be
3649 front-end-https[=off|on|auto]
3650 Enable the "Front-End-Https: On" header needed when
3651 using Squid as a SSL frontend in front of Microsoft OWA.
3652 See MS KB document Q307347 for details on this header.
3653 If set to auto the header will only be added if the
3654 request is forwarded as a https:// URL.
3656 tls-default-ca[=off]
3657 Whether to use the system Trusted CAs. Default is ON.
3659 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
3661 ==== GENERAL OPTIONS ====
3664 A peer-specific connect timeout.
3665 Also see the peer_connect_timeout directive.
3667 connect-fail-limit=N
3668 How many times connecting to a peer must fail before
3669 it is marked as down. Standby connection failures
3670 count towards this limit. Default is 10.
3672 allow-miss Disable Squid's use of only-if-cached when forwarding
3673 requests to siblings. This is primarily useful when
3674 icp_hit_stale is used by the sibling. Excessive use
3675 of this option may result in forwarding loops. One way
3676 to prevent peering loops when using this option, is to
3677 deny cache peer usage on requests from a peer:
3679 cache_peer_access peerName deny fromPeer
3681 max-conn=N Limit the number of concurrent connections the Squid
3682 may open to this peer, including already opened idle
3683 and standby connections. There is no peer-specific
3684 connection limit by default.
3686 A peer exceeding the limit is not used for new
3687 requests unless a standby connection is available.
3689 max-conn currently works poorly with idle persistent
3690 connections: When a peer reaches its max-conn limit,
3691 and there are idle persistent connections to the peer,
3692 the peer may not be selected because the limiting code
3693 does not know whether Squid can reuse those idle
3696 standby=N Maintain a pool of N "hot standby" connections to an
3697 UP peer, available for requests when no idle
3698 persistent connection is available (or safe) to use.
3699 By default and with zero N, no such pool is maintained.
3700 N must not exceed the max-conn limit (if any).
3702 At start or after reconfiguration, Squid opens new TCP
3703 standby connections until there are N connections
3704 available and then replenishes the standby pool as
3705 opened connections are used up for requests. A used
3706 connection never goes back to the standby pool, but
3707 may go to the regular idle persistent connection pool
3708 shared by all peers and origin servers.
3710 Squid never opens multiple new standby connections
3711 concurrently. This one-at-a-time approach minimizes
3712 flooding-like effect on peers. Furthermore, just a few
3713 standby connections should be sufficient in most cases
3714 to supply most new requests with a ready-to-use
3717 Standby connections obey server_idle_pconn_timeout.
3718 For the feature to work as intended, the peer must be
3719 configured to accept and keep them open longer than
3720 the idle timeout at the connecting Squid, to minimize
3721 race conditions typical to idle used persistent
3722 connections. Default request_timeout and
3723 server_idle_pconn_timeout values ensure such a
3726 name=xxx Unique name for the peer.
3727 Required if you have multiple peers on the same host
3728 but different ports.
3729 This name can be used in cache_peer_access and similar
3730 directives to identify the peer.
3731 Can be used by outgoing access controls through the
3734 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3735 requests to this peer. Use normal address selection instead.
3736 This overrides the spoof_client_ip ACL.
3738 proxy-only objects fetched from the peer will not be stored locally.
3742 NAME: cache_peer_access
3745 DEFAULT_DOC: No peer usage restrictions.
3748 Restricts usage of cache_peer proxies.
3751 cache_peer_access peer-name allow|deny [!]aclname ...
3753 For the required peer-name parameter, use either the value of the
3754 cache_peer name=value parameter or, if name=value is missing, the
3755 cache_peer hostname parameter.
3757 This directive narrows down the selection of peering candidates, but
3758 does not determine the order in which the selected candidates are
3759 contacted. That order is determined by the peer selection algorithms
3760 (see PEER SELECTION sections in the cache_peer documentation).
3762 If a deny rule matches, the corresponding peer will not be contacted
3763 for the current transaction -- Squid will not send ICP queries and
3764 will not forward HTTP requests to that peer. An allow match leaves
3765 the corresponding peer in the selection. The first match for a given
3766 peer wins for that peer.
3768 The relative order of cache_peer_access directives for the same peer
3769 matters. The relative order of any two cache_peer_access directives
3770 for different peers does not matter. To ease interpretation, it is a
3771 good idea to group cache_peer_access directives for the same peer
3774 A single cache_peer_access directive may be evaluated multiple times
3775 for a given transaction because individual peer selection algorithms
3776 may check it independently from each other. These redundant checks
3777 may be optimized away in future Squid versions.
3779 This clause only supports fast acl types.
3780 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3784 NAME: neighbor_type_domain
3785 TYPE: hostdomaintype
3787 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3790 Modify the cache_peer neighbor type when passing requests
3791 about specific domains to the peer.
3794 neighbor_type_domain neighbor parent|sibling domain domain ...
3797 cache_peer foo.example.com parent 3128 3130
3798 neighbor_type_domain foo.example.com sibling .au .de
3800 The above configuration treats all requests to foo.example.com as a
3801 parent proxy unless the request is for a .au or .de ccTLD domain name.
3804 NAME: dead_peer_timeout
3808 LOC: Config.Timeout.deadPeer
3810 This controls how long Squid waits to declare a peer cache
3811 as "dead." If there are no ICP replies received in this
3812 amount of time, Squid will declare the peer dead and not
3813 expect to receive any further ICP replies. However, it
3814 continues to send ICP queries, and will mark the peer as
3815 alive upon receipt of the first subsequent ICP reply.
3817 This timeout also affects when Squid expects to receive ICP
3818 replies from peers. If more than 'dead_peer' seconds have
3819 passed since the last ICP reply was received, Squid will not
3820 expect to receive an ICP reply on the next query. Thus, if
3821 your time between requests is greater than this timeout, you
3822 will see a lot of requests sent DIRECT to origin servers
3823 instead of to your parents.
3826 NAME: forward_max_tries
3829 LOC: Config.forward_max_tries
3831 Controls how many different forward paths Squid will try
3832 before giving up. See also forward_timeout.
3834 NOTE: connect_retries (default: none) can make each of these
3835 possible forwarding paths be tried multiple times.
3839 MEMORY CACHE OPTIONS
3840 -----------------------------------------------------------------------------
3847 LOC: Config.memMaxSize
3849 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3850 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3851 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3852 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3854 'cache_mem' specifies the ideal amount of memory to be used
3856 * In-Transit objects
3858 * Negative-Cached objects
3860 Data for these objects are stored in 4 KB blocks. This
3861 parameter specifies the ideal upper limit on the total size of
3862 4 KB blocks allocated. In-Transit objects take the highest
3865 In-transit objects have priority over the others. When
3866 additional space is needed for incoming data, negative-cached
3867 and hot objects will be released. In other words, the
3868 negative-cached and hot objects will fill up any unused space
3869 not needed for in-transit objects.
3871 If circumstances require, this limit will be exceeded.
3872 Specifically, if your incoming request rate requires more than
3873 'cache_mem' of memory to hold in-transit objects, Squid will
3874 exceed this limit to satisfy the new requests. When the load
3875 decreases, blocks will be freed until the high-water mark is
3876 reached. Thereafter, blocks will be used to store hot
3879 If shared memory caching is enabled, Squid does not use the shared
3880 cache space for in-transit objects, but they still consume as much
3881 local memory as they need. For more details about the shared memory
3882 cache, see memory_cache_shared.
3885 NAME: maximum_object_size_in_memory
3889 LOC: Config.Store.maxInMemObjSize
3891 Objects greater than this size will not be attempted to kept in
3892 the memory cache. This should be set high enough to keep objects
3893 accessed frequently in memory to improve performance whilst low
3894 enough to keep larger objects from hoarding cache_mem.
3897 NAME: memory_cache_shared
3900 LOC: Config.memShared
3902 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3904 Controls whether the memory cache is shared among SMP workers.
3906 The shared memory cache is meant to occupy cache_mem bytes and replace
3907 the non-shared memory cache, although some entities may still be
3908 cached locally by workers for now (e.g., internal and in-transit
3909 objects may be served from a local memory cache even if shared memory
3910 caching is enabled).
3912 By default, the memory cache is shared if and only if all of the
3913 following conditions are satisfied: Squid runs in SMP mode with
3914 multiple workers, cache_mem is positive, and Squid environment
3915 supports required IPC primitives (e.g., POSIX shared memory segments
3916 and GCC-style atomic operations).
3918 To avoid blocking locks, shared memory uses opportunistic algorithms
3919 that do not guarantee that every cachable entity that could have been
3920 shared among SMP workers will actually be shared.
3923 NAME: memory_cache_mode
3927 DEFAULT_DOC: Keep the most recently fetched objects in memory
3929 Controls which objects to keep in the memory cache (cache_mem)
3931 always Keep most recently fetched objects in memory (default)
3933 disk Only disk cache hits are kept in memory, which means
3934 an object must first be cached on disk and then hit
3935 a second time before cached in memory.
3937 network Only objects fetched from network is kept in memory
3940 NAME: memory_replacement_policy
3942 LOC: Config.memPolicy
3945 The memory replacement policy parameter determines which
3946 objects are purged from memory when memory space is needed.
3948 See cache_replacement_policy for details on algorithms.
3953 -----------------------------------------------------------------------------
3956 NAME: cache_replacement_policy
3958 LOC: Config.replPolicy
3961 The cache replacement policy parameter determines which
3962 objects are evicted (replaced) when disk space is needed.
3964 lru : Squid's original list based LRU policy
3965 heap GDSF : Greedy-Dual Size Frequency
3966 heap LFUDA: Least Frequently Used with Dynamic Aging
3967 heap LRU : LRU policy implemented using a heap
3969 Applies to any cache_dir lines listed below this directive.
3971 The LRU policies keeps recently referenced objects.
3973 The heap GDSF policy optimizes object hit rate by keeping smaller
3974 popular objects in cache so it has a better chance of getting a
3975 hit. It achieves a lower byte hit rate than LFUDA though since
3976 it evicts larger (possibly popular) objects.
3978 The heap LFUDA policy keeps popular objects in cache regardless of
3979 their size and thus optimizes byte hit rate at the expense of
3980 hit rate since one large, popular object will prevent many
3981 smaller, slightly less popular objects from being cached.
3983 Both policies utilize a dynamic aging mechanism that prevents
3984 cache pollution that can otherwise occur with frequency-based
3985 replacement policies.
3987 NOTE: if using the LFUDA replacement policy you should increase
3988 the value of maximum_object_size above its default of 4 MB to
3989 to maximize the potential byte hit rate improvement of LFUDA.
3991 For more information about the GDSF and LFUDA cache replacement
3992 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3993 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3996 NAME: minimum_object_size
4000 DEFAULT_DOC: no limit
4001 LOC: Config.Store.minObjectSize
4003 Objects smaller than this size will NOT be saved on disk. The
4004 value is specified in bytes, and the default is 0 KB, which
4005 means all responses can be stored.
4008 NAME: maximum_object_size
4012 LOC: Config.Store.maxObjectSize
4014 Set the default value for max-size parameter on any cache_dir.
4015 The value is specified in bytes, and the default is 4 MB.
4017 If you wish to get a high BYTES hit ratio, you should probably
4018 increase this (one 32 MB object hit counts for 3200 10KB
4021 If you wish to increase hit ratio more than you want to
4022 save bandwidth you should leave this low.
4024 NOTE: if using the LFUDA replacement policy you should increase
4025 this value to maximize the byte hit rate improvement of LFUDA!
4026 See cache_replacement_policy for a discussion of this policy.
4032 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
4033 LOC: Config.cacheSwap
4036 cache_dir Type Directory-Name Fs-specific-data [options]
4038 You can specify multiple cache_dir lines to spread the
4039 cache among different disk partitions.
4041 Type specifies the kind of storage system to use. Only "ufs"
4042 is built by default. To enable any of the other storage systems
4043 see the --enable-storeio configure option.
4045 'Directory' is a top-level directory where cache swap
4046 files will be stored. If you want to use an entire disk
4047 for caching, this can be the mount-point directory.
4048 The directory must exist and be writable by the Squid
4049 process. Squid will NOT create this directory for you.
4051 In SMP configurations, cache_dir must not precede the workers option
4052 and should use configuration macros or conditionals to give each
4053 worker interested in disk caching a dedicated cache directory.
4056 ==== The ufs store type ====
4058 "ufs" is the old well-known Squid storage format that has always
4062 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
4064 'Mbytes' is the amount of disk space (MB) to use under this
4065 directory. The default is 100 MB. Change this to suit your
4066 configuration. Do NOT put the size of your disk drive here.
4067 Instead, if you want Squid to use the entire disk drive,
4068 subtract 20% and use that value.
4070 'L1' is the number of first-level subdirectories which
4071 will be created under the 'Directory'. The default is 16.
4073 'L2' is the number of second-level subdirectories which
4074 will be created under each first-level directory. The default
4078 ==== The aufs store type ====
4080 "aufs" uses the same storage format as "ufs", utilizing
4081 POSIX-threads to avoid blocking the main Squid process on
4082 disk-I/O. This was formerly known in Squid as async-io.
4085 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
4087 see argument descriptions under ufs above
4090 ==== The diskd store type ====
4092 "diskd" uses the same storage format as "ufs", utilizing a
4093 separate process to avoid blocking the main Squid process on
4097 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
4099 see argument descriptions under ufs above
4101 Q1 specifies the number of unacknowledged I/O requests when Squid
4102 stops opening new files. If this many messages are in the queues,
4103 Squid won't open new files. Default is 64
4105 Q2 specifies the number of unacknowledged messages when Squid
4106 starts blocking. If this many messages are in the queues,
4107 Squid blocks until it receives some replies. Default is 72
4109 When Q1 < Q2 (the default), the cache directory is optimized
4110 for lower response time at the expense of a decrease in hit
4111 ratio. If Q1 > Q2, the cache directory is optimized for
4112 higher hit ratio at the expense of an increase in response
4116 ==== The rock store type ====
4119 cache_dir rock Directory-Name Mbytes [options]
4121 The Rock Store type is a database-style storage. All cached
4122 entries are stored in a "database" file, using fixed-size slots.
4123 A single entry occupies one or more slots.
4125 If possible, Squid using Rock Store creates a dedicated kid
4126 process called "disker" to avoid blocking Squid worker(s) on disk
4127 I/O. One disker kid is created for each rock cache_dir. Diskers
4128 are created only when Squid, running in daemon mode, has support
4129 for the IpcIo disk I/O module.
4131 swap-timeout=msec: Squid will not start writing a miss to or
4132 reading a hit from disk if it estimates that the swap operation
4133 will take more than the specified number of milliseconds. By
4134 default and when set to zero, disables the disk I/O time limit
4135 enforcement. Ignored when using blocking I/O module because
4136 blocking synchronous I/O does not allow Squid to estimate the
4137 expected swap wait time.
4139 max-swap-rate=swaps/sec: Artificially limits disk access using
4140 the specified I/O rate limit. Swap out requests that
4141 would cause the average I/O rate to exceed the limit are
4142 delayed. Individual swap in requests (i.e., hits or reads) are
4143 not delayed, but they do contribute to measured swap rate and
4144 since they are placed in the same FIFO queue as swap out
4145 requests, they may wait longer if max-swap-rate is smaller.
4146 This is necessary on file systems that buffer "too
4147 many" writes and then start blocking Squid and other processes
4148 while committing those writes to disk. Usually used together
4149 with swap-timeout to avoid excessive delays and queue overflows
4150 when disk demand exceeds available disk "bandwidth". By default
4151 and when set to zero, disables the disk I/O rate limit
4152 enforcement. Currently supported by IpcIo module only.
4154 slot-size=bytes: The size of a database "record" used for
4155 storing cached responses. A cached response occupies at least
4156 one slot and all database I/O is done using individual slots so
4157 increasing this parameter leads to more disk space waste while
4158 decreasing it leads to more disk I/O overheads. Should be a
4159 multiple of your operating system I/O page size. Defaults to
4160 16KBytes. A housekeeping header is stored with each slot and
4161 smaller slot-sizes will be rejected. The header is smaller than
4165 ==== COMMON OPTIONS ====
4167 no-store no new objects should be stored to this cache_dir.
4169 min-size=n the minimum object size in bytes this cache_dir
4170 will accept. It's used to restrict a cache_dir
4171 to only store large objects (e.g. AUFS) while
4172 other stores are optimized for smaller objects
4176 max-size=n the maximum object size in bytes this cache_dir
4178 The value in maximum_object_size directive sets
4179 the default unless more specific details are
4180 available (ie a small store capacity).
4182 Note: To make optimal use of the max-size limits you should order
4183 the cache_dir lines with the smallest max-size value first.
4187 # Uncomment and adjust the following to add a disk cache directory.
4188 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
4192 NAME: store_dir_select_algorithm
4194 LOC: Config.store_dir_select_algorithm
4197 How Squid selects which cache_dir to use when the response
4198 object will fit into more than one.
4200 Regardless of which algorithm is used the cache_dir min-size
4201 and max-size parameters are obeyed. As such they can affect
4202 the selection algorithm by limiting the set of considered
4209 This algorithm is suited to caches with similar cache_dir
4210 sizes and disk speeds.
4212 The disk with the least I/O pending is selected.
4213 When there are multiple disks with the same I/O load ranking
4214 the cache_dir with most available capacity is selected.
4216 When a mix of cache_dir sizes are configured the faster disks
4217 have a naturally lower I/O loading and larger disks have more
4218 capacity. So space used to store objects and data throughput
4219 may be very unbalanced towards larger disks.
4224 This algorithm is suited to caches with unequal cache_dir
4227 Each cache_dir is selected in a rotation. The next suitable
4230 Available cache_dir capacity is only considered in relation
4231 to whether the object will fit and meets the min-size and
4232 max-size parameters.
4234 Disk I/O loading is only considered to prevent overload on slow
4235 disks. This algorithm does not spread objects by size, so any
4236 I/O loading per-disk may appear very unbalanced and volatile.
4238 If several cache_dirs use similar min-size, max-size, or other
4239 limits to to reject certain responses, then do not group such
4240 cache_dir lines together, to avoid round-robin selection bias
4241 towards the first cache_dir after the group. Instead, interleave
4242 cache_dir lines from different groups. For example:
4244 store_dir_select_algorithm round-robin
4245 cache_dir rock /hdd1 ... min-size=100000
4246 cache_dir rock /ssd1 ... max-size=99999
4247 cache_dir rock /hdd2 ... min-size=100000
4248 cache_dir rock /ssd2 ... max-size=99999
4249 cache_dir rock /hdd3 ... min-size=100000
4250 cache_dir rock /ssd3 ... max-size=99999
4253 NAME: max_open_disk_fds
4255 LOC: Config.max_open_disk_fds
4257 DEFAULT_DOC: no limit
4259 To avoid having disk as the I/O bottleneck Squid can optionally
4260 bypass the on-disk cache if more than this amount of disk file
4261 descriptors are open.
4263 A value of 0 indicates no limit.
4266 NAME: cache_swap_low
4267 COMMENT: (percent, 0-100)
4270 LOC: Config.Swap.lowWaterMark
4272 The low-water mark for AUFS/UFS/diskd cache object eviction by
4273 the cache_replacement_policy algorithm.
4275 Removal begins when the swap (disk) usage of a cache_dir is
4276 above this low-water mark and attempts to maintain utilization
4277 near the low-water mark.
4279 As swap utilization increases towards the high-water mark set
4280 by cache_swap_high object eviction becomes more agressive.
4282 The value difference in percentages between low- and high-water
4283 marks represent an eviction rate of 300 objects per second and
4284 the rate continues to scale in agressiveness by multiples of
4285 this above the high-water mark.
4287 Defaults are 90% and 95%. If you have a large cache, 5% could be
4288 hundreds of MB. If this is the case you may wish to set these
4289 numbers closer together.
4291 See also cache_swap_high and cache_replacement_policy
4294 NAME: cache_swap_high
4295 COMMENT: (percent, 0-100)
4298 LOC: Config.Swap.highWaterMark
4300 The high-water mark for AUFS/UFS/diskd cache object eviction by
4301 the cache_replacement_policy algorithm.
4303 Removal begins when the swap (disk) usage of a cache_dir is
4304 above the low-water mark set by cache_swap_low and attempts to
4305 maintain utilization near the low-water mark.
4307 As swap utilization increases towards this high-water mark object
4308 eviction becomes more agressive.
4310 The value difference in percentages between low- and high-water
4311 marks represent an eviction rate of 300 objects per second and
4312 the rate continues to scale in agressiveness by multiples of
4313 this above the high-water mark.
4315 Defaults are 90% and 95%. If you have a large cache, 5% could be
4316 hundreds of MB. If this is the case you may wish to set these
4317 numbers closer together.
4319 See also cache_swap_low and cache_replacement_policy
4324 -----------------------------------------------------------------------------
4331 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
4335 logformat <name> <format specification>
4337 Defines an access log format.
4339 The <format specification> is a string with embedded % format codes
4341 % format codes all follow the same basic structure where all but
4342 the formatcode is optional. Output strings are automatically escaped
4343 as required according to their context and the output format
4344 modifiers are usually not needed, but can be specified if an explicit
4345 output format is desired.
4347 % ["|[|'|#|/] [-] [[0]width] [{arg}] formatcode [{arg}]
4349 " output in quoted string format
4350 [ output in squid text log format as used by log_mime_hdrs
4351 # output in URL quoted format
4352 / output in shell \-escaped format
4357 width minimum and/or maximum field width:
4358 [width_min][.width_max]
4359 When minimum starts with 0, the field is zero-padded.
4360 String values exceeding maximum width are truncated.
4362 {arg} argument such as header name etc. This field may be
4363 placed before or after the token, but not both at once.
4367 % a literal % character
4368 sn Unique sequence number per log line entry
4369 err_code The ID of an error response served by Squid or
4370 a similar internal error identifier.
4371 err_detail Additional err_code-dependent error information.
4372 note The annotation specified by the argument. Also
4373 logs the adaptation meta headers set by the
4374 adaptation_meta configuration parameter.
4375 If no argument given all annotations logged.
4376 The argument may include a separator to use with
4379 By default, multiple note values are separated with ","
4380 and multiple notes are separated with "\r\n".
4381 When logging named notes with %{name}note, the
4382 explicitly configured separator is used between note
4383 values. When logging all notes with %note, the
4384 explicitly configured separator is used between
4385 individual notes. There is currently no way to
4386 specify both value and notes separators when logging
4387 all notes with %note.
4389 Connection related format codes:
4391 >a Client source IP address
4393 >p Client source port
4394 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
4395 >la Local IP address the client connected to
4396 >lp Local port number the client connected to
4397 >qos Client connection TOS/DSCP value set by Squid
4398 >nfmark Client connection netfilter mark set by Squid
4400 la Local listening IP address the client connection was connected to.
4401 lp Local listening port number the client connection was connected to.
4403 <a Server IP address of the last server or peer connection
4404 <A Server FQDN or peer name
4405 <p Server port number of the last server or peer connection
4406 <la Local IP address of the last server or peer connection
4407 <lp Local port number of the last server or peer connection
4408 <qos Server connection TOS/DSCP value set by Squid
4409 <nfmark Server connection netfilter mark set by Squid
4411 Time related format codes:
4413 ts Seconds since epoch
4414 tu subsecond time (milliseconds)
4415 tl Local time. Optional strftime format argument
4416 default %d/%b/%Y:%H:%M:%S %z
4417 tg GMT time. Optional strftime format argument
4418 default %d/%b/%Y:%H:%M:%S %z
4419 tr Response time (milliseconds)
4420 dt Total time spent making DNS lookups (milliseconds)
4421 tS Approximate master transaction start time in
4422 <full seconds since epoch>.<fractional seconds> format.
4423 Currently, Squid considers the master transaction
4424 started when a complete HTTP request header initiating
4425 the transaction is received from the client. This is
4426 the same value that Squid uses to calculate transaction
4427 response time when logging %tr to access.log. Currently,
4428 Squid uses millisecond resolution for %tS values,
4429 similar to the default access.log "current time" field
4432 Access Control related format codes:
4434 et Tag returned by external acl
4435 ea Log string returned by external acl
4436 un User name (any available)
4437 ul User name from authentication
4438 ue User name from external acl helper
4439 ui User name from ident
4440 un A user name. Expands to the first available name
4441 from the following list of information sources:
4442 - authenticated user name, like %ul
4443 - user name supplied by an external ACL, like %ue
4444 - SSL client name, like %us
4445 - ident user name, like %ui
4446 credentials Client credentials. The exact meaning depends on
4447 the authentication scheme: For Basic authentication,
4448 it is the password; for Digest, the realm sent by the
4449 client; for NTLM and Negotiate, the client challenge
4450 or client credentials prefixed with "YR " or "KK ".
4452 HTTP related format codes:
4456 [http::]rm Request method (GET/POST etc)
4457 [http::]>rm Request method from client
4458 [http::]<rm Request method sent to server or peer
4459 [http::]ru Request URL from client (historic, filtered for logging)
4460 [http::]>ru Request URL from client
4461 [http::]<ru Request URL sent to server or peer
4462 [http::]>rs Request URL scheme from client
4463 [http::]<rs Request URL scheme sent to server or peer
4464 [http::]>rd Request URL domain from client
4465 [http::]<rd Request URL domain sent to server or peer
4466 [http::]>rP Request URL port from client
4467 [http::]<rP Request URL port sent to server or peer
4468 [http::]rp Request URL path excluding hostname
4469 [http::]>rp Request URL path excluding hostname from client
4470 [http::]<rp Request URL path excluding hostname sent to server or peer
4471 [http::]rv Request protocol version
4472 [http::]>rv Request protocol version from client
4473 [http::]<rv Request protocol version sent to server or peer
4475 [http::]>h Original received request header.
4476 Usually differs from the request header sent by
4477 Squid, although most fields are often preserved.
4478 Accepts optional header field name/value filter
4479 argument using name[:[separator]element] format.
4480 [http::]>ha Received request header after adaptation and
4481 redirection (pre-cache REQMOD vectoring point).
4482 Usually differs from the request header sent by
4483 Squid, although most fields are often preserved.
4484 Optional header name argument as for >h
4488 [http::]<Hs HTTP status code received from the next hop
4489 [http::]>Hs HTTP status code sent to the client
4491 [http::]<h Reply header. Optional header name argument
4494 [http::]mt MIME content type
4499 [http::]st Total size of request + reply traffic with client
4500 [http::]>st Total size of request received from client.
4501 Excluding chunked encoding bytes.
4502 [http::]<st Total size of reply sent to client (after adaptation)
4504 [http::]>sh Size of request headers received from client
4505 [http::]<sh Size of reply headers sent to client (after adaptation)
4507 [http::]<sH Reply high offset sent
4508 [http::]<sS Upstream object size
4510 [http::]<bs Number of HTTP-equivalent message body bytes
4511 received from the next hop, excluding chunked
4512 transfer encoding and control messages.
4513 Generated FTP/Gopher listings are treated as
4518 [http::]<pt Peer response time in milliseconds. The timer starts
4519 when the last request byte is sent to the next hop
4520 and stops when the last response byte is received.
4521 [http::]<tt Total time in milliseconds. The timer
4522 starts with the first connect request (or write I/O)
4523 sent to the first selected peer. The timer stops
4524 with the last I/O with the last peer.
4526 Squid handling related format codes:
4528 Ss Squid request status (TCP_MISS etc)
4529 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4531 SSL-related format codes:
4533 ssl::bump_mode SslBump decision for the transaction:
4535 For CONNECT requests that initiated bumping of
4536 a connection and for any request received on
4537 an already bumped connection, Squid logs the
4538 corresponding SslBump mode ("splice", "bump",
4539 "peek", "stare", "terminate", "server-first"
4540 or "client-first"). See the ssl_bump option
4541 for more information about these modes.
4543 A "none" token is logged for requests that
4544 triggered "ssl_bump" ACL evaluation matching
4547 In all other cases, a single dash ("-") is
4550 ssl::>sni SSL client SNI sent to Squid.
4553 The Subject field of the received client
4554 SSL certificate or a dash ('-') if Squid has
4555 received an invalid/malformed certificate or
4556 no certificate at all. Consider encoding the
4557 logged value because Subject often has spaces.
4560 The Issuer field of the received client
4561 SSL certificate or a dash ('-') if Squid has
4562 received an invalid/malformed certificate or
4563 no certificate at all. Consider encoding the
4564 logged value because Issuer often has spaces.
4567 The Subject field of the received server
4568 TLS certificate or a dash ('-') if this is
4569 not available. Consider encoding the logged
4570 value because Subject often has spaces.
4573 The Issuer field of the received server
4574 TLS certificate or a dash ('-') if this is
4575 not available. Consider encoding the logged
4576 value because Issuer often has spaces.
4579 The list of certificate validation errors
4580 detected by Squid (including OpenSSL and
4581 certificate validation helper components). The
4582 errors are listed in the discovery order. By
4583 default, the error codes are separated by ':'.
4584 Accepts an optional separator argument.
4586 %ssl::>negotiated_version The negotiated TLS version of the
4589 %ssl::<negotiated_version The negotiated TLS version of the
4590 last server or peer connection.
4592 %ssl::>received_hello_version The TLS version of the Hello
4593 message received from TLS client.
4595 %ssl::<received_hello_version The TLS version of the Hello
4596 message received from TLS server.
4598 %ssl::>received_supported_version The maximum TLS version
4599 supported by the TLS client.
4601 %ssl::<received_supported_version The maximum TLS version
4602 supported by the TLS server.
4604 %ssl::>negotiated_cipher The negotiated cipher of the
4607 %ssl::<negotiated_cipher The negotiated cipher of the
4608 last server or peer connection.
4610 If ICAP is enabled, the following code becomes available (as
4611 well as ICAP log codes documented with the icap_log option):
4613 icap::tt Total ICAP processing time for the HTTP
4614 transaction. The timer ticks when ICAP
4615 ACLs are checked and when ICAP
4616 transaction is in progress.
4618 If adaptation is enabled the following codes become available:
4620 adapt::<last_h The header of the last ICAP response or
4621 meta-information from the last eCAP
4622 transaction related to the HTTP transaction.
4623 Like <h, accepts an optional header name
4626 adapt::sum_trs Summed adaptation transaction response
4627 times recorded as a comma-separated list in
4628 the order of transaction start time. Each time
4629 value is recorded as an integer number,
4630 representing response time of one or more
4631 adaptation (ICAP or eCAP) transaction in
4632 milliseconds. When a failed transaction is
4633 being retried or repeated, its time is not
4634 logged individually but added to the
4635 replacement (next) transaction. See also:
4638 adapt::all_trs All adaptation transaction response times.
4639 Same as adaptation_strs but response times of
4640 individual transactions are never added
4641 together. Instead, all transaction response
4642 times are recorded individually.
4644 You can prefix adapt::*_trs format codes with adaptation
4645 service name in curly braces to record response time(s) specific
4646 to that service. For example: %{my_service}adapt::sum_trs
4648 The default formats available (which do not need re-defining) are:
4650 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
4651 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
4652 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
4653 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
4654 logformat useragent %>a [%tl] "%{User-Agent}>h"
4656 NOTE: When the log_mime_hdrs directive is set to ON.
4657 The squid, common and combined formats have a safely encoded copy
4658 of the mime headers appended to each line within a pair of brackets.
4660 NOTE: The common and combined formats are not quite true to the Apache definition.
4661 The logs from Squid contain an extra status and hierarchy code appended.
4665 NAME: access_log cache_access_log
4667 LOC: Config.Log.accesslogs
4668 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
4670 Configures whether and how Squid logs HTTP and ICP transactions.
4671 If access logging is enabled, a single line is logged for every
4672 matching HTTP or ICP request. The recommended directive formats are:
4674 access_log <module>:<place> [option ...] [acl acl ...]
4675 access_log none [acl acl ...]
4677 The following directive format is accepted but may be deprecated:
4678 access_log <module>:<place> [<logformat name> [acl acl ...]]
4680 In most cases, the first ACL name must not contain the '=' character
4681 and should not be equal to an existing logformat name. You can always
4682 start with an 'all' ACL to work around those restrictions.
4684 Will log to the specified module:place using the specified format (which
4685 must be defined in a logformat directive) those entries which match
4686 ALL the acl's specified (which must be defined in acl clauses).
4687 If no acl is specified, all requests will be logged to this destination.
4689 ===== Available options for the recommended directive format =====
4691 logformat=name Names log line format (either built-in or
4692 defined by a logformat directive). Defaults
4695 buffer-size=64KB Defines approximate buffering limit for log
4696 records (see buffered_logs). Squid should not
4697 keep more than the specified size and, hence,
4698 should flush records before the buffer becomes
4699 full to avoid overflows under normal
4700 conditions (the exact flushing algorithm is
4701 module-dependent though). The on-error option
4702 controls overflow handling.
4704 on-error=die|drop Defines action on unrecoverable errors. The
4705 'drop' action ignores (i.e., does not log)
4706 affected log records. The default 'die' action
4707 kills the affected worker. The drop action
4708 support has not been tested for modules other
4711 rotate=N Specifies the number of log file rotations to
4712 make when you run 'squid -k rotate'. The default
4713 is to obey the logfile_rotate directive. Setting
4714 rotate=0 will disable the file name rotation,
4715 but the log files are still closed and re-opened.
4716 This will enable you to rename the logfiles
4717 yourself just before sending the rotate signal.
4718 Only supported by the stdio module.
4720 ===== Modules Currently available =====
4722 none Do not log any requests matching these ACL.
4723 Do not specify Place or logformat name.
4725 stdio Write each log line to disk immediately at the completion of
4727 Place: the filename and path to be written.
4729 daemon Very similar to stdio. But instead of writing to disk the log
4730 line is passed to a daemon helper for asychronous handling instead.
4731 Place: varies depending on the daemon.
4733 log_file_daemon Place: the file name and path to be written.
4735 syslog To log each request via syslog facility.
4736 Place: The syslog facility and priority level for these entries.
4737 Place Format: facility.priority
4739 where facility could be any of:
4740 authpriv, daemon, local0 ... local7 or user.
4742 And priority could be any of:
4743 err, warning, notice, info, debug.
4745 udp To send each log line as text data to a UDP receiver.
4746 Place: The destination host name or IP and port.
4747 Place Format: //host:port
4749 tcp To send each log line as text data to a TCP receiver.
4750 Lines may be accumulated before sending (see buffered_logs).
4751 Place: The destination host name or IP and port.
4752 Place Format: //host:port
4755 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4761 LOC: Config.Log.icaplogs
4764 ICAP log files record ICAP transaction summaries, one line per
4767 The icap_log option format is:
4768 icap_log <filepath> [<logformat name> [acl acl ...]]
4769 icap_log none [acl acl ...]]
4771 Please see access_log option documentation for details. The two
4772 kinds of logs share the overall configuration approach and many
4775 ICAP processing of a single HTTP message or transaction may
4776 require multiple ICAP transactions. In such cases, multiple
4777 ICAP transaction log lines will correspond to a single access
4780 ICAP log supports many access.log logformat %codes. In ICAP context,
4781 HTTP message-related %codes are applied to the HTTP message embedded
4782 in an ICAP message. Logformat "%http::>..." codes are used for HTTP
4783 messages embedded in ICAP requests while "%http::<..." codes are used
4784 for HTTP messages embedded in ICAP responses. For example:
4786 http::>h To-be-adapted HTTP message headers sent by Squid to
4787 the ICAP service. For REQMOD transactions, these are
4788 HTTP request headers. For RESPMOD, these are HTTP
4789 response headers, but Squid currently cannot log them
4790 (i.e., %http::>h will expand to "-" for RESPMOD).
4792 http::<h Adapted HTTP message headers sent by the ICAP
4793 service to Squid (i.e., HTTP request headers in regular
4794 REQMOD; HTTP response headers in RESPMOD and during
4795 request satisfaction in REQMOD).
4797 ICAP OPTIONS transactions do not embed HTTP messages.
4799 Several logformat codes below deal with ICAP message bodies. An ICAP
4800 message body, if any, typically includes a complete HTTP message
4801 (required HTTP headers plus optional HTTP message body). When
4802 computing HTTP message body size for these logformat codes, Squid
4803 either includes or excludes chunked encoding overheads; see
4804 code-specific documentation for details.
4806 For Secure ICAP services, all size-related information is currently
4807 computed before/after TLS encryption/decryption, as if TLS was not
4810 The following format codes are also available for ICAP logs:
4812 icap::<A ICAP server IP address. Similar to <A.
4814 icap::<service_name ICAP service name from the icap_service
4815 option in Squid configuration file.
4817 icap::ru ICAP Request-URI. Similar to ru.
4819 icap::rm ICAP request method (REQMOD, RESPMOD, or
4820 OPTIONS). Similar to existing rm.
4822 icap::>st The total size of the ICAP request sent to the ICAP
4823 server (ICAP headers + ICAP body), including chunking
4826 icap::<st The total size of the ICAP response received from the
4827 ICAP server (ICAP headers + ICAP body), including
4828 chunking metadata (if any).
4830 icap::<bs The size of the ICAP response body received from the
4831 ICAP server, excluding chunking metadata (if any).
4833 icap::tr Transaction response time (in
4834 milliseconds). The timer starts when
4835 the ICAP transaction is created and
4836 stops when the transaction is completed.
4839 icap::tio Transaction I/O time (in milliseconds). The
4840 timer starts when the first ICAP request
4841 byte is scheduled for sending. The timers
4842 stops when the last byte of the ICAP response
4845 icap::to Transaction outcome: ICAP_ERR* for all
4846 transaction errors, ICAP_OPT for OPTION
4847 transactions, ICAP_ECHO for 204
4848 responses, ICAP_MOD for message
4849 modification, and ICAP_SAT for request
4850 satisfaction. Similar to Ss.
4852 icap::Hs ICAP response status code. Similar to Hs.
4854 icap::>h ICAP request header(s). Similar to >h.
4856 icap::<h ICAP response header(s). Similar to <h.
4858 The default ICAP log format, which can be used without an explicit
4859 definition, is called icap_squid:
4861 logformat icap_squid %ts.%03tu %6icap::tr %>A %icap::to/%03icap::Hs %icap::<st %icap::rm %icap::ru %un -/%icap::<A -
4863 See also: logformat and %adapt::<last_h
4866 NAME: logfile_daemon
4868 DEFAULT: @DEFAULT_LOGFILED@
4869 LOC: Log::TheConfig.logfile_daemon
4871 Specify the path to the logfile-writing daemon. This daemon is
4872 used to write the access and store logs, if configured.
4874 Squid sends a number of commands to the log daemon:
4875 L<data>\n - logfile data
4880 r<n>\n - set rotate count to <n>
4881 b<n>\n - 1 = buffer output, 0 = don't buffer output
4883 No responses is expected.
4886 NAME: stats_collection
4888 LOC: Config.accessList.stats_collection
4890 DEFAULT_DOC: Allow logging for all transactions.
4891 COMMENT: allow|deny acl acl...
4893 This options allows you to control which requests gets accounted
4894 in performance counters.
4896 This clause only supports fast acl types.
4897 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4900 NAME: cache_store_log
4903 LOC: Config.Log.store
4905 Logs the activities of the storage manager. Shows which
4906 objects are ejected from the cache, and which objects are
4907 saved and for how long.
4908 There are not really utilities to analyze this data, so you can safely
4909 disable it (the default).
4911 Store log uses modular logging outputs. See access_log for the list
4912 of modules supported.
4915 cache_store_log stdio:@DEFAULT_STORE_LOG@
4916 cache_store_log daemon:@DEFAULT_STORE_LOG@
4919 NAME: cache_swap_state cache_swap_log
4921 LOC: Config.Log.swap
4923 DEFAULT_DOC: Store the journal inside its cache_dir
4925 Location for the cache "swap.state" file. This index file holds
4926 the metadata of objects saved on disk. It is used to rebuild
4927 the cache during startup. Normally this file resides in each
4928 'cache_dir' directory, but you may specify an alternate
4929 pathname here. Note you must give a full filename, not just
4930 a directory. Since this is the index for the whole object
4931 list you CANNOT periodically rotate it!
4933 If %s can be used in the file name it will be replaced with a
4934 a representation of the cache_dir name where each / is replaced
4935 with '.'. This is needed to allow adding/removing cache_dir
4936 lines when cache_swap_log is being used.
4938 If have more than one 'cache_dir', and %s is not used in the name
4939 these swap logs will have names such as:
4945 The numbered extension (which is added automatically)
4946 corresponds to the order of the 'cache_dir' lines in this
4947 configuration file. If you change the order of the 'cache_dir'
4948 lines in this file, these index files will NOT correspond to
4949 the correct 'cache_dir' entry (unless you manually rename
4950 them). We recommend you do NOT use this option. It is
4951 better to keep these index files in each 'cache_dir' directory.
4954 NAME: logfile_rotate
4957 LOC: Config.Log.rotateNumber
4959 Specifies the default number of logfile rotations to make when you
4960 type 'squid -k rotate'. The default is 10, which will rotate
4961 with extensions 0 through 9. Setting logfile_rotate to 0 will
4962 disable the file name rotation, but the logfiles are still closed
4963 and re-opened. This will enable you to rename the logfiles
4964 yourself just before sending the rotate signal.
4966 Note, from Squid-3.1 this option is only a default for cache.log,
4967 that log can be rotated separately by using debug_options.
4969 Note, from Squid-3.6 this option is only a default for access.log
4970 recorded by stdio: module. Those logs can be rotated separately by
4971 using the rotate=N option on their access_log directive.
4973 Note, the 'squid -k rotate' command normally sends a USR1
4974 signal to the running squid process. In certain situations
4975 (e.g. on Linux with Async I/O), USR1 is used for other
4976 purposes, so -k rotate uses another signal. It is best to get
4977 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4984 DEFAULT: @DEFAULT_MIME_TABLE@
4985 LOC: Config.mimeTablePathname
4987 Path to Squid's icon configuration file.
4989 You shouldn't need to change this, but the default file contains
4990 examples and formatting information if you do.
4996 LOC: Config.onoff.log_mime_hdrs
4999 The Cache can record both the request and the response MIME
5000 headers for each HTTP transaction. The headers are encoded
5001 safely and will appear as two bracketed fields at the end of
5002 the access log (for either the native or httpd-emulated log
5003 formats). To enable this logging set log_mime_hdrs to 'on'.
5008 DEFAULT: @DEFAULT_PID_FILE@
5009 LOC: Config.pidFilename
5011 A filename to write the process-id to. To disable, enter "none".
5014 NAME: client_netmask
5016 LOC: Config.Addrs.client_netmask
5018 DEFAULT_DOC: Log full client IP address
5020 A netmask for client addresses in logfiles and cachemgr output.
5021 Change this to protect the privacy of your cache clients.
5022 A netmask of 255.255.255.0 will log all IP's in that range with
5023 the last digit set to '0'.
5026 NAME: strip_query_terms
5028 LOC: Config.onoff.strip_query_terms
5031 By default, Squid strips query terms from requested URLs before
5032 logging. This protects your user's privacy and reduces log size.
5034 When investigating HIT/MISS or other caching behaviour you
5035 will need to disable this to see the full URL used by Squid.
5042 LOC: Config.onoff.buffered_logs
5044 Whether to write/send access_log records ASAP or accumulate them and
5045 then write/send them in larger chunks. Buffering may improve
5046 performance because it decreases the number of I/Os. However,
5047 buffering increases the delay before log records become available to
5048 the final recipient (e.g., a disk file or logging daemon) and,
5049 hence, increases the risk of log records loss.
5051 Note that even when buffered_logs are off, Squid may have to buffer
5052 records if it cannot write/send them immediately due to pending I/Os
5053 (e.g., the I/O writing the previous log record) or connectivity loss.
5055 Currently honored by 'daemon' and 'tcp' access_log modules only.
5058 NAME: netdb_filename
5060 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
5061 LOC: Config.netdbFilename
5064 Where Squid stores it's netdb journal.
5065 When enabled this journal preserves netdb state between restarts.
5067 To disable, enter "none".
5071 OPTIONS FOR TROUBLESHOOTING
5072 -----------------------------------------------------------------------------
5077 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
5078 LOC: Debug::cache_log
5080 Squid administrative logging file.
5082 This is where general information about Squid behavior goes. You can
5083 increase the amount of data logged to this file and how often it is
5084 rotated with "debug_options"
5090 DEFAULT_DOC: Log all critical and important messages.
5091 LOC: Debug::debugOptions
5093 Logging options are set as section,level where each source file
5094 is assigned a unique section. Lower levels result in less
5095 output, Full debugging (level 9) can result in a very large
5096 log file, so be careful.
5098 The magic word "ALL" sets debugging levels for all sections.
5099 The default is to run with "ALL,1" to record important warnings.
5101 The rotate=N option can be used to keep more or less of these logs
5102 than would otherwise be kept by logfile_rotate.
5103 For most uses a single log should be enough to monitor current
5104 events affecting Squid.
5109 LOC: Config.coredump_dir
5110 DEFAULT_IF_NONE: none
5111 DEFAULT_DOC: Use the directory from where Squid was started.
5113 By default Squid leaves core files in the directory from where
5114 it was started. If you set 'coredump_dir' to a directory
5115 that exists, Squid will chdir() to that directory at startup
5116 and coredump files will be left there.
5120 # Leave coredumps in the first cache dir
5121 coredump_dir @DEFAULT_SWAP_DIR@
5127 OPTIONS FOR FTP GATEWAYING
5128 -----------------------------------------------------------------------------
5134 LOC: Config.Ftp.anon_user
5136 If you want the anonymous login password to be more informative
5137 (and enable the use of picky FTP servers), set this to something
5138 reasonable for your domain, like wwwuser@somewhere.net
5140 The reason why this is domainless by default is the
5141 request can be made on the behalf of a user in any domain,
5142 depending on how the cache is used.
5143 Some FTP server also validate the email address is valid
5144 (for example perl.com).
5150 LOC: Config.Ftp.passive
5152 If your firewall does not allow Squid to use passive
5153 connections, turn off this option.
5155 Use of ftp_epsv_all option requires this to be ON.
5161 LOC: Config.Ftp.epsv_all
5163 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
5165 NATs may be able to put the connection on a "fast path" through the
5166 translator, as the EPRT command will never be used and therefore,
5167 translation of the data portion of the segments will never be needed.
5169 When a client only expects to do two-way FTP transfers this may be
5171 If squid finds that it must do a three-way FTP transfer after issuing
5172 an EPSV ALL command, the FTP session will fail.
5174 If you have any doubts about this option do not use it.
5175 Squid will nicely attempt all other connection methods.
5177 Requires ftp_passive to be ON (default) for any effect.
5183 LOC: Config.accessList.ftp_epsv
5185 FTP Protocol extensions permit the use of a special "EPSV" command.
5187 NATs may be able to put the connection on a "fast path" through the
5188 translator using EPSV, as the EPRT command will never be used
5189 and therefore, translation of the data portion of the segments
5190 will never be needed.
5192 EPSV is often required to interoperate with FTP servers on IPv6
5193 networks. On the other hand, it may break some IPv4 servers.
5195 By default, EPSV may try EPSV with any FTP server. To fine tune
5196 that decision, you may restrict EPSV to certain clients or servers
5199 ftp_epsv allow|deny al1 acl2 ...
5201 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
5203 Only fast ACLs are supported.
5204 Requires ftp_passive to be ON (default) for any effect.
5210 LOC: Config.Ftp.eprt
5212 FTP Protocol extensions permit the use of a special "EPRT" command.
5214 This extension provides a protocol neutral alternative to the
5215 IPv4-only PORT command. When supported it enables active FTP data
5216 channels over IPv6 and efficient NAT handling.
5218 Turning this OFF will prevent EPRT being attempted and will skip
5219 straight to using PORT for IPv4 servers.
5221 Some devices are known to not handle this extension correctly and
5222 may result in crashes. Devices which suport EPRT enough to fail
5223 cleanly will result in Squid attempting PORT anyway. This directive
5224 should only be disabled when EPRT results in device failures.
5226 WARNING: Doing so will convert Squid back to the old behavior with all
5227 the related problems with external NAT devices/layers and IPv4-only FTP.
5230 NAME: ftp_sanitycheck
5233 LOC: Config.Ftp.sanitycheck
5235 For security and data integrity reasons Squid by default performs
5236 sanity checks of the addresses of FTP data connections ensure the
5237 data connection is to the requested server. If you need to allow
5238 FTP connections to servers using another IP address for the data
5239 connection turn this off.
5242 NAME: ftp_telnet_protocol
5245 LOC: Config.Ftp.telnet
5247 The FTP protocol is officially defined to use the telnet protocol
5248 as transport channel for the control connection. However, many
5249 implementations are broken and does not respect this aspect of
5252 If you have trouble accessing files with ASCII code 255 in the
5253 path or similar problems involving this ASCII code you can
5254 try setting this directive to off. If that helps, report to the
5255 operator of the FTP server in question that their FTP server
5256 is broken and does not follow the FTP standard.
5260 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
5261 -----------------------------------------------------------------------------
5266 DEFAULT: @DEFAULT_DISKD@
5267 LOC: Config.Program.diskd
5269 Specify the location of the diskd executable.
5270 Note this is only useful if you have compiled in
5271 diskd as one of the store io modules.
5274 NAME: unlinkd_program
5277 DEFAULT: @DEFAULT_UNLINKD@
5278 LOC: Config.Program.unlinkd
5280 Specify the location of the executable for file deletion process.
5283 NAME: pinger_program
5286 DEFAULT: @DEFAULT_PINGER@
5289 Specify the location of the executable for the pinger process.
5298 Control whether the pinger is active at run-time.
5299 Enables turning ICMP pinger on and off with a simple
5300 squid -k reconfigure.
5305 OPTIONS FOR URL REWRITING
5306 -----------------------------------------------------------------------------
5309 NAME: url_rewrite_program redirect_program
5311 LOC: Config.Program.redirect
5314 Specify the location of the executable URL rewriter to use.
5315 Since they can perform almost any function there isn't one included.
5317 For each requested URL, the rewriter will receive on line with the format
5319 [channel-ID <SP>] URL [<SP> extras]<NL>
5321 See url_rewrite_extras on how to send "extras" with optional values to
5323 After processing the request the helper must reply using the following format:
5325 [channel-ID <SP>] result [<SP> kv-pairs]
5327 The result code can be:
5329 OK status=30N url="..."
5330 Redirect the URL to the one supplied in 'url='.
5331 'status=' is optional and contains the status code to send
5332 the client in Squids HTTP response. It must be one of the
5333 HTTP redirect status codes: 301, 302, 303, 307, 308.
5334 When no status is given Squid will use 302.
5336 OK rewrite-url="..."
5337 Rewrite the URL to the one supplied in 'rewrite-url='.
5338 The new URL is fetched directly by Squid and returned to
5339 the client as the response to its request.
5342 When neither of url= and rewrite-url= are sent Squid does
5346 Do not change the URL.
5349 An internal error occurred in the helper, preventing
5350 a result being identified. The 'message=' key name is
5351 reserved for delivering a log message.
5354 In addition to the above kv-pairs Squid also understands the following
5355 optional kv-pairs received from URL rewriters:
5357 Associates a TAG with the client TCP connection.
5358 The TAG is treated as a regular annotation but persists across
5359 future requests on the client connection rather than just the
5360 current request. A helper may update the TAG during subsequent
5361 requests be returning a new kv-pair.
5363 When using the concurrency= option the protocol is changed by
5364 introducing a query channel tag in front of the request/response.
5365 The query channel tag is a number between 0 and concurrency-1.
5366 This value must be echoed back unchanged to Squid as the first part
5367 of the response relating to its request.
5369 WARNING: URL re-writing ability should be avoided whenever possible.
5370 Use the URL redirect form of response instead.
5372 Re-write creates a difference in the state held by the client
5373 and server. Possibly causing confusion when the server response
5374 contains snippets of its view state. Embeded URLs, response
5375 and content Location headers, etc. are not re-written by this
5378 By default, a URL rewriter is not used.
5381 NAME: url_rewrite_children redirect_children
5382 TYPE: HelperChildConfig
5383 DEFAULT: 20 startup=0 idle=1 concurrency=0
5384 LOC: Config.redirectChildren
5386 The maximum number of redirector processes to spawn. If you limit
5387 it too few Squid will have to wait for them to process a backlog of
5388 URLs, slowing it down. If you allow too many they will use RAM
5389 and other system resources noticably.
5391 The startup= and idle= options allow some measure of skew in your
5396 Sets a minimum of how many processes are to be spawned when Squid
5397 starts or reconfigures. When set to zero the first request will
5398 cause spawning of the first child process to handle it.
5400 Starting too few will cause an initial slowdown in traffic as Squid
5401 attempts to simultaneously spawn enough processes to cope.
5405 Sets a minimum of how many processes Squid is to try and keep available
5406 at all times. When traffic begins to rise above what the existing
5407 processes can handle this many more will be spawned up to the maximum
5408 configured. A minimum setting of 1 is required.
5412 The number of requests each redirector helper can handle in
5413 parallel. Defaults to 0 which indicates the redirector
5414 is a old-style single threaded redirector.
5416 When this directive is set to a value >= 1 then the protocol
5417 used to communicate with the helper is modified to include
5418 an ID in front of the request/response. The ID from the request
5419 must be echoed back with the response to that request.
5423 Sets the maximum number of queued requests to N. The default maximum
5424 is 2*numberofchildren. If the queued requests exceed queue size and
5425 redirector_bypass configuration option is set, then redirector is bypassed.
5426 Otherwise, Squid is allowed to temporarily exceed the configured maximum,
5427 marking the affected helper as "overloaded". If the helper overload lasts
5428 more than 3 minutes, the action prescribed by the on-persistent-overload
5431 on-persistent-overload=action
5433 Specifies Squid reaction to a new helper request arriving when the helper
5434 has been overloaded for more that 3 minutes already. The number of queued
5435 requests determines whether the helper is overloaded (see the queue-size
5438 Two actions are supported:
5440 die Squid worker quits. This is the default behavior.
5442 ERR Squid treats the helper request as if it was
5443 immediately submitted, and the helper immediately
5444 replied with an ERR response. This action has no effect
5445 on the already queued and in-progress helper requests.
5448 NAME: url_rewrite_host_header redirect_rewrites_host_header
5451 LOC: Config.onoff.redir_rewrites_host
5453 To preserve same-origin security policies in browsers and
5454 prevent Host: header forgery by redirectors Squid rewrites
5455 any Host: header in redirected requests.
5457 If you are running an accelerator this may not be a wanted
5458 effect of a redirector. This directive enables you disable
5459 Host: alteration in reverse-proxy traffic.
5461 WARNING: Entries are cached on the result of the URL rewriting
5462 process, so be careful if you have domain-virtual hosts.
5464 WARNING: Squid and other software verifies the URL and Host
5465 are matching, so be careful not to relay through other proxies
5466 or inspecting firewalls with this disabled.
5469 NAME: url_rewrite_access redirector_access
5472 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5473 LOC: Config.accessList.redirector
5475 If defined, this access list specifies which requests are
5476 sent to the redirector processes.
5478 This clause supports both fast and slow acl types.
5479 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5482 NAME: url_rewrite_bypass redirector_bypass
5484 LOC: Config.onoff.redirector_bypass
5487 When this is 'on', a request will not go through the
5488 redirector if all the helpers are busy. If this is 'off' and the
5489 redirector queue grows too large, the action is prescribed by the
5490 on-persistent-overload option. You should only enable this if the
5491 redirectors are not critical to your caching system. If you use
5492 redirectors for access control, and you enable this option,
5493 users may have access to pages they should not
5494 be allowed to request.
5495 This options sets default queue-size option of the url_rewrite_children
5499 NAME: url_rewrite_extras
5500 TYPE: TokenOrQuotedString
5501 LOC: Config.redirector_extras
5502 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5504 Specifies a string to be append to request line format for the
5505 rewriter helper. "Quoted" format values may contain spaces and
5506 logformat %macros. In theory, any logformat %macro can be used.
5507 In practice, a %macro expands as a dash (-) if the helper request is
5508 sent before the required macro information is available to Squid.
5511 NAME: url_rewrite_timeout
5512 TYPE: UrlHelperTimeout
5513 LOC: Config.onUrlRewriteTimeout
5515 DEFAULT_DOC: Squid waits for the helper response forever
5517 Squid times active requests to redirector. The timeout value and Squid
5518 reaction to a timed out request are configurable using the following
5521 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
5523 supported timeout actions:
5524 fail Squid return a ERR_GATEWAY_FAILURE error page
5526 bypass Do not re-write the URL
5528 retry Send the lookup to the helper again
5530 use_configured_response
5531 Use the <quoted-response> as helper response
5535 OPTIONS FOR STORE ID
5536 -----------------------------------------------------------------------------
5539 NAME: store_id_program storeurl_rewrite_program
5541 LOC: Config.Program.store_id
5544 Specify the location of the executable StoreID helper to use.
5545 Since they can perform almost any function there isn't one included.
5547 For each requested URL, the helper will receive one line with the format
5549 [channel-ID <SP>] URL [<SP> extras]<NL>
5552 After processing the request the helper must reply using the following format:
5554 [channel-ID <SP>] result [<SP> kv-pairs]
5556 The result code can be:
5559 Use the StoreID supplied in 'store-id='.
5562 The default is to use HTTP request URL as the store ID.
5565 An internal error occured in the helper, preventing
5566 a result being identified.
5568 In addition to the above kv-pairs Squid also understands the following
5569 optional kv-pairs received from URL rewriters:
5571 Associates a TAG with the client TCP connection.
5572 Please see url_rewrite_program related documentation for this
5575 Helper programs should be prepared to receive and possibly ignore
5576 additional whitespace-separated tokens on each input line.
5578 When using the concurrency= option the protocol is changed by
5579 introducing a query channel tag in front of the request/response.
5580 The query channel tag is a number between 0 and concurrency-1.
5581 This value must be echoed back unchanged to Squid as the first part
5582 of the response relating to its request.
5584 NOTE: when using StoreID refresh_pattern will apply to the StoreID
5585 returned from the helper and not the URL.
5587 WARNING: Wrong StoreID value returned by a careless helper may result
5588 in the wrong cached response returned to the user.
5590 By default, a StoreID helper is not used.
5593 NAME: store_id_extras
5594 TYPE: TokenOrQuotedString
5595 LOC: Config.storeId_extras
5596 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5598 Specifies a string to be append to request line format for the
5599 StoreId helper. "Quoted" format values may contain spaces and
5600 logformat %macros. In theory, any logformat %macro can be used.
5601 In practice, a %macro expands as a dash (-) if the helper request is
5602 sent before the required macro information is available to Squid.
5605 NAME: store_id_children storeurl_rewrite_children
5606 TYPE: HelperChildConfig
5607 DEFAULT: 20 startup=0 idle=1 concurrency=0
5608 LOC: Config.storeIdChildren
5610 The maximum number of StoreID helper processes to spawn. If you limit
5611 it too few Squid will have to wait for them to process a backlog of
5612 requests, slowing it down. If you allow too many they will use RAM
5613 and other system resources noticably.
5615 The startup= and idle= options allow some measure of skew in your
5620 Sets a minimum of how many processes are to be spawned when Squid
5621 starts or reconfigures. When set to zero the first request will
5622 cause spawning of the first child process to handle it.
5624 Starting too few will cause an initial slowdown in traffic as Squid
5625 attempts to simultaneously spawn enough processes to cope.
5629 Sets a minimum of how many processes Squid is to try and keep available
5630 at all times. When traffic begins to rise above what the existing
5631 processes can handle this many more will be spawned up to the maximum
5632 configured. A minimum setting of 1 is required.
5636 The number of requests each storeID helper can handle in
5637 parallel. Defaults to 0 which indicates the helper
5638 is a old-style single threaded program.
5640 When this directive is set to a value >= 1 then the protocol
5641 used to communicate with the helper is modified to include
5642 an ID in front of the request/response. The ID from the request
5643 must be echoed back with the response to that request.
5647 Sets the maximum number of queued requests to N. The default maximum
5648 is 2*numberofchildren. If the queued requests exceed queue size and
5649 redirector_bypass configuration option is set, then redirector is bypassed.
5650 Otherwise, Squid is allowed to temporarily exceed the configured maximum,
5651 marking the affected helper as "overloaded". If the helper overload lasts
5652 more than 3 minutes, the action prescribed by the on-persistent-overload
5655 on-persistent-overload=action
5657 Specifies Squid reaction to a new helper request arriving when the helper
5658 has been overloaded for more that 3 minutes already. The number of queued
5659 requests determines whether the helper is overloaded (see the queue-size
5662 Two actions are supported:
5664 die Squid worker quits. This is the default behavior.
5666 ERR Squid treats the helper request as if it was
5667 immediately submitted, and the helper immediately
5668 replied with an ERR response. This action has no effect
5669 on the already queued and in-progress helper requests.
5672 NAME: store_id_access storeurl_rewrite_access
5675 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5676 LOC: Config.accessList.store_id
5678 If defined, this access list specifies which requests are
5679 sent to the StoreID processes. By default all requests
5682 This clause supports both fast and slow acl types.
5683 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5686 NAME: store_id_bypass storeurl_rewrite_bypass
5688 LOC: Config.onoff.store_id_bypass
5691 When this is 'on', a request will not go through the
5692 helper if all helpers are busy. If this is 'off' and the helper
5693 queue grows too large, the action is prescribed by the
5694 on-persistent-overload option. You should only enable this if the
5695 helpers are not critical to your caching system. If you use
5696 helpers for critical caching components, and you enable this
5697 option, users may not get objects from cache.
5698 This options sets default queue-size option of the store_id_children
5703 OPTIONS FOR TUNING THE CACHE
5704 -----------------------------------------------------------------------------
5707 NAME: cache no_cache
5710 DEFAULT_DOC: By default, this directive is unused and has no effect.
5711 LOC: Config.accessList.noCache
5713 Requests denied by this directive will not be served from the cache
5714 and their responses will not be stored in the cache. This directive
5715 has no effect on other transactions and on already cached responses.
5717 This clause supports both fast and slow acl types.
5718 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5720 This and the two other similar caching directives listed below are
5721 checked at different transaction processing stages, have different
5722 access to response information, affect different cache operations,
5723 and differ in slow ACLs support:
5725 * cache: Checked before Squid makes a hit/miss determination.
5726 No access to reply information!
5727 Denies both serving a hit and storing a miss.
5728 Supports both fast and slow ACLs.
5729 * send_hit: Checked after a hit was detected.
5730 Has access to reply (hit) information.
5731 Denies serving a hit only.
5732 Supports fast ACLs only.
5733 * store_miss: Checked before storing a cachable miss.
5734 Has access to reply (miss) information.
5735 Denies storing a miss only.
5736 Supports fast ACLs only.
5738 If you are not sure which of the three directives to use, apply the
5739 following decision logic:
5741 * If your ACL(s) are of slow type _and_ need response info, redesign.
5742 Squid does not support that particular combination at this time.
5744 * If your directive ACL(s) are of slow type, use "cache"; and/or
5745 * if your directive ACL(s) need no response info, use "cache".
5747 * If you do not want the response cached, use store_miss; and/or
5748 * if you do not want a hit on a cached response, use send_hit.
5754 DEFAULT_DOC: By default, this directive is unused and has no effect.
5755 LOC: Config.accessList.sendHit
5757 Responses denied by this directive will not be served from the cache
5758 (but may still be cached, see store_miss). This directive has no
5759 effect on the responses it allows and on the cached objects.
5761 Please see the "cache" directive for a summary of differences among
5762 store_miss, send_hit, and cache directives.
5764 Unlike the "cache" directive, send_hit only supports fast acl
5765 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5769 # apply custom Store ID mapping to some URLs
5770 acl MapMe dstdomain .c.example.com
5771 store_id_program ...
5772 store_id_access allow MapMe
5774 # but prevent caching of special responses
5775 # such as 302 redirects that cause StoreID loops
5776 acl Ordinary http_status 200-299
5777 store_miss deny MapMe !Ordinary
5779 # and do not serve any previously stored special responses
5780 # from the cache (in case they were already cached before
5781 # the above store_miss rule was in effect).
5782 send_hit deny MapMe !Ordinary
5788 DEFAULT_DOC: By default, this directive is unused and has no effect.
5789 LOC: Config.accessList.storeMiss
5791 Responses denied by this directive will not be cached (but may still
5792 be served from the cache, see send_hit). This directive has no
5793 effect on the responses it allows and on the already cached responses.
5795 Please see the "cache" directive for a summary of differences among
5796 store_miss, send_hit, and cache directives. See the
5797 send_hit directive for a usage example.
5799 Unlike the "cache" directive, store_miss only supports fast acl
5800 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5806 LOC: Config.maxStale
5809 This option puts an upper limit on how stale content Squid
5810 will serve from the cache if cache validation fails.
5811 Can be overriden by the refresh_pattern max-stale option.
5814 NAME: refresh_pattern
5815 TYPE: refreshpattern
5819 usage: refresh_pattern [-i] regex min percent max [options]
5821 By default, regular expressions are CASE-SENSITIVE. To make
5822 them case-insensitive, use the -i option.
5824 'Min' is the time (in minutes) an object without an explicit
5825 expiry time should be considered fresh. The recommended
5826 value is 0, any higher values may cause dynamic applications
5827 to be erroneously cached unless the application designer
5828 has taken the appropriate actions.
5830 'Percent' is a percentage of the objects age (time since last
5831 modification age) an object without explicit expiry time
5832 will be considered fresh.
5834 'Max' is an upper limit on how long objects without an explicit
5835 expiry time will be considered fresh. The value is also used
5836 to form Cache-Control: max-age header for a request sent from
5837 Squid to origin/parent.
5839 options: override-expire
5849 override-expire enforces min age even if the server
5850 sent an explicit expiry time (e.g., with the
5851 Expires: header or Cache-Control: max-age). Doing this
5852 VIOLATES the HTTP standard. Enabling this feature
5853 could make you liable for problems which it causes.
5855 Note: override-expire does not enforce staleness - it only extends
5856 freshness / min. If the server returns a Expires time which
5857 is longer than your max time, Squid will still consider
5858 the object fresh for that period of time.
5860 override-lastmod enforces min age even on objects
5861 that were modified recently.
5863 reload-into-ims changes a client no-cache or ``reload''
5864 request for a cached entry into a conditional request using
5865 If-Modified-Since and/or If-None-Match headers, provided the
5866 cached entry has a Last-Modified and/or a strong ETag header.
5867 Doing this VIOLATES the HTTP standard. Enabling this feature
5868 could make you liable for problems which it causes.
5870 ignore-reload ignores a client no-cache or ``reload''
5871 header. Doing this VIOLATES the HTTP standard. Enabling
5872 this feature could make you liable for problems which
5875 ignore-no-store ignores any ``Cache-control: no-store''
5876 headers received from a server. Doing this VIOLATES
5877 the HTTP standard. Enabling this feature could make you
5878 liable for problems which it causes.
5880 ignore-private ignores any ``Cache-control: private''
5881 headers received from a server. Doing this VIOLATES
5882 the HTTP standard. Enabling this feature could make you
5883 liable for problems which it causes.
5885 refresh-ims causes squid to contact the origin server
5886 when a client issues an If-Modified-Since request. This
5887 ensures that the client will receive an updated version
5888 if one is available.
5890 store-stale stores responses even if they don't have explicit
5891 freshness or a validator (i.e., Last-Modified or an ETag)
5892 present, or if they're already stale. By default, Squid will
5893 not cache such responses because they usually can't be
5894 reused. Note that such responses will be stale by default.
5896 max-stale=NN provide a maximum staleness factor. Squid won't
5897 serve objects more stale than this even if it failed to
5898 validate the object. Default: use the max_stale global limit.
5900 Basically a cached object is:
5902 FRESH if expire > now, else STALE
5904 FRESH if lm-factor < percent, else STALE
5908 The refresh_pattern lines are checked in the order listed here.
5909 The first entry which matches is used. If none of the entries
5910 match the default will be used.
5912 Note, you must uncomment all the default lines if you want
5913 to change one. The default setting is only active if none is
5919 # Add any of your own refresh_pattern entries above these.
5921 refresh_pattern ^ftp: 1440 20% 10080
5922 refresh_pattern ^gopher: 1440 0% 1440
5923 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5924 refresh_pattern . 0 20% 4320
5928 NAME: quick_abort_min
5932 LOC: Config.quickAbort.min
5935 NAME: quick_abort_max
5939 LOC: Config.quickAbort.max
5942 NAME: quick_abort_pct
5946 LOC: Config.quickAbort.pct
5948 The cache by default continues downloading aborted requests
5949 which are almost completed (less than 16 KB remaining). This
5950 may be undesirable on slow (e.g. SLIP) links and/or very busy
5951 caches. Impatient users may tie up file descriptors and
5952 bandwidth by repeatedly requesting and immediately aborting
5955 When the user aborts a request, Squid will check the
5956 quick_abort values to the amount of data transferred until
5959 If the transfer has less than 'quick_abort_min' KB remaining,
5960 it will finish the retrieval.
5962 If the transfer has more than 'quick_abort_max' KB remaining,
5963 it will abort the retrieval.
5965 If more than 'quick_abort_pct' of the transfer has completed,
5966 it will finish the retrieval.
5968 If you do not want any retrieval to continue after the client
5969 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5972 If you want retrievals to always continue if they are being
5973 cached set 'quick_abort_min' to '-1 KB'.
5976 NAME: read_ahead_gap
5977 COMMENT: buffer-size
5979 LOC: Config.readAheadGap
5982 The amount of data the cache will buffer ahead of what has been
5983 sent to the client when retrieving an object from another server.
5987 IFDEF: USE_HTTP_VIOLATIONS
5990 LOC: Config.negativeTtl
5993 Set the Default Time-to-Live (TTL) for failed requests.
5994 Certain types of failures (such as "connection refused" and
5995 "404 Not Found") are able to be negatively-cached for a short time.
5996 Modern web servers should provide Expires: header, however if they
5997 do not this can provide a minimum TTL.
5998 The default is not to cache errors with unknown expiry details.
6000 Note that this is different from negative caching of DNS lookups.
6002 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6003 this feature could make you liable for problems which it
6007 NAME: positive_dns_ttl
6010 LOC: Config.positiveDnsTtl
6013 Upper limit on how long Squid will cache positive DNS responses.
6014 Default is 6 hours (360 minutes). This directive must be set
6015 larger than negative_dns_ttl.
6018 NAME: negative_dns_ttl
6021 LOC: Config.negativeDnsTtl
6024 Time-to-Live (TTL) for negative caching of failed DNS lookups.
6025 This also sets the lower cache limit on positive lookups.
6026 Minimum value is 1 second, and it is not recommendable to go
6027 much below 10 seconds.
6030 NAME: range_offset_limit
6031 COMMENT: size [acl acl...]
6033 LOC: Config.rangeOffsetLimit
6036 usage: (size) [units] [[!]aclname]
6038 Sets an upper limit on how far (number of bytes) into the file
6039 a Range request may be to cause Squid to prefetch the whole file.
6040 If beyond this limit, Squid forwards the Range request as it is and
6041 the result is NOT cached.
6043 This is to stop a far ahead range request (lets say start at 17MB)
6044 from making Squid fetch the whole object up to that point before
6045 sending anything to the client.
6047 Multiple range_offset_limit lines may be specified, and they will
6048 be searched from top to bottom on each request until a match is found.
6049 The first match found will be used. If no line matches a request, the
6050 default limit of 0 bytes will be used.
6052 'size' is the limit specified as a number of units.
6054 'units' specifies whether to use bytes, KB, MB, etc.
6055 If no units are specified bytes are assumed.
6057 A size of 0 causes Squid to never fetch more than the
6058 client requested. (default)
6060 A size of 'none' causes Squid to always fetch the object from the
6061 beginning so it may cache the result. (2.0 style)
6063 'aclname' is the name of a defined ACL.
6065 NP: Using 'none' as the byte value here will override any quick_abort settings
6066 that may otherwise apply to the range request. The range request will
6067 be fully fetched from start to finish regardless of the client
6068 actions. This affects bandwidth usage.
6071 NAME: minimum_expiry_time
6074 LOC: Config.minimum_expiry_time
6077 The minimum caching time according to (Expires - Date)
6078 headers Squid honors if the object can't be revalidated.
6079 The default is 60 seconds.
6081 In reverse proxy environments it might be desirable to honor
6082 shorter object lifetimes. It is most likely better to make
6083 your server return a meaningful Last-Modified header however.
6085 In ESI environments where page fragments often have short
6086 lifetimes, this will often be best set to 0.
6089 NAME: store_avg_object_size
6093 LOC: Config.Store.avgObjectSize
6095 Average object size, used to estimate number of objects your
6096 cache can hold. The default is 13 KB.
6098 This is used to pre-seed the cache index memory allocation to
6099 reduce expensive reallocate operations while handling clients
6100 traffic. Too-large values may result in memory allocation during
6101 peak traffic, too-small values will result in wasted memory.
6103 Check the cache manager 'info' report metrics for the real
6104 object sizes seen by your Squid before tuning this.
6107 NAME: store_objects_per_bucket
6110 LOC: Config.Store.objectsPerBucket
6112 Target number of objects per bucket in the store hash table.
6113 Lowering this value increases the total number of buckets and
6114 also the storage maintenance rate. The default is 20.
6119 -----------------------------------------------------------------------------
6122 NAME: request_header_max_size
6126 LOC: Config.maxRequestHeaderSize
6128 This specifies the maximum size for HTTP headers in a request.
6129 Request headers are usually relatively small (about 512 bytes).
6130 Placing a limit on the request header size will catch certain
6131 bugs (for example with persistent connections) and possibly
6132 buffer-overflow or denial-of-service attacks.
6135 NAME: reply_header_max_size
6139 LOC: Config.maxReplyHeaderSize
6141 This specifies the maximum size for HTTP headers in a reply.
6142 Reply headers are usually relatively small (about 512 bytes).
6143 Placing a limit on the reply header size will catch certain
6144 bugs (for example with persistent connections) and possibly
6145 buffer-overflow or denial-of-service attacks.
6148 NAME: request_body_max_size
6152 DEFAULT_DOC: No limit.
6153 LOC: Config.maxRequestBodySize
6155 This specifies the maximum size for an HTTP request body.
6156 In other words, the maximum size of a PUT/POST request.
6157 A user who attempts to send a request with a body larger
6158 than this limit receives an "Invalid Request" error message.
6159 If you set this parameter to a zero (the default), there will
6160 be no limit imposed.
6162 See also client_request_buffer_max_size for an alternative
6163 limitation on client uploads which can be configured.
6166 NAME: client_request_buffer_max_size
6170 LOC: Config.maxRequestBufferSize
6172 This specifies the maximum buffer size of a client request.
6173 It prevents squid eating too much memory when somebody uploads
6178 IFDEF: USE_HTTP_VIOLATIONS
6181 DEFAULT_DOC: Obey RFC 2616.
6182 LOC: Config.accessList.brokenPosts
6184 A list of ACL elements which, if matched, causes Squid to send
6185 an extra CRLF pair after the body of a PUT/POST request.
6187 Some HTTP servers has broken implementations of PUT/POST,
6188 and rely on an extra CRLF pair sent by some WWW clients.
6190 Quote from RFC2616 section 4.1 on this matter:
6192 Note: certain buggy HTTP/1.0 client implementations generate an
6193 extra CRLF's after a POST request. To restate what is explicitly
6194 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
6195 a request with an extra CRLF.
6197 This clause only supports fast acl types.
6198 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6201 acl buggy_server url_regex ^http://....
6202 broken_posts allow buggy_server
6205 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
6208 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
6210 LOC: Adaptation::Config::use_indirect_client
6212 Controls whether the indirect client IP address (instead of the direct
6213 client IP address) is passed to adaptation services.
6215 See also: follow_x_forwarded_for adaptation_send_client_ip
6219 IFDEF: USE_HTTP_VIOLATIONS
6223 LOC: Config.onoff.via
6225 If set (default), Squid will include a Via header in requests and
6226 replies as required by RFC2616.
6229 NAME: vary_ignore_expire
6232 LOC: Config.onoff.vary_ignore_expire
6235 Many HTTP servers supporting Vary gives such objects
6236 immediate expiry time with no cache-control header
6237 when requested by a HTTP/1.0 client. This option
6238 enables Squid to ignore such expiry times until
6239 HTTP/1.1 is fully implemented.
6241 WARNING: If turned on this may eventually cause some
6242 varying objects not intended for caching to get cached.
6245 NAME: request_entities
6247 LOC: Config.onoff.request_entities
6250 Squid defaults to deny GET and HEAD requests with request entities,
6251 as the meaning of such requests are undefined in the HTTP standard
6252 even if not explicitly forbidden.
6254 Set this directive to on if you have clients which insists
6255 on sending request entities in GET or HEAD requests. But be warned
6256 that there is server software (both proxies and web servers) which
6257 can fail to properly process this kind of request which may make you
6258 vulnerable to cache pollution attacks if enabled.
6261 NAME: request_header_access
6262 IFDEF: USE_HTTP_VIOLATIONS
6263 TYPE: http_header_access
6264 LOC: Config.request_header_access
6266 DEFAULT_DOC: No limits.
6268 Usage: request_header_access header_name allow|deny [!]aclname ...
6270 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6271 this feature could make you liable for problems which it
6274 This option replaces the old 'anonymize_headers' and the
6275 older 'http_anonymizer' option with something that is much
6276 more configurable. A list of ACLs for each header name allows
6277 removal of specific header fields under specific conditions.
6279 This option only applies to outgoing HTTP request headers (i.e.,
6280 headers sent by Squid to the next HTTP hop such as a cache peer
6281 or an origin server). The option has no effect during cache hit
6282 detection. The equivalent adaptation vectoring point in ICAP
6283 terminology is post-cache REQMOD.
6285 The option is applied to individual outgoing request header
6286 fields. For each request header field F, Squid uses the first
6287 qualifying sets of request_header_access rules:
6289 1. Rules with header_name equal to F's name.
6290 2. Rules with header_name 'Other', provided F's name is not
6291 on the hard-coded list of commonly used HTTP header names.
6292 3. Rules with header_name 'All'.
6294 Within that qualifying rule set, rule ACLs are checked as usual.
6295 If ACLs of an "allow" rule match, the header field is allowed to
6296 go through as is. If ACLs of a "deny" rule match, the header is
6297 removed and request_header_replace is then checked to identify
6298 if the removed header has a replacement. If no rules within the
6299 set have matching ACLs, the header field is left as is.
6301 For example, to achieve the same behavior as the old
6302 'http_anonymizer standard' option, you should use:
6304 request_header_access From deny all
6305 request_header_access Referer deny all
6306 request_header_access User-Agent deny all
6308 Or, to reproduce the old 'http_anonymizer paranoid' feature
6311 request_header_access Authorization allow all
6312 request_header_access Proxy-Authorization allow all
6313 request_header_access Cache-Control allow all
6314 request_header_access Content-Length allow all
6315 request_header_access Content-Type allow all
6316 request_header_access Date allow all
6317 request_header_access Host allow all
6318 request_header_access If-Modified-Since allow all
6319 request_header_access Pragma allow all
6320 request_header_access Accept allow all
6321 request_header_access Accept-Charset allow all
6322 request_header_access Accept-Encoding allow all
6323 request_header_access Accept-Language allow all
6324 request_header_access Connection allow all
6325 request_header_access All deny all
6327 HTTP reply headers are controlled with the reply_header_access directive.
6329 By default, all headers are allowed (no anonymizing is performed).
6332 NAME: reply_header_access
6333 IFDEF: USE_HTTP_VIOLATIONS
6334 TYPE: http_header_access
6335 LOC: Config.reply_header_access
6337 DEFAULT_DOC: No limits.
6339 Usage: reply_header_access header_name allow|deny [!]aclname ...
6341 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6342 this feature could make you liable for problems which it
6345 This option only applies to reply headers, i.e., from the
6346 server to the client.
6348 This is the same as request_header_access, but in the other
6349 direction. Please see request_header_access for detailed
6352 For example, to achieve the same behavior as the old
6353 'http_anonymizer standard' option, you should use:
6355 reply_header_access Server deny all
6356 reply_header_access WWW-Authenticate deny all
6357 reply_header_access Link deny all
6359 Or, to reproduce the old 'http_anonymizer paranoid' feature
6362 reply_header_access Allow allow all
6363 reply_header_access WWW-Authenticate allow all
6364 reply_header_access Proxy-Authenticate allow all
6365 reply_header_access Cache-Control allow all
6366 reply_header_access Content-Encoding allow all
6367 reply_header_access Content-Length allow all
6368 reply_header_access Content-Type allow all
6369 reply_header_access Date allow all
6370 reply_header_access Expires allow all
6371 reply_header_access Last-Modified allow all
6372 reply_header_access Location allow all
6373 reply_header_access Pragma allow all
6374 reply_header_access Content-Language allow all
6375 reply_header_access Retry-After allow all
6376 reply_header_access Title allow all
6377 reply_header_access Content-Disposition allow all
6378 reply_header_access Connection allow all
6379 reply_header_access All deny all
6381 HTTP request headers are controlled with the request_header_access directive.
6383 By default, all headers are allowed (no anonymizing is
6387 NAME: request_header_replace header_replace
6388 IFDEF: USE_HTTP_VIOLATIONS
6389 TYPE: http_header_replace
6390 LOC: Config.request_header_access
6393 Usage: request_header_replace header_name message
6394 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
6396 This option allows you to change the contents of headers
6397 denied with request_header_access above, by replacing them
6398 with some fixed string.
6400 This only applies to request headers, not reply headers.
6402 By default, headers are removed if denied.
6405 NAME: reply_header_replace
6406 IFDEF: USE_HTTP_VIOLATIONS
6407 TYPE: http_header_replace
6408 LOC: Config.reply_header_access
6411 Usage: reply_header_replace header_name message
6412 Example: reply_header_replace Server Foo/1.0
6414 This option allows you to change the contents of headers
6415 denied with reply_header_access above, by replacing them
6416 with some fixed string.
6418 This only applies to reply headers, not request headers.
6420 By default, headers are removed if denied.
6423 NAME: request_header_add
6424 TYPE: HeaderWithAclList
6425 LOC: Config.request_header_add
6428 Usage: request_header_add field-name field-value [ acl ... ]
6429 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6431 This option adds header fields to outgoing HTTP requests (i.e.,
6432 request headers sent by Squid to the next HTTP hop such as a
6433 cache peer or an origin server). The option has no effect during
6434 cache hit detection. The equivalent adaptation vectoring point
6435 in ICAP terminology is post-cache REQMOD.
6437 Field-name is a token specifying an HTTP header name. If a
6438 standard HTTP header name is used, Squid does not check whether
6439 the new header conflicts with any existing headers or violates
6440 HTTP rules. If the request to be modified already contains a
6441 field with the same name, the old field is preserved but the
6442 header field values are not merged.
6444 Field-value is either a token or a quoted string. If quoted
6445 string format is used, then the surrounding quotes are removed
6446 while escape sequences and %macros are processed.
6448 One or more Squid ACLs may be specified to restrict header
6449 injection to matching requests. As always in squid.conf, all
6450 ACLs in the ACL list must be satisfied for the insertion to
6451 happen. The request_header_add supports fast ACLs only.
6453 See also: reply_header_add.
6456 NAME: reply_header_add
6457 TYPE: HeaderWithAclList
6458 LOC: Config.reply_header_add
6461 Usage: reply_header_add field-name field-value [ acl ... ]
6462 Example: reply_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6464 This option adds header fields to outgoing HTTP responses (i.e., response
6465 headers delivered by Squid to the client). This option has no effect on
6466 cache hit detection. The equivalent adaptation vectoring point in
6467 ICAP terminology is post-cache RESPMOD. This option does not apply to
6468 successful CONNECT replies.
6470 Field-name is a token specifying an HTTP header name. If a
6471 standard HTTP header name is used, Squid does not check whether
6472 the new header conflicts with any existing headers or violates
6473 HTTP rules. If the response to be modified already contains a
6474 field with the same name, the old field is preserved but the
6475 header field values are not merged.
6477 Field-value is either a token or a quoted string. If quoted
6478 string format is used, then the surrounding quotes are removed
6479 while escape sequences and %macros are processed.
6481 One or more Squid ACLs may be specified to restrict header
6482 injection to matching responses. As always in squid.conf, all
6483 ACLs in the ACL list must be satisfied for the insertion to
6484 happen. The reply_header_add option supports fast ACLs only.
6486 See also: request_header_add.
6494 This option used to log custom information about the master
6495 transaction. For example, an admin may configure Squid to log
6496 which "user group" the transaction belongs to, where "user group"
6497 will be determined based on a set of ACLs and not [just]
6498 authentication information.
6499 Values of key/value pairs can be logged using %{key}note macros:
6501 note key value acl ...
6502 logformat myFormat ... %{key}note ...
6505 NAME: relaxed_header_parser
6506 COMMENT: on|off|warn
6508 LOC: Config.onoff.relaxed_header_parser
6511 In the default "on" setting Squid accepts certain forms
6512 of non-compliant HTTP messages where it is unambiguous
6513 what the sending application intended even if the message
6514 is not correctly formatted. The messages is then normalized
6515 to the correct form when forwarded by Squid.
6517 If set to "warn" then a warning will be emitted in cache.log
6518 each time such HTTP error is encountered.
6520 If set to "off" then such HTTP errors will cause the request
6521 or response to be rejected.
6524 NAME: collapsed_forwarding
6527 LOC: Config.onoff.collapsed_forwarding
6530 This option controls whether Squid is allowed to merge multiple
6531 potentially cachable requests for the same URI before Squid knows
6532 whether the response is going to be cachable.
6534 When enabled, instead of forwarding each concurrent request for
6535 the same URL, Squid just sends the first of them. The other, so
6536 called "collapsed" requests, wait for the response to the first
6537 request and, if it happens to be cachable, use that response.
6538 Here, "concurrent requests" means "received after the first
6539 request headers were parsed and before the corresponding response
6540 headers were parsed".
6542 This feature is disabled by default: enabling collapsed
6543 forwarding needlessly delays forwarding requests that look
6544 cachable (when they are collapsed) but then need to be forwarded
6545 individually anyway because they end up being for uncachable
6546 content. However, in some cases, such as acceleration of highly
6547 cachable content with periodic or grouped expiration times, the
6548 gains from collapsing [large volumes of simultaneous refresh
6549 requests] outweigh losses from such delays.
6551 Squid collapses two kinds of requests: regular client requests
6552 received on one of the listening ports and internal "cache
6553 revalidation" requests which are triggered by those regular
6554 requests hitting a stale cached object. Revalidation collapsing
6555 is currently disabled for Squid instances containing SMP-aware
6556 disk or memory caches and for Vary-controlled cached objects.
6559 NAME: collapsed_forwarding_shared_entries_limit
6560 COMMENT: (number of entries)
6562 LOC: Config.collapsed_forwarding_shared_entries_limit
6565 This limits the size of a table used for sharing information
6566 about collapsible entries among SMP workers. Limiting sharing
6567 too much results in cache content duplication and missed
6568 collapsing opportunities. Using excessively large values
6569 wastes shared memory.
6571 The limit should be significantly larger then the number of
6572 concurrent collapsible entries one wants to share. For a cache
6573 that handles less than 5000 concurrent requests, the default
6574 setting of 16384 should be plenty.
6576 If the limit is set to zero, it disables sharing of collapsed
6577 forwarding between SMP workers.
6582 -----------------------------------------------------------------------------
6585 NAME: forward_timeout
6588 LOC: Config.Timeout.forward
6591 This parameter specifies how long Squid should at most attempt in
6592 finding a forwarding path for the request before giving up.
6595 NAME: connect_timeout
6598 LOC: Config.Timeout.connect
6601 This parameter specifies how long to wait for the TCP connect to
6602 the requested server or peer to complete before Squid should
6603 attempt to find another path where to forward the request.
6606 NAME: peer_connect_timeout
6609 LOC: Config.Timeout.peer_connect
6612 This parameter specifies how long to wait for a pending TCP
6613 connection to a peer cache. The default is 30 seconds. You
6614 may also set different timeout values for individual neighbors
6615 with the 'connect-timeout' option on a 'cache_peer' line.
6621 LOC: Config.Timeout.read
6624 Applied on peer server connections.
6626 After each successful read(), the timeout will be extended by this
6627 amount. If no data is read again after this amount of time,
6628 the request is aborted and logged with ERR_READ_TIMEOUT.
6630 The default is 15 minutes.
6636 LOC: Config.Timeout.write
6639 This timeout is tracked for all connections that have data
6640 available for writing and are waiting for the socket to become
6641 ready. After each successful write, the timeout is extended by
6642 the configured amount. If Squid has data to write but the
6643 connection is not ready for the configured duration, the
6644 transaction associated with the connection is terminated. The
6645 default is 15 minutes.
6648 NAME: request_timeout
6650 LOC: Config.Timeout.request
6653 How long to wait for complete HTTP request headers after initial
6654 connection establishment.
6657 NAME: request_start_timeout
6659 LOC: Config.Timeout.request_start_timeout
6662 How long to wait for the first request byte after initial
6663 connection establishment.
6666 NAME: client_idle_pconn_timeout persistent_request_timeout
6668 LOC: Config.Timeout.clientIdlePconn
6671 How long to wait for the next HTTP request on a persistent
6672 client connection after the previous request completes.
6675 NAME: ftp_client_idle_timeout
6677 LOC: Config.Timeout.ftpClientIdle
6680 How long to wait for an FTP request on a connection to Squid ftp_port.
6681 Many FTP clients do not deal with idle connection closures well,
6682 necessitating a longer default timeout than client_idle_pconn_timeout
6683 used for incoming HTTP requests.
6686 NAME: client_lifetime
6689 LOC: Config.Timeout.lifetime
6692 The maximum amount of time a client (browser) is allowed to
6693 remain connected to the cache process. This protects the Cache
6694 from having a lot of sockets (and hence file descriptors) tied up
6695 in a CLOSE_WAIT state from remote clients that go away without
6696 properly shutting down (either because of a network failure or
6697 because of a poor client implementation). The default is one
6700 NOTE: The default value is intended to be much larger than any
6701 client would ever need to be connected to your cache. You
6702 should probably change client_lifetime only as a last resort.
6703 If you seem to have many client connections tying up
6704 filedescriptors, we recommend first tuning the read_timeout,
6705 request_timeout, persistent_request_timeout and quick_abort values.
6708 NAME: pconn_lifetime
6711 LOC: Config.Timeout.pconnLifetime
6714 Desired maximum lifetime of a persistent connection.
6715 When set, Squid will close a now-idle persistent connection that
6716 exceeded configured lifetime instead of moving the connection into
6717 the idle connection pool (or equivalent). No effect on ongoing/active
6718 transactions. Connection lifetime is the time period from the
6719 connection acceptance or opening time until "now".
6721 This limit is useful in environments with long-lived connections
6722 where Squid configuration or environmental factors change during a
6723 single connection lifetime. If unrestricted, some connections may
6724 last for hours and even days, ignoring those changes that should
6725 have affected their behavior or their existence.
6727 Currently, a new lifetime value supplied via Squid reconfiguration
6728 has no effect on already idle connections unless they become busy.
6730 When set to '0' this limit is not used.
6733 NAME: half_closed_clients
6735 LOC: Config.onoff.half_closed_clients
6738 Some clients may shutdown the sending side of their TCP
6739 connections, while leaving their receiving sides open. Sometimes,
6740 Squid can not tell the difference between a half-closed and a
6741 fully-closed TCP connection.
6743 By default, Squid will immediately close client connections when
6744 read(2) returns "no more data to read."
6746 Change this option to 'on' and Squid will keep open connections
6747 until a read(2) or write(2) on the socket returns an error.
6748 This may show some benefits for reverse proxies. But if not
6749 it is recommended to leave OFF.
6752 NAME: server_idle_pconn_timeout pconn_timeout
6754 LOC: Config.Timeout.serverIdlePconn
6757 Timeout for idle persistent connections to servers and other
6764 LOC: Ident::TheConfig.timeout
6767 Maximum time to wait for IDENT lookups to complete.
6769 If this is too high, and you enabled IDENT lookups from untrusted
6770 users, you might be susceptible to denial-of-service by having
6771 many ident requests going at once.
6774 NAME: shutdown_lifetime
6777 LOC: Config.shutdownLifetime
6780 When SIGTERM or SIGHUP is received, the cache is put into
6781 "shutdown pending" mode until all active sockets are closed.
6782 This value is the lifetime to set for all open descriptors
6783 during shutdown mode. Any active clients after this many
6784 seconds will receive a 'timeout' message.
6788 ADMINISTRATIVE PARAMETERS
6789 -----------------------------------------------------------------------------
6795 LOC: Config.adminEmail
6797 Email-address of local cache manager who will receive
6798 mail if the cache dies. The default is "webmaster".
6804 LOC: Config.EmailFrom
6806 From: email-address for mail sent when the cache dies.
6807 The default is to use 'squid@unique_hostname'.
6809 See also: unique_hostname directive.
6815 LOC: Config.EmailProgram
6817 Email program used to send mail if the cache dies.
6818 The default is "mail". The specified program must comply
6819 with the standard Unix mail syntax:
6820 mail-program recipient < mailfile
6822 Optional command line options can be specified.
6825 NAME: cache_effective_user
6827 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
6828 LOC: Config.effectiveUser
6830 If you start Squid as root, it will change its effective/real
6831 UID/GID to the user specified below. The default is to change
6832 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
6833 see also; cache_effective_group
6836 NAME: cache_effective_group
6839 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
6840 LOC: Config.effectiveGroup
6842 Squid sets the GID to the effective user's default group ID
6843 (taken from the password file) and supplementary group list
6844 from the groups membership.
6846 If you want Squid to run with a specific GID regardless of
6847 the group memberships of the effective user then set this
6848 to the group (or GID) you want Squid to run as. When set
6849 all other group privileges of the effective user are ignored
6850 and only this GID is effective. If Squid is not started as
6851 root the user starting Squid MUST be member of the specified
6854 This option is not recommended by the Squid Team.
6855 Our preference is for administrators to configure a secure
6856 user account for squid with UID/GID matching system policies.
6859 NAME: httpd_suppress_version_string
6863 LOC: Config.onoff.httpd_suppress_version_string
6865 Suppress Squid version string info in HTTP headers and HTML error pages.
6868 NAME: visible_hostname
6870 LOC: Config.visibleHostname
6872 DEFAULT_DOC: Automatically detect the system host name
6874 If you want to present a special hostname in error messages, etc,
6875 define this. Otherwise, the return value of gethostname()
6876 will be used. If you have multiple caches in a cluster and
6877 get errors about IP-forwarding you must set them to have individual
6878 names with this setting.
6881 NAME: unique_hostname
6883 LOC: Config.uniqueHostname
6885 DEFAULT_DOC: Copy the value from visible_hostname
6887 If you want to have multiple machines with the same
6888 'visible_hostname' you must give each machine a different
6889 'unique_hostname' so forwarding loops can be detected.
6892 NAME: hostname_aliases
6894 LOC: Config.hostnameAliases
6897 A list of other DNS names your cache has.
6905 Minimum umask which should be enforced while the proxy
6906 is running, in addition to the umask set at startup.
6908 For a traditional octal representation of umasks, start
6913 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6914 -----------------------------------------------------------------------------
6916 This section contains parameters for the (optional) cache
6917 announcement service. This service is provided to help
6918 cache administrators locate one another in order to join or
6919 create cache hierarchies.
6921 An 'announcement' message is sent (via UDP) to the registration
6922 service by Squid. By default, the announcement message is NOT
6923 SENT unless you enable it with 'announce_period' below.
6925 The announcement message includes your hostname, plus the
6926 following information from this configuration file:
6932 All current information is processed regularly and made
6933 available on the Web at http://www.ircache.net/Cache/Tracker/.
6936 NAME: announce_period
6938 LOC: Config.Announce.period
6940 DEFAULT_DOC: Announcement messages disabled.
6942 This is how frequently to send cache announcements.
6944 To enable announcing your cache, just set an announce period.
6947 announce_period 1 day
6952 DEFAULT: tracker.ircache.net
6953 LOC: Config.Announce.host
6955 Set the hostname where announce registration messages will be sent.
6957 See also announce_port and announce_file
6963 LOC: Config.Announce.file
6965 The contents of this file will be included in the announce
6966 registration messages.
6972 LOC: Config.Announce.port
6974 Set the port where announce registration messages will be sent.
6976 See also announce_host and announce_file
6980 HTTPD-ACCELERATOR OPTIONS
6981 -----------------------------------------------------------------------------
6984 NAME: httpd_accel_surrogate_id
6987 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6988 LOC: Config.Accel.surrogate_id
6990 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6991 need an identification token to allow control targeting. Because
6992 a farm of surrogates may all perform the same tasks, they may share
6993 an identification token.
6996 NAME: http_accel_surrogate_remote
7000 LOC: Config.onoff.surrogate_is_remote
7002 Remote surrogates (such as those in a CDN) honour the header
7003 "Surrogate-Control: no-store-remote".
7005 Set this to on to have squid behave as a remote surrogate.
7009 IFDEF: USE_SQUID_ESI
7010 COMMENT: libxml2|expat|custom
7012 LOC: ESIParser::Type
7015 ESI markup is not strictly XML compatible. The custom ESI parser
7016 will give higher performance, but cannot handle non ASCII character
7021 DELAY POOL PARAMETERS
7022 -----------------------------------------------------------------------------
7026 TYPE: delay_pool_count
7028 IFDEF: USE_DELAY_POOLS
7031 This represents the number of delay pools to be used. For example,
7032 if you have one class 2 delay pool and one class 3 delays pool, you
7033 have a total of 2 delay pools.
7035 See also delay_parameters, delay_class, delay_access for pool
7036 configuration details.
7040 TYPE: delay_pool_class
7042 IFDEF: USE_DELAY_POOLS
7045 This defines the class of each delay pool. There must be exactly one
7046 delay_class line for each delay pool. For example, to define two
7047 delay pools, one of class 2 and one of class 3, the settings above
7051 delay_pools 4 # 4 delay pools
7052 delay_class 1 2 # pool 1 is a class 2 pool
7053 delay_class 2 3 # pool 2 is a class 3 pool
7054 delay_class 3 4 # pool 3 is a class 4 pool
7055 delay_class 4 5 # pool 4 is a class 5 pool
7057 The delay pool classes are:
7059 class 1 Everything is limited by a single aggregate
7062 class 2 Everything is limited by a single aggregate
7063 bucket as well as an "individual" bucket chosen
7064 from bits 25 through 32 of the IPv4 address.
7066 class 3 Everything is limited by a single aggregate
7067 bucket as well as a "network" bucket chosen
7068 from bits 17 through 24 of the IP address and a
7069 "individual" bucket chosen from bits 17 through
7070 32 of the IPv4 address.
7072 class 4 Everything in a class 3 delay pool, with an
7073 additional limit on a per user basis. This
7074 only takes effect if the username is established
7075 in advance - by forcing authentication in your
7078 class 5 Requests are grouped according their tag (see
7079 external_acl's tag= reply).
7082 Each pool also requires a delay_parameters directive to configure the pool size
7083 and speed limits used whenever the pool is applied to a request. Along with
7084 a set of delay_access directives to determine when it is used.
7086 NOTE: If an IP address is a.b.c.d
7087 -> bits 25 through 32 are "d"
7088 -> bits 17 through 24 are "c"
7089 -> bits 17 through 32 are "c * 256 + d"
7091 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
7092 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
7094 This clause only supports fast acl types.
7095 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7097 See also delay_parameters and delay_access.
7101 TYPE: delay_pool_access
7103 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
7104 IFDEF: USE_DELAY_POOLS
7107 This is used to determine which delay pool a request falls into.
7109 delay_access is sorted per pool and the matching starts with pool 1,
7110 then pool 2, ..., and finally pool N. The first delay pool where the
7111 request is allowed is selected for the request. If it does not allow
7112 the request to any pool then the request is not delayed (default).
7114 For example, if you want some_big_clients in delay
7115 pool 1 and lotsa_little_clients in delay pool 2:
7117 delay_access 1 allow some_big_clients
7118 delay_access 1 deny all
7119 delay_access 2 allow lotsa_little_clients
7120 delay_access 2 deny all
7121 delay_access 3 allow authenticated_clients
7123 See also delay_parameters and delay_class.
7127 NAME: delay_parameters
7128 TYPE: delay_pool_rates
7130 IFDEF: USE_DELAY_POOLS
7133 This defines the parameters for a delay pool. Each delay pool has
7134 a number of "buckets" associated with it, as explained in the
7135 description of delay_class.
7137 For a class 1 delay pool, the syntax is:
7139 delay_parameters pool aggregate
7141 For a class 2 delay pool:
7143 delay_parameters pool aggregate individual
7145 For a class 3 delay pool:
7147 delay_parameters pool aggregate network individual
7149 For a class 4 delay pool:
7151 delay_parameters pool aggregate network individual user
7153 For a class 5 delay pool:
7155 delay_parameters pool tagrate
7157 The option variables are:
7159 pool a pool number - ie, a number between 1 and the
7160 number specified in delay_pools as used in
7163 aggregate the speed limit parameters for the aggregate bucket
7166 individual the speed limit parameters for the individual
7167 buckets (class 2, 3).
7169 network the speed limit parameters for the network buckets
7172 user the speed limit parameters for the user buckets
7175 tagrate the speed limit parameters for the tag buckets
7178 A pair of delay parameters is written restore/maximum, where restore is
7179 the number of bytes (not bits - modem and network speeds are usually
7180 quoted in bits) per second placed into the bucket, and maximum is the
7181 maximum number of bytes which can be in the bucket at any time.
7183 There must be one delay_parameters line for each delay pool.
7186 For example, if delay pool number 1 is a class 2 delay pool as in the
7187 above example, and is being used to strictly limit each host to 64Kbit/sec
7188 (plus overheads), with no overall limit, the line is:
7190 delay_parameters 1 none 8000/8000
7192 Note that 8 x 8K Byte/sec -> 64K bit/sec.
7194 Note that the word 'none' is used to represent no limit.
7197 And, if delay pool number 2 is a class 3 delay pool as in the above
7198 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
7199 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
7200 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
7201 to permit a decent web page to be downloaded at a decent speed
7202 (if the network is not being limited due to overuse) but slow down
7203 large downloads more significantly:
7205 delay_parameters 2 32000/32000 8000/8000 600/8000
7207 Note that 8 x 32K Byte/sec -> 256K bit/sec.
7208 8 x 8K Byte/sec -> 64K bit/sec.
7209 8 x 600 Byte/sec -> 4800 bit/sec.
7212 Finally, for a class 4 delay pool as in the example - each user will
7213 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
7215 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
7218 See also delay_class and delay_access.
7222 NAME: delay_initial_bucket_level
7223 COMMENT: (percent, 0-100)
7226 IFDEF: USE_DELAY_POOLS
7227 LOC: Config.Delay.initial
7229 The initial bucket percentage is used to determine how much is put
7230 in each bucket when squid starts, is reconfigured, or first notices
7231 a host accessing it (in class 2 and class 3, individual hosts and
7232 networks only have buckets associated with them once they have been
7237 CLIENT DELAY POOL PARAMETERS
7238 -----------------------------------------------------------------------------
7241 NAME: client_delay_pools
7242 TYPE: client_delay_pool_count
7244 IFDEF: USE_DELAY_POOLS
7245 LOC: Config.ClientDelay
7247 This option specifies the number of client delay pools used. It must
7248 preceed other client_delay_* options.
7251 client_delay_pools 2
7253 See also client_delay_parameters and client_delay_access.
7256 NAME: client_delay_initial_bucket_level
7257 COMMENT: (percent, 0-no_limit)
7260 IFDEF: USE_DELAY_POOLS
7261 LOC: Config.ClientDelay.initial
7263 This option determines the initial bucket size as a percentage of
7264 max_bucket_size from client_delay_parameters. Buckets are created
7265 at the time of the "first" connection from the matching IP. Idle
7266 buckets are periodically deleted up.
7268 You can specify more than 100 percent but note that such "oversized"
7269 buckets are not refilled until their size goes down to max_bucket_size
7270 from client_delay_parameters.
7273 client_delay_initial_bucket_level 50
7276 NAME: client_delay_parameters
7277 TYPE: client_delay_pool_rates
7279 IFDEF: USE_DELAY_POOLS
7280 LOC: Config.ClientDelay
7283 This option configures client-side bandwidth limits using the
7286 client_delay_parameters pool speed_limit max_bucket_size
7288 pool is an integer ID used for client_delay_access matching.
7290 speed_limit is bytes added to the bucket per second.
7292 max_bucket_size is the maximum size of a bucket, enforced after any
7293 speed_limit additions.
7295 Please see the delay_parameters option for more information and
7299 client_delay_parameters 1 1024 2048
7300 client_delay_parameters 2 51200 16384
7302 See also client_delay_access.
7306 NAME: client_delay_access
7307 TYPE: client_delay_pool_access
7309 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
7310 IFDEF: USE_DELAY_POOLS
7311 LOC: Config.ClientDelay
7313 This option determines the client-side delay pool for the
7316 client_delay_access pool_ID allow|deny acl_name
7318 All client_delay_access options are checked in their pool ID
7319 order, starting with pool 1. The first checked pool with allowed
7320 request is selected for the request. If no ACL matches or there
7321 are no client_delay_access options, the request bandwidth is not
7324 The ACL-selected pool is then used to find the
7325 client_delay_parameters for the request. Client-side pools are
7326 not used to aggregate clients. Clients are always aggregated
7327 based on their source IP addresses (one bucket per source IP).
7329 This clause only supports fast acl types.
7330 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7331 Additionally, only the client TCP connection details are available.
7332 ACLs testing HTTP properties will not work.
7334 Please see delay_access for more examples.
7337 client_delay_access 1 allow low_rate_network
7338 client_delay_access 2 allow vips_network
7341 See also client_delay_parameters and client_delay_pools.
7344 NAME: response_delay_pool
7345 TYPE: response_delay_pool_parameters
7347 IFDEF: USE_DELAY_POOLS
7348 LOC: Config.MessageDelay
7350 This option configures client response bandwidth limits using the
7353 response_delay_pool name [option=value] ...
7355 name the response delay pool name
7359 individual-restore The speed limit of an individual
7360 bucket(bytes/s). To be used in conjunction
7361 with 'individual-maximum'.
7363 individual-maximum The maximum number of bytes which can
7364 be placed into the individual bucket. To be used
7365 in conjunction with 'individual-restore'.
7367 aggregate-restore The speed limit for the aggregate
7368 bucket(bytes/s). To be used in conjunction with
7369 'aggregate-maximum'.
7371 aggregate-maximum The maximum number of bytes which can
7372 be placed into the aggregate bucket. To be used
7373 in conjunction with 'aggregate-restore'.
7375 initial-bucket-level The initial bucket size as a percentage
7376 of individual-maximum.
7378 Individual and(or) aggregate bucket options may not be specified,
7379 meaning no individual and(or) aggregate speed limitation.
7380 See also response_delay_pool_access and delay_parameters for
7381 terminology details.
7384 NAME: response_delay_pool_access
7385 TYPE: response_delay_pool_access
7387 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
7388 IFDEF: USE_DELAY_POOLS
7389 LOC: Config.MessageDelay
7391 Determines whether a specific named response delay pool is used
7392 for the transaction. The syntax for this directive is:
7394 response_delay_pool_access pool_name allow|deny acl_name
7396 All response_delay_pool_access options are checked in the order
7397 they appear in this configuration file. The first rule with a
7398 matching ACL wins. If (and only if) an "allow" rule won, Squid
7399 assigns the response to the corresponding named delay pool.
7403 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
7404 -----------------------------------------------------------------------------
7409 LOC: Config.Wccp.router
7411 DEFAULT_DOC: WCCP disabled.
7414 Use this option to define your WCCP ``home'' router for
7417 wccp_router supports a single WCCP(v1) router
7419 wccp2_router supports multiple WCCPv2 routers
7421 only one of the two may be used at the same time and defines
7422 which version of WCCP to use.
7426 TYPE: IpAddress_list
7427 LOC: Config.Wccp2.router
7429 DEFAULT_DOC: WCCPv2 disabled.
7432 Use this option to define your WCCP ``home'' router for
7435 wccp_router supports a single WCCP(v1) router
7437 wccp2_router supports multiple WCCPv2 routers
7439 only one of the two may be used at the same time and defines
7440 which version of WCCP to use.
7445 LOC: Config.Wccp.version
7449 This directive is only relevant if you need to set up WCCP(v1)
7450 to some very old and end-of-life Cisco routers. In all other
7451 setups it must be left unset or at the default setting.
7452 It defines an internal version in the WCCP(v1) protocol,
7453 with version 4 being the officially documented protocol.
7455 According to some users, Cisco IOS 11.2 and earlier only
7456 support WCCP version 3. If you're using that or an earlier
7457 version of IOS, you may need to change this value to 3, otherwise
7458 do not specify this parameter.
7461 NAME: wccp2_rebuild_wait
7463 LOC: Config.Wccp2.rebuildwait
7467 If this is enabled Squid will wait for the cache dir rebuild to finish
7468 before sending the first wccp2 HereIAm packet
7471 NAME: wccp2_forwarding_method
7473 LOC: Config.Wccp2.forwarding_method
7477 WCCP2 allows the setting of forwarding methods between the
7478 router/switch and the cache. Valid values are as follows:
7480 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
7481 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
7483 Currently (as of IOS 12.4) cisco routers only support GRE.
7484 Cisco switches only support the L2 redirect assignment method.
7487 NAME: wccp2_return_method
7489 LOC: Config.Wccp2.return_method
7493 WCCP2 allows the setting of return methods between the
7494 router/switch and the cache for packets that the cache
7495 decides not to handle. Valid values are as follows:
7497 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
7498 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
7500 Currently (as of IOS 12.4) cisco routers only support GRE.
7501 Cisco switches only support the L2 redirect assignment.
7503 If the "ip wccp redirect exclude in" command has been
7504 enabled on the cache interface, then it is still safe for
7505 the proxy server to use a l2 redirect method even if this
7506 option is set to GRE.
7509 NAME: wccp2_assignment_method
7511 LOC: Config.Wccp2.assignment_method
7515 WCCP2 allows the setting of methods to assign the WCCP hash
7516 Valid values are as follows:
7518 hash - Hash assignment
7519 mask - Mask assignment
7521 As a general rule, cisco routers support the hash assignment method
7522 and cisco switches support the mask assignment method.
7527 LOC: Config.Wccp2.info
7528 DEFAULT_IF_NONE: standard 0
7529 DEFAULT_DOC: Use the 'web-cache' standard service.
7532 WCCP2 allows for multiple traffic services. There are two
7533 types: "standard" and "dynamic". The standard type defines
7534 one service id - http (id 0). The dynamic service ids can be from
7535 51 to 255 inclusive. In order to use a dynamic service id
7536 one must define the type of traffic to be redirected; this is done
7537 using the wccp2_service_info option.
7539 The "standard" type does not require a wccp2_service_info option,
7540 just specifying the service id will suffice.
7542 MD5 service authentication can be enabled by adding
7543 "password=<password>" to the end of this service declaration.
7547 wccp2_service standard 0 # for the 'web-cache' standard service
7548 wccp2_service dynamic 80 # a dynamic service type which will be
7549 # fleshed out with subsequent options.
7550 wccp2_service standard 0 password=foo
7553 NAME: wccp2_service_info
7554 TYPE: wccp2_service_info
7555 LOC: Config.Wccp2.info
7559 Dynamic WCCPv2 services require further information to define the
7560 traffic you wish to have diverted.
7564 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
7565 priority=<priority> ports=<port>,<port>..
7567 The relevant WCCPv2 flags:
7568 + src_ip_hash, dst_ip_hash
7569 + source_port_hash, dst_port_hash
7570 + src_ip_alt_hash, dst_ip_alt_hash
7571 + src_port_alt_hash, dst_port_alt_hash
7574 The port list can be one to eight entries.
7578 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
7579 priority=240 ports=80
7581 Note: the service id must have been defined by a previous
7582 'wccp2_service dynamic <id>' entry.
7587 LOC: Config.Wccp2.weight
7591 Each cache server gets assigned a set of the destination
7592 hash proportional to their weight.
7597 LOC: Config.Wccp.address
7599 DEFAULT_DOC: Address selected by the operating system.
7602 Use this option if you require WCCPv2 to use a specific
7605 The default behavior is to not bind to any specific address.
7610 LOC: Config.Wccp2.address
7612 DEFAULT_DOC: Address selected by the operating system.
7615 Use this option if you require WCCP to use a specific
7618 The default behavior is to not bind to any specific address.
7622 PERSISTENT CONNECTION HANDLING
7623 -----------------------------------------------------------------------------
7625 Also see "pconn_timeout" in the TIMEOUTS section
7628 NAME: client_persistent_connections
7630 LOC: Config.onoff.client_pconns
7633 Persistent connection support for clients.
7634 Squid uses persistent connections (when allowed). You can use
7635 this option to disable persistent connections with clients.
7638 NAME: server_persistent_connections
7640 LOC: Config.onoff.server_pconns
7643 Persistent connection support for servers.
7644 Squid uses persistent connections (when allowed). You can use
7645 this option to disable persistent connections with servers.
7648 NAME: persistent_connection_after_error
7650 LOC: Config.onoff.error_pconns
7653 With this directive the use of persistent connections after
7654 HTTP errors can be disabled. Useful if you have clients
7655 who fail to handle errors on persistent connections proper.
7658 NAME: detect_broken_pconn
7660 LOC: Config.onoff.detect_broken_server_pconns
7663 Some servers have been found to incorrectly signal the use
7664 of HTTP/1.0 persistent connections even on replies not
7665 compatible, causing significant delays. This server problem
7666 has mostly been seen on redirects.
7668 By enabling this directive Squid attempts to detect such
7669 broken replies and automatically assume the reply is finished
7670 after 10 seconds timeout.
7674 CACHE DIGEST OPTIONS
7675 -----------------------------------------------------------------------------
7678 NAME: digest_generation
7679 IFDEF: USE_CACHE_DIGESTS
7681 LOC: Config.onoff.digest_generation
7684 This controls whether the server will generate a Cache Digest
7685 of its contents. By default, Cache Digest generation is
7686 enabled if Squid is compiled with --enable-cache-digests defined.
7689 NAME: digest_bits_per_entry
7690 IFDEF: USE_CACHE_DIGESTS
7692 LOC: Config.digest.bits_per_entry
7695 This is the number of bits of the server's Cache Digest which
7696 will be associated with the Digest entry for a given HTTP
7697 Method and URL (public key) combination. The default is 5.
7700 NAME: digest_rebuild_period
7701 IFDEF: USE_CACHE_DIGESTS
7704 LOC: Config.digest.rebuild_period
7707 This is the wait time between Cache Digest rebuilds.
7710 NAME: digest_rewrite_period
7712 IFDEF: USE_CACHE_DIGESTS
7714 LOC: Config.digest.rewrite_period
7717 This is the wait time between Cache Digest writes to
7721 NAME: digest_swapout_chunk_size
7724 IFDEF: USE_CACHE_DIGESTS
7725 LOC: Config.digest.swapout_chunk_size
7728 This is the number of bytes of the Cache Digest to write to
7729 disk at a time. It defaults to 4096 bytes (4KB), the Squid
7733 NAME: digest_rebuild_chunk_percentage
7734 COMMENT: (percent, 0-100)
7735 IFDEF: USE_CACHE_DIGESTS
7737 LOC: Config.digest.rebuild_chunk_percentage
7740 This is the percentage of the Cache Digest to be scanned at a
7741 time. By default it is set to 10% of the Cache Digest.
7746 -----------------------------------------------------------------------------
7751 LOC: Config.Port.snmp
7753 DEFAULT_DOC: SNMP disabled.
7756 The port number where Squid listens for SNMP requests. To enable
7757 SNMP support set this to a suitable port number. Port number
7758 3401 is often used for the Squid SNMP agent. By default it's
7759 set to "0" (disabled)
7767 LOC: Config.accessList.snmp
7769 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
7772 Allowing or denying access to the SNMP port.
7774 All access to the agent is denied by default.
7777 snmp_access allow|deny [!]aclname ...
7779 This clause only supports fast acl types.
7780 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7783 snmp_access allow snmppublic localhost
7784 snmp_access deny all
7787 NAME: snmp_incoming_address
7789 LOC: Config.Addrs.snmp_incoming
7791 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
7794 Just like 'udp_incoming_address', but for the SNMP port.
7796 snmp_incoming_address is used for the SNMP socket receiving
7797 messages from SNMP agents.
7799 The default snmp_incoming_address is to listen on all
7800 available network interfaces.
7803 NAME: snmp_outgoing_address
7805 LOC: Config.Addrs.snmp_outgoing
7807 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
7810 Just like 'udp_outgoing_address', but for the SNMP port.
7812 snmp_outgoing_address is used for SNMP packets returned to SNMP
7815 If snmp_outgoing_address is not set it will use the same socket
7816 as snmp_incoming_address. Only change this if you want to have
7817 SNMP replies sent using another address than where this Squid
7818 listens for SNMP queries.
7820 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
7821 the same value since they both use the same port.
7826 -----------------------------------------------------------------------------
7829 NAME: icp_port udp_port
7832 DEFAULT_DOC: ICP disabled.
7833 LOC: Config.Port.icp
7835 The port number where Squid sends and receives ICP queries to
7836 and from neighbor caches. The standard UDP port for ICP is 3130.
7839 icp_port @DEFAULT_ICP_PORT@
7846 DEFAULT_DOC: HTCP disabled.
7847 LOC: Config.Port.htcp
7849 The port number where Squid sends and receives HTCP queries to
7850 and from neighbor caches. To turn it on you want to set it to
7857 NAME: log_icp_queries
7861 LOC: Config.onoff.log_udp
7863 If set, ICP queries are logged to access.log. You may wish
7864 do disable this if your ICP load is VERY high to speed things
7865 up or to simplify log analysis.
7868 NAME: udp_incoming_address
7870 LOC:Config.Addrs.udp_incoming
7872 DEFAULT_DOC: Accept packets from all machine interfaces.
7874 udp_incoming_address is used for UDP packets received from other
7877 The default behavior is to not bind to any specific address.
7879 Only change this if you want to have all UDP queries received on
7880 a specific interface/address.
7882 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
7883 modules. Altering it will affect all of them in the same manner.
7885 see also; udp_outgoing_address
7887 NOTE, udp_incoming_address and udp_outgoing_address can not
7888 have the same value since they both use the same port.
7891 NAME: udp_outgoing_address
7893 LOC: Config.Addrs.udp_outgoing
7895 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
7897 udp_outgoing_address is used for UDP packets sent out to other
7900 The default behavior is to not bind to any specific address.
7902 Instead it will use the same socket as udp_incoming_address.
7903 Only change this if you want to have UDP queries sent using another
7904 address than where this Squid listens for UDP queries from other
7907 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
7908 modules. Altering it will affect all of them in the same manner.
7910 see also; udp_incoming_address
7912 NOTE, udp_incoming_address and udp_outgoing_address can not
7913 have the same value since they both use the same port.
7920 LOC: Config.onoff.icp_hit_stale
7922 If you want to return ICP_HIT for stale cache objects, set this
7923 option to 'on'. If you have sibling relationships with caches
7924 in other administrative domains, this should be 'off'. If you only
7925 have sibling relationships with caches under your control,
7926 it is probably okay to set this to 'on'.
7927 If set to 'on', your siblings should use the option "allow-miss"
7928 on their cache_peer lines for connecting to you.
7931 NAME: minimum_direct_hops
7934 LOC: Config.minDirectHops
7936 If using the ICMP pinging stuff, do direct fetches for sites
7937 which are no more than this many hops away.
7940 NAME: minimum_direct_rtt
7944 LOC: Config.minDirectRtt
7946 If using the ICMP pinging stuff, do direct fetches for sites
7947 which are no more than this many rtt milliseconds away.
7953 LOC: Config.Netdb.low
7955 The low water mark for the ICMP measurement database.
7957 Note: high watermark controlled by netdb_high directive.
7959 These watermarks are counts, not percents. The defaults are
7960 (low) 900 and (high) 1000. When the high water mark is
7961 reached, database entries will be deleted until the low
7968 LOC: Config.Netdb.high
7970 The high water mark for the ICMP measurement database.
7972 Note: low watermark controlled by netdb_low directive.
7974 These watermarks are counts, not percents. The defaults are
7975 (low) 900 and (high) 1000. When the high water mark is
7976 reached, database entries will be deleted until the low
7980 NAME: netdb_ping_period
7982 LOC: Config.Netdb.period
7985 The minimum period for measuring a site. There will be at
7986 least this much delay between successive pings to the same
7987 network. The default is five minutes.
7994 LOC: Config.onoff.query_icmp
7996 If you want to ask your peers to include ICMP data in their ICP
7997 replies, enable this option.
7999 If your peer has configured Squid (during compilation) with
8000 '--enable-icmp' that peer will send ICMP pings to origin server
8001 sites of the URLs it receives. If you enable this option the
8002 ICP replies from that peer will include the ICMP data (if available).
8003 Then, when choosing a parent cache, Squid will choose the parent with
8004 the minimal RTT to the origin server. When this happens, the
8005 hierarchy field of the access.log will be
8006 "CLOSEST_PARENT_MISS". This option is off by default.
8009 NAME: test_reachability
8013 LOC: Config.onoff.test_reachability
8015 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
8016 instead of ICP_MISS if the target host is NOT in the ICMP
8017 database, or has a zero RTT.
8020 NAME: icp_query_timeout
8023 DEFAULT_DOC: Dynamic detection.
8025 LOC: Config.Timeout.icp_query
8027 Normally Squid will automatically determine an optimal ICP
8028 query timeout value based on the round-trip-time of recent ICP
8029 queries. If you want to override the value determined by
8030 Squid, set this 'icp_query_timeout' to a non-zero value. This
8031 value is specified in MILLISECONDS, so, to use a 2-second
8032 timeout (the old default), you would write:
8034 icp_query_timeout 2000
8037 NAME: maximum_icp_query_timeout
8041 LOC: Config.Timeout.icp_query_max
8043 Normally the ICP query timeout is determined dynamically. But
8044 sometimes it can lead to very large values (say 5 seconds).
8045 Use this option to put an upper limit on the dynamic timeout
8046 value. Do NOT use this option to always use a fixed (instead
8047 of a dynamic) timeout value. To set a fixed timeout see the
8048 'icp_query_timeout' directive.
8051 NAME: minimum_icp_query_timeout
8055 LOC: Config.Timeout.icp_query_min
8057 Normally the ICP query timeout is determined dynamically. But
8058 sometimes it can lead to very small timeouts, even lower than
8059 the normal latency variance on your link due to traffic.
8060 Use this option to put an lower limit on the dynamic timeout
8061 value. Do NOT use this option to always use a fixed (instead
8062 of a dynamic) timeout value. To set a fixed timeout see the
8063 'icp_query_timeout' directive.
8066 NAME: background_ping_rate
8070 LOC: Config.backgroundPingRate
8072 Controls how often the ICP pings are sent to siblings that
8073 have background-ping set.
8077 MULTICAST ICP OPTIONS
8078 -----------------------------------------------------------------------------
8083 LOC: Config.mcast_group_list
8086 This tag specifies a list of multicast groups which your server
8087 should join to receive multicasted ICP queries.
8089 NOTE! Be very careful what you put here! Be sure you
8090 understand the difference between an ICP _query_ and an ICP
8091 _reply_. This option is to be set only if you want to RECEIVE
8092 multicast queries. Do NOT set this option to SEND multicast
8093 ICP (use cache_peer for that). ICP replies are always sent via
8094 unicast, so this option does not affect whether or not you will
8095 receive replies from multicast group members.
8097 You must be very careful to NOT use a multicast address which
8098 is already in use by another group of caches.
8100 If you are unsure about multicast, please read the Multicast
8101 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
8103 Usage: mcast_groups 239.128.16.128 224.0.1.20
8105 By default, Squid doesn't listen on any multicast groups.
8108 NAME: mcast_miss_addr
8109 IFDEF: MULTICAST_MISS_STREAM
8111 LOC: Config.mcast_miss.addr
8113 DEFAULT_DOC: disabled.
8115 If you enable this option, every "cache miss" URL will
8116 be sent out on the specified multicast address.
8118 Do not enable this option unless you are are absolutely
8119 certain you understand what you are doing.
8122 NAME: mcast_miss_ttl
8123 IFDEF: MULTICAST_MISS_STREAM
8125 LOC: Config.mcast_miss.ttl
8128 This is the time-to-live value for packets multicasted
8129 when multicasting off cache miss URLs is enabled. By
8130 default this is set to 'site scope', i.e. 16.
8133 NAME: mcast_miss_port
8134 IFDEF: MULTICAST_MISS_STREAM
8136 LOC: Config.mcast_miss.port
8139 This is the port number to be used in conjunction with
8143 NAME: mcast_miss_encode_key
8144 IFDEF: MULTICAST_MISS_STREAM
8146 LOC: Config.mcast_miss.encode_key
8147 DEFAULT: XXXXXXXXXXXXXXXX
8149 The URLs that are sent in the multicast miss stream are
8150 encrypted. This is the encryption key.
8153 NAME: mcast_icp_query_timeout
8157 LOC: Config.Timeout.mcast_icp_query
8159 For multicast peers, Squid regularly sends out ICP "probes" to
8160 count how many other peers are listening on the given multicast
8161 address. This value specifies how long Squid should wait to
8162 count all the replies. The default is 2000 msec, or 2
8167 INTERNAL ICON OPTIONS
8168 -----------------------------------------------------------------------------
8171 NAME: icon_directory
8173 LOC: Config.icons.directory
8174 DEFAULT: @DEFAULT_ICON_DIR@
8176 Where the icons are stored. These are normally kept in
8180 NAME: global_internal_static
8182 LOC: Config.onoff.global_internal_static
8185 This directive controls is Squid should intercept all requests for
8186 /squid-internal-static/ no matter which host the URL is requesting
8187 (default on setting), or if nothing special should be done for
8188 such URLs (off setting). The purpose of this directive is to make
8189 icons etc work better in complex cache hierarchies where it may
8190 not always be possible for all corners in the cache mesh to reach
8191 the server generating a directory listing.
8194 NAME: short_icon_urls
8196 LOC: Config.icons.use_short_names
8199 If this is enabled Squid will use short URLs for icons.
8200 If disabled it will revert to the old behavior of including
8201 it's own name and port in the URL.
8203 If you run a complex cache hierarchy with a mix of Squid and
8204 other proxies you may need to disable this directive.
8209 -----------------------------------------------------------------------------
8212 NAME: error_directory
8214 LOC: Config.errorDirectory
8216 DEFAULT_DOC: Send error pages in the clients preferred language
8218 If you wish to create your own versions of the default
8219 error files to customize them to suit your company copy
8220 the error/template files to another directory and point
8223 WARNING: This option will disable multi-language support
8224 on error pages if used.
8226 The squid developers are interested in making squid available in
8227 a wide variety of languages. If you are making translations for a
8228 language that Squid does not currently provide please consider
8229 contributing your translation back to the project.
8230 http://wiki.squid-cache.org/Translations
8232 The squid developers working on translations are happy to supply drop-in
8233 translated error files in exchange for any new language contributions.
8236 NAME: error_default_language
8237 IFDEF: USE_ERR_LOCALES
8239 LOC: Config.errorDefaultLanguage
8241 DEFAULT_DOC: Generate English language pages.
8243 Set the default language which squid will send error pages in
8244 if no existing translation matches the clients language
8247 If unset (default) generic English will be used.
8249 The squid developers are interested in making squid available in
8250 a wide variety of languages. If you are interested in making
8251 translations for any language see the squid wiki for details.
8252 http://wiki.squid-cache.org/Translations
8255 NAME: error_log_languages
8256 IFDEF: USE_ERR_LOCALES
8258 LOC: Config.errorLogMissingLanguages
8261 Log to cache.log what languages users are attempting to
8262 auto-negotiate for translations.
8264 Successful negotiations are not logged. Only failures
8265 have meaning to indicate that Squid may need an upgrade
8266 of its error page translations.
8269 NAME: err_page_stylesheet
8271 LOC: Config.errorStylesheet
8272 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
8274 CSS Stylesheet to pattern the display of Squid default error pages.
8276 For information on CSS see http://www.w3.org/Style/CSS/
8281 LOC: Config.errHtmlText
8284 HTML text to include in error messages. Make this a "mailto"
8285 URL to your admin address, or maybe just a link to your
8286 organizations Web page.
8288 To include this in your error messages, you must rewrite
8289 the error template files (found in the "errors" directory).
8290 Wherever you want the 'err_html_text' line to appear,
8291 insert a %L tag in the error template file.
8294 NAME: email_err_data
8297 LOC: Config.onoff.emailErrData
8300 If enabled, information about the occurred error will be
8301 included in the mailto links of the ERR pages (if %W is set)
8302 so that the email body contains the data.
8303 Syntax is <A HREF="mailto:%w%W">%w</A>
8308 LOC: Config.denyInfoList
8311 Usage: deny_info err_page_name acl
8312 or deny_info http://... acl
8313 or deny_info TCP_RESET acl
8315 This can be used to return a ERR_ page for requests which
8316 do not pass the 'http_access' rules. Squid remembers the last
8317 acl it evaluated in http_access, and if a 'deny_info' line exists
8318 for that ACL Squid returns a corresponding error page.
8320 The acl is typically the last acl on the http_access deny line which
8321 denied access. The exceptions to this rule are:
8322 - When Squid needs to request authentication credentials. It's then
8323 the first authentication related acl encountered
8324 - When none of the http_access lines matches. It's then the last
8325 acl processed on the last http_access line.
8326 - When the decision to deny access was made by an adaptation service,
8327 the acl name is the corresponding eCAP or ICAP service_name.
8329 NP: If providing your own custom error pages with error_directory
8330 you may also specify them by your custom file name:
8331 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
8333 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
8334 may be specified by prefixing the file name with the code and a colon.
8335 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
8337 Alternatively you can tell Squid to reset the TCP connection
8338 by specifying TCP_RESET.
8340 Or you can specify an error URL or URL pattern. The browsers will
8341 get redirected to the specified URL after formatting tags have
8342 been replaced. Redirect will be done with 302 or 307 according to
8343 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
8344 the URL. e.g. 303:http://example.com/
8347 %a - username (if available. Password NOT included)
8350 %E - Error description
8352 %H - Request domain name
8353 %i - Client IP Address
8355 %O - Unescaped message result from external ACL helper
8356 %o - Message result from external ACL helper
8357 %p - Request Port number
8358 %P - Request Protocol name
8359 %R - Request URL path
8360 %T - Timestamp in RFC 1123 format
8361 %U - Full canonical URL from client
8362 (HTTPS URLs terminate with *)
8363 %u - Full canonical URL from client
8364 %w - Admin email from squid.conf
8366 %% - Literal percent (%) code
8371 OPTIONS INFLUENCING REQUEST FORWARDING
8372 -----------------------------------------------------------------------------
8375 NAME: nonhierarchical_direct
8377 LOC: Config.onoff.nonhierarchical_direct
8380 By default, Squid will send any non-hierarchical requests
8381 (not cacheable request type) direct to origin servers.
8383 When this is set to "off", Squid will prefer to send these
8384 requests to parents.
8386 Note that in most configurations, by turning this off you will only
8387 add latency to these request without any improvement in global hit
8390 This option only sets a preference. If the parent is unavailable a
8391 direct connection to the origin server may still be attempted. To
8392 completely prevent direct connections use never_direct.
8397 LOC: Config.onoff.prefer_direct
8400 Normally Squid tries to use parents for most requests. If you for some
8401 reason like it to first try going direct and only use a parent if
8402 going direct fails set this to on.
8404 By combining nonhierarchical_direct off and prefer_direct on you
8405 can set up Squid to use a parent as a backup path if going direct
8408 Note: If you want Squid to use parents for all requests see
8409 the never_direct directive. prefer_direct only modifies how Squid
8410 acts on cacheable requests.
8413 NAME: cache_miss_revalidate
8417 LOC: Config.onoff.cache_miss_revalidate
8419 RFC 7232 defines a conditional request mechanism to prevent
8420 response objects being unnecessarily transferred over the network.
8421 If that mechanism is used by the client and a cache MISS occurs
8422 it can prevent new cache entries being created.
8424 This option determines whether Squid on cache MISS will pass the
8425 client revalidation request to the server or tries to fetch new
8426 content for caching. It can be useful while the cache is mostly
8427 empty to more quickly have the cache populated by generating
8428 non-conditional GETs.
8430 When set to 'on' (default), Squid will pass all client If-* headers
8431 to the server. This permits server responses without a cacheable
8432 payload to be delivered and on MISS no new cache entry is created.
8434 When set to 'off' and if the request is cacheable, Squid will
8435 remove the clients If-Modified-Since and If-None-Match headers from
8436 the request sent to the server. This requests a 200 status response
8437 from the server to create a new cache entry with.
8442 LOC: Config.accessList.AlwaysDirect
8444 DEFAULT_DOC: Prevent any cache_peer being used for this request.
8446 Usage: always_direct allow|deny [!]aclname ...
8448 Here you can use ACL elements to specify requests which should
8449 ALWAYS be forwarded by Squid to the origin servers without using
8450 any peers. For example, to always directly forward requests for
8451 local servers ignoring any parents or siblings you may have use
8454 acl local-servers dstdomain my.domain.net
8455 always_direct allow local-servers
8457 To always forward FTP requests directly, use
8460 always_direct allow FTP
8462 NOTE: There is a similar, but opposite option named
8463 'never_direct'. You need to be aware that "always_direct deny
8464 foo" is NOT the same thing as "never_direct allow foo". You
8465 may need to use a deny rule to exclude a more-specific case of
8466 some other rule. Example:
8468 acl local-external dstdomain external.foo.net
8469 acl local-servers dstdomain .foo.net
8470 always_direct deny local-external
8471 always_direct allow local-servers
8473 NOTE: If your goal is to make the client forward the request
8474 directly to the origin server bypassing Squid then this needs
8475 to be done in the client configuration. Squid configuration
8476 can only tell Squid how Squid should fetch the object.
8478 NOTE: This directive is not related to caching. The replies
8479 is cached as usual even if you use always_direct. To not cache
8480 the replies see the 'cache' directive.
8482 This clause supports both fast and slow acl types.
8483 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8488 LOC: Config.accessList.NeverDirect
8490 DEFAULT_DOC: Allow DNS results to be used for this request.
8492 Usage: never_direct allow|deny [!]aclname ...
8494 never_direct is the opposite of always_direct. Please read
8495 the description for always_direct if you have not already.
8497 With 'never_direct' you can use ACL elements to specify
8498 requests which should NEVER be forwarded directly to origin
8499 servers. For example, to force the use of a proxy for all
8500 requests, except those in your local domain use something like:
8502 acl local-servers dstdomain .foo.net
8503 never_direct deny local-servers
8504 never_direct allow all
8506 or if Squid is inside a firewall and there are local intranet
8507 servers inside the firewall use something like:
8509 acl local-intranet dstdomain .foo.net
8510 acl local-external dstdomain external.foo.net
8511 always_direct deny local-external
8512 always_direct allow local-intranet
8513 never_direct allow all
8515 This clause supports both fast and slow acl types.
8516 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8520 ADVANCED NETWORKING OPTIONS
8521 -----------------------------------------------------------------------------
8524 NAME: incoming_udp_average incoming_icp_average
8527 LOC: Config.comm_incoming.udp.average
8529 Heavy voodoo here. I can't even believe you are reading this.
8530 Are you crazy? Don't even think about adjusting these unless
8531 you understand the algorithms in comm_select.c first!
8534 NAME: incoming_tcp_average incoming_http_average
8537 LOC: Config.comm_incoming.tcp.average
8539 Heavy voodoo here. I can't even believe you are reading this.
8540 Are you crazy? Don't even think about adjusting these unless
8541 you understand the algorithms in comm_select.c first!
8544 NAME: incoming_dns_average
8547 LOC: Config.comm_incoming.dns.average
8549 Heavy voodoo here. I can't even believe you are reading this.
8550 Are you crazy? Don't even think about adjusting these unless
8551 you understand the algorithms in comm_select.c first!
8554 NAME: min_udp_poll_cnt min_icp_poll_cnt
8557 LOC: Config.comm_incoming.udp.min_poll
8559 Heavy voodoo here. I can't even believe you are reading this.
8560 Are you crazy? Don't even think about adjusting these unless
8561 you understand the algorithms in comm_select.c first!
8564 NAME: min_dns_poll_cnt
8567 LOC: Config.comm_incoming.dns.min_poll
8569 Heavy voodoo here. I can't even believe you are reading this.
8570 Are you crazy? Don't even think about adjusting these unless
8571 you understand the algorithms in comm_select.c first!
8574 NAME: min_tcp_poll_cnt min_http_poll_cnt
8577 LOC: Config.comm_incoming.tcp.min_poll
8579 Heavy voodoo here. I can't even believe you are reading this.
8580 Are you crazy? Don't even think about adjusting these unless
8581 you understand the algorithms in comm_select.c first!
8587 LOC: Config.accept_filter
8591 The name of an accept(2) filter to install on Squid's
8592 listen socket(s). This feature is perhaps specific to
8593 FreeBSD and requires support in the kernel.
8595 The 'httpready' filter delays delivering new connections
8596 to Squid until a full HTTP request has been received.
8597 See the accf_http(9) man page for details.
8599 The 'dataready' filter delays delivering new connections
8600 to Squid until there is some data to process.
8601 See the accf_dataready(9) man page for details.
8605 The 'data' filter delays delivering of new connections
8606 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
8607 You may optionally specify a number of seconds to wait by
8608 'data=N' where N is the number of seconds. Defaults to 30
8609 if not specified. See the tcp(7) man page for details.
8612 accept_filter httpready
8617 NAME: client_ip_max_connections
8619 LOC: Config.client_ip_max_connections
8621 DEFAULT_DOC: No limit.
8623 Set an absolute limit on the number of connections a single
8624 client IP can use. Any more than this and Squid will begin to drop
8625 new connections from the client until it closes some links.
8627 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
8628 connections from the client. For finer control use the ACL access controls.
8630 Requires client_db to be enabled (the default).
8632 WARNING: This may noticably slow down traffic received via external proxies
8633 or NAT devices and cause them to rebound error messages back to their clients.
8636 NAME: tcp_recv_bufsize
8640 DEFAULT_DOC: Use operating system TCP defaults.
8641 LOC: Config.tcpRcvBufsz
8643 Size of receive buffer to set for TCP sockets. Probably just
8644 as easy to change your kernel's default.
8645 Omit from squid.conf to use the default buffer size.
8650 -----------------------------------------------------------------------------
8657 LOC: Adaptation::Icap::TheConfig.onoff
8660 If you want to enable the ICAP module support, set this to on.
8663 NAME: icap_connect_timeout
8666 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
8669 This parameter specifies how long to wait for the TCP connect to
8670 the requested ICAP server to complete before giving up and either
8671 terminating the HTTP transaction or bypassing the failure.
8673 The default for optional services is peer_connect_timeout.
8674 The default for essential services is connect_timeout.
8675 If this option is explicitly set, its value applies to all services.
8678 NAME: icap_io_timeout
8682 DEFAULT_DOC: Use read_timeout.
8683 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
8686 This parameter specifies how long to wait for an I/O activity on
8687 an established, active ICAP connection before giving up and
8688 either terminating the HTTP transaction or bypassing the
8692 NAME: icap_service_failure_limit
8693 COMMENT: limit [in memory-depth time-units]
8694 TYPE: icap_service_failure_limit
8696 LOC: Adaptation::Icap::TheConfig
8699 The limit specifies the number of failures that Squid tolerates
8700 when establishing a new TCP connection with an ICAP service. If
8701 the number of failures exceeds the limit, the ICAP service is
8702 not used for new ICAP requests until it is time to refresh its
8705 A negative value disables the limit. Without the limit, an ICAP
8706 service will not be considered down due to connectivity failures
8707 between ICAP OPTIONS requests.
8709 Squid forgets ICAP service failures older than the specified
8710 value of memory-depth. The memory fading algorithm
8711 is approximate because Squid does not remember individual
8712 errors but groups them instead, splitting the option
8713 value into ten time slots of equal length.
8715 When memory-depth is 0 and by default this option has no
8716 effect on service failure expiration.
8718 Squid always forgets failures when updating service settings
8719 using an ICAP OPTIONS transaction, regardless of this option
8723 # suspend service usage after 10 failures in 5 seconds:
8724 icap_service_failure_limit 10 in 5 seconds
8727 NAME: icap_service_revival_delay
8730 LOC: Adaptation::Icap::TheConfig.service_revival_delay
8733 The delay specifies the number of seconds to wait after an ICAP
8734 OPTIONS request failure before requesting the options again. The
8735 failed ICAP service is considered "down" until fresh OPTIONS are
8738 The actual delay cannot be smaller than the hardcoded minimum
8739 delay of 30 seconds.
8742 NAME: icap_preview_enable
8746 LOC: Adaptation::Icap::TheConfig.preview_enable
8749 The ICAP Preview feature allows the ICAP server to handle the
8750 HTTP message by looking only at the beginning of the message body
8751 or even without receiving the body at all. In some environments,
8752 previews greatly speedup ICAP processing.
8754 During an ICAP OPTIONS transaction, the server may tell Squid what
8755 HTTP messages should be previewed and how big the preview should be.
8756 Squid will not use Preview if the server did not request one.
8758 To disable ICAP Preview for all ICAP services, regardless of
8759 individual ICAP server OPTIONS responses, set this option to "off".
8761 icap_preview_enable off
8764 NAME: icap_preview_size
8767 LOC: Adaptation::Icap::TheConfig.preview_size
8769 DEFAULT_DOC: No preview sent.
8771 The default size of preview data to be sent to the ICAP server.
8772 This value might be overwritten on a per server basis by OPTIONS requests.
8775 NAME: icap_206_enable
8779 LOC: Adaptation::Icap::TheConfig.allow206_enable
8782 206 (Partial Content) responses is an ICAP extension that allows the
8783 ICAP agents to optionally combine adapted and original HTTP message
8784 content. The decision to combine is postponed until the end of the
8785 ICAP response. Squid supports Partial Content extension by default.
8787 Activation of the Partial Content extension is negotiated with each
8788 ICAP service during OPTIONS exchange. Most ICAP servers should handle
8789 negotation correctly even if they do not support the extension, but
8790 some might fail. To disable Partial Content support for all ICAP
8791 services and to avoid any negotiation, set this option to "off".
8797 NAME: icap_default_options_ttl
8800 LOC: Adaptation::Icap::TheConfig.default_options_ttl
8803 The default TTL value for ICAP OPTIONS responses that don't have
8804 an Options-TTL header.
8807 NAME: icap_persistent_connections
8811 LOC: Adaptation::Icap::TheConfig.reuse_connections
8814 Whether or not Squid should use persistent connections to
8818 NAME: adaptation_send_client_ip icap_send_client_ip
8820 IFDEF: USE_ADAPTATION
8822 LOC: Adaptation::Config::send_client_ip
8825 If enabled, Squid shares HTTP client IP information with adaptation
8826 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
8827 For eCAP, Squid sets the libecap::metaClientIp transaction option.
8829 See also: adaptation_uses_indirect_client
8832 NAME: adaptation_send_username icap_send_client_username
8834 IFDEF: USE_ADAPTATION
8836 LOC: Adaptation::Config::send_username
8839 This sends authenticated HTTP client username (if available) to
8840 the adaptation service.
8842 For ICAP, the username value is encoded based on the
8843 icap_client_username_encode option and is sent using the header
8844 specified by the icap_client_username_header option.
8847 NAME: icap_client_username_header
8850 LOC: Adaptation::Icap::TheConfig.client_username_header
8851 DEFAULT: X-Client-Username
8853 ICAP request header name to use for adaptation_send_username.
8856 NAME: icap_client_username_encode
8860 LOC: Adaptation::Icap::TheConfig.client_username_encode
8863 Whether to base64 encode the authenticated client username.
8867 TYPE: icap_service_type
8869 LOC: Adaptation::Icap::TheConfig
8872 Defines a single ICAP service using the following format:
8874 icap_service id vectoring_point uri [option ...]
8877 an opaque identifier or name which is used to direct traffic to
8878 this specific service. Must be unique among all adaptation
8879 services in squid.conf.
8881 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8882 This specifies at which point of transaction processing the
8883 ICAP service should be activated. *_postcache vectoring points
8884 are not yet supported.
8886 uri: icap://servername:port/servicepath
8887 ICAP server and service location.
8888 icaps://servername:port/servicepath
8889 The "icap:" URI scheme is used for traditional ICAP server and
8890 service location (default port is 1344, connections are not
8891 encrypted). The "icaps:" URI scheme is for Secure ICAP
8892 services that use SSL/TLS-encrypted ICAP connections (by
8893 default, on port 11344).
8895 ICAP does not allow a single service to handle both REQMOD and RESPMOD
8896 transactions. Squid does not enforce that requirement. You can specify
8897 services with the same service_url and different vectoring_points. You
8898 can even specify multiple identical services as long as their
8899 service_names differ.
8901 To activate a service, use the adaptation_access directive. To group
8902 services, use adaptation_service_chain and adaptation_service_set.
8904 Service options are separated by white space. ICAP services support
8905 the following name=value options:
8908 If set to 'on' or '1', the ICAP service is treated as
8909 optional. If the service cannot be reached or malfunctions,
8910 Squid will try to ignore any errors and process the message as
8911 if the service was not enabled. No all ICAP errors can be
8912 bypassed. If set to 0, the ICAP service is treated as
8913 essential and all ICAP errors will result in an error page
8914 returned to the HTTP client.
8916 Bypass is off by default: services are treated as essential.
8919 If set to 'on' or '1', the ICAP service is allowed to
8920 dynamically change the current message adaptation plan by
8921 returning a chain of services to be used next. The services
8922 are specified using the X-Next-Services ICAP response header
8923 value, formatted as a comma-separated list of service names.
8924 Each named service should be configured in squid.conf. Other
8925 services are ignored. An empty X-Next-Services value results
8926 in an empty plan which ends the current adaptation.
8928 Dynamic adaptation plan may cross or cover multiple supported
8929 vectoring points in their natural processing order.
8931 Routing is not allowed by default: the ICAP X-Next-Services
8932 response header is ignored.
8935 Only has effect on split-stack systems. The default on those systems
8936 is to use IPv4-only connections. When set to 'on' this option will
8937 make Squid use IPv6-only connections to contact this ICAP service.
8939 on-overload=block|bypass|wait|force
8940 If the service Max-Connections limit has been reached, do
8941 one of the following for each new ICAP transaction:
8942 * block: send an HTTP error response to the client
8943 * bypass: ignore the "over-connected" ICAP service
8944 * wait: wait (in a FIFO queue) for an ICAP connection slot
8945 * force: proceed, ignoring the Max-Connections limit
8947 In SMP mode with N workers, each worker assumes the service
8948 connection limit is Max-Connections/N, even though not all
8949 workers may use a given service.
8951 The default value is "bypass" if service is bypassable,
8952 otherwise it is set to "wait".
8956 Use the given number as the Max-Connections limit, regardless
8957 of the Max-Connections value given by the service, if any.
8959 connection-encryption=on|off
8960 Determines the ICAP service effect on the connections_encrypted
8963 The default is "on" for Secure ICAP services (i.e., those
8964 with the icaps:// service URIs scheme) and "off" for plain ICAP
8967 Does not affect ICAP connections (e.g., does not turn Secure
8970 ==== ICAPS / TLS OPTIONS ====
8972 These options are used for Secure ICAP (icaps://....) services only.
8974 tls-cert=/path/to/ssl/certificate
8975 A client SSL certificate to use when connecting to
8978 tls-key=/path/to/ssl/key
8979 The private TLS/SSL key corresponding to sslcert above.
8980 If 'tls-key' is not specified 'tls-cert' is assumed to
8981 reference a combined PEM format file containing both the
8982 certificate and the key.
8984 tls-cipher=... The list of valid TLS/SSL ciphers to use when connecting
8985 to this icap server.
8988 The minimum TLS protocol version to permit. To control
8989 SSLv3 use the tls-options= parameter.
8990 Supported Values: 1.0 (default), 1.1, 1.2
8992 tls-options=... Specify various OpenSSL library options:
8994 NO_SSLv3 Disallow the use of SSLv3
8997 Always create a new key when using
8998 temporary/ephemeral DH key exchanges
9000 ALL Enable various bug workarounds
9001 suggested as "harmless" by OpenSSL
9002 Be warned that this reduces SSL/TLS
9003 strength to some attacks.
9005 See the OpenSSL SSL_CTX_set_options documentation for a
9006 more complete list. Options relevant only to SSLv2 are
9009 tls-cafile= PEM file containing CA certificates to use when verifying
9010 the icap server certificate.
9011 Use to specify intermediate CA certificate(s) if not sent
9012 by the server. Or the full CA chain for the server when
9013 using the tls-default-ca=off flag.
9014 May be repeated to load multiple files.
9016 tls-capath=... A directory containing additional CA certificates to
9017 use when verifying the icap server certificate.
9018 Requires OpenSSL or LibreSSL.
9020 tls-crlfile=... A certificate revocation list file to use when
9021 verifying the icap server certificate.
9023 tls-flags=... Specify various flags modifying the Squid TLS implementation:
9026 Accept certificates even if they fail to
9029 Don't verify the icap server certificate
9030 matches the server name
9032 tls-default-ca[=off]
9033 Whether to use the system Trusted CAs. Default is ON.
9035 tls-domain= The icap server name as advertised in it's certificate.
9036 Used for verifying the correctness of the received icap
9037 server certificate. If not specified the icap server
9038 hostname extracted from ICAP URI will be used.
9040 Older icap_service format without optional named parameters is
9041 deprecated but supported for backward compatibility.
9044 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
9045 icap_service svcLogger reqmod_precache icaps://icap2.mydomain.net:11344/reqmod routing=on
9049 TYPE: icap_class_type
9054 This deprecated option was documented to define an ICAP service
9055 chain, even though it actually defined a set of similar, redundant
9056 services, and the chains were not supported.
9058 To define a set of redundant services, please use the
9059 adaptation_service_set directive. For service chains, use
9060 adaptation_service_chain.
9064 TYPE: icap_access_type
9069 This option is deprecated. Please use adaptation_access, which
9070 has the same ICAP functionality, but comes with better
9071 documentation, and eCAP support.
9076 -----------------------------------------------------------------------------
9083 LOC: Adaptation::Ecap::TheConfig.onoff
9086 Controls whether eCAP support is enabled.
9090 TYPE: ecap_service_type
9092 LOC: Adaptation::Ecap::TheConfig
9095 Defines a single eCAP service
9097 ecap_service id vectoring_point uri [option ...]
9100 an opaque identifier or name which is used to direct traffic to
9101 this specific service. Must be unique among all adaptation
9102 services in squid.conf.
9104 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
9105 This specifies at which point of transaction processing the
9106 eCAP service should be activated. *_postcache vectoring points
9107 are not yet supported.
9109 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
9110 Squid uses the eCAP service URI to match this configuration
9111 line with one of the dynamically loaded services. Each loaded
9112 eCAP service must have a unique URI. Obtain the right URI from
9113 the service provider.
9115 To activate a service, use the adaptation_access directive. To group
9116 services, use adaptation_service_chain and adaptation_service_set.
9118 Service options are separated by white space. eCAP services support
9119 the following name=value options:
9122 If set to 'on' or '1', the eCAP service is treated as optional.
9123 If the service cannot be reached or malfunctions, Squid will try
9124 to ignore any errors and process the message as if the service
9125 was not enabled. No all eCAP errors can be bypassed.
9126 If set to 'off' or '0', the eCAP service is treated as essential
9127 and all eCAP errors will result in an error page returned to the
9130 Bypass is off by default: services are treated as essential.
9133 If set to 'on' or '1', the eCAP service is allowed to
9134 dynamically change the current message adaptation plan by
9135 returning a chain of services to be used next.
9137 Dynamic adaptation plan may cross or cover multiple supported
9138 vectoring points in their natural processing order.
9140 Routing is not allowed by default.
9142 connection-encryption=on|off
9143 Determines the eCAP service effect on the connections_encrypted
9146 Defaults to "on", which does not taint the master transaction
9149 Does not affect eCAP API calls.
9151 Older ecap_service format without optional named parameters is
9152 deprecated but supported for backward compatibility.
9156 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
9157 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
9160 NAME: loadable_modules
9162 IFDEF: USE_LOADABLE_MODULES
9163 LOC: Config.loadable_module_names
9166 Instructs Squid to load the specified dynamic module(s) or activate
9167 preloaded module(s).
9169 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
9173 MESSAGE ADAPTATION OPTIONS
9174 -----------------------------------------------------------------------------
9177 NAME: adaptation_service_set
9178 TYPE: adaptation_service_set_type
9179 IFDEF: USE_ADAPTATION
9184 Configures an ordered set of similar, redundant services. This is
9185 useful when hot standby or backup adaptation servers are available.
9187 adaptation_service_set set_name service_name1 service_name2 ...
9189 The named services are used in the set declaration order. The first
9190 applicable adaptation service from the set is used first. The next
9191 applicable service is tried if and only if the transaction with the
9192 previous service fails and the message waiting to be adapted is still
9195 When adaptation starts, broken services are ignored as if they were
9196 not a part of the set. A broken service is a down optional service.
9198 The services in a set must be attached to the same vectoring point
9199 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9201 If all services in a set are optional then adaptation failures are
9202 bypassable. If all services in the set are essential, then a
9203 transaction failure with one service may still be retried using
9204 another service from the set, but when all services fail, the master
9205 transaction fails as well.
9207 A set may contain a mix of optional and essential services, but that
9208 is likely to lead to surprising results because broken services become
9209 ignored (see above), making previously bypassable failures fatal.
9210 Technically, it is the bypassability of the last failed service that
9213 See also: adaptation_access adaptation_service_chain
9216 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
9217 adaptation service_set svcLogger loggerLocal loggerRemote
9220 NAME: adaptation_service_chain
9221 TYPE: adaptation_service_chain_type
9222 IFDEF: USE_ADAPTATION
9227 Configures a list of complementary services that will be applied
9228 one-by-one, forming an adaptation chain or pipeline. This is useful
9229 when Squid must perform different adaptations on the same message.
9231 adaptation_service_chain chain_name service_name1 svc_name2 ...
9233 The named services are used in the chain declaration order. The first
9234 applicable adaptation service from the chain is used first. The next
9235 applicable service is applied to the successful adaptation results of
9236 the previous service in the chain.
9238 When adaptation starts, broken services are ignored as if they were
9239 not a part of the chain. A broken service is a down optional service.
9241 Request satisfaction terminates the adaptation chain because Squid
9242 does not currently allow declaration of RESPMOD services at the
9243 "reqmod_precache" vectoring point (see icap_service or ecap_service).
9245 The services in a chain must be attached to the same vectoring point
9246 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
9248 A chain may contain a mix of optional and essential services. If an
9249 essential adaptation fails (or the failure cannot be bypassed for
9250 other reasons), the master transaction fails. Otherwise, the failure
9251 is bypassed as if the failed adaptation service was not in the chain.
9253 See also: adaptation_access adaptation_service_set
9256 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
9259 NAME: adaptation_access
9260 TYPE: adaptation_access_type
9261 IFDEF: USE_ADAPTATION
9264 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
9266 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
9268 adaptation_access service_name allow|deny [!]aclname...
9269 adaptation_access set_name allow|deny [!]aclname...
9271 At each supported vectoring point, the adaptation_access
9272 statements are processed in the order they appear in this
9273 configuration file. Statements pointing to the following services
9274 are ignored (i.e., skipped without checking their ACL):
9276 - services serving different vectoring points
9277 - "broken-but-bypassable" services
9278 - "up" services configured to ignore such transactions
9279 (e.g., based on the ICAP Transfer-Ignore header).
9281 When a set_name is used, all services in the set are checked
9282 using the same rules, to find the first applicable one. See
9283 adaptation_service_set for details.
9285 If an access list is checked and there is a match, the
9286 processing stops: For an "allow" rule, the corresponding
9287 adaptation service is used for the transaction. For a "deny"
9288 rule, no adaptation service is activated.
9290 It is currently not possible to apply more than one adaptation
9291 service at the same vectoring point to the same HTTP transaction.
9293 See also: icap_service and ecap_service
9296 adaptation_access service_1 allow all
9299 NAME: adaptation_service_iteration_limit
9301 IFDEF: USE_ADAPTATION
9302 LOC: Adaptation::Config::service_iteration_limit
9305 Limits the number of iterations allowed when applying adaptation
9306 services to a message. If your longest adaptation set or chain
9307 may have more than 16 services, increase the limit beyond its
9308 default value of 16. If detecting infinite iteration loops sooner
9309 is critical, make the iteration limit match the actual number
9310 of services in your longest adaptation set or chain.
9312 Infinite adaptation loops are most likely with routing services.
9314 See also: icap_service routing=1
9317 NAME: adaptation_masterx_shared_names
9319 IFDEF: USE_ADAPTATION
9320 LOC: Adaptation::Config::masterx_shared_name
9323 For each master transaction (i.e., the HTTP request and response
9324 sequence, including all related ICAP and eCAP exchanges), Squid
9325 maintains a table of metadata. The table entries are (name, value)
9326 pairs shared among eCAP and ICAP exchanges. The table is destroyed
9327 with the master transaction.
9329 This option specifies the table entry names that Squid must accept
9330 from and forward to the adaptation transactions.
9332 An ICAP REQMOD or RESPMOD transaction may set an entry in the
9333 shared table by returning an ICAP header field with a name
9334 specified in adaptation_masterx_shared_names.
9336 An eCAP REQMOD or RESPMOD transaction may set an entry in the
9337 shared table by implementing the libecap::visitEachOption() API
9338 to provide an option with a name specified in
9339 adaptation_masterx_shared_names.
9341 Squid will store and forward the set entry to subsequent adaptation
9342 transactions within the same master transaction scope.
9344 Only one shared entry name is supported at this time.
9347 # share authentication information among ICAP services
9348 adaptation_masterx_shared_names X-Subscriber-ID
9351 NAME: adaptation_meta
9353 IFDEF: USE_ADAPTATION
9354 LOC: Adaptation::Config::metaHeaders
9357 This option allows Squid administrator to add custom ICAP request
9358 headers or eCAP options to Squid ICAP requests or eCAP transactions.
9359 Use it to pass custom authentication tokens and other
9360 transaction-state related meta information to an ICAP/eCAP service.
9362 The addition of a meta header is ACL-driven:
9363 adaptation_meta name value [!]aclname ...
9365 Processing for a given header name stops after the first ACL list match.
9366 Thus, it is impossible to add two headers with the same name. If no ACL
9367 lists match for a given header name, no such header is added. For
9370 # do not debug transactions except for those that need debugging
9371 adaptation_meta X-Debug 1 needs_debugging
9373 # log all transactions except for those that must remain secret
9374 adaptation_meta X-Log 1 !keep_secret
9376 # mark transactions from users in the "G 1" group
9377 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
9379 The "value" parameter may be a regular squid.conf token or a "double
9380 quoted string". Within the quoted string, use backslash (\) to escape
9381 any character, which is currently only useful for escaping backslashes
9382 and double quotes. For example,
9383 "this string has one backslash (\\) and two \"quotes\""
9385 Used adaptation_meta header values may be logged via %note
9386 logformat code. If multiple adaptation_meta headers with the same name
9387 are used during master transaction lifetime, the header values are
9388 logged in the order they were used and duplicate values are ignored
9389 (only the first repeated value will be logged).
9395 LOC: Adaptation::Icap::TheConfig.repeat
9396 DEFAULT_IF_NONE: deny all
9398 This ACL determines which retriable ICAP transactions are
9399 retried. Transactions that received a complete ICAP response
9400 and did not have to consume or produce HTTP bodies to receive
9401 that response are usually retriable.
9403 icap_retry allow|deny [!]aclname ...
9405 Squid automatically retries some ICAP I/O timeouts and errors
9406 due to persistent connection race conditions.
9408 See also: icap_retry_limit
9411 NAME: icap_retry_limit
9414 LOC: Adaptation::Icap::TheConfig.repeat_limit
9416 DEFAULT_DOC: No retries are allowed.
9418 Limits the number of retries allowed.
9420 Communication errors due to persistent connection race
9421 conditions are unavoidable, automatically retried, and do not
9422 count against this limit.
9424 See also: icap_retry
9430 -----------------------------------------------------------------------------
9433 NAME: check_hostnames
9436 LOC: Config.onoff.check_hostnames
9438 For security and stability reasons Squid can check
9439 hostnames for Internet standard RFC compliance. If you want
9440 Squid to perform these checks turn this directive on.
9443 NAME: allow_underscore
9446 LOC: Config.onoff.allow_underscore
9448 Underscore characters is not strictly allowed in Internet hostnames
9449 but nevertheless used by many sites. Set this to off if you want
9450 Squid to be strict about the standard.
9451 This check is performed only when check_hostnames is set to on.
9454 NAME: dns_retransmit_interval
9457 LOC: Config.Timeout.idns_retransmit
9459 Initial retransmit interval for DNS queries. The interval is
9460 doubled each time all configured DNS servers have been tried.
9466 LOC: Config.Timeout.idns_query
9468 DNS Query timeout. If no response is received to a DNS query
9469 within this time all DNS servers for the queried domain
9470 are assumed to be unavailable.
9473 NAME: dns_packet_max
9475 DEFAULT_DOC: EDNS disabled
9477 LOC: Config.dns.packet_max
9479 Maximum number of bytes packet size to advertise via EDNS.
9480 Set to "none" to disable EDNS large packet support.
9482 For legacy reasons DNS UDP replies will default to 512 bytes which
9483 is too small for many responses. EDNS provides a means for Squid to
9484 negotiate receiving larger responses back immediately without having
9485 to failover with repeat requests. Responses larger than this limit
9486 will retain the old behaviour of failover to TCP DNS.
9488 Squid has no real fixed limit internally, but allowing packet sizes
9489 over 1500 bytes requires network jumbogram support and is usually not
9492 WARNING: The RFC also indicates that some older resolvers will reply
9493 with failure of the whole request if the extension is added. Some
9494 resolvers have already been identified which will reply with mangled
9495 EDNS response on occasion. Usually in response to many-KB jumbogram
9496 sizes being advertised by Squid.
9497 Squid will currently treat these both as an unable-to-resolve domain
9498 even if it would be resolvable without EDNS.
9505 DEFAULT_DOC: Search for single-label domain names is disabled.
9506 LOC: Config.onoff.res_defnames
9508 Normally the RES_DEFNAMES resolver option is disabled
9509 (see res_init(3)). This prevents caches in a hierarchy
9510 from interpreting single-component hostnames locally. To allow
9511 Squid to handle single-component names, enable this option.
9514 NAME: dns_multicast_local
9518 DEFAULT_DOC: Search for .local and .arpa names is disabled.
9519 LOC: Config.onoff.dns_mdns
9521 When set to on, Squid sends multicast DNS lookups on the local
9522 network for domains ending in .local and .arpa.
9523 This enables local servers and devices to be contacted in an
9524 ad-hoc or zero-configuration network environment.
9527 NAME: dns_nameservers
9530 DEFAULT_DOC: Use operating system definitions
9531 LOC: Config.dns.nameservers
9533 Use this if you want to specify a list of DNS name servers
9534 (IP addresses) to use instead of those given in your
9535 /etc/resolv.conf file.
9537 On Windows platforms, if no value is specified here or in
9538 the /etc/resolv.conf file, the list of DNS name servers are
9539 taken from the Windows registry, both static and dynamic DHCP
9540 configurations are supported.
9542 Example: dns_nameservers 10.0.0.1 192.172.0.4
9547 DEFAULT: @DEFAULT_HOSTS@
9548 LOC: Config.etcHostsPath
9550 Location of the host-local IP name-address associations
9551 database. Most Operating Systems have such a file on different
9553 - Un*X & Linux: /etc/hosts
9554 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
9555 (%SystemRoot% value install default is c:\winnt)
9556 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
9557 (%SystemRoot% value install default is c:\windows)
9558 - Windows 9x/Me: %windir%\hosts
9559 (%windir% value is usually c:\windows)
9560 - Cygwin: /etc/hosts
9562 The file contains newline-separated definitions, in the
9563 form ip_address_in_dotted_form name [name ...] names are
9564 whitespace-separated. Lines beginning with an hash (#)
9565 character are comments.
9567 The file is checked at startup and upon configuration.
9568 If set to 'none', it won't be checked.
9569 If append_domain is used, that domain will be added to
9570 domain-local (i.e. not containing any dot character) host
9576 LOC: Config.appendDomain
9578 DEFAULT_DOC: Use operating system definitions
9580 Appends local domain name to hostnames without any dots in
9581 them. append_domain must begin with a period.
9583 Be warned there are now Internet names with no dots in
9584 them using only top-domain names, so setting this may
9585 cause some Internet sites to become unavailable.
9588 append_domain .yourdomain.com
9591 NAME: ignore_unknown_nameservers
9593 LOC: Config.onoff.ignore_unknown_nameservers
9596 By default Squid checks that DNS responses are received
9597 from the same IP addresses they are sent to. If they
9598 don't match, Squid ignores the response and writes a warning
9599 message to cache.log. You can allow responses from unknown
9600 nameservers by setting this option to 'off'.
9606 LOC: Config.dns.v4_first
9608 With the IPv6 Internet being as fast or faster than IPv4 Internet
9609 for most networks Squid prefers to contact websites over IPv6.
9611 This option reverses the order of preference to make Squid contact
9612 dual-stack websites over IPv4 first. Squid will still perform both
9613 IPv6 and IPv4 DNS lookups before connecting.
9616 This option will restrict the situations under which IPv6
9617 connectivity is used (and tested). Hiding network problems
9618 which would otherwise be detected and warned about.
9622 COMMENT: (number of entries)
9625 LOC: Config.ipcache.size
9627 Maximum number of DNS IP cache entries.
9634 LOC: Config.ipcache.low
9641 LOC: Config.ipcache.high
9643 The size, low-, and high-water marks for the IP cache.
9646 NAME: fqdncache_size
9647 COMMENT: (number of entries)
9650 LOC: Config.fqdncache.size
9652 Maximum number of FQDN cache entries.
9657 -----------------------------------------------------------------------------
9660 NAME: configuration_includes_quoted_values
9662 TYPE: configuration_includes_quoted_values
9664 LOC: ConfigParser::RecognizeQuotedValues
9666 If set, Squid will recognize each "quoted string" after a configuration
9667 directive as a single parameter. The quotes are stripped before the
9668 parameter value is interpreted or used.
9669 See "Values with spaces, quotes, and other special characters"
9670 section for more details.
9677 LOC: Config.onoff.mem_pools
9679 If set, Squid will keep pools of allocated (but unused) memory
9680 available for future use. If memory is a premium on your
9681 system and you believe your malloc library outperforms Squid
9682 routines, disable this.
9685 NAME: memory_pools_limit
9689 LOC: Config.MemPools.limit
9691 Used only with memory_pools on:
9692 memory_pools_limit 50 MB
9694 If set to a non-zero value, Squid will keep at most the specified
9695 limit of allocated (but unused) memory in memory pools. All free()
9696 requests that exceed this limit will be handled by your malloc
9697 library. Squid does not pre-allocate any memory, just safe-keeps
9698 objects that otherwise would be free()d. Thus, it is safe to set
9699 memory_pools_limit to a reasonably high value even if your
9700 configuration will use less memory.
9702 If set to none, Squid will keep all memory it can. That is, there
9703 will be no limit on the total amount of memory used for safe-keeping.
9705 To disable memory allocation optimization, do not set
9706 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
9708 An overhead for maintaining memory pools is not taken into account
9709 when the limit is checked. This overhead is close to four bytes per
9710 object kept. However, pools may actually _save_ memory because of
9711 reduced memory thrashing in your malloc library.
9715 COMMENT: on|off|transparent|truncate|delete
9718 LOC: opt_forwarded_for
9720 If set to "on", Squid will append your client's IP address
9721 in the HTTP requests it forwards. By default it looks like:
9723 X-Forwarded-For: 192.1.2.3
9725 If set to "off", it will appear as
9727 X-Forwarded-For: unknown
9729 If set to "transparent", Squid will not alter the
9730 X-Forwarded-For header in any way.
9732 If set to "delete", Squid will delete the entire
9733 X-Forwarded-For header.
9735 If set to "truncate", Squid will remove all existing
9736 X-Forwarded-For entries, and place the client IP as the sole entry.
9739 NAME: cachemgr_passwd
9740 TYPE: cachemgrpasswd
9742 DEFAULT_DOC: No password. Actions which require password are denied.
9743 LOC: Config.passwd_list
9745 Specify passwords for cachemgr operations.
9747 Usage: cachemgr_passwd password action action ...
9749 Some valid actions are (see cache manager menu for a full list):
9789 * Indicates actions which will not be performed without a
9790 valid password, others can be performed if not listed here.
9792 To disable an action, set the password to "disable".
9793 To allow performing an action without a password, set the
9796 Use the keyword "all" to set the same password for all actions.
9799 cachemgr_passwd secret shutdown
9800 cachemgr_passwd lesssssssecret info stats/objects
9801 cachemgr_passwd disable all
9808 LOC: Config.onoff.client_db
9810 If you want to disable collecting per-client statistics,
9811 turn off client_db here.
9814 NAME: refresh_all_ims
9818 LOC: Config.onoff.refresh_all_ims
9820 When you enable this option, squid will always check
9821 the origin server for an update when a client sends an
9822 If-Modified-Since request. Many browsers use IMS
9823 requests when the user requests a reload, and this
9824 ensures those clients receive the latest version.
9826 By default (off), squid may return a Not Modified response
9827 based on the age of the cached version.
9830 NAME: reload_into_ims
9831 IFDEF: USE_HTTP_VIOLATIONS
9835 LOC: Config.onoff.reload_into_ims
9837 When you enable this option, client no-cache or ``reload''
9838 requests will be changed to If-Modified-Since requests.
9839 Doing this VIOLATES the HTTP standard. Enabling this
9840 feature could make you liable for problems which it
9843 see also refresh_pattern for a more selective approach.
9846 NAME: connect_retries
9848 LOC: Config.connect_retries
9850 DEFAULT_DOC: Do not retry failed connections.
9852 This sets the maximum number of connection attempts made for each
9853 TCP connection. The connect_retries attempts must all still
9854 complete within the connection timeout period.
9856 The default is not to re-try if the first connection attempt fails.
9857 The (not recommended) maximum is 10 tries.
9859 A warning message will be generated if it is set to a too-high
9860 value and the configured value will be over-ridden.
9862 Note: These re-tries are in addition to forward_max_tries
9863 which limit how many different addresses may be tried to find
9867 NAME: retry_on_error
9869 LOC: Config.retry.onerror
9872 If set to ON Squid will automatically retry requests when
9873 receiving an error response with status 403 (Forbidden),
9874 500 (Internal Error), 501 or 503 (Service not available).
9875 Status 502 and 504 (Gateway errors) are always retried.
9877 This is mainly useful if you are in a complex cache hierarchy to
9878 work around access control errors.
9880 NOTE: This retry will attempt to find another working destination.
9881 Which is different from the server which just failed.
9884 NAME: as_whois_server
9886 LOC: Config.as_whois_server
9887 DEFAULT: whois.ra.net
9889 WHOIS server to query for AS numbers. NOTE: AS numbers are
9890 queried only when Squid starts up, not for every request.
9895 LOC: Config.onoff.offline
9898 Enable this option and Squid will never try to validate cached
9902 NAME: uri_whitespace
9903 TYPE: uri_whitespace
9904 LOC: Config.uri_whitespace
9907 What to do with requests that have whitespace characters in the
9910 strip: The whitespace characters are stripped out of the URL.
9911 This is the behavior recommended by RFC2396 and RFC3986
9912 for tolerant handling of generic URI.
9913 NOTE: This is one difference between generic URI and HTTP URLs.
9915 deny: The request is denied. The user receives an "Invalid
9917 This is the behaviour recommended by RFC2616 for safe
9918 handling of HTTP request URL.
9920 allow: The request is allowed and the URI is not changed. The
9921 whitespace characters remain in the URI. Note the
9922 whitespace is passed to redirector processes if they
9924 Note this may be considered a violation of RFC2616
9925 request parsing where whitespace is prohibited in the
9928 encode: The request is allowed and the whitespace characters are
9929 encoded according to RFC1738.
9931 chop: The request is allowed and the URI is chopped at the
9935 NOTE the current Squid implementation of encode and chop violates
9936 RFC2616 by not using a 301 redirect after altering the URL.
9941 LOC: Config.chroot_dir
9944 Specifies a directory where Squid should do a chroot() while
9945 initializing. This also causes Squid to fully drop root
9946 privileges after initializing. This means, for example, if you
9947 use a HTTP port less than 1024 and try to reconfigure, you may
9948 get an error saying that Squid can not open the port.
9951 NAME: balance_on_multiple_ip
9953 LOC: Config.onoff.balance_on_multiple_ip
9956 Modern IP resolvers in squid sort lookup results by preferred access.
9957 By default squid will use these IP in order and only rotates to
9958 the next listed when the most preffered fails.
9960 Some load balancing servers based on round robin DNS have been
9961 found not to preserve user session state across requests
9962 to different IP addresses.
9964 Enabling this directive Squid rotates IP's per request.
9967 NAME: pipeline_prefetch
9968 TYPE: pipelinePrefetch
9969 LOC: Config.pipeline_max_prefetch
9971 DEFAULT_DOC: Do not pre-parse pipelined requests.
9973 HTTP clients may send a pipeline of 1+N requests to Squid using a
9974 single connection, without waiting for Squid to respond to the first
9975 of those requests. This option limits the number of concurrent
9976 requests Squid will try to handle in parallel. If set to N, Squid
9977 will try to receive and process up to 1+N requests on the same
9978 connection concurrently.
9980 Defaults to 0 (off) for bandwidth management and access logging
9983 NOTE: pipelining requires persistent connections to clients.
9985 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
9988 NAME: high_response_time_warning
9991 LOC: Config.warnings.high_rptm
9993 DEFAULT_DOC: disabled.
9995 If the one-minute median response time exceeds this value,
9996 Squid prints a WARNING with debug level 0 to get the
9997 administrators attention. The value is in milliseconds.
10000 NAME: high_page_fault_warning
10002 LOC: Config.warnings.high_pf
10004 DEFAULT_DOC: disabled.
10006 If the one-minute average page fault rate exceeds this
10007 value, Squid prints a WARNING with debug level 0 to get
10008 the administrators attention. The value is in page faults
10012 NAME: high_memory_warning
10014 LOC: Config.warnings.high_memory
10015 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
10017 DEFAULT_DOC: disabled.
10019 If the memory usage (as determined by gnumalloc, if available and used)
10020 exceeds this amount, Squid prints a WARNING with debug level 0 to get
10021 the administrators attention.
10023 # TODO: link high_memory_warning to mempools?
10025 NAME: sleep_after_fork
10026 COMMENT: (microseconds)
10028 LOC: Config.sleep_after_fork
10031 When this is set to a non-zero value, the main Squid process
10032 sleeps the specified number of microseconds after a fork()
10033 system call. This sleep may help the situation where your
10034 system reports fork() failures due to lack of (virtual)
10035 memory. Note, however, if you have a lot of child
10036 processes, these sleep delays will add up and your
10037 Squid will not service requests for some amount of time
10038 until all the child processes have been started.
10039 On Windows value less then 1000 (1 milliseconds) are
10043 NAME: windows_ipaddrchangemonitor
10044 IFDEF: _SQUID_WINDOWS_
10048 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
10050 On Windows Squid by default will monitor IP address changes and will
10051 reconfigure itself after any detected event. This is very useful for
10052 proxies connected to internet with dial-up interfaces.
10053 In some cases (a Proxy server acting as VPN gateway is one) it could be
10054 desiderable to disable this behaviour setting this to 'off'.
10055 Note: after changing this, Squid service must be restarted.
10060 IFDEF: USE_SQUID_EUI
10062 LOC: Eui::TheConfig.euiLookup
10064 Whether to lookup the EUI or MAC address of a connected client.
10067 NAME: max_filedescriptors max_filedesc
10070 DEFAULT_DOC: Use operating system limits set by ulimit.
10071 LOC: Config.max_filedescriptors
10073 Reduce the maximum number of filedescriptors supported below
10074 the usual operating system defaults.
10076 Remove from squid.conf to inherit the current ulimit setting.
10078 Note: Changing this requires a restart of Squid. Also
10079 not all I/O types supports large values (eg on Windows).
10082 NAME: force_request_body_continuation
10084 LOC: Config.accessList.forceRequestBodyContinuation
10086 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
10088 This option controls how Squid handles data upload requests from HTTP
10089 and FTP agents that require a "Please Continue" control message response
10090 to actually send the request body to Squid. It is mostly useful in
10091 adaptation environments.
10093 When Squid receives an HTTP request with an "Expect: 100-continue"
10094 header or an FTP upload command (e.g., STOR), Squid normally sends the
10095 request headers or FTP command information to an adaptation service (or
10096 peer) and waits for a response. Most adaptation services (and some
10097 broken peers) may not respond to Squid at that stage because they may
10098 decide to wait for the HTTP request body or FTP data transfer. However,
10099 that request body or data transfer may never come because Squid has not
10100 responded with the HTTP 100 or FTP 150 (Please Continue) control message
10101 to the request sender yet!
10103 An allow match tells Squid to respond with the HTTP 100 or FTP 150
10104 (Please Continue) control message on its own, before forwarding the
10105 request to an adaptation service or peer. Such a response usually forces
10106 the request sender to proceed with sending the body. A deny match tells
10107 Squid to delay that control response until the origin server confirms
10108 that the request body is needed. Delaying is the default behavior.
10111 NAME: server_pconn_for_nonretriable
10114 DEFAULT_DOC: Open new connections for forwarding requests Squid cannot retry safely.
10115 LOC: Config.accessList.serverPconnForNonretriable
10117 This option provides fine-grained control over persistent connection
10118 reuse when forwarding HTTP requests that Squid cannot retry. It is useful
10119 in environments where opening new connections is very expensive
10120 (e.g., all connections are secured with TLS with complex client and server
10121 certificate validation) and race conditions associated with persistent
10122 connections are very rare and/or only cause minor problems.
10124 HTTP prohibits retrying unsafe and non-idempotent requests (e.g., POST).
10125 Squid limitations also prohibit retrying all requests with bodies (e.g., PUT).
10126 By default, when forwarding such "risky" requests, Squid opens a new
10127 connection to the server or cache_peer, even if there is an idle persistent
10128 connection available. When Squid is configured to risk sending a non-retriable
10129 request on a previously used persistent connection, and the server closes
10130 the connection before seeing that risky request, the user gets an error response
10131 from Squid. In most cases, that error response will be HTTP 502 (Bad Gateway)
10132 with ERR_ZERO_SIZE_OBJECT or ERR_WRITE_ERROR (peer connection reset) error detail.
10134 If an allow rule matches, Squid reuses an available idle persistent connection
10135 (if any) for the request that Squid cannot retry. If a deny rule matches, then
10136 Squid opens a new connection for the request that Squid cannot retry.
10138 This option does not affect requests that Squid can retry. They will reuse idle
10139 persistent connections (if any).
10141 This clause only supports fast acl types.
10142 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
10145 acl SpeedIsWorthTheRisk method POST
10146 server_pconn_for_nonretriable allow SpeedIsWorthTheRisk