1 ## Copyright (C) 1996-2016 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 3.6
167 NAME: cache_peer_domain cache_host_domain
170 Replace with dstdomain ACLs and cache_peer_access.
173 NAME: sslproxy_cafile
176 Remove this line. Use tls_outgoing_options cafile= instead.
179 NAME: sslproxy_capath
182 Remove this line. Use tls_outgoing_options capath= instead.
185 NAME: sslproxy_cipher
188 Remove this line. Use tls_outgoing_options cipher= instead.
191 NAME: sslproxy_client_certificate
194 Remove this line. Use tls_outgoing_options cert= instead.
197 NAME: sslproxy_client_key
200 Remove this line. Use tls_outgoing_options key= instead.
206 Remove this line. Use tls_outgoing_options flags= instead.
209 NAME: sslproxy_options
212 Remove this line. Use tls_outgoing_options options= instead.
215 NAME: sslproxy_version
218 Remove this line. Use tls_outgoing_options options= instead.
221 # Options removed in 3.5
222 NAME: hierarchy_stoplist
225 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
228 # Options removed in 3.4
232 Remove this line. Use acls with access_log directives to control access logging
238 Remove this line. Use acls with icap_log directives to control icap logging
241 # Options Removed in 3.3
242 NAME: ignore_ims_on_miss
245 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
248 # Options Removed in 3.2
249 NAME: chunked_request_body_max_size
252 Remove this line. Squid is now HTTP/1.1 compliant.
255 NAME: dns_v4_fallback
258 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
261 NAME: emulate_httpd_log
264 Replace this with an access_log directive using the format 'common' or 'combined'.
270 Use a regular access.log with ACL limiting it to MISS events.
276 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
279 NAME: ignore_expect_100
282 Remove this line. The HTTP/1.1 feature is now fully supported by default.
288 Remove this option from your config. To log FQDN use %>A in the log format.
291 NAME: log_ip_on_direct
294 Remove this option from your config. To log server or peer names use %<A in the log format.
297 NAME: maximum_single_addr_tries
300 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
303 NAME: referer_log referrer_log
306 Replace this with an access_log directive using the format 'referrer'.
312 Remove this line. The feature is supported by default in storage types where update is implemented.
315 NAME: url_rewrite_concurrency
318 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
324 Replace this with an access_log directive using the format 'useragent'.
327 # Options Removed in 3.1
331 Remove this line. DNS is no longer tested on startup.
334 NAME: extension_methods
337 Remove this line. All valid methods for HTTP are accepted by default.
340 # 2.7 Options Removed/Replaced in 3.2
345 # 2.7 Options Removed/Replaced in 3.1
353 Remove this line. HTTP/1.1 is supported by default.
356 NAME: upgrade_http0.9
359 Remove this line. ICY/1.0 streaming protocol is supported by default.
362 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
365 Alter these entries. Use the qos_flows directive instead.
368 # Options Removed in 3.0
372 Since squid-3.0 replace with request_header_access or reply_header_access
373 depending on whether you wish to match client requests or server replies.
376 NAME: httpd_accel_no_pmtu_disc
379 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
382 NAME: wais_relay_host
385 Replace this line with 'cache_peer' configuration.
388 NAME: wais_relay_port
391 Replace this line with 'cache_peer' configuration.
396 -----------------------------------------------------------------------------
403 DEFAULT_DOC: SMP support disabled.
405 Number of main Squid processes or "workers" to fork and maintain.
406 0: "no daemon" mode, like running "squid -N ..."
407 1: "no SMP" mode, start one main Squid process daemon (default)
408 N: start N main Squid process daemons (i.e., SMP mode)
410 In SMP mode, each worker does nearly all what a single Squid daemon
411 does (e.g., listen on http_port and forward HTTP requests).
414 NAME: cpu_affinity_map
416 LOC: Config.cpuAffinityMap
418 DEFAULT_DOC: Let operating system decide.
420 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
422 Sets 1:1 mapping between Squid processes and CPU cores. For example,
424 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
426 affects processes 1 through 4 only and places them on the first
427 four even cores, starting with core #1.
429 CPU cores are numbered starting from 1. Requires support for
430 sched_getaffinity(2) and sched_setaffinity(2) system calls.
432 Multiple cpu_affinity_map options are merged.
437 NAME: shared_memory_locking
440 LOC: Config.shmLocking
443 Whether to ensure that all required shared memory is available by
444 "locking" that shared memory into RAM when Squid starts. The
445 alternative is faster startup time followed by slightly slower
446 performance and, if not enough RAM is actually available during
447 runtime, mysterious crashes.
449 SMP Squid uses many shared memory segments. These segments are
450 brought into Squid memory space using an mmap(2) system call. During
451 Squid startup, the mmap() call often succeeds regardless of whether
452 the system has enough RAM. In general, Squid cannot tell whether the
453 kernel applies this "optimistic" memory allocation policy (but
454 popular modern kernels usually use it).
456 Later, if Squid attempts to actually access the mapped memory
457 regions beyond what the kernel is willing to allocate, the
458 "optimistic" kernel simply kills Squid kid with a SIGBUS signal.
459 Some of the memory limits enforced by the kernel are currently
460 poorly understood: We do not know how to detect and check them. This
461 option ensures that the mapped memory will be available.
463 This option may have a positive performance side-effect: Locking
464 memory at start avoids runtime paging I/O. Paging slows Squid down.
466 Locking memory may require a large enough RLIMIT_MEMLOCK OS limit,
467 CAP_IPC_LOCK capability, or equivalent.
471 OPTIONS FOR AUTHENTICATION
472 -----------------------------------------------------------------------------
481 This is used to define parameters for the various authentication
482 schemes supported by Squid.
484 format: auth_param scheme parameter [setting]
486 The order in which authentication schemes are presented to the client is
487 dependent on the order the scheme first appears in config file. IE
488 has a bug (it's not RFC 2617 compliant) in that it will use the basic
489 scheme if basic is the first entry presented, even if more secure
490 schemes are presented. For now use the order in the recommended
491 settings section below. If other browsers have difficulties (don't
492 recognize the schemes offered even if you are using basic) either
493 put basic first, or disable the other schemes (by commenting out their
496 Once an authentication scheme is fully configured, it can only be
497 shutdown by shutting squid down and restarting. Changes can be made on
498 the fly and activated with a reconfigure. I.E. You can change to a
499 different helper, but not unconfigure the helper completely.
501 Please note that while this directive defines how Squid processes
502 authentication it does not automatically activate authentication.
503 To use authentication you must in addition make use of ACLs based
504 on login name in http_access (proxy_auth, proxy_auth_regex or
505 external with %LOGIN used in the format tag). The browser will be
506 challenged for authentication on the first such acl encountered
507 in http_access processing and will also be re-challenged for new
508 login credentials if the request is being denied by a proxy_auth
511 WARNING: authentication can't be used in a transparently intercepting
512 proxy as the client then thinks it is talking to an origin server and
513 not the proxy. This is a limitation of bending the TCP/IP protocol to
514 transparently intercepting port 80, not a limitation in Squid.
515 Ports flagged 'transparent', 'intercept', or 'tproxy' have
516 authentication disabled.
518 === Parameters common to all schemes. ===
521 Specifies the command for the external authenticator.
523 By default, each authentication scheme is not used unless a
524 program is specified.
526 See http://wiki.squid-cache.org/Features/AddonHelpers for
527 more details on helper operations and creating your own.
530 Specifies a string to be append to request line format for
531 the authentication helper. "Quoted" format values may contain
532 spaces and logformat %macros. In theory, any logformat %macro
533 can be used. In practice, a %macro expands as a dash (-) if
534 the helper request is sent before the required macro
535 information is available to Squid.
537 By default, Squid uses request formats provided in
538 scheme-specific examples below (search for %credentials).
540 The expanded key_extras value is added to the Squid credentials
541 cache and, hence, will affect authentication. It can be used to
542 autenticate different users with identical user names (e.g.,
543 when user authentication depends on http_port).
545 Avoid adding frequently changing information to key_extras. For
546 example, if you add user source IP, and it changes frequently
547 in your environment, then max_user_ip ACL is going to treat
548 every user+IP combination as a unique "user", breaking the ACL
549 and wasting a lot of memory on those user records. It will also
550 force users to authenticate from scratch whenever their IP
554 Specifies the protection scope (aka realm name) which is to be
555 reported to the client for the authentication scheme. It is
556 commonly part of the text the user will see when prompted for
557 their username and password.
559 For Basic the default is "Squid proxy-caching web server".
560 For Digest there is no default, this parameter is mandatory.
561 For NTLM and Negotiate this parameter is ignored.
563 "children" numberofchildren [startup=N] [idle=N] [concurrency=N] [queue-size=N]
565 The maximum number of authenticator processes to spawn. If
566 you start too few Squid will have to wait for them to process
567 a backlog of credential verifications, slowing it down. When
568 password verifications are done via a (slow) network you are
569 likely to need lots of authenticator processes.
571 The startup= and idle= options permit some skew in the exact
572 amount run. A minimum of startup=N will begin during startup
573 and reconfigure. Squid will start more in groups of up to
574 idle=N in an attempt to meet traffic needs and to keep idle=N
575 free above those traffic needs up to the maximum.
577 The concurrency= option sets the number of concurrent requests
578 the helper can process. The default of 0 is used for helpers
579 who only supports one request at a time. Setting this to a
580 number greater than 0 changes the protocol used to include a
581 channel ID field first on the request/response line, allowing
582 multiple requests to be sent to the same helper in parallel
583 without waiting for the response.
585 Concurrency must not be set unless it's known the helper
586 supports the input format with channel-ID fields.
588 The queue-size= option sets the maximum number of queued
589 requests. If the queued requests exceed queue size for more
590 than 3 minutes then squid aborts its operation.
591 The default value is set to 2*numberofchildren/
593 NOTE: NTLM and Negotiate schemes do not support concurrency
594 in the Squid code module even though some helpers can.
597 IF HAVE_AUTH_MODULE_BASIC
598 === Basic authentication parameters ===
601 HTTP uses iso-latin-1 as character set, while some
602 authentication backends such as LDAP expects UTF-8. If this is
603 set to on Squid will translate the HTTP iso-latin-1 charset to
604 UTF-8 before sending the username and password to the helper.
606 "credentialsttl" timetolive
607 Specifies how long squid assumes an externally validated
608 username:password pair is valid for - in other words how
609 often the helper program is called for that user. Set this
610 low to force revalidation with short lived passwords.
612 NOTE: setting this high does not impact your susceptibility
613 to replay attacks unless you are using an one-time password
614 system (such as SecureID). If you are using such a system,
615 you will be vulnerable to replay attacks unless you also
616 use the max_user_ip ACL in an http_access rule.
618 "casesensitive" on|off
619 Specifies if usernames are case sensitive. Most user databases
620 are case insensitive allowing the same username to be spelled
621 using both lower and upper case letters, but some are case
622 sensitive. This makes a big difference for user_max_ip ACL
623 processing and similar.
626 IF HAVE_AUTH_MODULE_DIGEST
627 === Digest authentication parameters ===
630 HTTP uses iso-latin-1 as character set, while some
631 authentication backends such as LDAP expects UTF-8. If this is
632 set to on Squid will translate the HTTP iso-latin-1 charset to
633 UTF-8 before sending the username and password to the helper.
635 "nonce_garbage_interval" timeinterval
636 Specifies the interval that nonces that have been issued
637 to client_agent's are checked for validity.
639 "nonce_max_duration" timeinterval
640 Specifies the maximum length of time a given nonce will be
643 "nonce_max_count" number
644 Specifies the maximum number of times a given nonce can be
647 "nonce_strictness" on|off
648 Determines if squid requires strict increment-by-1 behavior
649 for nonce counts, or just incrementing (off - for use when
650 user agents generate nonce counts that occasionally miss 1
651 (ie, 1,2,4,6)). Default off.
653 "check_nonce_count" on|off
654 This directive if set to off can disable the nonce count check
655 completely to work around buggy digest qop implementations in
656 certain mainstream browser versions. Default on to check the
657 nonce count to protect from authentication replay attacks.
659 "post_workaround" on|off
660 This is a workaround to certain buggy browsers who send an
661 incorrect request digest in POST requests when reusing the
662 same nonce as acquired earlier on a GET request.
665 IF HAVE_AUTH_MODULE_NEGOTIATE
666 === Negotiate authentication parameters ===
669 If you experience problems with PUT/POST requests when using
670 the this authentication scheme then you can try setting this
671 to off. This will cause Squid to forcibly close the connection
672 on the initial request where the browser asks which schemes
673 are supported by the proxy.
676 IF HAVE_AUTH_MODULE_NTLM
677 === NTLM authentication parameters ===
680 If you experience problems with PUT/POST requests when using
681 the this authentication scheme then you can try setting this
682 to off. This will cause Squid to forcibly close the connection
683 on the initial request where the browser asks which schemes
684 are supported by the proxy.
687 === Example Configuration ===
689 This configuration displays the recommended authentication scheme
690 order from most to least secure with recommended minimum configuration
691 settings for each scheme:
693 #auth_param negotiate program <uncomment and complete this line to activate>
694 #auth_param negotiate children 20 startup=0 idle=1
695 #auth_param negotiate keep_alive on
697 #auth_param digest program <uncomment and complete this line to activate>
698 #auth_param digest children 20 startup=0 idle=1
699 #auth_param digest realm Squid proxy-caching web server
700 #auth_param digest nonce_garbage_interval 5 minutes
701 #auth_param digest nonce_max_duration 30 minutes
702 #auth_param digest nonce_max_count 50
704 #auth_param ntlm program <uncomment and complete this line to activate>
705 #auth_param ntlm children 20 startup=0 idle=1
706 #auth_param ntlm keep_alive on
708 #auth_param basic program <uncomment and complete this line>
709 #auth_param basic children 5 startup=5 idle=1
710 #auth_param basic realm Squid proxy-caching web server
711 #auth_param basic credentialsttl 2 hours
714 NAME: authenticate_cache_garbage_interval
717 LOC: Config.authenticateGCInterval
719 The time period between garbage collection across the username cache.
720 This is a trade-off between memory utilization (long intervals - say
721 2 days) and CPU (short intervals - say 1 minute). Only change if you
725 NAME: authenticate_ttl
728 LOC: Config.authenticateTTL
730 The time a user & their credentials stay in the logged in
731 user cache since their last request. When the garbage
732 interval passes, all user credentials that have passed their
733 TTL are removed from memory.
736 NAME: authenticate_ip_ttl
738 LOC: Config.authenticateIpTTL
741 If you use proxy authentication and the 'max_user_ip' ACL,
742 this directive controls how long Squid remembers the IP
743 addresses associated with each user. Use a small value
744 (e.g., 60 seconds) if your users might change addresses
745 quickly, as is the case with dialup. You might be safe
746 using a larger value (e.g., 2 hours) in a corporate LAN
747 environment with relatively static address assignments.
752 -----------------------------------------------------------------------------
755 NAME: external_acl_type
756 TYPE: externalAclHelper
757 LOC: Config.externalAclHelperList
760 This option defines external acl classes using a helper program
761 to look up the status
763 external_acl_type name [options] FORMAT /path/to/helper [helper arguments]
767 ttl=n TTL in seconds for cached results (defaults to 3600
771 TTL for cached negative lookups (default same
774 grace=n Percentage remaining of TTL where a refresh of a
775 cached entry should be initiated without needing to
776 wait for a new reply. (default is for no grace period)
778 cache=n The maximum number of entries in the result cache. The
779 default limit is 262144 entries. Each cache entry usually
780 consumes at least 256 bytes. Squid currently does not remove
781 expired cache entries until the limit is reached, so a proxy
782 will sooner or later reach the limit. The expanded FORMAT
783 value is used as the cache key, so if the details in FORMAT
784 are highly variable, a larger cache may be needed to produce
785 reduction in helper load.
788 Maximum number of acl helper processes spawned to service
789 external acl lookups of this type. (default 20)
792 Minimum number of acl helper processes to spawn during
793 startup and reconfigure to service external acl lookups
794 of this type. (default 0)
797 Number of acl helper processes to keep ahead of traffic
798 loads. Squid will spawn this many at once whenever load
799 rises above the capabilities of existing processes.
800 Up to the value of children-max. (default 1)
802 concurrency=n concurrency level per process. Only used with helpers
803 capable of processing more than one query at a time.
805 queue-size=N The queue-size= option sets the maximum number of queued
806 requests. If the queued requests exceed queue size
808 The default value is set to 2*children-max.
810 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
812 ipv4 / ipv6 IP protocol used to communicate with this helper.
813 The default is to auto-detect IPv6 and use it when available.
816 FORMAT is a series of %macro codes. See logformat directive for a full list
817 of the accepted codes. Although note that at the time of any external ACL
818 being tested data may not be available and thus some %macro expand to '-'.
820 In addition to the logformat codes; when processing external ACLs these
821 additional macros are made available:
823 %ACL The name of the ACL being tested.
825 %DATA The ACL arguments specified in the referencing config
826 'acl ... external' line, separated by spaces (an
827 "argument string"). see acl external.
829 If there are no ACL arguments %DATA expands to '-'.
831 If you do not specify a DATA macro inside FORMAT,
832 Squid automatically appends %DATA to your FORMAT.
834 By default, Squid applies URL-encoding to each ACL
835 argument inside the argument string. If an explicit
836 encoding modifier is used (e.g., %#DATA), then Squid
837 encodes the whole argument string as a single token
838 (e.g., with %#DATA, spaces between arguments become
841 If SSL is enabled, the following formating codes become available:
843 %USER_CERT SSL User certificate in PEM format
844 %USER_CERTCHAIN SSL User certificate chain in PEM format
845 %USER_CERT_xx SSL User certificate subject attribute xx
846 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
849 NOTE: all other format codes accepted by older Squid versions
853 General request syntax:
855 [channel-ID] FORMAT-values
858 FORMAT-values consists of transaction details expanded with
859 whitespace separation per the config file FORMAT specification
860 using the FORMAT macros listed above.
862 Request values sent to the helper are URL escaped to protect
863 each value in requests against whitespaces.
865 If using protocol=2.5 then the request sent to the helper is not
866 URL escaped to protect against whitespace.
868 NOTE: protocol=3.0 is deprecated as no longer necessary.
870 When using the concurrency= option the protocol is changed by
871 introducing a query channel tag in front of the request/response.
872 The query channel tag is a number between 0 and concurrency-1.
873 This value must be echoed back unchanged to Squid as the first part
874 of the response relating to its request.
877 The helper receives lines expanded per the above format specification
878 and for each input line returns 1 line starting with OK/ERR/BH result
879 code and optionally followed by additional keywords with more details.
882 General result syntax:
884 [channel-ID] result keyword=value ...
886 Result consists of one of the codes:
889 the ACL test produced a match.
892 the ACL test does not produce a match.
895 An internal error occurred in the helper, preventing
896 a result being identified.
898 The meaning of 'a match' is determined by your squid.conf
899 access control configuration. See the Squid wiki for details.
903 user= The users name (login)
905 password= The users password (for login= cache_peer option)
907 message= Message describing the reason for this response.
908 Available as %o in error pages.
909 Useful on (ERR and BH results).
911 tag= Apply a tag to a request. Only sets a tag once,
912 does not alter existing tags.
914 log= String to be logged in access.log. Available as
915 %ea in logformat specifications.
917 clt_conn_tag= Associates a TAG with the client TCP connection.
918 Please see url_rewrite_program related documentation
921 Any keywords may be sent on any response whether OK, ERR or BH.
923 All response keyword values need to be a single token with URL
924 escaping, or enclosed in double quotes (") and escaped using \ on
925 any double quotes or \ characters within the value. The wrapping
926 double quotes are removed before the value is interpreted by Squid.
927 \r and \n are also replace by CR and LF.
929 Some example key values:
933 user="J. \"Bob\" Smith"
940 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
941 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
942 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
943 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
944 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
947 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
948 DEFAULT: localhost src 127.0.0.1/32 ::1
949 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
950 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
952 Defining an Access List
954 Every access list definition must begin with an aclname and acltype,
955 followed by either type-specific arguments or a quoted filename that
958 acl aclname acltype argument ...
959 acl aclname acltype "file" ...
961 When using "file", the file should contain one item per line.
966 Some acl types supports options which changes their default behaviour:
968 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
969 case-insensitive, use the -i option. To return case-sensitive
970 use the +i option between patterns, or make a new ACL line
973 -n Disable lookups and address type conversions. If lookup or
974 conversion is required because the parameter type (IP or
975 domain name) does not match the message address type (domain
976 name or IP), then the ACL would immediately declare a mismatch
977 without any warnings or lookups.
980 Perform a list membership test, interpreting values as
981 comma-separated token lists and matching against individual
982 tokens instead of whole values.
983 The optional "delimiters" parameter specifies one or more
984 alternative non-alphanumeric delimiter characters.
985 non-alphanumeric delimiter characters.
987 -- Used to stop processing all options, in the case the first acl
988 value has '-' character as first character (for example the '-'
989 is a valid domain name)
991 Some acl types require suspending the current request in order
992 to access some external data source.
993 Those which do are marked with the tag [slow], those which
994 don't are marked as [fast].
995 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
996 for further information
998 ***** ACL TYPES AVAILABLE *****
1000 acl aclname src ip-address/mask ... # clients IP address [fast]
1001 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
1002 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
1003 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
1005 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
1007 # The 'arp' ACL code is not portable to all operating systems.
1008 # It works on Linux, Solaris, Windows, FreeBSD, and some other
1011 # NOTE: Squid can only determine the MAC/EUI address for IPv4
1012 # clients that are on the same subnet. If the client is on a
1013 # different subnet, then Squid cannot find out its address.
1015 # NOTE 2: IPv6 protocol does not contain ARP. MAC/EUI is either
1016 # encoded directly in the IPv6 address or not available.
1018 acl aclname srcdomain .foo.com ...
1019 # reverse lookup, from client IP [slow]
1020 acl aclname dstdomain [-n] .foo.com ...
1021 # Destination server from URL [fast]
1022 acl aclname srcdom_regex [-i] \.foo\.com ...
1023 # regex matching client name [slow]
1024 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
1025 # regex matching server [fast]
1027 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
1028 # based URL is used and no match is found. The name "none" is used
1029 # if the reverse lookup fails.
1031 acl aclname src_as number ...
1032 acl aclname dst_as number ...
1034 # Except for access control, AS numbers can be used for
1035 # routing of requests to specific caches. Here's an
1036 # example for routing all requests for AS#1241 and only
1037 # those to mycache.mydomain.net:
1038 # acl asexample dst_as 1241
1039 # cache_peer_access mycache.mydomain.net allow asexample
1040 # cache_peer_access mycache_mydomain.net deny all
1042 acl aclname peername myPeer ...
1044 # match against a named cache_peer entry
1045 # set unique name= on cache_peer lines for reliable use.
1047 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
1057 # h1:m1 must be less than h2:m2
1059 acl aclname url_regex [-i] ^http:// ...
1060 # regex matching on whole URL [fast]
1061 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
1062 # regex matching on URL login field
1063 acl aclname urlpath_regex [-i] \.gif$ ...
1064 # regex matching on URL path [fast]
1066 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
1068 acl aclname localport 3128 ... # TCP port the client connected to [fast]
1069 # NP: for interception mode this is usually '80'
1071 acl aclname myportname 3128 ... # *_port name [fast]
1073 acl aclname proto HTTP FTP ... # request protocol [fast]
1075 acl aclname method GET POST ... # HTTP request method [fast]
1077 acl aclname http_status 200 301 500- 400-403 ...
1078 # status code in reply [fast]
1080 acl aclname browser [-i] regexp ...
1081 # pattern match on User-Agent header (see also req_header below) [fast]
1083 acl aclname referer_regex [-i] regexp ...
1084 # pattern match on Referer header [fast]
1085 # Referer is highly unreliable, so use with care
1087 acl aclname ident username ...
1088 acl aclname ident_regex [-i] pattern ...
1089 # string match on ident output [slow]
1090 # use REQUIRED to accept any non-null ident.
1092 acl aclname proxy_auth [-i] username ...
1093 acl aclname proxy_auth_regex [-i] pattern ...
1094 # perform http authentication challenge to the client and match against
1095 # supplied credentials [slow]
1097 # takes a list of allowed usernames.
1098 # use REQUIRED to accept any valid username.
1100 # Will use proxy authentication in forward-proxy scenarios, and plain
1101 # http authenticaiton in reverse-proxy scenarios
1103 # NOTE: when a Proxy-Authentication header is sent but it is not
1104 # needed during ACL checking the username is NOT logged
1107 # NOTE: proxy_auth requires a EXTERNAL authentication program
1108 # to check username/password combinations (see
1109 # auth_param directive).
1111 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1112 # as the browser needs to be configured for using a proxy in order
1113 # to respond to proxy authentication.
1115 acl aclname snmp_community string ...
1116 # A community string to limit access to your SNMP Agent [fast]
1119 # acl snmppublic snmp_community public
1121 acl aclname maxconn number
1122 # This will be matched when the client's IP address has
1123 # more than <number> TCP connections established. [fast]
1124 # NOTE: This only measures direct TCP links so X-Forwarded-For
1125 # indirect clients are not counted.
1127 acl aclname max_user_ip [-s] number
1128 # This will be matched when the user attempts to log in from more
1129 # than <number> different ip addresses. The authenticate_ip_ttl
1130 # parameter controls the timeout on the ip entries. [fast]
1131 # If -s is specified the limit is strict, denying browsing
1132 # from any further IP addresses until the ttl has expired. Without
1133 # -s Squid will just annoy the user by "randomly" denying requests.
1134 # (the counter is reset each time the limit is reached and a
1135 # request is denied)
1136 # NOTE: in acceleration mode or where there is mesh of child proxies,
1137 # clients may appear to come from multiple addresses if they are
1138 # going through proxy farms, so a limit of 1 may cause user problems.
1140 acl aclname random probability
1141 # Pseudo-randomly match requests. Based on the probability given.
1142 # Probability may be written as a decimal (0.333), fraction (1/3)
1143 # or ratio of matches:non-matches (3:5).
1145 acl aclname req_mime_type [-i] mime-type ...
1146 # regex match against the mime type of the request generated
1147 # by the client. Can be used to detect file upload or some
1148 # types HTTP tunneling requests [fast]
1149 # NOTE: This does NOT match the reply. You cannot use this
1150 # to match the returned file type.
1152 acl aclname req_header header-name [-i] any\.regex\.here
1153 # regex match against any of the known request headers. May be
1154 # thought of as a superset of "browser", "referer" and "mime-type"
1157 acl aclname rep_mime_type [-i] mime-type ...
1158 # regex match against the mime type of the reply received by
1159 # squid. Can be used to detect file download or some
1160 # types HTTP tunneling requests. [fast]
1161 # NOTE: This has no effect in http_access rules. It only has
1162 # effect in rules that affect the reply data stream such as
1163 # http_reply_access.
1165 acl aclname rep_header header-name [-i] any\.regex\.here
1166 # regex match against any of the known reply headers. May be
1167 # thought of as a superset of "browser", "referer" and "mime-type"
1170 acl aclname external class_name [arguments...]
1171 # external ACL lookup via a helper class defined by the
1172 # external_acl_type directive [slow]
1174 acl aclname user_cert attribute values...
1175 # match against attributes in a user SSL certificate
1176 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1178 acl aclname ca_cert attribute values...
1179 # match against attributes a users issuing CA SSL certificate
1180 # attribute is one of DN/C/O/CN/L/ST or a numerical OID [fast]
1182 acl aclname ext_user username ...
1183 acl aclname ext_user_regex [-i] pattern ...
1184 # string match on username returned by external acl helper [slow]
1185 # use REQUIRED to accept any non-null user name.
1187 acl aclname tag tagvalue ...
1188 # string match on tag returned by external acl helper [fast]
1189 # DEPRECATED. Only the first tag will match with this ACL.
1190 # Use the 'note' ACL instead for handling multiple tag values.
1192 acl aclname hier_code codename ...
1193 # string match against squid hierarchy code(s); [fast]
1194 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1196 # NOTE: This has no effect in http_access rules. It only has
1197 # effect in rules that affect the reply data stream such as
1198 # http_reply_access.
1200 acl aclname note [-m[=delimiters]] name [value ...]
1201 # match transaction annotation [fast]
1202 # Without values, matches any annotation with a given name.
1203 # With value(s), matches any annotation with a given name that
1204 # also has one of the given values.
1205 # If the -m flag is used, then the value of the named
1206 # annotation is interpreted as a list of tokens, and the ACL
1207 # matches individual name=token pairs rather than whole
1208 # name=value pairs. See "ACL Options" above for more info.
1209 # Annotation sources include note and adaptation_meta directives
1210 # as well as helper and eCAP responses.
1212 acl aclname adaptation_service service ...
1213 # Matches the name of any icap_service, ecap_service,
1214 # adaptation_service_set, or adaptation_service_chain that Squid
1215 # has used (or attempted to use) for the master transaction.
1216 # This ACL must be defined after the corresponding adaptation
1217 # service is named in squid.conf. This ACL is usable with
1218 # adaptation_meta because it starts matching immediately after
1219 # the service has been selected for adaptation.
1222 acl aclname ssl_error errorname
1223 # match against SSL certificate validation error [fast]
1225 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1228 # The following can be used as shortcuts for certificate properties:
1229 # [ssl::]certHasExpired: the "not after" field is in the past
1230 # [ssl::]certNotYetValid: the "not before" field is in the future
1231 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1232 # [ssl::]certSelfSigned: The certificate is self signed.
1233 # [ssl::]certDomainMismatch: The certificate CN domain does not
1234 # match the name the name of the host we are connecting to.
1236 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1237 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1238 # predefined ACLs, just like the 'all' ACL.
1240 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1241 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1243 acl aclname server_cert_fingerprint [-sha1] fingerprint
1244 # match against server SSL certificate fingerprint [fast]
1246 # The fingerprint is the digest of the DER encoded version
1247 # of the whole certificate. The user should use the form: XX:XX:...
1248 # Optional argument specifies the digest algorithm to use.
1249 # The SHA1 digest algorithm is the default and is currently
1250 # the only algorithm supported (-sha1).
1252 acl aclname at_step step
1253 # match against the current step during ssl_bump evaluation [fast]
1254 # Never matches and should not be used outside the ssl_bump context.
1256 # At each SslBump step, Squid evaluates ssl_bump directives to find
1257 # the next bumping action (e.g., peek or splice). Valid SslBump step
1258 # values and the corresponding ssl_bump evaluation moments are:
1259 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1260 # SslBump2: After getting SSL Client Hello info.
1261 # SslBump3: After getting SSL Server Hello info.
1263 acl aclname ssl::server_name .foo.com ...
1264 # matches server name obtained from various sources [fast]
1266 # The server name is obtained during Ssl-Bump steps from such sources
1267 # as CONNECT request URI, client SNI, and SSL server certificate CN.
1268 # During each Ssl-Bump step, Squid may improve its understanding of a
1269 # "true server name". Unlike dstdomain, this ACL does not perform
1272 acl aclname ssl::server_name_regex [-i] \.foo\.com ...
1273 # regex matches server name obtained from various sources [fast]
1275 acl aclname connections_encrypted
1276 # matches transactions with all HTTP messages received over TLS
1277 # transport connections. [fast]
1279 # The master transaction deals with HTTP messages received from
1280 # various sources. All sources used by the master transaction in the
1281 # past are considered by the ACL. The following rules define whether
1282 # a given message source taints the entire master transaction,
1283 # resulting in ACL mismatches:
1285 # * The HTTP client transport connection is not TLS.
1286 # * An adaptation service connection-encryption flag is off.
1287 # * The peer or origin server transport connection is not TLS.
1289 # Caching currently does not affect these rules. This cache ignorance
1290 # implies that only the current HTTP client transport and REQMOD
1291 # services status determine whether this ACL matches a from-cache
1292 # transaction. The source of the cached response does not have any
1293 # effect on future transaction that use the cached response without
1294 # revalidation. This may change.
1296 # DNS, ICP, and HTCP exchanges during the master transaction do not
1297 # affect these rules.
1299 acl aclname any-of acl1 acl2 ...
1300 # match any one of the acls [fast or slow]
1301 # The first matching ACL stops further ACL evaluation.
1303 # ACLs from multiple any-of lines with the same name are ORed.
1304 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1305 # acl A any-of a1 a2
1306 # acl A any-of a3 a4
1308 # This group ACL is fast if all evaluated ACLs in the group are fast
1309 # and slow otherwise.
1311 acl aclname all-of acl1 acl2 ...
1312 # match all of the acls [fast or slow]
1313 # The first mismatching ACL stops further ACL evaluation.
1315 # ACLs from multiple all-of lines with the same name are ORed.
1316 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1317 # acl B all-of b1 b2
1318 # acl B all-of b3 b4
1320 # This group ACL is fast if all evaluated ACLs in the group are fast
1321 # and slow otherwise.
1324 acl macaddress arp 09:00:2b:23:45:67
1325 acl myexample dst_as 1241
1326 acl password proxy_auth REQUIRED
1327 acl fileupload req_mime_type -i ^multipart/form-data$
1328 acl javascript rep_mime_type -i ^application/x-javascript$
1332 # Recommended minimum configuration:
1335 # Example rule allowing access from your local networks.
1336 # Adapt to list your (internal) IP networks from where browsing
1338 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1339 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1340 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1341 acl localhet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1342 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1343 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1344 acl localnet src fc00::/7 # RFC 4193 local private network range
1345 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1347 acl SSL_ports port 443
1348 acl Safe_ports port 80 # http
1349 acl Safe_ports port 21 # ftp
1350 acl Safe_ports port 443 # https
1351 acl Safe_ports port 70 # gopher
1352 acl Safe_ports port 210 # wais
1353 acl Safe_ports port 1025-65535 # unregistered ports
1354 acl Safe_ports port 280 # http-mgmt
1355 acl Safe_ports port 488 # gss-http
1356 acl Safe_ports port 591 # filemaker
1357 acl Safe_ports port 777 # multiling http
1358 acl CONNECT method CONNECT
1362 NAME: proxy_protocol_access
1364 LOC: Config.accessList.proxyProtocol
1366 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1368 Determine which client proxies can be trusted to provide correct
1369 information regarding real client IP address using PROXY protocol.
1371 Requests may pass through a chain of several other proxies
1372 before reaching us. The original source details may by sent in:
1373 * HTTP message Forwarded header, or
1374 * HTTP message X-Forwarded-For header, or
1375 * PROXY protocol connection header.
1377 This directive is solely for validating new PROXY protocol
1378 connections received from a port flagged with require-proxy-header.
1379 It is checked only once after TCP connection setup.
1381 A deny match results in TCP connection closure.
1383 An allow match is required for Squid to permit the corresponding
1384 TCP connection, before Squid even looks for HTTP request headers.
1385 If there is an allow match, Squid starts using PROXY header information
1386 to determine the source address of the connection for all future ACL
1387 checks, logging, etc.
1389 SECURITY CONSIDERATIONS:
1391 Any host from which we accept client IP details can place
1392 incorrect information in the relevant header, and Squid
1393 will use the incorrect information as if it were the
1394 source address of the request. This may enable remote
1395 hosts to bypass any access control restrictions that are
1396 based on the client's source addresses.
1398 This clause only supports fast acl types.
1399 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1402 NAME: follow_x_forwarded_for
1404 IFDEF: FOLLOW_X_FORWARDED_FOR
1405 LOC: Config.accessList.followXFF
1406 DEFAULT_IF_NONE: deny all
1407 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1409 Determine which client proxies can be trusted to provide correct
1410 information regarding real client IP address.
1412 Requests may pass through a chain of several other proxies
1413 before reaching us. The original source details may by sent in:
1414 * HTTP message Forwarded header, or
1415 * HTTP message X-Forwarded-For header, or
1416 * PROXY protocol connection header.
1418 PROXY protocol connections are controlled by the proxy_protocol_access
1419 directive which is checked before this.
1421 If a request reaches us from a source that is allowed by this
1422 directive, then we trust the information it provides regarding
1423 the IP of the client it received from (if any).
1425 For the purpose of ACLs used in this directive the src ACL type always
1426 matches the address we are testing and srcdomain matches its rDNS.
1428 On each HTTP request Squid checks for X-Forwarded-For header fields.
1429 If found the header values are iterated in reverse order and an allow
1430 match is required for Squid to continue on to the next value.
1431 The verification ends when a value receives a deny match, cannot be
1432 tested, or there are no more values to test.
1433 NOTE: Squid does not yet follow the Forwarded HTTP header.
1435 The end result of this process is an IP address that we will
1436 refer to as the indirect client address. This address may
1437 be treated as the client address for access control, ICAP, delay
1438 pools and logging, depending on the acl_uses_indirect_client,
1439 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1440 log_uses_indirect_client and tproxy_uses_indirect_client options.
1442 This clause only supports fast acl types.
1443 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1445 SECURITY CONSIDERATIONS:
1447 Any host from which we accept client IP details can place
1448 incorrect information in the relevant header, and Squid
1449 will use the incorrect information as if it were the
1450 source address of the request. This may enable remote
1451 hosts to bypass any access control restrictions that are
1452 based on the client's source addresses.
1456 acl localhost src 127.0.0.1
1457 acl my_other_proxy srcdomain .proxy.example.com
1458 follow_x_forwarded_for allow localhost
1459 follow_x_forwarded_for allow my_other_proxy
1462 NAME: acl_uses_indirect_client
1465 IFDEF: FOLLOW_X_FORWARDED_FOR
1467 LOC: Config.onoff.acl_uses_indirect_client
1469 Controls whether the indirect client address
1470 (see follow_x_forwarded_for) is used instead of the
1471 direct client address in acl matching.
1473 NOTE: maxconn ACL considers direct TCP links and indirect
1474 clients will always have zero. So no match.
1477 NAME: delay_pool_uses_indirect_client
1480 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1482 LOC: Config.onoff.delay_pool_uses_indirect_client
1484 Controls whether the indirect client address
1485 (see follow_x_forwarded_for) is used instead of the
1486 direct client address in delay pools.
1489 NAME: log_uses_indirect_client
1492 IFDEF: FOLLOW_X_FORWARDED_FOR
1494 LOC: Config.onoff.log_uses_indirect_client
1496 Controls whether the indirect client address
1497 (see follow_x_forwarded_for) is used instead of the
1498 direct client address in the access log.
1501 NAME: tproxy_uses_indirect_client
1504 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1506 LOC: Config.onoff.tproxy_uses_indirect_client
1508 Controls whether the indirect client address
1509 (see follow_x_forwarded_for) is used instead of the
1510 direct client address when spoofing the outgoing client.
1512 This has no effect on requests arriving in non-tproxy
1515 SECURITY WARNING: Usage of this option is dangerous
1516 and should not be used trivially. Correct configuration
1517 of follow_x_forwarded_for with a limited set of trusted
1518 sources is required to prevent abuse of your proxy.
1521 NAME: spoof_client_ip
1523 LOC: Config.accessList.spoof_client_ip
1525 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1527 Control client IP address spoofing of TPROXY traffic based on
1528 defined access lists.
1530 spoof_client_ip allow|deny [!]aclname ...
1532 If there are no "spoof_client_ip" lines present, the default
1533 is to "allow" spoofing of any suitable request.
1535 Note that the cache_peer "no-tproxy" option overrides this ACL.
1537 This clause supports fast acl types.
1538 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1543 LOC: Config.accessList.http
1544 DEFAULT_IF_NONE: deny all
1545 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1547 Allowing or Denying access based on defined access lists
1549 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1550 http_access allow|deny [!]aclname ...
1552 NOTE on default values:
1554 If there are no "access" lines present, the default is to deny
1557 If none of the "access" lines cause a match, the default is the
1558 opposite of the last line in the list. If the last line was
1559 deny, the default is allow. Conversely, if the last line
1560 is allow, the default will be deny. For these reasons, it is a
1561 good idea to have an "deny all" entry at the end of your access
1562 lists to avoid potential confusion.
1564 This clause supports both fast and slow acl types.
1565 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1570 # Recommended minimum Access Permission configuration:
1572 # Deny requests to certain unsafe ports
1573 http_access deny !Safe_ports
1575 # Deny CONNECT to other than secure SSL ports
1576 http_access deny CONNECT !SSL_ports
1578 # Only allow cachemgr access from localhost
1579 http_access allow localhost manager
1580 http_access deny manager
1582 # We strongly recommend the following be uncommented to protect innocent
1583 # web applications running on the proxy server who think the only
1584 # one who can access services on "localhost" is a local user
1585 #http_access deny to_localhost
1588 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1591 # Example rule allowing access from your local networks.
1592 # Adapt localnet in the ACL section to list your (internal) IP networks
1593 # from where browsing should be allowed
1594 http_access allow localnet
1595 http_access allow localhost
1597 # And finally deny all other access to this proxy
1598 http_access deny all
1602 NAME: adapted_http_access http_access2
1604 LOC: Config.accessList.adapted_http
1606 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1608 Allowing or Denying access based on defined access lists
1610 Essentially identical to http_access, but runs after redirectors
1611 and ICAP/eCAP adaptation. Allowing access control based on their
1614 If not set then only http_access is used.
1617 NAME: http_reply_access
1619 LOC: Config.accessList.reply
1621 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1623 Allow replies to client requests. This is complementary to http_access.
1625 http_reply_access allow|deny [!] aclname ...
1627 NOTE: if there are no access lines present, the default is to allow
1630 If none of the access lines cause a match the opposite of the
1631 last line will apply. Thus it is good practice to end the rules
1632 with an "allow all" or "deny all" entry.
1634 This clause supports both fast and slow acl types.
1635 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1640 LOC: Config.accessList.icp
1642 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1644 Allowing or Denying access to the ICP port based on defined
1647 icp_access allow|deny [!]aclname ...
1649 NOTE: The default if no icp_access lines are present is to
1650 deny all traffic. This default may cause problems with peers
1653 This clause only supports fast acl types.
1654 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1656 # Allow ICP queries from local networks only
1657 #icp_access allow localnet
1658 #icp_access deny all
1664 LOC: Config.accessList.htcp
1666 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1668 Allowing or Denying access to the HTCP port based on defined
1671 htcp_access allow|deny [!]aclname ...
1673 See also htcp_clr_access for details on access control for
1674 cache purge (CLR) HTCP messages.
1676 NOTE: The default if no htcp_access lines are present is to
1677 deny all traffic. This default may cause problems with peers
1678 using the htcp option.
1680 This clause only supports fast acl types.
1681 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1683 # Allow HTCP queries from local networks only
1684 #htcp_access allow localnet
1685 #htcp_access deny all
1688 NAME: htcp_clr_access
1691 LOC: Config.accessList.htcp_clr
1693 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1695 Allowing or Denying access to purge content using HTCP based
1696 on defined access lists.
1697 See htcp_access for details on general HTCP access control.
1699 htcp_clr_access allow|deny [!]aclname ...
1701 This clause only supports fast acl types.
1702 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1704 # Allow HTCP CLR requests from trusted peers
1705 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1706 htcp_clr_access allow htcp_clr_peer
1707 htcp_clr_access deny all
1712 LOC: Config.accessList.miss
1714 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1716 Determines whether network access is permitted when satisfying a request.
1719 to force your neighbors to use you as a sibling instead of
1722 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1723 miss_access deny !localclients
1724 miss_access allow all
1726 This means only your local clients are allowed to fetch relayed/MISS
1727 replies from the network and all other clients can only fetch cached
1730 The default for this setting allows all clients who passed the
1731 http_access rules to relay via this proxy.
1733 This clause only supports fast acl types.
1734 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1737 NAME: ident_lookup_access
1741 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1742 LOC: Ident::TheConfig.identLookup
1744 A list of ACL elements which, if matched, cause an ident
1745 (RFC 931) lookup to be performed for this request. For
1746 example, you might choose to always perform ident lookups
1747 for your main multi-user Unix boxes, but not for your Macs
1748 and PCs. By default, ident lookups are not performed for
1751 To enable ident lookups for specific client addresses, you
1752 can follow this example:
1754 acl ident_aware_hosts src 198.168.1.0/24
1755 ident_lookup_access allow ident_aware_hosts
1756 ident_lookup_access deny all
1758 Only src type ACL checks are fully supported. A srcdomain
1759 ACL might work at times, but it will not always provide
1762 This clause only supports fast acl types.
1763 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1766 NAME: reply_body_max_size
1767 COMMENT: size [acl acl...]
1770 DEFAULT_DOC: No limit is applied.
1771 LOC: Config.ReplyBodySize
1773 This option specifies the maximum size of a reply body. It can be
1774 used to prevent users from downloading very large files, such as
1775 MP3's and movies. When the reply headers are received, the
1776 reply_body_max_size lines are processed, and the first line where
1777 all (if any) listed ACLs are true is used as the maximum body size
1780 This size is checked twice. First when we get the reply headers,
1781 we check the content-length value. If the content length value exists
1782 and is larger than the allowed size, the request is denied and the
1783 user receives an error message that says "the request or reply
1784 is too large." If there is no content-length, and the reply
1785 size exceeds this limit, the client's connection is just closed
1786 and they will receive a partial reply.
1788 WARNING: downstream caches probably can not detect a partial reply
1789 if there is no content-length header, so they will cache
1790 partial responses and give them out as hits. You should NOT
1791 use this option if you have downstream caches.
1793 WARNING: A maximum size smaller than the size of squid's error messages
1794 will cause an infinite loop and crash squid. Ensure that the smallest
1795 non-zero value you use is greater that the maximum header size plus
1796 the size of your largest error page.
1798 If you set this parameter none (the default), there will be
1801 Configuration Format is:
1802 reply_body_max_size SIZE UNITS [acl ...]
1804 reply_body_max_size 10 MB
1808 NAME: on_unsupported_protocol
1809 TYPE: on_unsupported_protocol
1810 LOC: Config.accessList.on_unsupported_protocol
1812 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
1814 Determines Squid behavior when encountering strange requests at the
1815 beginning of an accepted TCP connection or the beginning of a bumped
1816 CONNECT tunnel. Controlling Squid reaction to unexpected traffic is
1817 especially useful in interception environments where Squid is likely
1818 to see connections for unsupported protocols that Squid should either
1819 terminate or tunnel at TCP level.
1821 on_unsupported_protocol <action> [!]acl ...
1823 The first matching action wins. Only fast ACLs are supported.
1825 Supported actions are:
1827 tunnel: Establish a TCP connection with the intended server and
1828 blindly shovel TCP packets between the client and server.
1830 respond: Respond with an error message, using the transfer protocol
1831 for the Squid port that received the request (e.g., HTTP
1832 for connections intercepted at the http_port). This is the
1835 Squid expects the following traffic patterns:
1837 http_port: a plain HTTP request
1838 https_port: SSL/TLS handshake followed by an [encrypted] HTTP request
1839 ftp_port: a plain FTP command (no on_unsupported_protocol support yet!)
1840 CONNECT tunnel on http_port: same as https_port
1841 CONNECT tunnel on https_port: same as https_port
1843 Currently, this directive has effect on intercepted connections and
1844 bumped tunnels only. Other cases are not supported because Squid
1845 cannot know the intended destination of other traffic.
1848 # define what Squid errors indicate receiving non-HTTP traffic:
1849 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
1850 # define what Squid errors indicate receiving nothing:
1851 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
1852 # tunnel everything that does not look like HTTP:
1853 on_unsupported_protocol tunnel foreignProtocol
1854 # tunnel if we think the client waits for the server to talk first:
1855 on_unsupported_protocol tunnel serverTalksFirstProtocol
1856 # in all other error cases, just send an HTTP "error page" response:
1857 on_unsupported_protocol respond all
1859 See also: squid_error ACL
1864 -----------------------------------------------------------------------------
1867 NAME: http_port ascii_port
1872 Usage: port [mode] [options]
1873 hostname:port [mode] [options]
1874 1.2.3.4:port [mode] [options]
1876 The socket addresses where Squid will listen for HTTP client
1877 requests. You may specify multiple socket addresses.
1878 There are three forms: port alone, hostname with port, and
1879 IP address with port. If you specify a hostname or IP
1880 address, Squid binds the socket to that specific
1881 address. Most likely, you do not need to bind to a specific
1882 address, so you can use the port number alone.
1884 If you are running Squid in accelerator mode, you
1885 probably want to listen on port 80 also, or instead.
1887 The -a command line option may be used to specify additional
1888 port(s) where Squid listens for proxy request. Such ports will
1889 be plain proxy ports with no options.
1891 You may specify multiple socket addresses on multiple lines.
1895 intercept Support for IP-Layer NAT interception delivering
1896 traffic to this Squid port.
1897 NP: disables authentication on the port.
1899 tproxy Support Linux TPROXY (or BSD divert-to) with spoofing
1900 of outgoing connections using the client IP address.
1901 NP: disables authentication on the port.
1903 accel Accelerator / reverse proxy mode
1905 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1906 establish secure connection with the client and with
1907 the server, decrypt HTTPS messages as they pass through
1908 Squid, and treat them as unencrypted HTTP messages,
1909 becoming the man-in-the-middle.
1911 The ssl_bump option is required to fully enable
1912 bumping of CONNECT requests.
1914 Omitting the mode flag causes default forward proxy mode to be used.
1917 Accelerator Mode Options:
1919 defaultsite=domainname
1920 What to use for the Host: header if it is not present
1921 in a request. Determines what site (not origin server)
1922 accelerators should consider the default.
1924 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1926 protocol= Protocol to reconstruct accelerated and intercepted
1927 requests with. Defaults to HTTP/1.1 for http_port and
1928 HTTPS/1.1 for https_port.
1929 When an unsupported value is configured Squid will
1930 produce a FATAL error.
1931 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1933 vport Virtual host port support. Using the http_port number
1934 instead of the port passed on Host: headers.
1936 vport=NN Virtual host port support. Using the specified port
1937 number instead of the port passed on Host: headers.
1940 Act as if this Squid is the origin server.
1941 This currently means generate new Date: and Expires:
1942 headers on HIT instead of adding Age:.
1944 ignore-cc Ignore request Cache-Control headers.
1946 WARNING: This option violates HTTP specifications if
1947 used in non-accelerator setups.
1949 allow-direct Allow direct forwarding in accelerator mode. Normally
1950 accelerated requests are denied direct forwarding as if
1951 never_direct was used.
1953 WARNING: this option opens accelerator mode to security
1954 vulnerabilities usually only affecting in interception
1955 mode. Make sure to protect forwarding with suitable
1956 http_access rules when using this.
1959 SSL Bump Mode Options:
1960 In addition to these options ssl-bump requires TLS/SSL options.
1962 generate-host-certificates[=<on|off>]
1963 Dynamically create SSL server certificates for the
1964 destination hosts of bumped CONNECT requests.When
1965 enabled, the cert and key options are used to sign
1966 generated certificates. Otherwise generated
1967 certificate will be selfsigned.
1968 If there is a CA certificate lifetime of the generated
1969 certificate equals lifetime of the CA certificate. If
1970 generated certificate is selfsigned lifetime is three
1972 This option is enabled by default when ssl-bump is used.
1973 See the ssl-bump option above for more information.
1975 dynamic_cert_mem_cache_size=SIZE
1976 Approximate total RAM size spent on cached generated
1977 certificates. If set to zero, caching is disabled. The
1978 default value is 4MB.
1982 cert= Path to SSL certificate (PEM format).
1984 key= Path to SSL private key file (PEM format)
1985 if not specified, the certificate file is
1986 assumed to be a combined certificate and
1989 cipher= Colon separated list of supported ciphers.
1990 NOTE: some ciphers such as EDH ciphers depend on
1991 additional settings. If those settings are
1992 omitted the ciphers may be silently ignored
1993 by the OpenSSL library.
1995 options= Various SSL implementation options. The most important
1998 NO_SSLv3 Disallow the use of SSLv3
2000 NO_TLSv1 Disallow the use of TLSv1.0
2002 NO_TLSv1_1 Disallow the use of TLSv1.1
2004 NO_TLSv1_2 Disallow the use of TLSv1.2
2007 Always create a new key when using
2008 temporary/ephemeral DH key exchanges
2011 Enable ephemeral ECDH key exchange.
2012 The adopted curve should be specified
2013 using the tls-dh option.
2016 Disable use of RFC5077 session tickets.
2017 Some servers may have problems
2018 understanding the TLS extension due
2019 to ambiguous specification in RFC4507.
2021 ALL Enable various bug workarounds
2022 suggested as "harmless" by OpenSSL
2023 Be warned that this reduces SSL/TLS
2024 strength to some attacks.
2026 See the OpenSSL SSL_CTX_set_options documentation for a
2029 clientca= File containing the list of CAs to use when
2030 requesting a client certificate.
2032 tls-cafile= PEM file containing CA certificates to use when verifying
2033 client certificates. If not configured clientca will be
2034 used. May be repeated to load multiple files.
2036 capath= Directory containing additional CA certificates
2037 and CRL lists to use when verifying client certificates.
2038 Requires OpenSSL or LibreSSL.
2040 crlfile= File of additional CRL lists to use when verifying
2041 the client certificate, in addition to CRLs stored in
2042 the capath. Implies VERIFY_CRL flag below.
2045 File containing DH parameters for temporary/ephemeral DH key
2046 exchanges, optionally prefixed by a curve for ephemeral ECDH
2048 See OpenSSL documentation for details on how to create the
2049 DH parameter file. Supported curves for ECDH can be listed
2050 using the "openssl ecparam -list_curves" command.
2051 WARNING: EDH and EECDH ciphers will be silently disabled if
2052 this option is not set.
2054 sslflags= Various flags modifying the use of SSL:
2056 Don't request client certificates
2057 immediately, but wait until acl processing
2058 requires a certificate (not yet implemented).
2060 Don't allow for session reuse. Each connection
2061 will result in a new SSL session.
2063 Verify CRL lists when accepting client
2066 Verify CRL lists for all certificates in the
2067 client certificate chain.
2069 tls-default-ca[=off]
2070 Whether to use the system Trusted CAs. Default is OFF.
2072 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
2074 sslcontext= SSL session ID context identifier.
2078 connection-auth[=on|off]
2079 use connection-auth=off to tell Squid to prevent
2080 forwarding Microsoft connection oriented authentication
2081 (NTLM, Negotiate and Kerberos)
2083 disable-pmtu-discovery=
2084 Control Path-MTU discovery usage:
2085 off lets OS decide on what to do (default).
2086 transparent disable PMTU discovery when transparent
2088 always disable always PMTU discovery.
2090 In many setups of transparently intercepting proxies
2091 Path-MTU discovery can not work on traffic towards the
2092 clients. This is the case when the intercepting device
2093 does not fully track connections and fails to forward
2094 ICMP must fragment messages to the cache server. If you
2095 have such setup and experience that certain clients
2096 sporadically hang or never complete requests set
2097 disable-pmtu-discovery option to 'transparent'.
2099 name= Specifies a internal name for the port. Defaults to
2100 the port specification (port or addr:port)
2102 tcpkeepalive[=idle,interval,timeout]
2103 Enable TCP keepalive probes of idle connections.
2104 In seconds; idle is the initial time before TCP starts
2105 probing the connection, interval how often to probe, and
2106 timeout the time before giving up.
2108 require-proxy-header
2109 Require PROXY protocol version 1 or 2 connections.
2110 The proxy_protocol_access is required to whitelist
2111 downstream proxies which can be trusted.
2113 If you run Squid on a dual-homed machine with an internal
2114 and an external interface we recommend you to specify the
2115 internal address:port in http_port. This way Squid will only be
2116 visible on the internal address.
2120 # Squid normally listens to port 3128
2121 http_port @DEFAULT_HTTP_PORT@
2126 IFDEF: USE_GNUTLS||USE_OPENSSL
2131 Usage: [ip:]port [mode] cert=certificate.pem [options]
2133 The socket address where Squid will listen for client requests made
2134 over TLS or SSL connections. Commonly referred to as HTTPS.
2136 This is most useful for situations where you are running squid in
2137 accelerator mode and you want to do the TLS work at the accelerator level.
2139 You may specify multiple socket addresses on multiple lines,
2140 each with their own certificate and/or options.
2142 The TLS cert= option is mandatory on HTTPS ports.
2144 See http_port for a list of modes and options.
2152 Enables Native FTP proxy by specifying the socket address where Squid
2153 listens for FTP client requests. See http_port directive for various
2154 ways to specify the listening address and mode.
2156 Usage: ftp_port address [mode] [options]
2158 WARNING: This is a new, experimental, complex feature that has seen
2159 limited production exposure. Some Squid modules (e.g., caching) do not
2160 currently work with native FTP proxying, and many features have not
2161 even been tested for compatibility. Test well before deploying!
2163 Native FTP proxying differs substantially from proxying HTTP requests
2164 with ftp:// URIs because Squid works as an FTP server and receives
2165 actual FTP commands (rather than HTTP requests with FTP URLs).
2167 Native FTP commands accepted at ftp_port are internally converted or
2168 wrapped into HTTP-like messages. The same happens to Native FTP
2169 responses received from FTP origin servers. Those HTTP-like messages
2170 are shoveled through regular access control and adaptation layers
2171 between the FTP client and the FTP origin server. This allows Squid to
2172 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2173 mechanisms when shoveling wrapped FTP messages. For example,
2174 http_access and adaptation_access directives are used.
2178 intercept Same as http_port intercept. The FTP origin address is
2179 determined based on the intended destination of the
2180 intercepted connection.
2182 tproxy Support Linux TPROXY for spoofing outgoing
2183 connections using the client IP address.
2184 NP: disables authentication and maybe IPv6 on the port.
2186 By default (i.e., without an explicit mode option), Squid extracts the
2187 FTP origin address from the login@origin parameter of the FTP USER
2188 command. Many popular FTP clients support such native FTP proxying.
2192 name=token Specifies an internal name for the port. Defaults to
2193 the port address. Usable with myportname ACL.
2196 Enables tracking of FTP directories by injecting extra
2197 PWD commands and adjusting Request-URI (in wrapping
2198 HTTP requests) to reflect the current FTP server
2199 directory. Tracking is disabled by default.
2201 protocol=FTP Protocol to reconstruct accelerated and intercepted
2202 requests with. Defaults to FTP. No other accepted
2203 values have been tested with. An unsupported value
2204 results in a FATAL error. Accepted values are FTP,
2205 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2207 Other http_port modes and options that are not specific to HTTP and
2208 HTTPS may also work.
2211 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2214 LOC: Ip::Qos::TheConfig.tosToServer
2216 Allows you to select a TOS/Diffserv value for packets outgoing
2217 on the server side, based on an ACL.
2219 tcp_outgoing_tos ds-field [!]aclname ...
2221 Example where normal_service_net uses the TOS value 0x00
2222 and good_service_net uses 0x20
2224 acl normal_service_net src 10.0.0.0/24
2225 acl good_service_net src 10.0.1.0/24
2226 tcp_outgoing_tos 0x00 normal_service_net
2227 tcp_outgoing_tos 0x20 good_service_net
2229 TOS/DSCP values really only have local significance - so you should
2230 know what you're specifying. For more information, see RFC2474,
2231 RFC2475, and RFC3260.
2233 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2234 "default" to use whatever default your host has.
2235 Note that only multiples of 4 are usable as the two rightmost bits have
2236 been redefined for use by ECN (RFC 3168 section 23.1).
2237 The squid parser will enforce this by masking away the ECN bits.
2239 Processing proceeds in the order specified, and stops at first fully
2242 Only fast ACLs are supported.
2245 NAME: clientside_tos
2248 LOC: Ip::Qos::TheConfig.tosToClient
2250 Allows you to select a TOS/DSCP value for packets being transmitted
2251 on the client-side, based on an ACL.
2253 clientside_tos ds-field [!]aclname ...
2255 Example where normal_service_net uses the TOS value 0x00
2256 and good_service_net uses 0x20
2258 acl normal_service_net src 10.0.0.0/24
2259 acl good_service_net src 10.0.1.0/24
2260 clientside_tos 0x00 normal_service_net
2261 clientside_tos 0x20 good_service_net
2263 Note: This feature is incompatible with qos_flows. Any TOS values set here
2264 will be overwritten by TOS values in qos_flows.
2266 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2267 "default" to use whatever default your host has.
2268 Note that only multiples of 4 are usable as the two rightmost bits have
2269 been redefined for use by ECN (RFC 3168 section 23.1).
2270 The squid parser will enforce this by masking away the ECN bits.
2274 NAME: tcp_outgoing_mark
2276 IFDEF: SO_MARK&&USE_LIBCAP
2278 LOC: Ip::Qos::TheConfig.nfmarkToServer
2280 Allows you to apply a Netfilter mark value to outgoing packets
2281 on the server side, based on an ACL.
2283 tcp_outgoing_mark mark-value [!]aclname ...
2285 Example where normal_service_net uses the mark value 0x00
2286 and good_service_net uses 0x20
2288 acl normal_service_net src 10.0.0.0/24
2289 acl good_service_net src 10.0.1.0/24
2290 tcp_outgoing_mark 0x00 normal_service_net
2291 tcp_outgoing_mark 0x20 good_service_net
2293 Only fast ACLs are supported.
2296 NAME: clientside_mark
2298 IFDEF: SO_MARK&&USE_LIBCAP
2300 LOC: Ip::Qos::TheConfig.nfmarkToClient
2302 Allows you to apply a Netfilter mark value to packets being transmitted
2303 on the client-side, based on an ACL.
2305 clientside_mark mark-value [!]aclname ...
2307 Example where normal_service_net uses the mark value 0x00
2308 and good_service_net uses 0x20
2310 acl normal_service_net src 10.0.0.0/24
2311 acl good_service_net src 10.0.1.0/24
2312 clientside_mark 0x00 normal_service_net
2313 clientside_mark 0x20 good_service_net
2315 Note: This feature is incompatible with qos_flows. Any mark values set here
2316 will be overwritten by mark values in qos_flows.
2323 LOC: Ip::Qos::TheConfig
2325 Allows you to select a TOS/DSCP value to mark outgoing
2326 connections to the client, based on where the reply was sourced.
2327 For platforms using netfilter, allows you to set a netfilter mark
2328 value instead of, or in addition to, a TOS value.
2330 By default this functionality is disabled. To enable it with the default
2331 settings simply use "qos_flows mark" or "qos_flows tos". Default
2332 settings will result in the netfilter mark or TOS value being copied
2333 from the upstream connection to the client. Note that it is the connection
2334 CONNMARK value not the packet MARK value that is copied.
2336 It is not currently possible to copy the mark or TOS value from the
2337 client to the upstream connection request.
2339 TOS values really only have local significance - so you should
2340 know what you're specifying. For more information, see RFC2474,
2341 RFC2475, and RFC3260.
2343 The TOS/DSCP byte must be exactly that - a octet value 0 - 255.
2344 Note that only multiples of 4 are usable as the two rightmost bits have
2345 been redefined for use by ECN (RFC 3168 section 23.1).
2346 The squid parser will enforce this by masking away the ECN bits.
2348 Mark values can be any unsigned 32-bit integer value.
2350 This setting is configured by setting the following values:
2352 tos|mark Whether to set TOS or netfilter mark values
2354 local-hit=0xFF Value to mark local cache hits.
2356 sibling-hit=0xFF Value to mark hits from sibling peers.
2358 parent-hit=0xFF Value to mark hits from parent peers.
2360 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2361 over the preserve-miss feature (see below), unless
2362 mask is specified, in which case only the bits
2363 specified in the mask are written.
2365 The TOS variant of the following features are only possible on Linux
2366 and require your kernel to be patched with the TOS preserving ZPH
2367 patch, available from http://zph.bratcheda.org
2368 No patch is needed to preserve the netfilter mark, which will work
2369 with all variants of netfilter.
2371 disable-preserve-miss
2372 This option disables the preservation of the TOS or netfilter
2373 mark. By default, the existing TOS or netfilter mark value of
2374 the response coming from the remote server will be retained
2375 and masked with miss-mark.
2376 NOTE: in the case of a netfilter mark, the mark must be set on
2377 the connection (using the CONNMARK target) not on the packet
2381 Allows you to mask certain bits in the TOS or mark value
2382 received from the remote server, before copying the value to
2383 the TOS sent towards clients.
2384 Default for tos: 0xFF (TOS from server is not changed).
2385 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2387 All of these features require the --enable-zph-qos compilation flag
2388 (enabled by default). Netfilter marking also requires the
2389 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2390 libcap 2.09+ (--with-libcap).
2394 NAME: tcp_outgoing_address
2397 DEFAULT_DOC: Address selection is performed by the operating system.
2398 LOC: Config.accessList.outgoing_address
2400 Allows you to map requests to different outgoing IP addresses
2401 based on the username or source address of the user making
2404 tcp_outgoing_address ipaddr [[!]aclname] ...
2407 Forwarding clients with dedicated IPs for certain subnets.
2409 acl normal_service_net src 10.0.0.0/24
2410 acl good_service_net src 10.0.2.0/24
2412 tcp_outgoing_address 2001:db8::c001 good_service_net
2413 tcp_outgoing_address 10.1.0.2 good_service_net
2415 tcp_outgoing_address 2001:db8::beef normal_service_net
2416 tcp_outgoing_address 10.1.0.1 normal_service_net
2418 tcp_outgoing_address 2001:db8::1
2419 tcp_outgoing_address 10.1.0.3
2421 Processing proceeds in the order specified, and stops at first fully
2424 Squid will add an implicit IP version test to each line.
2425 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2426 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2429 NOTE: The use of this directive using client dependent ACLs is
2430 incompatible with the use of server side persistent connections. To
2431 ensure correct results it is best to set server_persistent_connections
2432 to off when using this directive in such configurations.
2434 NOTE: The use of this directive to set a local IP on outgoing TCP links
2435 is incompatible with using TPROXY to set client IP out outbound TCP links.
2436 When needing to contact peers use the no-tproxy cache_peer option and the
2437 client_dst_passthru directive re-enable normal forwarding such as this.
2441 NAME: host_verify_strict
2444 LOC: Config.onoff.hostStrictVerify
2446 Regardless of this option setting, when dealing with intercepted
2447 traffic, Squid always verifies that the destination IP address matches
2448 the Host header domain or IP (called 'authority form URL').
2450 This enforcement is performed to satisfy a MUST-level requirement in
2451 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2452 authority of the origin server or gateway given by the original URL".
2455 Squid always responds with an HTTP 409 (Conflict) error
2456 page and logs a security warning if there is no match.
2458 Squid verifies that the destination IP address matches
2459 the Host header for forward-proxy and reverse-proxy traffic
2460 as well. For those traffic types, Squid also enables the
2461 following checks, comparing the corresponding Host header
2462 and Request-URI components:
2464 * The host names (domain or IP) must be identical,
2465 but valueless or missing Host header disables all checks.
2466 For the two host names to match, both must be either IP
2469 * Port numbers must be identical, but if a port is missing
2470 the scheme-default port is assumed.
2473 When set to OFF (the default):
2474 Squid allows suspicious requests to continue but logs a
2475 security warning and blocks caching of the response.
2477 * Forward-proxy traffic is not checked at all.
2479 * Reverse-proxy traffic is not checked at all.
2481 * Intercepted traffic which passes verification is handled
2482 according to client_dst_passthru.
2484 * Intercepted requests which fail verification are sent
2485 to the client original destination instead of DIRECT.
2486 This overrides 'client_dst_passthru off'.
2488 For now suspicious intercepted CONNECT requests are always
2489 responded to with an HTTP 409 (Conflict) error page.
2494 As described in CVE-2009-0801 when the Host: header alone is used
2495 to determine the destination of a request it becomes trivial for
2496 malicious scripts on remote websites to bypass browser same-origin
2497 security policy and sandboxing protections.
2499 The cause of this is that such applets are allowed to perform their
2500 own HTTP stack, in which case the same-origin policy of the browser
2501 sandbox only verifies that the applet tries to contact the same IP
2502 as from where it was loaded at the IP level. The Host: header may
2503 be different from the connected IP and approved origin.
2507 NAME: client_dst_passthru
2510 LOC: Config.onoff.client_dst_passthru
2512 With NAT or TPROXY intercepted traffic Squid may pass the request
2513 directly to the original client destination IP or seek a faster
2514 source using the HTTP Host header.
2516 Using Host to locate alternative servers can provide faster
2517 connectivity with a range of failure recovery options.
2518 But can also lead to connectivity trouble when the client and
2519 server are attempting stateful interactions unaware of the proxy.
2521 This option (on by default) prevents alternative DNS entries being
2522 located to send intercepted traffic DIRECT to an origin server.
2523 The clients original destination IP and port will be used instead.
2525 Regardless of this option setting, when dealing with intercepted
2526 traffic Squid will verify the Host: header and any traffic which
2527 fails Host verification will be treated as if this option were ON.
2529 see host_verify_strict for details on the verification process.
2534 -----------------------------------------------------------------------------
2537 NAME: tls_outgoing_options
2538 IFDEF: USE_GNUTLS||USE_OPENSSL
2539 TYPE: securePeerOptions
2540 DEFAULT: min-version=1.0
2541 LOC: Security::ProxyOutgoingConfig
2543 disable Do not support https:// URLs.
2545 cert=/path/to/client/certificate
2546 A client TLS certificate to use when connecting.
2548 key=/path/to/client/private_key
2549 The private TLS key corresponding to the cert= above.
2550 If key= is not specified cert= is assumed to reference
2551 a PEM file containing both the certificate and the key.
2553 cipher=... The list of valid TLS ciphers to use.
2556 The minimum TLS protocol version to permit.
2557 To control SSLv3 use the options= parameter.
2558 Supported Values: 1.0 (default), 1.1, 1.2
2560 options=... Specify various TLS/SSL implementation options:
2562 NO_SSLv3 Disallow the use of SSLv3
2564 NO_TLSv1 Disallow the use of TLSv1.0
2566 NO_TLSv1_1 Disallow the use of TLSv1.1
2568 NO_TLSv1_2 Disallow the use of TLSv1.2
2571 Always create a new key when using
2572 temporary/ephemeral DH key exchanges
2575 Disable use of RFC5077 session tickets.
2576 Some servers may have problems
2577 understanding the TLS extension due
2578 to ambiguous specification in RFC4507.
2580 ALL Enable various bug workarounds
2581 suggested as "harmless" by OpenSSL
2582 Be warned that this reduces SSL/TLS
2583 strength to some attacks.
2585 See the OpenSSL SSL_CTX_set_options documentation for a
2588 cafile= PEM file containing CA certificates to use when verifying
2589 the peer certificate. May be repeated to load multiple files.
2591 capath= A directory containing additional CA certificates to
2592 use when verifying the peer certificate.
2593 Requires OpenSSL or LibreSSL.
2595 crlfile=... A certificate revocation list file to use when
2596 verifying the peer certificate.
2598 flags=... Specify various flags modifying the TLS implementation:
2601 Accept certificates even if they fail to
2604 Don't verify the peer certificate
2605 matches the server name
2608 Whether to use the system Trusted CAs. Default is ON.
2610 domain= The peer name as advertised in its certificate.
2611 Used for verifying the correctness of the received peer
2612 certificate. If not specified the peer hostname will be
2618 -----------------------------------------------------------------------------
2621 NAME: ssl_unclean_shutdown
2625 LOC: Config.SSL.unclean_shutdown
2627 Some browsers (especially MSIE) bugs out on SSL shutdown
2634 LOC: Config.SSL.ssl_engine
2637 The OpenSSL engine to use. You will need to set this if you
2638 would like to use hardware SSL acceleration for example.
2641 NAME: sslproxy_session_ttl
2644 LOC: Config.SSL.session_ttl
2647 Sets the timeout value for SSL sessions
2650 NAME: sslproxy_session_cache_size
2653 LOC: Config.SSL.sessionCacheSize
2656 Sets the cache size to use for ssl session
2659 NAME: sslproxy_foreign_intermediate_certs
2662 LOC: Config.ssl_client.foreignIntermediateCertsPath
2665 Many origin servers fail to send their full server certificate
2666 chain for verification, assuming the client already has or can
2667 easily locate any missing intermediate certificates.
2669 Squid uses the certificates from the specified file to fill in
2670 these missing chains when trying to validate origin server
2673 The file is expected to contain zero or more PEM-encoded
2674 intermediate certificates. These certificates are not treated
2675 as trusted root certificates, and any self-signed certificate in
2676 this file will be ignored.
2679 NAME: sslproxy_cert_sign_hash
2682 LOC: Config.SSL.certSignHash
2685 Sets the hashing algorithm to use when signing generated certificates.
2686 Valid algorithm names depend on the OpenSSL library used. The following
2687 names are usually available: sha1, sha256, sha512, and md5. Please see
2688 your OpenSSL library manual for the available hashes. By default, Squids
2689 that support this option use sha256 hashes.
2691 Squid does not forcefully purge cached certificates that were generated
2692 with an algorithm other than the currently configured one. They remain
2693 in the cache, subject to the regular cache eviction policy, and become
2694 useful if the algorithm changes again.
2699 TYPE: sslproxy_ssl_bump
2700 LOC: Config.accessList.ssl_bump
2701 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
2704 This option is consulted when a CONNECT request is received on
2705 an http_port (or a new connection is intercepted at an
2706 https_port), provided that port was configured with an ssl-bump
2707 flag. The subsequent data on the connection is either treated as
2708 HTTPS and decrypted OR tunneled at TCP level without decryption,
2709 depending on the first matching bumping "action".
2711 ssl_bump <action> [!]acl ...
2713 The following bumping actions are currently supported:
2716 Become a TCP tunnel without decrypting proxied traffic.
2717 This is the default action.
2720 Establish a secure connection with the server and, using a
2721 mimicked server certificate, with the client.
2724 Receive client (step SslBump1) or server (step SslBump2)
2725 certificate while preserving the possibility of splicing the
2726 connection. Peeking at the server certificate (during step 2)
2727 usually precludes bumping of the connection at step 3.
2730 Receive client (step SslBump1) or server (step SslBump2)
2731 certificate while preserving the possibility of bumping the
2732 connection. Staring at the server certificate (during step 2)
2733 usually precludes splicing of the connection at step 3.
2736 Close client and server connections.
2738 Backward compatibility actions available at step SslBump1:
2741 Bump the connection. Establish a secure connection with the
2742 client first, then connect to the server. This old mode does
2743 not allow Squid to mimic server SSL certificate and does not
2744 work with intercepted SSL connections.
2747 Bump the connection. Establish a secure connection with the
2748 server first, then establish a secure connection with the
2749 client, using a mimicked server certificate. Works with both
2750 CONNECT requests and intercepted SSL connections, but does
2751 not allow to make decisions based on SSL handshake info.
2754 Decide whether to bump or splice the connection based on
2755 client-to-squid and server-to-squid SSL hello messages.
2759 Same as the "splice" action.
2761 All ssl_bump rules are evaluated at each of the supported bumping
2762 steps. Rules with actions that are impossible at the current step are
2763 ignored. The first matching ssl_bump action wins and is applied at the
2764 end of the current step. If no rules match, the splice action is used.
2765 See the at_step ACL for a list of the supported SslBump steps.
2767 This clause supports both fast and slow acl types.
2768 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2770 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
2773 # Example: Bump all TLS connections except those originating from
2774 # localhost or those going to example.com.
2776 acl broken_sites ssl::server_name .example.com
2777 ssl_bump splice localhost
2778 ssl_bump splice broken_sites
2782 NAME: sslproxy_cert_error
2785 DEFAULT_DOC: Server certificate errors terminate the transaction.
2786 LOC: Config.ssl_client.cert_error
2789 Use this ACL to bypass server certificate validation errors.
2791 For example, the following lines will bypass all validation errors
2792 when talking to servers for example.com. All other
2793 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2795 acl BrokenButTrustedServers dstdomain example.com
2796 sslproxy_cert_error allow BrokenButTrustedServers
2797 sslproxy_cert_error deny all
2799 This clause only supports fast acl types.
2800 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2801 Using slow acl types may result in server crashes
2803 Without this option, all server certificate validation errors
2804 terminate the transaction to protect Squid and the client.
2806 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2807 but should not happen unless your OpenSSL library is buggy.
2810 Bypassing validation errors is dangerous because an
2811 error usually implies that the server cannot be trusted
2812 and the connection may be insecure.
2814 See also: sslproxy_flags and DONT_VERIFY_PEER.
2817 NAME: sslproxy_cert_sign
2820 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2821 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2822 POSTSCRIPTUM: signTrusted all
2823 TYPE: sslproxy_cert_sign
2824 LOC: Config.ssl_client.cert_sign
2827 sslproxy_cert_sign <signing algorithm> acl ...
2829 The following certificate signing algorithms are supported:
2832 Sign using the configured CA certificate which is usually
2833 placed in and trusted by end-user browsers. This is the
2834 default for trusted origin server certificates.
2837 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2838 This is the default for untrusted origin server certificates
2839 that are not self-signed (see ssl::certUntrusted).
2842 Sign using a self-signed certificate with the right CN to
2843 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2844 browser. This is the default for self-signed origin server
2845 certificates (see ssl::certSelfSigned).
2847 This clause only supports fast acl types.
2849 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2850 signing algorithm to generate the certificate and ignores all
2851 subsequent sslproxy_cert_sign options (the first match wins). If no
2852 acl(s) match, the default signing algorithm is determined by errors
2853 detected when obtaining and validating the origin server certificate.
2855 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2856 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2857 CONNECT request that carries a domain name. In all other cases (CONNECT
2858 to an IP address or an intercepted SSL connection), Squid cannot detect
2859 the domain mismatch at certificate generation time when
2860 bump-server-first is used.
2863 NAME: sslproxy_cert_adapt
2866 TYPE: sslproxy_cert_adapt
2867 LOC: Config.ssl_client.cert_adapt
2870 sslproxy_cert_adapt <adaptation algorithm> acl ...
2872 The following certificate adaptation algorithms are supported:
2875 Sets the "Not After" property to the "Not After" property of
2876 the CA certificate used to sign generated certificates.
2879 Sets the "Not Before" property to the "Not Before" property of
2880 the CA certificate used to sign generated certificates.
2882 setCommonName or setCommonName{CN}
2883 Sets Subject.CN property to the host name specified as a
2884 CN parameter or, if no explicit CN parameter was specified,
2885 extracted from the CONNECT request. It is a misconfiguration
2886 to use setCommonName without an explicit parameter for
2887 intercepted or tproxied SSL connections.
2889 This clause only supports fast acl types.
2891 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2892 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2893 corresponding adaptation algorithm to generate the certificate and
2894 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2895 group (i.e., the first match wins within each algorithm group). If no
2896 acl(s) match, the default mimicking action takes place.
2898 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2899 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2900 CONNECT request that carries a domain name. In all other cases (CONNECT
2901 to an IP address or an intercepted SSL connection), Squid cannot detect
2902 the domain mismatch at certificate generation time when
2903 bump-server-first is used.
2906 NAME: sslpassword_program
2909 LOC: Config.Program.ssl_password
2912 Specify a program used for entering SSL key passphrases
2913 when using encrypted SSL certificate keys. If not specified
2914 keys must either be unencrypted, or Squid started with the -N
2915 option to allow it to query interactively for the passphrase.
2917 The key file name is given as argument to the program allowing
2918 selection of the right password if you have multiple encrypted
2923 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2924 -----------------------------------------------------------------------------
2927 NAME: sslcrtd_program
2930 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2931 LOC: Ssl::TheConfig.ssl_crtd
2933 Specify the location and options of the executable for certificate
2935 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2936 For more information use:
2937 @DEFAULT_SSL_CRTD@ -h
2940 NAME: sslcrtd_children
2941 TYPE: HelperChildConfig
2943 DEFAULT: 32 startup=5 idle=1
2944 LOC: Ssl::TheConfig.ssl_crtdChildren
2946 The maximum number of processes spawn to service ssl server.
2947 The maximum this may be safely set to is 32.
2949 The startup= and idle= options allow some measure of skew in your
2954 Sets the minimum number of processes to spawn when Squid
2955 starts or reconfigures. When set to zero the first request will
2956 cause spawning of the first child process to handle it.
2958 Starting too few children temporary slows Squid under load while it
2959 tries to spawn enough additional processes to cope with traffic.
2963 Sets a minimum of how many processes Squid is to try and keep available
2964 at all times. When traffic begins to rise above what the existing
2965 processes can handle this many more will be spawned up to the maximum
2966 configured. A minimum setting of 1 is required.
2970 Sets the maximum number of queued requests.
2971 If the queued requests exceed queue size for more than 3 minutes
2972 squid aborts its operation.
2973 The default value is set to 2*numberofchildren.
2975 You must have at least one ssl_crtd process.
2978 NAME: sslcrtvalidator_program
2982 LOC: Ssl::TheConfig.ssl_crt_validator
2984 Specify the location and options of the executable for ssl_crt_validator
2987 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2990 ttl=n TTL in seconds for cached results. The default is 60 secs
2991 cache=n limit the result cache size. The default value is 2048
2994 NAME: sslcrtvalidator_children
2995 TYPE: HelperChildConfig
2997 DEFAULT: 32 startup=5 idle=1 concurrency=1
2998 LOC: Ssl::TheConfig.ssl_crt_validator_Children
3000 The maximum number of processes spawn to service SSL server.
3001 The maximum this may be safely set to is 32.
3003 The startup= and idle= options allow some measure of skew in your
3008 Sets the minimum number of processes to spawn when Squid
3009 starts or reconfigures. When set to zero the first request will
3010 cause spawning of the first child process to handle it.
3012 Starting too few children temporary slows Squid under load while it
3013 tries to spawn enough additional processes to cope with traffic.
3017 Sets a minimum of how many processes Squid is to try and keep available
3018 at all times. When traffic begins to rise above what the existing
3019 processes can handle this many more will be spawned up to the maximum
3020 configured. A minimum setting of 1 is required.
3024 The number of requests each certificate validator helper can handle in
3025 parallel. A value of 0 indicates the certficate validator does not
3026 support concurrency. Defaults to 1.
3028 When this directive is set to a value >= 1 then the protocol
3029 used to communicate with the helper is modified to include
3030 a request ID in front of the request/response. The request
3031 ID from the request must be echoed back with the response
3036 Sets the maximum number of queued requests.
3037 If the queued requests exceed queue size for more than 3 minutes
3038 squid aborts its operation.
3039 The default value is set to 2*numberofchildren.
3041 You must have at least one ssl_crt_validator process.
3045 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
3046 -----------------------------------------------------------------------------
3054 To specify other caches in a hierarchy, use the format:
3056 cache_peer hostname type http-port icp-port [options]
3061 # hostname type port port options
3062 # -------------------- -------- ----- ----- -----------
3063 cache_peer parent.foo.net parent 3128 3130 default
3064 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
3065 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
3066 cache_peer example.com parent 80 0 default
3067 cache_peer cdn.example.com sibling 3128 0
3069 type: either 'parent', 'sibling', or 'multicast'.
3071 proxy-port: The port number where the peer accept HTTP requests.
3072 For other Squid proxies this is usually 3128
3073 For web servers this is usually 80
3075 icp-port: Used for querying neighbor caches about objects.
3076 Set to 0 if the peer does not support ICP or HTCP.
3077 See ICP and HTCP options below for additional details.
3080 ==== ICP OPTIONS ====
3082 You MUST also set icp_port and icp_access explicitly when using these options.
3083 The defaults will prevent peer traffic using ICP.
3086 no-query Disable ICP queries to this neighbor.
3089 Indicates the named peer is a member of a multicast group.
3090 ICP queries will not be sent directly to the peer, but ICP
3091 replies will be accepted from it.
3093 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
3094 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
3097 To only send ICP queries to this neighbor infrequently.
3098 This is used to keep the neighbor round trip time updated
3099 and is usually used in conjunction with weighted-round-robin.
3102 ==== HTCP OPTIONS ====
3104 You MUST also set htcp_port and htcp_access explicitly when using these options.
3105 The defaults will prevent peer traffic using HTCP.
3108 htcp Send HTCP, instead of ICP, queries to the neighbor.
3109 You probably also want to set the "icp-port" to 4827
3110 instead of 3130. This directive accepts a comma separated
3111 list of options described below.
3113 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3115 htcp=no-clr Send HTCP to the neighbor but without
3116 sending any CLR requests. This cannot be used with
3119 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3120 This cannot be used with no-clr.
3123 Send HTCP to the neighbor including CLRs but only when
3124 they do not result from PURGE requests.
3127 Forward any HTCP CLR requests this proxy receives to the peer.
3130 ==== PEER SELECTION METHODS ====
3132 The default peer selection method is ICP, with the first responding peer
3133 being used as source. These options can be used for better load balancing.
3136 default This is a parent cache which can be used as a "last-resort"
3137 if a peer cannot be located by any of the peer-selection methods.
3138 If specified more than once, only the first is used.
3140 round-robin Load-Balance parents which should be used in a round-robin
3141 fashion in the absence of any ICP queries.
3142 weight=N can be used to add bias.
3144 weighted-round-robin
3145 Load-Balance parents which should be used in a round-robin
3146 fashion with the frequency of each parent being based on the
3147 round trip time. Closer parents are used more often.
3148 Usually used for background-ping parents.
3149 weight=N can be used to add bias.
3151 carp Load-Balance parents which should be used as a CARP array.
3152 The requests will be distributed among the parents based on the
3153 CARP load balancing hash function based on their weight.
3155 userhash Load-balance parents based on the client proxy_auth or ident username.
3157 sourcehash Load-balance parents based on the client source IP.
3160 To be used only for cache peers of type "multicast".
3161 ALL members of this multicast group have "sibling"
3162 relationship with it, not "parent". This is to a multicast
3163 group when the requested object would be fetched only from
3164 a "parent" cache, anyway. It's useful, e.g., when
3165 configuring a pool of redundant Squid proxies, being
3166 members of the same multicast group.
3169 ==== PEER SELECTION OPTIONS ====
3171 weight=N use to affect the selection of a peer during any weighted
3172 peer-selection mechanisms.
3173 The weight must be an integer; default is 1,
3174 larger weights are favored more.
3175 This option does not affect parent selection if a peering
3176 protocol is not in use.
3178 basetime=N Specify a base amount to be subtracted from round trip
3180 It is subtracted before division by weight in calculating
3181 which parent to fectch from. If the rtt is less than the
3182 base time the rtt is set to a minimal value.
3184 ttl=N Specify a TTL to use when sending multicast ICP queries
3186 Only useful when sending to a multicast group.
3187 Because we don't accept ICP replies from random
3188 hosts, you must configure other group members as
3189 peers with the 'multicast-responder' option.
3191 no-delay To prevent access to this neighbor from influencing the
3194 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3195 enabled) for this host from the specified URL rather
3196 than the Squid default location.
3199 ==== CARP OPTIONS ====
3201 carp-key=key-specification
3202 use a different key than the full URL to hash against the peer.
3203 the key-specification is a comma-separated list of the keywords
3204 scheme, host, port, path, params
3205 Order is not important.
3207 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3209 originserver Causes this parent to be contacted as an origin server.
3210 Meant to be used in accelerator setups when the peer
3214 Set the Host header of requests forwarded to this peer.
3215 Useful in accelerator setups where the server (peer)
3216 expects a certain domain name but clients may request
3217 others. ie example.com or www.example.com
3219 no-digest Disable request of cache digests.
3222 Disables requesting ICMP RTT database (NetDB).
3225 ==== AUTHENTICATION OPTIONS ====
3228 If this is a personal/workgroup proxy and your parent
3229 requires proxy authentication.
3231 Note: The string can include URL escapes (i.e. %20 for
3232 spaces). This also means % must be written as %%.
3235 Send login details received from client to this peer.
3236 Both Proxy- and WWW-Authorization headers are passed
3237 without alteration to the peer.
3238 Authentication is not required by Squid for this to work.
3240 Note: This will pass any form of authentication but
3241 only Basic auth will work through a proxy unless the
3242 connection-auth options are also used.
3244 login=PASS Send login details received from client to this peer.
3245 Authentication is not required by this option.
3247 If there are no client-provided authentication headers
3248 to pass on, but username and password are available
3249 from an external ACL user= and password= result tags
3250 they may be sent instead.
3252 Note: To combine this with proxy_auth both proxies must
3253 share the same user database as HTTP only allows for
3254 a single login (one for proxy, one for origin server).
3255 Also be warned this will expose your users proxy
3256 password to the peer. USE WITH CAUTION
3259 Send the username to the upstream cache, but with a
3260 fixed password. This is meant to be used when the peer
3261 is in another administrative domain, but it is still
3262 needed to identify each user.
3263 The star can optionally be followed by some extra
3264 information which is added to the username. This can
3265 be used to identify this proxy to the peer, similar to
3266 the login=username:password option above.
3269 If this is a personal/workgroup proxy and your parent
3270 requires a secure proxy authentication.
3271 The first principal from the default keytab or defined by
3272 the environment variable KRB5_KTNAME will be used.
3274 WARNING: The connection may transmit requests from multiple
3275 clients. Negotiate often assumes end-to-end authentication
3276 and a single-client. Which is not strictly true here.
3278 login=NEGOTIATE:principal_name
3279 If this is a personal/workgroup proxy and your parent
3280 requires a secure proxy authentication.
3281 The principal principal_name from the default keytab or
3282 defined by the environment variable KRB5_KTNAME will be
3285 WARNING: The connection may transmit requests from multiple
3286 clients. Negotiate often assumes end-to-end authentication
3287 and a single-client. Which is not strictly true here.
3289 connection-auth=on|off
3290 Tell Squid that this peer does or not support Microsoft
3291 connection oriented authentication, and any such
3292 challenges received from there should be ignored.
3293 Default is auto to automatically determine the status
3297 Do not use a keytab to authenticate to a peer when
3298 login=NEGOTIATE is specified. Let the GSSAPI
3299 implementation determine which already existing
3300 credentials cache to use instead.
3303 ==== SSL / HTTPS / TLS OPTIONS ====
3305 ssl Encrypt connections to this peer with SSL/TLS.
3307 sslcert=/path/to/ssl/certificate
3308 A client SSL certificate to use when connecting to
3311 sslkey=/path/to/ssl/key
3312 The private SSL key corresponding to sslcert above.
3313 If 'sslkey' is not specified 'sslcert' is assumed to
3314 reference a combined file containing both the
3315 certificate and the key.
3317 sslcipher=... The list of valid SSL ciphers to use when connecting
3321 The minimum TLS protocol version to permit. To control
3322 SSLv3 use the ssloptions= parameter.
3323 Supported Values: 1.0 (default), 1.1, 1.2
3325 ssloptions=... Specify various SSL implementation options:
3327 NO_SSLv3 Disallow the use of SSLv3
3329 NO_TLSv1 Disallow the use of TLSv1.0
3331 NO_TLSv1_1 Disallow the use of TLSv1.1
3333 NO_TLSv1_2 Disallow the use of TLSv1.2
3336 Always create a new key when using
3337 temporary/ephemeral DH key exchanges
3340 Disable use of RFC5077 session tickets.
3341 Some servers may have problems
3342 understanding the TLS extension due
3343 to ambiguous specification in RFC4507.
3345 ALL Enable various bug workarounds
3346 suggested as "harmless" by OpenSSL
3347 Be warned that this reduces SSL/TLS
3348 strength to some attacks.
3350 See the OpenSSL SSL_CTX_set_options documentation for a
3353 tls-cafile= PEM file containing CA certificates to use when verifying
3354 the peer certificate. May be repeated to load multiple files.
3356 sslcapath=... A directory containing additional CA certificates to
3357 use when verifying the peer certificate.
3358 Requires OpenSSL or LibreSSL.
3360 sslcrlfile=... A certificate revocation list file to use when
3361 verifying the peer certificate.
3363 sslflags=... Specify various flags modifying the SSL implementation:
3366 Accept certificates even if they fail to
3370 Don't verify the peer certificate
3371 matches the server name
3373 ssldomain= The peer name as advertised in it's certificate.
3374 Used for verifying the correctness of the received peer
3375 certificate. If not specified the peer hostname will be
3379 Enable the "Front-End-Https: On" header needed when
3380 using Squid as a SSL frontend in front of Microsoft OWA.
3381 See MS KB document Q307347 for details on this header.
3382 If set to auto the header will only be added if the
3383 request is forwarded as a https:// URL.
3385 tls-default-ca[=off]
3386 Whether to use the system Trusted CAs. Default is ON.
3388 tls-no-npn Do not use the TLS NPN extension to advertise HTTP/1.1.
3390 ==== GENERAL OPTIONS ====
3393 A peer-specific connect timeout.
3394 Also see the peer_connect_timeout directive.
3396 connect-fail-limit=N
3397 How many times connecting to a peer must fail before
3398 it is marked as down. Standby connection failures
3399 count towards this limit. Default is 10.
3401 allow-miss Disable Squid's use of only-if-cached when forwarding
3402 requests to siblings. This is primarily useful when
3403 icp_hit_stale is used by the sibling. Excessive use
3404 of this option may result in forwarding loops. One way
3405 to prevent peering loops when using this option, is to
3406 deny cache peer usage on requests from a peer:
3408 cache_peer_access peerName deny fromPeer
3410 max-conn=N Limit the number of concurrent connections the Squid
3411 may open to this peer, including already opened idle
3412 and standby connections. There is no peer-specific
3413 connection limit by default.
3415 A peer exceeding the limit is not used for new
3416 requests unless a standby connection is available.
3418 max-conn currently works poorly with idle persistent
3419 connections: When a peer reaches its max-conn limit,
3420 and there are idle persistent connections to the peer,
3421 the peer may not be selected because the limiting code
3422 does not know whether Squid can reuse those idle
3425 standby=N Maintain a pool of N "hot standby" connections to an
3426 UP peer, available for requests when no idle
3427 persistent connection is available (or safe) to use.
3428 By default and with zero N, no such pool is maintained.
3429 N must not exceed the max-conn limit (if any).
3431 At start or after reconfiguration, Squid opens new TCP
3432 standby connections until there are N connections
3433 available and then replenishes the standby pool as
3434 opened connections are used up for requests. A used
3435 connection never goes back to the standby pool, but
3436 may go to the regular idle persistent connection pool
3437 shared by all peers and origin servers.
3439 Squid never opens multiple new standby connections
3440 concurrently. This one-at-a-time approach minimizes
3441 flooding-like effect on peers. Furthermore, just a few
3442 standby connections should be sufficient in most cases
3443 to supply most new requests with a ready-to-use
3446 Standby connections obey server_idle_pconn_timeout.
3447 For the feature to work as intended, the peer must be
3448 configured to accept and keep them open longer than
3449 the idle timeout at the connecting Squid, to minimize
3450 race conditions typical to idle used persistent
3451 connections. Default request_timeout and
3452 server_idle_pconn_timeout values ensure such a
3455 name=xxx Unique name for the peer.
3456 Required if you have multiple peers on the same host
3457 but different ports.
3458 This name can be used in cache_peer_access and similar
3459 directives to identify the peer.
3460 Can be used by outgoing access controls through the
3463 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3464 requests to this peer. Use normal address selection instead.
3465 This overrides the spoof_client_ip ACL.
3467 proxy-only objects fetched from the peer will not be stored locally.
3471 NAME: cache_peer_access
3474 DEFAULT_DOC: No peer usage restrictions.
3477 Restricts usage of cache_peer proxies.
3480 cache_peer_access peer-name allow|deny [!]aclname ...
3482 For the required peer-name parameter, use either the value of the
3483 cache_peer name=value parameter or, if name=value is missing, the
3484 cache_peer hostname parameter.
3486 This directive narrows down the selection of peering candidates, but
3487 does not determine the order in which the selected candidates are
3488 contacted. That order is determined by the peer selection algorithms
3489 (see PEER SELECTION sections in the cache_peer documentation).
3491 If a deny rule matches, the corresponding peer will not be contacted
3492 for the current transaction -- Squid will not send ICP queries and
3493 will not forward HTTP requests to that peer. An allow match leaves
3494 the corresponding peer in the selection. The first match for a given
3495 peer wins for that peer.
3497 The relative order of cache_peer_access directives for the same peer
3498 matters. The relative order of any two cache_peer_access directives
3499 for different peers does not matter. To ease interpretation, it is a
3500 good idea to group cache_peer_access directives for the same peer
3503 A single cache_peer_access directive may be evaluated multiple times
3504 for a given transaction because individual peer selection algorithms
3505 may check it independently from each other. These redundant checks
3506 may be optimized away in future Squid versions.
3508 This clause only supports fast acl types.
3509 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3513 NAME: neighbor_type_domain
3514 TYPE: hostdomaintype
3516 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3519 Modify the cache_peer neighbor type when passing requests
3520 about specific domains to the peer.
3523 neighbor_type_domain neighbor parent|sibling domain domain ...
3526 cache_peer foo.example.com parent 3128 3130
3527 neighbor_type_domain foo.example.com sibling .au .de
3529 The above configuration treats all requests to foo.example.com as a
3530 parent proxy unless the request is for a .au or .de ccTLD domain name.
3533 NAME: dead_peer_timeout
3537 LOC: Config.Timeout.deadPeer
3539 This controls how long Squid waits to declare a peer cache
3540 as "dead." If there are no ICP replies received in this
3541 amount of time, Squid will declare the peer dead and not
3542 expect to receive any further ICP replies. However, it
3543 continues to send ICP queries, and will mark the peer as
3544 alive upon receipt of the first subsequent ICP reply.
3546 This timeout also affects when Squid expects to receive ICP
3547 replies from peers. If more than 'dead_peer' seconds have
3548 passed since the last ICP reply was received, Squid will not
3549 expect to receive an ICP reply on the next query. Thus, if
3550 your time between requests is greater than this timeout, you
3551 will see a lot of requests sent DIRECT to origin servers
3552 instead of to your parents.
3555 NAME: forward_max_tries
3558 LOC: Config.forward_max_tries
3560 Controls how many different forward paths Squid will try
3561 before giving up. See also forward_timeout.
3563 NOTE: connect_retries (default: none) can make each of these
3564 possible forwarding paths be tried multiple times.
3568 MEMORY CACHE OPTIONS
3569 -----------------------------------------------------------------------------
3576 LOC: Config.memMaxSize
3578 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3579 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3580 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3581 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3583 'cache_mem' specifies the ideal amount of memory to be used
3585 * In-Transit objects
3587 * Negative-Cached objects
3589 Data for these objects are stored in 4 KB blocks. This
3590 parameter specifies the ideal upper limit on the total size of
3591 4 KB blocks allocated. In-Transit objects take the highest
3594 In-transit objects have priority over the others. When
3595 additional space is needed for incoming data, negative-cached
3596 and hot objects will be released. In other words, the
3597 negative-cached and hot objects will fill up any unused space
3598 not needed for in-transit objects.
3600 If circumstances require, this limit will be exceeded.
3601 Specifically, if your incoming request rate requires more than
3602 'cache_mem' of memory to hold in-transit objects, Squid will
3603 exceed this limit to satisfy the new requests. When the load
3604 decreases, blocks will be freed until the high-water mark is
3605 reached. Thereafter, blocks will be used to store hot
3608 If shared memory caching is enabled, Squid does not use the shared
3609 cache space for in-transit objects, but they still consume as much
3610 local memory as they need. For more details about the shared memory
3611 cache, see memory_cache_shared.
3614 NAME: maximum_object_size_in_memory
3618 LOC: Config.Store.maxInMemObjSize
3620 Objects greater than this size will not be attempted to kept in
3621 the memory cache. This should be set high enough to keep objects
3622 accessed frequently in memory to improve performance whilst low
3623 enough to keep larger objects from hoarding cache_mem.
3626 NAME: memory_cache_shared
3629 LOC: Config.memShared
3631 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3633 Controls whether the memory cache is shared among SMP workers.
3635 The shared memory cache is meant to occupy cache_mem bytes and replace
3636 the non-shared memory cache, although some entities may still be
3637 cached locally by workers for now (e.g., internal and in-transit
3638 objects may be served from a local memory cache even if shared memory
3639 caching is enabled).
3641 By default, the memory cache is shared if and only if all of the
3642 following conditions are satisfied: Squid runs in SMP mode with
3643 multiple workers, cache_mem is positive, and Squid environment
3644 supports required IPC primitives (e.g., POSIX shared memory segments
3645 and GCC-style atomic operations).
3647 To avoid blocking locks, shared memory uses opportunistic algorithms
3648 that do not guarantee that every cachable entity that could have been
3649 shared among SMP workers will actually be shared.
3651 Currently, entities exceeding 32KB in size cannot be shared.
3654 NAME: memory_cache_mode
3658 DEFAULT_DOC: Keep the most recently fetched objects in memory
3660 Controls which objects to keep in the memory cache (cache_mem)
3662 always Keep most recently fetched objects in memory (default)
3664 disk Only disk cache hits are kept in memory, which means
3665 an object must first be cached on disk and then hit
3666 a second time before cached in memory.
3668 network Only objects fetched from network is kept in memory
3671 NAME: memory_replacement_policy
3673 LOC: Config.memPolicy
3676 The memory replacement policy parameter determines which
3677 objects are purged from memory when memory space is needed.
3679 See cache_replacement_policy for details on algorithms.
3684 -----------------------------------------------------------------------------
3687 NAME: cache_replacement_policy
3689 LOC: Config.replPolicy
3692 The cache replacement policy parameter determines which
3693 objects are evicted (replaced) when disk space is needed.
3695 lru : Squid's original list based LRU policy
3696 heap GDSF : Greedy-Dual Size Frequency
3697 heap LFUDA: Least Frequently Used with Dynamic Aging
3698 heap LRU : LRU policy implemented using a heap
3700 Applies to any cache_dir lines listed below this directive.
3702 The LRU policies keeps recently referenced objects.
3704 The heap GDSF policy optimizes object hit rate by keeping smaller
3705 popular objects in cache so it has a better chance of getting a
3706 hit. It achieves a lower byte hit rate than LFUDA though since
3707 it evicts larger (possibly popular) objects.
3709 The heap LFUDA policy keeps popular objects in cache regardless of
3710 their size and thus optimizes byte hit rate at the expense of
3711 hit rate since one large, popular object will prevent many
3712 smaller, slightly less popular objects from being cached.
3714 Both policies utilize a dynamic aging mechanism that prevents
3715 cache pollution that can otherwise occur with frequency-based
3716 replacement policies.
3718 NOTE: if using the LFUDA replacement policy you should increase
3719 the value of maximum_object_size above its default of 4 MB to
3720 to maximize the potential byte hit rate improvement of LFUDA.
3722 For more information about the GDSF and LFUDA cache replacement
3723 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3724 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3727 NAME: minimum_object_size
3731 DEFAULT_DOC: no limit
3732 LOC: Config.Store.minObjectSize
3734 Objects smaller than this size will NOT be saved on disk. The
3735 value is specified in bytes, and the default is 0 KB, which
3736 means all responses can be stored.
3739 NAME: maximum_object_size
3743 LOC: Config.Store.maxObjectSize
3745 Set the default value for max-size parameter on any cache_dir.
3746 The value is specified in bytes, and the default is 4 MB.
3748 If you wish to get a high BYTES hit ratio, you should probably
3749 increase this (one 32 MB object hit counts for 3200 10KB
3752 If you wish to increase hit ratio more than you want to
3753 save bandwidth you should leave this low.
3755 NOTE: if using the LFUDA replacement policy you should increase
3756 this value to maximize the byte hit rate improvement of LFUDA!
3757 See cache_replacement_policy for a discussion of this policy.
3763 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3764 LOC: Config.cacheSwap
3767 cache_dir Type Directory-Name Fs-specific-data [options]
3769 You can specify multiple cache_dir lines to spread the
3770 cache among different disk partitions.
3772 Type specifies the kind of storage system to use. Only "ufs"
3773 is built by default. To enable any of the other storage systems
3774 see the --enable-storeio configure option.
3776 'Directory' is a top-level directory where cache swap
3777 files will be stored. If you want to use an entire disk
3778 for caching, this can be the mount-point directory.
3779 The directory must exist and be writable by the Squid
3780 process. Squid will NOT create this directory for you.
3782 In SMP configurations, cache_dir must not precede the workers option
3783 and should use configuration macros or conditionals to give each
3784 worker interested in disk caching a dedicated cache directory.
3787 ==== The ufs store type ====
3789 "ufs" is the old well-known Squid storage format that has always
3793 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3795 'Mbytes' is the amount of disk space (MB) to use under this
3796 directory. The default is 100 MB. Change this to suit your
3797 configuration. Do NOT put the size of your disk drive here.
3798 Instead, if you want Squid to use the entire disk drive,
3799 subtract 20% and use that value.
3801 'L1' is the number of first-level subdirectories which
3802 will be created under the 'Directory'. The default is 16.
3804 'L2' is the number of second-level subdirectories which
3805 will be created under each first-level directory. The default
3809 ==== The aufs store type ====
3811 "aufs" uses the same storage format as "ufs", utilizing
3812 POSIX-threads to avoid blocking the main Squid process on
3813 disk-I/O. This was formerly known in Squid as async-io.
3816 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3818 see argument descriptions under ufs above
3821 ==== The diskd store type ====
3823 "diskd" uses the same storage format as "ufs", utilizing a
3824 separate process to avoid blocking the main Squid process on
3828 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3830 see argument descriptions under ufs above
3832 Q1 specifies the number of unacknowledged I/O requests when Squid
3833 stops opening new files. If this many messages are in the queues,
3834 Squid won't open new files. Default is 64
3836 Q2 specifies the number of unacknowledged messages when Squid
3837 starts blocking. If this many messages are in the queues,
3838 Squid blocks until it receives some replies. Default is 72
3840 When Q1 < Q2 (the default), the cache directory is optimized
3841 for lower response time at the expense of a decrease in hit
3842 ratio. If Q1 > Q2, the cache directory is optimized for
3843 higher hit ratio at the expense of an increase in response
3847 ==== The rock store type ====
3850 cache_dir rock Directory-Name Mbytes [options]
3852 The Rock Store type is a database-style storage. All cached
3853 entries are stored in a "database" file, using fixed-size slots.
3854 A single entry occupies one or more slots.
3856 If possible, Squid using Rock Store creates a dedicated kid
3857 process called "disker" to avoid blocking Squid worker(s) on disk
3858 I/O. One disker kid is created for each rock cache_dir. Diskers
3859 are created only when Squid, running in daemon mode, has support
3860 for the IpcIo disk I/O module.
3862 swap-timeout=msec: Squid will not start writing a miss to or
3863 reading a hit from disk if it estimates that the swap operation
3864 will take more than the specified number of milliseconds. By
3865 default and when set to zero, disables the disk I/O time limit
3866 enforcement. Ignored when using blocking I/O module because
3867 blocking synchronous I/O does not allow Squid to estimate the
3868 expected swap wait time.
3870 max-swap-rate=swaps/sec: Artificially limits disk access using
3871 the specified I/O rate limit. Swap out requests that
3872 would cause the average I/O rate to exceed the limit are
3873 delayed. Individual swap in requests (i.e., hits or reads) are
3874 not delayed, but they do contribute to measured swap rate and
3875 since they are placed in the same FIFO queue as swap out
3876 requests, they may wait longer if max-swap-rate is smaller.
3877 This is necessary on file systems that buffer "too
3878 many" writes and then start blocking Squid and other processes
3879 while committing those writes to disk. Usually used together
3880 with swap-timeout to avoid excessive delays and queue overflows
3881 when disk demand exceeds available disk "bandwidth". By default
3882 and when set to zero, disables the disk I/O rate limit
3883 enforcement. Currently supported by IpcIo module only.
3885 slot-size=bytes: The size of a database "record" used for
3886 storing cached responses. A cached response occupies at least
3887 one slot and all database I/O is done using individual slots so
3888 increasing this parameter leads to more disk space waste while
3889 decreasing it leads to more disk I/O overheads. Should be a
3890 multiple of your operating system I/O page size. Defaults to
3891 16KBytes. A housekeeping header is stored with each slot and
3892 smaller slot-sizes will be rejected. The header is smaller than
3896 ==== COMMON OPTIONS ====
3898 no-store no new objects should be stored to this cache_dir.
3900 min-size=n the minimum object size in bytes this cache_dir
3901 will accept. It's used to restrict a cache_dir
3902 to only store large objects (e.g. AUFS) while
3903 other stores are optimized for smaller objects
3907 max-size=n the maximum object size in bytes this cache_dir
3909 The value in maximum_object_size directive sets
3910 the default unless more specific details are
3911 available (ie a small store capacity).
3913 Note: To make optimal use of the max-size limits you should order
3914 the cache_dir lines with the smallest max-size value first.
3918 # Uncomment and adjust the following to add a disk cache directory.
3919 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3923 NAME: store_dir_select_algorithm
3925 LOC: Config.store_dir_select_algorithm
3928 How Squid selects which cache_dir to use when the response
3929 object will fit into more than one.
3931 Regardless of which algorithm is used the cache_dir min-size
3932 and max-size parameters are obeyed. As such they can affect
3933 the selection algorithm by limiting the set of considered
3940 This algorithm is suited to caches with similar cache_dir
3941 sizes and disk speeds.
3943 The disk with the least I/O pending is selected.
3944 When there are multiple disks with the same I/O load ranking
3945 the cache_dir with most available capacity is selected.
3947 When a mix of cache_dir sizes are configured the faster disks
3948 have a naturally lower I/O loading and larger disks have more
3949 capacity. So space used to store objects and data throughput
3950 may be very unbalanced towards larger disks.
3955 This algorithm is suited to caches with unequal cache_dir
3958 Each cache_dir is selected in a rotation. The next suitable
3961 Available cache_dir capacity is only considered in relation
3962 to whether the object will fit and meets the min-size and
3963 max-size parameters.
3965 Disk I/O loading is only considered to prevent overload on slow
3966 disks. This algorithm does not spread objects by size, so any
3967 I/O loading per-disk may appear very unbalanced and volatile.
3969 If several cache_dirs use similar min-size, max-size, or other
3970 limits to to reject certain responses, then do not group such
3971 cache_dir lines together, to avoid round-robin selection bias
3972 towards the first cache_dir after the group. Instead, interleave
3973 cache_dir lines from different groups. For example:
3975 store_dir_select_algorithm round-robin
3976 cache_dir rock /hdd1 ... min-size=100000
3977 cache_dir rock /ssd1 ... max-size=99999
3978 cache_dir rock /hdd2 ... min-size=100000
3979 cache_dir rock /ssd2 ... max-size=99999
3980 cache_dir rock /hdd3 ... min-size=100000
3981 cache_dir rock /ssd3 ... max-size=99999
3984 NAME: max_open_disk_fds
3986 LOC: Config.max_open_disk_fds
3988 DEFAULT_DOC: no limit
3990 To avoid having disk as the I/O bottleneck Squid can optionally
3991 bypass the on-disk cache if more than this amount of disk file
3992 descriptors are open.
3994 A value of 0 indicates no limit.
3997 NAME: cache_swap_low
3998 COMMENT: (percent, 0-100)
4001 LOC: Config.Swap.lowWaterMark
4003 The low-water mark for AUFS/UFS/diskd cache object eviction by
4004 the cache_replacement_policy algorithm.
4006 Removal begins when the swap (disk) usage of a cache_dir is
4007 above this low-water mark and attempts to maintain utilization
4008 near the low-water mark.
4010 As swap utilization increases towards the high-water mark set
4011 by cache_swap_high object eviction becomes more agressive.
4013 The value difference in percentages between low- and high-water
4014 marks represent an eviction rate of 300 objects per second and
4015 the rate continues to scale in agressiveness by multiples of
4016 this above the high-water mark.
4018 Defaults are 90% and 95%. If you have a large cache, 5% could be
4019 hundreds of MB. If this is the case you may wish to set these
4020 numbers closer together.
4022 See also cache_swap_high and cache_replacement_policy
4025 NAME: cache_swap_high
4026 COMMENT: (percent, 0-100)
4029 LOC: Config.Swap.highWaterMark
4031 The high-water mark for AUFS/UFS/diskd cache object eviction by
4032 the cache_replacement_policy algorithm.
4034 Removal begins when the swap (disk) usage of a cache_dir is
4035 above the low-water mark set by cache_swap_low and attempts to
4036 maintain utilization near the low-water mark.
4038 As swap utilization increases towards this high-water mark object
4039 eviction becomes more agressive.
4041 The value difference in percentages between low- and high-water
4042 marks represent an eviction rate of 300 objects per second and
4043 the rate continues to scale in agressiveness by multiples of
4044 this above the high-water mark.
4046 Defaults are 90% and 95%. If you have a large cache, 5% could be
4047 hundreds of MB. If this is the case you may wish to set these
4048 numbers closer together.
4050 See also cache_swap_low and cache_replacement_policy
4055 -----------------------------------------------------------------------------
4062 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
4066 logformat <name> <format specification>
4068 Defines an access log format.
4070 The <format specification> is a string with embedded % format codes
4072 % format codes all follow the same basic structure where all but
4073 the formatcode is optional. Output strings are automatically escaped
4074 as required according to their context and the output format
4075 modifiers are usually not needed, but can be specified if an explicit
4076 output format is desired.
4078 % ["|[|'|#|/] [-] [[0]width] [{arg}] formatcode [{arg}]
4080 " output in quoted string format
4081 [ output in squid text log format as used by log_mime_hdrs
4082 # output in URL quoted format
4083 / output in shell \-escaped format
4088 width minimum and/or maximum field width:
4089 [width_min][.width_max]
4090 When minimum starts with 0, the field is zero-padded.
4091 String values exceeding maximum width are truncated.
4093 {arg} argument such as header name etc. This field may be
4094 placed before or after the token, but not both at once.
4098 % a literal % character
4099 sn Unique sequence number per log line entry
4100 err_code The ID of an error response served by Squid or
4101 a similar internal error identifier.
4102 err_detail Additional err_code-dependent error information.
4103 note The annotation specified by the argument. Also
4104 logs the adaptation meta headers set by the
4105 adaptation_meta configuration parameter.
4106 If no argument given all annotations logged.
4107 The argument may include a separator to use with
4110 By default, multiple note values are separated with ","
4111 and multiple notes are separated with "\r\n".
4112 When logging named notes with %{name}note, the
4113 explicitly configured separator is used between note
4114 values. When logging all notes with %note, the
4115 explicitly configured separator is used between
4116 individual notes. There is currently no way to
4117 specify both value and notes separators when logging
4118 all notes with %note.
4120 Connection related format codes:
4122 >a Client source IP address
4124 >p Client source port
4125 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
4126 >la Local IP address the client connected to
4127 >lp Local port number the client connected to
4128 >qos Client connection TOS/DSCP value set by Squid
4129 >nfmark Client connection netfilter mark set by Squid
4131 la Local listening IP address the client connection was connected to.
4132 lp Local listening port number the client connection was connected to.
4134 <a Server IP address of the last server or peer connection
4135 <A Server FQDN or peer name
4136 <p Server port number of the last server or peer connection
4137 <la Local IP address of the last server or peer connection
4138 <lp Local port number of the last server or peer connection
4139 <qos Server connection TOS/DSCP value set by Squid
4140 <nfmark Server connection netfilter mark set by Squid
4142 Time related format codes:
4144 ts Seconds since epoch
4145 tu subsecond time (milliseconds)
4146 tl Local time. Optional strftime format argument
4147 default %d/%b/%Y:%H:%M:%S %z
4148 tg GMT time. Optional strftime format argument
4149 default %d/%b/%Y:%H:%M:%S %z
4150 tr Response time (milliseconds)
4151 dt Total time spent making DNS lookups (milliseconds)
4152 tS Approximate master transaction start time in
4153 <full seconds since epoch>.<fractional seconds> format.
4154 Currently, Squid considers the master transaction
4155 started when a complete HTTP request header initiating
4156 the transaction is received from the client. This is
4157 the same value that Squid uses to calculate transaction
4158 response time when logging %tr to access.log. Currently,
4159 Squid uses millisecond resolution for %tS values,
4160 similar to the default access.log "current time" field
4163 Access Control related format codes:
4165 et Tag returned by external acl
4166 ea Log string returned by external acl
4167 un User name (any available)
4168 ul User name from authentication
4169 ue User name from external acl helper
4170 ui User name from ident
4171 un A user name. Expands to the first available name
4172 from the following list of information sources:
4173 - authenticated user name, like %ul
4174 - user name supplied by an external ACL, like %ue
4175 - SSL client name, like %us
4176 - ident user name, like %ui
4177 credentials Client credentials. The exact meaning depends on
4178 the authentication scheme: For Basic authentication,
4179 it is the password; for Digest, the realm sent by the
4180 client; for NTLM and Negotiate, the client challenge
4181 or client credentials prefixed with "YR " or "KK ".
4183 HTTP related format codes:
4187 [http::]rm Request method (GET/POST etc)
4188 [http::]>rm Request method from client
4189 [http::]<rm Request method sent to server or peer
4190 [http::]ru Request URL from client (historic, filtered for logging)
4191 [http::]>ru Request URL from client
4192 [http::]<ru Request URL sent to server or peer
4193 [http::]>rs Request URL scheme from client
4194 [http::]<rs Request URL scheme sent to server or peer
4195 [http::]>rd Request URL domain from client
4196 [http::]<rd Request URL domain sent to server or peer
4197 [http::]>rP Request URL port from client
4198 [http::]<rP Request URL port sent to server or peer
4199 [http::]rp Request URL path excluding hostname
4200 [http::]>rp Request URL path excluding hostname from client
4201 [http::]<rp Request URL path excluding hostname sent to server or peer
4202 [http::]rv Request protocol version
4203 [http::]>rv Request protocol version from client
4204 [http::]<rv Request protocol version sent to server or peer
4206 [http::]>h Original received request header.
4207 Usually differs from the request header sent by
4208 Squid, although most fields are often preserved.
4209 Accepts optional header field name/value filter
4210 argument using name[:[separator]element] format.
4211 [http::]>ha Received request header after adaptation and
4212 redirection (pre-cache REQMOD vectoring point).
4213 Usually differs from the request header sent by
4214 Squid, although most fields are often preserved.
4215 Optional header name argument as for >h
4219 [http::]<Hs HTTP status code received from the next hop
4220 [http::]>Hs HTTP status code sent to the client
4222 [http::]<h Reply header. Optional header name argument
4225 [http::]mt MIME content type
4230 [http::]st Total size of request + reply traffic with client
4231 [http::]>st Total size of request received from client.
4232 Excluding chunked encoding bytes.
4233 [http::]<st Total size of reply sent to client (after adaptation)
4235 [http::]>sh Size of request headers received from client
4236 [http::]<sh Size of reply headers sent to client (after adaptation)
4238 [http::]<sH Reply high offset sent
4239 [http::]<sS Upstream object size
4241 [http::]<bs Number of HTTP-equivalent message body bytes
4242 received from the next hop, excluding chunked
4243 transfer encoding and control messages.
4244 Generated FTP/Gopher listings are treated as
4249 [http::]<pt Peer response time in milliseconds. The timer starts
4250 when the last request byte is sent to the next hop
4251 and stops when the last response byte is received.
4252 [http::]<tt Total time in milliseconds. The timer
4253 starts with the first connect request (or write I/O)
4254 sent to the first selected peer. The timer stops
4255 with the last I/O with the last peer.
4257 Squid handling related format codes:
4259 Ss Squid request status (TCP_MISS etc)
4260 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4262 SSL-related format codes:
4264 ssl::bump_mode SslBump decision for the transaction:
4266 For CONNECT requests that initiated bumping of
4267 a connection and for any request received on
4268 an already bumped connection, Squid logs the
4269 corresponding SslBump mode ("server-first" or
4270 "client-first"). See the ssl_bump option for
4271 more information about these modes.
4273 A "none" token is logged for requests that
4274 triggered "ssl_bump" ACL evaluation matching
4275 either a "none" rule or no rules at all.
4277 In all other cases, a single dash ("-") is
4280 ssl::>sni SSL client SNI sent to Squid. Available only
4281 after the peek, stare, or splice SSL bumping
4285 The Subject field of the received client
4286 SSL certificate or a dash ('-') if Squid has
4287 received an invalid/malformed certificate or
4288 no certificate at all. Consider encoding the
4289 logged value because Subject often has spaces.
4292 The Issuer field of the received client
4293 SSL certificate or a dash ('-') if Squid has
4294 received an invalid/malformed certificate or
4295 no certificate at all. Consider encoding the
4296 logged value because Issuer often has spaces.
4299 The list of certificate validation errors
4300 detected by Squid (including OpenSSL and
4301 certificate validation helper components). The
4302 errors are listed in the discovery order. By
4303 default, the error codes are separated by ':'.
4304 Accepts an optional separator argument.
4306 %ssl::>negotiated_version The negotiated TLS version of the
4309 %ssl::<negotiated_version The negotiated TLS version of the
4310 last server or peer connection.
4312 %ssl::>received_hello_version The TLS version of the Hello
4313 message received from TLS client.
4315 %ssl::<received_hello_version The TLS version of the Hello
4316 message received from TLS server.
4318 %ssl::>received_supported_version The maximum TLS version
4319 supported by the TLS client.
4321 %ssl::<received_supported_version The maximum TLS version
4322 supported by the TLS server.
4324 %ssl::>negotiated_cipher The negotiated cipher of the
4327 %ssl::<negotiated_cipher The negotiated cipher of the
4328 last server or peer connection.
4330 If ICAP is enabled, the following code becomes available (as
4331 well as ICAP log codes documented with the icap_log option):
4333 icap::tt Total ICAP processing time for the HTTP
4334 transaction. The timer ticks when ICAP
4335 ACLs are checked and when ICAP
4336 transaction is in progress.
4338 If adaptation is enabled the following codes become available:
4340 adapt::<last_h The header of the last ICAP response or
4341 meta-information from the last eCAP
4342 transaction related to the HTTP transaction.
4343 Like <h, accepts an optional header name
4346 adapt::sum_trs Summed adaptation transaction response
4347 times recorded as a comma-separated list in
4348 the order of transaction start time. Each time
4349 value is recorded as an integer number,
4350 representing response time of one or more
4351 adaptation (ICAP or eCAP) transaction in
4352 milliseconds. When a failed transaction is
4353 being retried or repeated, its time is not
4354 logged individually but added to the
4355 replacement (next) transaction. See also:
4358 adapt::all_trs All adaptation transaction response times.
4359 Same as adaptation_strs but response times of
4360 individual transactions are never added
4361 together. Instead, all transaction response
4362 times are recorded individually.
4364 You can prefix adapt::*_trs format codes with adaptation
4365 service name in curly braces to record response time(s) specific
4366 to that service. For example: %{my_service}adapt::sum_trs
4368 The default formats available (which do not need re-defining) are:
4370 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
4371 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
4372 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
4373 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
4374 logformat useragent %>a [%tl] "%{User-Agent}>h"
4376 NOTE: When the log_mime_hdrs directive is set to ON.
4377 The squid, common and combined formats have a safely encoded copy
4378 of the mime headers appended to each line within a pair of brackets.
4380 NOTE: The common and combined formats are not quite true to the Apache definition.
4381 The logs from Squid contain an extra status and hierarchy code appended.
4385 NAME: access_log cache_access_log
4387 LOC: Config.Log.accesslogs
4388 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
4390 Configures whether and how Squid logs HTTP and ICP transactions.
4391 If access logging is enabled, a single line is logged for every
4392 matching HTTP or ICP request. The recommended directive formats are:
4394 access_log <module>:<place> [option ...] [acl acl ...]
4395 access_log none [acl acl ...]
4397 The following directive format is accepted but may be deprecated:
4398 access_log <module>:<place> [<logformat name> [acl acl ...]]
4400 In most cases, the first ACL name must not contain the '=' character
4401 and should not be equal to an existing logformat name. You can always
4402 start with an 'all' ACL to work around those restrictions.
4404 Will log to the specified module:place using the specified format (which
4405 must be defined in a logformat directive) those entries which match
4406 ALL the acl's specified (which must be defined in acl clauses).
4407 If no acl is specified, all requests will be logged to this destination.
4409 ===== Available options for the recommended directive format =====
4411 logformat=name Names log line format (either built-in or
4412 defined by a logformat directive). Defaults
4415 buffer-size=64KB Defines approximate buffering limit for log
4416 records (see buffered_logs). Squid should not
4417 keep more than the specified size and, hence,
4418 should flush records before the buffer becomes
4419 full to avoid overflows under normal
4420 conditions (the exact flushing algorithm is
4421 module-dependent though). The on-error option
4422 controls overflow handling.
4424 on-error=die|drop Defines action on unrecoverable errors. The
4425 'drop' action ignores (i.e., does not log)
4426 affected log records. The default 'die' action
4427 kills the affected worker. The drop action
4428 support has not been tested for modules other
4431 rotate=N Specifies the number of log file rotations to
4432 make when you run 'squid -k rotate'. The default
4433 is to obey the logfile_rotate directive. Setting
4434 rotate=0 will disable the file name rotation,
4435 but the log files are still closed and re-opened.
4436 This will enable you to rename the logfiles
4437 yourself just before sending the rotate signal.
4438 Only supported by the stdio module.
4440 ===== Modules Currently available =====
4442 none Do not log any requests matching these ACL.
4443 Do not specify Place or logformat name.
4445 stdio Write each log line to disk immediately at the completion of
4447 Place: the filename and path to be written.
4449 daemon Very similar to stdio. But instead of writing to disk the log
4450 line is passed to a daemon helper for asychronous handling instead.
4451 Place: varies depending on the daemon.
4453 log_file_daemon Place: the file name and path to be written.
4455 syslog To log each request via syslog facility.
4456 Place: The syslog facility and priority level for these entries.
4457 Place Format: facility.priority
4459 where facility could be any of:
4460 authpriv, daemon, local0 ... local7 or user.
4462 And priority could be any of:
4463 err, warning, notice, info, debug.
4465 udp To send each log line as text data to a UDP receiver.
4466 Place: The destination host name or IP and port.
4467 Place Format: //host:port
4469 tcp To send each log line as text data to a TCP receiver.
4470 Lines may be accumulated before sending (see buffered_logs).
4471 Place: The destination host name or IP and port.
4472 Place Format: //host:port
4475 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4481 LOC: Config.Log.icaplogs
4484 ICAP log files record ICAP transaction summaries, one line per
4487 The icap_log option format is:
4488 icap_log <filepath> [<logformat name> [acl acl ...]]
4489 icap_log none [acl acl ...]]
4491 Please see access_log option documentation for details. The two
4492 kinds of logs share the overall configuration approach and many
4495 ICAP processing of a single HTTP message or transaction may
4496 require multiple ICAP transactions. In such cases, multiple
4497 ICAP transaction log lines will correspond to a single access
4500 ICAP log uses logformat codes that make sense for an ICAP
4501 transaction. Header-related codes are applied to the HTTP header
4502 embedded in an ICAP server response, with the following caveats:
4503 For REQMOD, there is no HTTP response header unless the ICAP
4504 server performed request satisfaction. For RESPMOD, the HTTP
4505 request header is the header sent to the ICAP server. For
4506 OPTIONS, there are no HTTP headers.
4508 The following format codes are also available for ICAP logs:
4510 icap::<A ICAP server IP address. Similar to <A.
4512 icap::<service_name ICAP service name from the icap_service
4513 option in Squid configuration file.
4515 icap::ru ICAP Request-URI. Similar to ru.
4517 icap::rm ICAP request method (REQMOD, RESPMOD, or
4518 OPTIONS). Similar to existing rm.
4520 icap::>st Bytes sent to the ICAP server (TCP payload
4521 only; i.e., what Squid writes to the socket).
4523 icap::<st Bytes received from the ICAP server (TCP
4524 payload only; i.e., what Squid reads from
4527 icap::<bs Number of message body bytes received from the
4528 ICAP server. ICAP message body, if any, usually
4529 includes encapsulated HTTP message headers and
4530 possibly encapsulated HTTP message body. The
4531 HTTP body part is dechunked before its size is
4534 icap::tr Transaction response time (in
4535 milliseconds). The timer starts when
4536 the ICAP transaction is created and
4537 stops when the transaction is completed.
4540 icap::tio Transaction I/O time (in milliseconds). The
4541 timer starts when the first ICAP request
4542 byte is scheduled for sending. The timers
4543 stops when the last byte of the ICAP response
4546 icap::to Transaction outcome: ICAP_ERR* for all
4547 transaction errors, ICAP_OPT for OPTION
4548 transactions, ICAP_ECHO for 204
4549 responses, ICAP_MOD for message
4550 modification, and ICAP_SAT for request
4551 satisfaction. Similar to Ss.
4553 icap::Hs ICAP response status code. Similar to Hs.
4555 icap::>h ICAP request header(s). Similar to >h.
4557 icap::<h ICAP response header(s). Similar to <h.
4559 The default ICAP log format, which can be used without an explicit
4560 definition, is called icap_squid:
4562 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4564 See also: logformat, log_icap, and %adapt::<last_h
4567 NAME: logfile_daemon
4569 DEFAULT: @DEFAULT_LOGFILED@
4570 LOC: Log::TheConfig.logfile_daemon
4572 Specify the path to the logfile-writing daemon. This daemon is
4573 used to write the access and store logs, if configured.
4575 Squid sends a number of commands to the log daemon:
4576 L<data>\n - logfile data
4581 r<n>\n - set rotate count to <n>
4582 b<n>\n - 1 = buffer output, 0 = don't buffer output
4584 No responses is expected.
4587 NAME: stats_collection
4589 LOC: Config.accessList.stats_collection
4591 DEFAULT_DOC: Allow logging for all transactions.
4592 COMMENT: allow|deny acl acl...
4594 This options allows you to control which requests gets accounted
4595 in performance counters.
4597 This clause only supports fast acl types.
4598 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4601 NAME: cache_store_log
4604 LOC: Config.Log.store
4606 Logs the activities of the storage manager. Shows which
4607 objects are ejected from the cache, and which objects are
4608 saved and for how long.
4609 There are not really utilities to analyze this data, so you can safely
4610 disable it (the default).
4612 Store log uses modular logging outputs. See access_log for the list
4613 of modules supported.
4616 cache_store_log stdio:@DEFAULT_STORE_LOG@
4617 cache_store_log daemon:@DEFAULT_STORE_LOG@
4620 NAME: cache_swap_state cache_swap_log
4622 LOC: Config.Log.swap
4624 DEFAULT_DOC: Store the journal inside its cache_dir
4626 Location for the cache "swap.state" file. This index file holds
4627 the metadata of objects saved on disk. It is used to rebuild
4628 the cache during startup. Normally this file resides in each
4629 'cache_dir' directory, but you may specify an alternate
4630 pathname here. Note you must give a full filename, not just
4631 a directory. Since this is the index for the whole object
4632 list you CANNOT periodically rotate it!
4634 If %s can be used in the file name it will be replaced with a
4635 a representation of the cache_dir name where each / is replaced
4636 with '.'. This is needed to allow adding/removing cache_dir
4637 lines when cache_swap_log is being used.
4639 If have more than one 'cache_dir', and %s is not used in the name
4640 these swap logs will have names such as:
4646 The numbered extension (which is added automatically)
4647 corresponds to the order of the 'cache_dir' lines in this
4648 configuration file. If you change the order of the 'cache_dir'
4649 lines in this file, these index files will NOT correspond to
4650 the correct 'cache_dir' entry (unless you manually rename
4651 them). We recommend you do NOT use this option. It is
4652 better to keep these index files in each 'cache_dir' directory.
4655 NAME: logfile_rotate
4658 LOC: Config.Log.rotateNumber
4660 Specifies the default number of logfile rotations to make when you
4661 type 'squid -k rotate'. The default is 10, which will rotate
4662 with extensions 0 through 9. Setting logfile_rotate to 0 will
4663 disable the file name rotation, but the logfiles are still closed
4664 and re-opened. This will enable you to rename the logfiles
4665 yourself just before sending the rotate signal.
4667 Note, from Squid-3.1 this option is only a default for cache.log,
4668 that log can be rotated separately by using debug_options.
4670 Note, from Squid-3.6 this option is only a default for access.log
4671 recorded by stdio: module. Those logs can be rotated separately by
4672 using the rotate=N option on their access_log directive.
4674 Note, the 'squid -k rotate' command normally sends a USR1
4675 signal to the running squid process. In certain situations
4676 (e.g. on Linux with Async I/O), USR1 is used for other
4677 purposes, so -k rotate uses another signal. It is best to get
4678 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4685 DEFAULT: @DEFAULT_MIME_TABLE@
4686 LOC: Config.mimeTablePathname
4688 Path to Squid's icon configuration file.
4690 You shouldn't need to change this, but the default file contains
4691 examples and formatting information if you do.
4697 LOC: Config.onoff.log_mime_hdrs
4700 The Cache can record both the request and the response MIME
4701 headers for each HTTP transaction. The headers are encoded
4702 safely and will appear as two bracketed fields at the end of
4703 the access log (for either the native or httpd-emulated log
4704 formats). To enable this logging set log_mime_hdrs to 'on'.
4709 DEFAULT: @DEFAULT_PID_FILE@
4710 LOC: Config.pidFilename
4712 A filename to write the process-id to. To disable, enter "none".
4715 NAME: client_netmask
4717 LOC: Config.Addrs.client_netmask
4719 DEFAULT_DOC: Log full client IP address
4721 A netmask for client addresses in logfiles and cachemgr output.
4722 Change this to protect the privacy of your cache clients.
4723 A netmask of 255.255.255.0 will log all IP's in that range with
4724 the last digit set to '0'.
4727 NAME: strip_query_terms
4729 LOC: Config.onoff.strip_query_terms
4732 By default, Squid strips query terms from requested URLs before
4733 logging. This protects your user's privacy and reduces log size.
4735 When investigating HIT/MISS or other caching behaviour you
4736 will need to disable this to see the full URL used by Squid.
4743 LOC: Config.onoff.buffered_logs
4745 Whether to write/send access_log records ASAP or accumulate them and
4746 then write/send them in larger chunks. Buffering may improve
4747 performance because it decreases the number of I/Os. However,
4748 buffering increases the delay before log records become available to
4749 the final recipient (e.g., a disk file or logging daemon) and,
4750 hence, increases the risk of log records loss.
4752 Note that even when buffered_logs are off, Squid may have to buffer
4753 records if it cannot write/send them immediately due to pending I/Os
4754 (e.g., the I/O writing the previous log record) or connectivity loss.
4756 Currently honored by 'daemon' and 'tcp' access_log modules only.
4759 NAME: netdb_filename
4761 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4762 LOC: Config.netdbFilename
4765 Where Squid stores it's netdb journal.
4766 When enabled this journal preserves netdb state between restarts.
4768 To disable, enter "none".
4772 OPTIONS FOR TROUBLESHOOTING
4773 -----------------------------------------------------------------------------
4778 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4779 LOC: Debug::cache_log
4781 Squid administrative logging file.
4783 This is where general information about Squid behavior goes. You can
4784 increase the amount of data logged to this file and how often it is
4785 rotated with "debug_options"
4791 DEFAULT_DOC: Log all critical and important messages.
4792 LOC: Debug::debugOptions
4794 Logging options are set as section,level where each source file
4795 is assigned a unique section. Lower levels result in less
4796 output, Full debugging (level 9) can result in a very large
4797 log file, so be careful.
4799 The magic word "ALL" sets debugging levels for all sections.
4800 The default is to run with "ALL,1" to record important warnings.
4802 The rotate=N option can be used to keep more or less of these logs
4803 than would otherwise be kept by logfile_rotate.
4804 For most uses a single log should be enough to monitor current
4805 events affecting Squid.
4810 LOC: Config.coredump_dir
4811 DEFAULT_IF_NONE: none
4812 DEFAULT_DOC: Use the directory from where Squid was started.
4814 By default Squid leaves core files in the directory from where
4815 it was started. If you set 'coredump_dir' to a directory
4816 that exists, Squid will chdir() to that directory at startup
4817 and coredump files will be left there.
4821 # Leave coredumps in the first cache dir
4822 coredump_dir @DEFAULT_SWAP_DIR@
4828 OPTIONS FOR FTP GATEWAYING
4829 -----------------------------------------------------------------------------
4835 LOC: Config.Ftp.anon_user
4837 If you want the anonymous login password to be more informative
4838 (and enable the use of picky FTP servers), set this to something
4839 reasonable for your domain, like wwwuser@somewhere.net
4841 The reason why this is domainless by default is the
4842 request can be made on the behalf of a user in any domain,
4843 depending on how the cache is used.
4844 Some FTP server also validate the email address is valid
4845 (for example perl.com).
4851 LOC: Config.Ftp.passive
4853 If your firewall does not allow Squid to use passive
4854 connections, turn off this option.
4856 Use of ftp_epsv_all option requires this to be ON.
4862 LOC: Config.Ftp.epsv_all
4864 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4866 NATs may be able to put the connection on a "fast path" through the
4867 translator, as the EPRT command will never be used and therefore,
4868 translation of the data portion of the segments will never be needed.
4870 When a client only expects to do two-way FTP transfers this may be
4872 If squid finds that it must do a three-way FTP transfer after issuing
4873 an EPSV ALL command, the FTP session will fail.
4875 If you have any doubts about this option do not use it.
4876 Squid will nicely attempt all other connection methods.
4878 Requires ftp_passive to be ON (default) for any effect.
4884 LOC: Config.accessList.ftp_epsv
4886 FTP Protocol extensions permit the use of a special "EPSV" command.
4888 NATs may be able to put the connection on a "fast path" through the
4889 translator using EPSV, as the EPRT command will never be used
4890 and therefore, translation of the data portion of the segments
4891 will never be needed.
4893 EPSV is often required to interoperate with FTP servers on IPv6
4894 networks. On the other hand, it may break some IPv4 servers.
4896 By default, EPSV may try EPSV with any FTP server. To fine tune
4897 that decision, you may restrict EPSV to certain clients or servers
4900 ftp_epsv allow|deny al1 acl2 ...
4902 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4904 Only fast ACLs are supported.
4905 Requires ftp_passive to be ON (default) for any effect.
4911 LOC: Config.Ftp.eprt
4913 FTP Protocol extensions permit the use of a special "EPRT" command.
4915 This extension provides a protocol neutral alternative to the
4916 IPv4-only PORT command. When supported it enables active FTP data
4917 channels over IPv6 and efficient NAT handling.
4919 Turning this OFF will prevent EPRT being attempted and will skip
4920 straight to using PORT for IPv4 servers.
4922 Some devices are known to not handle this extension correctly and
4923 may result in crashes. Devices which suport EPRT enough to fail
4924 cleanly will result in Squid attempting PORT anyway. This directive
4925 should only be disabled when EPRT results in device failures.
4927 WARNING: Doing so will convert Squid back to the old behavior with all
4928 the related problems with external NAT devices/layers and IPv4-only FTP.
4931 NAME: ftp_sanitycheck
4934 LOC: Config.Ftp.sanitycheck
4936 For security and data integrity reasons Squid by default performs
4937 sanity checks of the addresses of FTP data connections ensure the
4938 data connection is to the requested server. If you need to allow
4939 FTP connections to servers using another IP address for the data
4940 connection turn this off.
4943 NAME: ftp_telnet_protocol
4946 LOC: Config.Ftp.telnet
4948 The FTP protocol is officially defined to use the telnet protocol
4949 as transport channel for the control connection. However, many
4950 implementations are broken and does not respect this aspect of
4953 If you have trouble accessing files with ASCII code 255 in the
4954 path or similar problems involving this ASCII code you can
4955 try setting this directive to off. If that helps, report to the
4956 operator of the FTP server in question that their FTP server
4957 is broken and does not follow the FTP standard.
4961 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4962 -----------------------------------------------------------------------------
4967 DEFAULT: @DEFAULT_DISKD@
4968 LOC: Config.Program.diskd
4970 Specify the location of the diskd executable.
4971 Note this is only useful if you have compiled in
4972 diskd as one of the store io modules.
4975 NAME: unlinkd_program
4978 DEFAULT: @DEFAULT_UNLINKD@
4979 LOC: Config.Program.unlinkd
4981 Specify the location of the executable for file deletion process.
4984 NAME: pinger_program
4987 DEFAULT: @DEFAULT_PINGER@
4990 Specify the location of the executable for the pinger process.
4999 Control whether the pinger is active at run-time.
5000 Enables turning ICMP pinger on and off with a simple
5001 squid -k reconfigure.
5006 OPTIONS FOR URL REWRITING
5007 -----------------------------------------------------------------------------
5010 NAME: url_rewrite_program redirect_program
5012 LOC: Config.Program.redirect
5015 Specify the location of the executable URL rewriter to use.
5016 Since they can perform almost any function there isn't one included.
5018 For each requested URL, the rewriter will receive on line with the format
5020 [channel-ID <SP>] URL [<SP> extras]<NL>
5022 See url_rewrite_extras on how to send "extras" with optional values to
5024 After processing the request the helper must reply using the following format:
5026 [channel-ID <SP>] result [<SP> kv-pairs]
5028 The result code can be:
5030 OK status=30N url="..."
5031 Redirect the URL to the one supplied in 'url='.
5032 'status=' is optional and contains the status code to send
5033 the client in Squids HTTP response. It must be one of the
5034 HTTP redirect status codes: 301, 302, 303, 307, 308.
5035 When no status is given Squid will use 302.
5037 OK rewrite-url="..."
5038 Rewrite the URL to the one supplied in 'rewrite-url='.
5039 The new URL is fetched directly by Squid and returned to
5040 the client as the response to its request.
5043 When neither of url= and rewrite-url= are sent Squid does
5047 Do not change the URL.
5050 An internal error occurred in the helper, preventing
5051 a result being identified. The 'message=' key name is
5052 reserved for delivering a log message.
5055 In addition to the above kv-pairs Squid also understands the following
5056 optional kv-pairs received from URL rewriters:
5058 Associates a TAG with the client TCP connection.
5059 The TAG is treated as a regular annotation but persists across
5060 future requests on the client connection rather than just the
5061 current request. A helper may update the TAG during subsequent
5062 requests be returning a new kv-pair.
5064 When using the concurrency= option the protocol is changed by
5065 introducing a query channel tag in front of the request/response.
5066 The query channel tag is a number between 0 and concurrency-1.
5067 This value must be echoed back unchanged to Squid as the first part
5068 of the response relating to its request.
5070 WARNING: URL re-writing ability should be avoided whenever possible.
5071 Use the URL redirect form of response instead.
5073 Re-write creates a difference in the state held by the client
5074 and server. Possibly causing confusion when the server response
5075 contains snippets of its view state. Embeded URLs, response
5076 and content Location headers, etc. are not re-written by this
5079 By default, a URL rewriter is not used.
5082 NAME: url_rewrite_children redirect_children
5083 TYPE: HelperChildConfig
5084 DEFAULT: 20 startup=0 idle=1 concurrency=0
5085 LOC: Config.redirectChildren
5087 The maximum number of redirector processes to spawn. If you limit
5088 it too few Squid will have to wait for them to process a backlog of
5089 URLs, slowing it down. If you allow too many they will use RAM
5090 and other system resources noticably.
5092 The startup= and idle= options allow some measure of skew in your
5097 Sets a minimum of how many processes are to be spawned when Squid
5098 starts or reconfigures. When set to zero the first request will
5099 cause spawning of the first child process to handle it.
5101 Starting too few will cause an initial slowdown in traffic as Squid
5102 attempts to simultaneously spawn enough processes to cope.
5106 Sets a minimum of how many processes Squid is to try and keep available
5107 at all times. When traffic begins to rise above what the existing
5108 processes can handle this many more will be spawned up to the maximum
5109 configured. A minimum setting of 1 is required.
5113 The number of requests each redirector helper can handle in
5114 parallel. Defaults to 0 which indicates the redirector
5115 is a old-style single threaded redirector.
5117 When this directive is set to a value >= 1 then the protocol
5118 used to communicate with the helper is modified to include
5119 an ID in front of the request/response. The ID from the request
5120 must be echoed back with the response to that request.
5124 Sets the maximum number of queued requests.
5125 If the queued requests exceed queue size and redirector_bypass
5126 configuration option is set, then redirector is bypassed. Otherwise, if
5127 overloading persists squid may abort its operation.
5128 The default value is set to 2*numberofchildren.
5131 NAME: url_rewrite_host_header redirect_rewrites_host_header
5134 LOC: Config.onoff.redir_rewrites_host
5136 To preserve same-origin security policies in browsers and
5137 prevent Host: header forgery by redirectors Squid rewrites
5138 any Host: header in redirected requests.
5140 If you are running an accelerator this may not be a wanted
5141 effect of a redirector. This directive enables you disable
5142 Host: alteration in reverse-proxy traffic.
5144 WARNING: Entries are cached on the result of the URL rewriting
5145 process, so be careful if you have domain-virtual hosts.
5147 WARNING: Squid and other software verifies the URL and Host
5148 are matching, so be careful not to relay through other proxies
5149 or inspecting firewalls with this disabled.
5152 NAME: url_rewrite_access redirector_access
5155 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5156 LOC: Config.accessList.redirector
5158 If defined, this access list specifies which requests are
5159 sent to the redirector processes.
5161 This clause supports both fast and slow acl types.
5162 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5165 NAME: url_rewrite_bypass redirector_bypass
5167 LOC: Config.onoff.redirector_bypass
5170 When this is 'on', a request will not go through the
5171 redirector if all the helpers are busy. If this is 'off'
5172 and the redirector queue grows too large, Squid will exit
5173 with a FATAL error and ask you to increase the number of
5174 redirectors. You should only enable this if the redirectors
5175 are not critical to your caching system. If you use
5176 redirectors for access control, and you enable this option,
5177 users may have access to pages they should not
5178 be allowed to request.
5179 This options sets default queue-size option of the url_rewrite_children
5183 NAME: url_rewrite_extras
5184 TYPE: TokenOrQuotedString
5185 LOC: Config.redirector_extras
5186 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5188 Specifies a string to be append to request line format for the
5189 rewriter helper. "Quoted" format values may contain spaces and
5190 logformat %macros. In theory, any logformat %macro can be used.
5191 In practice, a %macro expands as a dash (-) if the helper request is
5192 sent before the required macro information is available to Squid.
5195 NAME: url_rewrite_timeout
5196 TYPE: UrlHelperTimeout
5197 LOC: Config.onUrlRewriteTimeout
5199 DEFAULT_DOC: Squid waits for the helper response forever
5201 Squid times active requests to redirector. The timeout value and Squid
5202 reaction to a timed out request are configurable using the following
5205 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
5207 supported timeout actions:
5208 fail Squid return a ERR_GATEWAY_FAILURE error page
5210 bypass Do not re-write the URL
5212 retry Send the lookup to the helper again
5214 use_configured_response
5215 Use the <quoted-response> as helper response
5219 OPTIONS FOR STORE ID
5220 -----------------------------------------------------------------------------
5223 NAME: store_id_program storeurl_rewrite_program
5225 LOC: Config.Program.store_id
5228 Specify the location of the executable StoreID helper to use.
5229 Since they can perform almost any function there isn't one included.
5231 For each requested URL, the helper will receive one line with the format
5233 [channel-ID <SP>] URL [<SP> extras]<NL>
5236 After processing the request the helper must reply using the following format:
5238 [channel-ID <SP>] result [<SP> kv-pairs]
5240 The result code can be:
5243 Use the StoreID supplied in 'store-id='.
5246 The default is to use HTTP request URL as the store ID.
5249 An internal error occured in the helper, preventing
5250 a result being identified.
5252 In addition to the above kv-pairs Squid also understands the following
5253 optional kv-pairs received from URL rewriters:
5255 Associates a TAG with the client TCP connection.
5256 Please see url_rewrite_program related documentation for this
5259 Helper programs should be prepared to receive and possibly ignore
5260 additional whitespace-separated tokens on each input line.
5262 When using the concurrency= option the protocol is changed by
5263 introducing a query channel tag in front of the request/response.
5264 The query channel tag is a number between 0 and concurrency-1.
5265 This value must be echoed back unchanged to Squid as the first part
5266 of the response relating to its request.
5268 NOTE: when using StoreID refresh_pattern will apply to the StoreID
5269 returned from the helper and not the URL.
5271 WARNING: Wrong StoreID value returned by a careless helper may result
5272 in the wrong cached response returned to the user.
5274 By default, a StoreID helper is not used.
5277 NAME: store_id_extras
5278 TYPE: TokenOrQuotedString
5279 LOC: Config.storeId_extras
5280 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5282 Specifies a string to be append to request line format for the
5283 StoreId helper. "Quoted" format values may contain spaces and
5284 logformat %macros. In theory, any logformat %macro can be used.
5285 In practice, a %macro expands as a dash (-) if the helper request is
5286 sent before the required macro information is available to Squid.
5289 NAME: store_id_children storeurl_rewrite_children
5290 TYPE: HelperChildConfig
5291 DEFAULT: 20 startup=0 idle=1 concurrency=0
5292 LOC: Config.storeIdChildren
5294 The maximum number of StoreID helper processes to spawn. If you limit
5295 it too few Squid will have to wait for them to process a backlog of
5296 requests, slowing it down. If you allow too many they will use RAM
5297 and other system resources noticably.
5299 The startup= and idle= options allow some measure of skew in your
5304 Sets a minimum of how many processes are to be spawned when Squid
5305 starts or reconfigures. When set to zero the first request will
5306 cause spawning of the first child process to handle it.
5308 Starting too few will cause an initial slowdown in traffic as Squid
5309 attempts to simultaneously spawn enough processes to cope.
5313 Sets a minimum of how many processes Squid is to try and keep available
5314 at all times. When traffic begins to rise above what the existing
5315 processes can handle this many more will be spawned up to the maximum
5316 configured. A minimum setting of 1 is required.
5320 The number of requests each storeID helper can handle in
5321 parallel. Defaults to 0 which indicates the helper
5322 is a old-style single threaded program.
5324 When this directive is set to a value >= 1 then the protocol
5325 used to communicate with the helper is modified to include
5326 an ID in front of the request/response. The ID from the request
5327 must be echoed back with the response to that request.
5331 Sets the maximum number of queued requests.
5332 If the queued requests exceed queue size and store_id_bypass
5333 configuration option is set, then storeID helper is bypassed. Otherwise,
5334 if overloading persists squid may abort its operation.
5335 The default value is set to 2*numberofchildren.
5338 NAME: store_id_access storeurl_rewrite_access
5341 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5342 LOC: Config.accessList.store_id
5344 If defined, this access list specifies which requests are
5345 sent to the StoreID processes. By default all requests
5348 This clause supports both fast and slow acl types.
5349 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5352 NAME: store_id_bypass storeurl_rewrite_bypass
5354 LOC: Config.onoff.store_id_bypass
5357 When this is 'on', a request will not go through the
5358 helper if all helpers are busy. If this is 'off'
5359 and the helper queue grows too large, Squid will exit
5360 with a FATAL error and ask you to increase the number of
5361 helpers. You should only enable this if the helperss
5362 are not critical to your caching system. If you use
5363 helpers for critical caching components, and you enable this
5364 option, users may not get objects from cache.
5365 This options sets default queue-size option of the store_id_children
5370 OPTIONS FOR TUNING THE CACHE
5371 -----------------------------------------------------------------------------
5374 NAME: cache no_cache
5377 DEFAULT_DOC: By default, this directive is unused and has no effect.
5378 LOC: Config.accessList.noCache
5380 Requests denied by this directive will not be served from the cache
5381 and their responses will not be stored in the cache. This directive
5382 has no effect on other transactions and on already cached responses.
5384 This clause supports both fast and slow acl types.
5385 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5387 This and the two other similar caching directives listed below are
5388 checked at different transaction processing stages, have different
5389 access to response information, affect different cache operations,
5390 and differ in slow ACLs support:
5392 * cache: Checked before Squid makes a hit/miss determination.
5393 No access to reply information!
5394 Denies both serving a hit and storing a miss.
5395 Supports both fast and slow ACLs.
5396 * send_hit: Checked after a hit was detected.
5397 Has access to reply (hit) information.
5398 Denies serving a hit only.
5399 Supports fast ACLs only.
5400 * store_miss: Checked before storing a cachable miss.
5401 Has access to reply (miss) information.
5402 Denies storing a miss only.
5403 Supports fast ACLs only.
5405 If you are not sure which of the three directives to use, apply the
5406 following decision logic:
5408 * If your ACL(s) are of slow type _and_ need response info, redesign.
5409 Squid does not support that particular combination at this time.
5411 * If your directive ACL(s) are of slow type, use "cache"; and/or
5412 * if your directive ACL(s) need no response info, use "cache".
5414 * If you do not want the response cached, use store_miss; and/or
5415 * if you do not want a hit on a cached response, use send_hit.
5421 DEFAULT_DOC: By default, this directive is unused and has no effect.
5422 LOC: Config.accessList.sendHit
5424 Responses denied by this directive will not be served from the cache
5425 (but may still be cached, see store_miss). This directive has no
5426 effect on the responses it allows and on the cached objects.
5428 Please see the "cache" directive for a summary of differences among
5429 store_miss, send_hit, and cache directives.
5431 Unlike the "cache" directive, send_hit only supports fast acl
5432 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5436 # apply custom Store ID mapping to some URLs
5437 acl MapMe dstdomain .c.example.com
5438 store_id_program ...
5439 store_id_access allow MapMe
5441 # but prevent caching of special responses
5442 # such as 302 redirects that cause StoreID loops
5443 acl Ordinary http_status 200-299
5444 store_miss deny MapMe !Ordinary
5446 # and do not serve any previously stored special responses
5447 # from the cache (in case they were already cached before
5448 # the above store_miss rule was in effect).
5449 send_hit deny MapMe !Ordinary
5455 DEFAULT_DOC: By default, this directive is unused and has no effect.
5456 LOC: Config.accessList.storeMiss
5458 Responses denied by this directive will not be cached (but may still
5459 be served from the cache, see send_hit). This directive has no
5460 effect on the responses it allows and on the already cached responses.
5462 Please see the "cache" directive for a summary of differences among
5463 store_miss, send_hit, and cache directives. See the
5464 send_hit directive for a usage example.
5466 Unlike the "cache" directive, store_miss only supports fast acl
5467 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5473 LOC: Config.maxStale
5476 This option puts an upper limit on how stale content Squid
5477 will serve from the cache if cache validation fails.
5478 Can be overriden by the refresh_pattern max-stale option.
5481 NAME: refresh_pattern
5482 TYPE: refreshpattern
5486 usage: refresh_pattern [-i] regex min percent max [options]
5488 By default, regular expressions are CASE-SENSITIVE. To make
5489 them case-insensitive, use the -i option.
5491 'Min' is the time (in minutes) an object without an explicit
5492 expiry time should be considered fresh. The recommended
5493 value is 0, any higher values may cause dynamic applications
5494 to be erroneously cached unless the application designer
5495 has taken the appropriate actions.
5497 'Percent' is a percentage of the objects age (time since last
5498 modification age) an object without explicit expiry time
5499 will be considered fresh.
5501 'Max' is an upper limit on how long objects without an explicit
5502 expiry time will be considered fresh.
5504 options: override-expire
5514 override-expire enforces min age even if the server
5515 sent an explicit expiry time (e.g., with the
5516 Expires: header or Cache-Control: max-age). Doing this
5517 VIOLATES the HTTP standard. Enabling this feature
5518 could make you liable for problems which it causes.
5520 Note: override-expire does not enforce staleness - it only extends
5521 freshness / min. If the server returns a Expires time which
5522 is longer than your max time, Squid will still consider
5523 the object fresh for that period of time.
5525 override-lastmod enforces min age even on objects
5526 that were modified recently.
5528 reload-into-ims changes a client no-cache or ``reload''
5529 request for a cached entry into a conditional request using
5530 If-Modified-Since and/or If-None-Match headers, provided the
5531 cached entry has a Last-Modified and/or a strong ETag header.
5532 Doing this VIOLATES the HTTP standard. Enabling this feature
5533 could make you liable for problems which it causes.
5535 ignore-reload ignores a client no-cache or ``reload''
5536 header. Doing this VIOLATES the HTTP standard. Enabling
5537 this feature could make you liable for problems which
5540 ignore-no-store ignores any ``Cache-control: no-store''
5541 headers received from a server. Doing this VIOLATES
5542 the HTTP standard. Enabling this feature could make you
5543 liable for problems which it causes.
5545 ignore-private ignores any ``Cache-control: private''
5546 headers received from a server. Doing this VIOLATES
5547 the HTTP standard. Enabling this feature could make you
5548 liable for problems which it causes.
5550 refresh-ims causes squid to contact the origin server
5551 when a client issues an If-Modified-Since request. This
5552 ensures that the client will receive an updated version
5553 if one is available.
5555 store-stale stores responses even if they don't have explicit
5556 freshness or a validator (i.e., Last-Modified or an ETag)
5557 present, or if they're already stale. By default, Squid will
5558 not cache such responses because they usually can't be
5559 reused. Note that such responses will be stale by default.
5561 max-stale=NN provide a maximum staleness factor. Squid won't
5562 serve objects more stale than this even if it failed to
5563 validate the object. Default: use the max_stale global limit.
5565 Basically a cached object is:
5567 FRESH if expire > now, else STALE
5569 FRESH if lm-factor < percent, else STALE
5573 The refresh_pattern lines are checked in the order listed here.
5574 The first entry which matches is used. If none of the entries
5575 match the default will be used.
5577 Note, you must uncomment all the default lines if you want
5578 to change one. The default setting is only active if none is
5584 # Add any of your own refresh_pattern entries above these.
5586 refresh_pattern ^ftp: 1440 20% 10080
5587 refresh_pattern ^gopher: 1440 0% 1440
5588 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5589 refresh_pattern . 0 20% 4320
5593 NAME: quick_abort_min
5597 LOC: Config.quickAbort.min
5600 NAME: quick_abort_max
5604 LOC: Config.quickAbort.max
5607 NAME: quick_abort_pct
5611 LOC: Config.quickAbort.pct
5613 The cache by default continues downloading aborted requests
5614 which are almost completed (less than 16 KB remaining). This
5615 may be undesirable on slow (e.g. SLIP) links and/or very busy
5616 caches. Impatient users may tie up file descriptors and
5617 bandwidth by repeatedly requesting and immediately aborting
5620 When the user aborts a request, Squid will check the
5621 quick_abort values to the amount of data transferred until
5624 If the transfer has less than 'quick_abort_min' KB remaining,
5625 it will finish the retrieval.
5627 If the transfer has more than 'quick_abort_max' KB remaining,
5628 it will abort the retrieval.
5630 If more than 'quick_abort_pct' of the transfer has completed,
5631 it will finish the retrieval.
5633 If you do not want any retrieval to continue after the client
5634 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5637 If you want retrievals to always continue if they are being
5638 cached set 'quick_abort_min' to '-1 KB'.
5641 NAME: read_ahead_gap
5642 COMMENT: buffer-size
5644 LOC: Config.readAheadGap
5647 The amount of data the cache will buffer ahead of what has been
5648 sent to the client when retrieving an object from another server.
5652 IFDEF: USE_HTTP_VIOLATIONS
5655 LOC: Config.negativeTtl
5658 Set the Default Time-to-Live (TTL) for failed requests.
5659 Certain types of failures (such as "connection refused" and
5660 "404 Not Found") are able to be negatively-cached for a short time.
5661 Modern web servers should provide Expires: header, however if they
5662 do not this can provide a minimum TTL.
5663 The default is not to cache errors with unknown expiry details.
5665 Note that this is different from negative caching of DNS lookups.
5667 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5668 this feature could make you liable for problems which it
5672 NAME: positive_dns_ttl
5675 LOC: Config.positiveDnsTtl
5678 Upper limit on how long Squid will cache positive DNS responses.
5679 Default is 6 hours (360 minutes). This directive must be set
5680 larger than negative_dns_ttl.
5683 NAME: negative_dns_ttl
5686 LOC: Config.negativeDnsTtl
5689 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5690 This also sets the lower cache limit on positive lookups.
5691 Minimum value is 1 second, and it is not recommendable to go
5692 much below 10 seconds.
5695 NAME: range_offset_limit
5696 COMMENT: size [acl acl...]
5698 LOC: Config.rangeOffsetLimit
5701 usage: (size) [units] [[!]aclname]
5703 Sets an upper limit on how far (number of bytes) into the file
5704 a Range request may be to cause Squid to prefetch the whole file.
5705 If beyond this limit, Squid forwards the Range request as it is and
5706 the result is NOT cached.
5708 This is to stop a far ahead range request (lets say start at 17MB)
5709 from making Squid fetch the whole object up to that point before
5710 sending anything to the client.
5712 Multiple range_offset_limit lines may be specified, and they will
5713 be searched from top to bottom on each request until a match is found.
5714 The first match found will be used. If no line matches a request, the
5715 default limit of 0 bytes will be used.
5717 'size' is the limit specified as a number of units.
5719 'units' specifies whether to use bytes, KB, MB, etc.
5720 If no units are specified bytes are assumed.
5722 A size of 0 causes Squid to never fetch more than the
5723 client requested. (default)
5725 A size of 'none' causes Squid to always fetch the object from the
5726 beginning so it may cache the result. (2.0 style)
5728 'aclname' is the name of a defined ACL.
5730 NP: Using 'none' as the byte value here will override any quick_abort settings
5731 that may otherwise apply to the range request. The range request will
5732 be fully fetched from start to finish regardless of the client
5733 actions. This affects bandwidth usage.
5736 NAME: minimum_expiry_time
5739 LOC: Config.minimum_expiry_time
5742 The minimum caching time according to (Expires - Date)
5743 headers Squid honors if the object can't be revalidated.
5744 The default is 60 seconds.
5746 In reverse proxy environments it might be desirable to honor
5747 shorter object lifetimes. It is most likely better to make
5748 your server return a meaningful Last-Modified header however.
5750 In ESI environments where page fragments often have short
5751 lifetimes, this will often be best set to 0.
5754 NAME: store_avg_object_size
5758 LOC: Config.Store.avgObjectSize
5760 Average object size, used to estimate number of objects your
5761 cache can hold. The default is 13 KB.
5763 This is used to pre-seed the cache index memory allocation to
5764 reduce expensive reallocate operations while handling clients
5765 traffic. Too-large values may result in memory allocation during
5766 peak traffic, too-small values will result in wasted memory.
5768 Check the cache manager 'info' report metrics for the real
5769 object sizes seen by your Squid before tuning this.
5772 NAME: store_objects_per_bucket
5775 LOC: Config.Store.objectsPerBucket
5777 Target number of objects per bucket in the store hash table.
5778 Lowering this value increases the total number of buckets and
5779 also the storage maintenance rate. The default is 20.
5784 -----------------------------------------------------------------------------
5787 NAME: request_header_max_size
5791 LOC: Config.maxRequestHeaderSize
5793 This specifies the maximum size for HTTP headers in a request.
5794 Request headers are usually relatively small (about 512 bytes).
5795 Placing a limit on the request header size will catch certain
5796 bugs (for example with persistent connections) and possibly
5797 buffer-overflow or denial-of-service attacks.
5800 NAME: reply_header_max_size
5804 LOC: Config.maxReplyHeaderSize
5806 This specifies the maximum size for HTTP headers in a reply.
5807 Reply headers are usually relatively small (about 512 bytes).
5808 Placing a limit on the reply header size will catch certain
5809 bugs (for example with persistent connections) and possibly
5810 buffer-overflow or denial-of-service attacks.
5813 NAME: request_body_max_size
5817 DEFAULT_DOC: No limit.
5818 LOC: Config.maxRequestBodySize
5820 This specifies the maximum size for an HTTP request body.
5821 In other words, the maximum size of a PUT/POST request.
5822 A user who attempts to send a request with a body larger
5823 than this limit receives an "Invalid Request" error message.
5824 If you set this parameter to a zero (the default), there will
5825 be no limit imposed.
5827 See also client_request_buffer_max_size for an alternative
5828 limitation on client uploads which can be configured.
5831 NAME: client_request_buffer_max_size
5835 LOC: Config.maxRequestBufferSize
5837 This specifies the maximum buffer size of a client request.
5838 It prevents squid eating too much memory when somebody uploads
5843 IFDEF: USE_HTTP_VIOLATIONS
5846 DEFAULT_DOC: Obey RFC 2616.
5847 LOC: Config.accessList.brokenPosts
5849 A list of ACL elements which, if matched, causes Squid to send
5850 an extra CRLF pair after the body of a PUT/POST request.
5852 Some HTTP servers has broken implementations of PUT/POST,
5853 and rely on an extra CRLF pair sent by some WWW clients.
5855 Quote from RFC2616 section 4.1 on this matter:
5857 Note: certain buggy HTTP/1.0 client implementations generate an
5858 extra CRLF's after a POST request. To restate what is explicitly
5859 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5860 a request with an extra CRLF.
5862 This clause only supports fast acl types.
5863 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5866 acl buggy_server url_regex ^http://....
5867 broken_posts allow buggy_server
5870 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5873 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5875 LOC: Adaptation::Config::use_indirect_client
5877 Controls whether the indirect client IP address (instead of the direct
5878 client IP address) is passed to adaptation services.
5880 See also: follow_x_forwarded_for adaptation_send_client_ip
5884 IFDEF: USE_HTTP_VIOLATIONS
5888 LOC: Config.onoff.via
5890 If set (default), Squid will include a Via header in requests and
5891 replies as required by RFC2616.
5897 LOC: Config.onoff.ie_refresh
5900 Microsoft Internet Explorer up until version 5.5 Service
5901 Pack 1 has an issue with transparent proxies, wherein it
5902 is impossible to force a refresh. Turning this on provides
5903 a partial fix to the problem, by causing all IMS-REFRESH
5904 requests from older IE versions to check the origin server
5905 for fresh content. This reduces hit ratio by some amount
5906 (~10% in my experience), but allows users to actually get
5907 fresh content when they want it. Note because Squid
5908 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5909 of 5.5 is unchanged from old versions of Squid (i.e. a
5910 forced refresh is impossible). Newer versions of IE will,
5911 hopefully, continue to have the new behavior and will be
5912 handled based on that assumption. This option defaults to
5913 the old Squid behavior, which is better for hit ratios but
5914 worse for clients using IE, if they need to be able to
5915 force fresh content.
5918 NAME: vary_ignore_expire
5921 LOC: Config.onoff.vary_ignore_expire
5924 Many HTTP servers supporting Vary gives such objects
5925 immediate expiry time with no cache-control header
5926 when requested by a HTTP/1.0 client. This option
5927 enables Squid to ignore such expiry times until
5928 HTTP/1.1 is fully implemented.
5930 WARNING: If turned on this may eventually cause some
5931 varying objects not intended for caching to get cached.
5934 NAME: request_entities
5936 LOC: Config.onoff.request_entities
5939 Squid defaults to deny GET and HEAD requests with request entities,
5940 as the meaning of such requests are undefined in the HTTP standard
5941 even if not explicitly forbidden.
5943 Set this directive to on if you have clients which insists
5944 on sending request entities in GET or HEAD requests. But be warned
5945 that there is server software (both proxies and web servers) which
5946 can fail to properly process this kind of request which may make you
5947 vulnerable to cache pollution attacks if enabled.
5950 NAME: request_header_access
5951 IFDEF: USE_HTTP_VIOLATIONS
5952 TYPE: http_header_access
5953 LOC: Config.request_header_access
5955 DEFAULT_DOC: No limits.
5957 Usage: request_header_access header_name allow|deny [!]aclname ...
5959 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5960 this feature could make you liable for problems which it
5963 This option replaces the old 'anonymize_headers' and the
5964 older 'http_anonymizer' option with something that is much
5965 more configurable. A list of ACLs for each header name allows
5966 removal of specific header fields under specific conditions.
5968 This option only applies to outgoing HTTP request headers (i.e.,
5969 headers sent by Squid to the next HTTP hop such as a cache peer
5970 or an origin server). The option has no effect during cache hit
5971 detection. The equivalent adaptation vectoring point in ICAP
5972 terminology is post-cache REQMOD.
5974 The option is applied to individual outgoing request header
5975 fields. For each request header field F, Squid uses the first
5976 qualifying sets of request_header_access rules:
5978 1. Rules with header_name equal to F's name.
5979 2. Rules with header_name 'Other', provided F's name is not
5980 on the hard-coded list of commonly used HTTP header names.
5981 3. Rules with header_name 'All'.
5983 Within that qualifying rule set, rule ACLs are checked as usual.
5984 If ACLs of an "allow" rule match, the header field is allowed to
5985 go through as is. If ACLs of a "deny" rule match, the header is
5986 removed and request_header_replace is then checked to identify
5987 if the removed header has a replacement. If no rules within the
5988 set have matching ACLs, the header field is left as is.
5990 For example, to achieve the same behavior as the old
5991 'http_anonymizer standard' option, you should use:
5993 request_header_access From deny all
5994 request_header_access Referer deny all
5995 request_header_access User-Agent deny all
5997 Or, to reproduce the old 'http_anonymizer paranoid' feature
6000 request_header_access Authorization allow all
6001 request_header_access Proxy-Authorization allow all
6002 request_header_access Cache-Control allow all
6003 request_header_access Content-Length allow all
6004 request_header_access Content-Type allow all
6005 request_header_access Date allow all
6006 request_header_access Host allow all
6007 request_header_access If-Modified-Since allow all
6008 request_header_access Pragma allow all
6009 request_header_access Accept allow all
6010 request_header_access Accept-Charset allow all
6011 request_header_access Accept-Encoding allow all
6012 request_header_access Accept-Language allow all
6013 request_header_access Connection allow all
6014 request_header_access All deny all
6016 HTTP reply headers are controlled with the reply_header_access directive.
6018 By default, all headers are allowed (no anonymizing is performed).
6021 NAME: reply_header_access
6022 IFDEF: USE_HTTP_VIOLATIONS
6023 TYPE: http_header_access
6024 LOC: Config.reply_header_access
6026 DEFAULT_DOC: No limits.
6028 Usage: reply_header_access header_name allow|deny [!]aclname ...
6030 WARNING: Doing this VIOLATES the HTTP standard. Enabling
6031 this feature could make you liable for problems which it
6034 This option only applies to reply headers, i.e., from the
6035 server to the client.
6037 This is the same as request_header_access, but in the other
6038 direction. Please see request_header_access for detailed
6041 For example, to achieve the same behavior as the old
6042 'http_anonymizer standard' option, you should use:
6044 reply_header_access Server deny all
6045 reply_header_access WWW-Authenticate deny all
6046 reply_header_access Link deny all
6048 Or, to reproduce the old 'http_anonymizer paranoid' feature
6051 reply_header_access Allow allow all
6052 reply_header_access WWW-Authenticate allow all
6053 reply_header_access Proxy-Authenticate allow all
6054 reply_header_access Cache-Control allow all
6055 reply_header_access Content-Encoding allow all
6056 reply_header_access Content-Length allow all
6057 reply_header_access Content-Type allow all
6058 reply_header_access Date allow all
6059 reply_header_access Expires allow all
6060 reply_header_access Last-Modified allow all
6061 reply_header_access Location allow all
6062 reply_header_access Pragma allow all
6063 reply_header_access Content-Language allow all
6064 reply_header_access Retry-After allow all
6065 reply_header_access Title allow all
6066 reply_header_access Content-Disposition allow all
6067 reply_header_access Connection allow all
6068 reply_header_access All deny all
6070 HTTP request headers are controlled with the request_header_access directive.
6072 By default, all headers are allowed (no anonymizing is
6076 NAME: request_header_replace header_replace
6077 IFDEF: USE_HTTP_VIOLATIONS
6078 TYPE: http_header_replace
6079 LOC: Config.request_header_access
6082 Usage: request_header_replace header_name message
6083 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
6085 This option allows you to change the contents of headers
6086 denied with request_header_access above, by replacing them
6087 with some fixed string.
6089 This only applies to request headers, not reply headers.
6091 By default, headers are removed if denied.
6094 NAME: reply_header_replace
6095 IFDEF: USE_HTTP_VIOLATIONS
6096 TYPE: http_header_replace
6097 LOC: Config.reply_header_access
6100 Usage: reply_header_replace header_name message
6101 Example: reply_header_replace Server Foo/1.0
6103 This option allows you to change the contents of headers
6104 denied with reply_header_access above, by replacing them
6105 with some fixed string.
6107 This only applies to reply headers, not request headers.
6109 By default, headers are removed if denied.
6112 NAME: request_header_add
6113 TYPE: HeaderWithAclList
6114 LOC: Config.request_header_add
6117 Usage: request_header_add field-name field-value [ acl ... ]
6118 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6120 This option adds header fields to outgoing HTTP requests (i.e.,
6121 request headers sent by Squid to the next HTTP hop such as a
6122 cache peer or an origin server). The option has no effect during
6123 cache hit detection. The equivalent adaptation vectoring point
6124 in ICAP terminology is post-cache REQMOD.
6126 Field-name is a token specifying an HTTP header name. If a
6127 standard HTTP header name is used, Squid does not check whether
6128 the new header conflicts with any existing headers or violates
6129 HTTP rules. If the request to be modified already contains a
6130 field with the same name, the old field is preserved but the
6131 header field values are not merged.
6133 Field-value is either a token or a quoted string. If quoted
6134 string format is used, then the surrounding quotes are removed
6135 while escape sequences and %macros are processed.
6137 One or more Squid ACLs may be specified to restrict header
6138 injection to matching requests. As always in squid.conf, all
6139 ACLs in the ACL list must be satisfied for the insertion to
6140 happen. The request_header_add supports fast ACLs only.
6142 See also: reply_header_add.
6145 NAME: reply_header_add
6146 TYPE: HeaderWithAclList
6147 LOC: Config.reply_header_add
6150 Usage: reply_header_add field-name field-value [ acl ... ]
6151 Example: reply_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6153 This option adds header fields to outgoing HTTP responses (i.e., response
6154 headers delivered by Squid to the client). This option has no effect on
6155 cache hit detection. The equivalent adaptation vectoring point in
6156 ICAP terminology is post-cache RESPMOD. This option does not apply to
6157 successful CONNECT replies.
6159 Field-name is a token specifying an HTTP header name. If a
6160 standard HTTP header name is used, Squid does not check whether
6161 the new header conflicts with any existing headers or violates
6162 HTTP rules. If the response to be modified already contains a
6163 field with the same name, the old field is preserved but the
6164 header field values are not merged.
6166 Field-value is either a token or a quoted string. If quoted
6167 string format is used, then the surrounding quotes are removed
6168 while escape sequences and %macros are processed.
6170 One or more Squid ACLs may be specified to restrict header
6171 injection to matching responses. As always in squid.conf, all
6172 ACLs in the ACL list must be satisfied for the insertion to
6173 happen. The reply_header_add option supports fast ACLs only.
6175 See also: request_header_add.
6183 This option used to log custom information about the master
6184 transaction. For example, an admin may configure Squid to log
6185 which "user group" the transaction belongs to, where "user group"
6186 will be determined based on a set of ACLs and not [just]
6187 authentication information.
6188 Values of key/value pairs can be logged using %{key}note macros:
6190 note key value acl ...
6191 logformat myFormat ... %{key}note ...
6194 NAME: relaxed_header_parser
6195 COMMENT: on|off|warn
6197 LOC: Config.onoff.relaxed_header_parser
6200 In the default "on" setting Squid accepts certain forms
6201 of non-compliant HTTP messages where it is unambiguous
6202 what the sending application intended even if the message
6203 is not correctly formatted. The messages is then normalized
6204 to the correct form when forwarded by Squid.
6206 If set to "warn" then a warning will be emitted in cache.log
6207 each time such HTTP error is encountered.
6209 If set to "off" then such HTTP errors will cause the request
6210 or response to be rejected.
6213 NAME: collapsed_forwarding
6216 LOC: Config.onoff.collapsed_forwarding
6219 This option controls whether Squid is allowed to merge multiple
6220 potentially cachable requests for the same URI before Squid knows
6221 whether the response is going to be cachable.
6223 This feature is disabled by default: Enabling collapsed forwarding
6224 needlessly delays forwarding requests that look cachable (when they are
6225 collapsed) but then need to be forwarded individually anyway because
6226 they end up being for uncachable content. However, in some cases, such
6227 as accelleration of highly cachable content with periodic or groupped
6228 expiration times, the gains from collapsing [large volumes of
6229 simultenous refresh requests] outweigh losses from such delays.
6232 NAME: collapsed_forwarding_shared_entries_limit
6233 COMMENT: (number of entries)
6235 LOC: Config.collapsed_forwarding_shared_entries_limit
6238 This limits the size of a table used for sharing information
6239 about collapsible entries among SMP workers. Limiting sharing
6240 too much results in cache content duplication and missed
6241 collapsing opportunities. Using excessively large values
6242 wastes shared memory.
6244 The limit should be significantly larger then the number of
6245 concurrent collapsible entries one wants to share. For a cache
6246 that handles less than 5000 concurrent requests, the default
6247 setting of 16384 should be plenty.
6249 If the limit is set to zero, it disables sharing of collapsed
6250 forwarding between SMP workers.
6255 -----------------------------------------------------------------------------
6258 NAME: forward_timeout
6261 LOC: Config.Timeout.forward
6264 This parameter specifies how long Squid should at most attempt in
6265 finding a forwarding path for the request before giving up.
6268 NAME: connect_timeout
6271 LOC: Config.Timeout.connect
6274 This parameter specifies how long to wait for the TCP connect to
6275 the requested server or peer to complete before Squid should
6276 attempt to find another path where to forward the request.
6279 NAME: peer_connect_timeout
6282 LOC: Config.Timeout.peer_connect
6285 This parameter specifies how long to wait for a pending TCP
6286 connection to a peer cache. The default is 30 seconds. You
6287 may also set different timeout values for individual neighbors
6288 with the 'connect-timeout' option on a 'cache_peer' line.
6294 LOC: Config.Timeout.read
6297 Applied on peer server connections.
6299 After each successful read(), the timeout will be extended by this
6300 amount. If no data is read again after this amount of time,
6301 the request is aborted and logged with ERR_READ_TIMEOUT.
6303 The default is 15 minutes.
6309 LOC: Config.Timeout.write
6312 This timeout is tracked for all connections that have data
6313 available for writing and are waiting for the socket to become
6314 ready. After each successful write, the timeout is extended by
6315 the configured amount. If Squid has data to write but the
6316 connection is not ready for the configured duration, the
6317 transaction associated with the connection is terminated. The
6318 default is 15 minutes.
6321 NAME: request_timeout
6323 LOC: Config.Timeout.request
6326 How long to wait for complete HTTP request headers after initial
6327 connection establishment.
6330 NAME: request_start_timeout
6332 LOC: Config.Timeout.request_start_timeout
6335 How long to wait for the first request byte after initial
6336 connection establishment.
6339 NAME: client_idle_pconn_timeout persistent_request_timeout
6341 LOC: Config.Timeout.clientIdlePconn
6344 How long to wait for the next HTTP request on a persistent
6345 client connection after the previous request completes.
6348 NAME: ftp_client_idle_timeout
6350 LOC: Config.Timeout.ftpClientIdle
6353 How long to wait for an FTP request on a connection to Squid ftp_port.
6354 Many FTP clients do not deal with idle connection closures well,
6355 necessitating a longer default timeout than client_idle_pconn_timeout
6356 used for incoming HTTP requests.
6359 NAME: client_lifetime
6362 LOC: Config.Timeout.lifetime
6365 The maximum amount of time a client (browser) is allowed to
6366 remain connected to the cache process. This protects the Cache
6367 from having a lot of sockets (and hence file descriptors) tied up
6368 in a CLOSE_WAIT state from remote clients that go away without
6369 properly shutting down (either because of a network failure or
6370 because of a poor client implementation). The default is one
6373 NOTE: The default value is intended to be much larger than any
6374 client would ever need to be connected to your cache. You
6375 should probably change client_lifetime only as a last resort.
6376 If you seem to have many client connections tying up
6377 filedescriptors, we recommend first tuning the read_timeout,
6378 request_timeout, persistent_request_timeout and quick_abort values.
6381 NAME: pconn_lifetime
6384 LOC: Config.Timeout.pconnLifetime
6387 Desired maximum lifetime of a persistent connection.
6388 When set, Squid will close a now-idle persistent connection that
6389 exceeded configured lifetime instead of moving the connection into
6390 the idle connection pool (or equivalent). No effect on ongoing/active
6391 transactions. Connection lifetime is the time period from the
6392 connection acceptance or opening time until "now".
6394 This limit is useful in environments with long-lived connections
6395 where Squid configuration or environmental factors change during a
6396 single connection lifetime. If unrestricted, some connections may
6397 last for hours and even days, ignoring those changes that should
6398 have affected their behavior or their existence.
6400 Currently, a new lifetime value supplied via Squid reconfiguration
6401 has no effect on already idle connections unless they become busy.
6403 When set to '0' this limit is not used.
6406 NAME: half_closed_clients
6408 LOC: Config.onoff.half_closed_clients
6411 Some clients may shutdown the sending side of their TCP
6412 connections, while leaving their receiving sides open. Sometimes,
6413 Squid can not tell the difference between a half-closed and a
6414 fully-closed TCP connection.
6416 By default, Squid will immediately close client connections when
6417 read(2) returns "no more data to read."
6419 Change this option to 'on' and Squid will keep open connections
6420 until a read(2) or write(2) on the socket returns an error.
6421 This may show some benefits for reverse proxies. But if not
6422 it is recommended to leave OFF.
6425 NAME: server_idle_pconn_timeout pconn_timeout
6427 LOC: Config.Timeout.serverIdlePconn
6430 Timeout for idle persistent connections to servers and other
6437 LOC: Ident::TheConfig.timeout
6440 Maximum time to wait for IDENT lookups to complete.
6442 If this is too high, and you enabled IDENT lookups from untrusted
6443 users, you might be susceptible to denial-of-service by having
6444 many ident requests going at once.
6447 NAME: shutdown_lifetime
6450 LOC: Config.shutdownLifetime
6453 When SIGTERM or SIGHUP is received, the cache is put into
6454 "shutdown pending" mode until all active sockets are closed.
6455 This value is the lifetime to set for all open descriptors
6456 during shutdown mode. Any active clients after this many
6457 seconds will receive a 'timeout' message.
6461 ADMINISTRATIVE PARAMETERS
6462 -----------------------------------------------------------------------------
6468 LOC: Config.adminEmail
6470 Email-address of local cache manager who will receive
6471 mail if the cache dies. The default is "webmaster".
6477 LOC: Config.EmailFrom
6479 From: email-address for mail sent when the cache dies.
6480 The default is to use 'squid@unique_hostname'.
6482 See also: unique_hostname directive.
6488 LOC: Config.EmailProgram
6490 Email program used to send mail if the cache dies.
6491 The default is "mail". The specified program must comply
6492 with the standard Unix mail syntax:
6493 mail-program recipient < mailfile
6495 Optional command line options can be specified.
6498 NAME: cache_effective_user
6500 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
6501 LOC: Config.effectiveUser
6503 If you start Squid as root, it will change its effective/real
6504 UID/GID to the user specified below. The default is to change
6505 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
6506 see also; cache_effective_group
6509 NAME: cache_effective_group
6512 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
6513 LOC: Config.effectiveGroup
6515 Squid sets the GID to the effective user's default group ID
6516 (taken from the password file) and supplementary group list
6517 from the groups membership.
6519 If you want Squid to run with a specific GID regardless of
6520 the group memberships of the effective user then set this
6521 to the group (or GID) you want Squid to run as. When set
6522 all other group privileges of the effective user are ignored
6523 and only this GID is effective. If Squid is not started as
6524 root the user starting Squid MUST be member of the specified
6527 This option is not recommended by the Squid Team.
6528 Our preference is for administrators to configure a secure
6529 user account for squid with UID/GID matching system policies.
6532 NAME: httpd_suppress_version_string
6536 LOC: Config.onoff.httpd_suppress_version_string
6538 Suppress Squid version string info in HTTP headers and HTML error pages.
6541 NAME: visible_hostname
6543 LOC: Config.visibleHostname
6545 DEFAULT_DOC: Automatically detect the system host name
6547 If you want to present a special hostname in error messages, etc,
6548 define this. Otherwise, the return value of gethostname()
6549 will be used. If you have multiple caches in a cluster and
6550 get errors about IP-forwarding you must set them to have individual
6551 names with this setting.
6554 NAME: unique_hostname
6556 LOC: Config.uniqueHostname
6558 DEFAULT_DOC: Copy the value from visible_hostname
6560 If you want to have multiple machines with the same
6561 'visible_hostname' you must give each machine a different
6562 'unique_hostname' so forwarding loops can be detected.
6565 NAME: hostname_aliases
6567 LOC: Config.hostnameAliases
6570 A list of other DNS names your cache has.
6578 Minimum umask which should be enforced while the proxy
6579 is running, in addition to the umask set at startup.
6581 For a traditional octal representation of umasks, start
6586 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6587 -----------------------------------------------------------------------------
6589 This section contains parameters for the (optional) cache
6590 announcement service. This service is provided to help
6591 cache administrators locate one another in order to join or
6592 create cache hierarchies.
6594 An 'announcement' message is sent (via UDP) to the registration
6595 service by Squid. By default, the announcement message is NOT
6596 SENT unless you enable it with 'announce_period' below.
6598 The announcement message includes your hostname, plus the
6599 following information from this configuration file:
6605 All current information is processed regularly and made
6606 available on the Web at http://www.ircache.net/Cache/Tracker/.
6609 NAME: announce_period
6611 LOC: Config.Announce.period
6613 DEFAULT_DOC: Announcement messages disabled.
6615 This is how frequently to send cache announcements.
6617 To enable announcing your cache, just set an announce period.
6620 announce_period 1 day
6625 DEFAULT: tracker.ircache.net
6626 LOC: Config.Announce.host
6628 Set the hostname where announce registration messages will be sent.
6630 See also announce_port and announce_file
6636 LOC: Config.Announce.file
6638 The contents of this file will be included in the announce
6639 registration messages.
6645 LOC: Config.Announce.port
6647 Set the port where announce registration messages will be sent.
6649 See also announce_host and announce_file
6653 HTTPD-ACCELERATOR OPTIONS
6654 -----------------------------------------------------------------------------
6657 NAME: httpd_accel_surrogate_id
6660 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6661 LOC: Config.Accel.surrogate_id
6663 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6664 need an identification token to allow control targeting. Because
6665 a farm of surrogates may all perform the same tasks, they may share
6666 an identification token.
6669 NAME: http_accel_surrogate_remote
6673 LOC: Config.onoff.surrogate_is_remote
6675 Remote surrogates (such as those in a CDN) honour the header
6676 "Surrogate-Control: no-store-remote".
6678 Set this to on to have squid behave as a remote surrogate.
6682 IFDEF: USE_SQUID_ESI
6683 COMMENT: libxml2|expat|custom
6685 LOC: ESIParser::Type
6688 ESI markup is not strictly XML compatible. The custom ESI parser
6689 will give higher performance, but cannot handle non ASCII character
6694 DELAY POOL PARAMETERS
6695 -----------------------------------------------------------------------------
6699 TYPE: delay_pool_count
6701 IFDEF: USE_DELAY_POOLS
6704 This represents the number of delay pools to be used. For example,
6705 if you have one class 2 delay pool and one class 3 delays pool, you
6706 have a total of 2 delay pools.
6708 See also delay_parameters, delay_class, delay_access for pool
6709 configuration details.
6713 TYPE: delay_pool_class
6715 IFDEF: USE_DELAY_POOLS
6718 This defines the class of each delay pool. There must be exactly one
6719 delay_class line for each delay pool. For example, to define two
6720 delay pools, one of class 2 and one of class 3, the settings above
6724 delay_pools 4 # 4 delay pools
6725 delay_class 1 2 # pool 1 is a class 2 pool
6726 delay_class 2 3 # pool 2 is a class 3 pool
6727 delay_class 3 4 # pool 3 is a class 4 pool
6728 delay_class 4 5 # pool 4 is a class 5 pool
6730 The delay pool classes are:
6732 class 1 Everything is limited by a single aggregate
6735 class 2 Everything is limited by a single aggregate
6736 bucket as well as an "individual" bucket chosen
6737 from bits 25 through 32 of the IPv4 address.
6739 class 3 Everything is limited by a single aggregate
6740 bucket as well as a "network" bucket chosen
6741 from bits 17 through 24 of the IP address and a
6742 "individual" bucket chosen from bits 17 through
6743 32 of the IPv4 address.
6745 class 4 Everything in a class 3 delay pool, with an
6746 additional limit on a per user basis. This
6747 only takes effect if the username is established
6748 in advance - by forcing authentication in your
6751 class 5 Requests are grouped according their tag (see
6752 external_acl's tag= reply).
6755 Each pool also requires a delay_parameters directive to configure the pool size
6756 and speed limits used whenever the pool is applied to a request. Along with
6757 a set of delay_access directives to determine when it is used.
6759 NOTE: If an IP address is a.b.c.d
6760 -> bits 25 through 32 are "d"
6761 -> bits 17 through 24 are "c"
6762 -> bits 17 through 32 are "c * 256 + d"
6764 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6765 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6767 This clause only supports fast acl types.
6768 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6770 See also delay_parameters and delay_access.
6774 TYPE: delay_pool_access
6776 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6777 IFDEF: USE_DELAY_POOLS
6780 This is used to determine which delay pool a request falls into.
6782 delay_access is sorted per pool and the matching starts with pool 1,
6783 then pool 2, ..., and finally pool N. The first delay pool where the
6784 request is allowed is selected for the request. If it does not allow
6785 the request to any pool then the request is not delayed (default).
6787 For example, if you want some_big_clients in delay
6788 pool 1 and lotsa_little_clients in delay pool 2:
6790 delay_access 1 allow some_big_clients
6791 delay_access 1 deny all
6792 delay_access 2 allow lotsa_little_clients
6793 delay_access 2 deny all
6794 delay_access 3 allow authenticated_clients
6796 See also delay_parameters and delay_class.
6800 NAME: delay_parameters
6801 TYPE: delay_pool_rates
6803 IFDEF: USE_DELAY_POOLS
6806 This defines the parameters for a delay pool. Each delay pool has
6807 a number of "buckets" associated with it, as explained in the
6808 description of delay_class.
6810 For a class 1 delay pool, the syntax is:
6812 delay_parameters pool aggregate
6814 For a class 2 delay pool:
6816 delay_parameters pool aggregate individual
6818 For a class 3 delay pool:
6820 delay_parameters pool aggregate network individual
6822 For a class 4 delay pool:
6824 delay_parameters pool aggregate network individual user
6826 For a class 5 delay pool:
6828 delay_parameters pool tagrate
6830 The option variables are:
6832 pool a pool number - ie, a number between 1 and the
6833 number specified in delay_pools as used in
6836 aggregate the speed limit parameters for the aggregate bucket
6839 individual the speed limit parameters for the individual
6840 buckets (class 2, 3).
6842 network the speed limit parameters for the network buckets
6845 user the speed limit parameters for the user buckets
6848 tagrate the speed limit parameters for the tag buckets
6851 A pair of delay parameters is written restore/maximum, where restore is
6852 the number of bytes (not bits - modem and network speeds are usually
6853 quoted in bits) per second placed into the bucket, and maximum is the
6854 maximum number of bytes which can be in the bucket at any time.
6856 There must be one delay_parameters line for each delay pool.
6859 For example, if delay pool number 1 is a class 2 delay pool as in the
6860 above example, and is being used to strictly limit each host to 64Kbit/sec
6861 (plus overheads), with no overall limit, the line is:
6863 delay_parameters 1 none 8000/8000
6865 Note that 8 x 8K Byte/sec -> 64K bit/sec.
6867 Note that the word 'none' is used to represent no limit.
6870 And, if delay pool number 2 is a class 3 delay pool as in the above
6871 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6872 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6873 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6874 to permit a decent web page to be downloaded at a decent speed
6875 (if the network is not being limited due to overuse) but slow down
6876 large downloads more significantly:
6878 delay_parameters 2 32000/32000 8000/8000 600/8000
6880 Note that 8 x 32K Byte/sec -> 256K bit/sec.
6881 8 x 8K Byte/sec -> 64K bit/sec.
6882 8 x 600 Byte/sec -> 4800 bit/sec.
6885 Finally, for a class 4 delay pool as in the example - each user will
6886 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6888 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6891 See also delay_class and delay_access.
6895 NAME: delay_initial_bucket_level
6896 COMMENT: (percent, 0-100)
6899 IFDEF: USE_DELAY_POOLS
6900 LOC: Config.Delay.initial
6902 The initial bucket percentage is used to determine how much is put
6903 in each bucket when squid starts, is reconfigured, or first notices
6904 a host accessing it (in class 2 and class 3, individual hosts and
6905 networks only have buckets associated with them once they have been
6910 CLIENT DELAY POOL PARAMETERS
6911 -----------------------------------------------------------------------------
6914 NAME: client_delay_pools
6915 TYPE: client_delay_pool_count
6917 IFDEF: USE_DELAY_POOLS
6918 LOC: Config.ClientDelay
6920 This option specifies the number of client delay pools used. It must
6921 preceed other client_delay_* options.
6924 client_delay_pools 2
6926 See also client_delay_parameters and client_delay_access.
6929 NAME: client_delay_initial_bucket_level
6930 COMMENT: (percent, 0-no_limit)
6933 IFDEF: USE_DELAY_POOLS
6934 LOC: Config.ClientDelay.initial
6936 This option determines the initial bucket size as a percentage of
6937 max_bucket_size from client_delay_parameters. Buckets are created
6938 at the time of the "first" connection from the matching IP. Idle
6939 buckets are periodically deleted up.
6941 You can specify more than 100 percent but note that such "oversized"
6942 buckets are not refilled until their size goes down to max_bucket_size
6943 from client_delay_parameters.
6946 client_delay_initial_bucket_level 50
6949 NAME: client_delay_parameters
6950 TYPE: client_delay_pool_rates
6952 IFDEF: USE_DELAY_POOLS
6953 LOC: Config.ClientDelay
6956 This option configures client-side bandwidth limits using the
6959 client_delay_parameters pool speed_limit max_bucket_size
6961 pool is an integer ID used for client_delay_access matching.
6963 speed_limit is bytes added to the bucket per second.
6965 max_bucket_size is the maximum size of a bucket, enforced after any
6966 speed_limit additions.
6968 Please see the delay_parameters option for more information and
6972 client_delay_parameters 1 1024 2048
6973 client_delay_parameters 2 51200 16384
6975 See also client_delay_access.
6979 NAME: client_delay_access
6980 TYPE: client_delay_pool_access
6982 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6983 IFDEF: USE_DELAY_POOLS
6984 LOC: Config.ClientDelay
6986 This option determines the client-side delay pool for the
6989 client_delay_access pool_ID allow|deny acl_name
6991 All client_delay_access options are checked in their pool ID
6992 order, starting with pool 1. The first checked pool with allowed
6993 request is selected for the request. If no ACL matches or there
6994 are no client_delay_access options, the request bandwidth is not
6997 The ACL-selected pool is then used to find the
6998 client_delay_parameters for the request. Client-side pools are
6999 not used to aggregate clients. Clients are always aggregated
7000 based on their source IP addresses (one bucket per source IP).
7002 This clause only supports fast acl types.
7003 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7004 Additionally, only the client TCP connection details are available.
7005 ACLs testing HTTP properties will not work.
7007 Please see delay_access for more examples.
7010 client_delay_access 1 allow low_rate_network
7011 client_delay_access 2 allow vips_network
7014 See also client_delay_parameters and client_delay_pools.
7018 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
7019 -----------------------------------------------------------------------------
7024 LOC: Config.Wccp.router
7026 DEFAULT_DOC: WCCP disabled.
7029 Use this option to define your WCCP ``home'' router for
7032 wccp_router supports a single WCCP(v1) router
7034 wccp2_router supports multiple WCCPv2 routers
7036 only one of the two may be used at the same time and defines
7037 which version of WCCP to use.
7041 TYPE: IpAddress_list
7042 LOC: Config.Wccp2.router
7044 DEFAULT_DOC: WCCPv2 disabled.
7047 Use this option to define your WCCP ``home'' router for
7050 wccp_router supports a single WCCP(v1) router
7052 wccp2_router supports multiple WCCPv2 routers
7054 only one of the two may be used at the same time and defines
7055 which version of WCCP to use.
7060 LOC: Config.Wccp.version
7064 This directive is only relevant if you need to set up WCCP(v1)
7065 to some very old and end-of-life Cisco routers. In all other
7066 setups it must be left unset or at the default setting.
7067 It defines an internal version in the WCCP(v1) protocol,
7068 with version 4 being the officially documented protocol.
7070 According to some users, Cisco IOS 11.2 and earlier only
7071 support WCCP version 3. If you're using that or an earlier
7072 version of IOS, you may need to change this value to 3, otherwise
7073 do not specify this parameter.
7076 NAME: wccp2_rebuild_wait
7078 LOC: Config.Wccp2.rebuildwait
7082 If this is enabled Squid will wait for the cache dir rebuild to finish
7083 before sending the first wccp2 HereIAm packet
7086 NAME: wccp2_forwarding_method
7088 LOC: Config.Wccp2.forwarding_method
7092 WCCP2 allows the setting of forwarding methods between the
7093 router/switch and the cache. Valid values are as follows:
7095 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
7096 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
7098 Currently (as of IOS 12.4) cisco routers only support GRE.
7099 Cisco switches only support the L2 redirect assignment method.
7102 NAME: wccp2_return_method
7104 LOC: Config.Wccp2.return_method
7108 WCCP2 allows the setting of return methods between the
7109 router/switch and the cache for packets that the cache
7110 decides not to handle. Valid values are as follows:
7112 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
7113 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
7115 Currently (as of IOS 12.4) cisco routers only support GRE.
7116 Cisco switches only support the L2 redirect assignment.
7118 If the "ip wccp redirect exclude in" command has been
7119 enabled on the cache interface, then it is still safe for
7120 the proxy server to use a l2 redirect method even if this
7121 option is set to GRE.
7124 NAME: wccp2_assignment_method
7126 LOC: Config.Wccp2.assignment_method
7130 WCCP2 allows the setting of methods to assign the WCCP hash
7131 Valid values are as follows:
7133 hash - Hash assignment
7134 mask - Mask assignment
7136 As a general rule, cisco routers support the hash assignment method
7137 and cisco switches support the mask assignment method.
7142 LOC: Config.Wccp2.info
7143 DEFAULT_IF_NONE: standard 0
7144 DEFAULT_DOC: Use the 'web-cache' standard service.
7147 WCCP2 allows for multiple traffic services. There are two
7148 types: "standard" and "dynamic". The standard type defines
7149 one service id - http (id 0). The dynamic service ids can be from
7150 51 to 255 inclusive. In order to use a dynamic service id
7151 one must define the type of traffic to be redirected; this is done
7152 using the wccp2_service_info option.
7154 The "standard" type does not require a wccp2_service_info option,
7155 just specifying the service id will suffice.
7157 MD5 service authentication can be enabled by adding
7158 "password=<password>" to the end of this service declaration.
7162 wccp2_service standard 0 # for the 'web-cache' standard service
7163 wccp2_service dynamic 80 # a dynamic service type which will be
7164 # fleshed out with subsequent options.
7165 wccp2_service standard 0 password=foo
7168 NAME: wccp2_service_info
7169 TYPE: wccp2_service_info
7170 LOC: Config.Wccp2.info
7174 Dynamic WCCPv2 services require further information to define the
7175 traffic you wish to have diverted.
7179 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
7180 priority=<priority> ports=<port>,<port>..
7182 The relevant WCCPv2 flags:
7183 + src_ip_hash, dst_ip_hash
7184 + source_port_hash, dst_port_hash
7185 + src_ip_alt_hash, dst_ip_alt_hash
7186 + src_port_alt_hash, dst_port_alt_hash
7189 The port list can be one to eight entries.
7193 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
7194 priority=240 ports=80
7196 Note: the service id must have been defined by a previous
7197 'wccp2_service dynamic <id>' entry.
7202 LOC: Config.Wccp2.weight
7206 Each cache server gets assigned a set of the destination
7207 hash proportional to their weight.
7212 LOC: Config.Wccp.address
7214 DEFAULT_DOC: Address selected by the operating system.
7217 Use this option if you require WCCPv2 to use a specific
7220 The default behavior is to not bind to any specific address.
7225 LOC: Config.Wccp2.address
7227 DEFAULT_DOC: Address selected by the operating system.
7230 Use this option if you require WCCP to use a specific
7233 The default behavior is to not bind to any specific address.
7237 PERSISTENT CONNECTION HANDLING
7238 -----------------------------------------------------------------------------
7240 Also see "pconn_timeout" in the TIMEOUTS section
7243 NAME: client_persistent_connections
7245 LOC: Config.onoff.client_pconns
7248 Persistent connection support for clients.
7249 Squid uses persistent connections (when allowed). You can use
7250 this option to disable persistent connections with clients.
7253 NAME: server_persistent_connections
7255 LOC: Config.onoff.server_pconns
7258 Persistent connection support for servers.
7259 Squid uses persistent connections (when allowed). You can use
7260 this option to disable persistent connections with servers.
7263 NAME: persistent_connection_after_error
7265 LOC: Config.onoff.error_pconns
7268 With this directive the use of persistent connections after
7269 HTTP errors can be disabled. Useful if you have clients
7270 who fail to handle errors on persistent connections proper.
7273 NAME: detect_broken_pconn
7275 LOC: Config.onoff.detect_broken_server_pconns
7278 Some servers have been found to incorrectly signal the use
7279 of HTTP/1.0 persistent connections even on replies not
7280 compatible, causing significant delays. This server problem
7281 has mostly been seen on redirects.
7283 By enabling this directive Squid attempts to detect such
7284 broken replies and automatically assume the reply is finished
7285 after 10 seconds timeout.
7289 CACHE DIGEST OPTIONS
7290 -----------------------------------------------------------------------------
7293 NAME: digest_generation
7294 IFDEF: USE_CACHE_DIGESTS
7296 LOC: Config.onoff.digest_generation
7299 This controls whether the server will generate a Cache Digest
7300 of its contents. By default, Cache Digest generation is
7301 enabled if Squid is compiled with --enable-cache-digests defined.
7304 NAME: digest_bits_per_entry
7305 IFDEF: USE_CACHE_DIGESTS
7307 LOC: Config.digest.bits_per_entry
7310 This is the number of bits of the server's Cache Digest which
7311 will be associated with the Digest entry for a given HTTP
7312 Method and URL (public key) combination. The default is 5.
7315 NAME: digest_rebuild_period
7316 IFDEF: USE_CACHE_DIGESTS
7319 LOC: Config.digest.rebuild_period
7322 This is the wait time between Cache Digest rebuilds.
7325 NAME: digest_rewrite_period
7327 IFDEF: USE_CACHE_DIGESTS
7329 LOC: Config.digest.rewrite_period
7332 This is the wait time between Cache Digest writes to
7336 NAME: digest_swapout_chunk_size
7339 IFDEF: USE_CACHE_DIGESTS
7340 LOC: Config.digest.swapout_chunk_size
7343 This is the number of bytes of the Cache Digest to write to
7344 disk at a time. It defaults to 4096 bytes (4KB), the Squid
7348 NAME: digest_rebuild_chunk_percentage
7349 COMMENT: (percent, 0-100)
7350 IFDEF: USE_CACHE_DIGESTS
7352 LOC: Config.digest.rebuild_chunk_percentage
7355 This is the percentage of the Cache Digest to be scanned at a
7356 time. By default it is set to 10% of the Cache Digest.
7361 -----------------------------------------------------------------------------
7366 LOC: Config.Port.snmp
7368 DEFAULT_DOC: SNMP disabled.
7371 The port number where Squid listens for SNMP requests. To enable
7372 SNMP support set this to a suitable port number. Port number
7373 3401 is often used for the Squid SNMP agent. By default it's
7374 set to "0" (disabled)
7382 LOC: Config.accessList.snmp
7384 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
7387 Allowing or denying access to the SNMP port.
7389 All access to the agent is denied by default.
7392 snmp_access allow|deny [!]aclname ...
7394 This clause only supports fast acl types.
7395 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7398 snmp_access allow snmppublic localhost
7399 snmp_access deny all
7402 NAME: snmp_incoming_address
7404 LOC: Config.Addrs.snmp_incoming
7406 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
7409 Just like 'udp_incoming_address', but for the SNMP port.
7411 snmp_incoming_address is used for the SNMP socket receiving
7412 messages from SNMP agents.
7414 The default snmp_incoming_address is to listen on all
7415 available network interfaces.
7418 NAME: snmp_outgoing_address
7420 LOC: Config.Addrs.snmp_outgoing
7422 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
7425 Just like 'udp_outgoing_address', but for the SNMP port.
7427 snmp_outgoing_address is used for SNMP packets returned to SNMP
7430 If snmp_outgoing_address is not set it will use the same socket
7431 as snmp_incoming_address. Only change this if you want to have
7432 SNMP replies sent using another address than where this Squid
7433 listens for SNMP queries.
7435 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
7436 the same value since they both use the same port.
7441 -----------------------------------------------------------------------------
7444 NAME: icp_port udp_port
7447 DEFAULT_DOC: ICP disabled.
7448 LOC: Config.Port.icp
7450 The port number where Squid sends and receives ICP queries to
7451 and from neighbor caches. The standard UDP port for ICP is 3130.
7454 icp_port @DEFAULT_ICP_PORT@
7461 DEFAULT_DOC: HTCP disabled.
7462 LOC: Config.Port.htcp
7464 The port number where Squid sends and receives HTCP queries to
7465 and from neighbor caches. To turn it on you want to set it to
7472 NAME: log_icp_queries
7476 LOC: Config.onoff.log_udp
7478 If set, ICP queries are logged to access.log. You may wish
7479 do disable this if your ICP load is VERY high to speed things
7480 up or to simplify log analysis.
7483 NAME: udp_incoming_address
7485 LOC:Config.Addrs.udp_incoming
7487 DEFAULT_DOC: Accept packets from all machine interfaces.
7489 udp_incoming_address is used for UDP packets received from other
7492 The default behavior is to not bind to any specific address.
7494 Only change this if you want to have all UDP queries received on
7495 a specific interface/address.
7497 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
7498 modules. Altering it will affect all of them in the same manner.
7500 see also; udp_outgoing_address
7502 NOTE, udp_incoming_address and udp_outgoing_address can not
7503 have the same value since they both use the same port.
7506 NAME: udp_outgoing_address
7508 LOC: Config.Addrs.udp_outgoing
7510 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
7512 udp_outgoing_address is used for UDP packets sent out to other
7515 The default behavior is to not bind to any specific address.
7517 Instead it will use the same socket as udp_incoming_address.
7518 Only change this if you want to have UDP queries sent using another
7519 address than where this Squid listens for UDP queries from other
7522 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
7523 modules. Altering it will affect all of them in the same manner.
7525 see also; udp_incoming_address
7527 NOTE, udp_incoming_address and udp_outgoing_address can not
7528 have the same value since they both use the same port.
7535 LOC: Config.onoff.icp_hit_stale
7537 If you want to return ICP_HIT for stale cache objects, set this
7538 option to 'on'. If you have sibling relationships with caches
7539 in other administrative domains, this should be 'off'. If you only
7540 have sibling relationships with caches under your control,
7541 it is probably okay to set this to 'on'.
7542 If set to 'on', your siblings should use the option "allow-miss"
7543 on their cache_peer lines for connecting to you.
7546 NAME: minimum_direct_hops
7549 LOC: Config.minDirectHops
7551 If using the ICMP pinging stuff, do direct fetches for sites
7552 which are no more than this many hops away.
7555 NAME: minimum_direct_rtt
7559 LOC: Config.minDirectRtt
7561 If using the ICMP pinging stuff, do direct fetches for sites
7562 which are no more than this many rtt milliseconds away.
7568 LOC: Config.Netdb.low
7570 The low water mark for the ICMP measurement database.
7572 Note: high watermark controlled by netdb_high directive.
7574 These watermarks are counts, not percents. The defaults are
7575 (low) 900 and (high) 1000. When the high water mark is
7576 reached, database entries will be deleted until the low
7583 LOC: Config.Netdb.high
7585 The high water mark for the ICMP measurement database.
7587 Note: low watermark controlled by netdb_low directive.
7589 These watermarks are counts, not percents. The defaults are
7590 (low) 900 and (high) 1000. When the high water mark is
7591 reached, database entries will be deleted until the low
7595 NAME: netdb_ping_period
7597 LOC: Config.Netdb.period
7600 The minimum period for measuring a site. There will be at
7601 least this much delay between successive pings to the same
7602 network. The default is five minutes.
7609 LOC: Config.onoff.query_icmp
7611 If you want to ask your peers to include ICMP data in their ICP
7612 replies, enable this option.
7614 If your peer has configured Squid (during compilation) with
7615 '--enable-icmp' that peer will send ICMP pings to origin server
7616 sites of the URLs it receives. If you enable this option the
7617 ICP replies from that peer will include the ICMP data (if available).
7618 Then, when choosing a parent cache, Squid will choose the parent with
7619 the minimal RTT to the origin server. When this happens, the
7620 hierarchy field of the access.log will be
7621 "CLOSEST_PARENT_MISS". This option is off by default.
7624 NAME: test_reachability
7628 LOC: Config.onoff.test_reachability
7630 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7631 instead of ICP_MISS if the target host is NOT in the ICMP
7632 database, or has a zero RTT.
7635 NAME: icp_query_timeout
7638 DEFAULT_DOC: Dynamic detection.
7640 LOC: Config.Timeout.icp_query
7642 Normally Squid will automatically determine an optimal ICP
7643 query timeout value based on the round-trip-time of recent ICP
7644 queries. If you want to override the value determined by
7645 Squid, set this 'icp_query_timeout' to a non-zero value. This
7646 value is specified in MILLISECONDS, so, to use a 2-second
7647 timeout (the old default), you would write:
7649 icp_query_timeout 2000
7652 NAME: maximum_icp_query_timeout
7656 LOC: Config.Timeout.icp_query_max
7658 Normally the ICP query timeout is determined dynamically. But
7659 sometimes it can lead to very large values (say 5 seconds).
7660 Use this option to put an upper limit on the dynamic timeout
7661 value. Do NOT use this option to always use a fixed (instead
7662 of a dynamic) timeout value. To set a fixed timeout see the
7663 'icp_query_timeout' directive.
7666 NAME: minimum_icp_query_timeout
7670 LOC: Config.Timeout.icp_query_min
7672 Normally the ICP query timeout is determined dynamically. But
7673 sometimes it can lead to very small timeouts, even lower than
7674 the normal latency variance on your link due to traffic.
7675 Use this option to put an lower limit on the dynamic timeout
7676 value. Do NOT use this option to always use a fixed (instead
7677 of a dynamic) timeout value. To set a fixed timeout see the
7678 'icp_query_timeout' directive.
7681 NAME: background_ping_rate
7685 LOC: Config.backgroundPingRate
7687 Controls how often the ICP pings are sent to siblings that
7688 have background-ping set.
7692 MULTICAST ICP OPTIONS
7693 -----------------------------------------------------------------------------
7698 LOC: Config.mcast_group_list
7701 This tag specifies a list of multicast groups which your server
7702 should join to receive multicasted ICP queries.
7704 NOTE! Be very careful what you put here! Be sure you
7705 understand the difference between an ICP _query_ and an ICP
7706 _reply_. This option is to be set only if you want to RECEIVE
7707 multicast queries. Do NOT set this option to SEND multicast
7708 ICP (use cache_peer for that). ICP replies are always sent via
7709 unicast, so this option does not affect whether or not you will
7710 receive replies from multicast group members.
7712 You must be very careful to NOT use a multicast address which
7713 is already in use by another group of caches.
7715 If you are unsure about multicast, please read the Multicast
7716 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7718 Usage: mcast_groups 239.128.16.128 224.0.1.20
7720 By default, Squid doesn't listen on any multicast groups.
7723 NAME: mcast_miss_addr
7724 IFDEF: MULTICAST_MISS_STREAM
7726 LOC: Config.mcast_miss.addr
7728 DEFAULT_DOC: disabled.
7730 If you enable this option, every "cache miss" URL will
7731 be sent out on the specified multicast address.
7733 Do not enable this option unless you are are absolutely
7734 certain you understand what you are doing.
7737 NAME: mcast_miss_ttl
7738 IFDEF: MULTICAST_MISS_STREAM
7740 LOC: Config.mcast_miss.ttl
7743 This is the time-to-live value for packets multicasted
7744 when multicasting off cache miss URLs is enabled. By
7745 default this is set to 'site scope', i.e. 16.
7748 NAME: mcast_miss_port
7749 IFDEF: MULTICAST_MISS_STREAM
7751 LOC: Config.mcast_miss.port
7754 This is the port number to be used in conjunction with
7758 NAME: mcast_miss_encode_key
7759 IFDEF: MULTICAST_MISS_STREAM
7761 LOC: Config.mcast_miss.encode_key
7762 DEFAULT: XXXXXXXXXXXXXXXX
7764 The URLs that are sent in the multicast miss stream are
7765 encrypted. This is the encryption key.
7768 NAME: mcast_icp_query_timeout
7772 LOC: Config.Timeout.mcast_icp_query
7774 For multicast peers, Squid regularly sends out ICP "probes" to
7775 count how many other peers are listening on the given multicast
7776 address. This value specifies how long Squid should wait to
7777 count all the replies. The default is 2000 msec, or 2
7782 INTERNAL ICON OPTIONS
7783 -----------------------------------------------------------------------------
7786 NAME: icon_directory
7788 LOC: Config.icons.directory
7789 DEFAULT: @DEFAULT_ICON_DIR@
7791 Where the icons are stored. These are normally kept in
7795 NAME: global_internal_static
7797 LOC: Config.onoff.global_internal_static
7800 This directive controls is Squid should intercept all requests for
7801 /squid-internal-static/ no matter which host the URL is requesting
7802 (default on setting), or if nothing special should be done for
7803 such URLs (off setting). The purpose of this directive is to make
7804 icons etc work better in complex cache hierarchies where it may
7805 not always be possible for all corners in the cache mesh to reach
7806 the server generating a directory listing.
7809 NAME: short_icon_urls
7811 LOC: Config.icons.use_short_names
7814 If this is enabled Squid will use short URLs for icons.
7815 If disabled it will revert to the old behavior of including
7816 it's own name and port in the URL.
7818 If you run a complex cache hierarchy with a mix of Squid and
7819 other proxies you may need to disable this directive.
7824 -----------------------------------------------------------------------------
7827 NAME: error_directory
7829 LOC: Config.errorDirectory
7831 DEFAULT_DOC: Send error pages in the clients preferred language
7833 If you wish to create your own versions of the default
7834 error files to customize them to suit your company copy
7835 the error/template files to another directory and point
7838 WARNING: This option will disable multi-language support
7839 on error pages if used.
7841 The squid developers are interested in making squid available in
7842 a wide variety of languages. If you are making translations for a
7843 language that Squid does not currently provide please consider
7844 contributing your translation back to the project.
7845 http://wiki.squid-cache.org/Translations
7847 The squid developers working on translations are happy to supply drop-in
7848 translated error files in exchange for any new language contributions.
7851 NAME: error_default_language
7852 IFDEF: USE_ERR_LOCALES
7854 LOC: Config.errorDefaultLanguage
7856 DEFAULT_DOC: Generate English language pages.
7858 Set the default language which squid will send error pages in
7859 if no existing translation matches the clients language
7862 If unset (default) generic English will be used.
7864 The squid developers are interested in making squid available in
7865 a wide variety of languages. If you are interested in making
7866 translations for any language see the squid wiki for details.
7867 http://wiki.squid-cache.org/Translations
7870 NAME: error_log_languages
7871 IFDEF: USE_ERR_LOCALES
7873 LOC: Config.errorLogMissingLanguages
7876 Log to cache.log what languages users are attempting to
7877 auto-negotiate for translations.
7879 Successful negotiations are not logged. Only failures
7880 have meaning to indicate that Squid may need an upgrade
7881 of its error page translations.
7884 NAME: err_page_stylesheet
7886 LOC: Config.errorStylesheet
7887 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7889 CSS Stylesheet to pattern the display of Squid default error pages.
7891 For information on CSS see http://www.w3.org/Style/CSS/
7896 LOC: Config.errHtmlText
7899 HTML text to include in error messages. Make this a "mailto"
7900 URL to your admin address, or maybe just a link to your
7901 organizations Web page.
7903 To include this in your error messages, you must rewrite
7904 the error template files (found in the "errors" directory).
7905 Wherever you want the 'err_html_text' line to appear,
7906 insert a %L tag in the error template file.
7909 NAME: email_err_data
7912 LOC: Config.onoff.emailErrData
7915 If enabled, information about the occurred error will be
7916 included in the mailto links of the ERR pages (if %W is set)
7917 so that the email body contains the data.
7918 Syntax is <A HREF="mailto:%w%W">%w</A>
7923 LOC: Config.denyInfoList
7926 Usage: deny_info err_page_name acl
7927 or deny_info http://... acl
7928 or deny_info TCP_RESET acl
7930 This can be used to return a ERR_ page for requests which
7931 do not pass the 'http_access' rules. Squid remembers the last
7932 acl it evaluated in http_access, and if a 'deny_info' line exists
7933 for that ACL Squid returns a corresponding error page.
7935 The acl is typically the last acl on the http_access deny line which
7936 denied access. The exceptions to this rule are:
7937 - When Squid needs to request authentication credentials. It's then
7938 the first authentication related acl encountered
7939 - When none of the http_access lines matches. It's then the last
7940 acl processed on the last http_access line.
7941 - When the decision to deny access was made by an adaptation service,
7942 the acl name is the corresponding eCAP or ICAP service_name.
7944 NP: If providing your own custom error pages with error_directory
7945 you may also specify them by your custom file name:
7946 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7948 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7949 may be specified by prefixing the file name with the code and a colon.
7950 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7952 Alternatively you can tell Squid to reset the TCP connection
7953 by specifying TCP_RESET.
7955 Or you can specify an error URL or URL pattern. The browsers will
7956 get redirected to the specified URL after formatting tags have
7957 been replaced. Redirect will be done with 302 or 307 according to
7958 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7959 the URL. e.g. 303:http://example.com/
7962 %a - username (if available. Password NOT included)
7965 %E - Error description
7967 %H - Request domain name
7968 %i - Client IP Address
7970 %o - Message result from external ACL helper
7971 %p - Request Port number
7972 %P - Request Protocol name
7973 %R - Request URL path
7974 %T - Timestamp in RFC 1123 format
7975 %U - Full canonical URL from client
7976 (HTTPS URLs terminate with *)
7977 %u - Full canonical URL from client
7978 %w - Admin email from squid.conf
7980 %% - Literal percent (%) code
7985 OPTIONS INFLUENCING REQUEST FORWARDING
7986 -----------------------------------------------------------------------------
7989 NAME: nonhierarchical_direct
7991 LOC: Config.onoff.nonhierarchical_direct
7994 By default, Squid will send any non-hierarchical requests
7995 (not cacheable request type) direct to origin servers.
7997 When this is set to "off", Squid will prefer to send these
7998 requests to parents.
8000 Note that in most configurations, by turning this off you will only
8001 add latency to these request without any improvement in global hit
8004 This option only sets a preference. If the parent is unavailable a
8005 direct connection to the origin server may still be attempted. To
8006 completely prevent direct connections use never_direct.
8011 LOC: Config.onoff.prefer_direct
8014 Normally Squid tries to use parents for most requests. If you for some
8015 reason like it to first try going direct and only use a parent if
8016 going direct fails set this to on.
8018 By combining nonhierarchical_direct off and prefer_direct on you
8019 can set up Squid to use a parent as a backup path if going direct
8022 Note: If you want Squid to use parents for all requests see
8023 the never_direct directive. prefer_direct only modifies how Squid
8024 acts on cacheable requests.
8027 NAME: cache_miss_revalidate
8031 LOC: Config.onoff.cache_miss_revalidate
8033 RFC 7232 defines a conditional request mechanism to prevent
8034 response objects being unnecessarily transferred over the network.
8035 If that mechanism is used by the client and a cache MISS occurs
8036 it can prevent new cache entries being created.
8038 This option determines whether Squid on cache MISS will pass the
8039 client revalidation request to the server or tries to fetch new
8040 content for caching. It can be useful while the cache is mostly
8041 empty to more quickly have the cache populated by generating
8042 non-conditional GETs.
8044 When set to 'on' (default), Squid will pass all client If-* headers
8045 to the server. This permits server responses without a cacheable
8046 payload to be delivered and on MISS no new cache entry is created.
8048 When set to 'off' and if the request is cacheable, Squid will
8049 remove the clients If-Modified-Since and If-None-Match headers from
8050 the request sent to the server. This requests a 200 status response
8051 from the server to create a new cache entry with.
8056 LOC: Config.accessList.AlwaysDirect
8058 DEFAULT_DOC: Prevent any cache_peer being used for this request.
8060 Usage: always_direct allow|deny [!]aclname ...
8062 Here you can use ACL elements to specify requests which should
8063 ALWAYS be forwarded by Squid to the origin servers without using
8064 any peers. For example, to always directly forward requests for
8065 local servers ignoring any parents or siblings you may have use
8068 acl local-servers dstdomain my.domain.net
8069 always_direct allow local-servers
8071 To always forward FTP requests directly, use
8074 always_direct allow FTP
8076 NOTE: There is a similar, but opposite option named
8077 'never_direct'. You need to be aware that "always_direct deny
8078 foo" is NOT the same thing as "never_direct allow foo". You
8079 may need to use a deny rule to exclude a more-specific case of
8080 some other rule. Example:
8082 acl local-external dstdomain external.foo.net
8083 acl local-servers dstdomain .foo.net
8084 always_direct deny local-external
8085 always_direct allow local-servers
8087 NOTE: If your goal is to make the client forward the request
8088 directly to the origin server bypassing Squid then this needs
8089 to be done in the client configuration. Squid configuration
8090 can only tell Squid how Squid should fetch the object.
8092 NOTE: This directive is not related to caching. The replies
8093 is cached as usual even if you use always_direct. To not cache
8094 the replies see the 'cache' directive.
8096 This clause supports both fast and slow acl types.
8097 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8102 LOC: Config.accessList.NeverDirect
8104 DEFAULT_DOC: Allow DNS results to be used for this request.
8106 Usage: never_direct allow|deny [!]aclname ...
8108 never_direct is the opposite of always_direct. Please read
8109 the description for always_direct if you have not already.
8111 With 'never_direct' you can use ACL elements to specify
8112 requests which should NEVER be forwarded directly to origin
8113 servers. For example, to force the use of a proxy for all
8114 requests, except those in your local domain use something like:
8116 acl local-servers dstdomain .foo.net
8117 never_direct deny local-servers
8118 never_direct allow all
8120 or if Squid is inside a firewall and there are local intranet
8121 servers inside the firewall use something like:
8123 acl local-intranet dstdomain .foo.net
8124 acl local-external dstdomain external.foo.net
8125 always_direct deny local-external
8126 always_direct allow local-intranet
8127 never_direct allow all
8129 This clause supports both fast and slow acl types.
8130 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8134 ADVANCED NETWORKING OPTIONS
8135 -----------------------------------------------------------------------------
8138 NAME: incoming_udp_average incoming_icp_average
8141 LOC: Config.comm_incoming.udp.average
8143 Heavy voodoo here. I can't even believe you are reading this.
8144 Are you crazy? Don't even think about adjusting these unless
8145 you understand the algorithms in comm_select.c first!
8148 NAME: incoming_tcp_average incoming_http_average
8151 LOC: Config.comm_incoming.tcp.average
8153 Heavy voodoo here. I can't even believe you are reading this.
8154 Are you crazy? Don't even think about adjusting these unless
8155 you understand the algorithms in comm_select.c first!
8158 NAME: incoming_dns_average
8161 LOC: Config.comm_incoming.dns.average
8163 Heavy voodoo here. I can't even believe you are reading this.
8164 Are you crazy? Don't even think about adjusting these unless
8165 you understand the algorithms in comm_select.c first!
8168 NAME: min_udp_poll_cnt min_icp_poll_cnt
8171 LOC: Config.comm_incoming.udp.min_poll
8173 Heavy voodoo here. I can't even believe you are reading this.
8174 Are you crazy? Don't even think about adjusting these unless
8175 you understand the algorithms in comm_select.c first!
8178 NAME: min_dns_poll_cnt
8181 LOC: Config.comm_incoming.dns.min_poll
8183 Heavy voodoo here. I can't even believe you are reading this.
8184 Are you crazy? Don't even think about adjusting these unless
8185 you understand the algorithms in comm_select.c first!
8188 NAME: min_tcp_poll_cnt min_http_poll_cnt
8191 LOC: Config.comm_incoming.tcp.min_poll
8193 Heavy voodoo here. I can't even believe you are reading this.
8194 Are you crazy? Don't even think about adjusting these unless
8195 you understand the algorithms in comm_select.c first!
8201 LOC: Config.accept_filter
8205 The name of an accept(2) filter to install on Squid's
8206 listen socket(s). This feature is perhaps specific to
8207 FreeBSD and requires support in the kernel.
8209 The 'httpready' filter delays delivering new connections
8210 to Squid until a full HTTP request has been received.
8211 See the accf_http(9) man page for details.
8213 The 'dataready' filter delays delivering new connections
8214 to Squid until there is some data to process.
8215 See the accf_dataready(9) man page for details.
8219 The 'data' filter delays delivering of new connections
8220 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
8221 You may optionally specify a number of seconds to wait by
8222 'data=N' where N is the number of seconds. Defaults to 30
8223 if not specified. See the tcp(7) man page for details.
8226 accept_filter httpready
8231 NAME: client_ip_max_connections
8233 LOC: Config.client_ip_max_connections
8235 DEFAULT_DOC: No limit.
8237 Set an absolute limit on the number of connections a single
8238 client IP can use. Any more than this and Squid will begin to drop
8239 new connections from the client until it closes some links.
8241 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
8242 connections from the client. For finer control use the ACL access controls.
8244 Requires client_db to be enabled (the default).
8246 WARNING: This may noticably slow down traffic received via external proxies
8247 or NAT devices and cause them to rebound error messages back to their clients.
8250 NAME: tcp_recv_bufsize
8254 DEFAULT_DOC: Use operating system TCP defaults.
8255 LOC: Config.tcpRcvBufsz
8257 Size of receive buffer to set for TCP sockets. Probably just
8258 as easy to change your kernel's default.
8259 Omit from squid.conf to use the default buffer size.
8264 -----------------------------------------------------------------------------
8271 LOC: Adaptation::Icap::TheConfig.onoff
8274 If you want to enable the ICAP module support, set this to on.
8277 NAME: icap_connect_timeout
8280 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
8283 This parameter specifies how long to wait for the TCP connect to
8284 the requested ICAP server to complete before giving up and either
8285 terminating the HTTP transaction or bypassing the failure.
8287 The default for optional services is peer_connect_timeout.
8288 The default for essential services is connect_timeout.
8289 If this option is explicitly set, its value applies to all services.
8292 NAME: icap_io_timeout
8296 DEFAULT_DOC: Use read_timeout.
8297 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
8300 This parameter specifies how long to wait for an I/O activity on
8301 an established, active ICAP connection before giving up and
8302 either terminating the HTTP transaction or bypassing the
8306 NAME: icap_service_failure_limit
8307 COMMENT: limit [in memory-depth time-units]
8308 TYPE: icap_service_failure_limit
8310 LOC: Adaptation::Icap::TheConfig
8313 The limit specifies the number of failures that Squid tolerates
8314 when establishing a new TCP connection with an ICAP service. If
8315 the number of failures exceeds the limit, the ICAP service is
8316 not used for new ICAP requests until it is time to refresh its
8319 A negative value disables the limit. Without the limit, an ICAP
8320 service will not be considered down due to connectivity failures
8321 between ICAP OPTIONS requests.
8323 Squid forgets ICAP service failures older than the specified
8324 value of memory-depth. The memory fading algorithm
8325 is approximate because Squid does not remember individual
8326 errors but groups them instead, splitting the option
8327 value into ten time slots of equal length.
8329 When memory-depth is 0 and by default this option has no
8330 effect on service failure expiration.
8332 Squid always forgets failures when updating service settings
8333 using an ICAP OPTIONS transaction, regardless of this option
8337 # suspend service usage after 10 failures in 5 seconds:
8338 icap_service_failure_limit 10 in 5 seconds
8341 NAME: icap_service_revival_delay
8344 LOC: Adaptation::Icap::TheConfig.service_revival_delay
8347 The delay specifies the number of seconds to wait after an ICAP
8348 OPTIONS request failure before requesting the options again. The
8349 failed ICAP service is considered "down" until fresh OPTIONS are
8352 The actual delay cannot be smaller than the hardcoded minimum
8353 delay of 30 seconds.
8356 NAME: icap_preview_enable
8360 LOC: Adaptation::Icap::TheConfig.preview_enable
8363 The ICAP Preview feature allows the ICAP server to handle the
8364 HTTP message by looking only at the beginning of the message body
8365 or even without receiving the body at all. In some environments,
8366 previews greatly speedup ICAP processing.
8368 During an ICAP OPTIONS transaction, the server may tell Squid what
8369 HTTP messages should be previewed and how big the preview should be.
8370 Squid will not use Preview if the server did not request one.
8372 To disable ICAP Preview for all ICAP services, regardless of
8373 individual ICAP server OPTIONS responses, set this option to "off".
8375 icap_preview_enable off
8378 NAME: icap_preview_size
8381 LOC: Adaptation::Icap::TheConfig.preview_size
8383 DEFAULT_DOC: No preview sent.
8385 The default size of preview data to be sent to the ICAP server.
8386 This value might be overwritten on a per server basis by OPTIONS requests.
8389 NAME: icap_206_enable
8393 LOC: Adaptation::Icap::TheConfig.allow206_enable
8396 206 (Partial Content) responses is an ICAP extension that allows the
8397 ICAP agents to optionally combine adapted and original HTTP message
8398 content. The decision to combine is postponed until the end of the
8399 ICAP response. Squid supports Partial Content extension by default.
8401 Activation of the Partial Content extension is negotiated with each
8402 ICAP service during OPTIONS exchange. Most ICAP servers should handle
8403 negotation correctly even if they do not support the extension, but
8404 some might fail. To disable Partial Content support for all ICAP
8405 services and to avoid any negotiation, set this option to "off".
8411 NAME: icap_default_options_ttl
8414 LOC: Adaptation::Icap::TheConfig.default_options_ttl
8417 The default TTL value for ICAP OPTIONS responses that don't have
8418 an Options-TTL header.
8421 NAME: icap_persistent_connections
8425 LOC: Adaptation::Icap::TheConfig.reuse_connections
8428 Whether or not Squid should use persistent connections to
8432 NAME: adaptation_send_client_ip icap_send_client_ip
8434 IFDEF: USE_ADAPTATION
8436 LOC: Adaptation::Config::send_client_ip
8439 If enabled, Squid shares HTTP client IP information with adaptation
8440 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
8441 For eCAP, Squid sets the libecap::metaClientIp transaction option.
8443 See also: adaptation_uses_indirect_client
8446 NAME: adaptation_send_username icap_send_client_username
8448 IFDEF: USE_ADAPTATION
8450 LOC: Adaptation::Config::send_username
8453 This sends authenticated HTTP client username (if available) to
8454 the adaptation service.
8456 For ICAP, the username value is encoded based on the
8457 icap_client_username_encode option and is sent using the header
8458 specified by the icap_client_username_header option.
8461 NAME: icap_client_username_header
8464 LOC: Adaptation::Icap::TheConfig.client_username_header
8465 DEFAULT: X-Client-Username
8467 ICAP request header name to use for adaptation_send_username.
8470 NAME: icap_client_username_encode
8474 LOC: Adaptation::Icap::TheConfig.client_username_encode
8477 Whether to base64 encode the authenticated client username.
8481 TYPE: icap_service_type
8483 LOC: Adaptation::Icap::TheConfig
8486 Defines a single ICAP service using the following format:
8488 icap_service id vectoring_point uri [option ...]
8491 an opaque identifier or name which is used to direct traffic to
8492 this specific service. Must be unique among all adaptation
8493 services in squid.conf.
8495 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8496 This specifies at which point of transaction processing the
8497 ICAP service should be activated. *_postcache vectoring points
8498 are not yet supported.
8500 uri: icap://servername:port/servicepath
8501 ICAP server and service location.
8502 icaps://servername:port/servicepath
8503 The "icap:" URI scheme is used for traditional ICAP server and
8504 service location (default port is 1344, connections are not
8505 encrypted). The "icaps:" URI scheme is for Secure ICAP
8506 services that use SSL/TLS-encrypted ICAP connections (by
8507 default, on port 11344).
8509 ICAP does not allow a single service to handle both REQMOD and RESPMOD
8510 transactions. Squid does not enforce that requirement. You can specify
8511 services with the same service_url and different vectoring_points. You
8512 can even specify multiple identical services as long as their
8513 service_names differ.
8515 To activate a service, use the adaptation_access directive. To group
8516 services, use adaptation_service_chain and adaptation_service_set.
8518 Service options are separated by white space. ICAP services support
8519 the following name=value options:
8522 If set to 'on' or '1', the ICAP service is treated as
8523 optional. If the service cannot be reached or malfunctions,
8524 Squid will try to ignore any errors and process the message as
8525 if the service was not enabled. No all ICAP errors can be
8526 bypassed. If set to 0, the ICAP service is treated as
8527 essential and all ICAP errors will result in an error page
8528 returned to the HTTP client.
8530 Bypass is off by default: services are treated as essential.
8533 If set to 'on' or '1', the ICAP service is allowed to
8534 dynamically change the current message adaptation plan by
8535 returning a chain of services to be used next. The services
8536 are specified using the X-Next-Services ICAP response header
8537 value, formatted as a comma-separated list of service names.
8538 Each named service should be configured in squid.conf. Other
8539 services are ignored. An empty X-Next-Services value results
8540 in an empty plan which ends the current adaptation.
8542 Dynamic adaptation plan may cross or cover multiple supported
8543 vectoring points in their natural processing order.
8545 Routing is not allowed by default: the ICAP X-Next-Services
8546 response header is ignored.
8549 Only has effect on split-stack systems. The default on those systems
8550 is to use IPv4-only connections. When set to 'on' this option will
8551 make Squid use IPv6-only connections to contact this ICAP service.
8553 on-overload=block|bypass|wait|force
8554 If the service Max-Connections limit has been reached, do
8555 one of the following for each new ICAP transaction:
8556 * block: send an HTTP error response to the client
8557 * bypass: ignore the "over-connected" ICAP service
8558 * wait: wait (in a FIFO queue) for an ICAP connection slot
8559 * force: proceed, ignoring the Max-Connections limit
8561 In SMP mode with N workers, each worker assumes the service
8562 connection limit is Max-Connections/N, even though not all
8563 workers may use a given service.
8565 The default value is "bypass" if service is bypassable,
8566 otherwise it is set to "wait".
8570 Use the given number as the Max-Connections limit, regardless
8571 of the Max-Connections value given by the service, if any.
8573 connection-encryption=on|off
8574 Determines the ICAP service effect on the connections_encrypted
8577 The default is "on" for Secure ICAP services (i.e., those
8578 with the icaps:// service URIs scheme) and "off" for plain ICAP
8581 Does not affect ICAP connections (e.g., does not turn Secure
8584 ==== ICAPS / TLS OPTIONS ====
8586 These options are used for Secure ICAP (icaps://....) services only.
8588 tls-cert=/path/to/ssl/certificate
8589 A client SSL certificate to use when connecting to
8592 tls-key=/path/to/ssl/key
8593 The private TLS/SSL key corresponding to sslcert above.
8594 If 'tls-key' is not specified 'tls-cert' is assumed to
8595 reference a combined PEM format file containing both the
8596 certificate and the key.
8598 tls-cipher=... The list of valid TLS/SSL ciphers to use when connecting
8599 to this icap server.
8602 The minimum TLS protocol version to permit. To control
8603 SSLv3 use the ssloptions= parameter.
8604 Supported Values: 1.0 (default), 1.1, 1.2
8606 tls-options=... Specify various OpenSSL library options:
8608 NO_SSLv3 Disallow the use of SSLv3
8610 NO_TLSv1 Disallow the use of TLSv1.0
8611 NO_TLSv1_1 Disallow the use of TLSv1.1
8612 NO_TLSv1_2 Disallow the use of TLSv1.2
8615 Always create a new key when using
8616 temporary/ephemeral DH key exchanges
8618 ALL Enable various bug workarounds
8619 suggested as "harmless" by OpenSSL
8620 Be warned that this reduces SSL/TLS
8621 strength to some attacks.
8623 See the OpenSSL SSL_CTX_set_options documentation for a
8624 more complete list. Options relevant only to SSLv2 are
8627 tls-cafile= PEM file containing CA certificates to use when verifying
8628 the icap server certificate.
8629 Use to specify intermediate CA certificate(s) if not sent
8630 by the server. Or the full CA chain for the server when
8631 using the tls-default-ca=off flag.
8632 May be repeated to load multiple files.
8634 tls-capath=... A directory containing additional CA certificates to
8635 use when verifying the icap server certificate.
8636 Requires OpenSSL or LibreSSL.
8638 tls-crlfile=... A certificate revocation list file to use when
8639 verifying the icap server certificate.
8641 tls-flags=... Specify various flags modifying the Squid TLS implementation:
8644 Accept certificates even if they fail to
8647 Don't verify the icap server certificate
8648 matches the server name
8650 tls-default-ca[=off]
8651 Whether to use the system Trusted CAs. Default is ON.
8653 tls-domain= The icap server name as advertised in it's certificate.
8654 Used for verifying the correctness of the received icap
8655 server certificate. If not specified the icap server
8656 hostname extracted from ICAP URI will be used.
8658 Older icap_service format without optional named parameters is
8659 deprecated but supported for backward compatibility.
8662 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8663 icap_service svcLogger reqmod_precache icaps://icap2.mydomain.net:11344/reqmod routing=on
8667 TYPE: icap_class_type
8672 This deprecated option was documented to define an ICAP service
8673 chain, even though it actually defined a set of similar, redundant
8674 services, and the chains were not supported.
8676 To define a set of redundant services, please use the
8677 adaptation_service_set directive. For service chains, use
8678 adaptation_service_chain.
8682 TYPE: icap_access_type
8687 This option is deprecated. Please use adaptation_access, which
8688 has the same ICAP functionality, but comes with better
8689 documentation, and eCAP support.
8694 -----------------------------------------------------------------------------
8701 LOC: Adaptation::Ecap::TheConfig.onoff
8704 Controls whether eCAP support is enabled.
8708 TYPE: ecap_service_type
8710 LOC: Adaptation::Ecap::TheConfig
8713 Defines a single eCAP service
8715 ecap_service id vectoring_point uri [option ...]
8718 an opaque identifier or name which is used to direct traffic to
8719 this specific service. Must be unique among all adaptation
8720 services in squid.conf.
8722 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8723 This specifies at which point of transaction processing the
8724 eCAP service should be activated. *_postcache vectoring points
8725 are not yet supported.
8727 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8728 Squid uses the eCAP service URI to match this configuration
8729 line with one of the dynamically loaded services. Each loaded
8730 eCAP service must have a unique URI. Obtain the right URI from
8731 the service provider.
8733 To activate a service, use the adaptation_access directive. To group
8734 services, use adaptation_service_chain and adaptation_service_set.
8736 Service options are separated by white space. eCAP services support
8737 the following name=value options:
8740 If set to 'on' or '1', the eCAP service is treated as optional.
8741 If the service cannot be reached or malfunctions, Squid will try
8742 to ignore any errors and process the message as if the service
8743 was not enabled. No all eCAP errors can be bypassed.
8744 If set to 'off' or '0', the eCAP service is treated as essential
8745 and all eCAP errors will result in an error page returned to the
8748 Bypass is off by default: services are treated as essential.
8751 If set to 'on' or '1', the eCAP service is allowed to
8752 dynamically change the current message adaptation plan by
8753 returning a chain of services to be used next.
8755 Dynamic adaptation plan may cross or cover multiple supported
8756 vectoring points in their natural processing order.
8758 Routing is not allowed by default.
8760 connection-encryption=on|off
8761 Determines the eCAP service effect on the connections_encrypted
8764 Defaults to "on", which does not taint the master transaction
8767 Does not affect eCAP API calls.
8769 Older ecap_service format without optional named parameters is
8770 deprecated but supported for backward compatibility.
8774 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8775 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8778 NAME: loadable_modules
8780 IFDEF: USE_LOADABLE_MODULES
8781 LOC: Config.loadable_module_names
8784 Instructs Squid to load the specified dynamic module(s) or activate
8785 preloaded module(s).
8787 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8791 MESSAGE ADAPTATION OPTIONS
8792 -----------------------------------------------------------------------------
8795 NAME: adaptation_service_set
8796 TYPE: adaptation_service_set_type
8797 IFDEF: USE_ADAPTATION
8802 Configures an ordered set of similar, redundant services. This is
8803 useful when hot standby or backup adaptation servers are available.
8805 adaptation_service_set set_name service_name1 service_name2 ...
8807 The named services are used in the set declaration order. The first
8808 applicable adaptation service from the set is used first. The next
8809 applicable service is tried if and only if the transaction with the
8810 previous service fails and the message waiting to be adapted is still
8813 When adaptation starts, broken services are ignored as if they were
8814 not a part of the set. A broken service is a down optional service.
8816 The services in a set must be attached to the same vectoring point
8817 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8819 If all services in a set are optional then adaptation failures are
8820 bypassable. If all services in the set are essential, then a
8821 transaction failure with one service may still be retried using
8822 another service from the set, but when all services fail, the master
8823 transaction fails as well.
8825 A set may contain a mix of optional and essential services, but that
8826 is likely to lead to surprising results because broken services become
8827 ignored (see above), making previously bypassable failures fatal.
8828 Technically, it is the bypassability of the last failed service that
8831 See also: adaptation_access adaptation_service_chain
8834 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8835 adaptation service_set svcLogger loggerLocal loggerRemote
8838 NAME: adaptation_service_chain
8839 TYPE: adaptation_service_chain_type
8840 IFDEF: USE_ADAPTATION
8845 Configures a list of complementary services that will be applied
8846 one-by-one, forming an adaptation chain or pipeline. This is useful
8847 when Squid must perform different adaptations on the same message.
8849 adaptation_service_chain chain_name service_name1 svc_name2 ...
8851 The named services are used in the chain declaration order. The first
8852 applicable adaptation service from the chain is used first. The next
8853 applicable service is applied to the successful adaptation results of
8854 the previous service in the chain.
8856 When adaptation starts, broken services are ignored as if they were
8857 not a part of the chain. A broken service is a down optional service.
8859 Request satisfaction terminates the adaptation chain because Squid
8860 does not currently allow declaration of RESPMOD services at the
8861 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8863 The services in a chain must be attached to the same vectoring point
8864 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8866 A chain may contain a mix of optional and essential services. If an
8867 essential adaptation fails (or the failure cannot be bypassed for
8868 other reasons), the master transaction fails. Otherwise, the failure
8869 is bypassed as if the failed adaptation service was not in the chain.
8871 See also: adaptation_access adaptation_service_set
8874 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8877 NAME: adaptation_access
8878 TYPE: adaptation_access_type
8879 IFDEF: USE_ADAPTATION
8882 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8884 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8886 adaptation_access service_name allow|deny [!]aclname...
8887 adaptation_access set_name allow|deny [!]aclname...
8889 At each supported vectoring point, the adaptation_access
8890 statements are processed in the order they appear in this
8891 configuration file. Statements pointing to the following services
8892 are ignored (i.e., skipped without checking their ACL):
8894 - services serving different vectoring points
8895 - "broken-but-bypassable" services
8896 - "up" services configured to ignore such transactions
8897 (e.g., based on the ICAP Transfer-Ignore header).
8899 When a set_name is used, all services in the set are checked
8900 using the same rules, to find the first applicable one. See
8901 adaptation_service_set for details.
8903 If an access list is checked and there is a match, the
8904 processing stops: For an "allow" rule, the corresponding
8905 adaptation service is used for the transaction. For a "deny"
8906 rule, no adaptation service is activated.
8908 It is currently not possible to apply more than one adaptation
8909 service at the same vectoring point to the same HTTP transaction.
8911 See also: icap_service and ecap_service
8914 adaptation_access service_1 allow all
8917 NAME: adaptation_service_iteration_limit
8919 IFDEF: USE_ADAPTATION
8920 LOC: Adaptation::Config::service_iteration_limit
8923 Limits the number of iterations allowed when applying adaptation
8924 services to a message. If your longest adaptation set or chain
8925 may have more than 16 services, increase the limit beyond its
8926 default value of 16. If detecting infinite iteration loops sooner
8927 is critical, make the iteration limit match the actual number
8928 of services in your longest adaptation set or chain.
8930 Infinite adaptation loops are most likely with routing services.
8932 See also: icap_service routing=1
8935 NAME: adaptation_masterx_shared_names
8937 IFDEF: USE_ADAPTATION
8938 LOC: Adaptation::Config::masterx_shared_name
8941 For each master transaction (i.e., the HTTP request and response
8942 sequence, including all related ICAP and eCAP exchanges), Squid
8943 maintains a table of metadata. The table entries are (name, value)
8944 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8945 with the master transaction.
8947 This option specifies the table entry names that Squid must accept
8948 from and forward to the adaptation transactions.
8950 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8951 shared table by returning an ICAP header field with a name
8952 specified in adaptation_masterx_shared_names.
8954 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8955 shared table by implementing the libecap::visitEachOption() API
8956 to provide an option with a name specified in
8957 adaptation_masterx_shared_names.
8959 Squid will store and forward the set entry to subsequent adaptation
8960 transactions within the same master transaction scope.
8962 Only one shared entry name is supported at this time.
8965 # share authentication information among ICAP services
8966 adaptation_masterx_shared_names X-Subscriber-ID
8969 NAME: adaptation_meta
8971 IFDEF: USE_ADAPTATION
8972 LOC: Adaptation::Config::metaHeaders
8975 This option allows Squid administrator to add custom ICAP request
8976 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8977 Use it to pass custom authentication tokens and other
8978 transaction-state related meta information to an ICAP/eCAP service.
8980 The addition of a meta header is ACL-driven:
8981 adaptation_meta name value [!]aclname ...
8983 Processing for a given header name stops after the first ACL list match.
8984 Thus, it is impossible to add two headers with the same name. If no ACL
8985 lists match for a given header name, no such header is added. For
8988 # do not debug transactions except for those that need debugging
8989 adaptation_meta X-Debug 1 needs_debugging
8991 # log all transactions except for those that must remain secret
8992 adaptation_meta X-Log 1 !keep_secret
8994 # mark transactions from users in the "G 1" group
8995 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8997 The "value" parameter may be a regular squid.conf token or a "double
8998 quoted string". Within the quoted string, use backslash (\) to escape
8999 any character, which is currently only useful for escaping backslashes
9000 and double quotes. For example,
9001 "this string has one backslash (\\) and two \"quotes\""
9003 Used adaptation_meta header values may be logged via %note
9004 logformat code. If multiple adaptation_meta headers with the same name
9005 are used during master transaction lifetime, the header values are
9006 logged in the order they were used and duplicate values are ignored
9007 (only the first repeated value will be logged).
9013 LOC: Adaptation::Icap::TheConfig.repeat
9014 DEFAULT_IF_NONE: deny all
9016 This ACL determines which retriable ICAP transactions are
9017 retried. Transactions that received a complete ICAP response
9018 and did not have to consume or produce HTTP bodies to receive
9019 that response are usually retriable.
9021 icap_retry allow|deny [!]aclname ...
9023 Squid automatically retries some ICAP I/O timeouts and errors
9024 due to persistent connection race conditions.
9026 See also: icap_retry_limit
9029 NAME: icap_retry_limit
9032 LOC: Adaptation::Icap::TheConfig.repeat_limit
9034 DEFAULT_DOC: No retries are allowed.
9036 Limits the number of retries allowed.
9038 Communication errors due to persistent connection race
9039 conditions are unavoidable, automatically retried, and do not
9040 count against this limit.
9042 See also: icap_retry
9048 -----------------------------------------------------------------------------
9051 NAME: check_hostnames
9054 LOC: Config.onoff.check_hostnames
9056 For security and stability reasons Squid can check
9057 hostnames for Internet standard RFC compliance. If you want
9058 Squid to perform these checks turn this directive on.
9061 NAME: allow_underscore
9064 LOC: Config.onoff.allow_underscore
9066 Underscore characters is not strictly allowed in Internet hostnames
9067 but nevertheless used by many sites. Set this to off if you want
9068 Squid to be strict about the standard.
9069 This check is performed only when check_hostnames is set to on.
9072 NAME: dns_retransmit_interval
9075 LOC: Config.Timeout.idns_retransmit
9077 Initial retransmit interval for DNS queries. The interval is
9078 doubled each time all configured DNS servers have been tried.
9084 LOC: Config.Timeout.idns_query
9086 DNS Query timeout. If no response is received to a DNS query
9087 within this time all DNS servers for the queried domain
9088 are assumed to be unavailable.
9091 NAME: dns_packet_max
9093 DEFAULT_DOC: EDNS disabled
9095 LOC: Config.dns.packet_max
9097 Maximum number of bytes packet size to advertise via EDNS.
9098 Set to "none" to disable EDNS large packet support.
9100 For legacy reasons DNS UDP replies will default to 512 bytes which
9101 is too small for many responses. EDNS provides a means for Squid to
9102 negotiate receiving larger responses back immediately without having
9103 to failover with repeat requests. Responses larger than this limit
9104 will retain the old behaviour of failover to TCP DNS.
9106 Squid has no real fixed limit internally, but allowing packet sizes
9107 over 1500 bytes requires network jumbogram support and is usually not
9110 WARNING: The RFC also indicates that some older resolvers will reply
9111 with failure of the whole request if the extension is added. Some
9112 resolvers have already been identified which will reply with mangled
9113 EDNS response on occasion. Usually in response to many-KB jumbogram
9114 sizes being advertised by Squid.
9115 Squid will currently treat these both as an unable-to-resolve domain
9116 even if it would be resolvable without EDNS.
9123 DEFAULT_DOC: Search for single-label domain names is disabled.
9124 LOC: Config.onoff.res_defnames
9126 Normally the RES_DEFNAMES resolver option is disabled
9127 (see res_init(3)). This prevents caches in a hierarchy
9128 from interpreting single-component hostnames locally. To allow
9129 Squid to handle single-component names, enable this option.
9132 NAME: dns_multicast_local
9136 DEFAULT_DOC: Search for .local and .arpa names is disabled.
9137 LOC: Config.onoff.dns_mdns
9139 When set to on, Squid sends multicast DNS lookups on the local
9140 network for domains ending in .local and .arpa.
9141 This enables local servers and devices to be contacted in an
9142 ad-hoc or zero-configuration network environment.
9145 NAME: dns_nameservers
9148 DEFAULT_DOC: Use operating system definitions
9149 LOC: Config.dns_nameservers
9151 Use this if you want to specify a list of DNS name servers
9152 (IP addresses) to use instead of those given in your
9153 /etc/resolv.conf file.
9155 On Windows platforms, if no value is specified here or in
9156 the /etc/resolv.conf file, the list of DNS name servers are
9157 taken from the Windows registry, both static and dynamic DHCP
9158 configurations are supported.
9160 Example: dns_nameservers 10.0.0.1 192.172.0.4
9165 DEFAULT: @DEFAULT_HOSTS@
9166 LOC: Config.etcHostsPath
9168 Location of the host-local IP name-address associations
9169 database. Most Operating Systems have such a file on different
9171 - Un*X & Linux: /etc/hosts
9172 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
9173 (%SystemRoot% value install default is c:\winnt)
9174 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
9175 (%SystemRoot% value install default is c:\windows)
9176 - Windows 9x/Me: %windir%\hosts
9177 (%windir% value is usually c:\windows)
9178 - Cygwin: /etc/hosts
9180 The file contains newline-separated definitions, in the
9181 form ip_address_in_dotted_form name [name ...] names are
9182 whitespace-separated. Lines beginning with an hash (#)
9183 character are comments.
9185 The file is checked at startup and upon configuration.
9186 If set to 'none', it won't be checked.
9187 If append_domain is used, that domain will be added to
9188 domain-local (i.e. not containing any dot character) host
9194 LOC: Config.appendDomain
9196 DEFAULT_DOC: Use operating system definitions
9198 Appends local domain name to hostnames without any dots in
9199 them. append_domain must begin with a period.
9201 Be warned there are now Internet names with no dots in
9202 them using only top-domain names, so setting this may
9203 cause some Internet sites to become unavailable.
9206 append_domain .yourdomain.com
9209 NAME: ignore_unknown_nameservers
9211 LOC: Config.onoff.ignore_unknown_nameservers
9214 By default Squid checks that DNS responses are received
9215 from the same IP addresses they are sent to. If they
9216 don't match, Squid ignores the response and writes a warning
9217 message to cache.log. You can allow responses from unknown
9218 nameservers by setting this option to 'off'.
9224 LOC: Config.dns.v4_first
9226 With the IPv6 Internet being as fast or faster than IPv4 Internet
9227 for most networks Squid prefers to contact websites over IPv6.
9229 This option reverses the order of preference to make Squid contact
9230 dual-stack websites over IPv4 first. Squid will still perform both
9231 IPv6 and IPv4 DNS lookups before connecting.
9234 This option will restrict the situations under which IPv6
9235 connectivity is used (and tested). Hiding network problems
9236 which would otherwise be detected and warned about.
9240 COMMENT: (number of entries)
9243 LOC: Config.ipcache.size
9245 Maximum number of DNS IP cache entries.
9252 LOC: Config.ipcache.low
9259 LOC: Config.ipcache.high
9261 The size, low-, and high-water marks for the IP cache.
9264 NAME: fqdncache_size
9265 COMMENT: (number of entries)
9268 LOC: Config.fqdncache.size
9270 Maximum number of FQDN cache entries.
9275 -----------------------------------------------------------------------------
9278 NAME: configuration_includes_quoted_values
9280 TYPE: configuration_includes_quoted_values
9282 LOC: ConfigParser::RecognizeQuotedValues
9284 If set, Squid will recognize each "quoted string" after a configuration
9285 directive as a single parameter. The quotes are stripped before the
9286 parameter value is interpreted or used.
9287 See "Values with spaces, quotes, and other special characters"
9288 section for more details.
9295 LOC: Config.onoff.mem_pools
9297 If set, Squid will keep pools of allocated (but unused) memory
9298 available for future use. If memory is a premium on your
9299 system and you believe your malloc library outperforms Squid
9300 routines, disable this.
9303 NAME: memory_pools_limit
9307 LOC: Config.MemPools.limit
9309 Used only with memory_pools on:
9310 memory_pools_limit 50 MB
9312 If set to a non-zero value, Squid will keep at most the specified
9313 limit of allocated (but unused) memory in memory pools. All free()
9314 requests that exceed this limit will be handled by your malloc
9315 library. Squid does not pre-allocate any memory, just safe-keeps
9316 objects that otherwise would be free()d. Thus, it is safe to set
9317 memory_pools_limit to a reasonably high value even if your
9318 configuration will use less memory.
9320 If set to none, Squid will keep all memory it can. That is, there
9321 will be no limit on the total amount of memory used for safe-keeping.
9323 To disable memory allocation optimization, do not set
9324 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
9326 An overhead for maintaining memory pools is not taken into account
9327 when the limit is checked. This overhead is close to four bytes per
9328 object kept. However, pools may actually _save_ memory because of
9329 reduced memory thrashing in your malloc library.
9333 COMMENT: on|off|transparent|truncate|delete
9336 LOC: opt_forwarded_for
9338 If set to "on", Squid will append your client's IP address
9339 in the HTTP requests it forwards. By default it looks like:
9341 X-Forwarded-For: 192.1.2.3
9343 If set to "off", it will appear as
9345 X-Forwarded-For: unknown
9347 If set to "transparent", Squid will not alter the
9348 X-Forwarded-For header in any way.
9350 If set to "delete", Squid will delete the entire
9351 X-Forwarded-For header.
9353 If set to "truncate", Squid will remove all existing
9354 X-Forwarded-For entries, and place the client IP as the sole entry.
9357 NAME: cachemgr_passwd
9358 TYPE: cachemgrpasswd
9360 DEFAULT_DOC: No password. Actions which require password are denied.
9361 LOC: Config.passwd_list
9363 Specify passwords for cachemgr operations.
9365 Usage: cachemgr_passwd password action action ...
9367 Some valid actions are (see cache manager menu for a full list):
9407 * Indicates actions which will not be performed without a
9408 valid password, others can be performed if not listed here.
9410 To disable an action, set the password to "disable".
9411 To allow performing an action without a password, set the
9414 Use the keyword "all" to set the same password for all actions.
9417 cachemgr_passwd secret shutdown
9418 cachemgr_passwd lesssssssecret info stats/objects
9419 cachemgr_passwd disable all
9426 LOC: Config.onoff.client_db
9428 If you want to disable collecting per-client statistics,
9429 turn off client_db here.
9432 NAME: refresh_all_ims
9436 LOC: Config.onoff.refresh_all_ims
9438 When you enable this option, squid will always check
9439 the origin server for an update when a client sends an
9440 If-Modified-Since request. Many browsers use IMS
9441 requests when the user requests a reload, and this
9442 ensures those clients receive the latest version.
9444 By default (off), squid may return a Not Modified response
9445 based on the age of the cached version.
9448 NAME: reload_into_ims
9449 IFDEF: USE_HTTP_VIOLATIONS
9453 LOC: Config.onoff.reload_into_ims
9455 When you enable this option, client no-cache or ``reload''
9456 requests will be changed to If-Modified-Since requests.
9457 Doing this VIOLATES the HTTP standard. Enabling this
9458 feature could make you liable for problems which it
9461 see also refresh_pattern for a more selective approach.
9464 NAME: connect_retries
9466 LOC: Config.connect_retries
9468 DEFAULT_DOC: Do not retry failed connections.
9470 This sets the maximum number of connection attempts made for each
9471 TCP connection. The connect_retries attempts must all still
9472 complete within the connection timeout period.
9474 The default is not to re-try if the first connection attempt fails.
9475 The (not recommended) maximum is 10 tries.
9477 A warning message will be generated if it is set to a too-high
9478 value and the configured value will be over-ridden.
9480 Note: These re-tries are in addition to forward_max_tries
9481 which limit how many different addresses may be tried to find
9485 NAME: retry_on_error
9487 LOC: Config.retry.onerror
9490 If set to ON Squid will automatically retry requests when
9491 receiving an error response with status 403 (Forbidden),
9492 500 (Internal Error), 501 or 503 (Service not available).
9493 Status 502 and 504 (Gateway errors) are always retried.
9495 This is mainly useful if you are in a complex cache hierarchy to
9496 work around access control errors.
9498 NOTE: This retry will attempt to find another working destination.
9499 Which is different from the server which just failed.
9502 NAME: as_whois_server
9504 LOC: Config.as_whois_server
9505 DEFAULT: whois.ra.net
9507 WHOIS server to query for AS numbers. NOTE: AS numbers are
9508 queried only when Squid starts up, not for every request.
9513 LOC: Config.onoff.offline
9516 Enable this option and Squid will never try to validate cached
9520 NAME: uri_whitespace
9521 TYPE: uri_whitespace
9522 LOC: Config.uri_whitespace
9525 What to do with requests that have whitespace characters in the
9528 strip: The whitespace characters are stripped out of the URL.
9529 This is the behavior recommended by RFC2396 and RFC3986
9530 for tolerant handling of generic URI.
9531 NOTE: This is one difference between generic URI and HTTP URLs.
9533 deny: The request is denied. The user receives an "Invalid
9535 This is the behaviour recommended by RFC2616 for safe
9536 handling of HTTP request URL.
9538 allow: The request is allowed and the URI is not changed. The
9539 whitespace characters remain in the URI. Note the
9540 whitespace is passed to redirector processes if they
9542 Note this may be considered a violation of RFC2616
9543 request parsing where whitespace is prohibited in the
9546 encode: The request is allowed and the whitespace characters are
9547 encoded according to RFC1738.
9549 chop: The request is allowed and the URI is chopped at the
9553 NOTE the current Squid implementation of encode and chop violates
9554 RFC2616 by not using a 301 redirect after altering the URL.
9559 LOC: Config.chroot_dir
9562 Specifies a directory where Squid should do a chroot() while
9563 initializing. This also causes Squid to fully drop root
9564 privileges after initializing. This means, for example, if you
9565 use a HTTP port less than 1024 and try to reconfigure, you may
9566 get an error saying that Squid can not open the port.
9569 NAME: balance_on_multiple_ip
9571 LOC: Config.onoff.balance_on_multiple_ip
9574 Modern IP resolvers in squid sort lookup results by preferred access.
9575 By default squid will use these IP in order and only rotates to
9576 the next listed when the most preffered fails.
9578 Some load balancing servers based on round robin DNS have been
9579 found not to preserve user session state across requests
9580 to different IP addresses.
9582 Enabling this directive Squid rotates IP's per request.
9585 NAME: pipeline_prefetch
9586 TYPE: pipelinePrefetch
9587 LOC: Config.pipeline_max_prefetch
9589 DEFAULT_DOC: Do not pre-parse pipelined requests.
9591 HTTP clients may send a pipeline of 1+N requests to Squid using a
9592 single connection, without waiting for Squid to respond to the first
9593 of those requests. This option limits the number of concurrent
9594 requests Squid will try to handle in parallel. If set to N, Squid
9595 will try to receive and process up to 1+N requests on the same
9596 connection concurrently.
9598 Defaults to 0 (off) for bandwidth management and access logging
9601 NOTE: pipelining requires persistent connections to clients.
9603 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
9606 NAME: high_response_time_warning
9609 LOC: Config.warnings.high_rptm
9611 DEFAULT_DOC: disabled.
9613 If the one-minute median response time exceeds this value,
9614 Squid prints a WARNING with debug level 0 to get the
9615 administrators attention. The value is in milliseconds.
9618 NAME: high_page_fault_warning
9620 LOC: Config.warnings.high_pf
9622 DEFAULT_DOC: disabled.
9624 If the one-minute average page fault rate exceeds this
9625 value, Squid prints a WARNING with debug level 0 to get
9626 the administrators attention. The value is in page faults
9630 NAME: high_memory_warning
9632 LOC: Config.warnings.high_memory
9633 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
9635 DEFAULT_DOC: disabled.
9637 If the memory usage (as determined by gnumalloc, if available and used)
9638 exceeds this amount, Squid prints a WARNING with debug level 0 to get
9639 the administrators attention.
9641 # TODO: link high_memory_warning to mempools?
9643 NAME: sleep_after_fork
9644 COMMENT: (microseconds)
9646 LOC: Config.sleep_after_fork
9649 When this is set to a non-zero value, the main Squid process
9650 sleeps the specified number of microseconds after a fork()
9651 system call. This sleep may help the situation where your
9652 system reports fork() failures due to lack of (virtual)
9653 memory. Note, however, if you have a lot of child
9654 processes, these sleep delays will add up and your
9655 Squid will not service requests for some amount of time
9656 until all the child processes have been started.
9657 On Windows value less then 1000 (1 milliseconds) are
9661 NAME: windows_ipaddrchangemonitor
9662 IFDEF: _SQUID_WINDOWS_
9666 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9668 On Windows Squid by default will monitor IP address changes and will
9669 reconfigure itself after any detected event. This is very useful for
9670 proxies connected to internet with dial-up interfaces.
9671 In some cases (a Proxy server acting as VPN gateway is one) it could be
9672 desiderable to disable this behaviour setting this to 'off'.
9673 Note: after changing this, Squid service must be restarted.
9678 IFDEF: USE_SQUID_EUI
9680 LOC: Eui::TheConfig.euiLookup
9682 Whether to lookup the EUI or MAC address of a connected client.
9685 NAME: max_filedescriptors max_filedesc
9688 DEFAULT_DOC: Use operating system limits set by ulimit.
9689 LOC: Config.max_filedescriptors
9691 Reduce the maximum number of filedescriptors supported below
9692 the usual operating system defaults.
9694 Remove from squid.conf to inherit the current ulimit setting.
9696 Note: Changing this requires a restart of Squid. Also
9697 not all I/O types supports large values (eg on Windows).
9700 NAME: force_request_body_continuation
9702 LOC: Config.accessList.forceRequestBodyContinuation
9704 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
9706 This option controls how Squid handles data upload requests from HTTP
9707 and FTP agents that require a "Please Continue" control message response
9708 to actually send the request body to Squid. It is mostly useful in
9709 adaptation environments.
9711 When Squid receives an HTTP request with an "Expect: 100-continue"
9712 header or an FTP upload command (e.g., STOR), Squid normally sends the
9713 request headers or FTP command information to an adaptation service (or
9714 peer) and waits for a response. Most adaptation services (and some
9715 broken peers) may not respond to Squid at that stage because they may
9716 decide to wait for the HTTP request body or FTP data transfer. However,
9717 that request body or data transfer may never come because Squid has not
9718 responded with the HTTP 100 or FTP 150 (Please Continue) control message
9719 to the request sender yet!
9721 An allow match tells Squid to respond with the HTTP 100 or FTP 150
9722 (Please Continue) control message on its own, before forwarding the
9723 request to an adaptation service or peer. Such a response usually forces
9724 the request sender to proceed with sending the body. A deny match tells
9725 Squid to delay that control response until the origin server confirms
9726 that the request body is needed. Delaying is the default behavior.
9729 NAME: server_pconn_for_nonretriable
9732 DEFAULT_DOC: Open new connections for forwarding requests Squid cannot retry safely.
9733 LOC: Config.accessList.serverPconnForNonretriable
9735 This option provides fine-grained control over persistent connection
9736 reuse when forwarding HTTP requests that Squid cannot retry. It is useful
9737 in environments where opening new connections is very expensive
9738 (e.g., all connections are secured with TLS with complex client and server
9739 certificate validation) and race conditions associated with persistent
9740 connections are very rare and/or only cause minor problems.
9742 HTTP prohibits retrying unsafe and non-idempotent requests (e.g., POST).
9743 Squid limitations also prohibit retrying all requests with bodies (e.g., PUT).
9744 By default, when forwarding such "risky" requests, Squid opens a new
9745 connection to the server or cache_peer, even if there is an idle persistent
9746 connection available. When Squid is configured to risk sending a non-retriable
9747 request on a previously used persistent connection, and the server closes
9748 the connection before seeing that risky request, the user gets an error response
9749 from Squid. In most cases, that error response will be HTTP 502 (Bad Gateway)
9750 with ERR_ZERO_SIZE_OBJECT or ERR_WRITE_ERROR (peer connection reset) error detail.
9752 If an allow rule matches, Squid reuses an available idle persistent connection
9753 (if any) for the request that Squid cannot retry. If a deny rule matches, then
9754 Squid opens a new connection for the request that Squid cannot retry.
9756 This option does not affect requests that Squid can retry. They will reuse idle
9757 persistent connections (if any).
9759 This clause only supports fast acl types.
9760 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
9763 acl SpeedIsWorthTheRisk method POST
9764 server_pconn_for_nonretriable allow SpeedIsWorthTheRisk