1 ## Copyright (C) 1996-2015 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 # options still not yet ported from 2.7 to 3.x
115 NAME: broken_vary_encoding
118 This option is not yet supported by Squid-3.
124 This option is not yet supported by Squid-3.
130 This option is not yet supported by Squid-3.
133 NAME: external_refresh_check
136 This option is not yet supported by Squid-3.
139 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
142 This option is not yet supported by Squid-3.
145 NAME: refresh_stale_hit
148 This option is not yet supported by Squid-3.
151 # Options removed in 3.6
152 NAME: cache_peer_domain cache_host_domain
155 Replace with dstdomain ACLs and cache_peer_access.
158 NAME: sslproxy_cafile
161 Remove this line. Use tls_outgoing_options cafile= instead.
164 NAME: sslproxy_capath
167 Remove this line. Use tls_outgoing_options capath= instead.
170 NAME: sslproxy_cipher
173 Remove this line. Use tls_outgoing_options cipher= instead.
176 NAME: sslproxy_client_certificate
179 Remove this line. Use tls_outgoing_options cert= instead.
182 NAME: sslproxy_client_key
185 Remove this line. Use tls_outgoing_options key= instead.
191 Remove this line. Use tls_outgoing_options flags= instead.
194 NAME: sslproxy_options
197 Remove this line. Use tls_outgoing_options options= instead.
200 NAME: sslproxy_version
203 Remove this line. Use tls_outgoing_options options= instead.
206 # Options removed in 3.5
207 NAME: hierarchy_stoplist
210 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
213 # Options removed in 3.4
217 Remove this line. Use acls with access_log directives to control access logging
223 Remove this line. Use acls with icap_log directives to control icap logging
226 # Options Removed in 3.3
227 NAME: ignore_ims_on_miss
230 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
233 # Options Removed in 3.2
234 NAME: chunked_request_body_max_size
237 Remove this line. Squid is now HTTP/1.1 compliant.
240 NAME: dns_v4_fallback
243 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
246 NAME: emulate_httpd_log
249 Replace this with an access_log directive using the format 'common' or 'combined'.
255 Use a regular access.log with ACL limiting it to MISS events.
261 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
264 NAME: ignore_expect_100
267 Remove this line. The HTTP/1.1 feature is now fully supported by default.
273 Remove this option from your config. To log FQDN use %>A in the log format.
276 NAME: log_ip_on_direct
279 Remove this option from your config. To log server or peer names use %<A in the log format.
282 NAME: maximum_single_addr_tries
285 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
288 NAME: referer_log referrer_log
291 Replace this with an access_log directive using the format 'referrer'.
297 Remove this line. The feature is supported by default in storage types where update is implemented.
300 NAME: url_rewrite_concurrency
303 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
309 Replace this with an access_log directive using the format 'useragent'.
312 # Options Removed in 3.1
316 Remove this line. DNS is no longer tested on startup.
319 NAME: extension_methods
322 Remove this line. All valid methods for HTTP are accepted by default.
325 # 2.7 Options Removed/Replaced in 3.2
330 # 2.7 Options Removed/Replaced in 3.1
338 Remove this line. HTTP/1.1 is supported by default.
341 NAME: upgrade_http0.9
344 Remove this line. ICY/1.0 streaming protocol is supported by default.
347 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
350 Alter these entries. Use the qos_flows directive instead.
353 # Options Removed in 3.0
357 Since squid-3.0 replace with request_header_access or reply_header_access
358 depending on whether you wish to match client requests or server replies.
361 NAME: httpd_accel_no_pmtu_disc
364 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
367 NAME: wais_relay_host
370 Replace this line with 'cache_peer' configuration.
373 NAME: wais_relay_port
376 Replace this line with 'cache_peer' configuration.
381 -----------------------------------------------------------------------------
388 DEFAULT_DOC: SMP support disabled.
390 Number of main Squid processes or "workers" to fork and maintain.
391 0: "no daemon" mode, like running "squid -N ..."
392 1: "no SMP" mode, start one main Squid process daemon (default)
393 N: start N main Squid process daemons (i.e., SMP mode)
395 In SMP mode, each worker does nearly all what a single Squid daemon
396 does (e.g., listen on http_port and forward HTTP requests).
399 NAME: cpu_affinity_map
401 LOC: Config.cpuAffinityMap
403 DEFAULT_DOC: Let operating system decide.
405 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
407 Sets 1:1 mapping between Squid processes and CPU cores. For example,
409 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
411 affects processes 1 through 4 only and places them on the first
412 four even cores, starting with core #1.
414 CPU cores are numbered starting from 1. Requires support for
415 sched_getaffinity(2) and sched_setaffinity(2) system calls.
417 Multiple cpu_affinity_map options are merged.
423 OPTIONS FOR AUTHENTICATION
424 -----------------------------------------------------------------------------
433 This is used to define parameters for the various authentication
434 schemes supported by Squid.
436 format: auth_param scheme parameter [setting]
438 The order in which authentication schemes are presented to the client is
439 dependent on the order the scheme first appears in config file. IE
440 has a bug (it's not RFC 2617 compliant) in that it will use the basic
441 scheme if basic is the first entry presented, even if more secure
442 schemes are presented. For now use the order in the recommended
443 settings section below. If other browsers have difficulties (don't
444 recognize the schemes offered even if you are using basic) either
445 put basic first, or disable the other schemes (by commenting out their
448 Once an authentication scheme is fully configured, it can only be
449 shutdown by shutting squid down and restarting. Changes can be made on
450 the fly and activated with a reconfigure. I.E. You can change to a
451 different helper, but not unconfigure the helper completely.
453 Please note that while this directive defines how Squid processes
454 authentication it does not automatically activate authentication.
455 To use authentication you must in addition make use of ACLs based
456 on login name in http_access (proxy_auth, proxy_auth_regex or
457 external with %LOGIN used in the format tag). The browser will be
458 challenged for authentication on the first such acl encountered
459 in http_access processing and will also be re-challenged for new
460 login credentials if the request is being denied by a proxy_auth
463 WARNING: authentication can't be used in a transparently intercepting
464 proxy as the client then thinks it is talking to an origin server and
465 not the proxy. This is a limitation of bending the TCP/IP protocol to
466 transparently intercepting port 80, not a limitation in Squid.
467 Ports flagged 'transparent', 'intercept', or 'tproxy' have
468 authentication disabled.
470 === Parameters common to all schemes. ===
473 Specifies the command for the external authenticator.
475 By default, each authentication scheme is not used unless a
476 program is specified.
478 See http://wiki.squid-cache.org/Features/AddonHelpers for
479 more details on helper operations and creating your own.
482 Specifies a string to be append to request line format for
483 the authentication helper. "Quoted" format values may contain
484 spaces and logformat %macros. In theory, any logformat %macro
485 can be used. In practice, a %macro expands as a dash (-) if
486 the helper request is sent before the required macro
487 information is available to Squid.
489 By default, Squid uses request formats provided in
490 scheme-specific examples below (search for %credentials).
492 The expanded key_extras value is added to the Squid credentials
493 cache and, hence, will affect authentication. It can be used to
494 autenticate different users with identical user names (e.g.,
495 when user authentication depends on http_port).
497 Avoid adding frequently changing information to key_extras. For
498 example, if you add user source IP, and it changes frequently
499 in your environment, then max_user_ip ACL is going to treat
500 every user+IP combination as a unique "user", breaking the ACL
501 and wasting a lot of memory on those user records. It will also
502 force users to authenticate from scratch whenever their IP
506 Specifies the protection scope (aka realm name) which is to be
507 reported to the client for the authentication scheme. It is
508 commonly part of the text the user will see when prompted for
509 their username and password.
511 For Basic the default is "Squid proxy-caching web server".
512 For Digest there is no default, this parameter is mandatory.
513 For NTLM and Negotiate this parameter is ignored.
515 "children" numberofchildren [startup=N] [idle=N] [concurrency=N] [queue-size=N]
517 The maximum number of authenticator processes to spawn. If
518 you start too few Squid will have to wait for them to process
519 a backlog of credential verifications, slowing it down. When
520 password verifications are done via a (slow) network you are
521 likely to need lots of authenticator processes.
523 The startup= and idle= options permit some skew in the exact
524 amount run. A minimum of startup=N will begin during startup
525 and reconfigure. Squid will start more in groups of up to
526 idle=N in an attempt to meet traffic needs and to keep idle=N
527 free above those traffic needs up to the maximum.
529 The concurrency= option sets the number of concurrent requests
530 the helper can process. The default of 0 is used for helpers
531 who only supports one request at a time. Setting this to a
532 number greater than 0 changes the protocol used to include a
533 channel ID field first on the request/response line, allowing
534 multiple requests to be sent to the same helper in parallel
535 without waiting for the response.
537 Concurrency must not be set unless it's known the helper
538 supports the input format with channel-ID fields.
540 The queue-size= option sets the maximum number of queued
541 requests. If the queued requests exceed queue size for more
542 than 3 minutes then squid aborts its operation.
543 The default value is set to 2*numberofchildren/
545 NOTE: NTLM and Negotiate schemes do not support concurrency
546 in the Squid code module even though some helpers can.
549 IF HAVE_AUTH_MODULE_BASIC
550 === Basic authentication parameters ===
553 HTTP uses iso-latin-1 as character set, while some
554 authentication backends such as LDAP expects UTF-8. If this is
555 set to on Squid will translate the HTTP iso-latin-1 charset to
556 UTF-8 before sending the username and password to the helper.
558 "credentialsttl" timetolive
559 Specifies how long squid assumes an externally validated
560 username:password pair is valid for - in other words how
561 often the helper program is called for that user. Set this
562 low to force revalidation with short lived passwords.
564 NOTE: setting this high does not impact your susceptibility
565 to replay attacks unless you are using an one-time password
566 system (such as SecureID). If you are using such a system,
567 you will be vulnerable to replay attacks unless you also
568 use the max_user_ip ACL in an http_access rule.
570 "casesensitive" on|off
571 Specifies if usernames are case sensitive. Most user databases
572 are case insensitive allowing the same username to be spelled
573 using both lower and upper case letters, but some are case
574 sensitive. This makes a big difference for user_max_ip ACL
575 processing and similar.
578 IF HAVE_AUTH_MODULE_DIGEST
579 === Digest authentication parameters ===
582 HTTP uses iso-latin-1 as character set, while some
583 authentication backends such as LDAP expects UTF-8. If this is
584 set to on Squid will translate the HTTP iso-latin-1 charset to
585 UTF-8 before sending the username and password to the helper.
587 "nonce_garbage_interval" timeinterval
588 Specifies the interval that nonces that have been issued
589 to client_agent's are checked for validity.
591 "nonce_max_duration" timeinterval
592 Specifies the maximum length of time a given nonce will be
595 "nonce_max_count" number
596 Specifies the maximum number of times a given nonce can be
599 "nonce_strictness" on|off
600 Determines if squid requires strict increment-by-1 behavior
601 for nonce counts, or just incrementing (off - for use when
602 user agents generate nonce counts that occasionally miss 1
603 (ie, 1,2,4,6)). Default off.
605 "check_nonce_count" on|off
606 This directive if set to off can disable the nonce count check
607 completely to work around buggy digest qop implementations in
608 certain mainstream browser versions. Default on to check the
609 nonce count to protect from authentication replay attacks.
611 "post_workaround" on|off
612 This is a workaround to certain buggy browsers who send an
613 incorrect request digest in POST requests when reusing the
614 same nonce as acquired earlier on a GET request.
617 IF HAVE_AUTH_MODULE_NEGOTIATE
618 === Negotiate authentication parameters ===
621 If you experience problems with PUT/POST requests when using
622 the this authentication scheme then you can try setting this
623 to off. This will cause Squid to forcibly close the connection
624 on the initial request where the browser asks which schemes
625 are supported by the proxy.
628 IF HAVE_AUTH_MODULE_NTLM
629 === NTLM authentication parameters ===
632 If you experience problems with PUT/POST requests when using
633 the this authentication scheme then you can try setting this
634 to off. This will cause Squid to forcibly close the connection
635 on the initial request where the browser asks which schemes
636 are supported by the proxy.
639 === Example Configuration ===
641 This configuration displays the recommended authentication scheme
642 order from most to least secure with recommended minimum configuration
643 settings for each scheme:
645 #auth_param negotiate program <uncomment and complete this line to activate>
646 #auth_param negotiate children 20 startup=0 idle=1
647 #auth_param negotiate keep_alive on
649 #auth_param digest program <uncomment and complete this line to activate>
650 #auth_param digest children 20 startup=0 idle=1
651 #auth_param digest realm Squid proxy-caching web server
652 #auth_param digest nonce_garbage_interval 5 minutes
653 #auth_param digest nonce_max_duration 30 minutes
654 #auth_param digest nonce_max_count 50
656 #auth_param ntlm program <uncomment and complete this line to activate>
657 #auth_param ntlm children 20 startup=0 idle=1
658 #auth_param ntlm keep_alive on
660 #auth_param basic program <uncomment and complete this line>
661 #auth_param basic children 5 startup=5 idle=1
662 #auth_param basic realm Squid proxy-caching web server
663 #auth_param basic credentialsttl 2 hours
666 NAME: authenticate_cache_garbage_interval
669 LOC: Config.authenticateGCInterval
671 The time period between garbage collection across the username cache.
672 This is a trade-off between memory utilization (long intervals - say
673 2 days) and CPU (short intervals - say 1 minute). Only change if you
677 NAME: authenticate_ttl
680 LOC: Config.authenticateTTL
682 The time a user & their credentials stay in the logged in
683 user cache since their last request. When the garbage
684 interval passes, all user credentials that have passed their
685 TTL are removed from memory.
688 NAME: authenticate_ip_ttl
690 LOC: Config.authenticateIpTTL
693 If you use proxy authentication and the 'max_user_ip' ACL,
694 this directive controls how long Squid remembers the IP
695 addresses associated with each user. Use a small value
696 (e.g., 60 seconds) if your users might change addresses
697 quickly, as is the case with dialup. You might be safe
698 using a larger value (e.g., 2 hours) in a corporate LAN
699 environment with relatively static address assignments.
704 -----------------------------------------------------------------------------
707 NAME: external_acl_type
708 TYPE: externalAclHelper
709 LOC: Config.externalAclHelperList
712 This option defines external acl classes using a helper program
713 to look up the status
715 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
719 ttl=n TTL in seconds for cached results (defaults to 3600
723 TTL for cached negative lookups (default same
726 grace=n Percentage remaining of TTL where a refresh of a
727 cached entry should be initiated without needing to
728 wait for a new reply. (default is for no grace period)
730 cache=n Limit the result cache size, default is 262144.
731 The expanded FORMAT value is used as the cache key, so
732 if the details in FORMAT are highly variable a larger
733 cache may be needed to produce reduction in helper load.
736 Maximum number of acl helper processes spawned to service
737 external acl lookups of this type. (default 20)
740 Minimum number of acl helper processes to spawn during
741 startup and reconfigure to service external acl lookups
742 of this type. (default 0)
745 Number of acl helper processes to keep ahead of traffic
746 loads. Squid will spawn this many at once whenever load
747 rises above the capabilities of existing processes.
748 Up to the value of children-max. (default 1)
750 concurrency=n concurrency level per process. Only used with helpers
751 capable of processing more than one query at a time.
753 queue-size=N The queue-size= option sets the maximum number of queued
754 requests. If the queued requests exceed queue size
756 The default value is set to 2*children-max.
758 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
760 ipv4 / ipv6 IP protocol used to communicate with this helper.
761 The default is to auto-detect IPv6 and use it when available.
764 FORMAT specifications
766 %LOGIN Authenticated user login name
767 %EXT_USER Username from previous external acl
768 %EXT_LOG Log details from previous external acl
769 %EXT_TAG Tag from previous external acl
770 %IDENT Ident user name
772 %SRCPORT Client source port
775 %PROTO Requested URL scheme
777 %PATH Requested URL path
778 %METHOD Request method
779 %MYADDR Squid interface address
780 %MYPORT Squid http_port number
781 %PATH Requested URL-path (including query-string if any)
782 %USER_CERT SSL User certificate in PEM format
783 %USER_CERTCHAIN SSL User certificate chain in PEM format
784 %USER_CERT_xx SSL User certificate subject attribute xx
785 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
786 %ssl::>sni SSL client SNI sent to Squid
787 %ssl::<cert_subject SSL server certificate DN
788 %ssl::<cert_issuer SSL server certificate issuer DN
790 %>{Header} HTTP request header "Header"
792 HTTP request header "Hdr" list member "member"
794 HTTP request header list member using ; as
795 list separator. ; can be any non-alphanumeric
798 %<{Header} HTTP reply header "Header"
800 HTTP reply header "Hdr" list member "member"
802 HTTP reply header list member using ; as
803 list separator. ; can be any non-alphanumeric
806 %ACL The name of the ACL being tested.
807 %DATA The ACL arguments. If not used then any arguments
808 is automatically added at the end of the line
810 NOTE: this will encode the arguments as one token,
811 whereas the default will pass each separately.
813 %% The percent sign. Useful for helpers which need
814 an unchanging input format.
817 General request syntax:
819 [channel-ID] FORMAT-values [acl-values ...]
822 FORMAT-values consists of transaction details expanded with
823 whitespace separation per the config file FORMAT specification
824 using the FORMAT macros listed above.
826 acl-values consists of any string specified in the referencing
827 config 'acl ... external' line. see the "acl external" directive.
829 Request values sent to the helper are URL escaped to protect
830 each value in requests against whitespaces.
832 If using protocol=2.5 then the request sent to the helper is not
833 URL escaped to protect against whitespace.
835 NOTE: protocol=3.0 is deprecated as no longer necessary.
837 When using the concurrency= option the protocol is changed by
838 introducing a query channel tag in front of the request/response.
839 The query channel tag is a number between 0 and concurrency-1.
840 This value must be echoed back unchanged to Squid as the first part
841 of the response relating to its request.
844 The helper receives lines expanded per the above format specification
845 and for each input line returns 1 line starting with OK/ERR/BH result
846 code and optionally followed by additional keywords with more details.
849 General result syntax:
851 [channel-ID] result keyword=value ...
853 Result consists of one of the codes:
856 the ACL test produced a match.
859 the ACL test does not produce a match.
862 An internal error occurred in the helper, preventing
863 a result being identified.
865 The meaning of 'a match' is determined by your squid.conf
866 access control configuration. See the Squid wiki for details.
870 user= The users name (login)
872 password= The users password (for login= cache_peer option)
874 message= Message describing the reason for this response.
875 Available as %o in error pages.
876 Useful on (ERR and BH results).
878 tag= Apply a tag to a request. Only sets a tag once,
879 does not alter existing tags.
881 log= String to be logged in access.log. Available as
882 %ea in logformat specifications.
884 clt_conn_tag= Associates a TAG with the client TCP connection.
885 Please see url_rewrite_program related documentation
888 Any keywords may be sent on any response whether OK, ERR or BH.
890 All response keyword values need to be a single token with URL
891 escaping, or enclosed in double quotes (") and escaped using \ on
892 any double quotes or \ characters within the value. The wrapping
893 double quotes are removed before the value is interpreted by Squid.
894 \r and \n are also replace by CR and LF.
896 Some example key values:
900 user="J. \"Bob\" Smith"
907 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
908 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
909 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
910 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
911 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
914 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
915 DEFAULT: localhost src 127.0.0.1/32 ::1
916 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
917 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
919 Defining an Access List
921 Every access list definition must begin with an aclname and acltype,
922 followed by either type-specific arguments or a quoted filename that
925 acl aclname acltype argument ...
926 acl aclname acltype "file" ...
928 When using "file", the file should contain one item per line.
930 Some acl types supports options which changes their default behaviour.
931 The available options are:
933 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
934 case-insensitive, use the -i option. To return case-sensitive
935 use the +i option between patterns, or make a new ACL line
938 -n Disable lookups and address type conversions. If lookup or
939 conversion is required because the parameter type (IP or
940 domain name) does not match the message address type (domain
941 name or IP), then the ACL would immediately declare a mismatch
942 without any warnings or lookups.
944 -- Used to stop processing all options, in the case the first acl
945 value has '-' character as first character (for example the '-'
946 is a valid domain name)
948 Some acl types require suspending the current request in order
949 to access some external data source.
950 Those which do are marked with the tag [slow], those which
951 don't are marked as [fast].
952 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
953 for further information
955 ***** ACL TYPES AVAILABLE *****
957 acl aclname src ip-address/mask ... # clients IP address [fast]
958 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
959 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
960 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
962 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
963 # The arp ACL requires the special configure option --enable-arp-acl.
964 # Furthermore, the ARP ACL code is not portable to all operating systems.
965 # It works on Linux, Solaris, Windows, FreeBSD, and some
966 # other *BSD variants.
969 # NOTE: Squid can only determine the MAC address for clients that are on
970 # the same subnet. If the client is on a different subnet,
971 # then Squid cannot find out its MAC address.
973 acl aclname srcdomain .foo.com ...
974 # reverse lookup, from client IP [slow]
975 acl aclname dstdomain [-n] .foo.com ...
976 # Destination server from URL [fast]
977 acl aclname srcdom_regex [-i] \.foo\.com ...
978 # regex matching client name [slow]
979 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
980 # regex matching server [fast]
982 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
983 # based URL is used and no match is found. The name "none" is used
984 # if the reverse lookup fails.
986 acl aclname src_as number ...
987 acl aclname dst_as number ...
989 # Except for access control, AS numbers can be used for
990 # routing of requests to specific caches. Here's an
991 # example for routing all requests for AS#1241 and only
992 # those to mycache.mydomain.net:
993 # acl asexample dst_as 1241
994 # cache_peer_access mycache.mydomain.net allow asexample
995 # cache_peer_access mycache_mydomain.net deny all
997 acl aclname peername myPeer ...
999 # match against a named cache_peer entry
1000 # set unique name= on cache_peer lines for reliable use.
1002 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
1012 # h1:m1 must be less than h2:m2
1014 acl aclname url_regex [-i] ^http:// ...
1015 # regex matching on whole URL [fast]
1016 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
1017 # regex matching on URL login field
1018 acl aclname urlpath_regex [-i] \.gif$ ...
1019 # regex matching on URL path [fast]
1021 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
1023 acl aclname localport 3128 ... # TCP port the client connected to [fast]
1024 # NP: for interception mode this is usually '80'
1026 acl aclname myportname 3128 ... # *_port name [fast]
1028 acl aclname proto HTTP FTP ... # request protocol [fast]
1030 acl aclname method GET POST ... # HTTP request method [fast]
1032 acl aclname http_status 200 301 500- 400-403 ...
1033 # status code in reply [fast]
1035 acl aclname browser [-i] regexp ...
1036 # pattern match on User-Agent header (see also req_header below) [fast]
1038 acl aclname referer_regex [-i] regexp ...
1039 # pattern match on Referer header [fast]
1040 # Referer is highly unreliable, so use with care
1042 acl aclname ident username ...
1043 acl aclname ident_regex [-i] pattern ...
1044 # string match on ident output [slow]
1045 # use REQUIRED to accept any non-null ident.
1047 acl aclname proxy_auth [-i] username ...
1048 acl aclname proxy_auth_regex [-i] pattern ...
1049 # perform http authentication challenge to the client and match against
1050 # supplied credentials [slow]
1052 # takes a list of allowed usernames.
1053 # use REQUIRED to accept any valid username.
1055 # Will use proxy authentication in forward-proxy scenarios, and plain
1056 # http authenticaiton in reverse-proxy scenarios
1058 # NOTE: when a Proxy-Authentication header is sent but it is not
1059 # needed during ACL checking the username is NOT logged
1062 # NOTE: proxy_auth requires a EXTERNAL authentication program
1063 # to check username/password combinations (see
1064 # auth_param directive).
1066 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
1067 # as the browser needs to be configured for using a proxy in order
1068 # to respond to proxy authentication.
1070 acl aclname snmp_community string ...
1071 # A community string to limit access to your SNMP Agent [fast]
1074 # acl snmppublic snmp_community public
1076 acl aclname maxconn number
1077 # This will be matched when the client's IP address has
1078 # more than <number> TCP connections established. [fast]
1079 # NOTE: This only measures direct TCP links so X-Forwarded-For
1080 # indirect clients are not counted.
1082 acl aclname max_user_ip [-s] number
1083 # This will be matched when the user attempts to log in from more
1084 # than <number> different ip addresses. The authenticate_ip_ttl
1085 # parameter controls the timeout on the ip entries. [fast]
1086 # If -s is specified the limit is strict, denying browsing
1087 # from any further IP addresses until the ttl has expired. Without
1088 # -s Squid will just annoy the user by "randomly" denying requests.
1089 # (the counter is reset each time the limit is reached and a
1090 # request is denied)
1091 # NOTE: in acceleration mode or where there is mesh of child proxies,
1092 # clients may appear to come from multiple addresses if they are
1093 # going through proxy farms, so a limit of 1 may cause user problems.
1095 acl aclname random probability
1096 # Pseudo-randomly match requests. Based on the probability given.
1097 # Probability may be written as a decimal (0.333), fraction (1/3)
1098 # or ratio of matches:non-matches (3:5).
1100 acl aclname req_mime_type [-i] mime-type ...
1101 # regex match against the mime type of the request generated
1102 # by the client. Can be used to detect file upload or some
1103 # types HTTP tunneling requests [fast]
1104 # NOTE: This does NOT match the reply. You cannot use this
1105 # to match the returned file type.
1107 acl aclname req_header header-name [-i] any\.regex\.here
1108 # regex match against any of the known request headers. May be
1109 # thought of as a superset of "browser", "referer" and "mime-type"
1112 acl aclname rep_mime_type [-i] mime-type ...
1113 # regex match against the mime type of the reply received by
1114 # squid. Can be used to detect file download or some
1115 # types HTTP tunneling requests. [fast]
1116 # NOTE: This has no effect in http_access rules. It only has
1117 # effect in rules that affect the reply data stream such as
1118 # http_reply_access.
1120 acl aclname rep_header header-name [-i] any\.regex\.here
1121 # regex match against any of the known reply headers. May be
1122 # thought of as a superset of "browser", "referer" and "mime-type"
1125 acl aclname external class_name [arguments...]
1126 # external ACL lookup via a helper class defined by the
1127 # external_acl_type directive [slow]
1129 acl aclname user_cert attribute values...
1130 # match against attributes in a user SSL certificate
1131 # attribute is one of DN/C/O/CN/L/ST [fast]
1133 acl aclname ca_cert attribute values...
1134 # match against attributes a users issuing CA SSL certificate
1135 # attribute is one of DN/C/O/CN/L/ST [fast]
1137 acl aclname ext_user username ...
1138 acl aclname ext_user_regex [-i] pattern ...
1139 # string match on username returned by external acl helper [slow]
1140 # use REQUIRED to accept any non-null user name.
1142 acl aclname tag tagvalue ...
1143 # string match on tag returned by external acl helper [fast]
1144 # DEPRECATED. Only the first tag will match with this ACL.
1145 # Use the 'note' ACL instead for handling multiple tag values.
1147 acl aclname hier_code codename ...
1148 # string match against squid hierarchy code(s); [fast]
1149 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1151 # NOTE: This has no effect in http_access rules. It only has
1152 # effect in rules that affect the reply data stream such as
1153 # http_reply_access.
1155 acl aclname note name [value ...]
1156 # match transaction annotation [fast]
1157 # Without values, matches any annotation with a given name.
1158 # With value(s), matches any annotation with a given name that
1159 # also has one of the given values.
1160 # Names and values are compared using a string equality test.
1161 # Annotation sources include note and adaptation_meta directives
1162 # as well as helper and eCAP responses.
1164 acl aclname adaptation_service service ...
1165 # Matches the name of any icap_service, ecap_service,
1166 # adaptation_service_set, or adaptation_service_chain that Squid
1167 # has used (or attempted to use) for the master transaction.
1168 # This ACL must be defined after the corresponding adaptation
1169 # service is named in squid.conf. This ACL is usable with
1170 # adaptation_meta because it starts matching immediately after
1171 # the service has been selected for adaptation.
1174 acl aclname ssl_error errorname
1175 # match against SSL certificate validation error [fast]
1177 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1180 # The following can be used as shortcuts for certificate properties:
1181 # [ssl::]certHasExpired: the "not after" field is in the past
1182 # [ssl::]certNotYetValid: the "not before" field is in the future
1183 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1184 # [ssl::]certSelfSigned: The certificate is self signed.
1185 # [ssl::]certDomainMismatch: The certificate CN domain does not
1186 # match the name the name of the host we are connecting to.
1188 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1189 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1190 # predefined ACLs, just like the 'all' ACL.
1192 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1193 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1195 acl aclname server_cert_fingerprint [-sha1] fingerprint
1196 # match against server SSL certificate fingerprint [fast]
1198 # The fingerprint is the digest of the DER encoded version
1199 # of the whole certificate. The user should use the form: XX:XX:...
1200 # Optional argument specifies the digest algorithm to use.
1201 # The SHA1 digest algorithm is the default and is currently
1202 # the only algorithm supported (-sha1).
1204 acl aclname at_step step
1205 # match against the current step during ssl_bump evaluation [fast]
1206 # Never matches and should not be used outside the ssl_bump context.
1208 # At each SslBump step, Squid evaluates ssl_bump directives to find
1209 # the next bumping action (e.g., peek or splice). Valid SslBump step
1210 # values and the corresponding ssl_bump evaluation moments are:
1211 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1212 # SslBump2: After getting SSL Client Hello info.
1213 # SslBump3: After getting SSL Server Hello info.
1215 acl aclname ssl::server_name .foo.com ...
1216 # matches server name obtained from various sources [fast]
1218 # The server name is obtained during Ssl-Bump steps from such sources
1219 # as CONNECT request URI, client SNI, and SSL server certificate CN.
1220 # During each Ssl-Bump step, Squid may improve its understanding of a
1221 # "true server name". Unlike dstdomain, this ACL does not perform
1224 acl aclname ssl::server_name_regex [-i] \.foo\.com ...
1225 # regex matches server name obtained from various sources [fast]
1227 acl aclname any-of acl1 acl2 ...
1228 # match any one of the acls [fast or slow]
1229 # The first matching ACL stops further ACL evaluation.
1231 # ACLs from multiple any-of lines with the same name are ORed.
1232 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1233 # acl A any-of a1 a2
1234 # acl A any-of a3 a4
1236 # This group ACL is fast if all evaluated ACLs in the group are fast
1237 # and slow otherwise.
1239 acl aclname all-of acl1 acl2 ...
1240 # match all of the acls [fast or slow]
1241 # The first mismatching ACL stops further ACL evaluation.
1243 # ACLs from multiple all-of lines with the same name are ORed.
1244 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1245 # acl B all-of b1 b2
1246 # acl B all-of b3 b4
1248 # This group ACL is fast if all evaluated ACLs in the group are fast
1249 # and slow otherwise.
1252 acl macaddress arp 09:00:2b:23:45:67
1253 acl myexample dst_as 1241
1254 acl password proxy_auth REQUIRED
1255 acl fileupload req_mime_type -i ^multipart/form-data$
1256 acl javascript rep_mime_type -i ^application/x-javascript$
1260 # Recommended minimum configuration:
1263 # Example rule allowing access from your local networks.
1264 # Adapt to list your (internal) IP networks from where browsing
1266 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1267 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1268 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1269 acl localhet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1270 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1271 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1272 acl localnet src fc00::/7 # RFC 4193 local private network range
1273 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1275 acl SSL_ports port 443
1276 acl Safe_ports port 80 # http
1277 acl Safe_ports port 21 # ftp
1278 acl Safe_ports port 443 # https
1279 acl Safe_ports port 70 # gopher
1280 acl Safe_ports port 210 # wais
1281 acl Safe_ports port 1025-65535 # unregistered ports
1282 acl Safe_ports port 280 # http-mgmt
1283 acl Safe_ports port 488 # gss-http
1284 acl Safe_ports port 591 # filemaker
1285 acl Safe_ports port 777 # multiling http
1286 acl CONNECT method CONNECT
1290 NAME: proxy_protocol_access
1292 LOC: Config.accessList.proxyProtocol
1294 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1296 Determine which client proxies can be trusted to provide correct
1297 information regarding real client IP address using PROXY protocol.
1299 Requests may pass through a chain of several other proxies
1300 before reaching us. The original source details may by sent in:
1301 * HTTP message Forwarded header, or
1302 * HTTP message X-Forwarded-For header, or
1303 * PROXY protocol connection header.
1305 This directive is solely for validating new PROXY protocol
1306 connections received from a port flagged with require-proxy-header.
1307 It is checked only once after TCP connection setup.
1309 A deny match results in TCP connection closure.
1311 An allow match is required for Squid to permit the corresponding
1312 TCP connection, before Squid even looks for HTTP request headers.
1313 If there is an allow match, Squid starts using PROXY header information
1314 to determine the source address of the connection for all future ACL
1315 checks, logging, etc.
1317 SECURITY CONSIDERATIONS:
1319 Any host from which we accept client IP details can place
1320 incorrect information in the relevant header, and Squid
1321 will use the incorrect information as if it were the
1322 source address of the request. This may enable remote
1323 hosts to bypass any access control restrictions that are
1324 based on the client's source addresses.
1326 This clause only supports fast acl types.
1327 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1330 NAME: follow_x_forwarded_for
1332 IFDEF: FOLLOW_X_FORWARDED_FOR
1333 LOC: Config.accessList.followXFF
1334 DEFAULT_IF_NONE: deny all
1335 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1337 Determine which client proxies can be trusted to provide correct
1338 information regarding real client IP address.
1340 Requests may pass through a chain of several other proxies
1341 before reaching us. The original source details may by sent in:
1342 * HTTP message Forwarded header, or
1343 * HTTP message X-Forwarded-For header, or
1344 * PROXY protocol connection header.
1346 PROXY protocol connections are controlled by the proxy_protocol_access
1347 directive which is checked before this.
1349 If a request reaches us from a source that is allowed by this
1350 directive, then we trust the information it provides regarding
1351 the IP of the client it received from (if any).
1353 For the purpose of ACLs used in this directive the src ACL type always
1354 matches the address we are testing and srcdomain matches its rDNS.
1356 On each HTTP request Squid checks for X-Forwarded-For header fields.
1357 If found the header values are iterated in reverse order and an allow
1358 match is required for Squid to continue on to the next value.
1359 The verification ends when a value receives a deny match, cannot be
1360 tested, or there are no more values to test.
1361 NOTE: Squid does not yet follow the Forwarded HTTP header.
1363 The end result of this process is an IP address that we will
1364 refer to as the indirect client address. This address may
1365 be treated as the client address for access control, ICAP, delay
1366 pools and logging, depending on the acl_uses_indirect_client,
1367 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1368 log_uses_indirect_client and tproxy_uses_indirect_client options.
1370 This clause only supports fast acl types.
1371 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1373 SECURITY CONSIDERATIONS:
1375 Any host from which we accept client IP details can place
1376 incorrect information in the relevant header, and Squid
1377 will use the incorrect information as if it were the
1378 source address of the request. This may enable remote
1379 hosts to bypass any access control restrictions that are
1380 based on the client's source addresses.
1384 acl localhost src 127.0.0.1
1385 acl my_other_proxy srcdomain .proxy.example.com
1386 follow_x_forwarded_for allow localhost
1387 follow_x_forwarded_for allow my_other_proxy
1390 NAME: acl_uses_indirect_client
1393 IFDEF: FOLLOW_X_FORWARDED_FOR
1395 LOC: Config.onoff.acl_uses_indirect_client
1397 Controls whether the indirect client address
1398 (see follow_x_forwarded_for) is used instead of the
1399 direct client address in acl matching.
1401 NOTE: maxconn ACL considers direct TCP links and indirect
1402 clients will always have zero. So no match.
1405 NAME: delay_pool_uses_indirect_client
1408 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1410 LOC: Config.onoff.delay_pool_uses_indirect_client
1412 Controls whether the indirect client address
1413 (see follow_x_forwarded_for) is used instead of the
1414 direct client address in delay pools.
1417 NAME: log_uses_indirect_client
1420 IFDEF: FOLLOW_X_FORWARDED_FOR
1422 LOC: Config.onoff.log_uses_indirect_client
1424 Controls whether the indirect client address
1425 (see follow_x_forwarded_for) is used instead of the
1426 direct client address in the access log.
1429 NAME: tproxy_uses_indirect_client
1432 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1434 LOC: Config.onoff.tproxy_uses_indirect_client
1436 Controls whether the indirect client address
1437 (see follow_x_forwarded_for) is used instead of the
1438 direct client address when spoofing the outgoing client.
1440 This has no effect on requests arriving in non-tproxy
1443 SECURITY WARNING: Usage of this option is dangerous
1444 and should not be used trivially. Correct configuration
1445 of follow_x_forewarded_for with a limited set of trusted
1446 sources is required to prevent abuse of your proxy.
1449 NAME: spoof_client_ip
1451 LOC: Config.accessList.spoof_client_ip
1453 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1455 Control client IP address spoofing of TPROXY traffic based on
1456 defined access lists.
1458 spoof_client_ip allow|deny [!]aclname ...
1460 If there are no "spoof_client_ip" lines present, the default
1461 is to "allow" spoofing of any suitable request.
1463 Note that the cache_peer "no-tproxy" option overrides this ACL.
1465 This clause supports fast acl types.
1466 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1471 LOC: Config.accessList.http
1472 DEFAULT_IF_NONE: deny all
1473 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1475 Allowing or Denying access based on defined access lists
1477 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1478 http_access allow|deny [!]aclname ...
1480 NOTE on default values:
1482 If there are no "access" lines present, the default is to deny
1485 If none of the "access" lines cause a match, the default is the
1486 opposite of the last line in the list. If the last line was
1487 deny, the default is allow. Conversely, if the last line
1488 is allow, the default will be deny. For these reasons, it is a
1489 good idea to have an "deny all" entry at the end of your access
1490 lists to avoid potential confusion.
1492 This clause supports both fast and slow acl types.
1493 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1498 # Recommended minimum Access Permission configuration:
1500 # Deny requests to certain unsafe ports
1501 http_access deny !Safe_ports
1503 # Deny CONNECT to other than secure SSL ports
1504 http_access deny CONNECT !SSL_ports
1506 # Only allow cachemgr access from localhost
1507 http_access allow localhost manager
1508 http_access deny manager
1510 # We strongly recommend the following be uncommented to protect innocent
1511 # web applications running on the proxy server who think the only
1512 # one who can access services on "localhost" is a local user
1513 #http_access deny to_localhost
1516 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1519 # Example rule allowing access from your local networks.
1520 # Adapt localnet in the ACL section to list your (internal) IP networks
1521 # from where browsing should be allowed
1522 http_access allow localnet
1523 http_access allow localhost
1525 # And finally deny all other access to this proxy
1526 http_access deny all
1530 NAME: adapted_http_access http_access2
1532 LOC: Config.accessList.adapted_http
1534 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1536 Allowing or Denying access based on defined access lists
1538 Essentially identical to http_access, but runs after redirectors
1539 and ICAP/eCAP adaptation. Allowing access control based on their
1542 If not set then only http_access is used.
1545 NAME: http_reply_access
1547 LOC: Config.accessList.reply
1549 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1551 Allow replies to client requests. This is complementary to http_access.
1553 http_reply_access allow|deny [!] aclname ...
1555 NOTE: if there are no access lines present, the default is to allow
1558 If none of the access lines cause a match the opposite of the
1559 last line will apply. Thus it is good practice to end the rules
1560 with an "allow all" or "deny all" entry.
1562 This clause supports both fast and slow acl types.
1563 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1568 LOC: Config.accessList.icp
1570 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1572 Allowing or Denying access to the ICP port based on defined
1575 icp_access allow|deny [!]aclname ...
1577 NOTE: The default if no icp_access lines are present is to
1578 deny all traffic. This default may cause problems with peers
1581 This clause only supports fast acl types.
1582 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1584 # Allow ICP queries from local networks only
1585 #icp_access allow localnet
1586 #icp_access deny all
1592 LOC: Config.accessList.htcp
1594 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1596 Allowing or Denying access to the HTCP port based on defined
1599 htcp_access allow|deny [!]aclname ...
1601 See also htcp_clr_access for details on access control for
1602 cache purge (CLR) HTCP messages.
1604 NOTE: The default if no htcp_access lines are present is to
1605 deny all traffic. This default may cause problems with peers
1606 using the htcp option.
1608 This clause only supports fast acl types.
1609 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1611 # Allow HTCP queries from local networks only
1612 #htcp_access allow localnet
1613 #htcp_access deny all
1616 NAME: htcp_clr_access
1619 LOC: Config.accessList.htcp_clr
1621 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1623 Allowing or Denying access to purge content using HTCP based
1624 on defined access lists.
1625 See htcp_access for details on general HTCP access control.
1627 htcp_clr_access allow|deny [!]aclname ...
1629 This clause only supports fast acl types.
1630 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1632 # Allow HTCP CLR requests from trusted peers
1633 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1634 htcp_clr_access allow htcp_clr_peer
1635 htcp_clr_access deny all
1640 LOC: Config.accessList.miss
1642 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1644 Determins whether network access is permitted when satisfying a request.
1647 to force your neighbors to use you as a sibling instead of
1650 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1651 miss_access deny !localclients
1652 miss_access allow all
1654 This means only your local clients are allowed to fetch relayed/MISS
1655 replies from the network and all other clients can only fetch cached
1658 The default for this setting allows all clients who passed the
1659 http_access rules to relay via this proxy.
1661 This clause only supports fast acl types.
1662 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1665 NAME: ident_lookup_access
1669 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1670 LOC: Ident::TheConfig.identLookup
1672 A list of ACL elements which, if matched, cause an ident
1673 (RFC 931) lookup to be performed for this request. For
1674 example, you might choose to always perform ident lookups
1675 for your main multi-user Unix boxes, but not for your Macs
1676 and PCs. By default, ident lookups are not performed for
1679 To enable ident lookups for specific client addresses, you
1680 can follow this example:
1682 acl ident_aware_hosts src 198.168.1.0/24
1683 ident_lookup_access allow ident_aware_hosts
1684 ident_lookup_access deny all
1686 Only src type ACL checks are fully supported. A srcdomain
1687 ACL might work at times, but it will not always provide
1690 This clause only supports fast acl types.
1691 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1694 NAME: reply_body_max_size
1695 COMMENT: size [acl acl...]
1698 DEFAULT_DOC: No limit is applied.
1699 LOC: Config.ReplyBodySize
1701 This option specifies the maximum size of a reply body. It can be
1702 used to prevent users from downloading very large files, such as
1703 MP3's and movies. When the reply headers are received, the
1704 reply_body_max_size lines are processed, and the first line where
1705 all (if any) listed ACLs are true is used as the maximum body size
1708 This size is checked twice. First when we get the reply headers,
1709 we check the content-length value. If the content length value exists
1710 and is larger than the allowed size, the request is denied and the
1711 user receives an error message that says "the request or reply
1712 is too large." If there is no content-length, and the reply
1713 size exceeds this limit, the client's connection is just closed
1714 and they will receive a partial reply.
1716 WARNING: downstream caches probably can not detect a partial reply
1717 if there is no content-length header, so they will cache
1718 partial responses and give them out as hits. You should NOT
1719 use this option if you have downstream caches.
1721 WARNING: A maximum size smaller than the size of squid's error messages
1722 will cause an infinite loop and crash squid. Ensure that the smallest
1723 non-zero value you use is greater that the maximum header size plus
1724 the size of your largest error page.
1726 If you set this parameter none (the default), there will be
1729 Configuration Format is:
1730 reply_body_max_size SIZE UNITS [acl ...]
1732 reply_body_max_size 10 MB
1736 NAME: on_unsupported_protocol
1737 TYPE: on_unsupported_protocol
1738 LOC: Config.accessList.on_unsupported_protocol
1740 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
1742 Determines Squid behavior when encountering strange requests at the
1743 beginning of an accepted TCP connection or the beginning of a bumped
1744 CONNECT tunnel. Controlling Squid reaction to unexpected traffic is
1745 especially useful in interception environments where Squid is likely
1746 to see connections for unsupported protocols that Squid should either
1747 terminate or tunnel at TCP level.
1749 on_unsupported_protocol <action> [!]acl ...
1751 The first matching action wins. Only fast ACLs are supported.
1753 Supported actions are:
1755 tunnel: Establish a TCP connection with the intended server and
1756 blindly shovel TCP packets between the client and server.
1758 respond: Respond with an error message, using the transfer protocol
1759 for the Squid port that received the request (e.g., HTTP
1760 for connections intercepted at the http_port). This is the
1763 Squid expects the following traffic patterns:
1765 http_port: a plain HTTP request
1766 https_port: SSL/TLS handshake followed by an [encrypted] HTTP request
1767 ftp_port: a plain FTP command (no on_unsupported_protocol support yet!)
1768 CONNECT tunnel on http_port: same as https_port
1769 CONNECT tunnel on https_port: same as https_port
1771 Currently, this directive has effect on intercepted connections and
1772 bumped tunnels only. Other cases are not supported because Squid
1773 cannot know the intended destination of other traffic.
1776 # define what Squid errors indicate receiving non-HTTP traffic:
1777 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
1778 # define what Squid errors indicate receiving nothing:
1779 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
1780 # tunnel everything that does not look like HTTP:
1781 on_unsupported_protocol tunnel foreignProtocol
1782 # tunnel if we think the client waits for the server to talk first:
1783 on_unsupported_protocol tunnel serverTalksFirstProtocol
1784 # in all other error cases, just send an HTTP "error page" response:
1785 on_unsupported_protocol respond all
1787 See also: squid_error ACL
1792 -----------------------------------------------------------------------------
1795 NAME: http_port ascii_port
1800 Usage: port [mode] [options]
1801 hostname:port [mode] [options]
1802 1.2.3.4:port [mode] [options]
1804 The socket addresses where Squid will listen for HTTP client
1805 requests. You may specify multiple socket addresses.
1806 There are three forms: port alone, hostname with port, and
1807 IP address with port. If you specify a hostname or IP
1808 address, Squid binds the socket to that specific
1809 address. Most likely, you do not need to bind to a specific
1810 address, so you can use the port number alone.
1812 If you are running Squid in accelerator mode, you
1813 probably want to listen on port 80 also, or instead.
1815 The -a command line option may be used to specify additional
1816 port(s) where Squid listens for proxy request. Such ports will
1817 be plain proxy ports with no options.
1819 You may specify multiple socket addresses on multiple lines.
1823 intercept Support for IP-Layer interception of
1824 outgoing requests without browser settings.
1825 NP: disables authentication and IPv6 on the port.
1827 tproxy Support Linux TPROXY for spoofing outgoing
1828 connections using the client IP address.
1829 NP: disables authentication and maybe IPv6 on the port.
1831 accel Accelerator / reverse proxy mode
1833 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1834 establish secure connection with the client and with
1835 the server, decrypt HTTPS messages as they pass through
1836 Squid, and treat them as unencrypted HTTP messages,
1837 becoming the man-in-the-middle.
1839 The ssl_bump option is required to fully enable
1840 bumping of CONNECT requests.
1842 Omitting the mode flag causes default forward proxy mode to be used.
1845 Accelerator Mode Options:
1847 defaultsite=domainname
1848 What to use for the Host: header if it is not present
1849 in a request. Determines what site (not origin server)
1850 accelerators should consider the default.
1852 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1854 protocol= Protocol to reconstruct accelerated and intercepted
1855 requests with. Defaults to HTTP/1.1 for http_port and
1856 HTTPS/1.1 for https_port.
1857 When an unsupported value is configured Squid will
1858 produce a FATAL error.
1859 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1861 vport Virtual host port support. Using the http_port number
1862 instead of the port passed on Host: headers.
1864 vport=NN Virtual host port support. Using the specified port
1865 number instead of the port passed on Host: headers.
1868 Act as if this Squid is the origin server.
1869 This currently means generate new Date: and Expires:
1870 headers on HIT instead of adding Age:.
1872 ignore-cc Ignore request Cache-Control headers.
1874 WARNING: This option violates HTTP specifications if
1875 used in non-accelerator setups.
1877 allow-direct Allow direct forwarding in accelerator mode. Normally
1878 accelerated requests are denied direct forwarding as if
1879 never_direct was used.
1881 WARNING: this option opens accelerator mode to security
1882 vulnerabilities usually only affecting in interception
1883 mode. Make sure to protect forwarding with suitable
1884 http_access rules when using this.
1887 SSL Bump Mode Options:
1888 In addition to these options ssl-bump requires TLS/SSL options.
1890 generate-host-certificates[=<on|off>]
1891 Dynamically create SSL server certificates for the
1892 destination hosts of bumped CONNECT requests.When
1893 enabled, the cert and key options are used to sign
1894 generated certificates. Otherwise generated
1895 certificate will be selfsigned.
1896 If there is a CA certificate lifetime of the generated
1897 certificate equals lifetime of the CA certificate. If
1898 generated certificate is selfsigned lifetime is three
1900 This option is enabled by default when ssl-bump is used.
1901 See the ssl-bump option above for more information.
1903 dynamic_cert_mem_cache_size=SIZE
1904 Approximate total RAM size spent on cached generated
1905 certificates. If set to zero, caching is disabled. The
1906 default value is 4MB.
1910 cert= Path to SSL certificate (PEM format).
1912 key= Path to SSL private key file (PEM format)
1913 if not specified, the certificate file is
1914 assumed to be a combined certificate and
1917 cipher= Colon separated list of supported ciphers.
1918 NOTE: some ciphers such as EDH ciphers depend on
1919 additional settings. If those settings are
1920 omitted the ciphers may be silently ignored
1921 by the OpenSSL library.
1923 options= Various SSL implementation options. The most important
1926 NO_SSLv3 Disallow the use of SSLv3
1928 NO_TLSv1 Disallow the use of TLSv1.0
1930 NO_TLSv1_1 Disallow the use of TLSv1.1
1932 NO_TLSv1_2 Disallow the use of TLSv1.2
1935 Always create a new key when using
1936 temporary/ephemeral DH key exchanges
1939 Disable use of RFC5077 session tickets.
1940 Some servers may have problems
1941 understanding the TLS extension due
1942 to ambiguous specification in RFC4507.
1944 ALL Enable various bug workarounds
1945 suggested as "harmless" by OpenSSL
1946 Be warned that this reduces SSL/TLS
1947 strength to some attacks.
1949 See the OpenSSL SSL_CTX_set_options documentation for a
1952 clientca= File containing the list of CAs to use when
1953 requesting a client certificate.
1955 cafile= File containing additional CA certificates to
1956 use when verifying client certificates. If unset
1957 clientca will be used.
1959 capath= Directory containing additional CA certificates
1960 and CRL lists to use when verifying client certificates.
1962 crlfile= File of additional CRL lists to use when verifying
1963 the client certificate, in addition to CRLs stored in
1964 the capath. Implies VERIFY_CRL flag below.
1966 dhparams= File containing DH parameters for temporary/ephemeral
1967 DH key exchanges. See OpenSSL documentation for details
1968 on how to create this file.
1969 WARNING: EDH ciphers will be silently disabled if this
1972 sslflags= Various flags modifying the use of SSL:
1974 Don't request client certificates
1975 immediately, but wait until acl processing
1976 requires a certificate (not yet implemented).
1978 Don't use the default CA lists built in
1981 Don't allow for session reuse. Each connection
1982 will result in a new SSL session.
1984 Verify CRL lists when accepting client
1987 Verify CRL lists for all certificates in the
1988 client certificate chain.
1990 sslcontext= SSL session ID context identifier.
1994 connection-auth[=on|off]
1995 use connection-auth=off to tell Squid to prevent
1996 forwarding Microsoft connection oriented authentication
1997 (NTLM, Negotiate and Kerberos)
1999 disable-pmtu-discovery=
2000 Control Path-MTU discovery usage:
2001 off lets OS decide on what to do (default).
2002 transparent disable PMTU discovery when transparent
2004 always disable always PMTU discovery.
2006 In many setups of transparently intercepting proxies
2007 Path-MTU discovery can not work on traffic towards the
2008 clients. This is the case when the intercepting device
2009 does not fully track connections and fails to forward
2010 ICMP must fragment messages to the cache server. If you
2011 have such setup and experience that certain clients
2012 sporadically hang or never complete requests set
2013 disable-pmtu-discovery option to 'transparent'.
2015 name= Specifies a internal name for the port. Defaults to
2016 the port specification (port or addr:port)
2018 tcpkeepalive[=idle,interval,timeout]
2019 Enable TCP keepalive probes of idle connections.
2020 In seconds; idle is the initial time before TCP starts
2021 probing the connection, interval how often to probe, and
2022 timeout the time before giving up.
2024 require-proxy-header
2025 Require PROXY protocol version 1 or 2 connections.
2026 The proxy_protocol_access is required to whitelist
2027 downstream proxies which can be trusted.
2029 If you run Squid on a dual-homed machine with an internal
2030 and an external interface we recommend you to specify the
2031 internal address:port in http_port. This way Squid will only be
2032 visible on the internal address.
2036 # Squid normally listens to port 3128
2037 http_port @DEFAULT_HTTP_PORT@
2047 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
2049 The socket address where Squid will listen for client requests made
2050 over TLS or SSL connections. Commonly referred to as HTTPS.
2052 This is most useful for situations where you are running squid in
2053 accelerator mode and you want to do the SSL work at the accelerator level.
2055 You may specify multiple socket addresses on multiple lines,
2056 each with their own SSL certificate and/or options.
2060 accel Accelerator / reverse proxy mode
2062 intercept Support for IP-Layer interception of
2063 outgoing requests without browser settings.
2064 NP: disables authentication and IPv6 on the port.
2066 tproxy Support Linux TPROXY for spoofing outgoing
2067 connections using the client IP address.
2068 NP: disables authentication and maybe IPv6 on the port.
2070 ssl-bump For each intercepted connection allowed by ssl_bump
2071 ACLs, establish a secure connection with the client and with
2072 the server, decrypt HTTPS messages as they pass through
2073 Squid, and treat them as unencrypted HTTP messages,
2074 becoming the man-in-the-middle.
2076 An "ssl_bump server-first" match is required to
2077 fully enable bumping of intercepted SSL connections.
2079 Requires tproxy or intercept.
2081 Omitting the mode flag causes default forward proxy mode to be used.
2084 See http_port for a list of generic options
2089 cert= Path to SSL certificate (PEM format).
2091 key= Path to SSL private key file (PEM format)
2092 if not specified, the certificate file is
2093 assumed to be a combined certificate and
2096 cipher= Colon separated list of supported ciphers.
2098 options= Various SSL engine options. The most important
2101 NO_SSLv3 Disallow the use of SSLv3
2103 NO_TLSv1 Disallow the use of TLSv1.0
2105 NO_TLSv1_1 Disallow the use of TLSv1.1
2107 NO_TLSv1_2 Disallow the use of TLSv1.2
2110 Always create a new key when using
2111 temporary/ephemeral DH key exchanges
2114 Disable use of RFC5077 session tickets.
2115 Some servers may have problems
2116 understanding the TLS extension due
2117 to ambiguous specification in RFC4507.
2119 ALL Enable various bug workarounds
2120 suggested as "harmless" by OpenSSL
2121 Be warned that this reduces SSL/TLS
2122 strength to some attacks.
2124 See the OpenSSL SSL_CTX_set_options documentation for a
2127 clientca= File containing the list of CAs to use when
2128 requesting a client certificate.
2130 cafile= File containing additional CA certificates to
2131 use when verifying client certificates. If unset
2132 clientca will be used.
2134 capath= Directory containing additional CA certificates
2135 and CRL lists to use when verifying client certificates.
2137 crlfile= File of additional CRL lists to use when verifying
2138 the client certificate, in addition to CRLs stored in
2139 the capath. Implies VERIFY_CRL flag below.
2141 dhparams= File containing DH parameters for temporary/ephemeral
2144 sslflags= Various flags modifying the use of SSL:
2146 Don't request client certificates
2147 immediately, but wait until acl processing
2148 requires a certificate (not yet implemented).
2150 Don't use the default CA lists built in
2153 Don't allow for session reuse. Each connection
2154 will result in a new SSL session.
2156 Verify CRL lists when accepting client
2159 Verify CRL lists for all certificates in the
2160 client certificate chain.
2162 sslcontext= SSL session ID context identifier.
2164 generate-host-certificates[=<on|off>]
2165 Dynamically create SSL server certificates for the
2166 destination hosts of bumped SSL requests.When
2167 enabled, the cert and key options are used to sign
2168 generated certificates. Otherwise generated
2169 certificate will be selfsigned.
2170 If there is CA certificate life time of generated
2171 certificate equals lifetime of CA certificate. If
2172 generated certificate is selfsigned lifetime is three
2174 This option is enabled by default when SslBump is used.
2175 See the sslBump option above for more information.
2177 dynamic_cert_mem_cache_size=SIZE
2178 Approximate total RAM size spent on cached generated
2179 certificates. If set to zero, caching is disabled. The
2180 default value is 4MB.
2182 See http_port for a list of available options.
2190 Enables Native FTP proxy by specifying the socket address where Squid
2191 listens for FTP client requests. See http_port directive for various
2192 ways to specify the listening address and mode.
2194 Usage: ftp_port address [mode] [options]
2196 WARNING: This is a new, experimental, complex feature that has seen
2197 limited production exposure. Some Squid modules (e.g., caching) do not
2198 currently work with native FTP proxying, and many features have not
2199 even been tested for compatibility. Test well before deploying!
2201 Native FTP proxying differs substantially from proxying HTTP requests
2202 with ftp:// URIs because Squid works as an FTP server and receives
2203 actual FTP commands (rather than HTTP requests with FTP URLs).
2205 Native FTP commands accepted at ftp_port are internally converted or
2206 wrapped into HTTP-like messages. The same happens to Native FTP
2207 responses received from FTP origin servers. Those HTTP-like messages
2208 are shoveled through regular access control and adaptation layers
2209 between the FTP client and the FTP origin server. This allows Squid to
2210 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2211 mechanisms when shoveling wrapped FTP messages. For example,
2212 http_access and adaptation_access directives are used.
2216 intercept Same as http_port intercept. The FTP origin address is
2217 determined based on the intended destination of the
2218 intercepted connection.
2220 tproxy Support Linux TPROXY for spoofing outgoing
2221 connections using the client IP address.
2222 NP: disables authentication and maybe IPv6 on the port.
2224 By default (i.e., without an explicit mode option), Squid extracts the
2225 FTP origin address from the login@origin parameter of the FTP USER
2226 command. Many popular FTP clients support such native FTP proxying.
2230 name=token Specifies an internal name for the port. Defaults to
2231 the port address. Usable with myportname ACL.
2234 Enables tracking of FTP directories by injecting extra
2235 PWD commands and adjusting Request-URI (in wrapping
2236 HTTP requests) to reflect the current FTP server
2237 directory. Tracking is disabled by default.
2239 protocol=FTP Protocol to reconstruct accelerated and intercepted
2240 requests with. Defaults to FTP. No other accepted
2241 values have been tested with. An unsupported value
2242 results in a FATAL error. Accepted values are FTP,
2243 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2245 Other http_port modes and options that are not specific to HTTP and
2246 HTTPS may also work.
2249 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2252 LOC: Ip::Qos::TheConfig.tosToServer
2254 Allows you to select a TOS/Diffserv value for packets outgoing
2255 on the server side, based on an ACL.
2257 tcp_outgoing_tos ds-field [!]aclname ...
2259 Example where normal_service_net uses the TOS value 0x00
2260 and good_service_net uses 0x20
2262 acl normal_service_net src 10.0.0.0/24
2263 acl good_service_net src 10.0.1.0/24
2264 tcp_outgoing_tos 0x00 normal_service_net
2265 tcp_outgoing_tos 0x20 good_service_net
2267 TOS/DSCP values really only have local significance - so you should
2268 know what you're specifying. For more information, see RFC2474,
2269 RFC2475, and RFC3260.
2271 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2272 "default" to use whatever default your host has.
2273 Note that only multiples of 4 are usable as the two rightmost bits have
2274 been redefined for use by ECN (RFC 3168 section 23.1).
2275 The squid parser will enforce this by masking away the ECN bits.
2277 Processing proceeds in the order specified, and stops at first fully
2280 Only fast ACLs are supported.
2283 NAME: clientside_tos
2286 LOC: Ip::Qos::TheConfig.tosToClient
2288 Allows you to select a TOS/DSCP value for packets being transmitted
2289 on the client-side, based on an ACL.
2291 clientside_tos ds-field [!]aclname ...
2293 Example where normal_service_net uses the TOS value 0x00
2294 and good_service_net uses 0x20
2296 acl normal_service_net src 10.0.0.0/24
2297 acl good_service_net src 10.0.1.0/24
2298 clientside_tos 0x00 normal_service_net
2299 clientside_tos 0x20 good_service_net
2301 Note: This feature is incompatible with qos_flows. Any TOS values set here
2302 will be overwritten by TOS values in qos_flows.
2304 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2305 "default" to use whatever default your host has.
2306 Note that only multiples of 4 are usable as the two rightmost bits have
2307 been redefined for use by ECN (RFC 3168 section 23.1).
2308 The squid parser will enforce this by masking away the ECN bits.
2312 NAME: tcp_outgoing_mark
2314 IFDEF: SO_MARK&&USE_LIBCAP
2316 LOC: Ip::Qos::TheConfig.nfmarkToServer
2318 Allows you to apply a Netfilter mark value to outgoing packets
2319 on the server side, based on an ACL.
2321 tcp_outgoing_mark mark-value [!]aclname ...
2323 Example where normal_service_net uses the mark value 0x00
2324 and good_service_net uses 0x20
2326 acl normal_service_net src 10.0.0.0/24
2327 acl good_service_net src 10.0.1.0/24
2328 tcp_outgoing_mark 0x00 normal_service_net
2329 tcp_outgoing_mark 0x20 good_service_net
2331 Only fast ACLs are supported.
2334 NAME: clientside_mark
2336 IFDEF: SO_MARK&&USE_LIBCAP
2338 LOC: Ip::Qos::TheConfig.nfmarkToClient
2340 Allows you to apply a Netfilter mark value to packets being transmitted
2341 on the client-side, based on an ACL.
2343 clientside_mark mark-value [!]aclname ...
2345 Example where normal_service_net uses the mark value 0x00
2346 and good_service_net uses 0x20
2348 acl normal_service_net src 10.0.0.0/24
2349 acl good_service_net src 10.0.1.0/24
2350 clientside_mark 0x00 normal_service_net
2351 clientside_mark 0x20 good_service_net
2353 Note: This feature is incompatible with qos_flows. Any mark values set here
2354 will be overwritten by mark values in qos_flows.
2361 LOC: Ip::Qos::TheConfig
2363 Allows you to select a TOS/DSCP value to mark outgoing
2364 connections to the client, based on where the reply was sourced.
2365 For platforms using netfilter, allows you to set a netfilter mark
2366 value instead of, or in addition to, a TOS value.
2368 By default this functionality is disabled. To enable it with the default
2369 settings simply use "qos_flows mark" or "qos_flows tos". Default
2370 settings will result in the netfilter mark or TOS value being copied
2371 from the upstream connection to the client. Note that it is the connection
2372 CONNMARK value not the packet MARK value that is copied.
2374 It is not currently possible to copy the mark or TOS value from the
2375 client to the upstream connection request.
2377 TOS values really only have local significance - so you should
2378 know what you're specifying. For more information, see RFC2474,
2379 RFC2475, and RFC3260.
2381 The TOS/DSCP byte must be exactly that - a octet value 0 - 255.
2382 Note that only multiples of 4 are usable as the two rightmost bits have
2383 been redefined for use by ECN (RFC 3168 section 23.1).
2384 The squid parser will enforce this by masking away the ECN bits.
2386 Mark values can be any unsigned 32-bit integer value.
2388 This setting is configured by setting the following values:
2390 tos|mark Whether to set TOS or netfilter mark values
2392 local-hit=0xFF Value to mark local cache hits.
2394 sibling-hit=0xFF Value to mark hits from sibling peers.
2396 parent-hit=0xFF Value to mark hits from parent peers.
2398 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2399 over the preserve-miss feature (see below), unless
2400 mask is specified, in which case only the bits
2401 specified in the mask are written.
2403 The TOS variant of the following features are only possible on Linux
2404 and require your kernel to be patched with the TOS preserving ZPH
2405 patch, available from http://zph.bratcheda.org
2406 No patch is needed to preserve the netfilter mark, which will work
2407 with all variants of netfilter.
2409 disable-preserve-miss
2410 This option disables the preservation of the TOS or netfilter
2411 mark. By default, the existing TOS or netfilter mark value of
2412 the response coming from the remote server will be retained
2413 and masked with miss-mark.
2414 NOTE: in the case of a netfilter mark, the mark must be set on
2415 the connection (using the CONNMARK target) not on the packet
2419 Allows you to mask certain bits in the TOS or mark value
2420 received from the remote server, before copying the value to
2421 the TOS sent towards clients.
2422 Default for tos: 0xFF (TOS from server is not changed).
2423 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2425 All of these features require the --enable-zph-qos compilation flag
2426 (enabled by default). Netfilter marking also requires the
2427 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2428 libcap 2.09+ (--with-libcap).
2432 NAME: tcp_outgoing_address
2435 DEFAULT_DOC: Address selection is performed by the operating system.
2436 LOC: Config.accessList.outgoing_address
2438 Allows you to map requests to different outgoing IP addresses
2439 based on the username or source address of the user making
2442 tcp_outgoing_address ipaddr [[!]aclname] ...
2445 Forwarding clients with dedicated IPs for certain subnets.
2447 acl normal_service_net src 10.0.0.0/24
2448 acl good_service_net src 10.0.2.0/24
2450 tcp_outgoing_address 2001:db8::c001 good_service_net
2451 tcp_outgoing_address 10.1.0.2 good_service_net
2453 tcp_outgoing_address 2001:db8::beef normal_service_net
2454 tcp_outgoing_address 10.1.0.1 normal_service_net
2456 tcp_outgoing_address 2001:db8::1
2457 tcp_outgoing_address 10.1.0.3
2459 Processing proceeds in the order specified, and stops at first fully
2462 Squid will add an implicit IP version test to each line.
2463 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2464 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2467 NOTE: The use of this directive using client dependent ACLs is
2468 incompatible with the use of server side persistent connections. To
2469 ensure correct results it is best to set server_persistent_connections
2470 to off when using this directive in such configurations.
2472 NOTE: The use of this directive to set a local IP on outgoing TCP links
2473 is incompatible with using TPROXY to set client IP out outbound TCP links.
2474 When needing to contact peers use the no-tproxy cache_peer option and the
2475 client_dst_passthru directive re-enable normal forwarding such as this.
2479 NAME: host_verify_strict
2482 LOC: Config.onoff.hostStrictVerify
2484 Regardless of this option setting, when dealing with intercepted
2485 traffic, Squid always verifies that the destination IP address matches
2486 the Host header domain or IP (called 'authority form URL').
2488 This enforcement is performed to satisfy a MUST-level requirement in
2489 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2490 authority of the origin server or gateway given by the original URL".
2493 Squid always responds with an HTTP 409 (Conflict) error
2494 page and logs a security warning if there is no match.
2496 Squid verifies that the destination IP address matches
2497 the Host header for forward-proxy and reverse-proxy traffic
2498 as well. For those traffic types, Squid also enables the
2499 following checks, comparing the corresponding Host header
2500 and Request-URI components:
2502 * The host names (domain or IP) must be identical,
2503 but valueless or missing Host header disables all checks.
2504 For the two host names to match, both must be either IP
2507 * Port numbers must be identical, but if a port is missing
2508 the scheme-default port is assumed.
2511 When set to OFF (the default):
2512 Squid allows suspicious requests to continue but logs a
2513 security warning and blocks caching of the response.
2515 * Forward-proxy traffic is not checked at all.
2517 * Reverse-proxy traffic is not checked at all.
2519 * Intercepted traffic which passes verification is handled
2520 according to client_dst_passthru.
2522 * Intercepted requests which fail verification are sent
2523 to the client original destination instead of DIRECT.
2524 This overrides 'client_dst_passthru off'.
2526 For now suspicious intercepted CONNECT requests are always
2527 responded to with an HTTP 409 (Conflict) error page.
2532 As described in CVE-2009-0801 when the Host: header alone is used
2533 to determine the destination of a request it becomes trivial for
2534 malicious scripts on remote websites to bypass browser same-origin
2535 security policy and sandboxing protections.
2537 The cause of this is that such applets are allowed to perform their
2538 own HTTP stack, in which case the same-origin policy of the browser
2539 sandbox only verifies that the applet tries to contact the same IP
2540 as from where it was loaded at the IP level. The Host: header may
2541 be different from the connected IP and approved origin.
2545 NAME: client_dst_passthru
2548 LOC: Config.onoff.client_dst_passthru
2550 With NAT or TPROXY intercepted traffic Squid may pass the request
2551 directly to the original client destination IP or seek a faster
2552 source using the HTTP Host header.
2554 Using Host to locate alternative servers can provide faster
2555 connectivity with a range of failure recovery options.
2556 But can also lead to connectivity trouble when the client and
2557 server are attempting stateful interactions unaware of the proxy.
2559 This option (on by default) prevents alternative DNS entries being
2560 located to send intercepted traffic DIRECT to an origin server.
2561 The clients original destination IP and port will be used instead.
2563 Regardless of this option setting, when dealing with intercepted
2564 traffic Squid will verify the Host: header and any traffic which
2565 fails Host verification will be treated as if this option were ON.
2567 see host_verify_strict for details on the verification process.
2572 -----------------------------------------------------------------------------
2575 NAME: tls_outgoing_options
2576 IFDEF: USE_GNUTLS||USE_OPENSSL
2577 TYPE: securePeerOptions
2579 LOC: Security::ProxyOutgoingConfig
2581 disable Do not support https:// URLs.
2583 cert=/path/to/client/certificate
2584 A client TLS certificate to use when connecting.
2586 key=/path/to/client/private_key
2587 The private TLS key corresponding to the cert= above.
2588 If key= is not specified cert= is assumed to reference
2589 a PEM file containing both the certificate and the key.
2591 cipher=... The list of valid TLS ciphers to use.
2594 The minimum TLS protocol version to permit. To control
2595 SSLv3 use the options= parameter.
2596 Supported Values: 1.0 (default), 1.1, 1.2
2598 options=... Specify various TLS/SSL implementation options:
2600 NO_SSLv3 Disallow the use of SSLv3
2602 NO_TLSv1 Disallow the use of TLSv1.0
2604 NO_TLSv1_1 Disallow the use of TLSv1.1
2606 NO_TLSv1_2 Disallow the use of TLSv1.2
2609 Always create a new key when using
2610 temporary/ephemeral DH key exchanges
2613 Disable use of RFC5077 session tickets.
2614 Some servers may have problems
2615 understanding the TLS extension due
2616 to ambiguous specification in RFC4507.
2618 ALL Enable various bug workarounds
2619 suggested as "harmless" by OpenSSL
2620 Be warned that this reduces SSL/TLS
2621 strength to some attacks.
2623 See the OpenSSL SSL_CTX_set_options documentation for a
2626 cafile=... A file containing additional CA certificates to use
2627 when verifying the peer certificate.
2629 capath=... A directory containing additional CA certificates to
2630 use when verifying the peer certificate.
2632 crlfile=... A certificate revocation list file to use when
2633 verifying the peer certificate.
2635 flags=... Specify various flags modifying the TLS implementation:
2638 Accept certificates even if they fail to
2641 Don't use the default CA list built in
2644 Don't verify the peer certificate
2645 matches the server name
2647 domain= The peer name as advertised in its certificate.
2648 Used for verifying the correctness of the received peer
2649 certificate. If not specified the peer hostname will be
2655 -----------------------------------------------------------------------------
2658 NAME: ssl_unclean_shutdown
2662 LOC: Config.SSL.unclean_shutdown
2664 Some browsers (especially MSIE) bugs out on SSL shutdown
2671 LOC: Config.SSL.ssl_engine
2674 The OpenSSL engine to use. You will need to set this if you
2675 would like to use hardware SSL acceleration for example.
2678 NAME: sslproxy_session_ttl
2681 LOC: Config.SSL.session_ttl
2684 Sets the timeout value for SSL sessions
2687 NAME: sslproxy_session_cache_size
2690 LOC: Config.SSL.sessionCacheSize
2693 Sets the cache size to use for ssl session
2696 NAME: sslproxy_cert_sign_hash
2699 LOC: Config.SSL.certSignHash
2702 Sets the hashing algorithm to use when signing generated certificates.
2703 Valid algorithm names depend on the OpenSSL library used. The following
2704 names are usually available: sha1, sha256, sha512, and md5. Please see
2705 your OpenSSL library manual for the available hashes. By default, Squids
2706 that support this option use sha256 hashes.
2708 Squid does not forcefully purge cached certificates that were generated
2709 with an algorithm other than the currently configured one. They remain
2710 in the cache, subject to the regular cache eviction policy, and become
2711 useful if the algorithm changes again.
2716 TYPE: sslproxy_ssl_bump
2717 LOC: Config.accessList.ssl_bump
2718 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
2721 This option is consulted when a CONNECT request is received on
2722 an http_port (or a new connection is intercepted at an
2723 https_port), provided that port was configured with an ssl-bump
2724 flag. The subsequent data on the connection is either treated as
2725 HTTPS and decrypted OR tunneled at TCP level without decryption,
2726 depending on the first matching bumping "action".
2728 ssl_bump <action> [!]acl ...
2730 The following bumping actions are currently supported:
2733 Become a TCP tunnel without decrypting proxied traffic.
2734 This is the default action.
2737 Establish a secure connection with the server and, using a
2738 mimicked server certificate, with the client.
2741 Receive client (step SslBump1) or server (step SslBump2)
2742 certificate while preserving the possibility of splicing the
2743 connection. Peeking at the server certificate (during step 2)
2744 usually precludes bumping of the connection at step 3.
2747 Receive client (step SslBump1) or server (step SslBump2)
2748 certificate while preserving the possibility of bumping the
2749 connection. Staring at the server certificate (during step 2)
2750 usually precludes splicing of the connection at step 3.
2753 Close client and server connections.
2755 Backward compatibility actions available at step SslBump1:
2758 Bump the connection. Establish a secure connection with the
2759 client first, then connect to the server. This old mode does
2760 not allow Squid to mimic server SSL certificate and does not
2761 work with intercepted SSL connections.
2764 Bump the connection. Establish a secure connection with the
2765 server first, then establish a secure connection with the
2766 client, using a mimicked server certificate. Works with both
2767 CONNECT requests and intercepted SSL connections, but does
2768 not allow to make decisions based on SSL handshake info.
2771 Decide whether to bump or splice the connection based on
2772 client-to-squid and server-to-squid SSL hello messages.
2776 Same as the "splice" action.
2778 All ssl_bump rules are evaluated at each of the supported bumping
2779 steps. Rules with actions that are impossible at the current step are
2780 ignored. The first matching ssl_bump action wins and is applied at the
2781 end of the current step. If no rules match, the splice action is used.
2782 See the at_step ACL for a list of the supported SslBump steps.
2784 This clause supports both fast and slow acl types.
2785 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2787 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
2790 # Example: Bump all requests except those originating from
2791 # localhost or those going to example.com.
2793 acl broken_sites dstdomain .example.com
2794 ssl_bump splice localhost
2795 ssl_bump splice broken_sites
2799 NAME: sslproxy_cert_error
2802 DEFAULT_DOC: Server certificate errors terminate the transaction.
2803 LOC: Config.ssl_client.cert_error
2806 Use this ACL to bypass server certificate validation errors.
2808 For example, the following lines will bypass all validation errors
2809 when talking to servers for example.com. All other
2810 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2812 acl BrokenButTrustedServers dstdomain example.com
2813 sslproxy_cert_error allow BrokenButTrustedServers
2814 sslproxy_cert_error deny all
2816 This clause only supports fast acl types.
2817 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2818 Using slow acl types may result in server crashes
2820 Without this option, all server certificate validation errors
2821 terminate the transaction to protect Squid and the client.
2823 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2824 but should not happen unless your OpenSSL library is buggy.
2827 Bypassing validation errors is dangerous because an
2828 error usually implies that the server cannot be trusted
2829 and the connection may be insecure.
2831 See also: sslproxy_flags and DONT_VERIFY_PEER.
2834 NAME: sslproxy_cert_sign
2837 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2838 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2839 POSTSCRIPTUM: signTrusted all
2840 TYPE: sslproxy_cert_sign
2841 LOC: Config.ssl_client.cert_sign
2844 sslproxy_cert_sign <signing algorithm> acl ...
2846 The following certificate signing algorithms are supported:
2849 Sign using the configured CA certificate which is usually
2850 placed in and trusted by end-user browsers. This is the
2851 default for trusted origin server certificates.
2854 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2855 This is the default for untrusted origin server certificates
2856 that are not self-signed (see ssl::certUntrusted).
2859 Sign using a self-signed certificate with the right CN to
2860 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2861 browser. This is the default for self-signed origin server
2862 certificates (see ssl::certSelfSigned).
2864 This clause only supports fast acl types.
2866 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2867 signing algorithm to generate the certificate and ignores all
2868 subsequent sslproxy_cert_sign options (the first match wins). If no
2869 acl(s) match, the default signing algorithm is determined by errors
2870 detected when obtaining and validating the origin server certificate.
2872 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2873 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2874 CONNECT request that carries a domain name. In all other cases (CONNECT
2875 to an IP address or an intercepted SSL connection), Squid cannot detect
2876 the domain mismatch at certificate generation time when
2877 bump-server-first is used.
2880 NAME: sslproxy_cert_adapt
2883 TYPE: sslproxy_cert_adapt
2884 LOC: Config.ssl_client.cert_adapt
2887 sslproxy_cert_adapt <adaptation algorithm> acl ...
2889 The following certificate adaptation algorithms are supported:
2892 Sets the "Not After" property to the "Not After" property of
2893 the CA certificate used to sign generated certificates.
2896 Sets the "Not Before" property to the "Not Before" property of
2897 the CA certificate used to sign generated certificates.
2899 setCommonName or setCommonName{CN}
2900 Sets Subject.CN property to the host name specified as a
2901 CN parameter or, if no explicit CN parameter was specified,
2902 extracted from the CONNECT request. It is a misconfiguration
2903 to use setCommonName without an explicit parameter for
2904 intercepted or tproxied SSL connections.
2906 This clause only supports fast acl types.
2908 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2909 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2910 corresponding adaptation algorithm to generate the certificate and
2911 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2912 group (i.e., the first match wins within each algorithm group). If no
2913 acl(s) match, the default mimicking action takes place.
2915 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2916 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2917 CONNECT request that carries a domain name. In all other cases (CONNECT
2918 to an IP address or an intercepted SSL connection), Squid cannot detect
2919 the domain mismatch at certificate generation time when
2920 bump-server-first is used.
2923 NAME: sslpassword_program
2926 LOC: Config.Program.ssl_password
2929 Specify a program used for entering SSL key passphrases
2930 when using encrypted SSL certificate keys. If not specified
2931 keys must either be unencrypted, or Squid started with the -N
2932 option to allow it to query interactively for the passphrase.
2934 The key file name is given as argument to the program allowing
2935 selection of the right password if you have multiple encrypted
2940 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2941 -----------------------------------------------------------------------------
2944 NAME: sslcrtd_program
2947 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2948 LOC: Ssl::TheConfig.ssl_crtd
2950 Specify the location and options of the executable for ssl_crtd process.
2951 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2952 For more information use:
2953 @DEFAULT_SSL_CRTD@ -h
2956 NAME: sslcrtd_children
2957 TYPE: HelperChildConfig
2959 DEFAULT: 32 startup=5 idle=1
2960 LOC: Ssl::TheConfig.ssl_crtdChildren
2962 The maximum number of processes spawn to service ssl server.
2963 The maximum this may be safely set to is 32.
2965 The startup= and idle= options allow some measure of skew in your
2970 Sets the minimum number of processes to spawn when Squid
2971 starts or reconfigures. When set to zero the first request will
2972 cause spawning of the first child process to handle it.
2974 Starting too few children temporary slows Squid under load while it
2975 tries to spawn enough additional processes to cope with traffic.
2979 Sets a minimum of how many processes Squid is to try and keep available
2980 at all times. When traffic begins to rise above what the existing
2981 processes can handle this many more will be spawned up to the maximum
2982 configured. A minimum setting of 1 is required.
2986 Sets the maximum number of queued requests.
2987 If the queued requests exceed queue size for more than 3 minutes
2988 squid aborts its operation.
2989 The default value is set to 2*numberofchildren.
2991 You must have at least one ssl_crtd process.
2994 NAME: sslcrtvalidator_program
2998 LOC: Ssl::TheConfig.ssl_crt_validator
3000 Specify the location and options of the executable for ssl_crt_validator
3003 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
3006 ttl=n TTL in seconds for cached results. The default is 60 secs
3007 cache=n limit the result cache size. The default value is 2048
3010 NAME: sslcrtvalidator_children
3011 TYPE: HelperChildConfig
3013 DEFAULT: 32 startup=5 idle=1 concurrency=1
3014 LOC: Ssl::TheConfig.ssl_crt_validator_Children
3016 The maximum number of processes spawn to service SSL server.
3017 The maximum this may be safely set to is 32.
3019 The startup= and idle= options allow some measure of skew in your
3024 Sets the minimum number of processes to spawn when Squid
3025 starts or reconfigures. When set to zero the first request will
3026 cause spawning of the first child process to handle it.
3028 Starting too few children temporary slows Squid under load while it
3029 tries to spawn enough additional processes to cope with traffic.
3033 Sets a minimum of how many processes Squid is to try and keep available
3034 at all times. When traffic begins to rise above what the existing
3035 processes can handle this many more will be spawned up to the maximum
3036 configured. A minimum setting of 1 is required.
3040 The number of requests each certificate validator helper can handle in
3041 parallel. A value of 0 indicates the certficate validator does not
3042 support concurrency. Defaults to 1.
3044 When this directive is set to a value >= 1 then the protocol
3045 used to communicate with the helper is modified to include
3046 a request ID in front of the request/response. The request
3047 ID from the request must be echoed back with the response
3052 Sets the maximum number of queued requests.
3053 If the queued requests exceed queue size for more than 3 minutes
3054 squid aborts its operation.
3055 The default value is set to 2*numberofchildren.
3057 You must have at least one ssl_crt_validator process.
3061 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
3062 -----------------------------------------------------------------------------
3070 To specify other caches in a hierarchy, use the format:
3072 cache_peer hostname type http-port icp-port [options]
3077 # hostname type port port options
3078 # -------------------- -------- ----- ----- -----------
3079 cache_peer parent.foo.net parent 3128 3130 default
3080 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
3081 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
3082 cache_peer example.com parent 80 0 default
3083 cache_peer cdn.example.com sibling 3128 0
3085 type: either 'parent', 'sibling', or 'multicast'.
3087 proxy-port: The port number where the peer accept HTTP requests.
3088 For other Squid proxies this is usually 3128
3089 For web servers this is usually 80
3091 icp-port: Used for querying neighbor caches about objects.
3092 Set to 0 if the peer does not support ICP or HTCP.
3093 See ICP and HTCP options below for additional details.
3096 ==== ICP OPTIONS ====
3098 You MUST also set icp_port and icp_access explicitly when using these options.
3099 The defaults will prevent peer traffic using ICP.
3102 no-query Disable ICP queries to this neighbor.
3105 Indicates the named peer is a member of a multicast group.
3106 ICP queries will not be sent directly to the peer, but ICP
3107 replies will be accepted from it.
3109 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
3110 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
3113 To only send ICP queries to this neighbor infrequently.
3114 This is used to keep the neighbor round trip time updated
3115 and is usually used in conjunction with weighted-round-robin.
3118 ==== HTCP OPTIONS ====
3120 You MUST also set htcp_port and htcp_access explicitly when using these options.
3121 The defaults will prevent peer traffic using HTCP.
3124 htcp Send HTCP, instead of ICP, queries to the neighbor.
3125 You probably also want to set the "icp-port" to 4827
3126 instead of 3130. This directive accepts a comma separated
3127 list of options described below.
3129 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3131 htcp=no-clr Send HTCP to the neighbor but without
3132 sending any CLR requests. This cannot be used with
3135 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3136 This cannot be used with no-clr.
3139 Send HTCP to the neighbor including CLRs but only when
3140 they do not result from PURGE requests.
3143 Forward any HTCP CLR requests this proxy receives to the peer.
3146 ==== PEER SELECTION METHODS ====
3148 The default peer selection method is ICP, with the first responding peer
3149 being used as source. These options can be used for better load balancing.
3152 default This is a parent cache which can be used as a "last-resort"
3153 if a peer cannot be located by any of the peer-selection methods.
3154 If specified more than once, only the first is used.
3156 round-robin Load-Balance parents which should be used in a round-robin
3157 fashion in the absence of any ICP queries.
3158 weight=N can be used to add bias.
3160 weighted-round-robin
3161 Load-Balance parents which should be used in a round-robin
3162 fashion with the frequency of each parent being based on the
3163 round trip time. Closer parents are used more often.
3164 Usually used for background-ping parents.
3165 weight=N can be used to add bias.
3167 carp Load-Balance parents which should be used as a CARP array.
3168 The requests will be distributed among the parents based on the
3169 CARP load balancing hash function based on their weight.
3171 userhash Load-balance parents based on the client proxy_auth or ident username.
3173 sourcehash Load-balance parents based on the client source IP.
3176 To be used only for cache peers of type "multicast".
3177 ALL members of this multicast group have "sibling"
3178 relationship with it, not "parent". This is to a multicast
3179 group when the requested object would be fetched only from
3180 a "parent" cache, anyway. It's useful, e.g., when
3181 configuring a pool of redundant Squid proxies, being
3182 members of the same multicast group.
3185 ==== PEER SELECTION OPTIONS ====
3187 weight=N use to affect the selection of a peer during any weighted
3188 peer-selection mechanisms.
3189 The weight must be an integer; default is 1,
3190 larger weights are favored more.
3191 This option does not affect parent selection if a peering
3192 protocol is not in use.
3194 basetime=N Specify a base amount to be subtracted from round trip
3196 It is subtracted before division by weight in calculating
3197 which parent to fectch from. If the rtt is less than the
3198 base time the rtt is set to a minimal value.
3200 ttl=N Specify a TTL to use when sending multicast ICP queries
3202 Only useful when sending to a multicast group.
3203 Because we don't accept ICP replies from random
3204 hosts, you must configure other group members as
3205 peers with the 'multicast-responder' option.
3207 no-delay To prevent access to this neighbor from influencing the
3210 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3211 enabled) for this host from the specified URL rather
3212 than the Squid default location.
3215 ==== CARP OPTIONS ====
3217 carp-key=key-specification
3218 use a different key than the full URL to hash against the peer.
3219 the key-specification is a comma-separated list of the keywords
3220 scheme, host, port, path, params
3221 Order is not important.
3223 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3225 originserver Causes this parent to be contacted as an origin server.
3226 Meant to be used in accelerator setups when the peer
3230 Set the Host header of requests forwarded to this peer.
3231 Useful in accelerator setups where the server (peer)
3232 expects a certain domain name but clients may request
3233 others. ie example.com or www.example.com
3235 no-digest Disable request of cache digests.
3238 Disables requesting ICMP RTT database (NetDB).
3241 ==== AUTHENTICATION OPTIONS ====
3244 If this is a personal/workgroup proxy and your parent
3245 requires proxy authentication.
3247 Note: The string can include URL escapes (i.e. %20 for
3248 spaces). This also means % must be written as %%.
3251 Send login details received from client to this peer.
3252 Both Proxy- and WWW-Authorization headers are passed
3253 without alteration to the peer.
3254 Authentication is not required by Squid for this to work.
3256 Note: This will pass any form of authentication but
3257 only Basic auth will work through a proxy unless the
3258 connection-auth options are also used.
3260 login=PASS Send login details received from client to this peer.
3261 Authentication is not required by this option.
3263 If there are no client-provided authentication headers
3264 to pass on, but username and password are available
3265 from an external ACL user= and password= result tags
3266 they may be sent instead.
3268 Note: To combine this with proxy_auth both proxies must
3269 share the same user database as HTTP only allows for
3270 a single login (one for proxy, one for origin server).
3271 Also be warned this will expose your users proxy
3272 password to the peer. USE WITH CAUTION
3275 Send the username to the upstream cache, but with a
3276 fixed password. This is meant to be used when the peer
3277 is in another administrative domain, but it is still
3278 needed to identify each user.
3279 The star can optionally be followed by some extra
3280 information which is added to the username. This can
3281 be used to identify this proxy to the peer, similar to
3282 the login=username:password option above.
3285 If this is a personal/workgroup proxy and your parent
3286 requires a secure proxy authentication.
3287 The first principal from the default keytab or defined by
3288 the environment variable KRB5_KTNAME will be used.
3290 WARNING: The connection may transmit requests from multiple
3291 clients. Negotiate often assumes end-to-end authentication
3292 and a single-client. Which is not strictly true here.
3294 login=NEGOTIATE:principal_name
3295 If this is a personal/workgroup proxy and your parent
3296 requires a secure proxy authentication.
3297 The principal principal_name from the default keytab or
3298 defined by the environment variable KRB5_KTNAME will be
3301 WARNING: The connection may transmit requests from multiple
3302 clients. Negotiate often assumes end-to-end authentication
3303 and a single-client. Which is not strictly true here.
3305 connection-auth=on|off
3306 Tell Squid that this peer does or not support Microsoft
3307 connection oriented authentication, and any such
3308 challenges received from there should be ignored.
3309 Default is auto to automatically determine the status
3313 ==== SSL / HTTPS / TLS OPTIONS ====
3315 ssl Encrypt connections to this peer with SSL/TLS.
3317 sslcert=/path/to/ssl/certificate
3318 A client SSL certificate to use when connecting to
3321 sslkey=/path/to/ssl/key
3322 The private SSL key corresponding to sslcert above.
3323 If 'sslkey' is not specified 'sslcert' is assumed to
3324 reference a combined file containing both the
3325 certificate and the key.
3327 sslcipher=... The list of valid SSL ciphers to use when connecting
3331 The minimum TLS protocol version to permit. To control
3332 SSLv3 use the ssloptions= parameter.
3333 Supported Values: 1.0 (default), 1.1, 1.2
3335 ssloptions=... Specify various SSL implementation options:
3337 NO_SSLv3 Disallow the use of SSLv3
3339 NO_TLSv1 Disallow the use of TLSv1.0
3341 NO_TLSv1_1 Disallow the use of TLSv1.1
3343 NO_TLSv1_2 Disallow the use of TLSv1.2
3346 Always create a new key when using
3347 temporary/ephemeral DH key exchanges
3350 Disable use of RFC5077 session tickets.
3351 Some servers may have problems
3352 understanding the TLS extension due
3353 to ambiguous specification in RFC4507.
3355 ALL Enable various bug workarounds
3356 suggested as "harmless" by OpenSSL
3357 Be warned that this reduces SSL/TLS
3358 strength to some attacks.
3360 See the OpenSSL SSL_CTX_set_options documentation for a
3363 sslcafile=... A file containing additional CA certificates to use
3364 when verifying the peer certificate.
3366 sslcapath=... A directory containing additional CA certificates to
3367 use when verifying the peer certificate.
3369 sslcrlfile=... A certificate revocation list file to use when
3370 verifying the peer certificate.
3372 sslflags=... Specify various flags modifying the SSL implementation:
3375 Accept certificates even if they fail to
3379 Don't use the default CA list built in
3383 Don't verify the peer certificate
3384 matches the server name
3386 ssldomain= The peer name as advertised in it's certificate.
3387 Used for verifying the correctness of the received peer
3388 certificate. If not specified the peer hostname will be
3392 Enable the "Front-End-Https: On" header needed when
3393 using Squid as a SSL frontend in front of Microsoft OWA.
3394 See MS KB document Q307347 for details on this header.
3395 If set to auto the header will only be added if the
3396 request is forwarded as a https:// URL.
3399 ==== GENERAL OPTIONS ====
3402 A peer-specific connect timeout.
3403 Also see the peer_connect_timeout directive.
3405 connect-fail-limit=N
3406 How many times connecting to a peer must fail before
3407 it is marked as down. Standby connection failures
3408 count towards this limit. Default is 10.
3410 allow-miss Disable Squid's use of only-if-cached when forwarding
3411 requests to siblings. This is primarily useful when
3412 icp_hit_stale is used by the sibling. To extensive use
3413 of this option may result in forwarding loops, and you
3414 should avoid having two-way peerings with this option.
3415 For example to deny peer usage on requests from peer
3416 by denying cache_peer_access if the source is a peer.
3418 max-conn=N Limit the number of concurrent connections the Squid
3419 may open to this peer, including already opened idle
3420 and standby connections. There is no peer-specific
3421 connection limit by default.
3423 A peer exceeding the limit is not used for new
3424 requests unless a standby connection is available.
3426 max-conn currently works poorly with idle persistent
3427 connections: When a peer reaches its max-conn limit,
3428 and there are idle persistent connections to the peer,
3429 the peer may not be selected because the limiting code
3430 does not know whether Squid can reuse those idle
3433 standby=N Maintain a pool of N "hot standby" connections to an
3434 UP peer, available for requests when no idle
3435 persistent connection is available (or safe) to use.
3436 By default and with zero N, no such pool is maintained.
3437 N must not exceed the max-conn limit (if any).
3439 At start or after reconfiguration, Squid opens new TCP
3440 standby connections until there are N connections
3441 available and then replenishes the standby pool as
3442 opened connections are used up for requests. A used
3443 connection never goes back to the standby pool, but
3444 may go to the regular idle persistent connection pool
3445 shared by all peers and origin servers.
3447 Squid never opens multiple new standby connections
3448 concurrently. This one-at-a-time approach minimizes
3449 flooding-like effect on peers. Furthermore, just a few
3450 standby connections should be sufficient in most cases
3451 to supply most new requests with a ready-to-use
3454 Standby connections obey server_idle_pconn_timeout.
3455 For the feature to work as intended, the peer must be
3456 configured to accept and keep them open longer than
3457 the idle timeout at the connecting Squid, to minimize
3458 race conditions typical to idle used persistent
3459 connections. Default request_timeout and
3460 server_idle_pconn_timeout values ensure such a
3463 name=xxx Unique name for the peer.
3464 Required if you have multiple peers on the same host
3465 but different ports.
3466 This name can be used in cache_peer_access and similar
3467 directives to dentify the peer.
3468 Can be used by outgoing access controls through the
3471 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3472 requests to this peer. Use normal address selection instead.
3473 This overrides the spoof_client_ip ACL.
3475 proxy-only objects fetched from the peer will not be stored locally.
3479 NAME: cache_peer_access
3484 Use to limit the requests for which a neighbor proxy will be
3485 queried. Peers with no restrictions are queried for all requests.
3488 cache_peer_access cache-host allow|deny [!]aclname ...
3490 The syntax is identical to 'http_access' and the other lists of
3491 ACL elements. See the comments for 'http_access', or the
3492 Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3495 NAME: neighbor_type_domain
3496 TYPE: hostdomaintype
3498 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3501 Modify the cache_peer neighbor type when passing requests
3502 about specific domains to the peer.
3505 neighbor_type_domain neighbor parent|sibling domain domain ...
3508 cache_peer foo.example.com parent 3128 3130
3509 neighbor_type_domain foo.example.com sibling .au .de
3511 The above configuration treats all requests to foo.example.com as a
3512 parent proxy unless the request is for a .au or .de ccTLD domain name.
3515 NAME: dead_peer_timeout
3519 LOC: Config.Timeout.deadPeer
3521 This controls how long Squid waits to declare a peer cache
3522 as "dead." If there are no ICP replies received in this
3523 amount of time, Squid will declare the peer dead and not
3524 expect to receive any further ICP replies. However, it
3525 continues to send ICP queries, and will mark the peer as
3526 alive upon receipt of the first subsequent ICP reply.
3528 This timeout also affects when Squid expects to receive ICP
3529 replies from peers. If more than 'dead_peer' seconds have
3530 passed since the last ICP reply was received, Squid will not
3531 expect to receive an ICP reply on the next query. Thus, if
3532 your time between requests is greater than this timeout, you
3533 will see a lot of requests sent DIRECT to origin servers
3534 instead of to your parents.
3537 NAME: forward_max_tries
3540 LOC: Config.forward_max_tries
3542 Controls how many different forward paths Squid will try
3543 before giving up. See also forward_timeout.
3545 NOTE: connect_retries (default: none) can make each of these
3546 possible forwarding paths be tried multiple times.
3550 MEMORY CACHE OPTIONS
3551 -----------------------------------------------------------------------------
3558 LOC: Config.memMaxSize
3560 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3561 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3562 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3563 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3565 'cache_mem' specifies the ideal amount of memory to be used
3567 * In-Transit objects
3569 * Negative-Cached objects
3571 Data for these objects are stored in 4 KB blocks. This
3572 parameter specifies the ideal upper limit on the total size of
3573 4 KB blocks allocated. In-Transit objects take the highest
3576 In-transit objects have priority over the others. When
3577 additional space is needed for incoming data, negative-cached
3578 and hot objects will be released. In other words, the
3579 negative-cached and hot objects will fill up any unused space
3580 not needed for in-transit objects.
3582 If circumstances require, this limit will be exceeded.
3583 Specifically, if your incoming request rate requires more than
3584 'cache_mem' of memory to hold in-transit objects, Squid will
3585 exceed this limit to satisfy the new requests. When the load
3586 decreases, blocks will be freed until the high-water mark is
3587 reached. Thereafter, blocks will be used to store hot
3590 If shared memory caching is enabled, Squid does not use the shared
3591 cache space for in-transit objects, but they still consume as much
3592 local memory as they need. For more details about the shared memory
3593 cache, see memory_cache_shared.
3596 NAME: maximum_object_size_in_memory
3600 LOC: Config.Store.maxInMemObjSize
3602 Objects greater than this size will not be attempted to kept in
3603 the memory cache. This should be set high enough to keep objects
3604 accessed frequently in memory to improve performance whilst low
3605 enough to keep larger objects from hoarding cache_mem.
3608 NAME: memory_cache_shared
3611 LOC: Config.memShared
3613 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3615 Controls whether the memory cache is shared among SMP workers.
3617 The shared memory cache is meant to occupy cache_mem bytes and replace
3618 the non-shared memory cache, although some entities may still be
3619 cached locally by workers for now (e.g., internal and in-transit
3620 objects may be served from a local memory cache even if shared memory
3621 caching is enabled).
3623 By default, the memory cache is shared if and only if all of the
3624 following conditions are satisfied: Squid runs in SMP mode with
3625 multiple workers, cache_mem is positive, and Squid environment
3626 supports required IPC primitives (e.g., POSIX shared memory segments
3627 and GCC-style atomic operations).
3629 To avoid blocking locks, shared memory uses opportunistic algorithms
3630 that do not guarantee that every cachable entity that could have been
3631 shared among SMP workers will actually be shared.
3633 Currently, entities exceeding 32KB in size cannot be shared.
3636 NAME: memory_cache_mode
3640 DEFAULT_DOC: Keep the most recently fetched objects in memory
3642 Controls which objects to keep in the memory cache (cache_mem)
3644 always Keep most recently fetched objects in memory (default)
3646 disk Only disk cache hits are kept in memory, which means
3647 an object must first be cached on disk and then hit
3648 a second time before cached in memory.
3650 network Only objects fetched from network is kept in memory
3653 NAME: memory_replacement_policy
3655 LOC: Config.memPolicy
3658 The memory replacement policy parameter determines which
3659 objects are purged from memory when memory space is needed.
3661 See cache_replacement_policy for details on algorithms.
3666 -----------------------------------------------------------------------------
3669 NAME: cache_replacement_policy
3671 LOC: Config.replPolicy
3674 The cache replacement policy parameter determines which
3675 objects are evicted (replaced) when disk space is needed.
3677 lru : Squid's original list based LRU policy
3678 heap GDSF : Greedy-Dual Size Frequency
3679 heap LFUDA: Least Frequently Used with Dynamic Aging
3680 heap LRU : LRU policy implemented using a heap
3682 Applies to any cache_dir lines listed below this directive.
3684 The LRU policies keeps recently referenced objects.
3686 The heap GDSF policy optimizes object hit rate by keeping smaller
3687 popular objects in cache so it has a better chance of getting a
3688 hit. It achieves a lower byte hit rate than LFUDA though since
3689 it evicts larger (possibly popular) objects.
3691 The heap LFUDA policy keeps popular objects in cache regardless of
3692 their size and thus optimizes byte hit rate at the expense of
3693 hit rate since one large, popular object will prevent many
3694 smaller, slightly less popular objects from being cached.
3696 Both policies utilize a dynamic aging mechanism that prevents
3697 cache pollution that can otherwise occur with frequency-based
3698 replacement policies.
3700 NOTE: if using the LFUDA replacement policy you should increase
3701 the value of maximum_object_size above its default of 4 MB to
3702 to maximize the potential byte hit rate improvement of LFUDA.
3704 For more information about the GDSF and LFUDA cache replacement
3705 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3706 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3709 NAME: minimum_object_size
3713 DEFAULT_DOC: no limit
3714 LOC: Config.Store.minObjectSize
3716 Objects smaller than this size will NOT be saved on disk. The
3717 value is specified in bytes, and the default is 0 KB, which
3718 means all responses can be stored.
3721 NAME: maximum_object_size
3725 LOC: Config.Store.maxObjectSize
3727 Set the default value for max-size parameter on any cache_dir.
3728 The value is specified in bytes, and the default is 4 MB.
3730 If you wish to get a high BYTES hit ratio, you should probably
3731 increase this (one 32 MB object hit counts for 3200 10KB
3734 If you wish to increase hit ratio more than you want to
3735 save bandwidth you should leave this low.
3737 NOTE: if using the LFUDA replacement policy you should increase
3738 this value to maximize the byte hit rate improvement of LFUDA!
3739 See cache_replacement_policy for a discussion of this policy.
3745 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3746 LOC: Config.cacheSwap
3749 cache_dir Type Directory-Name Fs-specific-data [options]
3751 You can specify multiple cache_dir lines to spread the
3752 cache among different disk partitions.
3754 Type specifies the kind of storage system to use. Only "ufs"
3755 is built by default. To enable any of the other storage systems
3756 see the --enable-storeio configure option.
3758 'Directory' is a top-level directory where cache swap
3759 files will be stored. If you want to use an entire disk
3760 for caching, this can be the mount-point directory.
3761 The directory must exist and be writable by the Squid
3762 process. Squid will NOT create this directory for you.
3764 In SMP configurations, cache_dir must not precede the workers option
3765 and should use configuration macros or conditionals to give each
3766 worker interested in disk caching a dedicated cache directory.
3769 ==== The ufs store type ====
3771 "ufs" is the old well-known Squid storage format that has always
3775 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3777 'Mbytes' is the amount of disk space (MB) to use under this
3778 directory. The default is 100 MB. Change this to suit your
3779 configuration. Do NOT put the size of your disk drive here.
3780 Instead, if you want Squid to use the entire disk drive,
3781 subtract 20% and use that value.
3783 'L1' is the number of first-level subdirectories which
3784 will be created under the 'Directory'. The default is 16.
3786 'L2' is the number of second-level subdirectories which
3787 will be created under each first-level directory. The default
3791 ==== The aufs store type ====
3793 "aufs" uses the same storage format as "ufs", utilizing
3794 POSIX-threads to avoid blocking the main Squid process on
3795 disk-I/O. This was formerly known in Squid as async-io.
3798 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3800 see argument descriptions under ufs above
3803 ==== The diskd store type ====
3805 "diskd" uses the same storage format as "ufs", utilizing a
3806 separate process to avoid blocking the main Squid process on
3810 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3812 see argument descriptions under ufs above
3814 Q1 specifies the number of unacknowledged I/O requests when Squid
3815 stops opening new files. If this many messages are in the queues,
3816 Squid won't open new files. Default is 64
3818 Q2 specifies the number of unacknowledged messages when Squid
3819 starts blocking. If this many messages are in the queues,
3820 Squid blocks until it receives some replies. Default is 72
3822 When Q1 < Q2 (the default), the cache directory is optimized
3823 for lower response time at the expense of a decrease in hit
3824 ratio. If Q1 > Q2, the cache directory is optimized for
3825 higher hit ratio at the expense of an increase in response
3829 ==== The rock store type ====
3832 cache_dir rock Directory-Name Mbytes [options]
3834 The Rock Store type is a database-style storage. All cached
3835 entries are stored in a "database" file, using fixed-size slots.
3836 A single entry occupies one or more slots.
3838 If possible, Squid using Rock Store creates a dedicated kid
3839 process called "disker" to avoid blocking Squid worker(s) on disk
3840 I/O. One disker kid is created for each rock cache_dir. Diskers
3841 are created only when Squid, running in daemon mode, has support
3842 for the IpcIo disk I/O module.
3844 swap-timeout=msec: Squid will not start writing a miss to or
3845 reading a hit from disk if it estimates that the swap operation
3846 will take more than the specified number of milliseconds. By
3847 default and when set to zero, disables the disk I/O time limit
3848 enforcement. Ignored when using blocking I/O module because
3849 blocking synchronous I/O does not allow Squid to estimate the
3850 expected swap wait time.
3852 max-swap-rate=swaps/sec: Artificially limits disk access using
3853 the specified I/O rate limit. Swap out requests that
3854 would cause the average I/O rate to exceed the limit are
3855 delayed. Individual swap in requests (i.e., hits or reads) are
3856 not delayed, but they do contribute to measured swap rate and
3857 since they are placed in the same FIFO queue as swap out
3858 requests, they may wait longer if max-swap-rate is smaller.
3859 This is necessary on file systems that buffer "too
3860 many" writes and then start blocking Squid and other processes
3861 while committing those writes to disk. Usually used together
3862 with swap-timeout to avoid excessive delays and queue overflows
3863 when disk demand exceeds available disk "bandwidth". By default
3864 and when set to zero, disables the disk I/O rate limit
3865 enforcement. Currently supported by IpcIo module only.
3867 slot-size=bytes: The size of a database "record" used for
3868 storing cached responses. A cached response occupies at least
3869 one slot and all database I/O is done using individual slots so
3870 increasing this parameter leads to more disk space waste while
3871 decreasing it leads to more disk I/O overheads. Should be a
3872 multiple of your operating system I/O page size. Defaults to
3873 16KBytes. A housekeeping header is stored with each slot and
3874 smaller slot-sizes will be rejected. The header is smaller than
3878 ==== COMMON OPTIONS ====
3880 no-store no new objects should be stored to this cache_dir.
3882 min-size=n the minimum object size in bytes this cache_dir
3883 will accept. It's used to restrict a cache_dir
3884 to only store large objects (e.g. AUFS) while
3885 other stores are optimized for smaller objects
3889 max-size=n the maximum object size in bytes this cache_dir
3891 The value in maximum_object_size directive sets
3892 the default unless more specific details are
3893 available (ie a small store capacity).
3895 Note: To make optimal use of the max-size limits you should order
3896 the cache_dir lines with the smallest max-size value first.
3900 # Uncomment and adjust the following to add a disk cache directory.
3901 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3905 NAME: store_dir_select_algorithm
3907 LOC: Config.store_dir_select_algorithm
3910 How Squid selects which cache_dir to use when the response
3911 object will fit into more than one.
3913 Regardless of which algorithm is used the cache_dir min-size
3914 and max-size parameters are obeyed. As such they can affect
3915 the selection algorithm by limiting the set of considered
3922 This algorithm is suited to caches with similar cache_dir
3923 sizes and disk speeds.
3925 The disk with the least I/O pending is selected.
3926 When there are multiple disks with the same I/O load ranking
3927 the cache_dir with most available capacity is selected.
3929 When a mix of cache_dir sizes are configured the faster disks
3930 have a naturally lower I/O loading and larger disks have more
3931 capacity. So space used to store objects and data throughput
3932 may be very unbalanced towards larger disks.
3937 This algorithm is suited to caches with unequal cache_dir
3940 Each cache_dir is selected in a rotation. The next suitable
3943 Available cache_dir capacity is only considered in relation
3944 to whether the object will fit and meets the min-size and
3945 max-size parameters.
3947 Disk I/O loading is only considered to prevent overload on slow
3948 disks. This algorithm does not spread objects by size, so any
3949 I/O loading per-disk may appear very unbalanced and volatile.
3951 If several cache_dirs use similar min-size, max-size, or other
3952 limits to to reject certain responses, then do not group such
3953 cache_dir lines together, to avoid round-robin selection bias
3954 towards the first cache_dir after the group. Instead, interleave
3955 cache_dir lines from different groups. For example:
3957 store_dir_select_algorithm round-robin
3958 cache_dir rock /hdd1 ... min-size=100000
3959 cache_dir rock /ssd1 ... max-size=99999
3960 cache_dir rock /hdd2 ... min-size=100000
3961 cache_dir rock /ssd2 ... max-size=99999
3962 cache_dir rock /hdd3 ... min-size=100000
3963 cache_dir rock /ssd3 ... max-size=99999
3966 NAME: max_open_disk_fds
3968 LOC: Config.max_open_disk_fds
3970 DEFAULT_DOC: no limit
3972 To avoid having disk as the I/O bottleneck Squid can optionally
3973 bypass the on-disk cache if more than this amount of disk file
3974 descriptors are open.
3976 A value of 0 indicates no limit.
3979 NAME: cache_swap_low
3980 COMMENT: (percent, 0-100)
3983 LOC: Config.Swap.lowWaterMark
3985 The low-water mark for cache object replacement.
3986 Replacement begins when the swap (disk) usage is above the
3987 low-water mark and attempts to maintain utilization near the
3988 low-water mark. As swap utilization gets close to high-water
3989 mark object eviction becomes more aggressive. If utilization is
3990 close to the low-water mark less replacement is done each time.
3992 Defaults are 90% and 95%. If you have a large cache, 5% could be
3993 hundreds of MB. If this is the case you may wish to set these
3994 numbers closer together.
3996 See also cache_swap_high
3999 NAME: cache_swap_high
4000 COMMENT: (percent, 0-100)
4003 LOC: Config.Swap.highWaterMark
4005 The high-water mark for cache object replacement.
4006 Replacement begins when the swap (disk) usage is above the
4007 low-water mark and attempts to maintain utilization near the
4008 low-water mark. As swap utilization gets close to high-water
4009 mark object eviction becomes more aggressive. If utilization is
4010 close to the low-water mark less replacement is done each time.
4012 Defaults are 90% and 95%. If you have a large cache, 5% could be
4013 hundreds of MB. If this is the case you may wish to set these
4014 numbers closer together.
4016 See also cache_swap_low
4021 -----------------------------------------------------------------------------
4028 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
4032 logformat <name> <format specification>
4034 Defines an access log format.
4036 The <format specification> is a string with embedded % format codes
4038 % format codes all follow the same basic structure where all but
4039 the formatcode is optional. Output strings are automatically escaped
4040 as required according to their context and the output format
4041 modifiers are usually not needed, but can be specified if an explicit
4042 output format is desired.
4044 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
4046 " output in quoted string format
4047 [ output in squid text log format as used by log_mime_hdrs
4048 # output in URL quoted format
4053 width minimum and/or maximum field width:
4054 [width_min][.width_max]
4055 When minimum starts with 0, the field is zero-padded.
4056 String values exceeding maximum width are truncated.
4058 {arg} argument such as header name etc
4062 % a literal % character
4063 sn Unique sequence number per log line entry
4064 err_code The ID of an error response served by Squid or
4065 a similar internal error identifier.
4066 err_detail Additional err_code-dependent error information.
4067 note The annotation specified by the argument. Also
4068 logs the adaptation meta headers set by the
4069 adaptation_meta configuration parameter.
4070 If no argument given all annotations logged.
4071 The argument may include a separator to use with
4074 By default, multiple note values are separated with ","
4075 and multiple notes are separated with "\r\n".
4076 When logging named notes with %{name}note, the
4077 explicitly configured separator is used between note
4078 values. When logging all notes with %note, the
4079 explicitly configured separator is used between
4080 individual notes. There is currently no way to
4081 specify both value and notes separators when logging
4082 all notes with %note.
4084 Connection related format codes:
4086 >a Client source IP address
4088 >p Client source port
4089 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
4090 >la Local IP address the client connected to
4091 >lp Local port number the client connected to
4092 >qos Client connection TOS/DSCP value set by Squid
4093 >nfmark Client connection netfilter mark set by Squid
4095 la Local listening IP address the client connection was connected to.
4096 lp Local listening port number the client connection was connected to.
4098 <a Server IP address of the last server or peer connection
4099 <A Server FQDN or peer name
4100 <p Server port number of the last server or peer connection
4101 <la Local IP address of the last server or peer connection
4102 <lp Local port number of the last server or peer connection
4103 <qos Server connection TOS/DSCP value set by Squid
4104 <nfmark Server connection netfilter mark set by Squid
4106 Time related format codes:
4108 ts Seconds since epoch
4109 tu subsecond time (milliseconds)
4110 tl Local time. Optional strftime format argument
4111 default %d/%b/%Y:%H:%M:%S %z
4112 tg GMT time. Optional strftime format argument
4113 default %d/%b/%Y:%H:%M:%S %z
4114 tr Response time (milliseconds)
4115 dt Total time spent making DNS lookups (milliseconds)
4116 tS Approximate master transaction start time in
4117 <full seconds since epoch>.<fractional seconds> format.
4118 Currently, Squid considers the master transaction
4119 started when a complete HTTP request header initiating
4120 the transaction is received from the client. This is
4121 the same value that Squid uses to calculate transaction
4122 response time when logging %tr to access.log. Currently,
4123 Squid uses millisecond resolution for %tS values,
4124 similar to the default access.log "current time" field
4127 Access Control related format codes:
4129 et Tag returned by external acl
4130 ea Log string returned by external acl
4131 un User name (any available)
4132 ul User name from authentication
4133 ue User name from external acl helper
4134 ui User name from ident
4135 us User name from SSL
4136 credentials Client credentials. The exact meaning depends on
4137 the authentication scheme: For Basic authentication,
4138 it is the password; for Digest, the realm sent by the
4139 client; for NTLM and Negotiate, the client challenge
4140 or client credentials prefixed with "YR " or "KK ".
4142 HTTP related format codes:
4146 [http::]rm Request method (GET/POST etc)
4147 [http::]>rm Request method from client
4148 [http::]<rm Request method sent to server or peer
4149 [http::]ru Request URL from client (historic, filtered for logging)
4150 [http::]>ru Request URL from client
4151 [http::]<ru Request URL sent to server or peer
4152 [http::]>rs Request URL scheme from client
4153 [http::]<rs Request URL scheme sent to server or peer
4154 [http::]>rd Request URL domain from client
4155 [http::]<rd Request URL domain sent to server or peer
4156 [http::]>rP Request URL port from client
4157 [http::]<rP Request URL port sent to server or peer
4158 [http::]rp Request URL path excluding hostname
4159 [http::]>rp Request URL path excluding hostname from client
4160 [http::]<rp Request URL path excluding hostname sent to server or peer
4161 [http::]rv Request protocol version
4162 [http::]>rv Request protocol version from client
4163 [http::]<rv Request protocol version sent to server or peer
4165 [http::]>h Original received request header.
4166 Usually differs from the request header sent by
4167 Squid, although most fields are often preserved.
4168 Accepts optional header field name/value filter
4169 argument using name[:[separator]element] format.
4170 [http::]>ha Received request header after adaptation and
4171 redirection (pre-cache REQMOD vectoring point).
4172 Usually differs from the request header sent by
4173 Squid, although most fields are often preserved.
4174 Optional header name argument as for >h
4179 [http::]<Hs HTTP status code received from the next hop
4180 [http::]>Hs HTTP status code sent to the client
4182 [http::]<h Reply header. Optional header name argument
4185 [http::]mt MIME content type
4190 [http::]st Total size of request + reply traffic with client
4191 [http::]>st Total size of request received from client.
4192 Excluding chunked encoding bytes.
4193 [http::]<st Total size of reply sent to client (after adaptation)
4195 [http::]>sh Size of request headers received from client
4196 [http::]<sh Size of reply headers sent to client (after adaptation)
4198 [http::]<sH Reply high offset sent
4199 [http::]<sS Upstream object size
4201 [http::]<bs Number of HTTP-equivalent message body bytes
4202 received from the next hop, excluding chunked
4203 transfer encoding and control messages.
4204 Generated FTP/Gopher listings are treated as
4210 [http::]<pt Peer response time in milliseconds. The timer starts
4211 when the last request byte is sent to the next hop
4212 and stops when the last response byte is received.
4213 [http::]<tt Total time in milliseconds. The timer
4214 starts with the first connect request (or write I/O)
4215 sent to the first selected peer. The timer stops
4216 with the last I/O with the last peer.
4218 Squid handling related format codes:
4220 Ss Squid request status (TCP_MISS etc)
4221 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4223 SSL-related format codes:
4225 ssl::bump_mode SslBump decision for the transaction:
4227 For CONNECT requests that initiated bumping of
4228 a connection and for any request received on
4229 an already bumped connection, Squid logs the
4230 corresponding SslBump mode ("server-first" or
4231 "client-first"). See the ssl_bump option for
4232 more information about these modes.
4234 A "none" token is logged for requests that
4235 triggered "ssl_bump" ACL evaluation matching
4236 either a "none" rule or no rules at all.
4238 In all other cases, a single dash ("-") is
4241 ssl::>sni SSL client SNI sent to Squid. Available only
4242 after the peek, stare, or splice SSL bumping
4245 If ICAP is enabled, the following code becomes available (as
4246 well as ICAP log codes documented with the icap_log option):
4248 icap::tt Total ICAP processing time for the HTTP
4249 transaction. The timer ticks when ICAP
4250 ACLs are checked and when ICAP
4251 transaction is in progress.
4253 If adaptation is enabled the following three codes become available:
4255 adapt::<last_h The header of the last ICAP response or
4256 meta-information from the last eCAP
4257 transaction related to the HTTP transaction.
4258 Like <h, accepts an optional header name
4261 adapt::sum_trs Summed adaptation transaction response
4262 times recorded as a comma-separated list in
4263 the order of transaction start time. Each time
4264 value is recorded as an integer number,
4265 representing response time of one or more
4266 adaptation (ICAP or eCAP) transaction in
4267 milliseconds. When a failed transaction is
4268 being retried or repeated, its time is not
4269 logged individually but added to the
4270 replacement (next) transaction. See also:
4273 adapt::all_trs All adaptation transaction response times.
4274 Same as adaptation_strs but response times of
4275 individual transactions are never added
4276 together. Instead, all transaction response
4277 times are recorded individually.
4279 You can prefix adapt::*_trs format codes with adaptation
4280 service name in curly braces to record response time(s) specific
4281 to that service. For example: %{my_service}adapt::sum_trs
4283 If SSL is enabled, the following formating codes become available:
4285 %ssl::>cert_subject 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.
4291 %ssl::>cert_issuer The Issuer field of the received client
4292 SSL certificate or a dash ('-') if Squid has
4293 received an invalid/malformed certificate or
4294 no certificate at all. Consider encoding the
4295 logged value because Issuer often has spaces.
4297 The default formats available (which do not need re-defining) are:
4299 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
4300 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
4301 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
4302 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
4303 logformat useragent %>a [%tl] "%{User-Agent}>h"
4305 NOTE: When the log_mime_hdrs directive is set to ON.
4306 The squid, common and combined formats have a safely encoded copy
4307 of the mime headers appended to each line within a pair of brackets.
4309 NOTE: The common and combined formats are not quite true to the Apache definition.
4310 The logs from Squid contain an extra status and hierarchy code appended.
4314 NAME: access_log cache_access_log
4316 LOC: Config.Log.accesslogs
4317 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
4319 Configures whether and how Squid logs HTTP and ICP transactions.
4320 If access logging is enabled, a single line is logged for every
4321 matching HTTP or ICP request. The recommended directive formats are:
4323 access_log <module>:<place> [option ...] [acl acl ...]
4324 access_log none [acl acl ...]
4326 The following directive format is accepted but may be deprecated:
4327 access_log <module>:<place> [<logformat name> [acl acl ...]]
4329 In most cases, the first ACL name must not contain the '=' character
4330 and should not be equal to an existing logformat name. You can always
4331 start with an 'all' ACL to work around those restrictions.
4333 Will log to the specified module:place using the specified format (which
4334 must be defined in a logformat directive) those entries which match
4335 ALL the acl's specified (which must be defined in acl clauses).
4336 If no acl is specified, all requests will be logged to this destination.
4338 ===== Available options for the recommended directive format =====
4340 logformat=name Names log line format (either built-in or
4341 defined by a logformat directive). Defaults
4344 buffer-size=64KB Defines approximate buffering limit for log
4345 records (see buffered_logs). Squid should not
4346 keep more than the specified size and, hence,
4347 should flush records before the buffer becomes
4348 full to avoid overflows under normal
4349 conditions (the exact flushing algorithm is
4350 module-dependent though). The on-error option
4351 controls overflow handling.
4353 on-error=die|drop Defines action on unrecoverable errors. The
4354 'drop' action ignores (i.e., does not log)
4355 affected log records. The default 'die' action
4356 kills the affected worker. The drop action
4357 support has not been tested for modules other
4360 rotate=N Specifies the number of log file rotations to
4361 make when you run 'squid -k rotate'. The default
4362 is to obey the logfile_rotate directive. Setting
4363 rotate=0 will disable the file name rotation,
4364 but the log files are still closed and re-opened.
4365 This will enable you to rename the logfiles
4366 yourself just before sending the rotate signal.
4367 Only supported by the stdio module.
4369 ===== Modules Currently available =====
4371 none Do not log any requests matching these ACL.
4372 Do not specify Place or logformat name.
4374 stdio Write each log line to disk immediately at the completion of
4376 Place: the filename and path to be written.
4378 daemon Very similar to stdio. But instead of writing to disk the log
4379 line is passed to a daemon helper for asychronous handling instead.
4380 Place: varies depending on the daemon.
4382 log_file_daemon Place: the file name and path to be written.
4384 syslog To log each request via syslog facility.
4385 Place: The syslog facility and priority level for these entries.
4386 Place Format: facility.priority
4388 where facility could be any of:
4389 authpriv, daemon, local0 ... local7 or user.
4391 And priority could be any of:
4392 err, warning, notice, info, debug.
4394 udp To send each log line as text data to a UDP receiver.
4395 Place: The destination host name or IP and port.
4396 Place Format: //host:port
4398 tcp To send each log line as text data to a TCP receiver.
4399 Lines may be accumulated before sending (see buffered_logs).
4400 Place: The destination host name or IP and port.
4401 Place Format: //host:port
4404 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4410 LOC: Config.Log.icaplogs
4413 ICAP log files record ICAP transaction summaries, one line per
4416 The icap_log option format is:
4417 icap_log <filepath> [<logformat name> [acl acl ...]]
4418 icap_log none [acl acl ...]]
4420 Please see access_log option documentation for details. The two
4421 kinds of logs share the overall configuration approach and many
4424 ICAP processing of a single HTTP message or transaction may
4425 require multiple ICAP transactions. In such cases, multiple
4426 ICAP transaction log lines will correspond to a single access
4429 ICAP log uses logformat codes that make sense for an ICAP
4430 transaction. Header-related codes are applied to the HTTP header
4431 embedded in an ICAP server response, with the following caveats:
4432 For REQMOD, there is no HTTP response header unless the ICAP
4433 server performed request satisfaction. For RESPMOD, the HTTP
4434 request header is the header sent to the ICAP server. For
4435 OPTIONS, there are no HTTP headers.
4437 The following format codes are also available for ICAP logs:
4439 icap::<A ICAP server IP address. Similar to <A.
4441 icap::<service_name ICAP service name from the icap_service
4442 option in Squid configuration file.
4444 icap::ru ICAP Request-URI. Similar to ru.
4446 icap::rm ICAP request method (REQMOD, RESPMOD, or
4447 OPTIONS). Similar to existing rm.
4449 icap::>st Bytes sent to the ICAP server (TCP payload
4450 only; i.e., what Squid writes to the socket).
4452 icap::<st Bytes received from the ICAP server (TCP
4453 payload only; i.e., what Squid reads from
4456 icap::<bs Number of message body bytes received from the
4457 ICAP server. ICAP message body, if any, usually
4458 includes encapsulated HTTP message headers and
4459 possibly encapsulated HTTP message body. The
4460 HTTP body part is dechunked before its size is
4463 icap::tr Transaction response time (in
4464 milliseconds). The timer starts when
4465 the ICAP transaction is created and
4466 stops when the transaction is completed.
4469 icap::tio Transaction I/O time (in milliseconds). The
4470 timer starts when the first ICAP request
4471 byte is scheduled for sending. The timers
4472 stops when the last byte of the ICAP response
4475 icap::to Transaction outcome: ICAP_ERR* for all
4476 transaction errors, ICAP_OPT for OPTION
4477 transactions, ICAP_ECHO for 204
4478 responses, ICAP_MOD for message
4479 modification, and ICAP_SAT for request
4480 satisfaction. Similar to Ss.
4482 icap::Hs ICAP response status code. Similar to Hs.
4484 icap::>h ICAP request header(s). Similar to >h.
4486 icap::<h ICAP response header(s). Similar to <h.
4488 The default ICAP log format, which can be used without an explicit
4489 definition, is called icap_squid:
4491 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4493 See also: logformat, log_icap, and %adapt::<last_h
4496 NAME: logfile_daemon
4498 DEFAULT: @DEFAULT_LOGFILED@
4499 LOC: Log::TheConfig.logfile_daemon
4501 Specify the path to the logfile-writing daemon. This daemon is
4502 used to write the access and store logs, if configured.
4504 Squid sends a number of commands to the log daemon:
4505 L<data>\n - logfile data
4510 r<n>\n - set rotate count to <n>
4511 b<n>\n - 1 = buffer output, 0 = don't buffer output
4513 No responses is expected.
4516 NAME: stats_collection
4518 LOC: Config.accessList.stats_collection
4520 DEFAULT_DOC: Allow logging for all transactions.
4521 COMMENT: allow|deny acl acl...
4523 This options allows you to control which requests gets accounted
4524 in performance counters.
4526 This clause only supports fast acl types.
4527 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4530 NAME: cache_store_log
4533 LOC: Config.Log.store
4535 Logs the activities of the storage manager. Shows which
4536 objects are ejected from the cache, and which objects are
4537 saved and for how long.
4538 There are not really utilities to analyze this data, so you can safely
4539 disable it (the default).
4541 Store log uses modular logging outputs. See access_log for the list
4542 of modules supported.
4545 cache_store_log stdio:@DEFAULT_STORE_LOG@
4546 cache_store_log daemon:@DEFAULT_STORE_LOG@
4549 NAME: cache_swap_state cache_swap_log
4551 LOC: Config.Log.swap
4553 DEFAULT_DOC: Store the journal inside its cache_dir
4555 Location for the cache "swap.state" file. This index file holds
4556 the metadata of objects saved on disk. It is used to rebuild
4557 the cache during startup. Normally this file resides in each
4558 'cache_dir' directory, but you may specify an alternate
4559 pathname here. Note you must give a full filename, not just
4560 a directory. Since this is the index for the whole object
4561 list you CANNOT periodically rotate it!
4563 If %s can be used in the file name it will be replaced with a
4564 a representation of the cache_dir name where each / is replaced
4565 with '.'. This is needed to allow adding/removing cache_dir
4566 lines when cache_swap_log is being used.
4568 If have more than one 'cache_dir', and %s is not used in the name
4569 these swap logs will have names such as:
4575 The numbered extension (which is added automatically)
4576 corresponds to the order of the 'cache_dir' lines in this
4577 configuration file. If you change the order of the 'cache_dir'
4578 lines in this file, these index files will NOT correspond to
4579 the correct 'cache_dir' entry (unless you manually rename
4580 them). We recommend you do NOT use this option. It is
4581 better to keep these index files in each 'cache_dir' directory.
4584 NAME: logfile_rotate
4587 LOC: Config.Log.rotateNumber
4589 Specifies the default number of logfile rotations to make when you
4590 type 'squid -k rotate'. The default is 10, which will rotate
4591 with extensions 0 through 9. Setting logfile_rotate to 0 will
4592 disable the file name rotation, but the logfiles are still closed
4593 and re-opened. This will enable you to rename the logfiles
4594 yourself just before sending the rotate signal.
4596 Note, from Squid-3.1 this option is only a default for cache.log,
4597 that log can be rotated separately by using debug_options.
4599 Note, from Squid-3.6 this option is only a default for access.log
4600 recorded by stdio: module. Those logs can be rotated separately by
4601 using the rotate=N option on their access_log directive.
4603 Note, the 'squid -k rotate' command normally sends a USR1
4604 signal to the running squid process. In certain situations
4605 (e.g. on Linux with Async I/O), USR1 is used for other
4606 purposes, so -k rotate uses another signal. It is best to get
4607 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4614 DEFAULT: @DEFAULT_MIME_TABLE@
4615 LOC: Config.mimeTablePathname
4617 Path to Squid's icon configuration file.
4619 You shouldn't need to change this, but the default file contains
4620 examples and formatting information if you do.
4626 LOC: Config.onoff.log_mime_hdrs
4629 The Cache can record both the request and the response MIME
4630 headers for each HTTP transaction. The headers are encoded
4631 safely and will appear as two bracketed fields at the end of
4632 the access log (for either the native or httpd-emulated log
4633 formats). To enable this logging set log_mime_hdrs to 'on'.
4638 DEFAULT: @DEFAULT_PID_FILE@
4639 LOC: Config.pidFilename
4641 A filename to write the process-id to. To disable, enter "none".
4644 NAME: client_netmask
4646 LOC: Config.Addrs.client_netmask
4648 DEFAULT_DOC: Log full client IP address
4650 A netmask for client addresses in logfiles and cachemgr output.
4651 Change this to protect the privacy of your cache clients.
4652 A netmask of 255.255.255.0 will log all IP's in that range with
4653 the last digit set to '0'.
4656 NAME: strip_query_terms
4658 LOC: Config.onoff.strip_query_terms
4661 By default, Squid strips query terms from requested URLs before
4662 logging. This protects your user's privacy and reduces log size.
4664 When investigating HIT/MISS or other caching behaviour you
4665 will need to disable this to see the full URL used by Squid.
4672 LOC: Config.onoff.buffered_logs
4674 Whether to write/send access_log records ASAP or accumulate them and
4675 then write/send them in larger chunks. Buffering may improve
4676 performance because it decreases the number of I/Os. However,
4677 buffering increases the delay before log records become available to
4678 the final recipient (e.g., a disk file or logging daemon) and,
4679 hence, increases the risk of log records loss.
4681 Note that even when buffered_logs are off, Squid may have to buffer
4682 records if it cannot write/send them immediately due to pending I/Os
4683 (e.g., the I/O writing the previous log record) or connectivity loss.
4685 Currently honored by 'daemon' and 'tcp' access_log modules only.
4688 NAME: netdb_filename
4690 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4691 LOC: Config.netdbFilename
4694 Where Squid stores it's netdb journal.
4695 When enabled this journal preserves netdb state between restarts.
4697 To disable, enter "none".
4701 OPTIONS FOR TROUBLESHOOTING
4702 -----------------------------------------------------------------------------
4707 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4708 LOC: Debug::cache_log
4710 Squid administrative logging file.
4712 This is where general information about Squid behavior goes. You can
4713 increase the amount of data logged to this file and how often it is
4714 rotated with "debug_options"
4720 DEFAULT_DOC: Log all critical and important messages.
4721 LOC: Debug::debugOptions
4723 Logging options are set as section,level where each source file
4724 is assigned a unique section. Lower levels result in less
4725 output, Full debugging (level 9) can result in a very large
4726 log file, so be careful.
4728 The magic word "ALL" sets debugging levels for all sections.
4729 The default is to run with "ALL,1" to record important warnings.
4731 The rotate=N option can be used to keep more or less of these logs
4732 than would otherwise be kept by logfile_rotate.
4733 For most uses a single log should be enough to monitor current
4734 events affecting Squid.
4739 LOC: Config.coredump_dir
4740 DEFAULT_IF_NONE: none
4741 DEFAULT_DOC: Use the directory from where Squid was started.
4743 By default Squid leaves core files in the directory from where
4744 it was started. If you set 'coredump_dir' to a directory
4745 that exists, Squid will chdir() to that directory at startup
4746 and coredump files will be left there.
4750 # Leave coredumps in the first cache dir
4751 coredump_dir @DEFAULT_SWAP_DIR@
4757 OPTIONS FOR FTP GATEWAYING
4758 -----------------------------------------------------------------------------
4764 LOC: Config.Ftp.anon_user
4766 If you want the anonymous login password to be more informative
4767 (and enable the use of picky FTP servers), set this to something
4768 reasonable for your domain, like wwwuser@somewhere.net
4770 The reason why this is domainless by default is the
4771 request can be made on the behalf of a user in any domain,
4772 depending on how the cache is used.
4773 Some FTP server also validate the email address is valid
4774 (for example perl.com).
4780 LOC: Config.Ftp.passive
4782 If your firewall does not allow Squid to use passive
4783 connections, turn off this option.
4785 Use of ftp_epsv_all option requires this to be ON.
4791 LOC: Config.Ftp.epsv_all
4793 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4795 NATs may be able to put the connection on a "fast path" through the
4796 translator, as the EPRT command will never be used and therefore,
4797 translation of the data portion of the segments will never be needed.
4799 When a client only expects to do two-way FTP transfers this may be
4801 If squid finds that it must do a three-way FTP transfer after issuing
4802 an EPSV ALL command, the FTP session will fail.
4804 If you have any doubts about this option do not use it.
4805 Squid will nicely attempt all other connection methods.
4807 Requires ftp_passive to be ON (default) for any effect.
4813 LOC: Config.accessList.ftp_epsv
4815 FTP Protocol extensions permit the use of a special "EPSV" command.
4817 NATs may be able to put the connection on a "fast path" through the
4818 translator using EPSV, as the EPRT command will never be used
4819 and therefore, translation of the data portion of the segments
4820 will never be needed.
4822 EPSV is often required to interoperate with FTP servers on IPv6
4823 networks. On the other hand, it may break some IPv4 servers.
4825 By default, EPSV may try EPSV with any FTP server. To fine tune
4826 that decision, you may restrict EPSV to certain clients or servers
4829 ftp_epsv allow|deny al1 acl2 ...
4831 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4833 Only fast ACLs are supported.
4834 Requires ftp_passive to be ON (default) for any effect.
4840 LOC: Config.Ftp.eprt
4842 FTP Protocol extensions permit the use of a special "EPRT" command.
4844 This extension provides a protocol neutral alternative to the
4845 IPv4-only PORT command. When supported it enables active FTP data
4846 channels over IPv6 and efficient NAT handling.
4848 Turning this OFF will prevent EPRT being attempted and will skip
4849 straight to using PORT for IPv4 servers.
4851 Some devices are known to not handle this extension correctly and
4852 may result in crashes. Devices which suport EPRT enough to fail
4853 cleanly will result in Squid attempting PORT anyway. This directive
4854 should only be disabled when EPRT results in device failures.
4856 WARNING: Doing so will convert Squid back to the old behavior with all
4857 the related problems with external NAT devices/layers and IPv4-only FTP.
4860 NAME: ftp_sanitycheck
4863 LOC: Config.Ftp.sanitycheck
4865 For security and data integrity reasons Squid by default performs
4866 sanity checks of the addresses of FTP data connections ensure the
4867 data connection is to the requested server. If you need to allow
4868 FTP connections to servers using another IP address for the data
4869 connection turn this off.
4872 NAME: ftp_telnet_protocol
4875 LOC: Config.Ftp.telnet
4877 The FTP protocol is officially defined to use the telnet protocol
4878 as transport channel for the control connection. However, many
4879 implementations are broken and does not respect this aspect of
4882 If you have trouble accessing files with ASCII code 255 in the
4883 path or similar problems involving this ASCII code you can
4884 try setting this directive to off. If that helps, report to the
4885 operator of the FTP server in question that their FTP server
4886 is broken and does not follow the FTP standard.
4890 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4891 -----------------------------------------------------------------------------
4896 DEFAULT: @DEFAULT_DISKD@
4897 LOC: Config.Program.diskd
4899 Specify the location of the diskd executable.
4900 Note this is only useful if you have compiled in
4901 diskd as one of the store io modules.
4904 NAME: unlinkd_program
4907 DEFAULT: @DEFAULT_UNLINKD@
4908 LOC: Config.Program.unlinkd
4910 Specify the location of the executable for file deletion process.
4913 NAME: pinger_program
4916 DEFAULT: @DEFAULT_PINGER@
4919 Specify the location of the executable for the pinger process.
4928 Control whether the pinger is active at run-time.
4929 Enables turning ICMP pinger on and off with a simple
4930 squid -k reconfigure.
4935 OPTIONS FOR URL REWRITING
4936 -----------------------------------------------------------------------------
4939 NAME: url_rewrite_program redirect_program
4941 LOC: Config.Program.redirect
4944 Specify the location of the executable URL rewriter to use.
4945 Since they can perform almost any function there isn't one included.
4947 For each requested URL, the rewriter will receive on line with the format
4949 [channel-ID <SP>] URL [<SP> extras]<NL>
4951 See url_rewrite_extras on how to send "extras" with optional values to
4953 After processing the request the helper must reply using the following format:
4955 [channel-ID <SP>] result [<SP> kv-pairs]
4957 The result code can be:
4959 OK status=30N url="..."
4960 Redirect the URL to the one supplied in 'url='.
4961 'status=' is optional and contains the status code to send
4962 the client in Squids HTTP response. It must be one of the
4963 HTTP redirect status codes: 301, 302, 303, 307, 308.
4964 When no status is given Squid will use 302.
4966 OK rewrite-url="..."
4967 Rewrite the URL to the one supplied in 'rewrite-url='.
4968 The new URL is fetched directly by Squid and returned to
4969 the client as the response to its request.
4972 When neither of url= and rewrite-url= are sent Squid does
4976 Do not change the URL.
4979 An internal error occurred in the helper, preventing
4980 a result being identified. The 'message=' key name is
4981 reserved for delivering a log message.
4984 In addition to the above kv-pairs Squid also understands the following
4985 optional kv-pairs received from URL rewriters:
4987 Associates a TAG with the client TCP connection.
4988 The TAG is treated as a regular annotation but persists across
4989 future requests on the client connection rather than just the
4990 current request. A helper may update the TAG during subsequent
4991 requests be returning a new kv-pair.
4993 When using the concurrency= option the protocol is changed by
4994 introducing a query channel tag in front of the request/response.
4995 The query channel tag is a number between 0 and concurrency-1.
4996 This value must be echoed back unchanged to Squid as the first part
4997 of the response relating to its request.
4999 WARNING: URL re-writing ability should be avoided whenever possible.
5000 Use the URL redirect form of response instead.
5002 Re-write creates a difference in the state held by the client
5003 and server. Possibly causing confusion when the server response
5004 contains snippets of its view state. Embeded URLs, response
5005 and content Location headers, etc. are not re-written by this
5008 By default, a URL rewriter is not used.
5011 NAME: url_rewrite_children redirect_children
5012 TYPE: HelperChildConfig
5013 DEFAULT: 20 startup=0 idle=1 concurrency=0
5014 LOC: Config.redirectChildren
5016 The maximum number of redirector processes to spawn. If you limit
5017 it too few Squid will have to wait for them to process a backlog of
5018 URLs, slowing it down. If you allow too many they will use RAM
5019 and other system resources noticably.
5021 The startup= and idle= options allow some measure of skew in your
5026 Sets a minimum of how many processes are to be spawned when Squid
5027 starts or reconfigures. When set to zero the first request will
5028 cause spawning of the first child process to handle it.
5030 Starting too few will cause an initial slowdown in traffic as Squid
5031 attempts to simultaneously spawn enough processes to cope.
5035 Sets a minimum of how many processes Squid is to try and keep available
5036 at all times. When traffic begins to rise above what the existing
5037 processes can handle this many more will be spawned up to the maximum
5038 configured. A minimum setting of 1 is required.
5042 The number of requests each redirector helper can handle in
5043 parallel. Defaults to 0 which indicates the redirector
5044 is a old-style single threaded redirector.
5046 When this directive is set to a value >= 1 then the protocol
5047 used to communicate with the helper is modified to include
5048 an ID in front of the request/response. The ID from the request
5049 must be echoed back with the response to that request.
5053 Sets the maximum number of queued requests.
5054 If the queued requests exceed queue size and redirector_bypass
5055 configuration option is set, then redirector is bypassed. Otherwise, if
5056 overloading persists squid may abort its operation.
5057 The default value is set to 2*numberofchildren.
5060 NAME: url_rewrite_host_header redirect_rewrites_host_header
5063 LOC: Config.onoff.redir_rewrites_host
5065 To preserve same-origin security policies in browsers and
5066 prevent Host: header forgery by redirectors Squid rewrites
5067 any Host: header in redirected requests.
5069 If you are running an accelerator this may not be a wanted
5070 effect of a redirector. This directive enables you disable
5071 Host: alteration in reverse-proxy traffic.
5073 WARNING: Entries are cached on the result of the URL rewriting
5074 process, so be careful if you have domain-virtual hosts.
5076 WARNING: Squid and other software verifies the URL and Host
5077 are matching, so be careful not to relay through other proxies
5078 or inspecting firewalls with this disabled.
5081 NAME: url_rewrite_access redirector_access
5084 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5085 LOC: Config.accessList.redirector
5087 If defined, this access list specifies which requests are
5088 sent to the redirector processes.
5090 This clause supports both fast and slow acl types.
5091 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5094 NAME: url_rewrite_bypass redirector_bypass
5096 LOC: Config.onoff.redirector_bypass
5099 When this is 'on', a request will not go through the
5100 redirector if all the helpers are busy. If this is 'off'
5101 and the redirector queue grows too large, Squid will exit
5102 with a FATAL error and ask you to increase the number of
5103 redirectors. You should only enable this if the redirectors
5104 are not critical to your caching system. If you use
5105 redirectors for access control, and you enable this option,
5106 users may have access to pages they should not
5107 be allowed to request.
5108 This options sets default queue-size option of the url_rewrite_children
5112 NAME: url_rewrite_extras
5113 TYPE: TokenOrQuotedString
5114 LOC: Config.redirector_extras
5115 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5117 Specifies a string to be append to request line format for the
5118 rewriter helper. "Quoted" format values may contain spaces and
5119 logformat %macros. In theory, any logformat %macro can be used.
5120 In practice, a %macro expands as a dash (-) if the helper request is
5121 sent before the required macro information is available to Squid.
5124 NAME: url_rewrite_timeout
5125 TYPE: UrlHelperTimeout
5126 LOC: Config.onUrlRewriteTimeout
5128 DEFAULT_DOC: Squid waits for the helper response forever
5130 Squid times active requests to redirector. The timeout value and Squid
5131 reaction to a timed out request are configurable using the following
5134 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
5136 supported timeout actions:
5137 fail Squid return a ERR_GATEWAY_FAILURE error page
5139 bypass Do not re-write the URL
5141 retry Send the lookup to the helper again
5143 use_configured_response
5144 Use the <quoted-response> as helper response
5148 OPTIONS FOR STORE ID
5149 -----------------------------------------------------------------------------
5152 NAME: store_id_program storeurl_rewrite_program
5154 LOC: Config.Program.store_id
5157 Specify the location of the executable StoreID helper to use.
5158 Since they can perform almost any function there isn't one included.
5160 For each requested URL, the helper will receive one line with the format
5162 [channel-ID <SP>] URL [<SP> extras]<NL>
5165 After processing the request the helper must reply using the following format:
5167 [channel-ID <SP>] result [<SP> kv-pairs]
5169 The result code can be:
5172 Use the StoreID supplied in 'store-id='.
5175 The default is to use HTTP request URL as the store ID.
5178 An internal error occured in the helper, preventing
5179 a result being identified.
5181 In addition to the above kv-pairs Squid also understands the following
5182 optional kv-pairs received from URL rewriters:
5184 Associates a TAG with the client TCP connection.
5185 Please see url_rewrite_program related documentation for this
5188 Helper programs should be prepared to receive and possibly ignore
5189 additional whitespace-separated tokens on each input line.
5191 When using the concurrency= option the protocol is changed by
5192 introducing a query channel tag in front of the request/response.
5193 The query channel tag is a number between 0 and concurrency-1.
5194 This value must be echoed back unchanged to Squid as the first part
5195 of the response relating to its request.
5197 NOTE: when using StoreID refresh_pattern will apply to the StoreID
5198 returned from the helper and not the URL.
5200 WARNING: Wrong StoreID value returned by a careless helper may result
5201 in the wrong cached response returned to the user.
5203 By default, a StoreID helper is not used.
5206 NAME: store_id_extras
5207 TYPE: TokenOrQuotedString
5208 LOC: Config.storeId_extras
5209 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5211 Specifies a string to be append to request line format for the
5212 StoreId helper. "Quoted" format values may contain spaces and
5213 logformat %macros. In theory, any logformat %macro can be used.
5214 In practice, a %macro expands as a dash (-) if the helper request is
5215 sent before the required macro information is available to Squid.
5218 NAME: store_id_children storeurl_rewrite_children
5219 TYPE: HelperChildConfig
5220 DEFAULT: 20 startup=0 idle=1 concurrency=0
5221 LOC: Config.storeIdChildren
5223 The maximum number of StoreID helper processes to spawn. If you limit
5224 it too few Squid will have to wait for them to process a backlog of
5225 requests, slowing it down. If you allow too many they will use RAM
5226 and other system resources noticably.
5228 The startup= and idle= options allow some measure of skew in your
5233 Sets a minimum of how many processes are to be spawned when Squid
5234 starts or reconfigures. When set to zero the first request will
5235 cause spawning of the first child process to handle it.
5237 Starting too few will cause an initial slowdown in traffic as Squid
5238 attempts to simultaneously spawn enough processes to cope.
5242 Sets a minimum of how many processes Squid is to try and keep available
5243 at all times. When traffic begins to rise above what the existing
5244 processes can handle this many more will be spawned up to the maximum
5245 configured. A minimum setting of 1 is required.
5249 The number of requests each storeID helper can handle in
5250 parallel. Defaults to 0 which indicates the helper
5251 is a old-style single threaded program.
5253 When this directive is set to a value >= 1 then the protocol
5254 used to communicate with the helper is modified to include
5255 an ID in front of the request/response. The ID from the request
5256 must be echoed back with the response to that request.
5260 Sets the maximum number of queued requests.
5261 If the queued requests exceed queue size and store_id_bypass
5262 configuration option is set, then storeID helper is bypassed. Otherwise,
5263 if overloading persists squid may abort its operation.
5264 The default value is set to 2*numberofchildren.
5267 NAME: store_id_access storeurl_rewrite_access
5270 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5271 LOC: Config.accessList.store_id
5273 If defined, this access list specifies which requests are
5274 sent to the StoreID processes. By default all requests
5277 This clause supports both fast and slow acl types.
5278 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5281 NAME: store_id_bypass storeurl_rewrite_bypass
5283 LOC: Config.onoff.store_id_bypass
5286 When this is 'on', a request will not go through the
5287 helper if all helpers are busy. If this is 'off'
5288 and the helper queue grows too large, Squid will exit
5289 with a FATAL error and ask you to increase the number of
5290 helpers. You should only enable this if the helperss
5291 are not critical to your caching system. If you use
5292 helpers for critical caching components, and you enable this
5293 option, users may not get objects from cache.
5294 This options sets default queue-size option of the store_id_children
5299 OPTIONS FOR TUNING THE CACHE
5300 -----------------------------------------------------------------------------
5303 NAME: cache no_cache
5306 DEFAULT_DOC: By default, this directive is unused and has no effect.
5307 LOC: Config.accessList.noCache
5309 Requests denied by this directive will not be served from the cache
5310 and their responses will not be stored in the cache. This directive
5311 has no effect on other transactions and on already cached responses.
5313 This clause supports both fast and slow acl types.
5314 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5316 This and the two other similar caching directives listed below are
5317 checked at different transaction processing stages, have different
5318 access to response information, affect different cache operations,
5319 and differ in slow ACLs support:
5321 * cache: Checked before Squid makes a hit/miss determination.
5322 No access to reply information!
5323 Denies both serving a hit and storing a miss.
5324 Supports both fast and slow ACLs.
5325 * send_hit: Checked after a hit was detected.
5326 Has access to reply (hit) information.
5327 Denies serving a hit only.
5328 Supports fast ACLs only.
5329 * store_miss: Checked before storing a cachable miss.
5330 Has access to reply (miss) information.
5331 Denies storing a miss only.
5332 Supports fast ACLs only.
5334 If you are not sure which of the three directives to use, apply the
5335 following decision logic:
5337 * If your ACL(s) are of slow type _and_ need response info, redesign.
5338 Squid does not support that particular combination at this time.
5340 * If your directive ACL(s) are of slow type, use "cache"; and/or
5341 * if your directive ACL(s) need no response info, use "cache".
5343 * If you do not want the response cached, use store_miss; and/or
5344 * if you do not want a hit on a cached response, use send_hit.
5350 DEFAULT_DOC: By default, this directive is unused and has no effect.
5351 LOC: Config.accessList.sendHit
5353 Responses denied by this directive will not be served from the cache
5354 (but may still be cached, see store_miss). This directive has no
5355 effect on the responses it allows and on the cached objects.
5357 Please see the "cache" directive for a summary of differences among
5358 store_miss, send_hit, and cache directives.
5360 Unlike the "cache" directive, send_hit only supports fast acl
5361 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5365 # apply custom Store ID mapping to some URLs
5366 acl MapMe dstdomain .c.example.com
5367 store_id_program ...
5368 store_id_access allow MapMe
5370 # but prevent caching of special responses
5371 # such as 302 redirects that cause StoreID loops
5372 acl Ordinary http_status 200-299
5373 store_miss deny MapMe !Ordinary
5375 # and do not serve any previously stored special responses
5376 # from the cache (in case they were already cached before
5377 # the above store_miss rule was in effect).
5378 send_hit deny MapMe !Ordinary
5384 DEFAULT_DOC: By default, this directive is unused and has no effect.
5385 LOC: Config.accessList.storeMiss
5387 Responses denied by this directive will not be cached (but may still
5388 be served from the cache, see send_hit). This directive has no
5389 effect on the responses it allows and on the already cached responses.
5391 Please see the "cache" directive for a summary of differences among
5392 store_miss, send_hit, and cache directives. See the
5393 send_hit directive for a usage example.
5395 Unlike the "cache" directive, store_miss only supports fast acl
5396 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5402 LOC: Config.maxStale
5405 This option puts an upper limit on how stale content Squid
5406 will serve from the cache if cache validation fails.
5407 Can be overriden by the refresh_pattern max-stale option.
5410 NAME: refresh_pattern
5411 TYPE: refreshpattern
5415 usage: refresh_pattern [-i] regex min percent max [options]
5417 By default, regular expressions are CASE-SENSITIVE. To make
5418 them case-insensitive, use the -i option.
5420 'Min' is the time (in minutes) an object without an explicit
5421 expiry time should be considered fresh. The recommended
5422 value is 0, any higher values may cause dynamic applications
5423 to be erroneously cached unless the application designer
5424 has taken the appropriate actions.
5426 'Percent' is a percentage of the objects age (time since last
5427 modification age) an object without explicit expiry time
5428 will be considered fresh.
5430 'Max' is an upper limit on how long objects without an explicit
5431 expiry time will be considered fresh.
5433 options: override-expire
5438 ignore-must-revalidate
5445 override-expire enforces min age even if the server
5446 sent an explicit expiry time (e.g., with the
5447 Expires: header or Cache-Control: max-age). Doing this
5448 VIOLATES the HTTP standard. Enabling this feature
5449 could make you liable for problems which it causes.
5451 Note: override-expire does not enforce staleness - it only extends
5452 freshness / min. If the server returns a Expires time which
5453 is longer than your max time, Squid will still consider
5454 the object fresh for that period of time.
5456 override-lastmod enforces min age even on objects
5457 that were modified recently.
5459 reload-into-ims changes a client no-cache or ``reload''
5460 request for a cached entry into a conditional request using
5461 If-Modified-Since and/or If-None-Match headers, provided the
5462 cached entry has a Last-Modified and/or a strong ETag header.
5463 Doing this VIOLATES the HTTP standard. Enabling this feature
5464 could make you liable for problems which it causes.
5466 ignore-reload ignores a client no-cache or ``reload''
5467 header. Doing this VIOLATES the HTTP standard. Enabling
5468 this feature could make you liable for problems which
5471 ignore-no-store ignores any ``Cache-control: no-store''
5472 headers received from a server. Doing this VIOLATES
5473 the HTTP standard. Enabling this feature could make you
5474 liable for problems which it causes.
5476 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
5477 headers received from a server. Doing this VIOLATES
5478 the HTTP standard. Enabling this feature could make you
5479 liable for problems which it causes.
5481 ignore-private ignores any ``Cache-control: private''
5482 headers received from a server. Doing this VIOLATES
5483 the HTTP standard. Enabling this feature could make you
5484 liable for problems which it causes.
5486 ignore-auth caches responses to requests with authorization,
5487 as if the originserver had sent ``Cache-control: public''
5488 in the response header. Doing this VIOLATES the HTTP standard.
5489 Enabling this feature could make you liable for problems which
5492 refresh-ims causes squid to contact the origin server
5493 when a client issues an If-Modified-Since request. This
5494 ensures that the client will receive an updated version
5495 if one is available.
5497 store-stale stores responses even if they don't have explicit
5498 freshness or a validator (i.e., Last-Modified or an ETag)
5499 present, or if they're already stale. By default, Squid will
5500 not cache such responses because they usually can't be
5501 reused. Note that such responses will be stale by default.
5503 max-stale=NN provide a maximum staleness factor. Squid won't
5504 serve objects more stale than this even if it failed to
5505 validate the object. Default: use the max_stale global limit.
5507 Basically a cached object is:
5509 FRESH if expires < now, else STALE
5511 FRESH if lm-factor < percent, else STALE
5515 The refresh_pattern lines are checked in the order listed here.
5516 The first entry which matches is used. If none of the entries
5517 match the default will be used.
5519 Note, you must uncomment all the default lines if you want
5520 to change one. The default setting is only active if none is
5526 # Add any of your own refresh_pattern entries above these.
5528 refresh_pattern ^ftp: 1440 20% 10080
5529 refresh_pattern ^gopher: 1440 0% 1440
5530 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5531 refresh_pattern . 0 20% 4320
5535 NAME: quick_abort_min
5539 LOC: Config.quickAbort.min
5542 NAME: quick_abort_max
5546 LOC: Config.quickAbort.max
5549 NAME: quick_abort_pct
5553 LOC: Config.quickAbort.pct
5555 The cache by default continues downloading aborted requests
5556 which are almost completed (less than 16 KB remaining). This
5557 may be undesirable on slow (e.g. SLIP) links and/or very busy
5558 caches. Impatient users may tie up file descriptors and
5559 bandwidth by repeatedly requesting and immediately aborting
5562 When the user aborts a request, Squid will check the
5563 quick_abort values to the amount of data transferred until
5566 If the transfer has less than 'quick_abort_min' KB remaining,
5567 it will finish the retrieval.
5569 If the transfer has more than 'quick_abort_max' KB remaining,
5570 it will abort the retrieval.
5572 If more than 'quick_abort_pct' of the transfer has completed,
5573 it will finish the retrieval.
5575 If you do not want any retrieval to continue after the client
5576 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5579 If you want retrievals to always continue if they are being
5580 cached set 'quick_abort_min' to '-1 KB'.
5583 NAME: read_ahead_gap
5584 COMMENT: buffer-size
5586 LOC: Config.readAheadGap
5589 The amount of data the cache will buffer ahead of what has been
5590 sent to the client when retrieving an object from another server.
5594 IFDEF: USE_HTTP_VIOLATIONS
5597 LOC: Config.negativeTtl
5600 Set the Default Time-to-Live (TTL) for failed requests.
5601 Certain types of failures (such as "connection refused" and
5602 "404 Not Found") are able to be negatively-cached for a short time.
5603 Modern web servers should provide Expires: header, however if they
5604 do not this can provide a minimum TTL.
5605 The default is not to cache errors with unknown expiry details.
5607 Note that this is different from negative caching of DNS lookups.
5609 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5610 this feature could make you liable for problems which it
5614 NAME: positive_dns_ttl
5617 LOC: Config.positiveDnsTtl
5620 Upper limit on how long Squid will cache positive DNS responses.
5621 Default is 6 hours (360 minutes). This directive must be set
5622 larger than negative_dns_ttl.
5625 NAME: negative_dns_ttl
5628 LOC: Config.negativeDnsTtl
5631 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5632 This also sets the lower cache limit on positive lookups.
5633 Minimum value is 1 second, and it is not recommendable to go
5634 much below 10 seconds.
5637 NAME: range_offset_limit
5638 COMMENT: size [acl acl...]
5640 LOC: Config.rangeOffsetLimit
5643 usage: (size) [units] [[!]aclname]
5645 Sets an upper limit on how far (number of bytes) into the file
5646 a Range request may be to cause Squid to prefetch the whole file.
5647 If beyond this limit, Squid forwards the Range request as it is and
5648 the result is NOT cached.
5650 This is to stop a far ahead range request (lets say start at 17MB)
5651 from making Squid fetch the whole object up to that point before
5652 sending anything to the client.
5654 Multiple range_offset_limit lines may be specified, and they will
5655 be searched from top to bottom on each request until a match is found.
5656 The first match found will be used. If no line matches a request, the
5657 default limit of 0 bytes will be used.
5659 'size' is the limit specified as a number of units.
5661 'units' specifies whether to use bytes, KB, MB, etc.
5662 If no units are specified bytes are assumed.
5664 A size of 0 causes Squid to never fetch more than the
5665 client requested. (default)
5667 A size of 'none' causes Squid to always fetch the object from the
5668 beginning so it may cache the result. (2.0 style)
5670 'aclname' is the name of a defined ACL.
5672 NP: Using 'none' as the byte value here will override any quick_abort settings
5673 that may otherwise apply to the range request. The range request will
5674 be fully fetched from start to finish regardless of the client
5675 actions. This affects bandwidth usage.
5678 NAME: minimum_expiry_time
5681 LOC: Config.minimum_expiry_time
5684 The minimum caching time according to (Expires - Date)
5685 headers Squid honors if the object can't be revalidated.
5686 The default is 60 seconds.
5688 In reverse proxy environments it might be desirable to honor
5689 shorter object lifetimes. It is most likely better to make
5690 your server return a meaningful Last-Modified header however.
5692 In ESI environments where page fragments often have short
5693 lifetimes, this will often be best set to 0.
5696 NAME: store_avg_object_size
5700 LOC: Config.Store.avgObjectSize
5702 Average object size, used to estimate number of objects your
5703 cache can hold. The default is 13 KB.
5705 This is used to pre-seed the cache index memory allocation to
5706 reduce expensive reallocate operations while handling clients
5707 traffic. Too-large values may result in memory allocation during
5708 peak traffic, too-small values will result in wasted memory.
5710 Check the cache manager 'info' report metrics for the real
5711 object sizes seen by your Squid before tuning this.
5714 NAME: store_objects_per_bucket
5717 LOC: Config.Store.objectsPerBucket
5719 Target number of objects per bucket in the store hash table.
5720 Lowering this value increases the total number of buckets and
5721 also the storage maintenance rate. The default is 20.
5726 -----------------------------------------------------------------------------
5729 NAME: request_header_max_size
5733 LOC: Config.maxRequestHeaderSize
5735 This specifies the maximum size for HTTP headers in a request.
5736 Request headers are usually relatively small (about 512 bytes).
5737 Placing a limit on the request header size will catch certain
5738 bugs (for example with persistent connections) and possibly
5739 buffer-overflow or denial-of-service attacks.
5742 NAME: reply_header_max_size
5746 LOC: Config.maxReplyHeaderSize
5748 This specifies the maximum size for HTTP headers in a reply.
5749 Reply headers are usually relatively small (about 512 bytes).
5750 Placing a limit on the reply header size will catch certain
5751 bugs (for example with persistent connections) and possibly
5752 buffer-overflow or denial-of-service attacks.
5755 NAME: request_body_max_size
5759 DEFAULT_DOC: No limit.
5760 LOC: Config.maxRequestBodySize
5762 This specifies the maximum size for an HTTP request body.
5763 In other words, the maximum size of a PUT/POST request.
5764 A user who attempts to send a request with a body larger
5765 than this limit receives an "Invalid Request" error message.
5766 If you set this parameter to a zero (the default), there will
5767 be no limit imposed.
5769 See also client_request_buffer_max_size for an alternative
5770 limitation on client uploads which can be configured.
5773 NAME: client_request_buffer_max_size
5777 LOC: Config.maxRequestBufferSize
5779 This specifies the maximum buffer size of a client request.
5780 It prevents squid eating too much memory when somebody uploads
5785 IFDEF: USE_HTTP_VIOLATIONS
5788 DEFAULT_DOC: Obey RFC 2616.
5789 LOC: Config.accessList.brokenPosts
5791 A list of ACL elements which, if matched, causes Squid to send
5792 an extra CRLF pair after the body of a PUT/POST request.
5794 Some HTTP servers has broken implementations of PUT/POST,
5795 and rely on an extra CRLF pair sent by some WWW clients.
5797 Quote from RFC2616 section 4.1 on this matter:
5799 Note: certain buggy HTTP/1.0 client implementations generate an
5800 extra CRLF's after a POST request. To restate what is explicitly
5801 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5802 a request with an extra CRLF.
5804 This clause only supports fast acl types.
5805 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5808 acl buggy_server url_regex ^http://....
5809 broken_posts allow buggy_server
5812 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5815 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5817 LOC: Adaptation::Config::use_indirect_client
5819 Controls whether the indirect client IP address (instead of the direct
5820 client IP address) is passed to adaptation services.
5822 See also: follow_x_forwarded_for adaptation_send_client_ip
5826 IFDEF: USE_HTTP_VIOLATIONS
5830 LOC: Config.onoff.via
5832 If set (default), Squid will include a Via header in requests and
5833 replies as required by RFC2616.
5839 LOC: Config.onoff.ie_refresh
5842 Microsoft Internet Explorer up until version 5.5 Service
5843 Pack 1 has an issue with transparent proxies, wherein it
5844 is impossible to force a refresh. Turning this on provides
5845 a partial fix to the problem, by causing all IMS-REFRESH
5846 requests from older IE versions to check the origin server
5847 for fresh content. This reduces hit ratio by some amount
5848 (~10% in my experience), but allows users to actually get
5849 fresh content when they want it. Note because Squid
5850 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5851 of 5.5 is unchanged from old versions of Squid (i.e. a
5852 forced refresh is impossible). Newer versions of IE will,
5853 hopefully, continue to have the new behavior and will be
5854 handled based on that assumption. This option defaults to
5855 the old Squid behavior, which is better for hit ratios but
5856 worse for clients using IE, if they need to be able to
5857 force fresh content.
5860 NAME: vary_ignore_expire
5863 LOC: Config.onoff.vary_ignore_expire
5866 Many HTTP servers supporting Vary gives such objects
5867 immediate expiry time with no cache-control header
5868 when requested by a HTTP/1.0 client. This option
5869 enables Squid to ignore such expiry times until
5870 HTTP/1.1 is fully implemented.
5872 WARNING: If turned on this may eventually cause some
5873 varying objects not intended for caching to get cached.
5876 NAME: request_entities
5878 LOC: Config.onoff.request_entities
5881 Squid defaults to deny GET and HEAD requests with request entities,
5882 as the meaning of such requests are undefined in the HTTP standard
5883 even if not explicitly forbidden.
5885 Set this directive to on if you have clients which insists
5886 on sending request entities in GET or HEAD requests. But be warned
5887 that there is server software (both proxies and web servers) which
5888 can fail to properly process this kind of request which may make you
5889 vulnerable to cache pollution attacks if enabled.
5892 NAME: request_header_access
5893 IFDEF: USE_HTTP_VIOLATIONS
5894 TYPE: http_header_access
5895 LOC: Config.request_header_access
5897 DEFAULT_DOC: No limits.
5899 Usage: request_header_access header_name allow|deny [!]aclname ...
5901 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5902 this feature could make you liable for problems which it
5905 This option replaces the old 'anonymize_headers' and the
5906 older 'http_anonymizer' option with something that is much
5907 more configurable. A list of ACLs for each header name allows
5908 removal of specific header fields under specific conditions.
5910 This option only applies to outgoing HTTP request headers (i.e.,
5911 headers sent by Squid to the next HTTP hop such as a cache peer
5912 or an origin server). The option has no effect during cache hit
5913 detection. The equivalent adaptation vectoring point in ICAP
5914 terminology is post-cache REQMOD.
5916 The option is applied to individual outgoing request header
5917 fields. For each request header field F, Squid uses the first
5918 qualifying sets of request_header_access rules:
5920 1. Rules with header_name equal to F's name.
5921 2. Rules with header_name 'Other', provided F's name is not
5922 on the hard-coded list of commonly used HTTP header names.
5923 3. Rules with header_name 'All'.
5925 Within that qualifying rule set, rule ACLs are checked as usual.
5926 If ACLs of an "allow" rule match, the header field is allowed to
5927 go through as is. If ACLs of a "deny" rule match, the header is
5928 removed and request_header_replace is then checked to identify
5929 if the removed header has a replacement. If no rules within the
5930 set have matching ACLs, the header field is left as is.
5932 For example, to achieve the same behavior as the old
5933 'http_anonymizer standard' option, you should use:
5935 request_header_access From deny all
5936 request_header_access Referer deny all
5937 request_header_access User-Agent deny all
5939 Or, to reproduce the old 'http_anonymizer paranoid' feature
5942 request_header_access Authorization allow all
5943 request_header_access Proxy-Authorization allow all
5944 request_header_access Cache-Control allow all
5945 request_header_access Content-Length allow all
5946 request_header_access Content-Type allow all
5947 request_header_access Date allow all
5948 request_header_access Host allow all
5949 request_header_access If-Modified-Since allow all
5950 request_header_access Pragma allow all
5951 request_header_access Accept allow all
5952 request_header_access Accept-Charset allow all
5953 request_header_access Accept-Encoding allow all
5954 request_header_access Accept-Language allow all
5955 request_header_access Connection allow all
5956 request_header_access All deny all
5958 HTTP reply headers are controlled with the reply_header_access directive.
5960 By default, all headers are allowed (no anonymizing is performed).
5963 NAME: reply_header_access
5964 IFDEF: USE_HTTP_VIOLATIONS
5965 TYPE: http_header_access
5966 LOC: Config.reply_header_access
5968 DEFAULT_DOC: No limits.
5970 Usage: reply_header_access header_name allow|deny [!]aclname ...
5972 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5973 this feature could make you liable for problems which it
5976 This option only applies to reply headers, i.e., from the
5977 server to the client.
5979 This is the same as request_header_access, but in the other
5980 direction. Please see request_header_access for detailed
5983 For example, to achieve the same behavior as the old
5984 'http_anonymizer standard' option, you should use:
5986 reply_header_access Server deny all
5987 reply_header_access WWW-Authenticate deny all
5988 reply_header_access Link deny all
5990 Or, to reproduce the old 'http_anonymizer paranoid' feature
5993 reply_header_access Allow allow all
5994 reply_header_access WWW-Authenticate allow all
5995 reply_header_access Proxy-Authenticate allow all
5996 reply_header_access Cache-Control allow all
5997 reply_header_access Content-Encoding allow all
5998 reply_header_access Content-Length allow all
5999 reply_header_access Content-Type allow all
6000 reply_header_access Date allow all
6001 reply_header_access Expires allow all
6002 reply_header_access Last-Modified allow all
6003 reply_header_access Location allow all
6004 reply_header_access Pragma allow all
6005 reply_header_access Content-Language allow all
6006 reply_header_access Retry-After allow all
6007 reply_header_access Title allow all
6008 reply_header_access Content-Disposition allow all
6009 reply_header_access Connection allow all
6010 reply_header_access All deny all
6012 HTTP request headers are controlled with the request_header_access directive.
6014 By default, all headers are allowed (no anonymizing is
6018 NAME: request_header_replace header_replace
6019 IFDEF: USE_HTTP_VIOLATIONS
6020 TYPE: http_header_replace
6021 LOC: Config.request_header_access
6024 Usage: request_header_replace header_name message
6025 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
6027 This option allows you to change the contents of headers
6028 denied with request_header_access above, by replacing them
6029 with some fixed string.
6031 This only applies to request headers, not reply headers.
6033 By default, headers are removed if denied.
6036 NAME: reply_header_replace
6037 IFDEF: USE_HTTP_VIOLATIONS
6038 TYPE: http_header_replace
6039 LOC: Config.reply_header_access
6042 Usage: reply_header_replace header_name message
6043 Example: reply_header_replace Server Foo/1.0
6045 This option allows you to change the contents of headers
6046 denied with reply_header_access above, by replacing them
6047 with some fixed string.
6049 This only applies to reply headers, not request headers.
6051 By default, headers are removed if denied.
6054 NAME: request_header_add
6055 TYPE: HeaderWithAclList
6056 LOC: Config.request_header_add
6059 Usage: request_header_add field-name field-value acl1 [acl2] ...
6060 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
6062 This option adds header fields to outgoing HTTP requests (i.e.,
6063 request headers sent by Squid to the next HTTP hop such as a
6064 cache peer or an origin server). The option has no effect during
6065 cache hit detection. The equivalent adaptation vectoring point
6066 in ICAP terminology is post-cache REQMOD.
6068 Field-name is a token specifying an HTTP header name. If a
6069 standard HTTP header name is used, Squid does not check whether
6070 the new header conflicts with any existing headers or violates
6071 HTTP rules. If the request to be modified already contains a
6072 field with the same name, the old field is preserved but the
6073 header field values are not merged.
6075 Field-value is either a token or a quoted string. If quoted
6076 string format is used, then the surrounding quotes are removed
6077 while escape sequences and %macros are processed.
6079 In theory, all of the logformat codes can be used as %macros.
6080 However, unlike logging (which happens at the very end of
6081 transaction lifetime), the transaction may not yet have enough
6082 information to expand a macro when the new header value is needed.
6083 And some information may already be available to Squid but not yet
6084 committed where the macro expansion code can access it (report
6085 such instances!). The macro will be expanded into a single dash
6086 ('-') in such cases. Not all macros have been tested.
6088 One or more Squid ACLs may be specified to restrict header
6089 injection to matching requests. As always in squid.conf, all
6090 ACLs in an option ACL list must be satisfied for the insertion
6091 to happen. The request_header_add option supports fast ACLs
6100 This option used to log custom information about the master
6101 transaction. For example, an admin may configure Squid to log
6102 which "user group" the transaction belongs to, where "user group"
6103 will be determined based on a set of ACLs and not [just]
6104 authentication information.
6105 Values of key/value pairs can be logged using %{key}note macros:
6107 note key value acl ...
6108 logformat myFormat ... %{key}note ...
6111 NAME: relaxed_header_parser
6112 COMMENT: on|off|warn
6114 LOC: Config.onoff.relaxed_header_parser
6117 In the default "on" setting Squid accepts certain forms
6118 of non-compliant HTTP messages where it is unambiguous
6119 what the sending application intended even if the message
6120 is not correctly formatted. The messages is then normalized
6121 to the correct form when forwarded by Squid.
6123 If set to "warn" then a warning will be emitted in cache.log
6124 each time such HTTP error is encountered.
6126 If set to "off" then such HTTP errors will cause the request
6127 or response to be rejected.
6130 NAME: collapsed_forwarding
6133 LOC: Config.onoff.collapsed_forwarding
6136 This option controls whether Squid is allowed to merge multiple
6137 potentially cachable requests for the same URI before Squid knows
6138 whether the response is going to be cachable.
6140 This feature is disabled by default: Enabling collapsed forwarding
6141 needlessly delays forwarding requests that look cachable (when they are
6142 collapsed) but then need to be forwarded individually anyway because
6143 they end up being for uncachable content. However, in some cases, such
6144 as accelleration of highly cachable content with periodic or groupped
6145 expiration times, the gains from collapsing [large volumes of
6146 simultenous refresh requests] outweigh losses from such delays.
6151 -----------------------------------------------------------------------------
6154 NAME: forward_timeout
6157 LOC: Config.Timeout.forward
6160 This parameter specifies how long Squid should at most attempt in
6161 finding a forwarding path for the request before giving up.
6164 NAME: connect_timeout
6167 LOC: Config.Timeout.connect
6170 This parameter specifies how long to wait for the TCP connect to
6171 the requested server or peer to complete before Squid should
6172 attempt to find another path where to forward the request.
6175 NAME: peer_connect_timeout
6178 LOC: Config.Timeout.peer_connect
6181 This parameter specifies how long to wait for a pending TCP
6182 connection to a peer cache. The default is 30 seconds. You
6183 may also set different timeout values for individual neighbors
6184 with the 'connect-timeout' option on a 'cache_peer' line.
6190 LOC: Config.Timeout.read
6193 Applied on peer server connections.
6195 After each successful read(), the timeout will be extended by this
6196 amount. If no data is read again after this amount of time,
6197 the request is aborted and logged with ERR_READ_TIMEOUT.
6199 The default is 15 minutes.
6205 LOC: Config.Timeout.write
6208 This timeout is tracked for all connections that have data
6209 available for writing and are waiting for the socket to become
6210 ready. After each successful write, the timeout is extended by
6211 the configured amount. If Squid has data to write but the
6212 connection is not ready for the configured duration, the
6213 transaction associated with the connection is terminated. The
6214 default is 15 minutes.
6217 NAME: request_timeout
6219 LOC: Config.Timeout.request
6222 How long to wait for complete HTTP request headers after initial
6223 connection establishment.
6226 NAME: request_start_timeout
6228 LOC: Config.Timeout.request_start_timeout
6231 How long to wait for the first request byte after initial
6232 connection establishment.
6235 NAME: client_idle_pconn_timeout persistent_request_timeout
6237 LOC: Config.Timeout.clientIdlePconn
6240 How long to wait for the next HTTP request on a persistent
6241 client connection after the previous request completes.
6244 NAME: ftp_client_idle_timeout
6246 LOC: Config.Timeout.ftpClientIdle
6249 How long to wait for an FTP request on a connection to Squid ftp_port.
6250 Many FTP clients do not deal with idle connection closures well,
6251 necessitating a longer default timeout than client_idle_pconn_timeout
6252 used for incoming HTTP requests.
6255 NAME: client_lifetime
6258 LOC: Config.Timeout.lifetime
6261 The maximum amount of time a client (browser) is allowed to
6262 remain connected to the cache process. This protects the Cache
6263 from having a lot of sockets (and hence file descriptors) tied up
6264 in a CLOSE_WAIT state from remote clients that go away without
6265 properly shutting down (either because of a network failure or
6266 because of a poor client implementation). The default is one
6269 NOTE: The default value is intended to be much larger than any
6270 client would ever need to be connected to your cache. You
6271 should probably change client_lifetime only as a last resort.
6272 If you seem to have many client connections tying up
6273 filedescriptors, we recommend first tuning the read_timeout,
6274 request_timeout, persistent_request_timeout and quick_abort values.
6277 NAME: pconn_lifetime
6280 LOC: Config.Timeout.pconnLifetime
6283 Desired maximum lifetime of a persistent connection.
6284 When set, Squid will close a now-idle persistent connection that
6285 exceeded configured lifetime instead of moving the connection into
6286 the idle connection pool (or equivalent). No effect on ongoing/active
6287 transactions. Connection lifetime is the time period from the
6288 connection acceptance or opening time until "now".
6290 This limit is useful in environments with long-lived connections
6291 where Squid configuration or environmental factors change during a
6292 single connection lifetime. If unrestricted, some connections may
6293 last for hours and even days, ignoring those changes that should
6294 have affected their behavior or their existence.
6296 Currently, a new lifetime value supplied via Squid reconfiguration
6297 has no effect on already idle connections unless they become busy.
6299 When set to '0' this limit is not used.
6302 NAME: half_closed_clients
6304 LOC: Config.onoff.half_closed_clients
6307 Some clients may shutdown the sending side of their TCP
6308 connections, while leaving their receiving sides open. Sometimes,
6309 Squid can not tell the difference between a half-closed and a
6310 fully-closed TCP connection.
6312 By default, Squid will immediately close client connections when
6313 read(2) returns "no more data to read."
6315 Change this option to 'on' and Squid will keep open connections
6316 until a read(2) or write(2) on the socket returns an error.
6317 This may show some benefits for reverse proxies. But if not
6318 it is recommended to leave OFF.
6321 NAME: server_idle_pconn_timeout pconn_timeout
6323 LOC: Config.Timeout.serverIdlePconn
6326 Timeout for idle persistent connections to servers and other
6333 LOC: Ident::TheConfig.timeout
6336 Maximum time to wait for IDENT lookups to complete.
6338 If this is too high, and you enabled IDENT lookups from untrusted
6339 users, you might be susceptible to denial-of-service by having
6340 many ident requests going at once.
6343 NAME: shutdown_lifetime
6346 LOC: Config.shutdownLifetime
6349 When SIGTERM or SIGHUP is received, the cache is put into
6350 "shutdown pending" mode until all active sockets are closed.
6351 This value is the lifetime to set for all open descriptors
6352 during shutdown mode. Any active clients after this many
6353 seconds will receive a 'timeout' message.
6357 ADMINISTRATIVE PARAMETERS
6358 -----------------------------------------------------------------------------
6364 LOC: Config.adminEmail
6366 Email-address of local cache manager who will receive
6367 mail if the cache dies. The default is "webmaster".
6373 LOC: Config.EmailFrom
6375 From: email-address for mail sent when the cache dies.
6376 The default is to use 'squid@unique_hostname'.
6378 See also: unique_hostname directive.
6384 LOC: Config.EmailProgram
6386 Email program used to send mail if the cache dies.
6387 The default is "mail". The specified program must comply
6388 with the standard Unix mail syntax:
6389 mail-program recipient < mailfile
6391 Optional command line options can be specified.
6394 NAME: cache_effective_user
6396 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
6397 LOC: Config.effectiveUser
6399 If you start Squid as root, it will change its effective/real
6400 UID/GID to the user specified below. The default is to change
6401 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
6402 see also; cache_effective_group
6405 NAME: cache_effective_group
6408 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
6409 LOC: Config.effectiveGroup
6411 Squid sets the GID to the effective user's default group ID
6412 (taken from the password file) and supplementary group list
6413 from the groups membership.
6415 If you want Squid to run with a specific GID regardless of
6416 the group memberships of the effective user then set this
6417 to the group (or GID) you want Squid to run as. When set
6418 all other group privileges of the effective user are ignored
6419 and only this GID is effective. If Squid is not started as
6420 root the user starting Squid MUST be member of the specified
6423 This option is not recommended by the Squid Team.
6424 Our preference is for administrators to configure a secure
6425 user account for squid with UID/GID matching system policies.
6428 NAME: httpd_suppress_version_string
6432 LOC: Config.onoff.httpd_suppress_version_string
6434 Suppress Squid version string info in HTTP headers and HTML error pages.
6437 NAME: visible_hostname
6439 LOC: Config.visibleHostname
6441 DEFAULT_DOC: Automatically detect the system host name
6443 If you want to present a special hostname in error messages, etc,
6444 define this. Otherwise, the return value of gethostname()
6445 will be used. If you have multiple caches in a cluster and
6446 get errors about IP-forwarding you must set them to have individual
6447 names with this setting.
6450 NAME: unique_hostname
6452 LOC: Config.uniqueHostname
6454 DEFAULT_DOC: Copy the value from visible_hostname
6456 If you want to have multiple machines with the same
6457 'visible_hostname' you must give each machine a different
6458 'unique_hostname' so forwarding loops can be detected.
6461 NAME: hostname_aliases
6463 LOC: Config.hostnameAliases
6466 A list of other DNS names your cache has.
6474 Minimum umask which should be enforced while the proxy
6475 is running, in addition to the umask set at startup.
6477 For a traditional octal representation of umasks, start
6482 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6483 -----------------------------------------------------------------------------
6485 This section contains parameters for the (optional) cache
6486 announcement service. This service is provided to help
6487 cache administrators locate one another in order to join or
6488 create cache hierarchies.
6490 An 'announcement' message is sent (via UDP) to the registration
6491 service by Squid. By default, the announcement message is NOT
6492 SENT unless you enable it with 'announce_period' below.
6494 The announcement message includes your hostname, plus the
6495 following information from this configuration file:
6501 All current information is processed regularly and made
6502 available on the Web at http://www.ircache.net/Cache/Tracker/.
6505 NAME: announce_period
6507 LOC: Config.Announce.period
6509 DEFAULT_DOC: Announcement messages disabled.
6511 This is how frequently to send cache announcements.
6513 To enable announcing your cache, just set an announce period.
6516 announce_period 1 day
6521 DEFAULT: tracker.ircache.net
6522 LOC: Config.Announce.host
6524 Set the hostname where announce registration messages will be sent.
6526 See also announce_port and announce_file
6532 LOC: Config.Announce.file
6534 The contents of this file will be included in the announce
6535 registration messages.
6541 LOC: Config.Announce.port
6543 Set the port where announce registration messages will be sent.
6545 See also announce_host and announce_file
6549 HTTPD-ACCELERATOR OPTIONS
6550 -----------------------------------------------------------------------------
6553 NAME: httpd_accel_surrogate_id
6556 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6557 LOC: Config.Accel.surrogate_id
6559 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6560 need an identification token to allow control targeting. Because
6561 a farm of surrogates may all perform the same tasks, they may share
6562 an identification token.
6565 NAME: http_accel_surrogate_remote
6569 LOC: Config.onoff.surrogate_is_remote
6571 Remote surrogates (such as those in a CDN) honour the header
6572 "Surrogate-Control: no-store-remote".
6574 Set this to on to have squid behave as a remote surrogate.
6578 IFDEF: USE_SQUID_ESI
6579 COMMENT: libxml2|expat|custom
6581 LOC: ESIParser::Type
6584 ESI markup is not strictly XML compatible. The custom ESI parser
6585 will give higher performance, but cannot handle non ASCII character
6590 DELAY POOL PARAMETERS
6591 -----------------------------------------------------------------------------
6595 TYPE: delay_pool_count
6597 IFDEF: USE_DELAY_POOLS
6600 This represents the number of delay pools to be used. For example,
6601 if you have one class 2 delay pool and one class 3 delays pool, you
6602 have a total of 2 delay pools.
6604 See also delay_parameters, delay_class, delay_access for pool
6605 configuration details.
6609 TYPE: delay_pool_class
6611 IFDEF: USE_DELAY_POOLS
6614 This defines the class of each delay pool. There must be exactly one
6615 delay_class line for each delay pool. For example, to define two
6616 delay pools, one of class 2 and one of class 3, the settings above
6620 delay_pools 4 # 4 delay pools
6621 delay_class 1 2 # pool 1 is a class 2 pool
6622 delay_class 2 3 # pool 2 is a class 3 pool
6623 delay_class 3 4 # pool 3 is a class 4 pool
6624 delay_class 4 5 # pool 4 is a class 5 pool
6626 The delay pool classes are:
6628 class 1 Everything is limited by a single aggregate
6631 class 2 Everything is limited by a single aggregate
6632 bucket as well as an "individual" bucket chosen
6633 from bits 25 through 32 of the IPv4 address.
6635 class 3 Everything is limited by a single aggregate
6636 bucket as well as a "network" bucket chosen
6637 from bits 17 through 24 of the IP address and a
6638 "individual" bucket chosen from bits 17 through
6639 32 of the IPv4 address.
6641 class 4 Everything in a class 3 delay pool, with an
6642 additional limit on a per user basis. This
6643 only takes effect if the username is established
6644 in advance - by forcing authentication in your
6647 class 5 Requests are grouped according their tag (see
6648 external_acl's tag= reply).
6651 Each pool also requires a delay_parameters directive to configure the pool size
6652 and speed limits used whenever the pool is applied to a request. Along with
6653 a set of delay_access directives to determine when it is used.
6655 NOTE: If an IP address is a.b.c.d
6656 -> bits 25 through 32 are "d"
6657 -> bits 17 through 24 are "c"
6658 -> bits 17 through 32 are "c * 256 + d"
6660 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6661 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6663 This clause only supports fast acl types.
6664 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6666 See also delay_parameters and delay_access.
6670 TYPE: delay_pool_access
6672 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6673 IFDEF: USE_DELAY_POOLS
6676 This is used to determine which delay pool a request falls into.
6678 delay_access is sorted per pool and the matching starts with pool 1,
6679 then pool 2, ..., and finally pool N. The first delay pool where the
6680 request is allowed is selected for the request. If it does not allow
6681 the request to any pool then the request is not delayed (default).
6683 For example, if you want some_big_clients in delay
6684 pool 1 and lotsa_little_clients in delay pool 2:
6686 delay_access 1 allow some_big_clients
6687 delay_access 1 deny all
6688 delay_access 2 allow lotsa_little_clients
6689 delay_access 2 deny all
6690 delay_access 3 allow authenticated_clients
6692 See also delay_parameters and delay_class.
6696 NAME: delay_parameters
6697 TYPE: delay_pool_rates
6699 IFDEF: USE_DELAY_POOLS
6702 This defines the parameters for a delay pool. Each delay pool has
6703 a number of "buckets" associated with it, as explained in the
6704 description of delay_class.
6706 For a class 1 delay pool, the syntax is:
6708 delay_parameters pool aggregate
6710 For a class 2 delay pool:
6712 delay_parameters pool aggregate individual
6714 For a class 3 delay pool:
6716 delay_parameters pool aggregate network individual
6718 For a class 4 delay pool:
6720 delay_parameters pool aggregate network individual user
6722 For a class 5 delay pool:
6724 delay_parameters pool tagrate
6726 The option variables are:
6728 pool a pool number - ie, a number between 1 and the
6729 number specified in delay_pools as used in
6732 aggregate the speed limit parameters for the aggregate bucket
6735 individual the speed limit parameters for the individual
6736 buckets (class 2, 3).
6738 network the speed limit parameters for the network buckets
6741 user the speed limit parameters for the user buckets
6744 tagrate the speed limit parameters for the tag buckets
6747 A pair of delay parameters is written restore/maximum, where restore is
6748 the number of bytes (not bits - modem and network speeds are usually
6749 quoted in bits) per second placed into the bucket, and maximum is the
6750 maximum number of bytes which can be in the bucket at any time.
6752 There must be one delay_parameters line for each delay pool.
6755 For example, if delay pool number 1 is a class 2 delay pool as in the
6756 above example, and is being used to strictly limit each host to 64Kbit/sec
6757 (plus overheads), with no overall limit, the line is:
6759 delay_parameters 1 none 8000/8000
6761 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6763 Note that the word 'none' is used to represent no limit.
6766 And, if delay pool number 2 is a class 3 delay pool as in the above
6767 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6768 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6769 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6770 to permit a decent web page to be downloaded at a decent speed
6771 (if the network is not being limited due to overuse) but slow down
6772 large downloads more significantly:
6774 delay_parameters 2 32000/32000 8000/8000 600/8000
6776 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6777 8 x 8000 KByte/sec -> 64Kbit/sec.
6778 8 x 600 Byte/sec -> 4800bit/sec.
6781 Finally, for a class 4 delay pool as in the example - each user will
6782 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6784 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6787 See also delay_class and delay_access.
6791 NAME: delay_initial_bucket_level
6792 COMMENT: (percent, 0-100)
6795 IFDEF: USE_DELAY_POOLS
6796 LOC: Config.Delay.initial
6798 The initial bucket percentage is used to determine how much is put
6799 in each bucket when squid starts, is reconfigured, or first notices
6800 a host accessing it (in class 2 and class 3, individual hosts and
6801 networks only have buckets associated with them once they have been
6806 CLIENT DELAY POOL PARAMETERS
6807 -----------------------------------------------------------------------------
6810 NAME: client_delay_pools
6811 TYPE: client_delay_pool_count
6813 IFDEF: USE_DELAY_POOLS
6814 LOC: Config.ClientDelay
6816 This option specifies the number of client delay pools used. It must
6817 preceed other client_delay_* options.
6820 client_delay_pools 2
6822 See also client_delay_parameters and client_delay_access.
6825 NAME: client_delay_initial_bucket_level
6826 COMMENT: (percent, 0-no_limit)
6829 IFDEF: USE_DELAY_POOLS
6830 LOC: Config.ClientDelay.initial
6832 This option determines the initial bucket size as a percentage of
6833 max_bucket_size from client_delay_parameters. Buckets are created
6834 at the time of the "first" connection from the matching IP. Idle
6835 buckets are periodically deleted up.
6837 You can specify more than 100 percent but note that such "oversized"
6838 buckets are not refilled until their size goes down to max_bucket_size
6839 from client_delay_parameters.
6842 client_delay_initial_bucket_level 50
6845 NAME: client_delay_parameters
6846 TYPE: client_delay_pool_rates
6848 IFDEF: USE_DELAY_POOLS
6849 LOC: Config.ClientDelay
6852 This option configures client-side bandwidth limits using the
6855 client_delay_parameters pool speed_limit max_bucket_size
6857 pool is an integer ID used for client_delay_access matching.
6859 speed_limit is bytes added to the bucket per second.
6861 max_bucket_size is the maximum size of a bucket, enforced after any
6862 speed_limit additions.
6864 Please see the delay_parameters option for more information and
6868 client_delay_parameters 1 1024 2048
6869 client_delay_parameters 2 51200 16384
6871 See also client_delay_access.
6875 NAME: client_delay_access
6876 TYPE: client_delay_pool_access
6878 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6879 IFDEF: USE_DELAY_POOLS
6880 LOC: Config.ClientDelay
6882 This option determines the client-side delay pool for the
6885 client_delay_access pool_ID allow|deny acl_name
6887 All client_delay_access options are checked in their pool ID
6888 order, starting with pool 1. The first checked pool with allowed
6889 request is selected for the request. If no ACL matches or there
6890 are no client_delay_access options, the request bandwidth is not
6893 The ACL-selected pool is then used to find the
6894 client_delay_parameters for the request. Client-side pools are
6895 not used to aggregate clients. Clients are always aggregated
6896 based on their source IP addresses (one bucket per source IP).
6898 This clause only supports fast acl types.
6899 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6900 Additionally, only the client TCP connection details are available.
6901 ACLs testing HTTP properties will not work.
6903 Please see delay_access for more examples.
6906 client_delay_access 1 allow low_rate_network
6907 client_delay_access 2 allow vips_network
6910 See also client_delay_parameters and client_delay_pools.
6914 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6915 -----------------------------------------------------------------------------
6920 LOC: Config.Wccp.router
6922 DEFAULT_DOC: WCCP disabled.
6925 Use this option to define your WCCP ``home'' router for
6928 wccp_router supports a single WCCP(v1) router
6930 wccp2_router supports multiple WCCPv2 routers
6932 only one of the two may be used at the same time and defines
6933 which version of WCCP to use.
6937 TYPE: IpAddress_list
6938 LOC: Config.Wccp2.router
6940 DEFAULT_DOC: WCCPv2 disabled.
6943 Use this option to define your WCCP ``home'' router for
6946 wccp_router supports a single WCCP(v1) router
6948 wccp2_router supports multiple WCCPv2 routers
6950 only one of the two may be used at the same time and defines
6951 which version of WCCP to use.
6956 LOC: Config.Wccp.version
6960 This directive is only relevant if you need to set up WCCP(v1)
6961 to some very old and end-of-life Cisco routers. In all other
6962 setups it must be left unset or at the default setting.
6963 It defines an internal version in the WCCP(v1) protocol,
6964 with version 4 being the officially documented protocol.
6966 According to some users, Cisco IOS 11.2 and earlier only
6967 support WCCP version 3. If you're using that or an earlier
6968 version of IOS, you may need to change this value to 3, otherwise
6969 do not specify this parameter.
6972 NAME: wccp2_rebuild_wait
6974 LOC: Config.Wccp2.rebuildwait
6978 If this is enabled Squid will wait for the cache dir rebuild to finish
6979 before sending the first wccp2 HereIAm packet
6982 NAME: wccp2_forwarding_method
6984 LOC: Config.Wccp2.forwarding_method
6988 WCCP2 allows the setting of forwarding methods between the
6989 router/switch and the cache. Valid values are as follows:
6991 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6992 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6994 Currently (as of IOS 12.4) cisco routers only support GRE.
6995 Cisco switches only support the L2 redirect assignment method.
6998 NAME: wccp2_return_method
7000 LOC: Config.Wccp2.return_method
7004 WCCP2 allows the setting of return methods between the
7005 router/switch and the cache for packets that the cache
7006 decides not to handle. Valid values are as follows:
7008 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
7009 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
7011 Currently (as of IOS 12.4) cisco routers only support GRE.
7012 Cisco switches only support the L2 redirect assignment.
7014 If the "ip wccp redirect exclude in" command has been
7015 enabled on the cache interface, then it is still safe for
7016 the proxy server to use a l2 redirect method even if this
7017 option is set to GRE.
7020 NAME: wccp2_assignment_method
7022 LOC: Config.Wccp2.assignment_method
7026 WCCP2 allows the setting of methods to assign the WCCP hash
7027 Valid values are as follows:
7029 hash - Hash assignment
7030 mask - Mask assignment
7032 As a general rule, cisco routers support the hash assignment method
7033 and cisco switches support the mask assignment method.
7038 LOC: Config.Wccp2.info
7039 DEFAULT_IF_NONE: standard 0
7040 DEFAULT_DOC: Use the 'web-cache' standard service.
7043 WCCP2 allows for multiple traffic services. There are two
7044 types: "standard" and "dynamic". The standard type defines
7045 one service id - http (id 0). The dynamic service ids can be from
7046 51 to 255 inclusive. In order to use a dynamic service id
7047 one must define the type of traffic to be redirected; this is done
7048 using the wccp2_service_info option.
7050 The "standard" type does not require a wccp2_service_info option,
7051 just specifying the service id will suffice.
7053 MD5 service authentication can be enabled by adding
7054 "password=<password>" to the end of this service declaration.
7058 wccp2_service standard 0 # for the 'web-cache' standard service
7059 wccp2_service dynamic 80 # a dynamic service type which will be
7060 # fleshed out with subsequent options.
7061 wccp2_service standard 0 password=foo
7064 NAME: wccp2_service_info
7065 TYPE: wccp2_service_info
7066 LOC: Config.Wccp2.info
7070 Dynamic WCCPv2 services require further information to define the
7071 traffic you wish to have diverted.
7075 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
7076 priority=<priority> ports=<port>,<port>..
7078 The relevant WCCPv2 flags:
7079 + src_ip_hash, dst_ip_hash
7080 + source_port_hash, dst_port_hash
7081 + src_ip_alt_hash, dst_ip_alt_hash
7082 + src_port_alt_hash, dst_port_alt_hash
7085 The port list can be one to eight entries.
7089 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
7090 priority=240 ports=80
7092 Note: the service id must have been defined by a previous
7093 'wccp2_service dynamic <id>' entry.
7098 LOC: Config.Wccp2.weight
7102 Each cache server gets assigned a set of the destination
7103 hash proportional to their weight.
7108 LOC: Config.Wccp.address
7110 DEFAULT_DOC: Address selected by the operating system.
7113 Use this option if you require WCCPv2 to use a specific
7116 The default behavior is to not bind to any specific address.
7121 LOC: Config.Wccp2.address
7123 DEFAULT_DOC: Address selected by the operating system.
7126 Use this option if you require WCCP to use a specific
7129 The default behavior is to not bind to any specific address.
7133 PERSISTENT CONNECTION HANDLING
7134 -----------------------------------------------------------------------------
7136 Also see "pconn_timeout" in the TIMEOUTS section
7139 NAME: client_persistent_connections
7141 LOC: Config.onoff.client_pconns
7144 Persistent connection support for clients.
7145 Squid uses persistent connections (when allowed). You can use
7146 this option to disable persistent connections with clients.
7149 NAME: server_persistent_connections
7151 LOC: Config.onoff.server_pconns
7154 Persistent connection support for servers.
7155 Squid uses persistent connections (when allowed). You can use
7156 this option to disable persistent connections with servers.
7159 NAME: persistent_connection_after_error
7161 LOC: Config.onoff.error_pconns
7164 With this directive the use of persistent connections after
7165 HTTP errors can be disabled. Useful if you have clients
7166 who fail to handle errors on persistent connections proper.
7169 NAME: detect_broken_pconn
7171 LOC: Config.onoff.detect_broken_server_pconns
7174 Some servers have been found to incorrectly signal the use
7175 of HTTP/1.0 persistent connections even on replies not
7176 compatible, causing significant delays. This server problem
7177 has mostly been seen on redirects.
7179 By enabling this directive Squid attempts to detect such
7180 broken replies and automatically assume the reply is finished
7181 after 10 seconds timeout.
7185 CACHE DIGEST OPTIONS
7186 -----------------------------------------------------------------------------
7189 NAME: digest_generation
7190 IFDEF: USE_CACHE_DIGESTS
7192 LOC: Config.onoff.digest_generation
7195 This controls whether the server will generate a Cache Digest
7196 of its contents. By default, Cache Digest generation is
7197 enabled if Squid is compiled with --enable-cache-digests defined.
7200 NAME: digest_bits_per_entry
7201 IFDEF: USE_CACHE_DIGESTS
7203 LOC: Config.digest.bits_per_entry
7206 This is the number of bits of the server's Cache Digest which
7207 will be associated with the Digest entry for a given HTTP
7208 Method and URL (public key) combination. The default is 5.
7211 NAME: digest_rebuild_period
7212 IFDEF: USE_CACHE_DIGESTS
7215 LOC: Config.digest.rebuild_period
7218 This is the wait time between Cache Digest rebuilds.
7221 NAME: digest_rewrite_period
7223 IFDEF: USE_CACHE_DIGESTS
7225 LOC: Config.digest.rewrite_period
7228 This is the wait time between Cache Digest writes to
7232 NAME: digest_swapout_chunk_size
7235 IFDEF: USE_CACHE_DIGESTS
7236 LOC: Config.digest.swapout_chunk_size
7239 This is the number of bytes of the Cache Digest to write to
7240 disk at a time. It defaults to 4096 bytes (4KB), the Squid
7244 NAME: digest_rebuild_chunk_percentage
7245 COMMENT: (percent, 0-100)
7246 IFDEF: USE_CACHE_DIGESTS
7248 LOC: Config.digest.rebuild_chunk_percentage
7251 This is the percentage of the Cache Digest to be scanned at a
7252 time. By default it is set to 10% of the Cache Digest.
7257 -----------------------------------------------------------------------------
7262 LOC: Config.Port.snmp
7264 DEFAULT_DOC: SNMP disabled.
7267 The port number where Squid listens for SNMP requests. To enable
7268 SNMP support set this to a suitable port number. Port number
7269 3401 is often used for the Squid SNMP agent. By default it's
7270 set to "0" (disabled)
7278 LOC: Config.accessList.snmp
7280 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
7283 Allowing or denying access to the SNMP port.
7285 All access to the agent is denied by default.
7288 snmp_access allow|deny [!]aclname ...
7290 This clause only supports fast acl types.
7291 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7294 snmp_access allow snmppublic localhost
7295 snmp_access deny all
7298 NAME: snmp_incoming_address
7300 LOC: Config.Addrs.snmp_incoming
7302 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
7305 Just like 'udp_incoming_address', but for the SNMP port.
7307 snmp_incoming_address is used for the SNMP socket receiving
7308 messages from SNMP agents.
7310 The default snmp_incoming_address is to listen on all
7311 available network interfaces.
7314 NAME: snmp_outgoing_address
7316 LOC: Config.Addrs.snmp_outgoing
7318 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
7321 Just like 'udp_outgoing_address', but for the SNMP port.
7323 snmp_outgoing_address is used for SNMP packets returned to SNMP
7326 If snmp_outgoing_address is not set it will use the same socket
7327 as snmp_incoming_address. Only change this if you want to have
7328 SNMP replies sent using another address than where this Squid
7329 listens for SNMP queries.
7331 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
7332 the same value since they both use the same port.
7337 -----------------------------------------------------------------------------
7340 NAME: icp_port udp_port
7343 DEFAULT_DOC: ICP disabled.
7344 LOC: Config.Port.icp
7346 The port number where Squid sends and receives ICP queries to
7347 and from neighbor caches. The standard UDP port for ICP is 3130.
7350 icp_port @DEFAULT_ICP_PORT@
7357 DEFAULT_DOC: HTCP disabled.
7358 LOC: Config.Port.htcp
7360 The port number where Squid sends and receives HTCP queries to
7361 and from neighbor caches. To turn it on you want to set it to
7368 NAME: log_icp_queries
7372 LOC: Config.onoff.log_udp
7374 If set, ICP queries are logged to access.log. You may wish
7375 do disable this if your ICP load is VERY high to speed things
7376 up or to simplify log analysis.
7379 NAME: udp_incoming_address
7381 LOC:Config.Addrs.udp_incoming
7383 DEFAULT_DOC: Accept packets from all machine interfaces.
7385 udp_incoming_address is used for UDP packets received from other
7388 The default behavior is to not bind to any specific address.
7390 Only change this if you want to have all UDP queries received on
7391 a specific interface/address.
7393 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
7394 modules. Altering it will affect all of them in the same manner.
7396 see also; udp_outgoing_address
7398 NOTE, udp_incoming_address and udp_outgoing_address can not
7399 have the same value since they both use the same port.
7402 NAME: udp_outgoing_address
7404 LOC: Config.Addrs.udp_outgoing
7406 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
7408 udp_outgoing_address is used for UDP packets sent out to other
7411 The default behavior is to not bind to any specific address.
7413 Instead it will use the same socket as udp_incoming_address.
7414 Only change this if you want to have UDP queries sent using another
7415 address than where this Squid listens for UDP queries from other
7418 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
7419 modules. Altering it will affect all of them in the same manner.
7421 see also; udp_incoming_address
7423 NOTE, udp_incoming_address and udp_outgoing_address can not
7424 have the same value since they both use the same port.
7431 LOC: Config.onoff.icp_hit_stale
7433 If you want to return ICP_HIT for stale cache objects, set this
7434 option to 'on'. If you have sibling relationships with caches
7435 in other administrative domains, this should be 'off'. If you only
7436 have sibling relationships with caches under your control,
7437 it is probably okay to set this to 'on'.
7438 If set to 'on', your siblings should use the option "allow-miss"
7439 on their cache_peer lines for connecting to you.
7442 NAME: minimum_direct_hops
7445 LOC: Config.minDirectHops
7447 If using the ICMP pinging stuff, do direct fetches for sites
7448 which are no more than this many hops away.
7451 NAME: minimum_direct_rtt
7455 LOC: Config.minDirectRtt
7457 If using the ICMP pinging stuff, do direct fetches for sites
7458 which are no more than this many rtt milliseconds away.
7464 LOC: Config.Netdb.low
7466 The low water mark for the ICMP measurement database.
7468 Note: high watermark controlled by netdb_high directive.
7470 These watermarks are counts, not percents. The defaults are
7471 (low) 900 and (high) 1000. When the high water mark is
7472 reached, database entries will be deleted until the low
7479 LOC: Config.Netdb.high
7481 The high water mark for the ICMP measurement database.
7483 Note: low watermark controlled by netdb_low directive.
7485 These watermarks are counts, not percents. The defaults are
7486 (low) 900 and (high) 1000. When the high water mark is
7487 reached, database entries will be deleted until the low
7491 NAME: netdb_ping_period
7493 LOC: Config.Netdb.period
7496 The minimum period for measuring a site. There will be at
7497 least this much delay between successive pings to the same
7498 network. The default is five minutes.
7505 LOC: Config.onoff.query_icmp
7507 If you want to ask your peers to include ICMP data in their ICP
7508 replies, enable this option.
7510 If your peer has configured Squid (during compilation) with
7511 '--enable-icmp' that peer will send ICMP pings to origin server
7512 sites of the URLs it receives. If you enable this option the
7513 ICP replies from that peer will include the ICMP data (if available).
7514 Then, when choosing a parent cache, Squid will choose the parent with
7515 the minimal RTT to the origin server. When this happens, the
7516 hierarchy field of the access.log will be
7517 "CLOSEST_PARENT_MISS". This option is off by default.
7520 NAME: test_reachability
7524 LOC: Config.onoff.test_reachability
7526 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7527 instead of ICP_MISS if the target host is NOT in the ICMP
7528 database, or has a zero RTT.
7531 NAME: icp_query_timeout
7534 DEFAULT_DOC: Dynamic detection.
7536 LOC: Config.Timeout.icp_query
7538 Normally Squid will automatically determine an optimal ICP
7539 query timeout value based on the round-trip-time of recent ICP
7540 queries. If you want to override the value determined by
7541 Squid, set this 'icp_query_timeout' to a non-zero value. This
7542 value is specified in MILLISECONDS, so, to use a 2-second
7543 timeout (the old default), you would write:
7545 icp_query_timeout 2000
7548 NAME: maximum_icp_query_timeout
7552 LOC: Config.Timeout.icp_query_max
7554 Normally the ICP query timeout is determined dynamically. But
7555 sometimes it can lead to very large values (say 5 seconds).
7556 Use this option to put an upper limit on the dynamic timeout
7557 value. Do NOT use this option to always use a fixed (instead
7558 of a dynamic) timeout value. To set a fixed timeout see the
7559 'icp_query_timeout' directive.
7562 NAME: minimum_icp_query_timeout
7566 LOC: Config.Timeout.icp_query_min
7568 Normally the ICP query timeout is determined dynamically. But
7569 sometimes it can lead to very small timeouts, even lower than
7570 the normal latency variance on your link due to traffic.
7571 Use this option to put an lower limit on the dynamic timeout
7572 value. Do NOT use this option to always use a fixed (instead
7573 of a dynamic) timeout value. To set a fixed timeout see the
7574 'icp_query_timeout' directive.
7577 NAME: background_ping_rate
7581 LOC: Config.backgroundPingRate
7583 Controls how often the ICP pings are sent to siblings that
7584 have background-ping set.
7588 MULTICAST ICP OPTIONS
7589 -----------------------------------------------------------------------------
7594 LOC: Config.mcast_group_list
7597 This tag specifies a list of multicast groups which your server
7598 should join to receive multicasted ICP queries.
7600 NOTE! Be very careful what you put here! Be sure you
7601 understand the difference between an ICP _query_ and an ICP
7602 _reply_. This option is to be set only if you want to RECEIVE
7603 multicast queries. Do NOT set this option to SEND multicast
7604 ICP (use cache_peer for that). ICP replies are always sent via
7605 unicast, so this option does not affect whether or not you will
7606 receive replies from multicast group members.
7608 You must be very careful to NOT use a multicast address which
7609 is already in use by another group of caches.
7611 If you are unsure about multicast, please read the Multicast
7612 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7614 Usage: mcast_groups 239.128.16.128 224.0.1.20
7616 By default, Squid doesn't listen on any multicast groups.
7619 NAME: mcast_miss_addr
7620 IFDEF: MULTICAST_MISS_STREAM
7622 LOC: Config.mcast_miss.addr
7624 DEFAULT_DOC: disabled.
7626 If you enable this option, every "cache miss" URL will
7627 be sent out on the specified multicast address.
7629 Do not enable this option unless you are are absolutely
7630 certain you understand what you are doing.
7633 NAME: mcast_miss_ttl
7634 IFDEF: MULTICAST_MISS_STREAM
7636 LOC: Config.mcast_miss.ttl
7639 This is the time-to-live value for packets multicasted
7640 when multicasting off cache miss URLs is enabled. By
7641 default this is set to 'site scope', i.e. 16.
7644 NAME: mcast_miss_port
7645 IFDEF: MULTICAST_MISS_STREAM
7647 LOC: Config.mcast_miss.port
7650 This is the port number to be used in conjunction with
7654 NAME: mcast_miss_encode_key
7655 IFDEF: MULTICAST_MISS_STREAM
7657 LOC: Config.mcast_miss.encode_key
7658 DEFAULT: XXXXXXXXXXXXXXXX
7660 The URLs that are sent in the multicast miss stream are
7661 encrypted. This is the encryption key.
7664 NAME: mcast_icp_query_timeout
7668 LOC: Config.Timeout.mcast_icp_query
7670 For multicast peers, Squid regularly sends out ICP "probes" to
7671 count how many other peers are listening on the given multicast
7672 address. This value specifies how long Squid should wait to
7673 count all the replies. The default is 2000 msec, or 2
7678 INTERNAL ICON OPTIONS
7679 -----------------------------------------------------------------------------
7682 NAME: icon_directory
7684 LOC: Config.icons.directory
7685 DEFAULT: @DEFAULT_ICON_DIR@
7687 Where the icons are stored. These are normally kept in
7691 NAME: global_internal_static
7693 LOC: Config.onoff.global_internal_static
7696 This directive controls is Squid should intercept all requests for
7697 /squid-internal-static/ no matter which host the URL is requesting
7698 (default on setting), or if nothing special should be done for
7699 such URLs (off setting). The purpose of this directive is to make
7700 icons etc work better in complex cache hierarchies where it may
7701 not always be possible for all corners in the cache mesh to reach
7702 the server generating a directory listing.
7705 NAME: short_icon_urls
7707 LOC: Config.icons.use_short_names
7710 If this is enabled Squid will use short URLs for icons.
7711 If disabled it will revert to the old behavior of including
7712 it's own name and port in the URL.
7714 If you run a complex cache hierarchy with a mix of Squid and
7715 other proxies you may need to disable this directive.
7720 -----------------------------------------------------------------------------
7723 NAME: error_directory
7725 LOC: Config.errorDirectory
7727 DEFAULT_DOC: Send error pages in the clients preferred language
7729 If you wish to create your own versions of the default
7730 error files to customize them to suit your company copy
7731 the error/template files to another directory and point
7734 WARNING: This option will disable multi-language support
7735 on error pages if used.
7737 The squid developers are interested in making squid available in
7738 a wide variety of languages. If you are making translations for a
7739 language that Squid does not currently provide please consider
7740 contributing your translation back to the project.
7741 http://wiki.squid-cache.org/Translations
7743 The squid developers working on translations are happy to supply drop-in
7744 translated error files in exchange for any new language contributions.
7747 NAME: error_default_language
7748 IFDEF: USE_ERR_LOCALES
7750 LOC: Config.errorDefaultLanguage
7752 DEFAULT_DOC: Generate English language pages.
7754 Set the default language which squid will send error pages in
7755 if no existing translation matches the clients language
7758 If unset (default) generic English will be used.
7760 The squid developers are interested in making squid available in
7761 a wide variety of languages. If you are interested in making
7762 translations for any language see the squid wiki for details.
7763 http://wiki.squid-cache.org/Translations
7766 NAME: error_log_languages
7767 IFDEF: USE_ERR_LOCALES
7769 LOC: Config.errorLogMissingLanguages
7772 Log to cache.log what languages users are attempting to
7773 auto-negotiate for translations.
7775 Successful negotiations are not logged. Only failures
7776 have meaning to indicate that Squid may need an upgrade
7777 of its error page translations.
7780 NAME: err_page_stylesheet
7782 LOC: Config.errorStylesheet
7783 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7785 CSS Stylesheet to pattern the display of Squid default error pages.
7787 For information on CSS see http://www.w3.org/Style/CSS/
7792 LOC: Config.errHtmlText
7795 HTML text to include in error messages. Make this a "mailto"
7796 URL to your admin address, or maybe just a link to your
7797 organizations Web page.
7799 To include this in your error messages, you must rewrite
7800 the error template files (found in the "errors" directory).
7801 Wherever you want the 'err_html_text' line to appear,
7802 insert a %L tag in the error template file.
7805 NAME: email_err_data
7808 LOC: Config.onoff.emailErrData
7811 If enabled, information about the occurred error will be
7812 included in the mailto links of the ERR pages (if %W is set)
7813 so that the email body contains the data.
7814 Syntax is <A HREF="mailto:%w%W">%w</A>
7819 LOC: Config.denyInfoList
7822 Usage: deny_info err_page_name acl
7823 or deny_info http://... acl
7824 or deny_info TCP_RESET acl
7826 This can be used to return a ERR_ page for requests which
7827 do not pass the 'http_access' rules. Squid remembers the last
7828 acl it evaluated in http_access, and if a 'deny_info' line exists
7829 for that ACL Squid returns a corresponding error page.
7831 The acl is typically the last acl on the http_access deny line which
7832 denied access. The exceptions to this rule are:
7833 - When Squid needs to request authentication credentials. It's then
7834 the first authentication related acl encountered
7835 - When none of the http_access lines matches. It's then the last
7836 acl processed on the last http_access line.
7837 - When the decision to deny access was made by an adaptation service,
7838 the acl name is the corresponding eCAP or ICAP service_name.
7840 NP: If providing your own custom error pages with error_directory
7841 you may also specify them by your custom file name:
7842 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7844 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7845 may be specified by prefixing the file name with the code and a colon.
7846 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7848 Alternatively you can tell Squid to reset the TCP connection
7849 by specifying TCP_RESET.
7851 Or you can specify an error URL or URL pattern. The browsers will
7852 get redirected to the specified URL after formatting tags have
7853 been replaced. Redirect will be done with 302 or 307 according to
7854 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7855 the URL. e.g. 303:http://example.com/
7858 %a - username (if available. Password NOT included)
7861 %E - Error description
7863 %H - Request domain name
7864 %i - Client IP Address
7866 %o - Message result from external ACL helper
7867 %p - Request Port number
7868 %P - Request Protocol name
7869 %R - Request URL path
7870 %T - Timestamp in RFC 1123 format
7871 %U - Full canonical URL from client
7872 (HTTPS URLs terminate with *)
7873 %u - Full canonical URL from client
7874 %w - Admin email from squid.conf
7876 %% - Literal percent (%) code
7881 OPTIONS INFLUENCING REQUEST FORWARDING
7882 -----------------------------------------------------------------------------
7885 NAME: nonhierarchical_direct
7887 LOC: Config.onoff.nonhierarchical_direct
7890 By default, Squid will send any non-hierarchical requests
7891 (not cacheable request type) direct to origin servers.
7893 When this is set to "off", Squid will prefer to send these
7894 requests to parents.
7896 Note that in most configurations, by turning this off you will only
7897 add latency to these request without any improvement in global hit
7900 This option only sets a preference. If the parent is unavailable a
7901 direct connection to the origin server may still be attempted. To
7902 completely prevent direct connections use never_direct.
7907 LOC: Config.onoff.prefer_direct
7910 Normally Squid tries to use parents for most requests. If you for some
7911 reason like it to first try going direct and only use a parent if
7912 going direct fails set this to on.
7914 By combining nonhierarchical_direct off and prefer_direct on you
7915 can set up Squid to use a parent as a backup path if going direct
7918 Note: If you want Squid to use parents for all requests see
7919 the never_direct directive. prefer_direct only modifies how Squid
7920 acts on cacheable requests.
7923 NAME: cache_miss_revalidate
7927 LOC: Config.onoff.cache_miss_revalidate
7929 RFC 7232 defines a conditional request mechanism to prevent
7930 response objects being unnecessarily transferred over the network.
7931 If that mechanism is used by the client and a cache MISS occurs
7932 it can prevent new cache entries being created.
7934 This option determines whether Squid on cache MISS will pass the
7935 client revalidation request to the server or tries to fetch new
7936 content for caching. It can be useful while the cache is mostly
7937 empty to more quickly have the cache populated by generating
7938 non-conditional GETs.
7940 When set to 'on' (default), Squid will pass all client If-* headers
7941 to the server. This permits server responses without a cacheable
7942 payload to be delivered and on MISS no new cache entry is created.
7944 When set to 'off' and if the request is cacheable, Squid will
7945 remove the clients If-Modified-Since and If-None-Match headers from
7946 the request sent to the server. This requests a 200 status response
7947 from the server to create a new cache entry with.
7952 LOC: Config.accessList.AlwaysDirect
7954 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7956 Usage: always_direct allow|deny [!]aclname ...
7958 Here you can use ACL elements to specify requests which should
7959 ALWAYS be forwarded by Squid to the origin servers without using
7960 any peers. For example, to always directly forward requests for
7961 local servers ignoring any parents or siblings you may have use
7964 acl local-servers dstdomain my.domain.net
7965 always_direct allow local-servers
7967 To always forward FTP requests directly, use
7970 always_direct allow FTP
7972 NOTE: There is a similar, but opposite option named
7973 'never_direct'. You need to be aware that "always_direct deny
7974 foo" is NOT the same thing as "never_direct allow foo". You
7975 may need to use a deny rule to exclude a more-specific case of
7976 some other rule. Example:
7978 acl local-external dstdomain external.foo.net
7979 acl local-servers dstdomain .foo.net
7980 always_direct deny local-external
7981 always_direct allow local-servers
7983 NOTE: If your goal is to make the client forward the request
7984 directly to the origin server bypassing Squid then this needs
7985 to be done in the client configuration. Squid configuration
7986 can only tell Squid how Squid should fetch the object.
7988 NOTE: This directive is not related to caching. The replies
7989 is cached as usual even if you use always_direct. To not cache
7990 the replies see the 'cache' directive.
7992 This clause supports both fast and slow acl types.
7993 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7998 LOC: Config.accessList.NeverDirect
8000 DEFAULT_DOC: Allow DNS results to be used for this request.
8002 Usage: never_direct allow|deny [!]aclname ...
8004 never_direct is the opposite of always_direct. Please read
8005 the description for always_direct if you have not already.
8007 With 'never_direct' you can use ACL elements to specify
8008 requests which should NEVER be forwarded directly to origin
8009 servers. For example, to force the use of a proxy for all
8010 requests, except those in your local domain use something like:
8012 acl local-servers dstdomain .foo.net
8013 never_direct deny local-servers
8014 never_direct allow all
8016 or if Squid is inside a firewall and there are local intranet
8017 servers inside the firewall use something like:
8019 acl local-intranet dstdomain .foo.net
8020 acl local-external dstdomain external.foo.net
8021 always_direct deny local-external
8022 always_direct allow local-intranet
8023 never_direct allow all
8025 This clause supports both fast and slow acl types.
8026 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
8030 ADVANCED NETWORKING OPTIONS
8031 -----------------------------------------------------------------------------
8034 NAME: incoming_udp_average incoming_icp_average
8037 LOC: Config.comm_incoming.udp.average
8039 Heavy voodoo here. I can't even believe you are reading this.
8040 Are you crazy? Don't even think about adjusting these unless
8041 you understand the algorithms in comm_select.c first!
8044 NAME: incoming_tcp_average incoming_http_average
8047 LOC: Config.comm_incoming.tcp.average
8049 Heavy voodoo here. I can't even believe you are reading this.
8050 Are you crazy? Don't even think about adjusting these unless
8051 you understand the algorithms in comm_select.c first!
8054 NAME: incoming_dns_average
8057 LOC: Config.comm_incoming.dns.average
8059 Heavy voodoo here. I can't even believe you are reading this.
8060 Are you crazy? Don't even think about adjusting these unless
8061 you understand the algorithms in comm_select.c first!
8064 NAME: min_udp_poll_cnt min_icp_poll_cnt
8067 LOC: Config.comm_incoming.udp.min_poll
8069 Heavy voodoo here. I can't even believe you are reading this.
8070 Are you crazy? Don't even think about adjusting these unless
8071 you understand the algorithms in comm_select.c first!
8074 NAME: min_dns_poll_cnt
8077 LOC: Config.comm_incoming.dns.min_poll
8079 Heavy voodoo here. I can't even believe you are reading this.
8080 Are you crazy? Don't even think about adjusting these unless
8081 you understand the algorithms in comm_select.c first!
8084 NAME: min_tcp_poll_cnt min_http_poll_cnt
8087 LOC: Config.comm_incoming.tcp.min_poll
8089 Heavy voodoo here. I can't even believe you are reading this.
8090 Are you crazy? Don't even think about adjusting these unless
8091 you understand the algorithms in comm_select.c first!
8097 LOC: Config.accept_filter
8101 The name of an accept(2) filter to install on Squid's
8102 listen socket(s). This feature is perhaps specific to
8103 FreeBSD and requires support in the kernel.
8105 The 'httpready' filter delays delivering new connections
8106 to Squid until a full HTTP request has been received.
8107 See the accf_http(9) man page for details.
8109 The 'dataready' filter delays delivering new connections
8110 to Squid until there is some data to process.
8111 See the accf_dataready(9) man page for details.
8115 The 'data' filter delays delivering of new connections
8116 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
8117 You may optionally specify a number of seconds to wait by
8118 'data=N' where N is the number of seconds. Defaults to 30
8119 if not specified. See the tcp(7) man page for details.
8122 accept_filter httpready
8127 NAME: client_ip_max_connections
8129 LOC: Config.client_ip_max_connections
8131 DEFAULT_DOC: No limit.
8133 Set an absolute limit on the number of connections a single
8134 client IP can use. Any more than this and Squid will begin to drop
8135 new connections from the client until it closes some links.
8137 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
8138 connections from the client. For finer control use the ACL access controls.
8140 Requires client_db to be enabled (the default).
8142 WARNING: This may noticably slow down traffic received via external proxies
8143 or NAT devices and cause them to rebound error messages back to their clients.
8146 NAME: tcp_recv_bufsize
8150 DEFAULT_DOC: Use operating system TCP defaults.
8151 LOC: Config.tcpRcvBufsz
8153 Size of receive buffer to set for TCP sockets. Probably just
8154 as easy to change your kernel's default.
8155 Omit from squid.conf to use the default buffer size.
8160 -----------------------------------------------------------------------------
8167 LOC: Adaptation::Icap::TheConfig.onoff
8170 If you want to enable the ICAP module support, set this to on.
8173 NAME: icap_connect_timeout
8176 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
8179 This parameter specifies how long to wait for the TCP connect to
8180 the requested ICAP server to complete before giving up and either
8181 terminating the HTTP transaction or bypassing the failure.
8183 The default for optional services is peer_connect_timeout.
8184 The default for essential services is connect_timeout.
8185 If this option is explicitly set, its value applies to all services.
8188 NAME: icap_io_timeout
8192 DEFAULT_DOC: Use read_timeout.
8193 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
8196 This parameter specifies how long to wait for an I/O activity on
8197 an established, active ICAP connection before giving up and
8198 either terminating the HTTP transaction or bypassing the
8202 NAME: icap_service_failure_limit
8203 COMMENT: limit [in memory-depth time-units]
8204 TYPE: icap_service_failure_limit
8206 LOC: Adaptation::Icap::TheConfig
8209 The limit specifies the number of failures that Squid tolerates
8210 when establishing a new TCP connection with an ICAP service. If
8211 the number of failures exceeds the limit, the ICAP service is
8212 not used for new ICAP requests until it is time to refresh its
8215 A negative value disables the limit. Without the limit, an ICAP
8216 service will not be considered down due to connectivity failures
8217 between ICAP OPTIONS requests.
8219 Squid forgets ICAP service failures older than the specified
8220 value of memory-depth. The memory fading algorithm
8221 is approximate because Squid does not remember individual
8222 errors but groups them instead, splitting the option
8223 value into ten time slots of equal length.
8225 When memory-depth is 0 and by default this option has no
8226 effect on service failure expiration.
8228 Squid always forgets failures when updating service settings
8229 using an ICAP OPTIONS transaction, regardless of this option
8233 # suspend service usage after 10 failures in 5 seconds:
8234 icap_service_failure_limit 10 in 5 seconds
8237 NAME: icap_service_revival_delay
8240 LOC: Adaptation::Icap::TheConfig.service_revival_delay
8243 The delay specifies the number of seconds to wait after an ICAP
8244 OPTIONS request failure before requesting the options again. The
8245 failed ICAP service is considered "down" until fresh OPTIONS are
8248 The actual delay cannot be smaller than the hardcoded minimum
8249 delay of 30 seconds.
8252 NAME: icap_preview_enable
8256 LOC: Adaptation::Icap::TheConfig.preview_enable
8259 The ICAP Preview feature allows the ICAP server to handle the
8260 HTTP message by looking only at the beginning of the message body
8261 or even without receiving the body at all. In some environments,
8262 previews greatly speedup ICAP processing.
8264 During an ICAP OPTIONS transaction, the server may tell Squid what
8265 HTTP messages should be previewed and how big the preview should be.
8266 Squid will not use Preview if the server did not request one.
8268 To disable ICAP Preview for all ICAP services, regardless of
8269 individual ICAP server OPTIONS responses, set this option to "off".
8271 icap_preview_enable off
8274 NAME: icap_preview_size
8277 LOC: Adaptation::Icap::TheConfig.preview_size
8279 DEFAULT_DOC: No preview sent.
8281 The default size of preview data to be sent to the ICAP server.
8282 This value might be overwritten on a per server basis by OPTIONS requests.
8285 NAME: icap_206_enable
8289 LOC: Adaptation::Icap::TheConfig.allow206_enable
8292 206 (Partial Content) responses is an ICAP extension that allows the
8293 ICAP agents to optionally combine adapted and original HTTP message
8294 content. The decision to combine is postponed until the end of the
8295 ICAP response. Squid supports Partial Content extension by default.
8297 Activation of the Partial Content extension is negotiated with each
8298 ICAP service during OPTIONS exchange. Most ICAP servers should handle
8299 negotation correctly even if they do not support the extension, but
8300 some might fail. To disable Partial Content support for all ICAP
8301 services and to avoid any negotiation, set this option to "off".
8307 NAME: icap_default_options_ttl
8310 LOC: Adaptation::Icap::TheConfig.default_options_ttl
8313 The default TTL value for ICAP OPTIONS responses that don't have
8314 an Options-TTL header.
8317 NAME: icap_persistent_connections
8321 LOC: Adaptation::Icap::TheConfig.reuse_connections
8324 Whether or not Squid should use persistent connections to
8328 NAME: adaptation_send_client_ip icap_send_client_ip
8330 IFDEF: USE_ADAPTATION
8332 LOC: Adaptation::Config::send_client_ip
8335 If enabled, Squid shares HTTP client IP information with adaptation
8336 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
8337 For eCAP, Squid sets the libecap::metaClientIp transaction option.
8339 See also: adaptation_uses_indirect_client
8342 NAME: adaptation_send_username icap_send_client_username
8344 IFDEF: USE_ADAPTATION
8346 LOC: Adaptation::Config::send_username
8349 This sends authenticated HTTP client username (if available) to
8350 the adaptation service.
8352 For ICAP, the username value is encoded based on the
8353 icap_client_username_encode option and is sent using the header
8354 specified by the icap_client_username_header option.
8357 NAME: icap_client_username_header
8360 LOC: Adaptation::Icap::TheConfig.client_username_header
8361 DEFAULT: X-Client-Username
8363 ICAP request header name to use for adaptation_send_username.
8366 NAME: icap_client_username_encode
8370 LOC: Adaptation::Icap::TheConfig.client_username_encode
8373 Whether to base64 encode the authenticated client username.
8377 TYPE: icap_service_type
8379 LOC: Adaptation::Icap::TheConfig
8382 Defines a single ICAP service using the following format:
8384 icap_service id vectoring_point uri [option ...]
8387 an opaque identifier or name which is used to direct traffic to
8388 this specific service. Must be unique among all adaptation
8389 services in squid.conf.
8391 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8392 This specifies at which point of transaction processing the
8393 ICAP service should be activated. *_postcache vectoring points
8394 are not yet supported.
8396 uri: icap://servername:port/servicepath
8397 ICAP server and service location.
8398 icaps://servername:port/servicepath
8399 The "icap:" URI scheme is used for traditional ICAP server and
8400 service location (default port is 1344, connections are not
8401 encrypted). The "icaps:" URI scheme is for Secure ICAP
8402 services that use SSL/TLS-encrypted ICAP connections (by
8403 default, on port 11344).
8405 ICAP does not allow a single service to handle both REQMOD and RESPMOD
8406 transactions. Squid does not enforce that requirement. You can specify
8407 services with the same service_url and different vectoring_points. You
8408 can even specify multiple identical services as long as their
8409 service_names differ.
8411 To activate a service, use the adaptation_access directive. To group
8412 services, use adaptation_service_chain and adaptation_service_set.
8414 Service options are separated by white space. ICAP services support
8415 the following name=value options:
8418 If set to 'on' or '1', the ICAP service is treated as
8419 optional. If the service cannot be reached or malfunctions,
8420 Squid will try to ignore any errors and process the message as
8421 if the service was not enabled. No all ICAP errors can be
8422 bypassed. If set to 0, the ICAP service is treated as
8423 essential and all ICAP errors will result in an error page
8424 returned to the HTTP client.
8426 Bypass is off by default: services are treated as essential.
8429 If set to 'on' or '1', the ICAP service is allowed to
8430 dynamically change the current message adaptation plan by
8431 returning a chain of services to be used next. The services
8432 are specified using the X-Next-Services ICAP response header
8433 value, formatted as a comma-separated list of service names.
8434 Each named service should be configured in squid.conf. Other
8435 services are ignored. An empty X-Next-Services value results
8436 in an empty plan which ends the current adaptation.
8438 Dynamic adaptation plan may cross or cover multiple supported
8439 vectoring points in their natural processing order.
8441 Routing is not allowed by default: the ICAP X-Next-Services
8442 response header is ignored.
8445 Only has effect on split-stack systems. The default on those systems
8446 is to use IPv4-only connections. When set to 'on' this option will
8447 make Squid use IPv6-only connections to contact this ICAP service.
8449 on-overload=block|bypass|wait|force
8450 If the service Max-Connections limit has been reached, do
8451 one of the following for each new ICAP transaction:
8452 * block: send an HTTP error response to the client
8453 * bypass: ignore the "over-connected" ICAP service
8454 * wait: wait (in a FIFO queue) for an ICAP connection slot
8455 * force: proceed, ignoring the Max-Connections limit
8457 In SMP mode with N workers, each worker assumes the service
8458 connection limit is Max-Connections/N, even though not all
8459 workers may use a given service.
8461 The default value is "bypass" if service is bypassable,
8462 otherwise it is set to "wait".
8466 Use the given number as the Max-Connections limit, regardless
8467 of the Max-Connections value given by the service, if any.
8469 ==== SSL / ICAPS / TLS OPTIONS ====
8471 These options are used for Secure ICAP (icaps://....) services only.
8473 sslcert=/path/to/ssl/certificate
8474 A client SSL certificate to use when connecting to
8477 sslkey=/path/to/ssl/key
8478 The private SSL key corresponding to sslcert above.
8479 If 'sslkey' is not specified 'sslcert' is assumed to
8480 reference a combined file containing both the
8481 certificate and the key.
8483 sslcipher=... The list of valid SSL ciphers to use when connecting
8484 to this icap server.
8487 The minimum TLS protocol version to permit. To control
8488 SSLv3 use the ssloptions= parameter.
8489 Supported Values: 1.0 (default), 1.1, 1.2
8491 ssloptions=... Specify various SSL implementation options:
8493 NO_SSLv3 Disallow the use of SSLv3
8494 NO_TLSv1 Disallow the use of TLSv1.0
8495 NO_TLSv1_1 Disallow the use of TLSv1.1
8496 NO_TLSv1_2 Disallow the use of TLSv1.2
8498 Always create a new key when using
8499 temporary/ephemeral DH key exchanges
8500 ALL Enable various bug workarounds
8501 suggested as "harmless" by OpenSSL
8502 Be warned that this reduces SSL/TLS
8503 strength to some attacks.
8505 See the OpenSSL SSL_CTX_set_options documentation for a
8508 sslcafile=... A file containing additional CA certificates to use
8509 when verifying the icap server certificate.
8511 sslcapath=... A directory containing additional CA certificates to
8512 use when verifying the icap server certificate.
8514 sslcrlfile=... A certificate revocation list file to use when
8515 verifying the icap server certificate.
8517 sslflags=... Specify various flags modifying the SSL implementation:
8520 Accept certificates even if they fail to
8523 Don't use the default CA list built in
8526 Don't verify the icap server certificate
8527 matches the server name
8529 ssldomain= The icap server name as advertised in it's certificate.
8530 Used for verifying the correctness of the received icap
8531 server certificate. If not specified the icap server
8532 hostname extracted from ICAP URI will be used.
8534 Older icap_service format without optional named parameters is
8535 deprecated but supported for backward compatibility.
8538 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8539 icap_service svcLogger reqmod_precache icaps://icap2.mydomain.net:11344/reqmod routing=on
8543 TYPE: icap_class_type
8548 This deprecated option was documented to define an ICAP service
8549 chain, even though it actually defined a set of similar, redundant
8550 services, and the chains were not supported.
8552 To define a set of redundant services, please use the
8553 adaptation_service_set directive. For service chains, use
8554 adaptation_service_chain.
8558 TYPE: icap_access_type
8563 This option is deprecated. Please use adaptation_access, which
8564 has the same ICAP functionality, but comes with better
8565 documentation, and eCAP support.
8570 -----------------------------------------------------------------------------
8577 LOC: Adaptation::Ecap::TheConfig.onoff
8580 Controls whether eCAP support is enabled.
8584 TYPE: ecap_service_type
8586 LOC: Adaptation::Ecap::TheConfig
8589 Defines a single eCAP service
8591 ecap_service id vectoring_point uri [option ...]
8594 an opaque identifier or name which is used to direct traffic to
8595 this specific service. Must be unique among all adaptation
8596 services in squid.conf.
8598 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8599 This specifies at which point of transaction processing the
8600 eCAP service should be activated. *_postcache vectoring points
8601 are not yet supported.
8603 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8604 Squid uses the eCAP service URI to match this configuration
8605 line with one of the dynamically loaded services. Each loaded
8606 eCAP service must have a unique URI. Obtain the right URI from
8607 the service provider.
8609 To activate a service, use the adaptation_access directive. To group
8610 services, use adaptation_service_chain and adaptation_service_set.
8612 Service options are separated by white space. eCAP services support
8613 the following name=value options:
8616 If set to 'on' or '1', the eCAP service is treated as optional.
8617 If the service cannot be reached or malfunctions, Squid will try
8618 to ignore any errors and process the message as if the service
8619 was not enabled. No all eCAP errors can be bypassed.
8620 If set to 'off' or '0', the eCAP service is treated as essential
8621 and all eCAP errors will result in an error page returned to the
8624 Bypass is off by default: services are treated as essential.
8627 If set to 'on' or '1', the eCAP service is allowed to
8628 dynamically change the current message adaptation plan by
8629 returning a chain of services to be used next.
8631 Dynamic adaptation plan may cross or cover multiple supported
8632 vectoring points in their natural processing order.
8634 Routing is not allowed by default.
8636 Older ecap_service format without optional named parameters is
8637 deprecated but supported for backward compatibility.
8641 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8642 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8645 NAME: loadable_modules
8647 IFDEF: USE_LOADABLE_MODULES
8648 LOC: Config.loadable_module_names
8651 Instructs Squid to load the specified dynamic module(s) or activate
8652 preloaded module(s).
8654 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8658 MESSAGE ADAPTATION OPTIONS
8659 -----------------------------------------------------------------------------
8662 NAME: adaptation_service_set
8663 TYPE: adaptation_service_set_type
8664 IFDEF: USE_ADAPTATION
8669 Configures an ordered set of similar, redundant services. This is
8670 useful when hot standby or backup adaptation servers are available.
8672 adaptation_service_set set_name service_name1 service_name2 ...
8674 The named services are used in the set declaration order. The first
8675 applicable adaptation service from the set is used first. The next
8676 applicable service is tried if and only if the transaction with the
8677 previous service fails and the message waiting to be adapted is still
8680 When adaptation starts, broken services are ignored as if they were
8681 not a part of the set. A broken service is a down optional service.
8683 The services in a set must be attached to the same vectoring point
8684 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8686 If all services in a set are optional then adaptation failures are
8687 bypassable. If all services in the set are essential, then a
8688 transaction failure with one service may still be retried using
8689 another service from the set, but when all services fail, the master
8690 transaction fails as well.
8692 A set may contain a mix of optional and essential services, but that
8693 is likely to lead to surprising results because broken services become
8694 ignored (see above), making previously bypassable failures fatal.
8695 Technically, it is the bypassability of the last failed service that
8698 See also: adaptation_access adaptation_service_chain
8701 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8702 adaptation service_set svcLogger loggerLocal loggerRemote
8705 NAME: adaptation_service_chain
8706 TYPE: adaptation_service_chain_type
8707 IFDEF: USE_ADAPTATION
8712 Configures a list of complementary services that will be applied
8713 one-by-one, forming an adaptation chain or pipeline. This is useful
8714 when Squid must perform different adaptations on the same message.
8716 adaptation_service_chain chain_name service_name1 svc_name2 ...
8718 The named services are used in the chain declaration order. The first
8719 applicable adaptation service from the chain is used first. The next
8720 applicable service is applied to the successful adaptation results of
8721 the previous service in the chain.
8723 When adaptation starts, broken services are ignored as if they were
8724 not a part of the chain. A broken service is a down optional service.
8726 Request satisfaction terminates the adaptation chain because Squid
8727 does not currently allow declaration of RESPMOD services at the
8728 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8730 The services in a chain must be attached to the same vectoring point
8731 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8733 A chain may contain a mix of optional and essential services. If an
8734 essential adaptation fails (or the failure cannot be bypassed for
8735 other reasons), the master transaction fails. Otherwise, the failure
8736 is bypassed as if the failed adaptation service was not in the chain.
8738 See also: adaptation_access adaptation_service_set
8741 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8744 NAME: adaptation_access
8745 TYPE: adaptation_access_type
8746 IFDEF: USE_ADAPTATION
8749 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8751 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8753 adaptation_access service_name allow|deny [!]aclname...
8754 adaptation_access set_name allow|deny [!]aclname...
8756 At each supported vectoring point, the adaptation_access
8757 statements are processed in the order they appear in this
8758 configuration file. Statements pointing to the following services
8759 are ignored (i.e., skipped without checking their ACL):
8761 - services serving different vectoring points
8762 - "broken-but-bypassable" services
8763 - "up" services configured to ignore such transactions
8764 (e.g., based on the ICAP Transfer-Ignore header).
8766 When a set_name is used, all services in the set are checked
8767 using the same rules, to find the first applicable one. See
8768 adaptation_service_set for details.
8770 If an access list is checked and there is a match, the
8771 processing stops: For an "allow" rule, the corresponding
8772 adaptation service is used for the transaction. For a "deny"
8773 rule, no adaptation service is activated.
8775 It is currently not possible to apply more than one adaptation
8776 service at the same vectoring point to the same HTTP transaction.
8778 See also: icap_service and ecap_service
8781 adaptation_access service_1 allow all
8784 NAME: adaptation_service_iteration_limit
8786 IFDEF: USE_ADAPTATION
8787 LOC: Adaptation::Config::service_iteration_limit
8790 Limits the number of iterations allowed when applying adaptation
8791 services to a message. If your longest adaptation set or chain
8792 may have more than 16 services, increase the limit beyond its
8793 default value of 16. If detecting infinite iteration loops sooner
8794 is critical, make the iteration limit match the actual number
8795 of services in your longest adaptation set or chain.
8797 Infinite adaptation loops are most likely with routing services.
8799 See also: icap_service routing=1
8802 NAME: adaptation_masterx_shared_names
8804 IFDEF: USE_ADAPTATION
8805 LOC: Adaptation::Config::masterx_shared_name
8808 For each master transaction (i.e., the HTTP request and response
8809 sequence, including all related ICAP and eCAP exchanges), Squid
8810 maintains a table of metadata. The table entries are (name, value)
8811 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8812 with the master transaction.
8814 This option specifies the table entry names that Squid must accept
8815 from and forward to the adaptation transactions.
8817 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8818 shared table by returning an ICAP header field with a name
8819 specified in adaptation_masterx_shared_names.
8821 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8822 shared table by implementing the libecap::visitEachOption() API
8823 to provide an option with a name specified in
8824 adaptation_masterx_shared_names.
8826 Squid will store and forward the set entry to subsequent adaptation
8827 transactions within the same master transaction scope.
8829 Only one shared entry name is supported at this time.
8832 # share authentication information among ICAP services
8833 adaptation_masterx_shared_names X-Subscriber-ID
8836 NAME: adaptation_meta
8838 IFDEF: USE_ADAPTATION
8839 LOC: Adaptation::Config::metaHeaders
8842 This option allows Squid administrator to add custom ICAP request
8843 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8844 Use it to pass custom authentication tokens and other
8845 transaction-state related meta information to an ICAP/eCAP service.
8847 The addition of a meta header is ACL-driven:
8848 adaptation_meta name value [!]aclname ...
8850 Processing for a given header name stops after the first ACL list match.
8851 Thus, it is impossible to add two headers with the same name. If no ACL
8852 lists match for a given header name, no such header is added. For
8855 # do not debug transactions except for those that need debugging
8856 adaptation_meta X-Debug 1 needs_debugging
8858 # log all transactions except for those that must remain secret
8859 adaptation_meta X-Log 1 !keep_secret
8861 # mark transactions from users in the "G 1" group
8862 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8864 The "value" parameter may be a regular squid.conf token or a "double
8865 quoted string". Within the quoted string, use backslash (\) to escape
8866 any character, which is currently only useful for escaping backslashes
8867 and double quotes. For example,
8868 "this string has one backslash (\\) and two \"quotes\""
8870 Used adaptation_meta header values may be logged via %note
8871 logformat code. If multiple adaptation_meta headers with the same name
8872 are used during master transaction lifetime, the header values are
8873 logged in the order they were used and duplicate values are ignored
8874 (only the first repeated value will be logged).
8880 LOC: Adaptation::Icap::TheConfig.repeat
8881 DEFAULT_IF_NONE: deny all
8883 This ACL determines which retriable ICAP transactions are
8884 retried. Transactions that received a complete ICAP response
8885 and did not have to consume or produce HTTP bodies to receive
8886 that response are usually retriable.
8888 icap_retry allow|deny [!]aclname ...
8890 Squid automatically retries some ICAP I/O timeouts and errors
8891 due to persistent connection race conditions.
8893 See also: icap_retry_limit
8896 NAME: icap_retry_limit
8899 LOC: Adaptation::Icap::TheConfig.repeat_limit
8901 DEFAULT_DOC: No retries are allowed.
8903 Limits the number of retries allowed.
8905 Communication errors due to persistent connection race
8906 conditions are unavoidable, automatically retried, and do not
8907 count against this limit.
8909 See also: icap_retry
8915 -----------------------------------------------------------------------------
8918 NAME: check_hostnames
8921 LOC: Config.onoff.check_hostnames
8923 For security and stability reasons Squid can check
8924 hostnames for Internet standard RFC compliance. If you want
8925 Squid to perform these checks turn this directive on.
8928 NAME: allow_underscore
8931 LOC: Config.onoff.allow_underscore
8933 Underscore characters is not strictly allowed in Internet hostnames
8934 but nevertheless used by many sites. Set this to off if you want
8935 Squid to be strict about the standard.
8936 This check is performed only when check_hostnames is set to on.
8939 NAME: dns_retransmit_interval
8942 LOC: Config.Timeout.idns_retransmit
8944 Initial retransmit interval for DNS queries. The interval is
8945 doubled each time all configured DNS servers have been tried.
8951 LOC: Config.Timeout.idns_query
8953 DNS Query timeout. If no response is received to a DNS query
8954 within this time all DNS servers for the queried domain
8955 are assumed to be unavailable.
8958 NAME: dns_packet_max
8960 DEFAULT_DOC: EDNS disabled
8962 LOC: Config.dns.packet_max
8964 Maximum number of bytes packet size to advertise via EDNS.
8965 Set to "none" to disable EDNS large packet support.
8967 For legacy reasons DNS UDP replies will default to 512 bytes which
8968 is too small for many responses. EDNS provides a means for Squid to
8969 negotiate receiving larger responses back immediately without having
8970 to failover with repeat requests. Responses larger than this limit
8971 will retain the old behaviour of failover to TCP DNS.
8973 Squid has no real fixed limit internally, but allowing packet sizes
8974 over 1500 bytes requires network jumbogram support and is usually not
8977 WARNING: The RFC also indicates that some older resolvers will reply
8978 with failure of the whole request if the extension is added. Some
8979 resolvers have already been identified which will reply with mangled
8980 EDNS response on occasion. Usually in response to many-KB jumbogram
8981 sizes being advertised by Squid.
8982 Squid will currently treat these both as an unable-to-resolve domain
8983 even if it would be resolvable without EDNS.
8990 DEFAULT_DOC: Search for single-label domain names is disabled.
8991 LOC: Config.onoff.res_defnames
8993 Normally the RES_DEFNAMES resolver option is disabled
8994 (see res_init(3)). This prevents caches in a hierarchy
8995 from interpreting single-component hostnames locally. To allow
8996 Squid to handle single-component names, enable this option.
8999 NAME: dns_multicast_local
9003 DEFAULT_DOC: Search for .local and .arpa names is disabled.
9004 LOC: Config.onoff.dns_mdns
9006 When set to on, Squid sends multicast DNS lookups on the local
9007 network for domains ending in .local and .arpa.
9008 This enables local servers and devices to be contacted in an
9009 ad-hoc or zero-configuration network environment.
9012 NAME: dns_nameservers
9015 DEFAULT_DOC: Use operating system definitions
9016 LOC: Config.dns_nameservers
9018 Use this if you want to specify a list of DNS name servers
9019 (IP addresses) to use instead of those given in your
9020 /etc/resolv.conf file.
9022 On Windows platforms, if no value is specified here or in
9023 the /etc/resolv.conf file, the list of DNS name servers are
9024 taken from the Windows registry, both static and dynamic DHCP
9025 configurations are supported.
9027 Example: dns_nameservers 10.0.0.1 192.172.0.4
9032 DEFAULT: @DEFAULT_HOSTS@
9033 LOC: Config.etcHostsPath
9035 Location of the host-local IP name-address associations
9036 database. Most Operating Systems have such a file on different
9038 - Un*X & Linux: /etc/hosts
9039 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
9040 (%SystemRoot% value install default is c:\winnt)
9041 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
9042 (%SystemRoot% value install default is c:\windows)
9043 - Windows 9x/Me: %windir%\hosts
9044 (%windir% value is usually c:\windows)
9045 - Cygwin: /etc/hosts
9047 The file contains newline-separated definitions, in the
9048 form ip_address_in_dotted_form name [name ...] names are
9049 whitespace-separated. Lines beginning with an hash (#)
9050 character are comments.
9052 The file is checked at startup and upon configuration.
9053 If set to 'none', it won't be checked.
9054 If append_domain is used, that domain will be added to
9055 domain-local (i.e. not containing any dot character) host
9061 LOC: Config.appendDomain
9063 DEFAULT_DOC: Use operating system definitions
9065 Appends local domain name to hostnames without any dots in
9066 them. append_domain must begin with a period.
9068 Be warned there are now Internet names with no dots in
9069 them using only top-domain names, so setting this may
9070 cause some Internet sites to become unavailable.
9073 append_domain .yourdomain.com
9076 NAME: ignore_unknown_nameservers
9078 LOC: Config.onoff.ignore_unknown_nameservers
9081 By default Squid checks that DNS responses are received
9082 from the same IP addresses they are sent to. If they
9083 don't match, Squid ignores the response and writes a warning
9084 message to cache.log. You can allow responses from unknown
9085 nameservers by setting this option to 'off'.
9091 LOC: Config.dns.v4_first
9093 With the IPv6 Internet being as fast or faster than IPv4 Internet
9094 for most networks Squid prefers to contact websites over IPv6.
9096 This option reverses the order of preference to make Squid contact
9097 dual-stack websites over IPv4 first. Squid will still perform both
9098 IPv6 and IPv4 DNS lookups before connecting.
9101 This option will restrict the situations under which IPv6
9102 connectivity is used (and tested). Hiding network problems
9103 which would otherwise be detected and warned about.
9107 COMMENT: (number of entries)
9110 LOC: Config.ipcache.size
9112 Maximum number of DNS IP cache entries.
9119 LOC: Config.ipcache.low
9126 LOC: Config.ipcache.high
9128 The size, low-, and high-water marks for the IP cache.
9131 NAME: fqdncache_size
9132 COMMENT: (number of entries)
9135 LOC: Config.fqdncache.size
9137 Maximum number of FQDN cache entries.
9142 -----------------------------------------------------------------------------
9145 NAME: configuration_includes_quoted_values
9147 TYPE: configuration_includes_quoted_values
9149 LOC: ConfigParser::RecognizeQuotedValues
9151 If set, Squid will recognize each "quoted string" after a configuration
9152 directive as a single parameter. The quotes are stripped before the
9153 parameter value is interpreted or used.
9154 See "Values with spaces, quotes, and other special characters"
9155 section for more details.
9162 LOC: Config.onoff.mem_pools
9164 If set, Squid will keep pools of allocated (but unused) memory
9165 available for future use. If memory is a premium on your
9166 system and you believe your malloc library outperforms Squid
9167 routines, disable this.
9170 NAME: memory_pools_limit
9174 LOC: Config.MemPools.limit
9176 Used only with memory_pools on:
9177 memory_pools_limit 50 MB
9179 If set to a non-zero value, Squid will keep at most the specified
9180 limit of allocated (but unused) memory in memory pools. All free()
9181 requests that exceed this limit will be handled by your malloc
9182 library. Squid does not pre-allocate any memory, just safe-keeps
9183 objects that otherwise would be free()d. Thus, it is safe to set
9184 memory_pools_limit to a reasonably high value even if your
9185 configuration will use less memory.
9187 If set to none, Squid will keep all memory it can. That is, there
9188 will be no limit on the total amount of memory used for safe-keeping.
9190 To disable memory allocation optimization, do not set
9191 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
9193 An overhead for maintaining memory pools is not taken into account
9194 when the limit is checked. This overhead is close to four bytes per
9195 object kept. However, pools may actually _save_ memory because of
9196 reduced memory thrashing in your malloc library.
9200 COMMENT: on|off|transparent|truncate|delete
9203 LOC: opt_forwarded_for
9205 If set to "on", Squid will append your client's IP address
9206 in the HTTP requests it forwards. By default it looks like:
9208 X-Forwarded-For: 192.1.2.3
9210 If set to "off", it will appear as
9212 X-Forwarded-For: unknown
9214 If set to "transparent", Squid will not alter the
9215 X-Forwarded-For header in any way.
9217 If set to "delete", Squid will delete the entire
9218 X-Forwarded-For header.
9220 If set to "truncate", Squid will remove all existing
9221 X-Forwarded-For entries, and place the client IP as the sole entry.
9224 NAME: cachemgr_passwd
9225 TYPE: cachemgrpasswd
9227 DEFAULT_DOC: No password. Actions which require password are denied.
9228 LOC: Config.passwd_list
9230 Specify passwords for cachemgr operations.
9232 Usage: cachemgr_passwd password action action ...
9234 Some valid actions are (see cache manager menu for a full list):
9274 * Indicates actions which will not be performed without a
9275 valid password, others can be performed if not listed here.
9277 To disable an action, set the password to "disable".
9278 To allow performing an action without a password, set the
9281 Use the keyword "all" to set the same password for all actions.
9284 cachemgr_passwd secret shutdown
9285 cachemgr_passwd lesssssssecret info stats/objects
9286 cachemgr_passwd disable all
9293 LOC: Config.onoff.client_db
9295 If you want to disable collecting per-client statistics,
9296 turn off client_db here.
9299 NAME: refresh_all_ims
9303 LOC: Config.onoff.refresh_all_ims
9305 When you enable this option, squid will always check
9306 the origin server for an update when a client sends an
9307 If-Modified-Since request. Many browsers use IMS
9308 requests when the user requests a reload, and this
9309 ensures those clients receive the latest version.
9311 By default (off), squid may return a Not Modified response
9312 based on the age of the cached version.
9315 NAME: reload_into_ims
9316 IFDEF: USE_HTTP_VIOLATIONS
9320 LOC: Config.onoff.reload_into_ims
9322 When you enable this option, client no-cache or ``reload''
9323 requests will be changed to If-Modified-Since requests.
9324 Doing this VIOLATES the HTTP standard. Enabling this
9325 feature could make you liable for problems which it
9328 see also refresh_pattern for a more selective approach.
9331 NAME: connect_retries
9333 LOC: Config.connect_retries
9335 DEFAULT_DOC: Do not retry failed connections.
9337 This sets the maximum number of connection attempts made for each
9338 TCP connection. The connect_retries attempts must all still
9339 complete within the connection timeout period.
9341 The default is not to re-try if the first connection attempt fails.
9342 The (not recommended) maximum is 10 tries.
9344 A warning message will be generated if it is set to a too-high
9345 value and the configured value will be over-ridden.
9347 Note: These re-tries are in addition to forward_max_tries
9348 which limit how many different addresses may be tried to find
9352 NAME: retry_on_error
9354 LOC: Config.retry.onerror
9357 If set to ON Squid will automatically retry requests when
9358 receiving an error response with status 403 (Forbidden),
9359 500 (Internal Error), 501 or 503 (Service not available).
9360 Status 502 and 504 (Gateway errors) are always retried.
9362 This is mainly useful if you are in a complex cache hierarchy to
9363 work around access control errors.
9365 NOTE: This retry will attempt to find another working destination.
9366 Which is different from the server which just failed.
9369 NAME: as_whois_server
9371 LOC: Config.as_whois_server
9372 DEFAULT: whois.ra.net
9374 WHOIS server to query for AS numbers. NOTE: AS numbers are
9375 queried only when Squid starts up, not for every request.
9380 LOC: Config.onoff.offline
9383 Enable this option and Squid will never try to validate cached
9387 NAME: uri_whitespace
9388 TYPE: uri_whitespace
9389 LOC: Config.uri_whitespace
9392 What to do with requests that have whitespace characters in the
9395 strip: The whitespace characters are stripped out of the URL.
9396 This is the behavior recommended by RFC2396 and RFC3986
9397 for tolerant handling of generic URI.
9398 NOTE: This is one difference between generic URI and HTTP URLs.
9400 deny: The request is denied. The user receives an "Invalid
9402 This is the behaviour recommended by RFC2616 for safe
9403 handling of HTTP request URL.
9405 allow: The request is allowed and the URI is not changed. The
9406 whitespace characters remain in the URI. Note the
9407 whitespace is passed to redirector processes if they
9409 Note this may be considered a violation of RFC2616
9410 request parsing where whitespace is prohibited in the
9413 encode: The request is allowed and the whitespace characters are
9414 encoded according to RFC1738.
9416 chop: The request is allowed and the URI is chopped at the
9420 NOTE the current Squid implementation of encode and chop violates
9421 RFC2616 by not using a 301 redirect after altering the URL.
9426 LOC: Config.chroot_dir
9429 Specifies a directory where Squid should do a chroot() while
9430 initializing. This also causes Squid to fully drop root
9431 privileges after initializing. This means, for example, if you
9432 use a HTTP port less than 1024 and try to reconfigure, you may
9433 get an error saying that Squid can not open the port.
9436 NAME: balance_on_multiple_ip
9438 LOC: Config.onoff.balance_on_multiple_ip
9441 Modern IP resolvers in squid sort lookup results by preferred access.
9442 By default squid will use these IP in order and only rotates to
9443 the next listed when the most preffered fails.
9445 Some load balancing servers based on round robin DNS have been
9446 found not to preserve user session state across requests
9447 to different IP addresses.
9449 Enabling this directive Squid rotates IP's per request.
9452 NAME: pipeline_prefetch
9453 TYPE: pipelinePrefetch
9454 LOC: Config.pipeline_max_prefetch
9456 DEFAULT_DOC: Do not pre-parse pipelined requests.
9458 HTTP clients may send a pipeline of 1+N requests to Squid using a
9459 single connection, without waiting for Squid to respond to the first
9460 of those requests. This option limits the number of concurrent
9461 requests Squid will try to handle in parallel. If set to N, Squid
9462 will try to receive and process up to 1+N requests on the same
9463 connection concurrently.
9465 Defaults to 0 (off) for bandwidth management and access logging
9468 NOTE: pipelining requires persistent connections to clients.
9470 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
9473 NAME: high_response_time_warning
9476 LOC: Config.warnings.high_rptm
9478 DEFAULT_DOC: disabled.
9480 If the one-minute median response time exceeds this value,
9481 Squid prints a WARNING with debug level 0 to get the
9482 administrators attention. The value is in milliseconds.
9485 NAME: high_page_fault_warning
9487 LOC: Config.warnings.high_pf
9489 DEFAULT_DOC: disabled.
9491 If the one-minute average page fault rate exceeds this
9492 value, Squid prints a WARNING with debug level 0 to get
9493 the administrators attention. The value is in page faults
9497 NAME: high_memory_warning
9499 LOC: Config.warnings.high_memory
9500 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
9502 DEFAULT_DOC: disabled.
9504 If the memory usage (as determined by gnumalloc, if available and used)
9505 exceeds this amount, Squid prints a WARNING with debug level 0 to get
9506 the administrators attention.
9508 # TODO: link high_memory_warning to mempools?
9510 NAME: sleep_after_fork
9511 COMMENT: (microseconds)
9513 LOC: Config.sleep_after_fork
9516 When this is set to a non-zero value, the main Squid process
9517 sleeps the specified number of microseconds after a fork()
9518 system call. This sleep may help the situation where your
9519 system reports fork() failures due to lack of (virtual)
9520 memory. Note, however, if you have a lot of child
9521 processes, these sleep delays will add up and your
9522 Squid will not service requests for some amount of time
9523 until all the child processes have been started.
9524 On Windows value less then 1000 (1 milliseconds) are
9528 NAME: windows_ipaddrchangemonitor
9529 IFDEF: _SQUID_WINDOWS_
9533 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9535 On Windows Squid by default will monitor IP address changes and will
9536 reconfigure itself after any detected event. This is very useful for
9537 proxies connected to internet with dial-up interfaces.
9538 In some cases (a Proxy server acting as VPN gateway is one) it could be
9539 desiderable to disable this behaviour setting this to 'off'.
9540 Note: after changing this, Squid service must be restarted.
9545 IFDEF: USE_SQUID_EUI
9547 LOC: Eui::TheConfig.euiLookup
9549 Whether to lookup the EUI or MAC address of a connected client.
9552 NAME: max_filedescriptors max_filedesc
9555 DEFAULT_DOC: Use operating system limits set by ulimit.
9556 LOC: Config.max_filedescriptors
9558 Reduce the maximum number of filedescriptors supported below
9559 the usual operating system defaults.
9561 Remove from squid.conf to inherit the current ulimit setting.
9563 Note: Changing this requires a restart of Squid. Also
9564 not all I/O types supports large values (eg on Windows).
9567 NAME: force_request_body_continuation
9569 LOC: Config.accessList.forceRequestBodyContinuation
9571 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
9573 This option controls how Squid handles data upload requests from HTTP
9574 and FTP agents that require a "Please Continue" control message response
9575 to actually send the request body to Squid. It is mostly useful in
9576 adaptation environments.
9578 When Squid receives an HTTP request with an "Expect: 100-continue"
9579 header or an FTP upload command (e.g., STOR), Squid normally sends the
9580 request headers or FTP command information to an adaptation service (or
9581 peer) and waits for a response. Most adaptation services (and some
9582 broken peers) may not respond to Squid at that stage because they may
9583 decide to wait for the HTTP request body or FTP data transfer. However,
9584 that request body or data transfer may never come because Squid has not
9585 responded with the HTTP 100 or FTP 150 (Please Continue) control message
9586 to the request sender yet!
9588 An allow match tells Squid to respond with the HTTP 100 or FTP 150
9589 (Please Continue) control message on its own, before forwarding the
9590 request to an adaptation service or peer. Such a response usually forces
9591 the request sender to proceed with sending the body. A deny match tells
9592 Squid to delay that control response until the origin server confirms
9593 that the request body is needed. Delaying is the default behavior.