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 3.5
152 NAME: hierarchy_stoplist
155 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
161 Remove this line. Use acls with access_log directives to control access logging
167 Remove this line. Use acls with icap_log directives to control icap logging
170 # Options Removed in 3.3
171 NAME: ignore_ims_on_miss
174 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
177 # Options Removed in 3.2
178 NAME: dns_v4_fallback
181 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
184 NAME: emulate_httpd_log
187 Replace this with an access_log directive using the format 'common' or 'combined'.
193 Use a regular access.log with ACL limiting it to MISS events.
199 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
202 NAME: ignore_expect_100
205 Remove this line. The HTTP/1.1 feature is now fully supported by default.
211 Remove this option from your config. To log FQDN use %>A in the log format.
214 NAME: log_ip_on_direct
217 Remove this option from your config. To log server or peer names use %<A in the log format.
220 NAME: maximum_single_addr_tries
223 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
226 NAME: referer_log referrer_log
229 Replace this with an access_log directive using the format 'referrer'.
235 Remove this line. The feature is supported by default in storage types where update is implemented.
238 NAME: url_rewrite_concurrency
241 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
247 Replace this with an access_log directive using the format 'useragent'.
250 # Options Removed in 3.1
254 Remove this line. DNS is no longer tested on startup.
257 NAME: extension_methods
260 Remove this line. All valid methods for HTTP are accepted by default.
263 # 2.7 Options Removed/Replaced in 3.2
268 # 2.7 Options Removed/Replaced in 3.1
276 Remove this line. HTTP/1.1 is supported by default.
279 NAME: upgrade_http0.9
282 Remove this line. ICY/1.0 streaming protocol is supported by default.
285 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
288 Alter these entries. Use the qos_flows directive instead.
291 # Options Removed in 3.0
295 Since squid-3.0 replace with request_header_access or reply_header_access
296 depending on whether you wish to match client requests or server replies.
299 NAME: httpd_accel_no_pmtu_disc
302 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
305 NAME: wais_relay_host
308 Replace this line with 'cache_peer' configuration.
311 NAME: wais_relay_port
314 Replace this line with 'cache_peer' configuration.
318 OPTIONS FOR AUTHENTICATION
319 -----------------------------------------------------------------------------
328 This is used to define parameters for the various authentication
329 schemes supported by Squid.
331 format: auth_param scheme parameter [setting]
333 The order in which authentication schemes are presented to the client is
334 dependent on the order the scheme first appears in config file. IE
335 has a bug (it's not RFC 2617 compliant) in that it will use the basic
336 scheme if basic is the first entry presented, even if more secure
337 schemes are presented. For now use the order in the recommended
338 settings section below. If other browsers have difficulties (don't
339 recognize the schemes offered even if you are using basic) either
340 put basic first, or disable the other schemes (by commenting out their
343 Once an authentication scheme is fully configured, it can only be
344 shutdown by shutting squid down and restarting. Changes can be made on
345 the fly and activated with a reconfigure. I.E. You can change to a
346 different helper, but not unconfigure the helper completely.
348 Please note that while this directive defines how Squid processes
349 authentication it does not automatically activate authentication.
350 To use authentication you must in addition make use of ACLs based
351 on login name in http_access (proxy_auth, proxy_auth_regex or
352 external with %LOGIN used in the format tag). The browser will be
353 challenged for authentication on the first such acl encountered
354 in http_access processing and will also be re-challenged for new
355 login credentials if the request is being denied by a proxy_auth
358 WARNING: authentication can't be used in a transparently intercepting
359 proxy as the client then thinks it is talking to an origin server and
360 not the proxy. This is a limitation of bending the TCP/IP protocol to
361 transparently intercepting port 80, not a limitation in Squid.
362 Ports flagged 'transparent', 'intercept', or 'tproxy' have
363 authentication disabled.
365 === Parameters common to all schemes. ===
368 Specifies the command for the external authenticator.
370 By default, each authentication scheme is not used unless a
371 program is specified.
373 See http://wiki.squid-cache.org/Features/AddonHelpers for
374 more details on helper operations and creating your own.
377 Specifies a string to be append to request line format for
378 the authentication helper. "Quoted" format values may contain
379 spaces and logformat %macros. In theory, any logformat %macro
380 can be used. In practice, a %macro expands as a dash (-) if
381 the helper request is sent before the required macro
382 information is available to Squid.
384 By default, Squid uses request formats provided in
385 scheme-specific examples below (search for %credentials).
387 The expanded key_extras value is added to the Squid credentials
388 cache and, hence, will affect authentication. It can be used to
389 autenticate different users with identical user names (e.g.,
390 when user authentication depends on http_port).
392 Avoid adding frequently changing information to key_extras. For
393 example, if you add user source IP, and it changes frequently
394 in your environment, then max_user_ip ACL is going to treat
395 every user+IP combination as a unique "user", breaking the ACL
396 and wasting a lot of memory on those user records. It will also
397 force users to authenticate from scratch whenever their IP
401 Specifies the protection scope (aka realm name) which is to be
402 reported to the client for the authentication scheme. It is
403 commonly part of the text the user will see when prompted for
404 their username and password.
406 For Basic the default is "Squid proxy-caching web server".
407 For Digest there is no default, this parameter is mandatory.
408 For NTLM and Negotiate this parameter is ignored.
410 "children" numberofchildren [startup=N] [idle=N] [concurrency=N] [queue-size=N]
412 The maximum number of authenticator processes to spawn. If
413 you start too few Squid will have to wait for them to process
414 a backlog of credential verifications, slowing it down. When
415 password verifications are done via a (slow) network you are
416 likely to need lots of authenticator processes.
418 The startup= and idle= options permit some skew in the exact
419 amount run. A minimum of startup=N will begin during startup
420 and reconfigure. Squid will start more in groups of up to
421 idle=N in an attempt to meet traffic needs and to keep idle=N
422 free above those traffic needs up to the maximum.
424 The concurrency= option sets the number of concurrent requests
425 the helper can process. The default of 0 is used for helpers
426 who only supports one request at a time. Setting this to a
427 number greater than 0 changes the protocol used to include a
428 channel ID field first on the request/response line, allowing
429 multiple requests to be sent to the same helper in parallel
430 without waiting for the response.
432 Concurrency must not be set unless it's known the helper
433 supports the input format with channel-ID fields.
435 The queue-size= option sets the maximum number of queued
436 requests. If the queued requests exceed queue size for more
437 than 3 minutes then squid aborts its operation.
438 The default value is set to 2*numberofchildren/
440 NOTE: NTLM and Negotiate schemes do not support concurrency
441 in the Squid code module even though some helpers can.
444 IF HAVE_AUTH_MODULE_BASIC
445 === Basic authentication parameters ===
448 HTTP uses iso-latin-1 as character set, while some
449 authentication backends such as LDAP expects UTF-8. If this is
450 set to on Squid will translate the HTTP iso-latin-1 charset to
451 UTF-8 before sending the username and password to the helper.
453 "credentialsttl" timetolive
454 Specifies how long squid assumes an externally validated
455 username:password pair is valid for - in other words how
456 often the helper program is called for that user. Set this
457 low to force revalidation with short lived passwords.
459 NOTE: setting this high does not impact your susceptibility
460 to replay attacks unless you are using an one-time password
461 system (such as SecureID). If you are using such a system,
462 you will be vulnerable to replay attacks unless you also
463 use the max_user_ip ACL in an http_access rule.
465 "casesensitive" on|off
466 Specifies if usernames are case sensitive. Most user databases
467 are case insensitive allowing the same username to be spelled
468 using both lower and upper case letters, but some are case
469 sensitive. This makes a big difference for user_max_ip ACL
470 processing and similar.
473 IF HAVE_AUTH_MODULE_DIGEST
474 === Digest authentication parameters ===
477 HTTP uses iso-latin-1 as character set, while some
478 authentication backends such as LDAP expects UTF-8. If this is
479 set to on Squid will translate the HTTP iso-latin-1 charset to
480 UTF-8 before sending the username and password to the helper.
482 "nonce_garbage_interval" timeinterval
483 Specifies the interval that nonces that have been issued
484 to client_agent's are checked for validity.
486 "nonce_max_duration" timeinterval
487 Specifies the maximum length of time a given nonce will be
490 "nonce_max_count" number
491 Specifies the maximum number of times a given nonce can be
494 "nonce_strictness" on|off
495 Determines if squid requires strict increment-by-1 behavior
496 for nonce counts, or just incrementing (off - for use when
497 user agents generate nonce counts that occasionally miss 1
498 (ie, 1,2,4,6)). Default off.
500 "check_nonce_count" on|off
501 This directive if set to off can disable the nonce count check
502 completely to work around buggy digest qop implementations in
503 certain mainstream browser versions. Default on to check the
504 nonce count to protect from authentication replay attacks.
506 "post_workaround" on|off
507 This is a workaround to certain buggy browsers who send an
508 incorrect request digest in POST requests when reusing the
509 same nonce as acquired earlier on a GET request.
512 IF HAVE_AUTH_MODULE_NEGOTIATE
513 === Negotiate authentication parameters ===
516 If you experience problems with PUT/POST requests when using
517 the this authentication scheme then you can try setting this
518 to off. This will cause Squid to forcibly close the connection
519 on the initial request where the browser asks which schemes
520 are supported by the proxy.
523 IF HAVE_AUTH_MODULE_NTLM
524 === NTLM authentication parameters ===
527 If you experience problems with PUT/POST requests when using
528 the this authentication scheme then you can try setting this
529 to off. This will cause Squid to forcibly close the connection
530 on the initial request where the browser asks which schemes
531 are supported by the proxy.
534 === Example Configuration ===
536 This configuration displays the recommended authentication scheme
537 order from most to least secure with recommended minimum configuration
538 settings for each scheme:
540 #auth_param negotiate program <uncomment and complete this line to activate>
541 #auth_param negotiate children 20 startup=0 idle=1
542 #auth_param negotiate keep_alive on
544 #auth_param digest program <uncomment and complete this line to activate>
545 #auth_param digest children 20 startup=0 idle=1
546 #auth_param digest realm Squid proxy-caching web server
547 #auth_param digest nonce_garbage_interval 5 minutes
548 #auth_param digest nonce_max_duration 30 minutes
549 #auth_param digest nonce_max_count 50
551 #auth_param ntlm program <uncomment and complete this line to activate>
552 #auth_param ntlm children 20 startup=0 idle=1
553 #auth_param ntlm keep_alive on
555 #auth_param basic program <uncomment and complete this line>
556 #auth_param basic children 5 startup=5 idle=1
557 #auth_param basic realm Squid proxy-caching web server
558 #auth_param basic credentialsttl 2 hours
561 NAME: authenticate_cache_garbage_interval
564 LOC: Config.authenticateGCInterval
566 The time period between garbage collection across the username cache.
567 This is a trade-off between memory utilization (long intervals - say
568 2 days) and CPU (short intervals - say 1 minute). Only change if you
572 NAME: authenticate_ttl
575 LOC: Config.authenticateTTL
577 The time a user & their credentials stay in the logged in
578 user cache since their last request. When the garbage
579 interval passes, all user credentials that have passed their
580 TTL are removed from memory.
583 NAME: authenticate_ip_ttl
585 LOC: Config.authenticateIpTTL
588 If you use proxy authentication and the 'max_user_ip' ACL,
589 this directive controls how long Squid remembers the IP
590 addresses associated with each user. Use a small value
591 (e.g., 60 seconds) if your users might change addresses
592 quickly, as is the case with dialup. You might be safe
593 using a larger value (e.g., 2 hours) in a corporate LAN
594 environment with relatively static address assignments.
599 -----------------------------------------------------------------------------
602 NAME: external_acl_type
603 TYPE: externalAclHelper
604 LOC: Config.externalAclHelperList
607 This option defines external acl classes using a helper program
608 to look up the status
610 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
614 ttl=n TTL in seconds for cached results (defaults to 3600
618 TTL for cached negative lookups (default same
621 grace=n Percentage remaining of TTL where a refresh of a
622 cached entry should be initiated without needing to
623 wait for a new reply. (default is for no grace period)
625 cache=n Limit the result cache size, default is 262144.
626 The expanded FORMAT value is used as the cache key, so
627 if the details in FORMAT are highly variable a larger
628 cache may be needed to produce reduction in helper load.
631 Maximum number of acl helper processes spawned to service
632 external acl lookups of this type. (default 20)
635 Minimum number of acl helper processes to spawn during
636 startup and reconfigure to service external acl lookups
637 of this type. (default 0)
640 Number of acl helper processes to keep ahead of traffic
641 loads. Squid will spawn this many at once whenever load
642 rises above the capabilities of existing processes.
643 Up to the value of children-max. (default 1)
645 concurrency=n concurrency level per process. Only used with helpers
646 capable of processing more than one query at a time.
648 queue-size=N The queue-size= option sets the maximum number of queued
649 requests. If the queued requests exceed queue size
651 The default value is set to 2*children-max.
653 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers.
655 ipv4 / ipv6 IP protocol used to communicate with this helper.
656 The default is to auto-detect IPv6 and use it when available.
659 FORMAT specifications
661 %LOGIN Authenticated user login name
662 %EXT_USER Username from previous external acl
663 %EXT_LOG Log details from previous external acl
664 %EXT_TAG Tag from previous external acl
665 %IDENT Ident user name
667 %SRCPORT Client source port
670 %PROTO Requested URL scheme
672 %PATH Requested URL path
673 %METHOD Request method
674 %MYADDR Squid interface address
675 %MYPORT Squid http_port number
676 %PATH Requested URL-path (including query-string if any)
677 %USER_CERT SSL User certificate in PEM format
678 %USER_CERTCHAIN SSL User certificate chain in PEM format
679 %USER_CERT_xx SSL User certificate subject attribute xx
680 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
681 %ssl::>sni SSL client SNI sent to Squid
682 %ssl::<cert_subject SSL server certificate DN
683 %ssl::<cert_issuer SSL server certificate issuer DN
685 %>{Header} HTTP request header "Header"
687 HTTP request header "Hdr" list member "member"
689 HTTP request header list member using ; as
690 list separator. ; can be any non-alphanumeric
693 %<{Header} HTTP reply header "Header"
695 HTTP reply header "Hdr" list member "member"
697 HTTP reply header list member using ; as
698 list separator. ; can be any non-alphanumeric
701 %ACL The name of the ACL being tested.
702 %DATA The ACL arguments. If not used then any arguments
703 is automatically added at the end of the line
705 NOTE: this will encode the arguments as one token,
706 whereas the default will pass each separately.
708 %% The percent sign. Useful for helpers which need
709 an unchanging input format.
712 General request syntax:
714 [channel-ID] FORMAT-values [acl-values ...]
717 FORMAT-values consists of transaction details expanded with
718 whitespace separation per the config file FORMAT specification
719 using the FORMAT macros listed above.
721 acl-values consists of any string specified in the referencing
722 config 'acl ... external' line. see the "acl external" directive.
724 Request values sent to the helper are URL escaped to protect
725 each value in requests against whitespaces.
727 If using protocol=2.5 then the request sent to the helper is not
728 URL escaped to protect against whitespace.
730 NOTE: protocol=3.0 is deprecated as no longer necessary.
732 When using the concurrency= option the protocol is changed by
733 introducing a query channel tag in front of the request/response.
734 The query channel tag is a number between 0 and concurrency-1.
735 This value must be echoed back unchanged to Squid as the first part
736 of the response relating to its request.
739 The helper receives lines expanded per the above format specification
740 and for each input line returns 1 line starting with OK/ERR/BH result
741 code and optionally followed by additional keywords with more details.
744 General result syntax:
746 [channel-ID] result keyword=value ...
748 Result consists of one of the codes:
751 the ACL test produced a match.
754 the ACL test does not produce a match.
757 An internal error occurred in the helper, preventing
758 a result being identified.
760 The meaning of 'a match' is determined by your squid.conf
761 access control configuration. See the Squid wiki for details.
765 user= The users name (login)
767 password= The users password (for login= cache_peer option)
769 message= Message describing the reason for this response.
770 Available as %o in error pages.
771 Useful on (ERR and BH results).
773 tag= Apply a tag to a request. Only sets a tag once,
774 does not alter existing tags.
776 log= String to be logged in access.log. Available as
777 %ea in logformat specifications.
779 clt_conn_tag= Associates a TAG with the client TCP connection.
780 Please see url_rewrite_program related documentation
783 Any keywords may be sent on any response whether OK, ERR or BH.
785 All response keyword values need to be a single token with URL
786 escaping, or enclosed in double quotes (") and escaped using \ on
787 any double quotes or \ characters within the value. The wrapping
788 double quotes are removed before the value is interpreted by Squid.
789 \r and \n are also replace by CR and LF.
791 Some example key values:
795 user="J. \"Bob\" Smith"
802 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
803 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
804 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
805 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
806 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
809 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
810 DEFAULT: localhost src 127.0.0.1/32 ::1
811 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
812 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
814 Defining an Access List
816 Every access list definition must begin with an aclname and acltype,
817 followed by either type-specific arguments or a quoted filename that
820 acl aclname acltype argument ...
821 acl aclname acltype "file" ...
823 When using "file", the file should contain one item per line.
825 Some acl types supports options which changes their default behaviour.
826 The available options are:
828 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
829 case-insensitive, use the -i option. To return case-sensitive
830 use the +i option between patterns, or make a new ACL line
833 -n Disable lookups and address type conversions. If lookup or
834 conversion is required because the parameter type (IP or
835 domain name) does not match the message address type (domain
836 name or IP), then the ACL would immediately declare a mismatch
837 without any warnings or lookups.
839 -- Used to stop processing all options, in the case the first acl
840 value has '-' character as first character (for example the '-'
841 is a valid domain name)
843 Some acl types require suspending the current request in order
844 to access some external data source.
845 Those which do are marked with the tag [slow], those which
846 don't are marked as [fast].
847 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
848 for further information
850 ***** ACL TYPES AVAILABLE *****
852 acl aclname src ip-address/mask ... # clients IP address [fast]
853 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
854 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
855 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
857 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
858 # The arp ACL requires the special configure option --enable-arp-acl.
859 # Furthermore, the ARP ACL code is not portable to all operating systems.
860 # It works on Linux, Solaris, Windows, FreeBSD, and some
861 # other *BSD variants.
864 # NOTE: Squid can only determine the MAC address for clients that are on
865 # the same subnet. If the client is on a different subnet,
866 # then Squid cannot find out its MAC address.
868 acl aclname srcdomain .foo.com ...
869 # reverse lookup, from client IP [slow]
870 acl aclname dstdomain [-n] .foo.com ...
871 # Destination server from URL [fast]
872 acl aclname srcdom_regex [-i] \.foo\.com ...
873 # regex matching client name [slow]
874 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
875 # regex matching server [fast]
877 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
878 # based URL is used and no match is found. The name "none" is used
879 # if the reverse lookup fails.
881 acl aclname src_as number ...
882 acl aclname dst_as number ...
884 # Except for access control, AS numbers can be used for
885 # routing of requests to specific caches. Here's an
886 # example for routing all requests for AS#1241 and only
887 # those to mycache.mydomain.net:
888 # acl asexample dst_as 1241
889 # cache_peer_access mycache.mydomain.net allow asexample
890 # cache_peer_access mycache_mydomain.net deny all
892 acl aclname peername myPeer ...
894 # match against a named cache_peer entry
895 # set unique name= on cache_peer lines for reliable use.
897 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
907 # h1:m1 must be less than h2:m2
909 acl aclname url_regex [-i] ^http:// ...
910 # regex matching on whole URL [fast]
911 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
912 # regex matching on URL login field
913 acl aclname urlpath_regex [-i] \.gif$ ...
914 # regex matching on URL path [fast]
916 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
918 acl aclname localport 3128 ... # TCP port the client connected to [fast]
919 # NP: for interception mode this is usually '80'
921 acl aclname myportname 3128 ... # *_port name [fast]
923 acl aclname proto HTTP FTP ... # request protocol [fast]
925 acl aclname method GET POST ... # HTTP request method [fast]
927 acl aclname http_status 200 301 500- 400-403 ...
928 # status code in reply [fast]
930 acl aclname browser [-i] regexp ...
931 # pattern match on User-Agent header (see also req_header below) [fast]
933 acl aclname referer_regex [-i] regexp ...
934 # pattern match on Referer header [fast]
935 # Referer is highly unreliable, so use with care
937 acl aclname ident username ...
938 acl aclname ident_regex [-i] pattern ...
939 # string match on ident output [slow]
940 # use REQUIRED to accept any non-null ident.
942 acl aclname proxy_auth [-i] username ...
943 acl aclname proxy_auth_regex [-i] pattern ...
944 # perform http authentication challenge to the client and match against
945 # supplied credentials [slow]
947 # takes a list of allowed usernames.
948 # use REQUIRED to accept any valid username.
950 # Will use proxy authentication in forward-proxy scenarios, and plain
951 # http authenticaiton in reverse-proxy scenarios
953 # NOTE: when a Proxy-Authentication header is sent but it is not
954 # needed during ACL checking the username is NOT logged
957 # NOTE: proxy_auth requires a EXTERNAL authentication program
958 # to check username/password combinations (see
959 # auth_param directive).
961 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
962 # as the browser needs to be configured for using a proxy in order
963 # to respond to proxy authentication.
965 acl aclname snmp_community string ...
966 # A community string to limit access to your SNMP Agent [fast]
969 # acl snmppublic snmp_community public
971 acl aclname maxconn number
972 # This will be matched when the client's IP address has
973 # more than <number> TCP connections established. [fast]
974 # NOTE: This only measures direct TCP links so X-Forwarded-For
975 # indirect clients are not counted.
977 acl aclname max_user_ip [-s] number
978 # This will be matched when the user attempts to log in from more
979 # than <number> different ip addresses. The authenticate_ip_ttl
980 # parameter controls the timeout on the ip entries. [fast]
981 # If -s is specified the limit is strict, denying browsing
982 # from any further IP addresses until the ttl has expired. Without
983 # -s Squid will just annoy the user by "randomly" denying requests.
984 # (the counter is reset each time the limit is reached and a
986 # NOTE: in acceleration mode or where there is mesh of child proxies,
987 # clients may appear to come from multiple addresses if they are
988 # going through proxy farms, so a limit of 1 may cause user problems.
990 acl aclname random probability
991 # Pseudo-randomly match requests. Based on the probability given.
992 # Probability may be written as a decimal (0.333), fraction (1/3)
993 # or ratio of matches:non-matches (3:5).
995 acl aclname req_mime_type [-i] mime-type ...
996 # regex match against the mime type of the request generated
997 # by the client. Can be used to detect file upload or some
998 # types HTTP tunneling requests [fast]
999 # NOTE: This does NOT match the reply. You cannot use this
1000 # to match the returned file type.
1002 acl aclname req_header header-name [-i] any\.regex\.here
1003 # regex match against any of the known request headers. May be
1004 # thought of as a superset of "browser", "referer" and "mime-type"
1007 acl aclname rep_mime_type [-i] mime-type ...
1008 # regex match against the mime type of the reply received by
1009 # squid. Can be used to detect file download or some
1010 # types HTTP tunneling requests. [fast]
1011 # NOTE: This has no effect in http_access rules. It only has
1012 # effect in rules that affect the reply data stream such as
1013 # http_reply_access.
1015 acl aclname rep_header header-name [-i] any\.regex\.here
1016 # regex match against any of the known reply headers. May be
1017 # thought of as a superset of "browser", "referer" and "mime-type"
1020 acl aclname external class_name [arguments...]
1021 # external ACL lookup via a helper class defined by the
1022 # external_acl_type directive [slow]
1024 acl aclname user_cert attribute values...
1025 # match against attributes in a user SSL certificate
1026 # attribute is one of DN/C/O/CN/L/ST [fast]
1028 acl aclname ca_cert attribute values...
1029 # match against attributes a users issuing CA SSL certificate
1030 # attribute is one of DN/C/O/CN/L/ST [fast]
1032 acl aclname ext_user username ...
1033 acl aclname ext_user_regex [-i] pattern ...
1034 # string match on username returned by external acl helper [slow]
1035 # use REQUIRED to accept any non-null user name.
1037 acl aclname tag tagvalue ...
1038 # string match on tag returned by external acl helper [fast]
1039 # DEPRECATED. Only the first tag will match with this ACL.
1040 # Use the 'note' ACL instead for handling multiple tag values.
1042 acl aclname hier_code codename ...
1043 # string match against squid hierarchy code(s); [fast]
1044 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1046 # NOTE: This has no effect in http_access rules. It only has
1047 # effect in rules that affect the reply data stream such as
1048 # http_reply_access.
1050 acl aclname note name [value ...]
1051 # match transaction annotation [fast]
1052 # Without values, matches any annotation with a given name.
1053 # With value(s), matches any annotation with a given name that
1054 # also has one of the given values.
1055 # Names and values are compared using a string equality test.
1056 # Annotation sources include note and adaptation_meta directives
1057 # as well as helper and eCAP responses.
1059 acl aclname adaptation_service service ...
1060 # Matches the name of any icap_service, ecap_service,
1061 # adaptation_service_set, or adaptation_service_chain that Squid
1062 # has used (or attempted to use) for the master transaction.
1063 # This ACL must be defined after the corresponding adaptation
1064 # service is named in squid.conf. This ACL is usable with
1065 # adaptation_meta because it starts matching immediately after
1066 # the service has been selected for adaptation.
1069 acl aclname ssl_error errorname
1070 # match against SSL certificate validation error [fast]
1072 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1075 # The following can be used as shortcuts for certificate properties:
1076 # [ssl::]certHasExpired: the "not after" field is in the past
1077 # [ssl::]certNotYetValid: the "not before" field is in the future
1078 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1079 # [ssl::]certSelfSigned: The certificate is self signed.
1080 # [ssl::]certDomainMismatch: The certificate CN domain does not
1081 # match the name the name of the host we are connecting to.
1083 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1084 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1085 # predefined ACLs, just like the 'all' ACL.
1087 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1088 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1090 acl aclname server_cert_fingerprint [-sha1] fingerprint
1091 # match against server SSL certificate fingerprint [fast]
1093 # The fingerprint is the digest of the DER encoded version
1094 # of the whole certificate. The user should use the form: XX:XX:...
1095 # Optional argument specifies the digest algorithm to use.
1096 # The SHA1 digest algorithm is the default and is currently
1097 # the only algorithm supported (-sha1).
1099 acl aclname at_step step
1100 # match against the current step during ssl_bump evaluation [fast]
1101 # Never matches and should not be used outside the ssl_bump context.
1103 # At each SslBump step, Squid evaluates ssl_bump directives to find
1104 # the next bumping action (e.g., peek or splice). Valid SslBump step
1105 # values and the corresponding ssl_bump evaluation moments are:
1106 # SslBump1: After getting TCP-level and HTTP CONNECT info.
1107 # SslBump2: After getting SSL Client Hello info.
1108 # SslBump3: After getting SSL Server Hello info.
1110 acl aclname any-of acl1 acl2 ...
1111 # match any one of the acls [fast or slow]
1112 # The first matching ACL stops further ACL evaluation.
1114 # ACLs from multiple any-of lines with the same name are ORed.
1115 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1116 # acl A any-of a1 a2
1117 # acl A any-of a3 a4
1119 # This group ACL is fast if all evaluated ACLs in the group are fast
1120 # and slow otherwise.
1122 acl aclname all-of acl1 acl2 ...
1123 # match all of the acls [fast or slow]
1124 # The first mismatching ACL stops further ACL evaluation.
1126 # ACLs from multiple all-of lines with the same name are ORed.
1127 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1128 # acl B all-of b1 b2
1129 # acl B all-of b3 b4
1131 # This group ACL is fast if all evaluated ACLs in the group are fast
1132 # and slow otherwise.
1135 acl macaddress arp 09:00:2b:23:45:67
1136 acl myexample dst_as 1241
1137 acl password proxy_auth REQUIRED
1138 acl fileupload req_mime_type -i ^multipart/form-data$
1139 acl javascript rep_mime_type -i ^application/x-javascript$
1143 # Recommended minimum configuration:
1146 # Example rule allowing access from your local networks.
1147 # Adapt to list your (internal) IP networks from where browsing
1149 acl localnet src 0.0.0.1-0.255.255.255 # RFC 1122 "this" network (LAN)
1150 acl localnet src 10.0.0.0/8 # RFC 1918 local private network (LAN)
1151 acl localnet src 100.64.0.0/10 # RFC 6598 shared address space (CGN)
1152 acl localhet src 169.254.0.0/16 # RFC 3927 link-local (directly plugged) machines
1153 acl localnet src 172.16.0.0/12 # RFC 1918 local private network (LAN)
1154 acl localnet src 192.168.0.0/16 # RFC 1918 local private network (LAN)
1155 acl localnet src fc00::/7 # RFC 4193 local private network range
1156 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1158 acl SSL_ports port 443
1159 acl Safe_ports port 80 # http
1160 acl Safe_ports port 21 # ftp
1161 acl Safe_ports port 443 # https
1162 acl Safe_ports port 70 # gopher
1163 acl Safe_ports port 210 # wais
1164 acl Safe_ports port 1025-65535 # unregistered ports
1165 acl Safe_ports port 280 # http-mgmt
1166 acl Safe_ports port 488 # gss-http
1167 acl Safe_ports port 591 # filemaker
1168 acl Safe_ports port 777 # multiling http
1169 acl CONNECT method CONNECT
1173 NAME: proxy_protocol_access
1175 LOC: Config.accessList.proxyProtocol
1177 DEFAULT_DOC: all TCP connections to ports with require-proxy-header will be denied
1179 Determine which client proxies can be trusted to provide correct
1180 information regarding real client IP address using PROXY protocol.
1182 Requests may pass through a chain of several other proxies
1183 before reaching us. The original source details may by sent in:
1184 * HTTP message Forwarded header, or
1185 * HTTP message X-Forwarded-For header, or
1186 * PROXY protocol connection header.
1188 This directive is solely for validating new PROXY protocol
1189 connections received from a port flagged with require-proxy-header.
1190 It is checked only once after TCP connection setup.
1192 A deny match results in TCP connection closure.
1194 An allow match is required for Squid to permit the corresponding
1195 TCP connection, before Squid even looks for HTTP request headers.
1196 If there is an allow match, Squid starts using PROXY header information
1197 to determine the source address of the connection for all future ACL
1198 checks, logging, etc.
1200 SECURITY CONSIDERATIONS:
1202 Any host from which we accept client IP details can place
1203 incorrect information in the relevant header, and Squid
1204 will use the incorrect information as if it were the
1205 source address of the request. This may enable remote
1206 hosts to bypass any access control restrictions that are
1207 based on the client's source addresses.
1209 This clause only supports fast acl types.
1210 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1213 NAME: follow_x_forwarded_for
1215 IFDEF: FOLLOW_X_FORWARDED_FOR
1216 LOC: Config.accessList.followXFF
1217 DEFAULT_IF_NONE: deny all
1218 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1220 Determine which client proxies can be trusted to provide correct
1221 information regarding real client IP address.
1223 Requests may pass through a chain of several other proxies
1224 before reaching us. The original source details may by sent in:
1225 * HTTP message Forwarded header, or
1226 * HTTP message X-Forwarded-For header, or
1227 * PROXY protocol connection header.
1229 PROXY protocol connections are controlled by the proxy_protocol_access
1230 directive which is checked before this.
1232 If a request reaches us from a source that is allowed by this
1233 directive, then we trust the information it provides regarding
1234 the IP of the client it received from (if any).
1236 For the purpose of ACLs used in this directive the src ACL type always
1237 matches the address we are testing and srcdomain matches its rDNS.
1239 On each HTTP request Squid checks for X-Forwarded-For header fields.
1240 If found the header values are iterated in reverse order and an allow
1241 match is required for Squid to continue on to the next value.
1242 The verification ends when a value receives a deny match, cannot be
1243 tested, or there are no more values to test.
1244 NOTE: Squid does not yet follow the Forwarded HTTP header.
1246 The end result of this process is an IP address that we will
1247 refer to as the indirect client address. This address may
1248 be treated as the client address for access control, ICAP, delay
1249 pools and logging, depending on the acl_uses_indirect_client,
1250 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1251 log_uses_indirect_client and tproxy_uses_indirect_client options.
1253 This clause only supports fast acl types.
1254 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1256 SECURITY CONSIDERATIONS:
1258 Any host from which we accept client IP details can place
1259 incorrect information in the relevant header, and Squid
1260 will use the incorrect information as if it were the
1261 source address of the request. This may enable remote
1262 hosts to bypass any access control restrictions that are
1263 based on the client's source addresses.
1267 acl localhost src 127.0.0.1
1268 acl my_other_proxy srcdomain .proxy.example.com
1269 follow_x_forwarded_for allow localhost
1270 follow_x_forwarded_for allow my_other_proxy
1273 NAME: acl_uses_indirect_client
1276 IFDEF: FOLLOW_X_FORWARDED_FOR
1278 LOC: Config.onoff.acl_uses_indirect_client
1280 Controls whether the indirect client address
1281 (see follow_x_forwarded_for) is used instead of the
1282 direct client address in acl matching.
1284 NOTE: maxconn ACL considers direct TCP links and indirect
1285 clients will always have zero. So no match.
1288 NAME: delay_pool_uses_indirect_client
1291 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1293 LOC: Config.onoff.delay_pool_uses_indirect_client
1295 Controls whether the indirect client address
1296 (see follow_x_forwarded_for) is used instead of the
1297 direct client address in delay pools.
1300 NAME: log_uses_indirect_client
1303 IFDEF: FOLLOW_X_FORWARDED_FOR
1305 LOC: Config.onoff.log_uses_indirect_client
1307 Controls whether the indirect client address
1308 (see follow_x_forwarded_for) is used instead of the
1309 direct client address in the access log.
1312 NAME: tproxy_uses_indirect_client
1315 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1317 LOC: Config.onoff.tproxy_uses_indirect_client
1319 Controls whether the indirect client address
1320 (see follow_x_forwarded_for) is used instead of the
1321 direct client address when spoofing the outgoing client.
1323 This has no effect on requests arriving in non-tproxy
1326 SECURITY WARNING: Usage of this option is dangerous
1327 and should not be used trivially. Correct configuration
1328 of follow_x_forewarded_for with a limited set of trusted
1329 sources is required to prevent abuse of your proxy.
1332 NAME: spoof_client_ip
1334 LOC: Config.accessList.spoof_client_ip
1336 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1338 Control client IP address spoofing of TPROXY traffic based on
1339 defined access lists.
1341 spoof_client_ip allow|deny [!]aclname ...
1343 If there are no "spoof_client_ip" lines present, the default
1344 is to "allow" spoofing of any suitable request.
1346 Note that the cache_peer "no-tproxy" option overrides this ACL.
1348 This clause supports fast acl types.
1349 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1354 LOC: Config.accessList.http
1355 DEFAULT_IF_NONE: deny all
1356 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1358 Allowing or Denying access based on defined access lists
1360 To allow or deny a message received on an HTTP, HTTPS, or FTP port:
1361 http_access allow|deny [!]aclname ...
1363 NOTE on default values:
1365 If there are no "access" lines present, the default is to deny
1368 If none of the "access" lines cause a match, the default is the
1369 opposite of the last line in the list. If the last line was
1370 deny, the default is allow. Conversely, if the last line
1371 is allow, the default will be deny. For these reasons, it is a
1372 good idea to have an "deny all" entry at the end of your access
1373 lists to avoid potential confusion.
1375 This clause supports both fast and slow acl types.
1376 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1381 # Recommended minimum Access Permission configuration:
1383 # Deny requests to certain unsafe ports
1384 http_access deny !Safe_ports
1386 # Deny CONNECT to other than secure SSL ports
1387 http_access deny CONNECT !SSL_ports
1389 # Only allow cachemgr access from localhost
1390 http_access allow localhost manager
1391 http_access deny manager
1393 # We strongly recommend the following be uncommented to protect innocent
1394 # web applications running on the proxy server who think the only
1395 # one who can access services on "localhost" is a local user
1396 #http_access deny to_localhost
1399 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1402 # Example rule allowing access from your local networks.
1403 # Adapt localnet in the ACL section to list your (internal) IP networks
1404 # from where browsing should be allowed
1405 http_access allow localnet
1406 http_access allow localhost
1408 # And finally deny all other access to this proxy
1409 http_access deny all
1413 NAME: adapted_http_access http_access2
1415 LOC: Config.accessList.adapted_http
1417 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1419 Allowing or Denying access based on defined access lists
1421 Essentially identical to http_access, but runs after redirectors
1422 and ICAP/eCAP adaptation. Allowing access control based on their
1425 If not set then only http_access is used.
1428 NAME: http_reply_access
1430 LOC: Config.accessList.reply
1432 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1434 Allow replies to client requests. This is complementary to http_access.
1436 http_reply_access allow|deny [!] aclname ...
1438 NOTE: if there are no access lines present, the default is to allow
1441 If none of the access lines cause a match the opposite of the
1442 last line will apply. Thus it is good practice to end the rules
1443 with an "allow all" or "deny all" entry.
1445 This clause supports both fast and slow acl types.
1446 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1451 LOC: Config.accessList.icp
1453 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1455 Allowing or Denying access to the ICP port based on defined
1458 icp_access allow|deny [!]aclname ...
1460 NOTE: The default if no icp_access lines are present is to
1461 deny all traffic. This default may cause problems with peers
1464 This clause only supports fast acl types.
1465 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1467 # Allow ICP queries from local networks only
1468 #icp_access allow localnet
1469 #icp_access deny all
1475 LOC: Config.accessList.htcp
1477 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1479 Allowing or Denying access to the HTCP port based on defined
1482 htcp_access allow|deny [!]aclname ...
1484 See also htcp_clr_access for details on access control for
1485 cache purge (CLR) HTCP messages.
1487 NOTE: The default if no htcp_access lines are present is to
1488 deny all traffic. This default may cause problems with peers
1489 using the htcp option.
1491 This clause only supports fast acl types.
1492 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1494 # Allow HTCP queries from local networks only
1495 #htcp_access allow localnet
1496 #htcp_access deny all
1499 NAME: htcp_clr_access
1502 LOC: Config.accessList.htcp_clr
1504 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1506 Allowing or Denying access to purge content using HTCP based
1507 on defined access lists.
1508 See htcp_access for details on general HTCP access control.
1510 htcp_clr_access allow|deny [!]aclname ...
1512 This clause only supports fast acl types.
1513 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1515 # Allow HTCP CLR requests from trusted peers
1516 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1517 htcp_clr_access allow htcp_clr_peer
1518 htcp_clr_access deny all
1523 LOC: Config.accessList.miss
1525 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1527 Determins whether network access is permitted when satisfying a request.
1530 to force your neighbors to use you as a sibling instead of
1533 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1534 miss_access deny !localclients
1535 miss_access allow all
1537 This means only your local clients are allowed to fetch relayed/MISS
1538 replies from the network and all other clients can only fetch cached
1541 The default for this setting allows all clients who passed the
1542 http_access rules to relay via this proxy.
1544 This clause only supports fast acl types.
1545 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1548 NAME: ident_lookup_access
1552 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1553 LOC: Ident::TheConfig.identLookup
1555 A list of ACL elements which, if matched, cause an ident
1556 (RFC 931) lookup to be performed for this request. For
1557 example, you might choose to always perform ident lookups
1558 for your main multi-user Unix boxes, but not for your Macs
1559 and PCs. By default, ident lookups are not performed for
1562 To enable ident lookups for specific client addresses, you
1563 can follow this example:
1565 acl ident_aware_hosts src 198.168.1.0/24
1566 ident_lookup_access allow ident_aware_hosts
1567 ident_lookup_access deny all
1569 Only src type ACL checks are fully supported. A srcdomain
1570 ACL might work at times, but it will not always provide
1573 This clause only supports fast acl types.
1574 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1577 NAME: reply_body_max_size
1578 COMMENT: size [acl acl...]
1581 DEFAULT_DOC: No limit is applied.
1582 LOC: Config.ReplyBodySize
1584 This option specifies the maximum size of a reply body. It can be
1585 used to prevent users from downloading very large files, such as
1586 MP3's and movies. When the reply headers are received, the
1587 reply_body_max_size lines are processed, and the first line where
1588 all (if any) listed ACLs are true is used as the maximum body size
1591 This size is checked twice. First when we get the reply headers,
1592 we check the content-length value. If the content length value exists
1593 and is larger than the allowed size, the request is denied and the
1594 user receives an error message that says "the request or reply
1595 is too large." If there is no content-length, and the reply
1596 size exceeds this limit, the client's connection is just closed
1597 and they will receive a partial reply.
1599 WARNING: downstream caches probably can not detect a partial reply
1600 if there is no content-length header, so they will cache
1601 partial responses and give them out as hits. You should NOT
1602 use this option if you have downstream caches.
1604 WARNING: A maximum size smaller than the size of squid's error messages
1605 will cause an infinite loop and crash squid. Ensure that the smallest
1606 non-zero value you use is greater that the maximum header size plus
1607 the size of your largest error page.
1609 If you set this parameter none (the default), there will be
1612 Configuration Format is:
1613 reply_body_max_size SIZE UNITS [acl ...]
1615 reply_body_max_size 10 MB
1619 NAME: on_unsupported_protocol
1620 TYPE: on_unsupported_protocol
1621 LOC: Config.accessList.on_unsupported_protocol
1623 DEFAULT_DOC: Respond with an error message to unidentifiable traffic
1625 Determines Squid behavior when encountering strange requests at the
1626 beginning of an accepted TCP connection. This is especially useful in
1627 interception environments where Squid is likely to see connections for
1628 unsupported protocols that Squid should either terminate or tunnel at
1631 on_unsupported_protocol <action> [!]acl ...
1633 The first matching action wins.
1635 Supported actions are:
1637 tunnel: Establish a TCP connection with the intended server and
1638 blindly shovel TCP packets between the client and server.
1640 respond: Respond with an error message, using the transfer protocol
1641 for the Squid port that received the request (e.g., HTTP
1642 for connections intercepted at the http_port). This is the
1645 Currently, this directive is ignored for non-intercepted connections
1646 because Squid cannot know what their intended destination is.
1649 # define what Squid errors indicate receiving non-HTTP traffic:
1650 acl foreignProtocol squid_error ERR_PROTOCOL_UNKNOWN ERR_TOO_BIG
1651 # define what Squid errors indicate receiving nothing:
1652 acl serverTalksFirstProtocol squid_error ERR_REQUEST_START_TIMEOUT
1653 # tunnel everything that does not look like HTTP:
1654 on_unsupported_protocol tunnel foreignProtocol
1655 # tunnel if we think the client waits for the server to talk first:
1656 on_unsupported_protocol tunnel serverTalksFirstProtocol
1657 # in all other error cases, just send an HTTP "error page" response:
1658 on_unsupported_protocol respond all
1660 See also: squid_error ACL
1665 -----------------------------------------------------------------------------
1668 NAME: http_port ascii_port
1673 Usage: port [mode] [options]
1674 hostname:port [mode] [options]
1675 1.2.3.4:port [mode] [options]
1677 The socket addresses where Squid will listen for HTTP client
1678 requests. You may specify multiple socket addresses.
1679 There are three forms: port alone, hostname with port, and
1680 IP address with port. If you specify a hostname or IP
1681 address, Squid binds the socket to that specific
1682 address. Most likely, you do not need to bind to a specific
1683 address, so you can use the port number alone.
1685 If you are running Squid in accelerator mode, you
1686 probably want to listen on port 80 also, or instead.
1688 The -a command line option may be used to specify additional
1689 port(s) where Squid listens for proxy request. Such ports will
1690 be plain proxy ports with no options.
1692 You may specify multiple socket addresses on multiple lines.
1696 intercept Support for IP-Layer interception of
1697 outgoing requests without browser settings.
1698 NP: disables authentication and IPv6 on the port.
1700 tproxy Support Linux TPROXY for spoofing outgoing
1701 connections using the client IP address.
1702 NP: disables authentication and maybe IPv6 on the port.
1704 accel Accelerator / reverse proxy mode
1706 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1707 establish secure connection with the client and with
1708 the server, decrypt HTTPS messages as they pass through
1709 Squid, and treat them as unencrypted HTTP messages,
1710 becoming the man-in-the-middle.
1712 The ssl_bump option is required to fully enable
1713 bumping of CONNECT requests.
1715 Omitting the mode flag causes default forward proxy mode to be used.
1718 Accelerator Mode Options:
1720 defaultsite=domainname
1721 What to use for the Host: header if it is not present
1722 in a request. Determines what site (not origin server)
1723 accelerators should consider the default.
1725 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1727 protocol= Protocol to reconstruct accelerated and intercepted
1728 requests with. Defaults to HTTP/1.1 for http_port and
1729 HTTPS/1.1 for https_port.
1730 When an unsupported value is configured Squid will
1731 produce a FATAL error.
1732 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1734 vport Virtual host port support. Using the http_port number
1735 instead of the port passed on Host: headers.
1737 vport=NN Virtual host port support. Using the specified port
1738 number instead of the port passed on Host: headers.
1741 Act as if this Squid is the origin server.
1742 This currently means generate new Date: and Expires:
1743 headers on HIT instead of adding Age:.
1745 ignore-cc Ignore request Cache-Control headers.
1747 WARNING: This option violates HTTP specifications if
1748 used in non-accelerator setups.
1750 allow-direct Allow direct forwarding in accelerator mode. Normally
1751 accelerated requests are denied direct forwarding as if
1752 never_direct was used.
1754 WARNING: this option opens accelerator mode to security
1755 vulnerabilities usually only affecting in interception
1756 mode. Make sure to protect forwarding with suitable
1757 http_access rules when using this.
1760 SSL Bump Mode Options:
1761 In addition to these options ssl-bump requires TLS/SSL options.
1763 generate-host-certificates[=<on|off>]
1764 Dynamically create SSL server certificates for the
1765 destination hosts of bumped CONNECT requests.When
1766 enabled, the cert and key options are used to sign
1767 generated certificates. Otherwise generated
1768 certificate will be selfsigned.
1769 If there is a CA certificate lifetime of the generated
1770 certificate equals lifetime of the CA certificate. If
1771 generated certificate is selfsigned lifetime is three
1773 This option is enabled by default when ssl-bump is used.
1774 See the ssl-bump option above for more information.
1776 dynamic_cert_mem_cache_size=SIZE
1777 Approximate total RAM size spent on cached generated
1778 certificates. If set to zero, caching is disabled. The
1779 default value is 4MB.
1783 cert= Path to SSL certificate (PEM format).
1785 key= Path to SSL private key file (PEM format)
1786 if not specified, the certificate file is
1787 assumed to be a combined certificate and
1790 version= The version of SSL/TLS supported
1791 1 automatic (default)
1797 cipher= Colon separated list of supported ciphers.
1798 NOTE: some ciphers such as EDH ciphers depend on
1799 additional settings. If those settings are
1800 omitted the ciphers may be silently ignored
1801 by the OpenSSL library.
1803 options= Various SSL implementation options. The most important
1805 NO_SSLv3 Disallow the use of SSLv3
1806 NO_TLSv1 Disallow the use of TLSv1.0
1807 NO_TLSv1_1 Disallow the use of TLSv1.1
1808 NO_TLSv1_2 Disallow the use of TLSv1.2
1809 SINGLE_DH_USE Always create a new key when using
1810 temporary/ephemeral DH key exchanges
1811 SSL_OP_NO_TICKET Disables TLS tickets extension
1812 ALL Enable various bug workarounds
1813 suggested as "harmless" by OpenSSL
1814 Be warned that this reduces SSL/TLS
1815 strength to some attacks.
1816 See OpenSSL SSL_CTX_set_options documentation for a
1817 complete list of options.
1819 clientca= File containing the list of CAs to use when
1820 requesting a client certificate.
1822 cafile= File containing additional CA certificates to
1823 use when verifying client certificates. If unset
1824 clientca will be used.
1826 capath= Directory containing additional CA certificates
1827 and CRL lists to use when verifying client certificates.
1829 crlfile= File of additional CRL lists to use when verifying
1830 the client certificate, in addition to CRLs stored in
1831 the capath. Implies VERIFY_CRL flag below.
1833 dhparams= File containing DH parameters for temporary/ephemeral
1834 DH key exchanges. See OpenSSL documentation for details
1835 on how to create this file.
1836 WARNING: EDH ciphers will be silently disabled if this
1839 sslflags= Various flags modifying the use of SSL:
1841 Don't request client certificates
1842 immediately, but wait until acl processing
1843 requires a certificate (not yet implemented).
1845 Don't use the default CA lists built in
1848 Don't allow for session reuse. Each connection
1849 will result in a new SSL session.
1851 Verify CRL lists when accepting client
1854 Verify CRL lists for all certificates in the
1855 client certificate chain.
1857 sslcontext= SSL session ID context identifier.
1861 connection-auth[=on|off]
1862 use connection-auth=off to tell Squid to prevent
1863 forwarding Microsoft connection oriented authentication
1864 (NTLM, Negotiate and Kerberos)
1866 disable-pmtu-discovery=
1867 Control Path-MTU discovery usage:
1868 off lets OS decide on what to do (default).
1869 transparent disable PMTU discovery when transparent
1871 always disable always PMTU discovery.
1873 In many setups of transparently intercepting proxies
1874 Path-MTU discovery can not work on traffic towards the
1875 clients. This is the case when the intercepting device
1876 does not fully track connections and fails to forward
1877 ICMP must fragment messages to the cache server. If you
1878 have such setup and experience that certain clients
1879 sporadically hang or never complete requests set
1880 disable-pmtu-discovery option to 'transparent'.
1882 name= Specifies a internal name for the port. Defaults to
1883 the port specification (port or addr:port)
1885 tcpkeepalive[=idle,interval,timeout]
1886 Enable TCP keepalive probes of idle connections.
1887 In seconds; idle is the initial time before TCP starts
1888 probing the connection, interval how often to probe, and
1889 timeout the time before giving up.
1891 require-proxy-header
1892 Require PROXY protocol version 1 or 2 connections.
1893 The proxy_protocol_access is required to whitelist
1894 downstream proxies which can be trusted.
1896 If you run Squid on a dual-homed machine with an internal
1897 and an external interface we recommend you to specify the
1898 internal address:port in http_port. This way Squid will only be
1899 visible on the internal address.
1903 # Squid normally listens to port 3128
1904 http_port @DEFAULT_HTTP_PORT@
1914 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1916 The socket address where Squid will listen for client requests made
1917 over TLS or SSL connections. Commonly referred to as HTTPS.
1919 This is most useful for situations where you are running squid in
1920 accelerator mode and you want to do the SSL work at the accelerator level.
1922 You may specify multiple socket addresses on multiple lines,
1923 each with their own SSL certificate and/or options.
1927 accel Accelerator / reverse proxy mode
1929 intercept Support for IP-Layer interception of
1930 outgoing requests without browser settings.
1931 NP: disables authentication and IPv6 on the port.
1933 tproxy Support Linux TPROXY for spoofing outgoing
1934 connections using the client IP address.
1935 NP: disables authentication and maybe IPv6 on the port.
1937 ssl-bump For each intercepted connection allowed by ssl_bump
1938 ACLs, establish a secure connection with the client and with
1939 the server, decrypt HTTPS messages as they pass through
1940 Squid, and treat them as unencrypted HTTP messages,
1941 becoming the man-in-the-middle.
1943 An "ssl_bump server-first" match is required to
1944 fully enable bumping of intercepted SSL connections.
1946 Requires tproxy or intercept.
1948 Omitting the mode flag causes default forward proxy mode to be used.
1951 See http_port for a list of generic options
1956 cert= Path to SSL certificate (PEM format).
1958 key= Path to SSL private key file (PEM format)
1959 if not specified, the certificate file is
1960 assumed to be a combined certificate and
1963 version= The version of SSL/TLS supported
1964 1 automatic (default)
1968 cipher= Colon separated list of supported ciphers.
1970 options= Various SSL engine options. The most important
1972 NO_SSLv3 Disallow the use of SSLv3
1973 NO_TLSv1 Disallow the use of TLSv1
1974 SINGLE_DH_USE Always create a new key when using
1975 temporary/ephemeral DH key exchanges
1976 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1977 documentation for a complete list of options.
1979 clientca= File containing the list of CAs to use when
1980 requesting a client certificate.
1982 cafile= File containing additional CA certificates to
1983 use when verifying client certificates. If unset
1984 clientca will be used.
1986 capath= Directory containing additional CA certificates
1987 and CRL lists to use when verifying client certificates.
1989 crlfile= File of additional CRL lists to use when verifying
1990 the client certificate, in addition to CRLs stored in
1991 the capath. Implies VERIFY_CRL flag below.
1993 dhparams= File containing DH parameters for temporary/ephemeral
1996 sslflags= Various flags modifying the use of SSL:
1998 Don't request client certificates
1999 immediately, but wait until acl processing
2000 requires a certificate (not yet implemented).
2002 Don't use the default CA lists built in
2005 Don't allow for session reuse. Each connection
2006 will result in a new SSL session.
2008 Verify CRL lists when accepting client
2011 Verify CRL lists for all certificates in the
2012 client certificate chain.
2014 sslcontext= SSL session ID context identifier.
2016 generate-host-certificates[=<on|off>]
2017 Dynamically create SSL server certificates for the
2018 destination hosts of bumped SSL requests.When
2019 enabled, the cert and key options are used to sign
2020 generated certificates. Otherwise generated
2021 certificate will be selfsigned.
2022 If there is CA certificate life time of generated
2023 certificate equals lifetime of CA certificate. If
2024 generated certificate is selfsigned lifetime is three
2026 This option is enabled by default when SslBump is used.
2027 See the sslBump option above for more information.
2029 dynamic_cert_mem_cache_size=SIZE
2030 Approximate total RAM size spent on cached generated
2031 certificates. If set to zero, caching is disabled. The
2032 default value is 4MB.
2034 See http_port for a list of available options.
2042 Enables Native FTP proxy by specifying the socket address where Squid
2043 listens for FTP client requests. See http_port directive for various
2044 ways to specify the listening address and mode.
2046 Usage: ftp_port address [mode] [options]
2048 WARNING: This is a new, experimental, complex feature that has seen
2049 limited production exposure. Some Squid modules (e.g., caching) do not
2050 currently work with native FTP proxying, and many features have not
2051 even been tested for compatibility. Test well before deploying!
2053 Native FTP proxying differs substantially from proxying HTTP requests
2054 with ftp:// URIs because Squid works as an FTP server and receives
2055 actual FTP commands (rather than HTTP requests with FTP URLs).
2057 Native FTP commands accepted at ftp_port are internally converted or
2058 wrapped into HTTP-like messages. The same happens to Native FTP
2059 responses received from FTP origin servers. Those HTTP-like messages
2060 are shoveled through regular access control and adaptation layers
2061 between the FTP client and the FTP origin server. This allows Squid to
2062 examine, adapt, block, and log FTP exchanges. Squid reuses most HTTP
2063 mechanisms when shoveling wrapped FTP messages. For example,
2064 http_access and adaptation_access directives are used.
2068 intercept Same as http_port intercept. The FTP origin address is
2069 determined based on the intended destination of the
2070 intercepted connection.
2072 tproxy Support Linux TPROXY for spoofing outgoing
2073 connections using the client IP address.
2074 NP: disables authentication and maybe IPv6 on the port.
2076 By default (i.e., without an explicit mode option), Squid extracts the
2077 FTP origin address from the login@origin parameter of the FTP USER
2078 command. Many popular FTP clients support such native FTP proxying.
2082 name=token Specifies an internal name for the port. Defaults to
2083 the port address. Usable with myportname ACL.
2086 Enables tracking of FTP directories by injecting extra
2087 PWD commands and adjusting Request-URI (in wrapping
2088 HTTP requests) to reflect the current FTP server
2089 directory. Tracking is disabled by default.
2091 protocol=FTP Protocol to reconstruct accelerated and intercepted
2092 requests with. Defaults to FTP. No other accepted
2093 values have been tested with. An unsupported value
2094 results in a FATAL error. Accepted values are FTP,
2095 HTTP (or HTTP/1.1), and HTTPS (or HTTPS/1.1).
2097 Other http_port modes and options that are not specific to HTTP and
2098 HTTPS may also work.
2101 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2104 LOC: Ip::Qos::TheConfig.tosToServer
2106 Allows you to select a TOS/Diffserv value for packets outgoing
2107 on the server side, based on an ACL.
2109 tcp_outgoing_tos ds-field [!]aclname ...
2111 Example where normal_service_net uses the TOS value 0x00
2112 and good_service_net uses 0x20
2114 acl normal_service_net src 10.0.0.0/24
2115 acl good_service_net src 10.0.1.0/24
2116 tcp_outgoing_tos 0x00 normal_service_net
2117 tcp_outgoing_tos 0x20 good_service_net
2119 TOS/DSCP values really only have local significance - so you should
2120 know what you're specifying. For more information, see RFC2474,
2121 RFC2475, and RFC3260.
2123 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
2124 "default" to use whatever default your host has. Note that in
2125 practice often only multiples of 4 is usable as the two rightmost bits
2126 have been redefined for use by ECN (RFC 3168 section 23.1).
2128 Processing proceeds in the order specified, and stops at first fully
2131 Only fast ACLs are supported.
2134 NAME: clientside_tos
2137 LOC: Ip::Qos::TheConfig.tosToClient
2139 Allows you to select a TOS/Diffserv value for packets being transmitted
2140 on the client-side, based on an ACL.
2142 clientside_tos ds-field [!]aclname ...
2144 Example where normal_service_net uses the TOS value 0x00
2145 and good_service_net uses 0x20
2147 acl normal_service_net src 10.0.0.0/24
2148 acl good_service_net src 10.0.1.0/24
2149 clientside_tos 0x00 normal_service_net
2150 clientside_tos 0x20 good_service_net
2152 Note: This feature is incompatible with qos_flows. Any TOS values set here
2153 will be overwritten by TOS values in qos_flows.
2156 NAME: tcp_outgoing_mark
2158 IFDEF: SO_MARK&&USE_LIBCAP
2160 LOC: Ip::Qos::TheConfig.nfmarkToServer
2162 Allows you to apply a Netfilter mark value to outgoing packets
2163 on the server side, based on an ACL.
2165 tcp_outgoing_mark mark-value [!]aclname ...
2167 Example where normal_service_net uses the mark value 0x00
2168 and good_service_net uses 0x20
2170 acl normal_service_net src 10.0.0.0/24
2171 acl good_service_net src 10.0.1.0/24
2172 tcp_outgoing_mark 0x00 normal_service_net
2173 tcp_outgoing_mark 0x20 good_service_net
2175 Only fast ACLs are supported.
2178 NAME: clientside_mark
2180 IFDEF: SO_MARK&&USE_LIBCAP
2182 LOC: Ip::Qos::TheConfig.nfmarkToClient
2184 Allows you to apply a Netfilter mark value to packets being transmitted
2185 on the client-side, based on an ACL.
2187 clientside_mark mark-value [!]aclname ...
2189 Example where normal_service_net uses the mark value 0x00
2190 and good_service_net uses 0x20
2192 acl normal_service_net src 10.0.0.0/24
2193 acl good_service_net src 10.0.1.0/24
2194 clientside_mark 0x00 normal_service_net
2195 clientside_mark 0x20 good_service_net
2197 Note: This feature is incompatible with qos_flows. Any mark values set here
2198 will be overwritten by mark values in qos_flows.
2205 LOC: Ip::Qos::TheConfig
2207 Allows you to select a TOS/DSCP value to mark outgoing
2208 connections to the client, based on where the reply was sourced.
2209 For platforms using netfilter, allows you to set a netfilter mark
2210 value instead of, or in addition to, a TOS value.
2212 By default this functionality is disabled. To enable it with the default
2213 settings simply use "qos_flows mark" or "qos_flows tos". Default
2214 settings will result in the netfilter mark or TOS value being copied
2215 from the upstream connection to the client. Note that it is the connection
2216 CONNMARK value not the packet MARK value that is copied.
2218 It is not currently possible to copy the mark or TOS value from the
2219 client to the upstream connection request.
2221 TOS values really only have local significance - so you should
2222 know what you're specifying. For more information, see RFC2474,
2223 RFC2475, and RFC3260.
2225 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
2226 in practice often only multiples of 4 is usable as the two rightmost bits
2227 have been redefined for use by ECN (RFC 3168 section 23.1).
2229 Mark values can be any unsigned 32-bit integer value.
2231 This setting is configured by setting the following values:
2233 tos|mark Whether to set TOS or netfilter mark values
2235 local-hit=0xFF Value to mark local cache hits.
2237 sibling-hit=0xFF Value to mark hits from sibling peers.
2239 parent-hit=0xFF Value to mark hits from parent peers.
2241 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2242 over the preserve-miss feature (see below), unless
2243 mask is specified, in which case only the bits
2244 specified in the mask are written.
2246 The TOS variant of the following features are only possible on Linux
2247 and require your kernel to be patched with the TOS preserving ZPH
2248 patch, available from http://zph.bratcheda.org
2249 No patch is needed to preserve the netfilter mark, which will work
2250 with all variants of netfilter.
2252 disable-preserve-miss
2253 This option disables the preservation of the TOS or netfilter
2254 mark. By default, the existing TOS or netfilter mark value of
2255 the response coming from the remote server will be retained
2256 and masked with miss-mark.
2257 NOTE: in the case of a netfilter mark, the mark must be set on
2258 the connection (using the CONNMARK target) not on the packet
2262 Allows you to mask certain bits in the TOS or mark value
2263 received from the remote server, before copying the value to
2264 the TOS sent towards clients.
2265 Default for tos: 0xFF (TOS from server is not changed).
2266 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2268 All of these features require the --enable-zph-qos compilation flag
2269 (enabled by default). Netfilter marking also requires the
2270 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2271 libcap 2.09+ (--with-libcap).
2275 NAME: tcp_outgoing_address
2278 DEFAULT_DOC: Address selection is performed by the operating system.
2279 LOC: Config.accessList.outgoing_address
2281 Allows you to map requests to different outgoing IP addresses
2282 based on the username or source address of the user making
2285 tcp_outgoing_address ipaddr [[!]aclname] ...
2288 Forwarding clients with dedicated IPs for certain subnets.
2290 acl normal_service_net src 10.0.0.0/24
2291 acl good_service_net src 10.0.2.0/24
2293 tcp_outgoing_address 2001:db8::c001 good_service_net
2294 tcp_outgoing_address 10.1.0.2 good_service_net
2296 tcp_outgoing_address 2001:db8::beef normal_service_net
2297 tcp_outgoing_address 10.1.0.1 normal_service_net
2299 tcp_outgoing_address 2001:db8::1
2300 tcp_outgoing_address 10.1.0.3
2302 Processing proceeds in the order specified, and stops at first fully
2305 Squid will add an implicit IP version test to each line.
2306 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2307 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2310 NOTE: The use of this directive using client dependent ACLs is
2311 incompatible with the use of server side persistent connections. To
2312 ensure correct results it is best to set server_persistent_connections
2313 to off when using this directive in such configurations.
2315 NOTE: The use of this directive to set a local IP on outgoing TCP links
2316 is incompatible with using TPROXY to set client IP out outbound TCP links.
2317 When needing to contact peers use the no-tproxy cache_peer option and the
2318 client_dst_passthru directive re-enable normal forwarding such as this.
2322 NAME: host_verify_strict
2325 LOC: Config.onoff.hostStrictVerify
2327 Regardless of this option setting, when dealing with intercepted
2328 traffic, Squid always verifies that the destination IP address matches
2329 the Host header domain or IP (called 'authority form URL').
2331 This enforcement is performed to satisfy a MUST-level requirement in
2332 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2333 authority of the origin server or gateway given by the original URL".
2336 Squid always responds with an HTTP 409 (Conflict) error
2337 page and logs a security warning if there is no match.
2339 Squid verifies that the destination IP address matches
2340 the Host header for forward-proxy and reverse-proxy traffic
2341 as well. For those traffic types, Squid also enables the
2342 following checks, comparing the corresponding Host header
2343 and Request-URI components:
2345 * The host names (domain or IP) must be identical,
2346 but valueless or missing Host header disables all checks.
2347 For the two host names to match, both must be either IP
2350 * Port numbers must be identical, but if a port is missing
2351 the scheme-default port is assumed.
2354 When set to OFF (the default):
2355 Squid allows suspicious requests to continue but logs a
2356 security warning and blocks caching of the response.
2358 * Forward-proxy traffic is not checked at all.
2360 * Reverse-proxy traffic is not checked at all.
2362 * Intercepted traffic which passes verification is handled
2363 according to client_dst_passthru.
2365 * Intercepted requests which fail verification are sent
2366 to the client original destination instead of DIRECT.
2367 This overrides 'client_dst_passthru off'.
2369 For now suspicious intercepted CONNECT requests are always
2370 responded to with an HTTP 409 (Conflict) error page.
2375 As described in CVE-2009-0801 when the Host: header alone is used
2376 to determine the destination of a request it becomes trivial for
2377 malicious scripts on remote websites to bypass browser same-origin
2378 security policy and sandboxing protections.
2380 The cause of this is that such applets are allowed to perform their
2381 own HTTP stack, in which case the same-origin policy of the browser
2382 sandbox only verifies that the applet tries to contact the same IP
2383 as from where it was loaded at the IP level. The Host: header may
2384 be different from the connected IP and approved origin.
2388 NAME: client_dst_passthru
2391 LOC: Config.onoff.client_dst_passthru
2393 With NAT or TPROXY intercepted traffic Squid may pass the request
2394 directly to the original client destination IP or seek a faster
2395 source using the HTTP Host header.
2397 Using Host to locate alternative servers can provide faster
2398 connectivity with a range of failure recovery options.
2399 But can also lead to connectivity trouble when the client and
2400 server are attempting stateful interactions unaware of the proxy.
2402 This option (on by default) prevents alternative DNS entries being
2403 located to send intercepted traffic DIRECT to an origin server.
2404 The clients original destination IP and port will be used instead.
2406 Regardless of this option setting, when dealing with intercepted
2407 traffic Squid will verify the Host: header and any traffic which
2408 fails Host verification will be treated as if this option were ON.
2410 see host_verify_strict for details on the verification process.
2415 -----------------------------------------------------------------------------
2418 NAME: ssl_unclean_shutdown
2422 LOC: Config.SSL.unclean_shutdown
2424 Some browsers (especially MSIE) bugs out on SSL shutdown
2431 LOC: Config.SSL.ssl_engine
2434 The OpenSSL engine to use. You will need to set this if you
2435 would like to use hardware SSL acceleration for example.
2438 NAME: sslproxy_client_certificate
2441 LOC: Config.ssl_client.cert
2444 Client SSL Certificate to use when proxying https:// URLs
2447 NAME: sslproxy_client_key
2450 LOC: Config.ssl_client.key
2453 Client SSL Key to use when proxying https:// URLs
2456 NAME: sslproxy_version
2459 DEFAULT_DOC: automatic SSL/TLS version negotiation
2460 LOC: Config.ssl_client.version
2463 SSL version level to use when proxying https:// URLs
2465 The versions of SSL/TLS supported:
2467 1 automatic (default)
2474 NAME: sslproxy_options
2477 LOC: Config.ssl_client.options
2480 SSL implementation options to use when proxying https:// URLs
2482 The most important being:
2484 NO_SSLv3 Disallow the use of SSLv3
2485 NO_TLSv1 Disallow the use of TLSv1.0
2486 NO_TLSv1_1 Disallow the use of TLSv1.1
2487 NO_TLSv1_2 Disallow the use of TLSv1.2
2489 Always create a new key when using temporary/ephemeral
2492 Disable use of RFC5077 session tickets. Some servers
2493 may have problems understanding the TLS extension due
2494 to ambiguous specification in RFC4507.
2495 ALL Enable various bug workarounds suggested as "harmless"
2496 by OpenSSL. Be warned that this may reduce SSL/TLS
2497 strength to some attacks.
2499 See the OpenSSL SSL_CTX_set_options documentation for a
2500 complete list of possible options.
2503 NAME: sslproxy_cipher
2506 LOC: Config.ssl_client.cipher
2509 SSL cipher list to use when proxying https:// URLs
2511 Colon separated list of supported ciphers.
2514 NAME: sslproxy_cafile
2517 LOC: Config.ssl_client.cafile
2520 file containing CA certificates to use when verifying server
2521 certificates while proxying https:// URLs
2524 NAME: sslproxy_capath
2527 LOC: Config.ssl_client.capath
2530 directory containing CA certificates to use when verifying
2531 server certificates while proxying https:// URLs
2534 NAME: sslproxy_session_ttl
2537 LOC: Config.SSL.session_ttl
2540 Sets the timeout value for SSL sessions
2543 NAME: sslproxy_session_cache_size
2546 LOC: Config.SSL.sessionCacheSize
2549 Sets the cache size to use for ssl session
2552 NAME: sslproxy_cert_sign_hash
2555 LOC: Config.SSL.certSignHash
2558 Sets the hashing algorithm to use when signing generated certificates.
2559 Valid algorithm names depend on the OpenSSL library used. The following
2560 names are usually available: sha1, sha256, sha512, and md5. Please see
2561 your OpenSSL library manual for the available hashes. By default, Squids
2562 that support this option use sha256 hashes.
2564 Squid does not forcefully purge cached certificates that were generated
2565 with an algorithm other than the currently configured one. They remain
2566 in the cache, subject to the regular cache eviction policy, and become
2567 useful if the algorithm changes again.
2572 TYPE: sslproxy_ssl_bump
2573 LOC: Config.accessList.ssl_bump
2574 DEFAULT_DOC: Become a TCP tunnel without decrypting proxied traffic.
2577 This option is consulted when a CONNECT request is received on
2578 an http_port (or a new connection is intercepted at an
2579 https_port), provided that port was configured with an ssl-bump
2580 flag. The subsequent data on the connection is either treated as
2581 HTTPS and decrypted OR tunneled at TCP level without decryption,
2582 depending on the first matching bumping "action".
2584 ssl_bump <action> [!]acl ...
2586 The following bumping actions are currently supported:
2589 Become a TCP tunnel without decrypting proxied traffic.
2590 This is the default action.
2593 Establish a secure connection with the server and, using a
2594 mimicked server certificate, with the client.
2597 Receive client (step SslBump1) or server (step SslBump2)
2598 certificate while preserving the possibility of splicing the
2599 connection. Peeking at the server certificate (during step 2)
2600 usually precludes bumping of the connection at step 3.
2603 Receive client (step SslBump1) or server (step SslBump2)
2604 certificate while preserving the possibility of bumping the
2605 connection. Staring at the server certificate (during step 2)
2606 usually precludes splicing of the connection at step 3.
2609 Close client and server connections.
2611 Backward compatibility actions available at step SslBump1:
2614 Bump the connection. Establish a secure connection with the
2615 client first, then connect to the server. This old mode does
2616 not allow Squid to mimic server SSL certificate and does not
2617 work with intercepted SSL connections.
2620 Bump the connection. Establish a secure connection with the
2621 server first, then establish a secure connection with the
2622 client, using a mimicked server certificate. Works with both
2623 CONNECT requests and intercepted SSL connections, but does
2624 not allow to make decisions based on SSL handshake info.
2627 Decide whether to bump or splice the connection based on
2628 client-to-squid and server-to-squid SSL hello messages.
2632 Same as the "splice" action.
2634 All ssl_bump rules are evaluated at each of the supported bumping
2635 steps. Rules with actions that are impossible at the current step are
2636 ignored. The first matching ssl_bump action wins and is applied at the
2637 end of the current step. If no rules match, the splice action is used.
2638 See the at_step ACL for a list of the supported SslBump steps.
2640 This clause supports both fast and slow acl types.
2641 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2643 See also: http_port ssl-bump, https_port ssl-bump, and acl at_step.
2646 # Example: Bump all requests except those originating from
2647 # localhost or those going to example.com.
2649 acl broken_sites dstdomain .example.com
2650 ssl_bump splice localhost
2651 ssl_bump splice broken_sites
2655 NAME: sslproxy_flags
2658 LOC: Config.ssl_client.flags
2661 Various flags modifying the use of SSL while proxying https:// URLs:
2662 DONT_VERIFY_PEER Accept certificates that fail verification.
2663 For refined control, see sslproxy_cert_error.
2664 NO_DEFAULT_CA Don't use the default CA list built in
2668 NAME: sslproxy_cert_error
2671 DEFAULT_DOC: Server certificate errors terminate the transaction.
2672 LOC: Config.ssl_client.cert_error
2675 Use this ACL to bypass server certificate validation errors.
2677 For example, the following lines will bypass all validation errors
2678 when talking to servers for example.com. All other
2679 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2681 acl BrokenButTrustedServers dstdomain example.com
2682 sslproxy_cert_error allow BrokenButTrustedServers
2683 sslproxy_cert_error deny all
2685 This clause only supports fast acl types.
2686 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2687 Using slow acl types may result in server crashes
2689 Without this option, all server certificate validation errors
2690 terminate the transaction to protect Squid and the client.
2692 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2693 but should not happen unless your OpenSSL library is buggy.
2696 Bypassing validation errors is dangerous because an
2697 error usually implies that the server cannot be trusted
2698 and the connection may be insecure.
2700 See also: sslproxy_flags and DONT_VERIFY_PEER.
2703 NAME: sslproxy_cert_sign
2706 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2707 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2708 POSTSCRIPTUM: signTrusted all
2709 TYPE: sslproxy_cert_sign
2710 LOC: Config.ssl_client.cert_sign
2713 sslproxy_cert_sign <signing algorithm> acl ...
2715 The following certificate signing algorithms are supported:
2718 Sign using the configured CA certificate which is usually
2719 placed in and trusted by end-user browsers. This is the
2720 default for trusted origin server certificates.
2723 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2724 This is the default for untrusted origin server certificates
2725 that are not self-signed (see ssl::certUntrusted).
2728 Sign using a self-signed certificate with the right CN to
2729 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2730 browser. This is the default for self-signed origin server
2731 certificates (see ssl::certSelfSigned).
2733 This clause only supports fast acl types.
2735 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2736 signing algorithm to generate the certificate and ignores all
2737 subsequent sslproxy_cert_sign options (the first match wins). If no
2738 acl(s) match, the default signing algorithm is determined by errors
2739 detected when obtaining and validating the origin server certificate.
2741 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2742 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2743 CONNECT request that carries a domain name. In all other cases (CONNECT
2744 to an IP address or an intercepted SSL connection), Squid cannot detect
2745 the domain mismatch at certificate generation time when
2746 bump-server-first is used.
2749 NAME: sslproxy_cert_adapt
2752 TYPE: sslproxy_cert_adapt
2753 LOC: Config.ssl_client.cert_adapt
2756 sslproxy_cert_adapt <adaptation algorithm> acl ...
2758 The following certificate adaptation algorithms are supported:
2761 Sets the "Not After" property to the "Not After" property of
2762 the CA certificate used to sign generated certificates.
2765 Sets the "Not Before" property to the "Not Before" property of
2766 the CA certificate used to sign generated certificates.
2768 setCommonName or setCommonName{CN}
2769 Sets Subject.CN property to the host name specified as a
2770 CN parameter or, if no explicit CN parameter was specified,
2771 extracted from the CONNECT request. It is a misconfiguration
2772 to use setCommonName without an explicit parameter for
2773 intercepted or tproxied SSL connections.
2775 This clause only supports fast acl types.
2777 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2778 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2779 corresponding adaptation algorithm to generate the certificate and
2780 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2781 group (i.e., the first match wins within each algorithm group). If no
2782 acl(s) match, the default mimicking action takes place.
2784 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2785 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2786 CONNECT request that carries a domain name. In all other cases (CONNECT
2787 to an IP address or an intercepted SSL connection), Squid cannot detect
2788 the domain mismatch at certificate generation time when
2789 bump-server-first is used.
2792 NAME: sslpassword_program
2795 LOC: Config.Program.ssl_password
2798 Specify a program used for entering SSL key passphrases
2799 when using encrypted SSL certificate keys. If not specified
2800 keys must either be unencrypted, or Squid started with the -N
2801 option to allow it to query interactively for the passphrase.
2803 The key file name is given as argument to the program allowing
2804 selection of the right password if you have multiple encrypted
2809 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2810 -----------------------------------------------------------------------------
2813 NAME: sslcrtd_program
2816 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2817 LOC: Ssl::TheConfig.ssl_crtd
2819 Specify the location and options of the executable for ssl_crtd process.
2820 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2821 For more information use:
2822 @DEFAULT_SSL_CRTD@ -h
2825 NAME: sslcrtd_children
2826 TYPE: HelperChildConfig
2828 DEFAULT: 32 startup=5 idle=1
2829 LOC: Ssl::TheConfig.ssl_crtdChildren
2831 The maximum number of processes spawn to service ssl server.
2832 The maximum this may be safely set to is 32.
2834 The startup= and idle= options allow some measure of skew in your
2839 Sets the minimum number of processes to spawn when Squid
2840 starts or reconfigures. When set to zero the first request will
2841 cause spawning of the first child process to handle it.
2843 Starting too few children temporary slows Squid under load while it
2844 tries to spawn enough additional processes to cope with traffic.
2848 Sets a minimum of how many processes Squid is to try and keep available
2849 at all times. When traffic begins to rise above what the existing
2850 processes can handle this many more will be spawned up to the maximum
2851 configured. A minimum setting of 1 is required.
2855 Sets the maximum number of queued requests.
2856 If the queued requests exceed queue size for more than 3 minutes
2857 squid aborts its operation.
2858 The default value is set to 2*numberofchildren.
2860 You must have at least one ssl_crtd process.
2863 NAME: sslcrtvalidator_program
2867 LOC: Ssl::TheConfig.ssl_crt_validator
2869 Specify the location and options of the executable for ssl_crt_validator
2872 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2875 ttl=n TTL in seconds for cached results. The default is 60 secs
2876 cache=n limit the result cache size. The default value is 2048
2879 NAME: sslcrtvalidator_children
2880 TYPE: HelperChildConfig
2882 DEFAULT: 32 startup=5 idle=1 concurrency=1
2883 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2885 The maximum number of processes spawn to service SSL server.
2886 The maximum this may be safely set to is 32.
2888 The startup= and idle= options allow some measure of skew in your
2893 Sets the minimum number of processes to spawn when Squid
2894 starts or reconfigures. When set to zero the first request will
2895 cause spawning of the first child process to handle it.
2897 Starting too few children temporary slows Squid under load while it
2898 tries to spawn enough additional processes to cope with traffic.
2902 Sets a minimum of how many processes Squid is to try and keep available
2903 at all times. When traffic begins to rise above what the existing
2904 processes can handle this many more will be spawned up to the maximum
2905 configured. A minimum setting of 1 is required.
2909 The number of requests each certificate validator helper can handle in
2910 parallel. A value of 0 indicates the certficate validator does not
2911 support concurrency. Defaults to 1.
2913 When this directive is set to a value >= 1 then the protocol
2914 used to communicate with the helper is modified to include
2915 a request ID in front of the request/response. The request
2916 ID from the request must be echoed back with the response
2921 Sets the maximum number of queued requests.
2922 If the queued requests exceed queue size for more than 3 minutes
2923 squid aborts its operation.
2924 The default value is set to 2*numberofchildren.
2926 You must have at least one ssl_crt_validator process.
2930 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2931 -----------------------------------------------------------------------------
2939 To specify other caches in a hierarchy, use the format:
2941 cache_peer hostname type http-port icp-port [options]
2946 # hostname type port port options
2947 # -------------------- -------- ----- ----- -----------
2948 cache_peer parent.foo.net parent 3128 3130 default
2949 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2950 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2951 cache_peer example.com parent 80 0 default
2952 cache_peer cdn.example.com sibling 3128 0
2954 type: either 'parent', 'sibling', or 'multicast'.
2956 proxy-port: The port number where the peer accept HTTP requests.
2957 For other Squid proxies this is usually 3128
2958 For web servers this is usually 80
2960 icp-port: Used for querying neighbor caches about objects.
2961 Set to 0 if the peer does not support ICP or HTCP.
2962 See ICP and HTCP options below for additional details.
2965 ==== ICP OPTIONS ====
2967 You MUST also set icp_port and icp_access explicitly when using these options.
2968 The defaults will prevent peer traffic using ICP.
2971 no-query Disable ICP queries to this neighbor.
2974 Indicates the named peer is a member of a multicast group.
2975 ICP queries will not be sent directly to the peer, but ICP
2976 replies will be accepted from it.
2978 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2979 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2982 To only send ICP queries to this neighbor infrequently.
2983 This is used to keep the neighbor round trip time updated
2984 and is usually used in conjunction with weighted-round-robin.
2987 ==== HTCP OPTIONS ====
2989 You MUST also set htcp_port and htcp_access explicitly when using these options.
2990 The defaults will prevent peer traffic using HTCP.
2993 htcp Send HTCP, instead of ICP, queries to the neighbor.
2994 You probably also want to set the "icp-port" to 4827
2995 instead of 3130. This directive accepts a comma separated
2996 list of options described below.
2998 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
3000 htcp=no-clr Send HTCP to the neighbor but without
3001 sending any CLR requests. This cannot be used with
3004 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
3005 This cannot be used with no-clr.
3008 Send HTCP to the neighbor including CLRs but only when
3009 they do not result from PURGE requests.
3012 Forward any HTCP CLR requests this proxy receives to the peer.
3015 ==== PEER SELECTION METHODS ====
3017 The default peer selection method is ICP, with the first responding peer
3018 being used as source. These options can be used for better load balancing.
3021 default This is a parent cache which can be used as a "last-resort"
3022 if a peer cannot be located by any of the peer-selection methods.
3023 If specified more than once, only the first is used.
3025 round-robin Load-Balance parents which should be used in a round-robin
3026 fashion in the absence of any ICP queries.
3027 weight=N can be used to add bias.
3029 weighted-round-robin
3030 Load-Balance parents which should be used in a round-robin
3031 fashion with the frequency of each parent being based on the
3032 round trip time. Closer parents are used more often.
3033 Usually used for background-ping parents.
3034 weight=N can be used to add bias.
3036 carp Load-Balance parents which should be used as a CARP array.
3037 The requests will be distributed among the parents based on the
3038 CARP load balancing hash function based on their weight.
3040 userhash Load-balance parents based on the client proxy_auth or ident username.
3042 sourcehash Load-balance parents based on the client source IP.
3045 To be used only for cache peers of type "multicast".
3046 ALL members of this multicast group have "sibling"
3047 relationship with it, not "parent". This is to a multicast
3048 group when the requested object would be fetched only from
3049 a "parent" cache, anyway. It's useful, e.g., when
3050 configuring a pool of redundant Squid proxies, being
3051 members of the same multicast group.
3054 ==== PEER SELECTION OPTIONS ====
3056 weight=N use to affect the selection of a peer during any weighted
3057 peer-selection mechanisms.
3058 The weight must be an integer; default is 1,
3059 larger weights are favored more.
3060 This option does not affect parent selection if a peering
3061 protocol is not in use.
3063 basetime=N Specify a base amount to be subtracted from round trip
3065 It is subtracted before division by weight in calculating
3066 which parent to fectch from. If the rtt is less than the
3067 base time the rtt is set to a minimal value.
3069 ttl=N Specify a TTL to use when sending multicast ICP queries
3071 Only useful when sending to a multicast group.
3072 Because we don't accept ICP replies from random
3073 hosts, you must configure other group members as
3074 peers with the 'multicast-responder' option.
3076 no-delay To prevent access to this neighbor from influencing the
3079 digest-url=URL Tell Squid to fetch the cache digest (if digests are
3080 enabled) for this host from the specified URL rather
3081 than the Squid default location.
3084 ==== CARP OPTIONS ====
3086 carp-key=key-specification
3087 use a different key than the full URL to hash against the peer.
3088 the key-specification is a comma-separated list of the keywords
3089 scheme, host, port, path, params
3090 Order is not important.
3092 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
3094 originserver Causes this parent to be contacted as an origin server.
3095 Meant to be used in accelerator setups when the peer
3099 Set the Host header of requests forwarded to this peer.
3100 Useful in accelerator setups where the server (peer)
3101 expects a certain domain name but clients may request
3102 others. ie example.com or www.example.com
3104 no-digest Disable request of cache digests.
3107 Disables requesting ICMP RTT database (NetDB).
3110 ==== AUTHENTICATION OPTIONS ====
3113 If this is a personal/workgroup proxy and your parent
3114 requires proxy authentication.
3116 Note: The string can include URL escapes (i.e. %20 for
3117 spaces). This also means % must be written as %%.
3120 Send login details received from client to this peer.
3121 Both Proxy- and WWW-Authorization headers are passed
3122 without alteration to the peer.
3123 Authentication is not required by Squid for this to work.
3125 Note: This will pass any form of authentication but
3126 only Basic auth will work through a proxy unless the
3127 connection-auth options are also used.
3129 login=PASS Send login details received from client to this peer.
3130 Authentication is not required by this option.
3132 If there are no client-provided authentication headers
3133 to pass on, but username and password are available
3134 from an external ACL user= and password= result tags
3135 they may be sent instead.
3137 Note: To combine this with proxy_auth both proxies must
3138 share the same user database as HTTP only allows for
3139 a single login (one for proxy, one for origin server).
3140 Also be warned this will expose your users proxy
3141 password to the peer. USE WITH CAUTION
3144 Send the username to the upstream cache, but with a
3145 fixed password. This is meant to be used when the peer
3146 is in another administrative domain, but it is still
3147 needed to identify each user.
3148 The star can optionally be followed by some extra
3149 information which is added to the username. This can
3150 be used to identify this proxy to the peer, similar to
3151 the login=username:password option above.
3154 If this is a personal/workgroup proxy and your parent
3155 requires a secure proxy authentication.
3156 The first principal from the default keytab or defined by
3157 the environment variable KRB5_KTNAME will be used.
3159 WARNING: The connection may transmit requests from multiple
3160 clients. Negotiate often assumes end-to-end authentication
3161 and a single-client. Which is not strictly true here.
3163 login=NEGOTIATE:principal_name
3164 If this is a personal/workgroup proxy and your parent
3165 requires a secure proxy authentication.
3166 The principal principal_name from the default keytab or
3167 defined by the environment variable KRB5_KTNAME will be
3170 WARNING: The connection may transmit requests from multiple
3171 clients. Negotiate often assumes end-to-end authentication
3172 and a single-client. Which is not strictly true here.
3174 connection-auth=on|off
3175 Tell Squid that this peer does or not support Microsoft
3176 connection oriented authentication, and any such
3177 challenges received from there should be ignored.
3178 Default is auto to automatically determine the status
3182 ==== SSL / HTTPS / TLS OPTIONS ====
3184 ssl Encrypt connections to this peer with SSL/TLS.
3186 sslcert=/path/to/ssl/certificate
3187 A client SSL certificate to use when connecting to
3190 sslkey=/path/to/ssl/key
3191 The private SSL key corresponding to sslcert above.
3192 If 'sslkey' is not specified 'sslcert' is assumed to
3193 reference a combined file containing both the
3194 certificate and the key.
3196 sslversion=1|3|4|5|6
3197 The SSL version to use when connecting to this peer
3198 1 = automatic (default)
3204 sslcipher=... The list of valid SSL ciphers to use when connecting
3207 ssloptions=... Specify various SSL implementation options:
3209 NO_SSLv3 Disallow the use of SSLv3
3210 NO_TLSv1 Disallow the use of TLSv1.0
3211 NO_TLSv1_1 Disallow the use of TLSv1.1
3212 NO_TLSv1_2 Disallow the use of TLSv1.2
3214 Always create a new key when using
3215 temporary/ephemeral DH key exchanges
3216 ALL Enable various bug workarounds
3217 suggested as "harmless" by OpenSSL
3218 Be warned that this reduces SSL/TLS
3219 strength to some attacks.
3221 See the OpenSSL SSL_CTX_set_options documentation for a
3224 sslcafile=... A file containing additional CA certificates to use
3225 when verifying the peer certificate.
3227 sslcapath=... A directory containing additional CA certificates to
3228 use when verifying the peer certificate.
3230 sslcrlfile=... A certificate revocation list file to use when
3231 verifying the peer certificate.
3233 sslflags=... Specify various flags modifying the SSL implementation:
3236 Accept certificates even if they fail to
3239 Don't use the default CA list built in
3242 Don't verify the peer certificate
3243 matches the server name
3245 ssldomain= The peer name as advertised in it's certificate.
3246 Used for verifying the correctness of the received peer
3247 certificate. If not specified the peer hostname will be
3251 Enable the "Front-End-Https: On" header needed when
3252 using Squid as a SSL frontend in front of Microsoft OWA.
3253 See MS KB document Q307347 for details on this header.
3254 If set to auto the header will only be added if the
3255 request is forwarded as a https:// URL.
3258 ==== GENERAL OPTIONS ====
3261 A peer-specific connect timeout.
3262 Also see the peer_connect_timeout directive.
3264 connect-fail-limit=N
3265 How many times connecting to a peer must fail before
3266 it is marked as down. Standby connection failures
3267 count towards this limit. Default is 10.
3269 allow-miss Disable Squid's use of only-if-cached when forwarding
3270 requests to siblings. This is primarily useful when
3271 icp_hit_stale is used by the sibling. To extensive use
3272 of this option may result in forwarding loops, and you
3273 should avoid having two-way peerings with this option.
3274 For example to deny peer usage on requests from peer
3275 by denying cache_peer_access if the source is a peer.
3277 max-conn=N Limit the number of concurrent connections the Squid
3278 may open to this peer, including already opened idle
3279 and standby connections. There is no peer-specific
3280 connection limit by default.
3282 A peer exceeding the limit is not used for new
3283 requests unless a standby connection is available.
3285 max-conn currently works poorly with idle persistent
3286 connections: When a peer reaches its max-conn limit,
3287 and there are idle persistent connections to the peer,
3288 the peer may not be selected because the limiting code
3289 does not know whether Squid can reuse those idle
3292 standby=N Maintain a pool of N "hot standby" connections to an
3293 UP peer, available for requests when no idle
3294 persistent connection is available (or safe) to use.
3295 By default and with zero N, no such pool is maintained.
3296 N must not exceed the max-conn limit (if any).
3298 At start or after reconfiguration, Squid opens new TCP
3299 standby connections until there are N connections
3300 available and then replenishes the standby pool as
3301 opened connections are used up for requests. A used
3302 connection never goes back to the standby pool, but
3303 may go to the regular idle persistent connection pool
3304 shared by all peers and origin servers.
3306 Squid never opens multiple new standby connections
3307 concurrently. This one-at-a-time approach minimizes
3308 flooding-like effect on peers. Furthermore, just a few
3309 standby connections should be sufficient in most cases
3310 to supply most new requests with a ready-to-use
3313 Standby connections obey server_idle_pconn_timeout.
3314 For the feature to work as intended, the peer must be
3315 configured to accept and keep them open longer than
3316 the idle timeout at the connecting Squid, to minimize
3317 race conditions typical to idle used persistent
3318 connections. Default request_timeout and
3319 server_idle_pconn_timeout values ensure such a
3322 name=xxx Unique name for the peer.
3323 Required if you have multiple peers on the same host
3324 but different ports.
3325 This name can be used in cache_peer_access and similar
3326 directives to dentify the peer.
3327 Can be used by outgoing access controls through the
3330 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3331 requests to this peer. Use normal address selection instead.
3332 This overrides the spoof_client_ip ACL.
3334 proxy-only objects fetched from the peer will not be stored locally.
3338 NAME: cache_peer_domain cache_host_domain
3343 Use to limit the domains for which a neighbor cache will be
3347 cache_peer_domain cache-host domain [domain ...]
3348 cache_peer_domain cache-host !domain
3350 For example, specifying
3352 cache_peer_domain parent.foo.net .edu
3354 has the effect such that UDP query packets are sent to
3355 'bigserver' only when the requested object exists on a
3356 server in the .edu domain. Prefixing the domainname
3357 with '!' means the cache will be queried for objects
3360 NOTE: * Any number of domains may be given for a cache-host,
3361 either on the same or separate lines.
3362 * When multiple domains are given for a particular
3363 cache-host, the first matched domain is applied.
3364 * Cache hosts with no domain restrictions are queried
3366 * There are no defaults.
3367 * There is also a 'cache_peer_access' tag in the ACL
3371 NAME: cache_peer_access
3376 Similar to 'cache_peer_domain' but provides more flexibility by
3380 cache_peer_access cache-host allow|deny [!]aclname ...
3382 The syntax is identical to 'http_access' and the other lists of
3383 ACL elements. See the comments for 'http_access' below, or
3384 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3387 NAME: neighbor_type_domain
3388 TYPE: hostdomaintype
3390 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3393 Modify the cache_peer neighbor type when passing requests
3394 about specific domains to the peer.
3397 neighbor_type_domain neighbor parent|sibling domain domain ...
3400 cache_peer foo.example.com parent 3128 3130
3401 neighbor_type_domain foo.example.com sibling .au .de
3403 The above configuration treats all requests to foo.example.com as a
3404 parent proxy unless the request is for a .au or .de ccTLD domain name.
3407 NAME: dead_peer_timeout
3411 LOC: Config.Timeout.deadPeer
3413 This controls how long Squid waits to declare a peer cache
3414 as "dead." If there are no ICP replies received in this
3415 amount of time, Squid will declare the peer dead and not
3416 expect to receive any further ICP replies. However, it
3417 continues to send ICP queries, and will mark the peer as
3418 alive upon receipt of the first subsequent ICP reply.
3420 This timeout also affects when Squid expects to receive ICP
3421 replies from peers. If more than 'dead_peer' seconds have
3422 passed since the last ICP reply was received, Squid will not
3423 expect to receive an ICP reply on the next query. Thus, if
3424 your time between requests is greater than this timeout, you
3425 will see a lot of requests sent DIRECT to origin servers
3426 instead of to your parents.
3429 NAME: forward_max_tries
3432 LOC: Config.forward_max_tries
3434 Controls how many different forward paths Squid will try
3435 before giving up. See also forward_timeout.
3437 NOTE: connect_retries (default: none) can make each of these
3438 possible forwarding paths be tried multiple times.
3442 MEMORY CACHE OPTIONS
3443 -----------------------------------------------------------------------------
3450 LOC: Config.memMaxSize
3452 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3453 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3454 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3455 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3457 'cache_mem' specifies the ideal amount of memory to be used
3459 * In-Transit objects
3461 * Negative-Cached objects
3463 Data for these objects are stored in 4 KB blocks. This
3464 parameter specifies the ideal upper limit on the total size of
3465 4 KB blocks allocated. In-Transit objects take the highest
3468 In-transit objects have priority over the others. When
3469 additional space is needed for incoming data, negative-cached
3470 and hot objects will be released. In other words, the
3471 negative-cached and hot objects will fill up any unused space
3472 not needed for in-transit objects.
3474 If circumstances require, this limit will be exceeded.
3475 Specifically, if your incoming request rate requires more than
3476 'cache_mem' of memory to hold in-transit objects, Squid will
3477 exceed this limit to satisfy the new requests. When the load
3478 decreases, blocks will be freed until the high-water mark is
3479 reached. Thereafter, blocks will be used to store hot
3482 If shared memory caching is enabled, Squid does not use the shared
3483 cache space for in-transit objects, but they still consume as much
3484 local memory as they need. For more details about the shared memory
3485 cache, see memory_cache_shared.
3488 NAME: maximum_object_size_in_memory
3492 LOC: Config.Store.maxInMemObjSize
3494 Objects greater than this size will not be attempted to kept in
3495 the memory cache. This should be set high enough to keep objects
3496 accessed frequently in memory to improve performance whilst low
3497 enough to keep larger objects from hoarding cache_mem.
3500 NAME: memory_cache_shared
3503 LOC: Config.memShared
3505 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3507 Controls whether the memory cache is shared among SMP workers.
3509 The shared memory cache is meant to occupy cache_mem bytes and replace
3510 the non-shared memory cache, although some entities may still be
3511 cached locally by workers for now (e.g., internal and in-transit
3512 objects may be served from a local memory cache even if shared memory
3513 caching is enabled).
3515 By default, the memory cache is shared if and only if all of the
3516 following conditions are satisfied: Squid runs in SMP mode with
3517 multiple workers, cache_mem is positive, and Squid environment
3518 supports required IPC primitives (e.g., POSIX shared memory segments
3519 and GCC-style atomic operations).
3521 To avoid blocking locks, shared memory uses opportunistic algorithms
3522 that do not guarantee that every cachable entity that could have been
3523 shared among SMP workers will actually be shared.
3525 Currently, entities exceeding 32KB in size cannot be shared.
3528 NAME: memory_cache_mode
3532 DEFAULT_DOC: Keep the most recently fetched objects in memory
3534 Controls which objects to keep in the memory cache (cache_mem)
3536 always Keep most recently fetched objects in memory (default)
3538 disk Only disk cache hits are kept in memory, which means
3539 an object must first be cached on disk and then hit
3540 a second time before cached in memory.
3542 network Only objects fetched from network is kept in memory
3545 NAME: memory_replacement_policy
3547 LOC: Config.memPolicy
3550 The memory replacement policy parameter determines which
3551 objects are purged from memory when memory space is needed.
3553 See cache_replacement_policy for details on algorithms.
3558 -----------------------------------------------------------------------------
3561 NAME: cache_replacement_policy
3563 LOC: Config.replPolicy
3566 The cache replacement policy parameter determines which
3567 objects are evicted (replaced) when disk space is needed.
3569 lru : Squid's original list based LRU policy
3570 heap GDSF : Greedy-Dual Size Frequency
3571 heap LFUDA: Least Frequently Used with Dynamic Aging
3572 heap LRU : LRU policy implemented using a heap
3574 Applies to any cache_dir lines listed below this directive.
3576 The LRU policies keeps recently referenced objects.
3578 The heap GDSF policy optimizes object hit rate by keeping smaller
3579 popular objects in cache so it has a better chance of getting a
3580 hit. It achieves a lower byte hit rate than LFUDA though since
3581 it evicts larger (possibly popular) objects.
3583 The heap LFUDA policy keeps popular objects in cache regardless of
3584 their size and thus optimizes byte hit rate at the expense of
3585 hit rate since one large, popular object will prevent many
3586 smaller, slightly less popular objects from being cached.
3588 Both policies utilize a dynamic aging mechanism that prevents
3589 cache pollution that can otherwise occur with frequency-based
3590 replacement policies.
3592 NOTE: if using the LFUDA replacement policy you should increase
3593 the value of maximum_object_size above its default of 4 MB to
3594 to maximize the potential byte hit rate improvement of LFUDA.
3596 For more information about the GDSF and LFUDA cache replacement
3597 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3598 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3601 NAME: minimum_object_size
3605 DEFAULT_DOC: no limit
3606 LOC: Config.Store.minObjectSize
3608 Objects smaller than this size will NOT be saved on disk. The
3609 value is specified in bytes, and the default is 0 KB, which
3610 means all responses can be stored.
3613 NAME: maximum_object_size
3617 LOC: Config.Store.maxObjectSize
3619 Set the default value for max-size parameter on any cache_dir.
3620 The value is specified in bytes, and the default is 4 MB.
3622 If you wish to get a high BYTES hit ratio, you should probably
3623 increase this (one 32 MB object hit counts for 3200 10KB
3626 If you wish to increase hit ratio more than you want to
3627 save bandwidth you should leave this low.
3629 NOTE: if using the LFUDA replacement policy you should increase
3630 this value to maximize the byte hit rate improvement of LFUDA!
3631 See cache_replacement_policy for a discussion of this policy.
3637 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3638 LOC: Config.cacheSwap
3641 cache_dir Type Directory-Name Fs-specific-data [options]
3643 You can specify multiple cache_dir lines to spread the
3644 cache among different disk partitions.
3646 Type specifies the kind of storage system to use. Only "ufs"
3647 is built by default. To enable any of the other storage systems
3648 see the --enable-storeio configure option.
3650 'Directory' is a top-level directory where cache swap
3651 files will be stored. If you want to use an entire disk
3652 for caching, this can be the mount-point directory.
3653 The directory must exist and be writable by the Squid
3654 process. Squid will NOT create this directory for you.
3656 In SMP configurations, cache_dir must not precede the workers option
3657 and should use configuration macros or conditionals to give each
3658 worker interested in disk caching a dedicated cache directory.
3661 ==== The ufs store type ====
3663 "ufs" is the old well-known Squid storage format that has always
3667 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3669 'Mbytes' is the amount of disk space (MB) to use under this
3670 directory. The default is 100 MB. Change this to suit your
3671 configuration. Do NOT put the size of your disk drive here.
3672 Instead, if you want Squid to use the entire disk drive,
3673 subtract 20% and use that value.
3675 'L1' is the number of first-level subdirectories which
3676 will be created under the 'Directory'. The default is 16.
3678 'L2' is the number of second-level subdirectories which
3679 will be created under each first-level directory. The default
3683 ==== The aufs store type ====
3685 "aufs" uses the same storage format as "ufs", utilizing
3686 POSIX-threads to avoid blocking the main Squid process on
3687 disk-I/O. This was formerly known in Squid as async-io.
3690 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3692 see argument descriptions under ufs above
3695 ==== The diskd store type ====
3697 "diskd" uses the same storage format as "ufs", utilizing a
3698 separate process to avoid blocking the main Squid process on
3702 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3704 see argument descriptions under ufs above
3706 Q1 specifies the number of unacknowledged I/O requests when Squid
3707 stops opening new files. If this many messages are in the queues,
3708 Squid won't open new files. Default is 64
3710 Q2 specifies the number of unacknowledged messages when Squid
3711 starts blocking. If this many messages are in the queues,
3712 Squid blocks until it receives some replies. Default is 72
3714 When Q1 < Q2 (the default), the cache directory is optimized
3715 for lower response time at the expense of a decrease in hit
3716 ratio. If Q1 > Q2, the cache directory is optimized for
3717 higher hit ratio at the expense of an increase in response
3721 ==== The rock store type ====
3724 cache_dir rock Directory-Name Mbytes [options]
3726 The Rock Store type is a database-style storage. All cached
3727 entries are stored in a "database" file, using fixed-size slots.
3728 A single entry occupies one or more slots.
3730 If possible, Squid using Rock Store creates a dedicated kid
3731 process called "disker" to avoid blocking Squid worker(s) on disk
3732 I/O. One disker kid is created for each rock cache_dir. Diskers
3733 are created only when Squid, running in daemon mode, has support
3734 for the IpcIo disk I/O module.
3736 swap-timeout=msec: Squid will not start writing a miss to or
3737 reading a hit from disk if it estimates that the swap operation
3738 will take more than the specified number of milliseconds. By
3739 default and when set to zero, disables the disk I/O time limit
3740 enforcement. Ignored when using blocking I/O module because
3741 blocking synchronous I/O does not allow Squid to estimate the
3742 expected swap wait time.
3744 max-swap-rate=swaps/sec: Artificially limits disk access using
3745 the specified I/O rate limit. Swap out requests that
3746 would cause the average I/O rate to exceed the limit are
3747 delayed. Individual swap in requests (i.e., hits or reads) are
3748 not delayed, but they do contribute to measured swap rate and
3749 since they are placed in the same FIFO queue as swap out
3750 requests, they may wait longer if max-swap-rate is smaller.
3751 This is necessary on file systems that buffer "too
3752 many" writes and then start blocking Squid and other processes
3753 while committing those writes to disk. Usually used together
3754 with swap-timeout to avoid excessive delays and queue overflows
3755 when disk demand exceeds available disk "bandwidth". By default
3756 and when set to zero, disables the disk I/O rate limit
3757 enforcement. Currently supported by IpcIo module only.
3759 slot-size=bytes: The size of a database "record" used for
3760 storing cached responses. A cached response occupies at least
3761 one slot and all database I/O is done using individual slots so
3762 increasing this parameter leads to more disk space waste while
3763 decreasing it leads to more disk I/O overheads. Should be a
3764 multiple of your operating system I/O page size. Defaults to
3765 16KBytes. A housekeeping header is stored with each slot and
3766 smaller slot-sizes will be rejected. The header is smaller than
3770 ==== COMMON OPTIONS ====
3772 no-store no new objects should be stored to this cache_dir.
3774 min-size=n the minimum object size in bytes this cache_dir
3775 will accept. It's used to restrict a cache_dir
3776 to only store large objects (e.g. AUFS) while
3777 other stores are optimized for smaller objects
3781 max-size=n the maximum object size in bytes this cache_dir
3783 The value in maximum_object_size directive sets
3784 the default unless more specific details are
3785 available (ie a small store capacity).
3787 Note: To make optimal use of the max-size limits you should order
3788 the cache_dir lines with the smallest max-size value first.
3792 # Uncomment and adjust the following to add a disk cache directory.
3793 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3797 NAME: store_dir_select_algorithm
3799 LOC: Config.store_dir_select_algorithm
3802 How Squid selects which cache_dir to use when the response
3803 object will fit into more than one.
3805 Regardless of which algorithm is used the cache_dir min-size
3806 and max-size parameters are obeyed. As such they can affect
3807 the selection algorithm by limiting the set of considered
3814 This algorithm is suited to caches with similar cache_dir
3815 sizes and disk speeds.
3817 The disk with the least I/O pending is selected.
3818 When there are multiple disks with the same I/O load ranking
3819 the cache_dir with most available capacity is selected.
3821 When a mix of cache_dir sizes are configured the faster disks
3822 have a naturally lower I/O loading and larger disks have more
3823 capacity. So space used to store objects and data throughput
3824 may be very unbalanced towards larger disks.
3829 This algorithm is suited to caches with unequal cache_dir
3832 Each cache_dir is selected in a rotation. The next suitable
3835 Available cache_dir capacity is only considered in relation
3836 to whether the object will fit and meets the min-size and
3837 max-size parameters.
3839 Disk I/O loading is only considered to prevent overload on slow
3840 disks. This algorithm does not spread objects by size, so any
3841 I/O loading per-disk may appear very unbalanced and volatile.
3843 If several cache_dirs use similar min-size, max-size, or other
3844 limits to to reject certain responses, then do not group such
3845 cache_dir lines together, to avoid round-robin selection bias
3846 towards the first cache_dir after the group. Instead, interleave
3847 cache_dir lines from different groups. For example:
3849 store_dir_select_algorithm round-robin
3850 cache_dir rock /hdd1 ... min-size=100000
3851 cache_dir rock /ssd1 ... max-size=99999
3852 cache_dir rock /hdd2 ... min-size=100000
3853 cache_dir rock /ssd2 ... max-size=99999
3854 cache_dir rock /hdd3 ... min-size=100000
3855 cache_dir rock /ssd3 ... max-size=99999
3858 NAME: max_open_disk_fds
3860 LOC: Config.max_open_disk_fds
3862 DEFAULT_DOC: no limit
3864 To avoid having disk as the I/O bottleneck Squid can optionally
3865 bypass the on-disk cache if more than this amount of disk file
3866 descriptors are open.
3868 A value of 0 indicates no limit.
3871 NAME: cache_swap_low
3872 COMMENT: (percent, 0-100)
3875 LOC: Config.Swap.lowWaterMark
3877 The low-water mark for cache object replacement.
3878 Replacement begins when the swap (disk) usage is above the
3879 low-water mark and attempts to maintain utilization near the
3880 low-water mark. As swap utilization gets close to high-water
3881 mark object eviction becomes more aggressive. If utilization is
3882 close to the low-water mark less replacement is done each time.
3884 Defaults are 90% and 95%. If you have a large cache, 5% could be
3885 hundreds of MB. If this is the case you may wish to set these
3886 numbers closer together.
3888 See also cache_swap_high
3891 NAME: cache_swap_high
3892 COMMENT: (percent, 0-100)
3895 LOC: Config.Swap.highWaterMark
3897 The high-water mark for cache object replacement.
3898 Replacement begins when the swap (disk) usage is above the
3899 low-water mark and attempts to maintain utilization near the
3900 low-water mark. As swap utilization gets close to high-water
3901 mark object eviction becomes more aggressive. If utilization is
3902 close to the low-water mark less replacement is done each time.
3904 Defaults are 90% and 95%. If you have a large cache, 5% could be
3905 hundreds of MB. If this is the case you may wish to set these
3906 numbers closer together.
3908 See also cache_swap_low
3913 -----------------------------------------------------------------------------
3920 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
3924 logformat <name> <format specification>
3926 Defines an access log format.
3928 The <format specification> is a string with embedded % format codes
3930 % format codes all follow the same basic structure where all but
3931 the formatcode is optional. Output strings are automatically escaped
3932 as required according to their context and the output format
3933 modifiers are usually not needed, but can be specified if an explicit
3934 output format is desired.
3936 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3938 " output in quoted string format
3939 [ output in squid text log format as used by log_mime_hdrs
3940 # output in URL quoted format
3945 width minimum and/or maximum field width:
3946 [width_min][.width_max]
3947 When minimum starts with 0, the field is zero-padded.
3948 String values exceeding maximum width are truncated.
3950 {arg} argument such as header name etc
3954 % a literal % character
3955 sn Unique sequence number per log line entry
3956 err_code The ID of an error response served by Squid or
3957 a similar internal error identifier.
3958 err_detail Additional err_code-dependent error information.
3959 note The annotation specified by the argument. Also
3960 logs the adaptation meta headers set by the
3961 adaptation_meta configuration parameter.
3962 If no argument given all annotations logged.
3963 The argument may include a separator to use with
3966 By default, multiple note values are separated with ","
3967 and multiple notes are separated with "\r\n".
3968 When logging named notes with %{name}note, the
3969 explicitly configured separator is used between note
3970 values. When logging all notes with %note, the
3971 explicitly configured separator is used between
3972 individual notes. There is currently no way to
3973 specify both value and notes separators when logging
3974 all notes with %note.
3976 Connection related format codes:
3978 >a Client source IP address
3980 >p Client source port
3981 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3982 >la Local IP address the client connected to
3983 >lp Local port number the client connected to
3984 >qos Client connection TOS/DSCP value set by Squid
3985 >nfmark Client connection netfilter mark set by Squid
3987 la Local listening IP address the client connection was connected to.
3988 lp Local listening port number the client connection was connected to.
3990 <a Server IP address of the last server or peer connection
3991 <A Server FQDN or peer name
3992 <p Server port number of the last server or peer connection
3993 <la Local IP address of the last server or peer connection
3994 <lp Local port number of the last server or peer connection
3995 <qos Server connection TOS/DSCP value set by Squid
3996 <nfmark Server connection netfilter mark set by Squid
3998 Time related format codes:
4000 ts Seconds since epoch
4001 tu subsecond time (milliseconds)
4002 tl Local time. Optional strftime format argument
4003 default %d/%b/%Y:%H:%M:%S %z
4004 tg GMT time. Optional strftime format argument
4005 default %d/%b/%Y:%H:%M:%S %z
4006 tr Response time (milliseconds)
4007 dt Total time spent making DNS lookups (milliseconds)
4008 tS Approximate master transaction start time in
4009 <full seconds since epoch>.<fractional seconds> format.
4010 Currently, Squid considers the master transaction
4011 started when a complete HTTP request header initiating
4012 the transaction is received from the client. This is
4013 the same value that Squid uses to calculate transaction
4014 response time when logging %tr to access.log. Currently,
4015 Squid uses millisecond resolution for %tS values,
4016 similar to the default access.log "current time" field
4019 Access Control related format codes:
4021 et Tag returned by external acl
4022 ea Log string returned by external acl
4023 un User name (any available)
4024 ul User name from authentication
4025 ue User name from external acl helper
4026 ui User name from ident
4027 us User name from SSL
4028 credentials Client credentials. The exact meaning depends on
4029 the authentication scheme: For Basic authentication,
4030 it is the password; for Digest, the realm sent by the
4031 client; for NTLM and Negotiate, the client challenge
4032 or client credentials prefixed with "YR " or "KK ".
4034 HTTP related format codes:
4038 [http::]rm Request method (GET/POST etc)
4039 [http::]>rm Request method from client
4040 [http::]<rm Request method sent to server or peer
4041 [http::]ru Request URL from client (historic, filtered for logging)
4042 [http::]>ru Request URL from client
4043 [http::]<ru Request URL sent to server or peer
4044 [http::]>rs Request URL scheme from client
4045 [http::]<rs Request URL scheme sent to server or peer
4046 [http::]>rd Request URL domain from client
4047 [http::]<rd Request URL domain sent to server or peer
4048 [http::]>rP Request URL port from client
4049 [http::]<rP Request URL port sent to server or peer
4050 [http::]rp Request URL path excluding hostname
4051 [http::]>rp Request URL path excluding hostname from client
4052 [http::]<rp Request URL path excluding hostname sent to server or peer
4053 [http::]rv Request protocol version
4054 [http::]>rv Request protocol version from client
4055 [http::]<rv Request protocol version sent to server or peer
4057 [http::]>h Original received request header.
4058 Usually differs from the request header sent by
4059 Squid, although most fields are often preserved.
4060 Accepts optional header field name/value filter
4061 argument using name[:[separator]element] format.
4062 [http::]>ha Received request header after adaptation and
4063 redirection (pre-cache REQMOD vectoring point).
4064 Usually differs from the request header sent by
4065 Squid, although most fields are often preserved.
4066 Optional header name argument as for >h
4071 [http::]<Hs HTTP status code received from the next hop
4072 [http::]>Hs HTTP status code sent to the client
4074 [http::]<h Reply header. Optional header name argument
4077 [http::]mt MIME content type
4082 [http::]st Total size of request + reply traffic with client
4083 [http::]>st Total size of request received from client.
4084 Excluding chunked encoding bytes.
4085 [http::]<st Total size of reply sent to client (after adaptation)
4087 [http::]>sh Size of request headers received from client
4088 [http::]<sh Size of reply headers sent to client (after adaptation)
4090 [http::]<sH Reply high offset sent
4091 [http::]<sS Upstream object size
4093 [http::]<bs Number of HTTP-equivalent message body bytes
4094 received from the next hop, excluding chunked
4095 transfer encoding and control messages.
4096 Generated FTP/Gopher listings are treated as
4102 [http::]<pt Peer response time in milliseconds. The timer starts
4103 when the last request byte is sent to the next hop
4104 and stops when the last response byte is received.
4105 [http::]<tt Total time in milliseconds. The timer
4106 starts with the first connect request (or write I/O)
4107 sent to the first selected peer. The timer stops
4108 with the last I/O with the last peer.
4110 Squid handling related format codes:
4112 Ss Squid request status (TCP_MISS etc)
4113 Sh Squid hierarchy status (DEFAULT_PARENT etc)
4115 SSL-related format codes:
4117 ssl::bump_mode SslBump decision for the transaction:
4119 For CONNECT requests that initiated bumping of
4120 a connection and for any request received on
4121 an already bumped connection, Squid logs the
4122 corresponding SslBump mode ("server-first" or
4123 "client-first"). See the ssl_bump option for
4124 more information about these modes.
4126 A "none" token is logged for requests that
4127 triggered "ssl_bump" ACL evaluation matching
4128 either a "none" rule or no rules at all.
4130 In all other cases, a single dash ("-") is
4133 ssl::>sni SSL client SNI sent to Squid. Available only
4134 after the peek, stare, or splice SSL bumping
4137 If ICAP is enabled, the following code becomes available (as
4138 well as ICAP log codes documented with the icap_log option):
4140 icap::tt Total ICAP processing time for the HTTP
4141 transaction. The timer ticks when ICAP
4142 ACLs are checked and when ICAP
4143 transaction is in progress.
4145 If adaptation is enabled the following three codes become available:
4147 adapt::<last_h The header of the last ICAP response or
4148 meta-information from the last eCAP
4149 transaction related to the HTTP transaction.
4150 Like <h, accepts an optional header name
4153 adapt::sum_trs Summed adaptation transaction response
4154 times recorded as a comma-separated list in
4155 the order of transaction start time. Each time
4156 value is recorded as an integer number,
4157 representing response time of one or more
4158 adaptation (ICAP or eCAP) transaction in
4159 milliseconds. When a failed transaction is
4160 being retried or repeated, its time is not
4161 logged individually but added to the
4162 replacement (next) transaction. See also:
4165 adapt::all_trs All adaptation transaction response times.
4166 Same as adaptation_strs but response times of
4167 individual transactions are never added
4168 together. Instead, all transaction response
4169 times are recorded individually.
4171 You can prefix adapt::*_trs format codes with adaptation
4172 service name in curly braces to record response time(s) specific
4173 to that service. For example: %{my_service}adapt::sum_trs
4175 If SSL is enabled, the following formating codes become available:
4177 %ssl::>cert_subject The Subject field of the received client
4178 SSL certificate or a dash ('-') if Squid has
4179 received an invalid/malformed certificate or
4180 no certificate at all. Consider encoding the
4181 logged value because Subject often has spaces.
4183 %ssl::>cert_issuer The Issuer field of the received client
4184 SSL certificate or a dash ('-') if Squid has
4185 received an invalid/malformed certificate or
4186 no certificate at all. Consider encoding the
4187 logged value because Issuer often has spaces.
4189 The default formats available (which do not need re-defining) are:
4191 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
4192 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
4193 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
4194 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
4195 logformat useragent %>a [%tl] "%{User-Agent}>h"
4197 NOTE: When the log_mime_hdrs directive is set to ON.
4198 The squid, common and combined formats have a safely encoded copy
4199 of the mime headers appended to each line within a pair of brackets.
4201 NOTE: The common and combined formats are not quite true to the Apache definition.
4202 The logs from Squid contain an extra status and hierarchy code appended.
4206 NAME: access_log cache_access_log
4208 LOC: Config.Log.accesslogs
4209 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
4211 Configures whether and how Squid logs HTTP and ICP transactions.
4212 If access logging is enabled, a single line is logged for every
4213 matching HTTP or ICP request. The recommended directive formats are:
4215 access_log <module>:<place> [option ...] [acl acl ...]
4216 access_log none [acl acl ...]
4218 The following directive format is accepted but may be deprecated:
4219 access_log <module>:<place> [<logformat name> [acl acl ...]]
4221 In most cases, the first ACL name must not contain the '=' character
4222 and should not be equal to an existing logformat name. You can always
4223 start with an 'all' ACL to work around those restrictions.
4225 Will log to the specified module:place using the specified format (which
4226 must be defined in a logformat directive) those entries which match
4227 ALL the acl's specified (which must be defined in acl clauses).
4228 If no acl is specified, all requests will be logged to this destination.
4230 ===== Available options for the recommended directive format =====
4232 logformat=name Names log line format (either built-in or
4233 defined by a logformat directive). Defaults
4236 buffer-size=64KB Defines approximate buffering limit for log
4237 records (see buffered_logs). Squid should not
4238 keep more than the specified size and, hence,
4239 should flush records before the buffer becomes
4240 full to avoid overflows under normal
4241 conditions (the exact flushing algorithm is
4242 module-dependent though). The on-error option
4243 controls overflow handling.
4245 on-error=die|drop Defines action on unrecoverable errors. The
4246 'drop' action ignores (i.e., does not log)
4247 affected log records. The default 'die' action
4248 kills the affected worker. The drop action
4249 support has not been tested for modules other
4252 rotate=N Specifies the number of log file rotations to
4253 make when you run 'squid -k rotate'. The default
4254 is to obey the logfile_rotate directive. Setting
4255 rotate=0 will disable the file name rotation,
4256 but the log files are still closed and re-opened.
4257 This will enable you to rename the logfiles
4258 yourself just before sending the rotate signal.
4259 Only supported by the stdio module.
4261 ===== Modules Currently available =====
4263 none Do not log any requests matching these ACL.
4264 Do not specify Place or logformat name.
4266 stdio Write each log line to disk immediately at the completion of
4268 Place: the filename and path to be written.
4270 daemon Very similar to stdio. But instead of writing to disk the log
4271 line is passed to a daemon helper for asychronous handling instead.
4272 Place: varies depending on the daemon.
4274 log_file_daemon Place: the file name and path to be written.
4276 syslog To log each request via syslog facility.
4277 Place: The syslog facility and priority level for these entries.
4278 Place Format: facility.priority
4280 where facility could be any of:
4281 authpriv, daemon, local0 ... local7 or user.
4283 And priority could be any of:
4284 err, warning, notice, info, debug.
4286 udp To send each log line as text data to a UDP receiver.
4287 Place: The destination host name or IP and port.
4288 Place Format: //host:port
4290 tcp To send each log line as text data to a TCP receiver.
4291 Lines may be accumulated before sending (see buffered_logs).
4292 Place: The destination host name or IP and port.
4293 Place Format: //host:port
4296 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4302 LOC: Config.Log.icaplogs
4305 ICAP log files record ICAP transaction summaries, one line per
4308 The icap_log option format is:
4309 icap_log <filepath> [<logformat name> [acl acl ...]]
4310 icap_log none [acl acl ...]]
4312 Please see access_log option documentation for details. The two
4313 kinds of logs share the overall configuration approach and many
4316 ICAP processing of a single HTTP message or transaction may
4317 require multiple ICAP transactions. In such cases, multiple
4318 ICAP transaction log lines will correspond to a single access
4321 ICAP log uses logformat codes that make sense for an ICAP
4322 transaction. Header-related codes are applied to the HTTP header
4323 embedded in an ICAP server response, with the following caveats:
4324 For REQMOD, there is no HTTP response header unless the ICAP
4325 server performed request satisfaction. For RESPMOD, the HTTP
4326 request header is the header sent to the ICAP server. For
4327 OPTIONS, there are no HTTP headers.
4329 The following format codes are also available for ICAP logs:
4331 icap::<A ICAP server IP address. Similar to <A.
4333 icap::<service_name ICAP service name from the icap_service
4334 option in Squid configuration file.
4336 icap::ru ICAP Request-URI. Similar to ru.
4338 icap::rm ICAP request method (REQMOD, RESPMOD, or
4339 OPTIONS). Similar to existing rm.
4341 icap::>st Bytes sent to the ICAP server (TCP payload
4342 only; i.e., what Squid writes to the socket).
4344 icap::<st Bytes received from the ICAP server (TCP
4345 payload only; i.e., what Squid reads from
4348 icap::<bs Number of message body bytes received from the
4349 ICAP server. ICAP message body, if any, usually
4350 includes encapsulated HTTP message headers and
4351 possibly encapsulated HTTP message body. The
4352 HTTP body part is dechunked before its size is
4355 icap::tr Transaction response time (in
4356 milliseconds). The timer starts when
4357 the ICAP transaction is created and
4358 stops when the transaction is completed.
4361 icap::tio Transaction I/O time (in milliseconds). The
4362 timer starts when the first ICAP request
4363 byte is scheduled for sending. The timers
4364 stops when the last byte of the ICAP response
4367 icap::to Transaction outcome: ICAP_ERR* for all
4368 transaction errors, ICAP_OPT for OPTION
4369 transactions, ICAP_ECHO for 204
4370 responses, ICAP_MOD for message
4371 modification, and ICAP_SAT for request
4372 satisfaction. Similar to Ss.
4374 icap::Hs ICAP response status code. Similar to Hs.
4376 icap::>h ICAP request header(s). Similar to >h.
4378 icap::<h ICAP response header(s). Similar to <h.
4380 The default ICAP log format, which can be used without an explicit
4381 definition, is called icap_squid:
4383 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4385 See also: logformat, log_icap, and %adapt::<last_h
4388 NAME: logfile_daemon
4390 DEFAULT: @DEFAULT_LOGFILED@
4391 LOC: Log::TheConfig.logfile_daemon
4393 Specify the path to the logfile-writing daemon. This daemon is
4394 used to write the access and store logs, if configured.
4396 Squid sends a number of commands to the log daemon:
4397 L<data>\n - logfile data
4402 r<n>\n - set rotate count to <n>
4403 b<n>\n - 1 = buffer output, 0 = don't buffer output
4405 No responses is expected.
4408 NAME: stats_collection
4410 LOC: Config.accessList.stats_collection
4412 DEFAULT_DOC: Allow logging for all transactions.
4413 COMMENT: allow|deny acl acl...
4415 This options allows you to control which requests gets accounted
4416 in performance counters.
4418 This clause only supports fast acl types.
4419 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4422 NAME: cache_store_log
4425 LOC: Config.Log.store
4427 Logs the activities of the storage manager. Shows which
4428 objects are ejected from the cache, and which objects are
4429 saved and for how long.
4430 There are not really utilities to analyze this data, so you can safely
4431 disable it (the default).
4433 Store log uses modular logging outputs. See access_log for the list
4434 of modules supported.
4437 cache_store_log stdio:@DEFAULT_STORE_LOG@
4438 cache_store_log daemon:@DEFAULT_STORE_LOG@
4441 NAME: cache_swap_state cache_swap_log
4443 LOC: Config.Log.swap
4445 DEFAULT_DOC: Store the journal inside its cache_dir
4447 Location for the cache "swap.state" file. This index file holds
4448 the metadata of objects saved on disk. It is used to rebuild
4449 the cache during startup. Normally this file resides in each
4450 'cache_dir' directory, but you may specify an alternate
4451 pathname here. Note you must give a full filename, not just
4452 a directory. Since this is the index for the whole object
4453 list you CANNOT periodically rotate it!
4455 If %s can be used in the file name it will be replaced with a
4456 a representation of the cache_dir name where each / is replaced
4457 with '.'. This is needed to allow adding/removing cache_dir
4458 lines when cache_swap_log is being used.
4460 If have more than one 'cache_dir', and %s is not used in the name
4461 these swap logs will have names such as:
4467 The numbered extension (which is added automatically)
4468 corresponds to the order of the 'cache_dir' lines in this
4469 configuration file. If you change the order of the 'cache_dir'
4470 lines in this file, these index files will NOT correspond to
4471 the correct 'cache_dir' entry (unless you manually rename
4472 them). We recommend you do NOT use this option. It is
4473 better to keep these index files in each 'cache_dir' directory.
4476 NAME: logfile_rotate
4479 LOC: Config.Log.rotateNumber
4481 Specifies the default number of logfile rotations to make when you
4482 type 'squid -k rotate'. The default is 10, which will rotate
4483 with extensions 0 through 9. Setting logfile_rotate to 0 will
4484 disable the file name rotation, but the logfiles are still closed
4485 and re-opened. This will enable you to rename the logfiles
4486 yourself just before sending the rotate signal.
4488 Note, from Squid-3.1 this option is only a default for cache.log,
4489 that log can be rotated separately by using debug_options.
4491 Note, from Squid-3.6 this option is only a default for access.log
4492 recorded by stdio: module. Those logs can be rotated separately by
4493 using the rotate=N option on their access_log directive.
4495 Note, the 'squid -k rotate' command normally sends a USR1
4496 signal to the running squid process. In certain situations
4497 (e.g. on Linux with Async I/O), USR1 is used for other
4498 purposes, so -k rotate uses another signal. It is best to get
4499 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4506 DEFAULT: @DEFAULT_MIME_TABLE@
4507 LOC: Config.mimeTablePathname
4509 Path to Squid's icon configuration file.
4511 You shouldn't need to change this, but the default file contains
4512 examples and formatting information if you do.
4518 LOC: Config.onoff.log_mime_hdrs
4521 The Cache can record both the request and the response MIME
4522 headers for each HTTP transaction. The headers are encoded
4523 safely and will appear as two bracketed fields at the end of
4524 the access log (for either the native or httpd-emulated log
4525 formats). To enable this logging set log_mime_hdrs to 'on'.
4530 DEFAULT: @DEFAULT_PID_FILE@
4531 LOC: Config.pidFilename
4533 A filename to write the process-id to. To disable, enter "none".
4536 NAME: client_netmask
4538 LOC: Config.Addrs.client_netmask
4540 DEFAULT_DOC: Log full client IP address
4542 A netmask for client addresses in logfiles and cachemgr output.
4543 Change this to protect the privacy of your cache clients.
4544 A netmask of 255.255.255.0 will log all IP's in that range with
4545 the last digit set to '0'.
4548 NAME: strip_query_terms
4550 LOC: Config.onoff.strip_query_terms
4553 By default, Squid strips query terms from requested URLs before
4554 logging. This protects your user's privacy and reduces log size.
4556 When investigating HIT/MISS or other caching behaviour you
4557 will need to disable this to see the full URL used by Squid.
4564 LOC: Config.onoff.buffered_logs
4566 Whether to write/send access_log records ASAP or accumulate them and
4567 then write/send them in larger chunks. Buffering may improve
4568 performance because it decreases the number of I/Os. However,
4569 buffering increases the delay before log records become available to
4570 the final recipient (e.g., a disk file or logging daemon) and,
4571 hence, increases the risk of log records loss.
4573 Note that even when buffered_logs are off, Squid may have to buffer
4574 records if it cannot write/send them immediately due to pending I/Os
4575 (e.g., the I/O writing the previous log record) or connectivity loss.
4577 Currently honored by 'daemon' and 'tcp' access_log modules only.
4580 NAME: netdb_filename
4582 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4583 LOC: Config.netdbFilename
4586 Where Squid stores it's netdb journal.
4587 When enabled this journal preserves netdb state between restarts.
4589 To disable, enter "none".
4593 OPTIONS FOR TROUBLESHOOTING
4594 -----------------------------------------------------------------------------
4599 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4600 LOC: Debug::cache_log
4602 Squid administrative logging file.
4604 This is where general information about Squid behavior goes. You can
4605 increase the amount of data logged to this file and how often it is
4606 rotated with "debug_options"
4612 DEFAULT_DOC: Log all critical and important messages.
4613 LOC: Debug::debugOptions
4615 Logging options are set as section,level where each source file
4616 is assigned a unique section. Lower levels result in less
4617 output, Full debugging (level 9) can result in a very large
4618 log file, so be careful.
4620 The magic word "ALL" sets debugging levels for all sections.
4621 The default is to run with "ALL,1" to record important warnings.
4623 The rotate=N option can be used to keep more or less of these logs
4624 than would otherwise be kept by logfile_rotate.
4625 For most uses a single log should be enough to monitor current
4626 events affecting Squid.
4631 LOC: Config.coredump_dir
4632 DEFAULT_IF_NONE: none
4633 DEFAULT_DOC: Use the directory from where Squid was started.
4635 By default Squid leaves core files in the directory from where
4636 it was started. If you set 'coredump_dir' to a directory
4637 that exists, Squid will chdir() to that directory at startup
4638 and coredump files will be left there.
4642 # Leave coredumps in the first cache dir
4643 coredump_dir @DEFAULT_SWAP_DIR@
4649 OPTIONS FOR FTP GATEWAYING
4650 -----------------------------------------------------------------------------
4656 LOC: Config.Ftp.anon_user
4658 If you want the anonymous login password to be more informative
4659 (and enable the use of picky FTP servers), set this to something
4660 reasonable for your domain, like wwwuser@somewhere.net
4662 The reason why this is domainless by default is the
4663 request can be made on the behalf of a user in any domain,
4664 depending on how the cache is used.
4665 Some FTP server also validate the email address is valid
4666 (for example perl.com).
4672 LOC: Config.Ftp.passive
4674 If your firewall does not allow Squid to use passive
4675 connections, turn off this option.
4677 Use of ftp_epsv_all option requires this to be ON.
4683 LOC: Config.Ftp.epsv_all
4685 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4687 NATs may be able to put the connection on a "fast path" through the
4688 translator, as the EPRT command will never be used and therefore,
4689 translation of the data portion of the segments will never be needed.
4691 When a client only expects to do two-way FTP transfers this may be
4693 If squid finds that it must do a three-way FTP transfer after issuing
4694 an EPSV ALL command, the FTP session will fail.
4696 If you have any doubts about this option do not use it.
4697 Squid will nicely attempt all other connection methods.
4699 Requires ftp_passive to be ON (default) for any effect.
4705 LOC: Config.accessList.ftp_epsv
4707 FTP Protocol extensions permit the use of a special "EPSV" command.
4709 NATs may be able to put the connection on a "fast path" through the
4710 translator using EPSV, as the EPRT command will never be used
4711 and therefore, translation of the data portion of the segments
4712 will never be needed.
4714 EPSV is often required to interoperate with FTP servers on IPv6
4715 networks. On the other hand, it may break some IPv4 servers.
4717 By default, EPSV may try EPSV with any FTP server. To fine tune
4718 that decision, you may restrict EPSV to certain clients or servers
4721 ftp_epsv allow|deny al1 acl2 ...
4723 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4725 Only fast ACLs are supported.
4726 Requires ftp_passive to be ON (default) for any effect.
4732 LOC: Config.Ftp.eprt
4734 FTP Protocol extensions permit the use of a special "EPRT" command.
4736 This extension provides a protocol neutral alternative to the
4737 IPv4-only PORT command. When supported it enables active FTP data
4738 channels over IPv6 and efficient NAT handling.
4740 Turning this OFF will prevent EPRT being attempted and will skip
4741 straight to using PORT for IPv4 servers.
4743 Some devices are known to not handle this extension correctly and
4744 may result in crashes. Devices which suport EPRT enough to fail
4745 cleanly will result in Squid attempting PORT anyway. This directive
4746 should only be disabled when EPRT results in device failures.
4748 WARNING: Doing so will convert Squid back to the old behavior with all
4749 the related problems with external NAT devices/layers and IPv4-only FTP.
4752 NAME: ftp_sanitycheck
4755 LOC: Config.Ftp.sanitycheck
4757 For security and data integrity reasons Squid by default performs
4758 sanity checks of the addresses of FTP data connections ensure the
4759 data connection is to the requested server. If you need to allow
4760 FTP connections to servers using another IP address for the data
4761 connection turn this off.
4764 NAME: ftp_telnet_protocol
4767 LOC: Config.Ftp.telnet
4769 The FTP protocol is officially defined to use the telnet protocol
4770 as transport channel for the control connection. However, many
4771 implementations are broken and does not respect this aspect of
4774 If you have trouble accessing files with ASCII code 255 in the
4775 path or similar problems involving this ASCII code you can
4776 try setting this directive to off. If that helps, report to the
4777 operator of the FTP server in question that their FTP server
4778 is broken and does not follow the FTP standard.
4782 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4783 -----------------------------------------------------------------------------
4788 DEFAULT: @DEFAULT_DISKD@
4789 LOC: Config.Program.diskd
4791 Specify the location of the diskd executable.
4792 Note this is only useful if you have compiled in
4793 diskd as one of the store io modules.
4796 NAME: unlinkd_program
4799 DEFAULT: @DEFAULT_UNLINKD@
4800 LOC: Config.Program.unlinkd
4802 Specify the location of the executable for file deletion process.
4805 NAME: pinger_program
4807 DEFAULT: @DEFAULT_PINGER@
4808 LOC: Config.pinger.program
4811 Specify the location of the executable for the pinger process.
4817 LOC: Config.pinger.enable
4820 Control whether the pinger is active at run-time.
4821 Enables turning ICMP pinger on and off with a simple
4822 squid -k reconfigure.
4827 OPTIONS FOR URL REWRITING
4828 -----------------------------------------------------------------------------
4831 NAME: url_rewrite_program redirect_program
4833 LOC: Config.Program.redirect
4836 Specify the location of the executable URL rewriter to use.
4837 Since they can perform almost any function there isn't one included.
4839 For each requested URL, the rewriter will receive on line with the format
4841 [channel-ID <SP>] URL [<SP> extras]<NL>
4843 See url_rewrite_extras on how to send "extras" with optional values to
4845 After processing the request the helper must reply using the following format:
4847 [channel-ID <SP>] result [<SP> kv-pairs]
4849 The result code can be:
4851 OK status=30N url="..."
4852 Redirect the URL to the one supplied in 'url='.
4853 'status=' is optional and contains the status code to send
4854 the client in Squids HTTP response. It must be one of the
4855 HTTP redirect status codes: 301, 302, 303, 307, 308.
4856 When no status is given Squid will use 302.
4858 OK rewrite-url="..."
4859 Rewrite the URL to the one supplied in 'rewrite-url='.
4860 The new URL is fetched directly by Squid and returned to
4861 the client as the response to its request.
4864 When neither of url= and rewrite-url= are sent Squid does
4868 Do not change the URL.
4871 An internal error occurred in the helper, preventing
4872 a result being identified. The 'message=' key name is
4873 reserved for delivering a log message.
4876 In addition to the above kv-pairs Squid also understands the following
4877 optional kv-pairs received from URL rewriters:
4879 Associates a TAG with the client TCP connection.
4880 The TAG is treated as a regular annotation but persists across
4881 future requests on the client connection rather than just the
4882 current request. A helper may update the TAG during subsequent
4883 requests be returning a new kv-pair.
4885 When using the concurrency= option the protocol is changed by
4886 introducing a query channel tag in front of the request/response.
4887 The query channel tag is a number between 0 and concurrency-1.
4888 This value must be echoed back unchanged to Squid as the first part
4889 of the response relating to its request.
4891 WARNING: URL re-writing ability should be avoided whenever possible.
4892 Use the URL redirect form of response instead.
4894 Re-write creates a difference in the state held by the client
4895 and server. Possibly causing confusion when the server response
4896 contains snippets of its view state. Embeded URLs, response
4897 and content Location headers, etc. are not re-written by this
4900 By default, a URL rewriter is not used.
4903 NAME: url_rewrite_children redirect_children
4904 TYPE: HelperChildConfig
4905 DEFAULT: 20 startup=0 idle=1 concurrency=0
4906 LOC: Config.redirectChildren
4908 The maximum number of redirector processes to spawn. If you limit
4909 it too few Squid will have to wait for them to process a backlog of
4910 URLs, slowing it down. If you allow too many they will use RAM
4911 and other system resources noticably.
4913 The startup= and idle= options allow some measure of skew in your
4918 Sets a minimum of how many processes are to be spawned when Squid
4919 starts or reconfigures. When set to zero the first request will
4920 cause spawning of the first child process to handle it.
4922 Starting too few will cause an initial slowdown in traffic as Squid
4923 attempts to simultaneously spawn enough processes to cope.
4927 Sets a minimum of how many processes Squid is to try and keep available
4928 at all times. When traffic begins to rise above what the existing
4929 processes can handle this many more will be spawned up to the maximum
4930 configured. A minimum setting of 1 is required.
4934 The number of requests each redirector helper can handle in
4935 parallel. Defaults to 0 which indicates the redirector
4936 is a old-style single threaded redirector.
4938 When this directive is set to a value >= 1 then the protocol
4939 used to communicate with the helper is modified to include
4940 an ID in front of the request/response. The ID from the request
4941 must be echoed back with the response to that request.
4945 Sets the maximum number of queued requests.
4946 If the queued requests exceed queue size and redirector_bypass
4947 configuration option is set, then redirector is bypassed. Otherwise, if
4948 overloading persists squid may abort its operation.
4949 The default value is set to 2*numberofchildren.
4952 NAME: url_rewrite_host_header redirect_rewrites_host_header
4955 LOC: Config.onoff.redir_rewrites_host
4957 To preserve same-origin security policies in browsers and
4958 prevent Host: header forgery by redirectors Squid rewrites
4959 any Host: header in redirected requests.
4961 If you are running an accelerator this may not be a wanted
4962 effect of a redirector. This directive enables you disable
4963 Host: alteration in reverse-proxy traffic.
4965 WARNING: Entries are cached on the result of the URL rewriting
4966 process, so be careful if you have domain-virtual hosts.
4968 WARNING: Squid and other software verifies the URL and Host
4969 are matching, so be careful not to relay through other proxies
4970 or inspecting firewalls with this disabled.
4973 NAME: url_rewrite_access redirector_access
4976 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4977 LOC: Config.accessList.redirector
4979 If defined, this access list specifies which requests are
4980 sent to the redirector processes.
4982 This clause supports both fast and slow acl types.
4983 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4986 NAME: url_rewrite_bypass redirector_bypass
4988 LOC: Config.onoff.redirector_bypass
4991 When this is 'on', a request will not go through the
4992 redirector if all the helpers are busy. If this is 'off'
4993 and the redirector queue grows too large, Squid will exit
4994 with a FATAL error and ask you to increase the number of
4995 redirectors. You should only enable this if the redirectors
4996 are not critical to your caching system. If you use
4997 redirectors for access control, and you enable this option,
4998 users may have access to pages they should not
4999 be allowed to request.
5000 This options sets default queue-size option of the url_rewrite_children
5004 NAME: url_rewrite_extras
5005 TYPE: TokenOrQuotedString
5006 LOC: Config.redirector_extras
5007 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5009 Specifies a string to be append to request line format for the
5010 rewriter helper. "Quoted" format values may contain spaces and
5011 logformat %macros. In theory, any logformat %macro can be used.
5012 In practice, a %macro expands as a dash (-) if the helper request is
5013 sent before the required macro information is available to Squid.
5016 NAME: url_rewrite_timeout
5017 TYPE: UrlHelperTimeout
5018 LOC: Config.onUrlRewriteTimeout
5020 DEFAULT_DOC: Squid waits for the helper response forever
5022 Squid times active requests to redirector. The timeout value and Squid
5023 reaction to a timed out request are configurable using the following
5026 url_rewrite_timeout timeout time-units on_timeout=<action> [response=<quoted-response>]
5028 supported timeout actions:
5029 fail Squid return a ERR_GATEWAY_FAILURE error page
5031 bypass Do not re-write the URL
5033 retry Send the lookup to the helper again
5035 use_configured_response
5036 Use the <quoted-response> as helper response
5040 OPTIONS FOR STORE ID
5041 -----------------------------------------------------------------------------
5044 NAME: store_id_program storeurl_rewrite_program
5046 LOC: Config.Program.store_id
5049 Specify the location of the executable StoreID helper to use.
5050 Since they can perform almost any function there isn't one included.
5052 For each requested URL, the helper will receive one line with the format
5054 [channel-ID <SP>] URL [<SP> extras]<NL>
5057 After processing the request the helper must reply using the following format:
5059 [channel-ID <SP>] result [<SP> kv-pairs]
5061 The result code can be:
5064 Use the StoreID supplied in 'store-id='.
5067 The default is to use HTTP request URL as the store ID.
5070 An internal error occured in the helper, preventing
5071 a result being identified.
5073 In addition to the above kv-pairs Squid also understands the following
5074 optional kv-pairs received from URL rewriters:
5076 Associates a TAG with the client TCP connection.
5077 Please see url_rewrite_program related documentation for this
5080 Helper programs should be prepared to receive and possibly ignore
5081 additional whitespace-separated tokens on each input line.
5083 When using the concurrency= option the protocol is changed by
5084 introducing a query channel tag in front of the request/response.
5085 The query channel tag is a number between 0 and concurrency-1.
5086 This value must be echoed back unchanged to Squid as the first part
5087 of the response relating to its request.
5089 NOTE: when using StoreID refresh_pattern will apply to the StoreID
5090 returned from the helper and not the URL.
5092 WARNING: Wrong StoreID value returned by a careless helper may result
5093 in the wrong cached response returned to the user.
5095 By default, a StoreID helper is not used.
5098 NAME: store_id_extras
5099 TYPE: TokenOrQuotedString
5100 LOC: Config.storeId_extras
5101 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
5103 Specifies a string to be append to request line format for the
5104 StoreId helper. "Quoted" format values may contain spaces and
5105 logformat %macros. In theory, any logformat %macro can be used.
5106 In practice, a %macro expands as a dash (-) if the helper request is
5107 sent before the required macro information is available to Squid.
5110 NAME: store_id_children storeurl_rewrite_children
5111 TYPE: HelperChildConfig
5112 DEFAULT: 20 startup=0 idle=1 concurrency=0
5113 LOC: Config.storeIdChildren
5115 The maximum number of StoreID helper processes to spawn. If you limit
5116 it too few Squid will have to wait for them to process a backlog of
5117 requests, slowing it down. If you allow too many they will use RAM
5118 and other system resources noticably.
5120 The startup= and idle= options allow some measure of skew in your
5125 Sets a minimum of how many processes are to be spawned when Squid
5126 starts or reconfigures. When set to zero the first request will
5127 cause spawning of the first child process to handle it.
5129 Starting too few will cause an initial slowdown in traffic as Squid
5130 attempts to simultaneously spawn enough processes to cope.
5134 Sets a minimum of how many processes Squid is to try and keep available
5135 at all times. When traffic begins to rise above what the existing
5136 processes can handle this many more will be spawned up to the maximum
5137 configured. A minimum setting of 1 is required.
5141 The number of requests each storeID helper can handle in
5142 parallel. Defaults to 0 which indicates the helper
5143 is a old-style single threaded program.
5145 When this directive is set to a value >= 1 then the protocol
5146 used to communicate with the helper is modified to include
5147 an ID in front of the request/response. The ID from the request
5148 must be echoed back with the response to that request.
5152 Sets the maximum number of queued requests.
5153 If the queued requests exceed queue size and store_id_bypass
5154 configuration option is set, then storeID helper is bypassed. Otherwise,
5155 if overloading persists squid may abort its operation.
5156 The default value is set to 2*numberofchildren.
5159 NAME: store_id_access storeurl_rewrite_access
5162 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
5163 LOC: Config.accessList.store_id
5165 If defined, this access list specifies which requests are
5166 sent to the StoreID processes. By default all requests
5169 This clause supports both fast and slow acl types.
5170 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5173 NAME: store_id_bypass storeurl_rewrite_bypass
5175 LOC: Config.onoff.store_id_bypass
5178 When this is 'on', a request will not go through the
5179 helper if all helpers are busy. If this is 'off'
5180 and the helper queue grows too large, Squid will exit
5181 with a FATAL error and ask you to increase the number of
5182 helpers. You should only enable this if the helperss
5183 are not critical to your caching system. If you use
5184 helpers for critical caching components, and you enable this
5185 option, users may not get objects from cache.
5186 This options sets default queue-size option of the store_id_children
5191 OPTIONS FOR TUNING THE CACHE
5192 -----------------------------------------------------------------------------
5195 NAME: cache no_cache
5198 DEFAULT_DOC: By default, this directive is unused and has no effect.
5199 LOC: Config.accessList.noCache
5201 Requests denied by this directive will not be served from the cache
5202 and their responses will not be stored in the cache. This directive
5203 has no effect on other transactions and on already cached responses.
5205 This clause supports both fast and slow acl types.
5206 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5208 This and the two other similar caching directives listed below are
5209 checked at different transaction processing stages, have different
5210 access to response information, affect different cache operations,
5211 and differ in slow ACLs support:
5213 * cache: Checked before Squid makes a hit/miss determination.
5214 No access to reply information!
5215 Denies both serving a hit and storing a miss.
5216 Supports both fast and slow ACLs.
5217 * send_hit: Checked after a hit was detected.
5218 Has access to reply (hit) information.
5219 Denies serving a hit only.
5220 Supports fast ACLs only.
5221 * store_miss: Checked before storing a cachable miss.
5222 Has access to reply (miss) information.
5223 Denies storing a miss only.
5224 Supports fast ACLs only.
5226 If you are not sure which of the three directives to use, apply the
5227 following decision logic:
5229 * If your ACL(s) are of slow type _and_ need response info, redesign.
5230 Squid does not support that particular combination at this time.
5232 * If your directive ACL(s) are of slow type, use "cache"; and/or
5233 * if your directive ACL(s) need no response info, use "cache".
5235 * If you do not want the response cached, use store_miss; and/or
5236 * if you do not want a hit on a cached response, use send_hit.
5242 DEFAULT_DOC: By default, this directive is unused and has no effect.
5243 LOC: Config.accessList.sendHit
5245 Responses denied by this directive will not be served from the cache
5246 (but may still be cached, see store_miss). This directive has no
5247 effect on the responses it allows and on the cached objects.
5249 Please see the "cache" directive for a summary of differences among
5250 store_miss, send_hit, and cache directives.
5252 Unlike the "cache" directive, send_hit only supports fast acl
5253 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5257 # apply custom Store ID mapping to some URLs
5258 acl MapMe dstdomain .c.example.com
5259 store_id_program ...
5260 store_id_access allow MapMe
5262 # but prevent caching of special responses
5263 # such as 302 redirects that cause StoreID loops
5264 acl Ordinary http_status 200-299
5265 store_miss deny MapMe !Ordinary
5267 # and do not serve any previously stored special responses
5268 # from the cache (in case they were already cached before
5269 # the above store_miss rule was in effect).
5270 send_hit deny MapMe !Ordinary
5276 DEFAULT_DOC: By default, this directive is unused and has no effect.
5277 LOC: Config.accessList.storeMiss
5279 Responses denied by this directive will not be cached (but may still
5280 be served from the cache, see send_hit). This directive has no
5281 effect on the responses it allows and on the already cached responses.
5283 Please see the "cache" directive for a summary of differences among
5284 store_miss, send_hit, and cache directives. See the
5285 send_hit directive for a usage example.
5287 Unlike the "cache" directive, store_miss only supports fast acl
5288 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5294 LOC: Config.maxStale
5297 This option puts an upper limit on how stale content Squid
5298 will serve from the cache if cache validation fails.
5299 Can be overriden by the refresh_pattern max-stale option.
5302 NAME: refresh_pattern
5303 TYPE: refreshpattern
5307 usage: refresh_pattern [-i] regex min percent max [options]
5309 By default, regular expressions are CASE-SENSITIVE. To make
5310 them case-insensitive, use the -i option.
5312 'Min' is the time (in minutes) an object without an explicit
5313 expiry time should be considered fresh. The recommended
5314 value is 0, any higher values may cause dynamic applications
5315 to be erroneously cached unless the application designer
5316 has taken the appropriate actions.
5318 'Percent' is a percentage of the objects age (time since last
5319 modification age) an object without explicit expiry time
5320 will be considered fresh.
5322 'Max' is an upper limit on how long objects without an explicit
5323 expiry time will be considered fresh.
5325 options: override-expire
5330 ignore-must-revalidate
5337 override-expire enforces min age even if the server
5338 sent an explicit expiry time (e.g., with the
5339 Expires: header or Cache-Control: max-age). Doing this
5340 VIOLATES the HTTP standard. Enabling this feature
5341 could make you liable for problems which it causes.
5343 Note: override-expire does not enforce staleness - it only extends
5344 freshness / min. If the server returns a Expires time which
5345 is longer than your max time, Squid will still consider
5346 the object fresh for that period of time.
5348 override-lastmod enforces min age even on objects
5349 that were modified recently.
5351 reload-into-ims changes a client no-cache or ``reload''
5352 request for a cached entry into a conditional request using
5353 If-Modified-Since and/or If-None-Match headers, provided the
5354 cached entry has a Last-Modified and/or a strong ETag header.
5355 Doing this VIOLATES the HTTP standard. Enabling this feature
5356 could make you liable for problems which it causes.
5358 ignore-reload ignores a client no-cache or ``reload''
5359 header. Doing this VIOLATES the HTTP standard. Enabling
5360 this feature could make you liable for problems which
5363 ignore-no-store ignores any ``Cache-control: no-store''
5364 headers received from a server. Doing this VIOLATES
5365 the HTTP standard. Enabling this feature could make you
5366 liable for problems which it causes.
5368 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
5369 headers received from a server. Doing this VIOLATES
5370 the HTTP standard. Enabling this feature could make you
5371 liable for problems which it causes.
5373 ignore-private ignores any ``Cache-control: private''
5374 headers received from a server. Doing this VIOLATES
5375 the HTTP standard. Enabling this feature could make you
5376 liable for problems which it causes.
5378 ignore-auth caches responses to requests with authorization,
5379 as if the originserver had sent ``Cache-control: public''
5380 in the response header. Doing this VIOLATES the HTTP standard.
5381 Enabling this feature could make you liable for problems which
5384 refresh-ims causes squid to contact the origin server
5385 when a client issues an If-Modified-Since request. This
5386 ensures that the client will receive an updated version
5387 if one is available.
5389 store-stale stores responses even if they don't have explicit
5390 freshness or a validator (i.e., Last-Modified or an ETag)
5391 present, or if they're already stale. By default, Squid will
5392 not cache such responses because they usually can't be
5393 reused. Note that such responses will be stale by default.
5395 max-stale=NN provide a maximum staleness factor. Squid won't
5396 serve objects more stale than this even if it failed to
5397 validate the object. Default: use the max_stale global limit.
5399 Basically a cached object is:
5401 FRESH if expires < now, else STALE
5403 FRESH if lm-factor < percent, else STALE
5407 The refresh_pattern lines are checked in the order listed here.
5408 The first entry which matches is used. If none of the entries
5409 match the default will be used.
5411 Note, you must uncomment all the default lines if you want
5412 to change one. The default setting is only active if none is
5418 # Add any of your own refresh_pattern entries above these.
5420 refresh_pattern ^ftp: 1440 20% 10080
5421 refresh_pattern ^gopher: 1440 0% 1440
5422 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5423 refresh_pattern . 0 20% 4320
5427 NAME: quick_abort_min
5431 LOC: Config.quickAbort.min
5434 NAME: quick_abort_max
5438 LOC: Config.quickAbort.max
5441 NAME: quick_abort_pct
5445 LOC: Config.quickAbort.pct
5447 The cache by default continues downloading aborted requests
5448 which are almost completed (less than 16 KB remaining). This
5449 may be undesirable on slow (e.g. SLIP) links and/or very busy
5450 caches. Impatient users may tie up file descriptors and
5451 bandwidth by repeatedly requesting and immediately aborting
5454 When the user aborts a request, Squid will check the
5455 quick_abort values to the amount of data transferred until
5458 If the transfer has less than 'quick_abort_min' KB remaining,
5459 it will finish the retrieval.
5461 If the transfer has more than 'quick_abort_max' KB remaining,
5462 it will abort the retrieval.
5464 If more than 'quick_abort_pct' of the transfer has completed,
5465 it will finish the retrieval.
5467 If you do not want any retrieval to continue after the client
5468 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5471 If you want retrievals to always continue if they are being
5472 cached set 'quick_abort_min' to '-1 KB'.
5475 NAME: read_ahead_gap
5476 COMMENT: buffer-size
5478 LOC: Config.readAheadGap
5481 The amount of data the cache will buffer ahead of what has been
5482 sent to the client when retrieving an object from another server.
5486 IFDEF: USE_HTTP_VIOLATIONS
5489 LOC: Config.negativeTtl
5492 Set the Default Time-to-Live (TTL) for failed requests.
5493 Certain types of failures (such as "connection refused" and
5494 "404 Not Found") are able to be negatively-cached for a short time.
5495 Modern web servers should provide Expires: header, however if they
5496 do not this can provide a minimum TTL.
5497 The default is not to cache errors with unknown expiry details.
5499 Note that this is different from negative caching of DNS lookups.
5501 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5502 this feature could make you liable for problems which it
5506 NAME: positive_dns_ttl
5509 LOC: Config.positiveDnsTtl
5512 Upper limit on how long Squid will cache positive DNS responses.
5513 Default is 6 hours (360 minutes). This directive must be set
5514 larger than negative_dns_ttl.
5517 NAME: negative_dns_ttl
5520 LOC: Config.negativeDnsTtl
5523 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5524 This also sets the lower cache limit on positive lookups.
5525 Minimum value is 1 second, and it is not recommendable to go
5526 much below 10 seconds.
5529 NAME: range_offset_limit
5530 COMMENT: size [acl acl...]
5532 LOC: Config.rangeOffsetLimit
5535 usage: (size) [units] [[!]aclname]
5537 Sets an upper limit on how far (number of bytes) into the file
5538 a Range request may be to cause Squid to prefetch the whole file.
5539 If beyond this limit, Squid forwards the Range request as it is and
5540 the result is NOT cached.
5542 This is to stop a far ahead range request (lets say start at 17MB)
5543 from making Squid fetch the whole object up to that point before
5544 sending anything to the client.
5546 Multiple range_offset_limit lines may be specified, and they will
5547 be searched from top to bottom on each request until a match is found.
5548 The first match found will be used. If no line matches a request, the
5549 default limit of 0 bytes will be used.
5551 'size' is the limit specified as a number of units.
5553 'units' specifies whether to use bytes, KB, MB, etc.
5554 If no units are specified bytes are assumed.
5556 A size of 0 causes Squid to never fetch more than the
5557 client requested. (default)
5559 A size of 'none' causes Squid to always fetch the object from the
5560 beginning so it may cache the result. (2.0 style)
5562 'aclname' is the name of a defined ACL.
5564 NP: Using 'none' as the byte value here will override any quick_abort settings
5565 that may otherwise apply to the range request. The range request will
5566 be fully fetched from start to finish regardless of the client
5567 actions. This affects bandwidth usage.
5570 NAME: minimum_expiry_time
5573 LOC: Config.minimum_expiry_time
5576 The minimum caching time according to (Expires - Date)
5577 headers Squid honors if the object can't be revalidated.
5578 The default is 60 seconds.
5580 In reverse proxy environments it might be desirable to honor
5581 shorter object lifetimes. It is most likely better to make
5582 your server return a meaningful Last-Modified header however.
5584 In ESI environments where page fragments often have short
5585 lifetimes, this will often be best set to 0.
5588 NAME: store_avg_object_size
5592 LOC: Config.Store.avgObjectSize
5594 Average object size, used to estimate number of objects your
5595 cache can hold. The default is 13 KB.
5597 This is used to pre-seed the cache index memory allocation to
5598 reduce expensive reallocate operations while handling clients
5599 traffic. Too-large values may result in memory allocation during
5600 peak traffic, too-small values will result in wasted memory.
5602 Check the cache manager 'info' report metrics for the real
5603 object sizes seen by your Squid before tuning this.
5606 NAME: store_objects_per_bucket
5609 LOC: Config.Store.objectsPerBucket
5611 Target number of objects per bucket in the store hash table.
5612 Lowering this value increases the total number of buckets and
5613 also the storage maintenance rate. The default is 20.
5618 -----------------------------------------------------------------------------
5621 NAME: request_header_max_size
5625 LOC: Config.maxRequestHeaderSize
5627 This specifies the maximum size for HTTP headers in a request.
5628 Request headers are usually relatively small (about 512 bytes).
5629 Placing a limit on the request header size will catch certain
5630 bugs (for example with persistent connections) and possibly
5631 buffer-overflow or denial-of-service attacks.
5634 NAME: reply_header_max_size
5638 LOC: Config.maxReplyHeaderSize
5640 This specifies the maximum size for HTTP headers in a reply.
5641 Reply headers are usually relatively small (about 512 bytes).
5642 Placing a limit on the reply header size will catch certain
5643 bugs (for example with persistent connections) and possibly
5644 buffer-overflow or denial-of-service attacks.
5647 NAME: request_body_max_size
5651 DEFAULT_DOC: No limit.
5652 LOC: Config.maxRequestBodySize
5654 This specifies the maximum size for an HTTP request body.
5655 In other words, the maximum size of a PUT/POST request.
5656 A user who attempts to send a request with a body larger
5657 than this limit receives an "Invalid Request" error message.
5658 If you set this parameter to a zero (the default), there will
5659 be no limit imposed.
5661 See also client_request_buffer_max_size for an alternative
5662 limitation on client uploads which can be configured.
5665 NAME: client_request_buffer_max_size
5669 LOC: Config.maxRequestBufferSize
5671 This specifies the maximum buffer size of a client request.
5672 It prevents squid eating too much memory when somebody uploads
5676 NAME: chunked_request_body_max_size
5680 LOC: Config.maxChunkedRequestBodySize
5682 A broken or confused HTTP/1.1 client may send a chunked HTTP
5683 request to Squid. Squid does not have full support for that
5684 feature yet. To cope with such requests, Squid buffers the
5685 entire request and then dechunks request body to create a
5686 plain HTTP/1.0 request with a known content length. The plain
5687 request is then used by the rest of Squid code as usual.
5689 The option value specifies the maximum size of the buffer used
5690 to hold the request before the conversion. If the chunked
5691 request size exceeds the specified limit, the conversion
5692 fails, and the client receives an "unsupported request" error,
5693 as if dechunking was disabled.
5695 Dechunking is enabled by default. To disable conversion of
5696 chunked requests, set the maximum to zero.
5698 Request dechunking feature and this option in particular are a
5699 temporary hack. When chunking requests and responses are fully
5700 supported, there will be no need to buffer a chunked request.
5704 IFDEF: USE_HTTP_VIOLATIONS
5707 DEFAULT_DOC: Obey RFC 2616.
5708 LOC: Config.accessList.brokenPosts
5710 A list of ACL elements which, if matched, causes Squid to send
5711 an extra CRLF pair after the body of a PUT/POST request.
5713 Some HTTP servers has broken implementations of PUT/POST,
5714 and rely on an extra CRLF pair sent by some WWW clients.
5716 Quote from RFC2616 section 4.1 on this matter:
5718 Note: certain buggy HTTP/1.0 client implementations generate an
5719 extra CRLF's after a POST request. To restate what is explicitly
5720 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5721 a request with an extra CRLF.
5723 This clause only supports fast acl types.
5724 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5727 acl buggy_server url_regex ^http://....
5728 broken_posts allow buggy_server
5731 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5734 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5736 LOC: Adaptation::Config::use_indirect_client
5738 Controls whether the indirect client IP address (instead of the direct
5739 client IP address) is passed to adaptation services.
5741 See also: follow_x_forwarded_for adaptation_send_client_ip
5745 IFDEF: USE_HTTP_VIOLATIONS
5749 LOC: Config.onoff.via
5751 If set (default), Squid will include a Via header in requests and
5752 replies as required by RFC2616.
5758 LOC: Config.onoff.ie_refresh
5761 Microsoft Internet Explorer up until version 5.5 Service
5762 Pack 1 has an issue with transparent proxies, wherein it
5763 is impossible to force a refresh. Turning this on provides
5764 a partial fix to the problem, by causing all IMS-REFRESH
5765 requests from older IE versions to check the origin server
5766 for fresh content. This reduces hit ratio by some amount
5767 (~10% in my experience), but allows users to actually get
5768 fresh content when they want it. Note because Squid
5769 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5770 of 5.5 is unchanged from old versions of Squid (i.e. a
5771 forced refresh is impossible). Newer versions of IE will,
5772 hopefully, continue to have the new behavior and will be
5773 handled based on that assumption. This option defaults to
5774 the old Squid behavior, which is better for hit ratios but
5775 worse for clients using IE, if they need to be able to
5776 force fresh content.
5779 NAME: vary_ignore_expire
5782 LOC: Config.onoff.vary_ignore_expire
5785 Many HTTP servers supporting Vary gives such objects
5786 immediate expiry time with no cache-control header
5787 when requested by a HTTP/1.0 client. This option
5788 enables Squid to ignore such expiry times until
5789 HTTP/1.1 is fully implemented.
5791 WARNING: If turned on this may eventually cause some
5792 varying objects not intended for caching to get cached.
5795 NAME: request_entities
5797 LOC: Config.onoff.request_entities
5800 Squid defaults to deny GET and HEAD requests with request entities,
5801 as the meaning of such requests are undefined in the HTTP standard
5802 even if not explicitly forbidden.
5804 Set this directive to on if you have clients which insists
5805 on sending request entities in GET or HEAD requests. But be warned
5806 that there is server software (both proxies and web servers) which
5807 can fail to properly process this kind of request which may make you
5808 vulnerable to cache pollution attacks if enabled.
5811 NAME: request_header_access
5812 IFDEF: USE_HTTP_VIOLATIONS
5813 TYPE: http_header_access
5814 LOC: Config.request_header_access
5816 DEFAULT_DOC: No limits.
5818 Usage: request_header_access header_name allow|deny [!]aclname ...
5820 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5821 this feature could make you liable for problems which it
5824 This option replaces the old 'anonymize_headers' and the
5825 older 'http_anonymizer' option with something that is much
5826 more configurable. A list of ACLs for each header name allows
5827 removal of specific header fields under specific conditions.
5829 This option only applies to outgoing HTTP request headers (i.e.,
5830 headers sent by Squid to the next HTTP hop such as a cache peer
5831 or an origin server). The option has no effect during cache hit
5832 detection. The equivalent adaptation vectoring point in ICAP
5833 terminology is post-cache REQMOD.
5835 The option is applied to individual outgoing request header
5836 fields. For each request header field F, Squid uses the first
5837 qualifying sets of request_header_access rules:
5839 1. Rules with header_name equal to F's name.
5840 2. Rules with header_name 'Other', provided F's name is not
5841 on the hard-coded list of commonly used HTTP header names.
5842 3. Rules with header_name 'All'.
5844 Within that qualifying rule set, rule ACLs are checked as usual.
5845 If ACLs of an "allow" rule match, the header field is allowed to
5846 go through as is. If ACLs of a "deny" rule match, the header is
5847 removed and request_header_replace is then checked to identify
5848 if the removed header has a replacement. If no rules within the
5849 set have matching ACLs, the header field is left as is.
5851 For example, to achieve the same behavior as the old
5852 'http_anonymizer standard' option, you should use:
5854 request_header_access From deny all
5855 request_header_access Referer deny all
5856 request_header_access User-Agent deny all
5858 Or, to reproduce the old 'http_anonymizer paranoid' feature
5861 request_header_access Authorization allow all
5862 request_header_access Proxy-Authorization allow all
5863 request_header_access Cache-Control allow all
5864 request_header_access Content-Length allow all
5865 request_header_access Content-Type allow all
5866 request_header_access Date allow all
5867 request_header_access Host allow all
5868 request_header_access If-Modified-Since allow all
5869 request_header_access Pragma allow all
5870 request_header_access Accept allow all
5871 request_header_access Accept-Charset allow all
5872 request_header_access Accept-Encoding allow all
5873 request_header_access Accept-Language allow all
5874 request_header_access Connection allow all
5875 request_header_access All deny all
5877 HTTP reply headers are controlled with the reply_header_access directive.
5879 By default, all headers are allowed (no anonymizing is performed).
5882 NAME: reply_header_access
5883 IFDEF: USE_HTTP_VIOLATIONS
5884 TYPE: http_header_access
5885 LOC: Config.reply_header_access
5887 DEFAULT_DOC: No limits.
5889 Usage: reply_header_access header_name allow|deny [!]aclname ...
5891 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5892 this feature could make you liable for problems which it
5895 This option only applies to reply headers, i.e., from the
5896 server to the client.
5898 This is the same as request_header_access, but in the other
5899 direction. Please see request_header_access for detailed
5902 For example, to achieve the same behavior as the old
5903 'http_anonymizer standard' option, you should use:
5905 reply_header_access Server deny all
5906 reply_header_access WWW-Authenticate deny all
5907 reply_header_access Link deny all
5909 Or, to reproduce the old 'http_anonymizer paranoid' feature
5912 reply_header_access Allow allow all
5913 reply_header_access WWW-Authenticate allow all
5914 reply_header_access Proxy-Authenticate allow all
5915 reply_header_access Cache-Control allow all
5916 reply_header_access Content-Encoding allow all
5917 reply_header_access Content-Length allow all
5918 reply_header_access Content-Type allow all
5919 reply_header_access Date allow all
5920 reply_header_access Expires allow all
5921 reply_header_access Last-Modified allow all
5922 reply_header_access Location allow all
5923 reply_header_access Pragma allow all
5924 reply_header_access Content-Language allow all
5925 reply_header_access Retry-After allow all
5926 reply_header_access Title allow all
5927 reply_header_access Content-Disposition allow all
5928 reply_header_access Connection allow all
5929 reply_header_access All deny all
5931 HTTP request headers are controlled with the request_header_access directive.
5933 By default, all headers are allowed (no anonymizing is
5937 NAME: request_header_replace header_replace
5938 IFDEF: USE_HTTP_VIOLATIONS
5939 TYPE: http_header_replace
5940 LOC: Config.request_header_access
5943 Usage: request_header_replace header_name message
5944 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5946 This option allows you to change the contents of headers
5947 denied with request_header_access above, by replacing them
5948 with some fixed string.
5950 This only applies to request headers, not reply headers.
5952 By default, headers are removed if denied.
5955 NAME: reply_header_replace
5956 IFDEF: USE_HTTP_VIOLATIONS
5957 TYPE: http_header_replace
5958 LOC: Config.reply_header_access
5961 Usage: reply_header_replace header_name message
5962 Example: reply_header_replace Server Foo/1.0
5964 This option allows you to change the contents of headers
5965 denied with reply_header_access above, by replacing them
5966 with some fixed string.
5968 This only applies to reply headers, not request headers.
5970 By default, headers are removed if denied.
5973 NAME: request_header_add
5974 TYPE: HeaderWithAclList
5975 LOC: Config.request_header_add
5978 Usage: request_header_add field-name field-value acl1 [acl2] ...
5979 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5981 This option adds header fields to outgoing HTTP requests (i.e.,
5982 request headers sent by Squid to the next HTTP hop such as a
5983 cache peer or an origin server). The option has no effect during
5984 cache hit detection. The equivalent adaptation vectoring point
5985 in ICAP terminology is post-cache REQMOD.
5987 Field-name is a token specifying an HTTP header name. If a
5988 standard HTTP header name is used, Squid does not check whether
5989 the new header conflicts with any existing headers or violates
5990 HTTP rules. If the request to be modified already contains a
5991 field with the same name, the old field is preserved but the
5992 header field values are not merged.
5994 Field-value is either a token or a quoted string. If quoted
5995 string format is used, then the surrounding quotes are removed
5996 while escape sequences and %macros are processed.
5998 In theory, all of the logformat codes can be used as %macros.
5999 However, unlike logging (which happens at the very end of
6000 transaction lifetime), the transaction may not yet have enough
6001 information to expand a macro when the new header value is needed.
6002 And some information may already be available to Squid but not yet
6003 committed where the macro expansion code can access it (report
6004 such instances!). The macro will be expanded into a single dash
6005 ('-') in such cases. Not all macros have been tested.
6007 One or more Squid ACLs may be specified to restrict header
6008 injection to matching requests. As always in squid.conf, all
6009 ACLs in an option ACL list must be satisfied for the insertion
6010 to happen. The request_header_add option supports fast ACLs
6019 This option used to log custom information about the master
6020 transaction. For example, an admin may configure Squid to log
6021 which "user group" the transaction belongs to, where "user group"
6022 will be determined based on a set of ACLs and not [just]
6023 authentication information.
6024 Values of key/value pairs can be logged using %{key}note macros:
6026 note key value acl ...
6027 logformat myFormat ... %{key}note ...
6030 NAME: relaxed_header_parser
6031 COMMENT: on|off|warn
6033 LOC: Config.onoff.relaxed_header_parser
6036 In the default "on" setting Squid accepts certain forms
6037 of non-compliant HTTP messages where it is unambiguous
6038 what the sending application intended even if the message
6039 is not correctly formatted. The messages is then normalized
6040 to the correct form when forwarded by Squid.
6042 If set to "warn" then a warning will be emitted in cache.log
6043 each time such HTTP error is encountered.
6045 If set to "off" then such HTTP errors will cause the request
6046 or response to be rejected.
6049 NAME: collapsed_forwarding
6052 LOC: Config.onoff.collapsed_forwarding
6055 This option controls whether Squid is allowed to merge multiple
6056 potentially cachable requests for the same URI before Squid knows
6057 whether the response is going to be cachable.
6059 This feature is disabled by default: Enabling collapsed forwarding
6060 needlessly delays forwarding requests that look cachable (when they are
6061 collapsed) but then need to be forwarded individually anyway because
6062 they end up being for uncachable content. However, in some cases, such
6063 as accelleration of highly cachable content with periodic or groupped
6064 expiration times, the gains from collapsing [large volumes of
6065 simultenous refresh requests] outweigh losses from such delays.
6070 -----------------------------------------------------------------------------
6073 NAME: forward_timeout
6076 LOC: Config.Timeout.forward
6079 This parameter specifies how long Squid should at most attempt in
6080 finding a forwarding path for the request before giving up.
6083 NAME: connect_timeout
6086 LOC: Config.Timeout.connect
6089 This parameter specifies how long to wait for the TCP connect to
6090 the requested server or peer to complete before Squid should
6091 attempt to find another path where to forward the request.
6094 NAME: peer_connect_timeout
6097 LOC: Config.Timeout.peer_connect
6100 This parameter specifies how long to wait for a pending TCP
6101 connection to a peer cache. The default is 30 seconds. You
6102 may also set different timeout values for individual neighbors
6103 with the 'connect-timeout' option on a 'cache_peer' line.
6109 LOC: Config.Timeout.read
6112 Applied on peer server connections.
6114 After each successful read(), the timeout will be extended by this
6115 amount. If no data is read again after this amount of time,
6116 the request is aborted and logged with ERR_READ_TIMEOUT.
6118 The default is 15 minutes.
6124 LOC: Config.Timeout.write
6127 This timeout is tracked for all connections that have data
6128 available for writing and are waiting for the socket to become
6129 ready. After each successful write, the timeout is extended by
6130 the configured amount. If Squid has data to write but the
6131 connection is not ready for the configured duration, the
6132 transaction associated with the connection is terminated. The
6133 default is 15 minutes.
6136 NAME: request_timeout
6138 LOC: Config.Timeout.request
6141 How long to wait for complete HTTP request headers after initial
6142 connection establishment.
6145 NAME: request_start_timeout
6147 LOC: Config.Timeout.request_start_timeout
6150 How long to wait for the first request byte after initial
6151 connection establishment.
6154 NAME: client_idle_pconn_timeout persistent_request_timeout
6156 LOC: Config.Timeout.clientIdlePconn
6159 How long to wait for the next HTTP request on a persistent
6160 client connection after the previous request completes.
6163 NAME: ftp_client_idle_timeout
6165 LOC: Config.Timeout.ftpClientIdle
6168 How long to wait for an FTP request on a connection to Squid ftp_port.
6169 Many FTP clients do not deal with idle connection closures well,
6170 necessitating a longer default timeout than client_idle_pconn_timeout
6171 used for incoming HTTP requests.
6174 NAME: client_lifetime
6177 LOC: Config.Timeout.lifetime
6180 The maximum amount of time a client (browser) is allowed to
6181 remain connected to the cache process. This protects the Cache
6182 from having a lot of sockets (and hence file descriptors) tied up
6183 in a CLOSE_WAIT state from remote clients that go away without
6184 properly shutting down (either because of a network failure or
6185 because of a poor client implementation). The default is one
6188 NOTE: The default value is intended to be much larger than any
6189 client would ever need to be connected to your cache. You
6190 should probably change client_lifetime only as a last resort.
6191 If you seem to have many client connections tying up
6192 filedescriptors, we recommend first tuning the read_timeout,
6193 request_timeout, persistent_request_timeout and quick_abort values.
6196 NAME: pconn_lifetime
6199 LOC: Config.Timeout.pconnLifetime
6202 Desired maximum lifetime of a persistent connection.
6203 When set, Squid will close a now-idle persistent connection that
6204 exceeded configured lifetime instead of moving the connection into
6205 the idle connection pool (or equivalent). No effect on ongoing/active
6206 transactions. Connection lifetime is the time period from the
6207 connection acceptance or opening time until "now".
6209 This limit is useful in environments with long-lived connections
6210 where Squid configuration or environmental factors change during a
6211 single connection lifetime. If unrestricted, some connections may
6212 last for hours and even days, ignoring those changes that should
6213 have affected their behavior or their existence.
6215 Currently, a new lifetime value supplied via Squid reconfiguration
6216 has no effect on already idle connections unless they become busy.
6218 When set to '0' this limit is not used.
6221 NAME: half_closed_clients
6223 LOC: Config.onoff.half_closed_clients
6226 Some clients may shutdown the sending side of their TCP
6227 connections, while leaving their receiving sides open. Sometimes,
6228 Squid can not tell the difference between a half-closed and a
6229 fully-closed TCP connection.
6231 By default, Squid will immediately close client connections when
6232 read(2) returns "no more data to read."
6234 Change this option to 'on' and Squid will keep open connections
6235 until a read(2) or write(2) on the socket returns an error.
6236 This may show some benefits for reverse proxies. But if not
6237 it is recommended to leave OFF.
6240 NAME: server_idle_pconn_timeout pconn_timeout
6242 LOC: Config.Timeout.serverIdlePconn
6245 Timeout for idle persistent connections to servers and other
6252 LOC: Ident::TheConfig.timeout
6255 Maximum time to wait for IDENT lookups to complete.
6257 If this is too high, and you enabled IDENT lookups from untrusted
6258 users, you might be susceptible to denial-of-service by having
6259 many ident requests going at once.
6262 NAME: shutdown_lifetime
6265 LOC: Config.shutdownLifetime
6268 When SIGTERM or SIGHUP is received, the cache is put into
6269 "shutdown pending" mode until all active sockets are closed.
6270 This value is the lifetime to set for all open descriptors
6271 during shutdown mode. Any active clients after this many
6272 seconds will receive a 'timeout' message.
6276 ADMINISTRATIVE PARAMETERS
6277 -----------------------------------------------------------------------------
6283 LOC: Config.adminEmail
6285 Email-address of local cache manager who will receive
6286 mail if the cache dies. The default is "webmaster".
6292 LOC: Config.EmailFrom
6294 From: email-address for mail sent when the cache dies.
6295 The default is to use 'squid@unique_hostname'.
6297 See also: unique_hostname directive.
6303 LOC: Config.EmailProgram
6305 Email program used to send mail if the cache dies.
6306 The default is "mail". The specified program must comply
6307 with the standard Unix mail syntax:
6308 mail-program recipient < mailfile
6310 Optional command line options can be specified.
6313 NAME: cache_effective_user
6315 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
6316 LOC: Config.effectiveUser
6318 If you start Squid as root, it will change its effective/real
6319 UID/GID to the user specified below. The default is to change
6320 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
6321 see also; cache_effective_group
6324 NAME: cache_effective_group
6327 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
6328 LOC: Config.effectiveGroup
6330 Squid sets the GID to the effective user's default group ID
6331 (taken from the password file) and supplementary group list
6332 from the groups membership.
6334 If you want Squid to run with a specific GID regardless of
6335 the group memberships of the effective user then set this
6336 to the group (or GID) you want Squid to run as. When set
6337 all other group privileges of the effective user are ignored
6338 and only this GID is effective. If Squid is not started as
6339 root the user starting Squid MUST be member of the specified
6342 This option is not recommended by the Squid Team.
6343 Our preference is for administrators to configure a secure
6344 user account for squid with UID/GID matching system policies.
6347 NAME: httpd_suppress_version_string
6351 LOC: Config.onoff.httpd_suppress_version_string
6353 Suppress Squid version string info in HTTP headers and HTML error pages.
6356 NAME: visible_hostname
6358 LOC: Config.visibleHostname
6360 DEFAULT_DOC: Automatically detect the system host name
6362 If you want to present a special hostname in error messages, etc,
6363 define this. Otherwise, the return value of gethostname()
6364 will be used. If you have multiple caches in a cluster and
6365 get errors about IP-forwarding you must set them to have individual
6366 names with this setting.
6369 NAME: unique_hostname
6371 LOC: Config.uniqueHostname
6373 DEFAULT_DOC: Copy the value from visible_hostname
6375 If you want to have multiple machines with the same
6376 'visible_hostname' you must give each machine a different
6377 'unique_hostname' so forwarding loops can be detected.
6380 NAME: hostname_aliases
6382 LOC: Config.hostnameAliases
6385 A list of other DNS names your cache has.
6393 Minimum umask which should be enforced while the proxy
6394 is running, in addition to the umask set at startup.
6396 For a traditional octal representation of umasks, start
6401 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6402 -----------------------------------------------------------------------------
6404 This section contains parameters for the (optional) cache
6405 announcement service. This service is provided to help
6406 cache administrators locate one another in order to join or
6407 create cache hierarchies.
6409 An 'announcement' message is sent (via UDP) to the registration
6410 service by Squid. By default, the announcement message is NOT
6411 SENT unless you enable it with 'announce_period' below.
6413 The announcement message includes your hostname, plus the
6414 following information from this configuration file:
6420 All current information is processed regularly and made
6421 available on the Web at http://www.ircache.net/Cache/Tracker/.
6424 NAME: announce_period
6426 LOC: Config.Announce.period
6428 DEFAULT_DOC: Announcement messages disabled.
6430 This is how frequently to send cache announcements.
6432 To enable announcing your cache, just set an announce period.
6435 announce_period 1 day
6440 DEFAULT: tracker.ircache.net
6441 LOC: Config.Announce.host
6443 Set the hostname where announce registration messages will be sent.
6445 See also announce_port and announce_file
6451 LOC: Config.Announce.file
6453 The contents of this file will be included in the announce
6454 registration messages.
6460 LOC: Config.Announce.port
6462 Set the port where announce registration messages will be sent.
6464 See also announce_host and announce_file
6468 HTTPD-ACCELERATOR OPTIONS
6469 -----------------------------------------------------------------------------
6472 NAME: httpd_accel_surrogate_id
6475 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6476 LOC: Config.Accel.surrogate_id
6478 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6479 need an identification token to allow control targeting. Because
6480 a farm of surrogates may all perform the same tasks, they may share
6481 an identification token.
6484 NAME: http_accel_surrogate_remote
6488 LOC: Config.onoff.surrogate_is_remote
6490 Remote surrogates (such as those in a CDN) honour the header
6491 "Surrogate-Control: no-store-remote".
6493 Set this to on to have squid behave as a remote surrogate.
6497 IFDEF: USE_SQUID_ESI
6498 COMMENT: libxml2|expat|custom
6500 LOC: ESIParser::Type
6503 ESI markup is not strictly XML compatible. The custom ESI parser
6504 will give higher performance, but cannot handle non ASCII character
6509 DELAY POOL PARAMETERS
6510 -----------------------------------------------------------------------------
6514 TYPE: delay_pool_count
6516 IFDEF: USE_DELAY_POOLS
6519 This represents the number of delay pools to be used. For example,
6520 if you have one class 2 delay pool and one class 3 delays pool, you
6521 have a total of 2 delay pools.
6523 See also delay_parameters, delay_class, delay_access for pool
6524 configuration details.
6528 TYPE: delay_pool_class
6530 IFDEF: USE_DELAY_POOLS
6533 This defines the class of each delay pool. There must be exactly one
6534 delay_class line for each delay pool. For example, to define two
6535 delay pools, one of class 2 and one of class 3, the settings above
6539 delay_pools 4 # 4 delay pools
6540 delay_class 1 2 # pool 1 is a class 2 pool
6541 delay_class 2 3 # pool 2 is a class 3 pool
6542 delay_class 3 4 # pool 3 is a class 4 pool
6543 delay_class 4 5 # pool 4 is a class 5 pool
6545 The delay pool classes are:
6547 class 1 Everything is limited by a single aggregate
6550 class 2 Everything is limited by a single aggregate
6551 bucket as well as an "individual" bucket chosen
6552 from bits 25 through 32 of the IPv4 address.
6554 class 3 Everything is limited by a single aggregate
6555 bucket as well as a "network" bucket chosen
6556 from bits 17 through 24 of the IP address and a
6557 "individual" bucket chosen from bits 17 through
6558 32 of the IPv4 address.
6560 class 4 Everything in a class 3 delay pool, with an
6561 additional limit on a per user basis. This
6562 only takes effect if the username is established
6563 in advance - by forcing authentication in your
6566 class 5 Requests are grouped according their tag (see
6567 external_acl's tag= reply).
6570 Each pool also requires a delay_parameters directive to configure the pool size
6571 and speed limits used whenever the pool is applied to a request. Along with
6572 a set of delay_access directives to determine when it is used.
6574 NOTE: If an IP address is a.b.c.d
6575 -> bits 25 through 32 are "d"
6576 -> bits 17 through 24 are "c"
6577 -> bits 17 through 32 are "c * 256 + d"
6579 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6580 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6582 This clause only supports fast acl types.
6583 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6585 See also delay_parameters and delay_access.
6589 TYPE: delay_pool_access
6591 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6592 IFDEF: USE_DELAY_POOLS
6595 This is used to determine which delay pool a request falls into.
6597 delay_access is sorted per pool and the matching starts with pool 1,
6598 then pool 2, ..., and finally pool N. The first delay pool where the
6599 request is allowed is selected for the request. If it does not allow
6600 the request to any pool then the request is not delayed (default).
6602 For example, if you want some_big_clients in delay
6603 pool 1 and lotsa_little_clients in delay pool 2:
6605 delay_access 1 allow some_big_clients
6606 delay_access 1 deny all
6607 delay_access 2 allow lotsa_little_clients
6608 delay_access 2 deny all
6609 delay_access 3 allow authenticated_clients
6611 See also delay_parameters and delay_class.
6615 NAME: delay_parameters
6616 TYPE: delay_pool_rates
6618 IFDEF: USE_DELAY_POOLS
6621 This defines the parameters for a delay pool. Each delay pool has
6622 a number of "buckets" associated with it, as explained in the
6623 description of delay_class.
6625 For a class 1 delay pool, the syntax is:
6627 delay_parameters pool aggregate
6629 For a class 2 delay pool:
6631 delay_parameters pool aggregate individual
6633 For a class 3 delay pool:
6635 delay_parameters pool aggregate network individual
6637 For a class 4 delay pool:
6639 delay_parameters pool aggregate network individual user
6641 For a class 5 delay pool:
6643 delay_parameters pool tagrate
6645 The option variables are:
6647 pool a pool number - ie, a number between 1 and the
6648 number specified in delay_pools as used in
6651 aggregate the speed limit parameters for the aggregate bucket
6654 individual the speed limit parameters for the individual
6655 buckets (class 2, 3).
6657 network the speed limit parameters for the network buckets
6660 user the speed limit parameters for the user buckets
6663 tagrate the speed limit parameters for the tag buckets
6666 A pair of delay parameters is written restore/maximum, where restore is
6667 the number of bytes (not bits - modem and network speeds are usually
6668 quoted in bits) per second placed into the bucket, and maximum is the
6669 maximum number of bytes which can be in the bucket at any time.
6671 There must be one delay_parameters line for each delay pool.
6674 For example, if delay pool number 1 is a class 2 delay pool as in the
6675 above example, and is being used to strictly limit each host to 64Kbit/sec
6676 (plus overheads), with no overall limit, the line is:
6678 delay_parameters 1 none 8000/8000
6680 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6682 Note that the word 'none' is used to represent no limit.
6685 And, if delay pool number 2 is a class 3 delay pool as in the above
6686 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6687 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6688 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6689 to permit a decent web page to be downloaded at a decent speed
6690 (if the network is not being limited due to overuse) but slow down
6691 large downloads more significantly:
6693 delay_parameters 2 32000/32000 8000/8000 600/8000
6695 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6696 8 x 8000 KByte/sec -> 64Kbit/sec.
6697 8 x 600 Byte/sec -> 4800bit/sec.
6700 Finally, for a class 4 delay pool as in the example - each user will
6701 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6703 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6706 See also delay_class and delay_access.
6710 NAME: delay_initial_bucket_level
6711 COMMENT: (percent, 0-100)
6714 IFDEF: USE_DELAY_POOLS
6715 LOC: Config.Delay.initial
6717 The initial bucket percentage is used to determine how much is put
6718 in each bucket when squid starts, is reconfigured, or first notices
6719 a host accessing it (in class 2 and class 3, individual hosts and
6720 networks only have buckets associated with them once they have been
6725 CLIENT DELAY POOL PARAMETERS
6726 -----------------------------------------------------------------------------
6729 NAME: client_delay_pools
6730 TYPE: client_delay_pool_count
6732 IFDEF: USE_DELAY_POOLS
6733 LOC: Config.ClientDelay
6735 This option specifies the number of client delay pools used. It must
6736 preceed other client_delay_* options.
6739 client_delay_pools 2
6741 See also client_delay_parameters and client_delay_access.
6744 NAME: client_delay_initial_bucket_level
6745 COMMENT: (percent, 0-no_limit)
6748 IFDEF: USE_DELAY_POOLS
6749 LOC: Config.ClientDelay.initial
6751 This option determines the initial bucket size as a percentage of
6752 max_bucket_size from client_delay_parameters. Buckets are created
6753 at the time of the "first" connection from the matching IP. Idle
6754 buckets are periodically deleted up.
6756 You can specify more than 100 percent but note that such "oversized"
6757 buckets are not refilled until their size goes down to max_bucket_size
6758 from client_delay_parameters.
6761 client_delay_initial_bucket_level 50
6764 NAME: client_delay_parameters
6765 TYPE: client_delay_pool_rates
6767 IFDEF: USE_DELAY_POOLS
6768 LOC: Config.ClientDelay
6771 This option configures client-side bandwidth limits using the
6774 client_delay_parameters pool speed_limit max_bucket_size
6776 pool is an integer ID used for client_delay_access matching.
6778 speed_limit is bytes added to the bucket per second.
6780 max_bucket_size is the maximum size of a bucket, enforced after any
6781 speed_limit additions.
6783 Please see the delay_parameters option for more information and
6787 client_delay_parameters 1 1024 2048
6788 client_delay_parameters 2 51200 16384
6790 See also client_delay_access.
6794 NAME: client_delay_access
6795 TYPE: client_delay_pool_access
6797 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6798 IFDEF: USE_DELAY_POOLS
6799 LOC: Config.ClientDelay
6801 This option determines the client-side delay pool for the
6804 client_delay_access pool_ID allow|deny acl_name
6806 All client_delay_access options are checked in their pool ID
6807 order, starting with pool 1. The first checked pool with allowed
6808 request is selected for the request. If no ACL matches or there
6809 are no client_delay_access options, the request bandwidth is not
6812 The ACL-selected pool is then used to find the
6813 client_delay_parameters for the request. Client-side pools are
6814 not used to aggregate clients. Clients are always aggregated
6815 based on their source IP addresses (one bucket per source IP).
6817 This clause only supports fast acl types.
6818 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6819 Additionally, only the client TCP connection details are available.
6820 ACLs testing HTTP properties will not work.
6822 Please see delay_access for more examples.
6825 client_delay_access 1 allow low_rate_network
6826 client_delay_access 2 allow vips_network
6829 See also client_delay_parameters and client_delay_pools.
6833 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6834 -----------------------------------------------------------------------------
6839 LOC: Config.Wccp.router
6841 DEFAULT_DOC: WCCP disabled.
6844 Use this option to define your WCCP ``home'' router for
6847 wccp_router supports a single WCCP(v1) router
6849 wccp2_router supports multiple WCCPv2 routers
6851 only one of the two may be used at the same time and defines
6852 which version of WCCP to use.
6856 TYPE: IpAddress_list
6857 LOC: Config.Wccp2.router
6859 DEFAULT_DOC: WCCPv2 disabled.
6862 Use this option to define your WCCP ``home'' router for
6865 wccp_router supports a single WCCP(v1) router
6867 wccp2_router supports multiple WCCPv2 routers
6869 only one of the two may be used at the same time and defines
6870 which version of WCCP to use.
6875 LOC: Config.Wccp.version
6879 This directive is only relevant if you need to set up WCCP(v1)
6880 to some very old and end-of-life Cisco routers. In all other
6881 setups it must be left unset or at the default setting.
6882 It defines an internal version in the WCCP(v1) protocol,
6883 with version 4 being the officially documented protocol.
6885 According to some users, Cisco IOS 11.2 and earlier only
6886 support WCCP version 3. If you're using that or an earlier
6887 version of IOS, you may need to change this value to 3, otherwise
6888 do not specify this parameter.
6891 NAME: wccp2_rebuild_wait
6893 LOC: Config.Wccp2.rebuildwait
6897 If this is enabled Squid will wait for the cache dir rebuild to finish
6898 before sending the first wccp2 HereIAm packet
6901 NAME: wccp2_forwarding_method
6903 LOC: Config.Wccp2.forwarding_method
6907 WCCP2 allows the setting of forwarding methods between the
6908 router/switch and the cache. Valid values are as follows:
6910 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6911 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6913 Currently (as of IOS 12.4) cisco routers only support GRE.
6914 Cisco switches only support the L2 redirect assignment method.
6917 NAME: wccp2_return_method
6919 LOC: Config.Wccp2.return_method
6923 WCCP2 allows the setting of return methods between the
6924 router/switch and the cache for packets that the cache
6925 decides not to handle. Valid values are as follows:
6927 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6928 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6930 Currently (as of IOS 12.4) cisco routers only support GRE.
6931 Cisco switches only support the L2 redirect assignment.
6933 If the "ip wccp redirect exclude in" command has been
6934 enabled on the cache interface, then it is still safe for
6935 the proxy server to use a l2 redirect method even if this
6936 option is set to GRE.
6939 NAME: wccp2_assignment_method
6941 LOC: Config.Wccp2.assignment_method
6945 WCCP2 allows the setting of methods to assign the WCCP hash
6946 Valid values are as follows:
6948 hash - Hash assignment
6949 mask - Mask assignment
6951 As a general rule, cisco routers support the hash assignment method
6952 and cisco switches support the mask assignment method.
6957 LOC: Config.Wccp2.info
6958 DEFAULT_IF_NONE: standard 0
6959 DEFAULT_DOC: Use the 'web-cache' standard service.
6962 WCCP2 allows for multiple traffic services. There are two
6963 types: "standard" and "dynamic". The standard type defines
6964 one service id - http (id 0). The dynamic service ids can be from
6965 51 to 255 inclusive. In order to use a dynamic service id
6966 one must define the type of traffic to be redirected; this is done
6967 using the wccp2_service_info option.
6969 The "standard" type does not require a wccp2_service_info option,
6970 just specifying the service id will suffice.
6972 MD5 service authentication can be enabled by adding
6973 "password=<password>" to the end of this service declaration.
6977 wccp2_service standard 0 # for the 'web-cache' standard service
6978 wccp2_service dynamic 80 # a dynamic service type which will be
6979 # fleshed out with subsequent options.
6980 wccp2_service standard 0 password=foo
6983 NAME: wccp2_service_info
6984 TYPE: wccp2_service_info
6985 LOC: Config.Wccp2.info
6989 Dynamic WCCPv2 services require further information to define the
6990 traffic you wish to have diverted.
6994 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6995 priority=<priority> ports=<port>,<port>..
6997 The relevant WCCPv2 flags:
6998 + src_ip_hash, dst_ip_hash
6999 + source_port_hash, dst_port_hash
7000 + src_ip_alt_hash, dst_ip_alt_hash
7001 + src_port_alt_hash, dst_port_alt_hash
7004 The port list can be one to eight entries.
7008 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
7009 priority=240 ports=80
7011 Note: the service id must have been defined by a previous
7012 'wccp2_service dynamic <id>' entry.
7017 LOC: Config.Wccp2.weight
7021 Each cache server gets assigned a set of the destination
7022 hash proportional to their weight.
7027 LOC: Config.Wccp.address
7029 DEFAULT_DOC: Address selected by the operating system.
7032 Use this option if you require WCCPv2 to use a specific
7035 The default behavior is to not bind to any specific address.
7040 LOC: Config.Wccp2.address
7042 DEFAULT_DOC: Address selected by the operating system.
7045 Use this option if you require WCCP to use a specific
7048 The default behavior is to not bind to any specific address.
7052 PERSISTENT CONNECTION HANDLING
7053 -----------------------------------------------------------------------------
7055 Also see "pconn_timeout" in the TIMEOUTS section
7058 NAME: client_persistent_connections
7060 LOC: Config.onoff.client_pconns
7063 Persistent connection support for clients.
7064 Squid uses persistent connections (when allowed). You can use
7065 this option to disable persistent connections with clients.
7068 NAME: server_persistent_connections
7070 LOC: Config.onoff.server_pconns
7073 Persistent connection support for servers.
7074 Squid uses persistent connections (when allowed). You can use
7075 this option to disable persistent connections with servers.
7078 NAME: persistent_connection_after_error
7080 LOC: Config.onoff.error_pconns
7083 With this directive the use of persistent connections after
7084 HTTP errors can be disabled. Useful if you have clients
7085 who fail to handle errors on persistent connections proper.
7088 NAME: detect_broken_pconn
7090 LOC: Config.onoff.detect_broken_server_pconns
7093 Some servers have been found to incorrectly signal the use
7094 of HTTP/1.0 persistent connections even on replies not
7095 compatible, causing significant delays. This server problem
7096 has mostly been seen on redirects.
7098 By enabling this directive Squid attempts to detect such
7099 broken replies and automatically assume the reply is finished
7100 after 10 seconds timeout.
7104 CACHE DIGEST OPTIONS
7105 -----------------------------------------------------------------------------
7108 NAME: digest_generation
7109 IFDEF: USE_CACHE_DIGESTS
7111 LOC: Config.onoff.digest_generation
7114 This controls whether the server will generate a Cache Digest
7115 of its contents. By default, Cache Digest generation is
7116 enabled if Squid is compiled with --enable-cache-digests defined.
7119 NAME: digest_bits_per_entry
7120 IFDEF: USE_CACHE_DIGESTS
7122 LOC: Config.digest.bits_per_entry
7125 This is the number of bits of the server's Cache Digest which
7126 will be associated with the Digest entry for a given HTTP
7127 Method and URL (public key) combination. The default is 5.
7130 NAME: digest_rebuild_period
7131 IFDEF: USE_CACHE_DIGESTS
7134 LOC: Config.digest.rebuild_period
7137 This is the wait time between Cache Digest rebuilds.
7140 NAME: digest_rewrite_period
7142 IFDEF: USE_CACHE_DIGESTS
7144 LOC: Config.digest.rewrite_period
7147 This is the wait time between Cache Digest writes to
7151 NAME: digest_swapout_chunk_size
7154 IFDEF: USE_CACHE_DIGESTS
7155 LOC: Config.digest.swapout_chunk_size
7158 This is the number of bytes of the Cache Digest to write to
7159 disk at a time. It defaults to 4096 bytes (4KB), the Squid
7163 NAME: digest_rebuild_chunk_percentage
7164 COMMENT: (percent, 0-100)
7165 IFDEF: USE_CACHE_DIGESTS
7167 LOC: Config.digest.rebuild_chunk_percentage
7170 This is the percentage of the Cache Digest to be scanned at a
7171 time. By default it is set to 10% of the Cache Digest.
7176 -----------------------------------------------------------------------------
7181 LOC: Config.Port.snmp
7183 DEFAULT_DOC: SNMP disabled.
7186 The port number where Squid listens for SNMP requests. To enable
7187 SNMP support set this to a suitable port number. Port number
7188 3401 is often used for the Squid SNMP agent. By default it's
7189 set to "0" (disabled)
7197 LOC: Config.accessList.snmp
7199 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
7202 Allowing or denying access to the SNMP port.
7204 All access to the agent is denied by default.
7207 snmp_access allow|deny [!]aclname ...
7209 This clause only supports fast acl types.
7210 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7213 snmp_access allow snmppublic localhost
7214 snmp_access deny all
7217 NAME: snmp_incoming_address
7219 LOC: Config.Addrs.snmp_incoming
7221 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
7224 Just like 'udp_incoming_address', but for the SNMP port.
7226 snmp_incoming_address is used for the SNMP socket receiving
7227 messages from SNMP agents.
7229 The default snmp_incoming_address is to listen on all
7230 available network interfaces.
7233 NAME: snmp_outgoing_address
7235 LOC: Config.Addrs.snmp_outgoing
7237 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
7240 Just like 'udp_outgoing_address', but for the SNMP port.
7242 snmp_outgoing_address is used for SNMP packets returned to SNMP
7245 If snmp_outgoing_address is not set it will use the same socket
7246 as snmp_incoming_address. Only change this if you want to have
7247 SNMP replies sent using another address than where this Squid
7248 listens for SNMP queries.
7250 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
7251 the same value since they both use the same port.
7256 -----------------------------------------------------------------------------
7259 NAME: icp_port udp_port
7262 DEFAULT_DOC: ICP disabled.
7263 LOC: Config.Port.icp
7265 The port number where Squid sends and receives ICP queries to
7266 and from neighbor caches. The standard UDP port for ICP is 3130.
7269 icp_port @DEFAULT_ICP_PORT@
7276 DEFAULT_DOC: HTCP disabled.
7277 LOC: Config.Port.htcp
7279 The port number where Squid sends and receives HTCP queries to
7280 and from neighbor caches. To turn it on you want to set it to
7287 NAME: log_icp_queries
7291 LOC: Config.onoff.log_udp
7293 If set, ICP queries are logged to access.log. You may wish
7294 do disable this if your ICP load is VERY high to speed things
7295 up or to simplify log analysis.
7298 NAME: udp_incoming_address
7300 LOC:Config.Addrs.udp_incoming
7302 DEFAULT_DOC: Accept packets from all machine interfaces.
7304 udp_incoming_address is used for UDP packets received from other
7307 The default behavior is to not bind to any specific address.
7309 Only change this if you want to have all UDP queries received on
7310 a specific interface/address.
7312 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
7313 modules. Altering it will affect all of them in the same manner.
7315 see also; udp_outgoing_address
7317 NOTE, udp_incoming_address and udp_outgoing_address can not
7318 have the same value since they both use the same port.
7321 NAME: udp_outgoing_address
7323 LOC: Config.Addrs.udp_outgoing
7325 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
7327 udp_outgoing_address is used for UDP packets sent out to other
7330 The default behavior is to not bind to any specific address.
7332 Instead it will use the same socket as udp_incoming_address.
7333 Only change this if you want to have UDP queries sent using another
7334 address than where this Squid listens for UDP queries from other
7337 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
7338 modules. Altering it will affect all of them in the same manner.
7340 see also; udp_incoming_address
7342 NOTE, udp_incoming_address and udp_outgoing_address can not
7343 have the same value since they both use the same port.
7350 LOC: Config.onoff.icp_hit_stale
7352 If you want to return ICP_HIT for stale cache objects, set this
7353 option to 'on'. If you have sibling relationships with caches
7354 in other administrative domains, this should be 'off'. If you only
7355 have sibling relationships with caches under your control,
7356 it is probably okay to set this to 'on'.
7357 If set to 'on', your siblings should use the option "allow-miss"
7358 on their cache_peer lines for connecting to you.
7361 NAME: minimum_direct_hops
7364 LOC: Config.minDirectHops
7366 If using the ICMP pinging stuff, do direct fetches for sites
7367 which are no more than this many hops away.
7370 NAME: minimum_direct_rtt
7374 LOC: Config.minDirectRtt
7376 If using the ICMP pinging stuff, do direct fetches for sites
7377 which are no more than this many rtt milliseconds away.
7383 LOC: Config.Netdb.low
7385 The low water mark for the ICMP measurement database.
7387 Note: high watermark controlled by netdb_high directive.
7389 These watermarks are counts, not percents. The defaults are
7390 (low) 900 and (high) 1000. When the high water mark is
7391 reached, database entries will be deleted until the low
7398 LOC: Config.Netdb.high
7400 The high water mark for the ICMP measurement database.
7402 Note: low watermark controlled by netdb_low directive.
7404 These watermarks are counts, not percents. The defaults are
7405 (low) 900 and (high) 1000. When the high water mark is
7406 reached, database entries will be deleted until the low
7410 NAME: netdb_ping_period
7412 LOC: Config.Netdb.period
7415 The minimum period for measuring a site. There will be at
7416 least this much delay between successive pings to the same
7417 network. The default is five minutes.
7424 LOC: Config.onoff.query_icmp
7426 If you want to ask your peers to include ICMP data in their ICP
7427 replies, enable this option.
7429 If your peer has configured Squid (during compilation) with
7430 '--enable-icmp' that peer will send ICMP pings to origin server
7431 sites of the URLs it receives. If you enable this option the
7432 ICP replies from that peer will include the ICMP data (if available).
7433 Then, when choosing a parent cache, Squid will choose the parent with
7434 the minimal RTT to the origin server. When this happens, the
7435 hierarchy field of the access.log will be
7436 "CLOSEST_PARENT_MISS". This option is off by default.
7439 NAME: test_reachability
7443 LOC: Config.onoff.test_reachability
7445 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7446 instead of ICP_MISS if the target host is NOT in the ICMP
7447 database, or has a zero RTT.
7450 NAME: icp_query_timeout
7453 DEFAULT_DOC: Dynamic detection.
7455 LOC: Config.Timeout.icp_query
7457 Normally Squid will automatically determine an optimal ICP
7458 query timeout value based on the round-trip-time of recent ICP
7459 queries. If you want to override the value determined by
7460 Squid, set this 'icp_query_timeout' to a non-zero value. This
7461 value is specified in MILLISECONDS, so, to use a 2-second
7462 timeout (the old default), you would write:
7464 icp_query_timeout 2000
7467 NAME: maximum_icp_query_timeout
7471 LOC: Config.Timeout.icp_query_max
7473 Normally the ICP query timeout is determined dynamically. But
7474 sometimes it can lead to very large values (say 5 seconds).
7475 Use this option to put an upper limit on the dynamic timeout
7476 value. Do NOT use this option to always use a fixed (instead
7477 of a dynamic) timeout value. To set a fixed timeout see the
7478 'icp_query_timeout' directive.
7481 NAME: minimum_icp_query_timeout
7485 LOC: Config.Timeout.icp_query_min
7487 Normally the ICP query timeout is determined dynamically. But
7488 sometimes it can lead to very small timeouts, even lower than
7489 the normal latency variance on your link due to traffic.
7490 Use this option to put an lower limit on the dynamic timeout
7491 value. Do NOT use this option to always use a fixed (instead
7492 of a dynamic) timeout value. To set a fixed timeout see the
7493 'icp_query_timeout' directive.
7496 NAME: background_ping_rate
7500 LOC: Config.backgroundPingRate
7502 Controls how often the ICP pings are sent to siblings that
7503 have background-ping set.
7507 MULTICAST ICP OPTIONS
7508 -----------------------------------------------------------------------------
7513 LOC: Config.mcast_group_list
7516 This tag specifies a list of multicast groups which your server
7517 should join to receive multicasted ICP queries.
7519 NOTE! Be very careful what you put here! Be sure you
7520 understand the difference between an ICP _query_ and an ICP
7521 _reply_. This option is to be set only if you want to RECEIVE
7522 multicast queries. Do NOT set this option to SEND multicast
7523 ICP (use cache_peer for that). ICP replies are always sent via
7524 unicast, so this option does not affect whether or not you will
7525 receive replies from multicast group members.
7527 You must be very careful to NOT use a multicast address which
7528 is already in use by another group of caches.
7530 If you are unsure about multicast, please read the Multicast
7531 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7533 Usage: mcast_groups 239.128.16.128 224.0.1.20
7535 By default, Squid doesn't listen on any multicast groups.
7538 NAME: mcast_miss_addr
7539 IFDEF: MULTICAST_MISS_STREAM
7541 LOC: Config.mcast_miss.addr
7543 DEFAULT_DOC: disabled.
7545 If you enable this option, every "cache miss" URL will
7546 be sent out on the specified multicast address.
7548 Do not enable this option unless you are are absolutely
7549 certain you understand what you are doing.
7552 NAME: mcast_miss_ttl
7553 IFDEF: MULTICAST_MISS_STREAM
7555 LOC: Config.mcast_miss.ttl
7558 This is the time-to-live value for packets multicasted
7559 when multicasting off cache miss URLs is enabled. By
7560 default this is set to 'site scope', i.e. 16.
7563 NAME: mcast_miss_port
7564 IFDEF: MULTICAST_MISS_STREAM
7566 LOC: Config.mcast_miss.port
7569 This is the port number to be used in conjunction with
7573 NAME: mcast_miss_encode_key
7574 IFDEF: MULTICAST_MISS_STREAM
7576 LOC: Config.mcast_miss.encode_key
7577 DEFAULT: XXXXXXXXXXXXXXXX
7579 The URLs that are sent in the multicast miss stream are
7580 encrypted. This is the encryption key.
7583 NAME: mcast_icp_query_timeout
7587 LOC: Config.Timeout.mcast_icp_query
7589 For multicast peers, Squid regularly sends out ICP "probes" to
7590 count how many other peers are listening on the given multicast
7591 address. This value specifies how long Squid should wait to
7592 count all the replies. The default is 2000 msec, or 2
7597 INTERNAL ICON OPTIONS
7598 -----------------------------------------------------------------------------
7601 NAME: icon_directory
7603 LOC: Config.icons.directory
7604 DEFAULT: @DEFAULT_ICON_DIR@
7606 Where the icons are stored. These are normally kept in
7610 NAME: global_internal_static
7612 LOC: Config.onoff.global_internal_static
7615 This directive controls is Squid should intercept all requests for
7616 /squid-internal-static/ no matter which host the URL is requesting
7617 (default on setting), or if nothing special should be done for
7618 such URLs (off setting). The purpose of this directive is to make
7619 icons etc work better in complex cache hierarchies where it may
7620 not always be possible for all corners in the cache mesh to reach
7621 the server generating a directory listing.
7624 NAME: short_icon_urls
7626 LOC: Config.icons.use_short_names
7629 If this is enabled Squid will use short URLs for icons.
7630 If disabled it will revert to the old behavior of including
7631 it's own name and port in the URL.
7633 If you run a complex cache hierarchy with a mix of Squid and
7634 other proxies you may need to disable this directive.
7639 -----------------------------------------------------------------------------
7642 NAME: error_directory
7644 LOC: Config.errorDirectory
7646 DEFAULT_DOC: Send error pages in the clients preferred language
7648 If you wish to create your own versions of the default
7649 error files to customize them to suit your company copy
7650 the error/template files to another directory and point
7653 WARNING: This option will disable multi-language support
7654 on error pages if used.
7656 The squid developers are interested in making squid available in
7657 a wide variety of languages. If you are making translations for a
7658 language that Squid does not currently provide please consider
7659 contributing your translation back to the project.
7660 http://wiki.squid-cache.org/Translations
7662 The squid developers working on translations are happy to supply drop-in
7663 translated error files in exchange for any new language contributions.
7666 NAME: error_default_language
7667 IFDEF: USE_ERR_LOCALES
7669 LOC: Config.errorDefaultLanguage
7671 DEFAULT_DOC: Generate English language pages.
7673 Set the default language which squid will send error pages in
7674 if no existing translation matches the clients language
7677 If unset (default) generic English will be used.
7679 The squid developers are interested in making squid available in
7680 a wide variety of languages. If you are interested in making
7681 translations for any language see the squid wiki for details.
7682 http://wiki.squid-cache.org/Translations
7685 NAME: error_log_languages
7686 IFDEF: USE_ERR_LOCALES
7688 LOC: Config.errorLogMissingLanguages
7691 Log to cache.log what languages users are attempting to
7692 auto-negotiate for translations.
7694 Successful negotiations are not logged. Only failures
7695 have meaning to indicate that Squid may need an upgrade
7696 of its error page translations.
7699 NAME: err_page_stylesheet
7701 LOC: Config.errorStylesheet
7702 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7704 CSS Stylesheet to pattern the display of Squid default error pages.
7706 For information on CSS see http://www.w3.org/Style/CSS/
7711 LOC: Config.errHtmlText
7714 HTML text to include in error messages. Make this a "mailto"
7715 URL to your admin address, or maybe just a link to your
7716 organizations Web page.
7718 To include this in your error messages, you must rewrite
7719 the error template files (found in the "errors" directory).
7720 Wherever you want the 'err_html_text' line to appear,
7721 insert a %L tag in the error template file.
7724 NAME: email_err_data
7727 LOC: Config.onoff.emailErrData
7730 If enabled, information about the occurred error will be
7731 included in the mailto links of the ERR pages (if %W is set)
7732 so that the email body contains the data.
7733 Syntax is <A HREF="mailto:%w%W">%w</A>
7738 LOC: Config.denyInfoList
7741 Usage: deny_info err_page_name acl
7742 or deny_info http://... acl
7743 or deny_info TCP_RESET acl
7745 This can be used to return a ERR_ page for requests which
7746 do not pass the 'http_access' rules. Squid remembers the last
7747 acl it evaluated in http_access, and if a 'deny_info' line exists
7748 for that ACL Squid returns a corresponding error page.
7750 The acl is typically the last acl on the http_access deny line which
7751 denied access. The exceptions to this rule are:
7752 - When Squid needs to request authentication credentials. It's then
7753 the first authentication related acl encountered
7754 - When none of the http_access lines matches. It's then the last
7755 acl processed on the last http_access line.
7756 - When the decision to deny access was made by an adaptation service,
7757 the acl name is the corresponding eCAP or ICAP service_name.
7759 NP: If providing your own custom error pages with error_directory
7760 you may also specify them by your custom file name:
7761 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7763 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7764 may be specified by prefixing the file name with the code and a colon.
7765 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7767 Alternatively you can tell Squid to reset the TCP connection
7768 by specifying TCP_RESET.
7770 Or you can specify an error URL or URL pattern. The browsers will
7771 get redirected to the specified URL after formatting tags have
7772 been replaced. Redirect will be done with 302 or 307 according to
7773 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7774 the URL. e.g. 303:http://example.com/
7777 %a - username (if available. Password NOT included)
7780 %E - Error description
7782 %H - Request domain name
7783 %i - Client IP Address
7785 %o - Message result from external ACL helper
7786 %p - Request Port number
7787 %P - Request Protocol name
7788 %R - Request URL path
7789 %T - Timestamp in RFC 1123 format
7790 %U - Full canonical URL from client
7791 (HTTPS URLs terminate with *)
7792 %u - Full canonical URL from client
7793 %w - Admin email from squid.conf
7795 %% - Literal percent (%) code
7800 OPTIONS INFLUENCING REQUEST FORWARDING
7801 -----------------------------------------------------------------------------
7804 NAME: nonhierarchical_direct
7806 LOC: Config.onoff.nonhierarchical_direct
7809 By default, Squid will send any non-hierarchical requests
7810 (not cacheable request type) direct to origin servers.
7812 When this is set to "off", Squid will prefer to send these
7813 requests to parents.
7815 Note that in most configurations, by turning this off you will only
7816 add latency to these request without any improvement in global hit
7819 This option only sets a preference. If the parent is unavailable a
7820 direct connection to the origin server may still be attempted. To
7821 completely prevent direct connections use never_direct.
7826 LOC: Config.onoff.prefer_direct
7829 Normally Squid tries to use parents for most requests. If you for some
7830 reason like it to first try going direct and only use a parent if
7831 going direct fails set this to on.
7833 By combining nonhierarchical_direct off and prefer_direct on you
7834 can set up Squid to use a parent as a backup path if going direct
7837 Note: If you want Squid to use parents for all requests see
7838 the never_direct directive. prefer_direct only modifies how Squid
7839 acts on cacheable requests.
7842 NAME: cache_miss_revalidate
7846 LOC: Config.onoff.cache_miss_revalidate
7848 RFC 7232 defines a conditional request mechanism to prevent
7849 response objects being unnecessarily transferred over the network.
7850 If that mechanism is used by the client and a cache MISS occurs
7851 it can prevent new cache entries being created.
7853 This option determines whether Squid on cache MISS will pass the
7854 client revalidation request to the server or tries to fetch new
7855 content for caching. It can be useful while the cache is mostly
7856 empty to more quickly have the cache populated by generating
7857 non-conditional GETs.
7859 When set to 'on' (default), Squid will pass all client If-* headers
7860 to the server. This permits server responses without a cacheable
7861 payload to be delivered and on MISS no new cache entry is created.
7863 When set to 'off' and if the request is cacheable, Squid will
7864 remove the clients If-Modified-Since and If-None-Match headers from
7865 the request sent to the server. This requests a 200 status response
7866 from the server to create a new cache entry with.
7871 LOC: Config.accessList.AlwaysDirect
7873 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7875 Usage: always_direct allow|deny [!]aclname ...
7877 Here you can use ACL elements to specify requests which should
7878 ALWAYS be forwarded by Squid to the origin servers without using
7879 any peers. For example, to always directly forward requests for
7880 local servers ignoring any parents or siblings you may have use
7883 acl local-servers dstdomain my.domain.net
7884 always_direct allow local-servers
7886 To always forward FTP requests directly, use
7889 always_direct allow FTP
7891 NOTE: There is a similar, but opposite option named
7892 'never_direct'. You need to be aware that "always_direct deny
7893 foo" is NOT the same thing as "never_direct allow foo". You
7894 may need to use a deny rule to exclude a more-specific case of
7895 some other rule. Example:
7897 acl local-external dstdomain external.foo.net
7898 acl local-servers dstdomain .foo.net
7899 always_direct deny local-external
7900 always_direct allow local-servers
7902 NOTE: If your goal is to make the client forward the request
7903 directly to the origin server bypassing Squid then this needs
7904 to be done in the client configuration. Squid configuration
7905 can only tell Squid how Squid should fetch the object.
7907 NOTE: This directive is not related to caching. The replies
7908 is cached as usual even if you use always_direct. To not cache
7909 the replies see the 'cache' directive.
7911 This clause supports both fast and slow acl types.
7912 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7917 LOC: Config.accessList.NeverDirect
7919 DEFAULT_DOC: Allow DNS results to be used for this request.
7921 Usage: never_direct allow|deny [!]aclname ...
7923 never_direct is the opposite of always_direct. Please read
7924 the description for always_direct if you have not already.
7926 With 'never_direct' you can use ACL elements to specify
7927 requests which should NEVER be forwarded directly to origin
7928 servers. For example, to force the use of a proxy for all
7929 requests, except those in your local domain use something like:
7931 acl local-servers dstdomain .foo.net
7932 never_direct deny local-servers
7933 never_direct allow all
7935 or if Squid is inside a firewall and there are local intranet
7936 servers inside the firewall use something like:
7938 acl local-intranet dstdomain .foo.net
7939 acl local-external dstdomain external.foo.net
7940 always_direct deny local-external
7941 always_direct allow local-intranet
7942 never_direct allow all
7944 This clause supports both fast and slow acl types.
7945 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7949 ADVANCED NETWORKING OPTIONS
7950 -----------------------------------------------------------------------------
7953 NAME: incoming_udp_average incoming_icp_average
7956 LOC: Config.comm_incoming.udp.average
7958 Heavy voodoo here. I can't even believe you are reading this.
7959 Are you crazy? Don't even think about adjusting these unless
7960 you understand the algorithms in comm_select.c first!
7963 NAME: incoming_tcp_average incoming_http_average
7966 LOC: Config.comm_incoming.tcp.average
7968 Heavy voodoo here. I can't even believe you are reading this.
7969 Are you crazy? Don't even think about adjusting these unless
7970 you understand the algorithms in comm_select.c first!
7973 NAME: incoming_dns_average
7976 LOC: Config.comm_incoming.dns.average
7978 Heavy voodoo here. I can't even believe you are reading this.
7979 Are you crazy? Don't even think about adjusting these unless
7980 you understand the algorithms in comm_select.c first!
7983 NAME: min_udp_poll_cnt min_icp_poll_cnt
7986 LOC: Config.comm_incoming.udp.min_poll
7988 Heavy voodoo here. I can't even believe you are reading this.
7989 Are you crazy? Don't even think about adjusting these unless
7990 you understand the algorithms in comm_select.c first!
7993 NAME: min_dns_poll_cnt
7996 LOC: Config.comm_incoming.dns.min_poll
7998 Heavy voodoo here. I can't even believe you are reading this.
7999 Are you crazy? Don't even think about adjusting these unless
8000 you understand the algorithms in comm_select.c first!
8003 NAME: min_tcp_poll_cnt min_http_poll_cnt
8006 LOC: Config.comm_incoming.tcp.min_poll
8008 Heavy voodoo here. I can't even believe you are reading this.
8009 Are you crazy? Don't even think about adjusting these unless
8010 you understand the algorithms in comm_select.c first!
8016 LOC: Config.accept_filter
8020 The name of an accept(2) filter to install on Squid's
8021 listen socket(s). This feature is perhaps specific to
8022 FreeBSD and requires support in the kernel.
8024 The 'httpready' filter delays delivering new connections
8025 to Squid until a full HTTP request has been received.
8026 See the accf_http(9) man page for details.
8028 The 'dataready' filter delays delivering new connections
8029 to Squid until there is some data to process.
8030 See the accf_dataready(9) man page for details.
8034 The 'data' filter delays delivering of new connections
8035 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
8036 You may optionally specify a number of seconds to wait by
8037 'data=N' where N is the number of seconds. Defaults to 30
8038 if not specified. See the tcp(7) man page for details.
8041 accept_filter httpready
8046 NAME: client_ip_max_connections
8048 LOC: Config.client_ip_max_connections
8050 DEFAULT_DOC: No limit.
8052 Set an absolute limit on the number of connections a single
8053 client IP can use. Any more than this and Squid will begin to drop
8054 new connections from the client until it closes some links.
8056 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
8057 connections from the client. For finer control use the ACL access controls.
8059 Requires client_db to be enabled (the default).
8061 WARNING: This may noticably slow down traffic received via external proxies
8062 or NAT devices and cause them to rebound error messages back to their clients.
8065 NAME: tcp_recv_bufsize
8069 DEFAULT_DOC: Use operating system TCP defaults.
8070 LOC: Config.tcpRcvBufsz
8072 Size of receive buffer to set for TCP sockets. Probably just
8073 as easy to change your kernel's default.
8074 Omit from squid.conf to use the default buffer size.
8079 -----------------------------------------------------------------------------
8086 LOC: Adaptation::Icap::TheConfig.onoff
8089 If you want to enable the ICAP module support, set this to on.
8092 NAME: icap_connect_timeout
8095 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
8098 This parameter specifies how long to wait for the TCP connect to
8099 the requested ICAP server to complete before giving up and either
8100 terminating the HTTP transaction or bypassing the failure.
8102 The default for optional services is peer_connect_timeout.
8103 The default for essential services is connect_timeout.
8104 If this option is explicitly set, its value applies to all services.
8107 NAME: icap_io_timeout
8111 DEFAULT_DOC: Use read_timeout.
8112 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
8115 This parameter specifies how long to wait for an I/O activity on
8116 an established, active ICAP connection before giving up and
8117 either terminating the HTTP transaction or bypassing the
8121 NAME: icap_service_failure_limit
8122 COMMENT: limit [in memory-depth time-units]
8123 TYPE: icap_service_failure_limit
8125 LOC: Adaptation::Icap::TheConfig
8128 The limit specifies the number of failures that Squid tolerates
8129 when establishing a new TCP connection with an ICAP service. If
8130 the number of failures exceeds the limit, the ICAP service is
8131 not used for new ICAP requests until it is time to refresh its
8134 A negative value disables the limit. Without the limit, an ICAP
8135 service will not be considered down due to connectivity failures
8136 between ICAP OPTIONS requests.
8138 Squid forgets ICAP service failures older than the specified
8139 value of memory-depth. The memory fading algorithm
8140 is approximate because Squid does not remember individual
8141 errors but groups them instead, splitting the option
8142 value into ten time slots of equal length.
8144 When memory-depth is 0 and by default this option has no
8145 effect on service failure expiration.
8147 Squid always forgets failures when updating service settings
8148 using an ICAP OPTIONS transaction, regardless of this option
8152 # suspend service usage after 10 failures in 5 seconds:
8153 icap_service_failure_limit 10 in 5 seconds
8156 NAME: icap_service_revival_delay
8159 LOC: Adaptation::Icap::TheConfig.service_revival_delay
8162 The delay specifies the number of seconds to wait after an ICAP
8163 OPTIONS request failure before requesting the options again. The
8164 failed ICAP service is considered "down" until fresh OPTIONS are
8167 The actual delay cannot be smaller than the hardcoded minimum
8168 delay of 30 seconds.
8171 NAME: icap_preview_enable
8175 LOC: Adaptation::Icap::TheConfig.preview_enable
8178 The ICAP Preview feature allows the ICAP server to handle the
8179 HTTP message by looking only at the beginning of the message body
8180 or even without receiving the body at all. In some environments,
8181 previews greatly speedup ICAP processing.
8183 During an ICAP OPTIONS transaction, the server may tell Squid what
8184 HTTP messages should be previewed and how big the preview should be.
8185 Squid will not use Preview if the server did not request one.
8187 To disable ICAP Preview for all ICAP services, regardless of
8188 individual ICAP server OPTIONS responses, set this option to "off".
8190 icap_preview_enable off
8193 NAME: icap_preview_size
8196 LOC: Adaptation::Icap::TheConfig.preview_size
8198 DEFAULT_DOC: No preview sent.
8200 The default size of preview data to be sent to the ICAP server.
8201 This value might be overwritten on a per server basis by OPTIONS requests.
8204 NAME: icap_206_enable
8208 LOC: Adaptation::Icap::TheConfig.allow206_enable
8211 206 (Partial Content) responses is an ICAP extension that allows the
8212 ICAP agents to optionally combine adapted and original HTTP message
8213 content. The decision to combine is postponed until the end of the
8214 ICAP response. Squid supports Partial Content extension by default.
8216 Activation of the Partial Content extension is negotiated with each
8217 ICAP service during OPTIONS exchange. Most ICAP servers should handle
8218 negotation correctly even if they do not support the extension, but
8219 some might fail. To disable Partial Content support for all ICAP
8220 services and to avoid any negotiation, set this option to "off".
8226 NAME: icap_default_options_ttl
8229 LOC: Adaptation::Icap::TheConfig.default_options_ttl
8232 The default TTL value for ICAP OPTIONS responses that don't have
8233 an Options-TTL header.
8236 NAME: icap_persistent_connections
8240 LOC: Adaptation::Icap::TheConfig.reuse_connections
8243 Whether or not Squid should use persistent connections to
8247 NAME: adaptation_send_client_ip icap_send_client_ip
8249 IFDEF: USE_ADAPTATION
8251 LOC: Adaptation::Config::send_client_ip
8254 If enabled, Squid shares HTTP client IP information with adaptation
8255 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
8256 For eCAP, Squid sets the libecap::metaClientIp transaction option.
8258 See also: adaptation_uses_indirect_client
8261 NAME: adaptation_send_username icap_send_client_username
8263 IFDEF: USE_ADAPTATION
8265 LOC: Adaptation::Config::send_username
8268 This sends authenticated HTTP client username (if available) to
8269 the adaptation service.
8271 For ICAP, the username value is encoded based on the
8272 icap_client_username_encode option and is sent using the header
8273 specified by the icap_client_username_header option.
8276 NAME: icap_client_username_header
8279 LOC: Adaptation::Icap::TheConfig.client_username_header
8280 DEFAULT: X-Client-Username
8282 ICAP request header name to use for adaptation_send_username.
8285 NAME: icap_client_username_encode
8289 LOC: Adaptation::Icap::TheConfig.client_username_encode
8292 Whether to base64 encode the authenticated client username.
8296 TYPE: icap_service_type
8298 LOC: Adaptation::Icap::TheConfig
8301 Defines a single ICAP service using the following format:
8303 icap_service id vectoring_point uri [option ...]
8306 an opaque identifier or name which is used to direct traffic to
8307 this specific service. Must be unique among all adaptation
8308 services in squid.conf.
8310 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8311 This specifies at which point of transaction processing the
8312 ICAP service should be activated. *_postcache vectoring points
8313 are not yet supported.
8315 uri: icap://servername:port/servicepath
8316 ICAP server and service location.
8318 ICAP does not allow a single service to handle both REQMOD and RESPMOD
8319 transactions. Squid does not enforce that requirement. You can specify
8320 services with the same service_url and different vectoring_points. You
8321 can even specify multiple identical services as long as their
8322 service_names differ.
8324 To activate a service, use the adaptation_access directive. To group
8325 services, use adaptation_service_chain and adaptation_service_set.
8327 Service options are separated by white space. ICAP services support
8328 the following name=value options:
8331 If set to 'on' or '1', the ICAP service is treated as
8332 optional. If the service cannot be reached or malfunctions,
8333 Squid will try to ignore any errors and process the message as
8334 if the service was not enabled. No all ICAP errors can be
8335 bypassed. If set to 0, the ICAP service is treated as
8336 essential and all ICAP errors will result in an error page
8337 returned to the HTTP client.
8339 Bypass is off by default: services are treated as essential.
8342 If set to 'on' or '1', the ICAP service is allowed to
8343 dynamically change the current message adaptation plan by
8344 returning a chain of services to be used next. The services
8345 are specified using the X-Next-Services ICAP response header
8346 value, formatted as a comma-separated list of service names.
8347 Each named service should be configured in squid.conf. Other
8348 services are ignored. An empty X-Next-Services value results
8349 in an empty plan which ends the current adaptation.
8351 Dynamic adaptation plan may cross or cover multiple supported
8352 vectoring points in their natural processing order.
8354 Routing is not allowed by default: the ICAP X-Next-Services
8355 response header is ignored.
8358 Only has effect on split-stack systems. The default on those systems
8359 is to use IPv4-only connections. When set to 'on' this option will
8360 make Squid use IPv6-only connections to contact this ICAP service.
8362 on-overload=block|bypass|wait|force
8363 If the service Max-Connections limit has been reached, do
8364 one of the following for each new ICAP transaction:
8365 * block: send an HTTP error response to the client
8366 * bypass: ignore the "over-connected" ICAP service
8367 * wait: wait (in a FIFO queue) for an ICAP connection slot
8368 * force: proceed, ignoring the Max-Connections limit
8370 In SMP mode with N workers, each worker assumes the service
8371 connection limit is Max-Connections/N, even though not all
8372 workers may use a given service.
8374 The default value is "bypass" if service is bypassable,
8375 otherwise it is set to "wait".
8379 Use the given number as the Max-Connections limit, regardless
8380 of the Max-Connections value given by the service, if any.
8382 Older icap_service format without optional named parameters is
8383 deprecated but supported for backward compatibility.
8386 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8387 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
8391 TYPE: icap_class_type
8396 This deprecated option was documented to define an ICAP service
8397 chain, even though it actually defined a set of similar, redundant
8398 services, and the chains were not supported.
8400 To define a set of redundant services, please use the
8401 adaptation_service_set directive. For service chains, use
8402 adaptation_service_chain.
8406 TYPE: icap_access_type
8411 This option is deprecated. Please use adaptation_access, which
8412 has the same ICAP functionality, but comes with better
8413 documentation, and eCAP support.
8418 -----------------------------------------------------------------------------
8425 LOC: Adaptation::Ecap::TheConfig.onoff
8428 Controls whether eCAP support is enabled.
8432 TYPE: ecap_service_type
8434 LOC: Adaptation::Ecap::TheConfig
8437 Defines a single eCAP service
8439 ecap_service id vectoring_point uri [option ...]
8442 an opaque identifier or name which is used to direct traffic to
8443 this specific service. Must be unique among all adaptation
8444 services in squid.conf.
8446 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8447 This specifies at which point of transaction processing the
8448 eCAP service should be activated. *_postcache vectoring points
8449 are not yet supported.
8451 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8452 Squid uses the eCAP service URI to match this configuration
8453 line with one of the dynamically loaded services. Each loaded
8454 eCAP service must have a unique URI. Obtain the right URI from
8455 the service provider.
8457 To activate a service, use the adaptation_access directive. To group
8458 services, use adaptation_service_chain and adaptation_service_set.
8460 Service options are separated by white space. eCAP services support
8461 the following name=value options:
8464 If set to 'on' or '1', the eCAP service is treated as optional.
8465 If the service cannot be reached or malfunctions, Squid will try
8466 to ignore any errors and process the message as if the service
8467 was not enabled. No all eCAP errors can be bypassed.
8468 If set to 'off' or '0', the eCAP service is treated as essential
8469 and all eCAP errors will result in an error page returned to the
8472 Bypass is off by default: services are treated as essential.
8475 If set to 'on' or '1', the eCAP service is allowed to
8476 dynamically change the current message adaptation plan by
8477 returning a chain of services to be used next.
8479 Dynamic adaptation plan may cross or cover multiple supported
8480 vectoring points in their natural processing order.
8482 Routing is not allowed by default.
8484 Older ecap_service format without optional named parameters is
8485 deprecated but supported for backward compatibility.
8489 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8490 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8493 NAME: loadable_modules
8495 IFDEF: USE_LOADABLE_MODULES
8496 LOC: Config.loadable_module_names
8499 Instructs Squid to load the specified dynamic module(s) or activate
8500 preloaded module(s).
8502 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8506 MESSAGE ADAPTATION OPTIONS
8507 -----------------------------------------------------------------------------
8510 NAME: adaptation_service_set
8511 TYPE: adaptation_service_set_type
8512 IFDEF: USE_ADAPTATION
8517 Configures an ordered set of similar, redundant services. This is
8518 useful when hot standby or backup adaptation servers are available.
8520 adaptation_service_set set_name service_name1 service_name2 ...
8522 The named services are used in the set declaration order. The first
8523 applicable adaptation service from the set is used first. The next
8524 applicable service is tried if and only if the transaction with the
8525 previous service fails and the message waiting to be adapted is still
8528 When adaptation starts, broken services are ignored as if they were
8529 not a part of the set. A broken service is a down optional service.
8531 The services in a set must be attached to the same vectoring point
8532 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8534 If all services in a set are optional then adaptation failures are
8535 bypassable. If all services in the set are essential, then a
8536 transaction failure with one service may still be retried using
8537 another service from the set, but when all services fail, the master
8538 transaction fails as well.
8540 A set may contain a mix of optional and essential services, but that
8541 is likely to lead to surprising results because broken services become
8542 ignored (see above), making previously bypassable failures fatal.
8543 Technically, it is the bypassability of the last failed service that
8546 See also: adaptation_access adaptation_service_chain
8549 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8550 adaptation service_set svcLogger loggerLocal loggerRemote
8553 NAME: adaptation_service_chain
8554 TYPE: adaptation_service_chain_type
8555 IFDEF: USE_ADAPTATION
8560 Configures a list of complementary services that will be applied
8561 one-by-one, forming an adaptation chain or pipeline. This is useful
8562 when Squid must perform different adaptations on the same message.
8564 adaptation_service_chain chain_name service_name1 svc_name2 ...
8566 The named services are used in the chain declaration order. The first
8567 applicable adaptation service from the chain is used first. The next
8568 applicable service is applied to the successful adaptation results of
8569 the previous service in the chain.
8571 When adaptation starts, broken services are ignored as if they were
8572 not a part of the chain. A broken service is a down optional service.
8574 Request satisfaction terminates the adaptation chain because Squid
8575 does not currently allow declaration of RESPMOD services at the
8576 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8578 The services in a chain must be attached to the same vectoring point
8579 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8581 A chain may contain a mix of optional and essential services. If an
8582 essential adaptation fails (or the failure cannot be bypassed for
8583 other reasons), the master transaction fails. Otherwise, the failure
8584 is bypassed as if the failed adaptation service was not in the chain.
8586 See also: adaptation_access adaptation_service_set
8589 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8592 NAME: adaptation_access
8593 TYPE: adaptation_access_type
8594 IFDEF: USE_ADAPTATION
8597 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8599 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8601 adaptation_access service_name allow|deny [!]aclname...
8602 adaptation_access set_name allow|deny [!]aclname...
8604 At each supported vectoring point, the adaptation_access
8605 statements are processed in the order they appear in this
8606 configuration file. Statements pointing to the following services
8607 are ignored (i.e., skipped without checking their ACL):
8609 - services serving different vectoring points
8610 - "broken-but-bypassable" services
8611 - "up" services configured to ignore such transactions
8612 (e.g., based on the ICAP Transfer-Ignore header).
8614 When a set_name is used, all services in the set are checked
8615 using the same rules, to find the first applicable one. See
8616 adaptation_service_set for details.
8618 If an access list is checked and there is a match, the
8619 processing stops: For an "allow" rule, the corresponding
8620 adaptation service is used for the transaction. For a "deny"
8621 rule, no adaptation service is activated.
8623 It is currently not possible to apply more than one adaptation
8624 service at the same vectoring point to the same HTTP transaction.
8626 See also: icap_service and ecap_service
8629 adaptation_access service_1 allow all
8632 NAME: adaptation_service_iteration_limit
8634 IFDEF: USE_ADAPTATION
8635 LOC: Adaptation::Config::service_iteration_limit
8638 Limits the number of iterations allowed when applying adaptation
8639 services to a message. If your longest adaptation set or chain
8640 may have more than 16 services, increase the limit beyond its
8641 default value of 16. If detecting infinite iteration loops sooner
8642 is critical, make the iteration limit match the actual number
8643 of services in your longest adaptation set or chain.
8645 Infinite adaptation loops are most likely with routing services.
8647 See also: icap_service routing=1
8650 NAME: adaptation_masterx_shared_names
8652 IFDEF: USE_ADAPTATION
8653 LOC: Adaptation::Config::masterx_shared_name
8656 For each master transaction (i.e., the HTTP request and response
8657 sequence, including all related ICAP and eCAP exchanges), Squid
8658 maintains a table of metadata. The table entries are (name, value)
8659 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8660 with the master transaction.
8662 This option specifies the table entry names that Squid must accept
8663 from and forward to the adaptation transactions.
8665 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8666 shared table by returning an ICAP header field with a name
8667 specified in adaptation_masterx_shared_names.
8669 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8670 shared table by implementing the libecap::visitEachOption() API
8671 to provide an option with a name specified in
8672 adaptation_masterx_shared_names.
8674 Squid will store and forward the set entry to subsequent adaptation
8675 transactions within the same master transaction scope.
8677 Only one shared entry name is supported at this time.
8680 # share authentication information among ICAP services
8681 adaptation_masterx_shared_names X-Subscriber-ID
8684 NAME: adaptation_meta
8686 IFDEF: USE_ADAPTATION
8687 LOC: Adaptation::Config::metaHeaders
8690 This option allows Squid administrator to add custom ICAP request
8691 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8692 Use it to pass custom authentication tokens and other
8693 transaction-state related meta information to an ICAP/eCAP service.
8695 The addition of a meta header is ACL-driven:
8696 adaptation_meta name value [!]aclname ...
8698 Processing for a given header name stops after the first ACL list match.
8699 Thus, it is impossible to add two headers with the same name. If no ACL
8700 lists match for a given header name, no such header is added. For
8703 # do not debug transactions except for those that need debugging
8704 adaptation_meta X-Debug 1 needs_debugging
8706 # log all transactions except for those that must remain secret
8707 adaptation_meta X-Log 1 !keep_secret
8709 # mark transactions from users in the "G 1" group
8710 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8712 The "value" parameter may be a regular squid.conf token or a "double
8713 quoted string". Within the quoted string, use backslash (\) to escape
8714 any character, which is currently only useful for escaping backslashes
8715 and double quotes. For example,
8716 "this string has one backslash (\\) and two \"quotes\""
8718 Used adaptation_meta header values may be logged via %note
8719 logformat code. If multiple adaptation_meta headers with the same name
8720 are used during master transaction lifetime, the header values are
8721 logged in the order they were used and duplicate values are ignored
8722 (only the first repeated value will be logged).
8728 LOC: Adaptation::Icap::TheConfig.repeat
8729 DEFAULT_IF_NONE: deny all
8731 This ACL determines which retriable ICAP transactions are
8732 retried. Transactions that received a complete ICAP response
8733 and did not have to consume or produce HTTP bodies to receive
8734 that response are usually retriable.
8736 icap_retry allow|deny [!]aclname ...
8738 Squid automatically retries some ICAP I/O timeouts and errors
8739 due to persistent connection race conditions.
8741 See also: icap_retry_limit
8744 NAME: icap_retry_limit
8747 LOC: Adaptation::Icap::TheConfig.repeat_limit
8749 DEFAULT_DOC: No retries are allowed.
8751 Limits the number of retries allowed.
8753 Communication errors due to persistent connection race
8754 conditions are unavoidable, automatically retried, and do not
8755 count against this limit.
8757 See also: icap_retry
8763 -----------------------------------------------------------------------------
8766 NAME: check_hostnames
8769 LOC: Config.onoff.check_hostnames
8771 For security and stability reasons Squid can check
8772 hostnames for Internet standard RFC compliance. If you want
8773 Squid to perform these checks turn this directive on.
8776 NAME: allow_underscore
8779 LOC: Config.onoff.allow_underscore
8781 Underscore characters is not strictly allowed in Internet hostnames
8782 but nevertheless used by many sites. Set this to off if you want
8783 Squid to be strict about the standard.
8784 This check is performed only when check_hostnames is set to on.
8787 NAME: dns_retransmit_interval
8790 LOC: Config.Timeout.idns_retransmit
8792 Initial retransmit interval for DNS queries. The interval is
8793 doubled each time all configured DNS servers have been tried.
8799 LOC: Config.Timeout.idns_query
8801 DNS Query timeout. If no response is received to a DNS query
8802 within this time all DNS servers for the queried domain
8803 are assumed to be unavailable.
8806 NAME: dns_packet_max
8808 DEFAULT_DOC: EDNS disabled
8810 LOC: Config.dns.packet_max
8812 Maximum number of bytes packet size to advertise via EDNS.
8813 Set to "none" to disable EDNS large packet support.
8815 For legacy reasons DNS UDP replies will default to 512 bytes which
8816 is too small for many responses. EDNS provides a means for Squid to
8817 negotiate receiving larger responses back immediately without having
8818 to failover with repeat requests. Responses larger than this limit
8819 will retain the old behaviour of failover to TCP DNS.
8821 Squid has no real fixed limit internally, but allowing packet sizes
8822 over 1500 bytes requires network jumbogram support and is usually not
8825 WARNING: The RFC also indicates that some older resolvers will reply
8826 with failure of the whole request if the extension is added. Some
8827 resolvers have already been identified which will reply with mangled
8828 EDNS response on occasion. Usually in response to many-KB jumbogram
8829 sizes being advertised by Squid.
8830 Squid will currently treat these both as an unable-to-resolve domain
8831 even if it would be resolvable without EDNS.
8838 DEFAULT_DOC: Search for single-label domain names is disabled.
8839 LOC: Config.onoff.res_defnames
8841 Normally the RES_DEFNAMES resolver option is disabled
8842 (see res_init(3)). This prevents caches in a hierarchy
8843 from interpreting single-component hostnames locally. To allow
8844 Squid to handle single-component names, enable this option.
8847 NAME: dns_multicast_local
8851 DEFAULT_DOC: Search for .local and .arpa names is disabled.
8852 LOC: Config.onoff.dns_mdns
8854 When set to on, Squid sends multicast DNS lookups on the local
8855 network for domains ending in .local and .arpa.
8856 This enables local servers and devices to be contacted in an
8857 ad-hoc or zero-configuration network environment.
8860 NAME: dns_nameservers
8863 DEFAULT_DOC: Use operating system definitions
8864 LOC: Config.dns_nameservers
8866 Use this if you want to specify a list of DNS name servers
8867 (IP addresses) to use instead of those given in your
8868 /etc/resolv.conf file.
8870 On Windows platforms, if no value is specified here or in
8871 the /etc/resolv.conf file, the list of DNS name servers are
8872 taken from the Windows registry, both static and dynamic DHCP
8873 configurations are supported.
8875 Example: dns_nameservers 10.0.0.1 192.172.0.4
8880 DEFAULT: @DEFAULT_HOSTS@
8881 LOC: Config.etcHostsPath
8883 Location of the host-local IP name-address associations
8884 database. Most Operating Systems have such a file on different
8886 - Un*X & Linux: /etc/hosts
8887 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
8888 (%SystemRoot% value install default is c:\winnt)
8889 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
8890 (%SystemRoot% value install default is c:\windows)
8891 - Windows 9x/Me: %windir%\hosts
8892 (%windir% value is usually c:\windows)
8893 - Cygwin: /etc/hosts
8895 The file contains newline-separated definitions, in the
8896 form ip_address_in_dotted_form name [name ...] names are
8897 whitespace-separated. Lines beginning with an hash (#)
8898 character are comments.
8900 The file is checked at startup and upon configuration.
8901 If set to 'none', it won't be checked.
8902 If append_domain is used, that domain will be added to
8903 domain-local (i.e. not containing any dot character) host
8909 LOC: Config.appendDomain
8911 DEFAULT_DOC: Use operating system definitions
8913 Appends local domain name to hostnames without any dots in
8914 them. append_domain must begin with a period.
8916 Be warned there are now Internet names with no dots in
8917 them using only top-domain names, so setting this may
8918 cause some Internet sites to become unavailable.
8921 append_domain .yourdomain.com
8924 NAME: ignore_unknown_nameservers
8926 LOC: Config.onoff.ignore_unknown_nameservers
8929 By default Squid checks that DNS responses are received
8930 from the same IP addresses they are sent to. If they
8931 don't match, Squid ignores the response and writes a warning
8932 message to cache.log. You can allow responses from unknown
8933 nameservers by setting this option to 'off'.
8939 LOC: Config.dns.v4_first
8941 With the IPv6 Internet being as fast or faster than IPv4 Internet
8942 for most networks Squid prefers to contact websites over IPv6.
8944 This option reverses the order of preference to make Squid contact
8945 dual-stack websites over IPv4 first. Squid will still perform both
8946 IPv6 and IPv4 DNS lookups before connecting.
8949 This option will restrict the situations under which IPv6
8950 connectivity is used (and tested). Hiding network problems
8951 which would otherwise be detected and warned about.
8955 COMMENT: (number of entries)
8958 LOC: Config.ipcache.size
8960 Maximum number of DNS IP cache entries.
8967 LOC: Config.ipcache.low
8974 LOC: Config.ipcache.high
8976 The size, low-, and high-water marks for the IP cache.
8979 NAME: fqdncache_size
8980 COMMENT: (number of entries)
8983 LOC: Config.fqdncache.size
8985 Maximum number of FQDN cache entries.
8990 -----------------------------------------------------------------------------
8993 NAME: configuration_includes_quoted_values
8995 TYPE: configuration_includes_quoted_values
8997 LOC: ConfigParser::RecognizeQuotedValues
8999 If set, Squid will recognize each "quoted string" after a configuration
9000 directive as a single parameter. The quotes are stripped before the
9001 parameter value is interpreted or used.
9002 See "Values with spaces, quotes, and other special characters"
9003 section for more details.
9010 LOC: Config.onoff.mem_pools
9012 If set, Squid will keep pools of allocated (but unused) memory
9013 available for future use. If memory is a premium on your
9014 system and you believe your malloc library outperforms Squid
9015 routines, disable this.
9018 NAME: memory_pools_limit
9022 LOC: Config.MemPools.limit
9024 Used only with memory_pools on:
9025 memory_pools_limit 50 MB
9027 If set to a non-zero value, Squid will keep at most the specified
9028 limit of allocated (but unused) memory in memory pools. All free()
9029 requests that exceed this limit will be handled by your malloc
9030 library. Squid does not pre-allocate any memory, just safe-keeps
9031 objects that otherwise would be free()d. Thus, it is safe to set
9032 memory_pools_limit to a reasonably high value even if your
9033 configuration will use less memory.
9035 If set to none, Squid will keep all memory it can. That is, there
9036 will be no limit on the total amount of memory used for safe-keeping.
9038 To disable memory allocation optimization, do not set
9039 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
9041 An overhead for maintaining memory pools is not taken into account
9042 when the limit is checked. This overhead is close to four bytes per
9043 object kept. However, pools may actually _save_ memory because of
9044 reduced memory thrashing in your malloc library.
9048 COMMENT: on|off|transparent|truncate|delete
9051 LOC: opt_forwarded_for
9053 If set to "on", Squid will append your client's IP address
9054 in the HTTP requests it forwards. By default it looks like:
9056 X-Forwarded-For: 192.1.2.3
9058 If set to "off", it will appear as
9060 X-Forwarded-For: unknown
9062 If set to "transparent", Squid will not alter the
9063 X-Forwarded-For header in any way.
9065 If set to "delete", Squid will delete the entire
9066 X-Forwarded-For header.
9068 If set to "truncate", Squid will remove all existing
9069 X-Forwarded-For entries, and place the client IP as the sole entry.
9072 NAME: cachemgr_passwd
9073 TYPE: cachemgrpasswd
9075 DEFAULT_DOC: No password. Actions which require password are denied.
9076 LOC: Config.passwd_list
9078 Specify passwords for cachemgr operations.
9080 Usage: cachemgr_passwd password action action ...
9082 Some valid actions are (see cache manager menu for a full list):
9122 * Indicates actions which will not be performed without a
9123 valid password, others can be performed if not listed here.
9125 To disable an action, set the password to "disable".
9126 To allow performing an action without a password, set the
9129 Use the keyword "all" to set the same password for all actions.
9132 cachemgr_passwd secret shutdown
9133 cachemgr_passwd lesssssssecret info stats/objects
9134 cachemgr_passwd disable all
9141 LOC: Config.onoff.client_db
9143 If you want to disable collecting per-client statistics,
9144 turn off client_db here.
9147 NAME: refresh_all_ims
9151 LOC: Config.onoff.refresh_all_ims
9153 When you enable this option, squid will always check
9154 the origin server for an update when a client sends an
9155 If-Modified-Since request. Many browsers use IMS
9156 requests when the user requests a reload, and this
9157 ensures those clients receive the latest version.
9159 By default (off), squid may return a Not Modified response
9160 based on the age of the cached version.
9163 NAME: reload_into_ims
9164 IFDEF: USE_HTTP_VIOLATIONS
9168 LOC: Config.onoff.reload_into_ims
9170 When you enable this option, client no-cache or ``reload''
9171 requests will be changed to If-Modified-Since requests.
9172 Doing this VIOLATES the HTTP standard. Enabling this
9173 feature could make you liable for problems which it
9176 see also refresh_pattern for a more selective approach.
9179 NAME: connect_retries
9181 LOC: Config.connect_retries
9183 DEFAULT_DOC: Do not retry failed connections.
9185 This sets the maximum number of connection attempts made for each
9186 TCP connection. The connect_retries attempts must all still
9187 complete within the connection timeout period.
9189 The default is not to re-try if the first connection attempt fails.
9190 The (not recommended) maximum is 10 tries.
9192 A warning message will be generated if it is set to a too-high
9193 value and the configured value will be over-ridden.
9195 Note: These re-tries are in addition to forward_max_tries
9196 which limit how many different addresses may be tried to find
9200 NAME: retry_on_error
9202 LOC: Config.retry.onerror
9205 If set to ON Squid will automatically retry requests when
9206 receiving an error response with status 403 (Forbidden),
9207 500 (Internal Error), 501 or 503 (Service not available).
9208 Status 502 and 504 (Gateway errors) are always retried.
9210 This is mainly useful if you are in a complex cache hierarchy to
9211 work around access control errors.
9213 NOTE: This retry will attempt to find another working destination.
9214 Which is different from the server which just failed.
9217 NAME: as_whois_server
9219 LOC: Config.as_whois_server
9220 DEFAULT: whois.ra.net
9222 WHOIS server to query for AS numbers. NOTE: AS numbers are
9223 queried only when Squid starts up, not for every request.
9228 LOC: Config.onoff.offline
9231 Enable this option and Squid will never try to validate cached
9235 NAME: uri_whitespace
9236 TYPE: uri_whitespace
9237 LOC: Config.uri_whitespace
9240 What to do with requests that have whitespace characters in the
9243 strip: The whitespace characters are stripped out of the URL.
9244 This is the behavior recommended by RFC2396 and RFC3986
9245 for tolerant handling of generic URI.
9246 NOTE: This is one difference between generic URI and HTTP URLs.
9248 deny: The request is denied. The user receives an "Invalid
9250 This is the behaviour recommended by RFC2616 for safe
9251 handling of HTTP request URL.
9253 allow: The request is allowed and the URI is not changed. The
9254 whitespace characters remain in the URI. Note the
9255 whitespace is passed to redirector processes if they
9257 Note this may be considered a violation of RFC2616
9258 request parsing where whitespace is prohibited in the
9261 encode: The request is allowed and the whitespace characters are
9262 encoded according to RFC1738.
9264 chop: The request is allowed and the URI is chopped at the
9268 NOTE the current Squid implementation of encode and chop violates
9269 RFC2616 by not using a 301 redirect after altering the URL.
9274 LOC: Config.chroot_dir
9277 Specifies a directory where Squid should do a chroot() while
9278 initializing. This also causes Squid to fully drop root
9279 privileges after initializing. This means, for example, if you
9280 use a HTTP port less than 1024 and try to reconfigure, you may
9281 get an error saying that Squid can not open the port.
9284 NAME: balance_on_multiple_ip
9286 LOC: Config.onoff.balance_on_multiple_ip
9289 Modern IP resolvers in squid sort lookup results by preferred access.
9290 By default squid will use these IP in order and only rotates to
9291 the next listed when the most preffered fails.
9293 Some load balancing servers based on round robin DNS have been
9294 found not to preserve user session state across requests
9295 to different IP addresses.
9297 Enabling this directive Squid rotates IP's per request.
9300 NAME: pipeline_prefetch
9301 TYPE: pipelinePrefetch
9302 LOC: Config.pipeline_max_prefetch
9304 DEFAULT_DOC: Do not pre-parse pipelined requests.
9306 HTTP clients may send a pipeline of 1+N requests to Squid using a
9307 single connection, without waiting for Squid to respond to the first
9308 of those requests. This option limits the number of concurrent
9309 requests Squid will try to handle in parallel. If set to N, Squid
9310 will try to receive and process up to 1+N requests on the same
9311 connection concurrently.
9313 Defaults to 0 (off) for bandwidth management and access logging
9316 NOTE: pipelining requires persistent connections to clients.
9318 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
9321 NAME: high_response_time_warning
9324 LOC: Config.warnings.high_rptm
9326 DEFAULT_DOC: disabled.
9328 If the one-minute median response time exceeds this value,
9329 Squid prints a WARNING with debug level 0 to get the
9330 administrators attention. The value is in milliseconds.
9333 NAME: high_page_fault_warning
9335 LOC: Config.warnings.high_pf
9337 DEFAULT_DOC: disabled.
9339 If the one-minute average page fault rate exceeds this
9340 value, Squid prints a WARNING with debug level 0 to get
9341 the administrators attention. The value is in page faults
9345 NAME: high_memory_warning
9347 LOC: Config.warnings.high_memory
9348 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
9350 DEFAULT_DOC: disabled.
9352 If the memory usage (as determined by gnumalloc, if available and used)
9353 exceeds this amount, Squid prints a WARNING with debug level 0 to get
9354 the administrators attention.
9356 # TODO: link high_memory_warning to mempools?
9358 NAME: sleep_after_fork
9359 COMMENT: (microseconds)
9361 LOC: Config.sleep_after_fork
9364 When this is set to a non-zero value, the main Squid process
9365 sleeps the specified number of microseconds after a fork()
9366 system call. This sleep may help the situation where your
9367 system reports fork() failures due to lack of (virtual)
9368 memory. Note, however, if you have a lot of child
9369 processes, these sleep delays will add up and your
9370 Squid will not service requests for some amount of time
9371 until all the child processes have been started.
9372 On Windows value less then 1000 (1 milliseconds) are
9376 NAME: windows_ipaddrchangemonitor
9377 IFDEF: _SQUID_WINDOWS_
9381 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9383 On Windows Squid by default will monitor IP address changes and will
9384 reconfigure itself after any detected event. This is very useful for
9385 proxies connected to internet with dial-up interfaces.
9386 In some cases (a Proxy server acting as VPN gateway is one) it could be
9387 desiderable to disable this behaviour setting this to 'off'.
9388 Note: after changing this, Squid service must be restarted.
9393 IFDEF: USE_SQUID_EUI
9395 LOC: Eui::TheConfig.euiLookup
9397 Whether to lookup the EUI or MAC address of a connected client.
9400 NAME: max_filedescriptors max_filedesc
9403 DEFAULT_DOC: Use operating system limits set by ulimit.
9404 LOC: Config.max_filedescriptors
9406 Reduce the maximum number of filedescriptors supported below
9407 the usual operating system defaults.
9409 Remove from squid.conf to inherit the current ulimit setting.
9411 Note: Changing this requires a restart of Squid. Also
9412 not all I/O types supports large values (eg on Windows).
9419 DEFAULT_DOC: SMP support disabled.
9421 Number of main Squid processes or "workers" to fork and maintain.
9422 0: "no daemon" mode, like running "squid -N ..."
9423 1: "no SMP" mode, start one main Squid process daemon (default)
9424 N: start N main Squid process daemons (i.e., SMP mode)
9426 In SMP mode, each worker does nearly all what a single Squid daemon
9427 does (e.g., listen on http_port and forward HTTP requests).
9430 NAME: cpu_affinity_map
9431 TYPE: CpuAffinityMap
9432 LOC: Config.cpuAffinityMap
9434 DEFAULT_DOC: Let operating system decide.
9436 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
9438 Sets 1:1 mapping between Squid processes and CPU cores. For example,
9440 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
9442 affects processes 1 through 4 only and places them on the first
9443 four even cores, starting with core #1.
9445 CPU cores are numbered starting from 1. Requires support for
9446 sched_getaffinity(2) and sched_setaffinity(2) system calls.
9448 Multiple cpu_affinity_map options are merged.
9453 NAME: force_request_body_continuation
9455 LOC: Config.accessList.forceRequestBodyContinuation
9457 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
9459 This option controls how Squid handles data upload requests from HTTP
9460 and FTP agents that require a "Please Continue" control message response
9461 to actually send the request body to Squid. It is mostly useful in
9462 adaptation environments.
9464 When Squid receives an HTTP request with an "Expect: 100-continue"
9465 header or an FTP upload command (e.g., STOR), Squid normally sends the
9466 request headers or FTP command information to an adaptation service (or
9467 peer) and waits for a response. Most adaptation services (and some
9468 broken peers) may not respond to Squid at that stage because they may
9469 decide to wait for the HTTP request body or FTP data transfer. However,
9470 that request body or data transfer may never come because Squid has not
9471 responded with the HTTP 100 or FTP 150 (Please Continue) control message
9472 to the request sender yet!
9474 An allow match tells Squid to respond with the HTTP 100 or FTP 150
9475 (Please Continue) control message on its own, before forwarding the
9476 request to an adaptation service or peer. Such a response usually forces
9477 the request sender to proceed with sending the body. A deny match tells
9478 Squid to delay that control response until the origin server confirms
9479 that the request body is needed. Delaying is the default behavior.