2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the documentation for the Squid configuration file.
34 This documentation can also be found online at:
35 http://www.squid-cache.org/Doc/config/
37 You may wish to look at the Squid home page and wiki for the
38 FAQ and other documentation:
39 http://www.squid-cache.org/
40 http://wiki.squid-cache.org/SquidFaq
41 http://wiki.squid-cache.org/ConfigExamples
43 This documentation shows what the defaults for various directives
44 happen to be. If you don't need to change the default, you should
45 leave the line out of your squid.conf in most cases.
47 In some cases "none" refers to no default setting at all,
48 while in other cases it refers to the value of the option
49 - the comments for that keyword indicate if this is the case.
54 Configuration options can be included using the "include" directive.
55 Include takes a list of files to include. Quoting and wildcards are
60 include /path/to/included/file/squid.acl.config
62 Includes can be nested up to a hard-coded depth of 16 levels.
63 This arbitrary restriction is to prevent recursive include references
64 from causing Squid entering an infinite loop whilst trying to load
67 Values with byte units
69 Squid accepts size units on some size related directives. All
70 such directives are documented with a default value displaying
73 Units accepted by Squid are:
75 KB - Kilobyte (1024 bytes)
79 Values with spaces, quotes, and other special characters
81 Squid supports directive parameters with spaces, quotes, and other
82 special characters. Surround such parameters with "double quotes". Use
83 the configuration_includes_quoted_values directive to enable or
86 Squid supports reading configuration option parameters from external
87 files using the syntax:
88 parameters("/path/filename")
90 acl whitelist dstdomain parameters("/etc/squid/whitelist.txt")
92 Conditional configuration
94 If-statements can be used to make configuration directives
98 ... regular configuration directives ...
100 ... regular configuration directives ...]
103 The else part is optional. The keywords "if", "else", and "endif"
104 must be typed on their own lines, as if they were regular
105 configuration directives.
107 NOTE: An else-if condition is not supported.
109 These individual conditions types are supported:
112 Always evaluates to true.
114 Always evaluates to false.
115 <integer> = <integer>
116 Equality comparison of two integer numbers.
121 The following SMP-related preprocessor macros can be used.
123 ${process_name} expands to the current Squid process "name"
124 (e.g., squid1, squid2, or cache1).
126 ${process_number} expands to the current Squid process
127 identifier, which is an integer number (e.g., 1, 2, 3) unique
128 across all Squid processes of the current service instance.
130 ${service_name} expands into the current Squid service instance
131 name identifier which is provided by -n on the command line.
135 # options still not yet ported from 2.7 to 3.x
136 NAME: broken_vary_encoding
139 This option is not yet supported by Squid-3.
145 This option is not yet supported by Squid-3.
151 This option is not yet supported by Squid-3.
154 NAME: external_refresh_check
157 This option is not yet supported by Squid-3.
160 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
163 This option is not yet supported by Squid-3.
166 NAME: refresh_stale_hit
169 This option is not yet supported by Squid-3.
172 # Options Removed in 3.3
173 NAME: ignore_ims_on_miss
176 Remove this line. The HTTP/1.1 feature is now fully supported by default.
179 # Options Removed in 3.2
180 NAME: ignore_expect_100
183 Remove this line. The HTTP/1.1 feature is now fully supported by default.
186 NAME: dns_v4_fallback
189 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
195 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
198 NAME: maximum_single_addr_tries
201 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
207 Remove this line. The feature is supported by default in storage types where update is implemented.
210 NAME: url_rewrite_concurrency
213 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
216 # Options Removed in 3.1
220 Remove this line. DNS is no longer tested on startup.
223 NAME: extension_methods
226 Remove this line. All valid methods for HTTP are accepted by default.
229 # 2.7 Options Removed/Replaced in 3.2
234 # 2.7 Options Removed/Replaced in 3.1
242 Remove this line. HTTP/1.1 is supported by default.
245 NAME: upgrade_http0.9
248 Remove this line. ICY/1.0 streaming protocol is supported by default.
251 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
254 Alter these entries. Use the qos_flows directive instead.
257 # Options Removed in 3.0
261 Since squid-3.0 replace with request_header_access or reply_header_access
262 depending on whether you wish to match client requests or server replies.
265 NAME: httpd_accel_no_pmtu_disc
268 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
271 NAME: wais_relay_host
274 Replace this line with 'cache_peer' configuration.
277 NAME: wais_relay_port
280 Replace this line with 'cache_peer' configuration.
284 OPTIONS FOR AUTHENTICATION
285 -----------------------------------------------------------------------------
294 This is used to define parameters for the various authentication
295 schemes supported by Squid.
297 format: auth_param scheme parameter [setting]
299 The order in which authentication schemes are presented to the client is
300 dependent on the order the scheme first appears in config file. IE
301 has a bug (it's not RFC 2617 compliant) in that it will use the basic
302 scheme if basic is the first entry presented, even if more secure
303 schemes are presented. For now use the order in the recommended
304 settings section below. If other browsers have difficulties (don't
305 recognize the schemes offered even if you are using basic) either
306 put basic first, or disable the other schemes (by commenting out their
309 Once an authentication scheme is fully configured, it can only be
310 shutdown by shutting squid down and restarting. Changes can be made on
311 the fly and activated with a reconfigure. I.E. You can change to a
312 different helper, but not unconfigure the helper completely.
314 Please note that while this directive defines how Squid processes
315 authentication it does not automatically activate authentication.
316 To use authentication you must in addition make use of ACLs based
317 on login name in http_access (proxy_auth, proxy_auth_regex or
318 external with %LOGIN used in the format tag). The browser will be
319 challenged for authentication on the first such acl encountered
320 in http_access processing and will also be re-challenged for new
321 login credentials if the request is being denied by a proxy_auth
324 WARNING: authentication can't be used in a transparently intercepting
325 proxy as the client then thinks it is talking to an origin server and
326 not the proxy. This is a limitation of bending the TCP/IP protocol to
327 transparently intercepting port 80, not a limitation in Squid.
328 Ports flagged 'transparent', 'intercept', or 'tproxy' have
329 authentication disabled.
331 === Parameters common to all schemes. ===
334 Specifies the command for the external authenticator.
336 By default, each authentication scheme is not used unless a
337 program is specified.
339 See http://wiki.squid-cache.org/Features/AddonHelpers for
340 more details on helper operations and creating your own.
343 Specifies a string to be append to request line format for
344 the authentication helper. "Quoted" format values may contain
345 spaces and logformat %macros. In theory, any logformat %macro
346 can be used. In practice, a %macro expands as a dash (-) if
347 the helper request is sent before the required macro
348 information is available to Squid.
350 By default, Squid uses request formats provided in
351 scheme-specific examples below (search for %credentials).
353 The expanded key_extras value is added to the Squid credentials
354 cache and, hence, will affect authentication. It can be used to
355 autenticate different users with identical user names (e.g.,
356 when user authentication depends on http_port).
358 Avoid adding frequently changing information to key_extras. For
359 example, if you add user source IP, and it changes frequently
360 in your environment, then max_user_ip ACL is going to treat
361 every user+IP combination as a unique "user", breaking the ACL
362 and wasting a lot of memory on those user records. It will also
363 force users to authenticate from scratch whenever their IP
367 Specifies the protection scope (aka realm name) which is to be
368 reported to the client for the authentication scheme. It is
369 commonly part of the text the user will see when prompted for
370 their username and password.
372 For Basic the default is "Squid proxy-caching web server".
373 For Digest there is no default, this parameter is mandatory.
374 For NTLM and Negotiate this parameter is ignored.
376 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
378 The maximum number of authenticator processes to spawn. If
379 you start too few Squid will have to wait for them to process
380 a backlog of credential verifications, slowing it down. When
381 password verifications are done via a (slow) network you are
382 likely to need lots of authenticator processes.
384 The startup= and idle= options permit some skew in the exact
385 amount run. A minimum of startup=N will begin during startup
386 and reconfigure. Squid will start more in groups of up to
387 idle=N in an attempt to meet traffic needs and to keep idle=N
388 free above those traffic needs up to the maximum.
390 The concurrency= option sets the number of concurrent requests
391 the helper can process. The default of 0 is used for helpers
392 who only supports one request at a time. Setting this to a
393 number greater than 0 changes the protocol used to include a
394 channel ID field first on the request/response line, allowing
395 multiple requests to be sent to the same helper in parallel
396 without waiting for the response.
398 Concurrency must not be set unless it's known the helper
399 supports the input format with channel-ID fields.
401 NOTE: NTLM and Negotiate schemes do not support concurrency
402 in the Squid code module even though some helpers can.
405 IF HAVE_AUTH_MODULE_BASIC
406 === Basic authentication parameters ===
409 HTTP uses iso-latin-1 as character set, while some
410 authentication backends such as LDAP expects UTF-8. If this is
411 set to on Squid will translate the HTTP iso-latin-1 charset to
412 UTF-8 before sending the username and password to the helper.
414 "credentialsttl" timetolive
415 Specifies how long squid assumes an externally validated
416 username:password pair is valid for - in other words how
417 often the helper program is called for that user. Set this
418 low to force revalidation with short lived passwords.
420 NOTE: setting this high does not impact your susceptibility
421 to replay attacks unless you are using an one-time password
422 system (such as SecureID). If you are using such a system,
423 you will be vulnerable to replay attacks unless you also
424 use the max_user_ip ACL in an http_access rule.
426 "casesensitive" on|off
427 Specifies if usernames are case sensitive. Most user databases
428 are case insensitive allowing the same username to be spelled
429 using both lower and upper case letters, but some are case
430 sensitive. This makes a big difference for user_max_ip ACL
431 processing and similar.
434 IF HAVE_AUTH_MODULE_DIGEST
435 === Digest authentication parameters ===
438 HTTP uses iso-latin-1 as character set, while some
439 authentication backends such as LDAP expects UTF-8. If this is
440 set to on Squid will translate the HTTP iso-latin-1 charset to
441 UTF-8 before sending the username and password to the helper.
443 "nonce_garbage_interval" timeinterval
444 Specifies the interval that nonces that have been issued
445 to client_agent's are checked for validity.
447 "nonce_max_duration" timeinterval
448 Specifies the maximum length of time a given nonce will be
451 "nonce_max_count" number
452 Specifies the maximum number of times a given nonce can be
455 "nonce_strictness" on|off
456 Determines if squid requires strict increment-by-1 behavior
457 for nonce counts, or just incrementing (off - for use when
458 user agents generate nonce counts that occasionally miss 1
459 (ie, 1,2,4,6)). Default off.
461 "check_nonce_count" on|off
462 This directive if set to off can disable the nonce count check
463 completely to work around buggy digest qop implementations in
464 certain mainstream browser versions. Default on to check the
465 nonce count to protect from authentication replay attacks.
467 "post_workaround" on|off
468 This is a workaround to certain buggy browsers who send an
469 incorrect request digest in POST requests when reusing the
470 same nonce as acquired earlier on a GET request.
473 IF HAVE_AUTH_MODULE_NEGOTIATE
474 === Negotiate authentication parameters ===
477 If you experience problems with PUT/POST requests when using
478 the this authentication scheme then you can try setting this
479 to off. This will cause Squid to forcibly close the connection
480 on the initial request where the browser asks which schemes
481 are supported by the proxy.
484 IF HAVE_AUTH_MODULE_NTLM
485 === NTLM authentication parameters ===
488 If you experience problems with PUT/POST requests when using
489 the this authentication scheme then you can try setting this
490 to off. This will cause Squid to forcibly close the connection
491 on the initial request where the browser asks which schemes
492 are supported by the proxy.
495 === Example Configuration ===
497 This configuration displays the recommended authentication scheme
498 order from most to least secure with recommended minimum configuration
499 settings for each scheme:
501 #auth_param negotiate program <uncomment and complete this line to activate>
502 #auth_param negotiate children 20 startup=0 idle=1
503 #auth_param negotiate keep_alive on
505 #auth_param digest program <uncomment and complete this line to activate>
506 #auth_param digest children 20 startup=0 idle=1
507 #auth_param digest realm Squid proxy-caching web server
508 #auth_param digest nonce_garbage_interval 5 minutes
509 #auth_param digest nonce_max_duration 30 minutes
510 #auth_param digest nonce_max_count 50
512 #auth_param ntlm program <uncomment and complete this line to activate>
513 #auth_param ntlm children 20 startup=0 idle=1
514 #auth_param ntlm keep_alive on
516 #auth_param basic program <uncomment and complete this line>
517 #auth_param basic children 5 startup=5 idle=1
518 #auth_param basic realm Squid proxy-caching web server
519 #auth_param basic credentialsttl 2 hours
522 NAME: authenticate_cache_garbage_interval
525 LOC: Config.authenticateGCInterval
527 The time period between garbage collection across the username cache.
528 This is a trade-off between memory utilization (long intervals - say
529 2 days) and CPU (short intervals - say 1 minute). Only change if you
533 NAME: authenticate_ttl
536 LOC: Config.authenticateTTL
538 The time a user & their credentials stay in the logged in
539 user cache since their last request. When the garbage
540 interval passes, all user credentials that have passed their
541 TTL are removed from memory.
544 NAME: authenticate_ip_ttl
546 LOC: Config.authenticateIpTTL
549 If you use proxy authentication and the 'max_user_ip' ACL,
550 this directive controls how long Squid remembers the IP
551 addresses associated with each user. Use a small value
552 (e.g., 60 seconds) if your users might change addresses
553 quickly, as is the case with dialup. You might be safe
554 using a larger value (e.g., 2 hours) in a corporate LAN
555 environment with relatively static address assignments.
560 -----------------------------------------------------------------------------
563 NAME: external_acl_type
564 TYPE: externalAclHelper
565 LOC: Config.externalAclHelperList
568 This option defines external acl classes using a helper program
569 to look up the status
571 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
575 ttl=n TTL in seconds for cached results (defaults to 3600
578 TTL for cached negative lookups (default same
581 Maximum number of acl helper processes spawned to service
582 external acl lookups of this type. (default 20)
584 Minimum number of acl helper processes to spawn during
585 startup and reconfigure to service external acl lookups
586 of this type. (default 0)
588 Number of acl helper processes to keep ahead of traffic
589 loads. Squid will spawn this many at once whenever load
590 rises above the capabilities of existing processes.
591 Up to the value of children-max. (default 1)
592 concurrency=n concurrency level per process. Only used with helpers
593 capable of processing more than one query at a time.
594 cache=n limit the result cache size, default is unbounded.
595 grace=n Percentage remaining of TTL where a refresh of a
596 cached entry should be initiated without needing to
597 wait for a new reply. (default is for no grace period)
598 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
599 ipv4 / ipv6 IP protocol used to communicate with this helper.
600 The default is to auto-detect IPv6 and use it when available.
602 FORMAT specifications
604 %LOGIN Authenticated user login name
605 %EXT_USER Username from previous external acl
606 %EXT_LOG Log details from previous external acl
607 %EXT_TAG Tag from previous external acl
608 %IDENT Ident user name
610 %SRCPORT Client source port
613 %PROTO Requested URL scheme
615 %PATH Requested URL path
616 %METHOD Request method
617 %MYADDR Squid interface address
618 %MYPORT Squid http_port number
619 %PATH Requested URL-path (including query-string if any)
620 %USER_CERT SSL User certificate in PEM format
621 %USER_CERTCHAIN SSL User certificate chain in PEM format
622 %USER_CERT_xx SSL User certificate subject attribute xx
623 %USER_CA_xx SSL User certificate issuer attribute xx
625 %>{Header} HTTP request header "Header"
627 HTTP request header "Hdr" list member "member"
629 HTTP request header list member using ; as
630 list separator. ; can be any non-alphanumeric
633 %<{Header} HTTP reply header "Header"
635 HTTP reply header "Hdr" list member "member"
637 HTTP reply header list member using ; as
638 list separator. ; can be any non-alphanumeric
641 %ACL The name of the ACL being tested.
642 %DATA The ACL arguments. If not used then any arguments
643 is automatically added at the end of the line
645 NOTE: this will encode the arguments as one token,
646 whereas the default will pass each separately.
648 %% The percent sign. Useful for helpers which need
649 an unchanging input format.
652 General request syntax:
654 [channel-ID] FORMAT-values [acl-values ...]
657 FORMAT-values consists of transaction details expanded with
658 whitespace separation per the config file FORMAT specification
659 using the FORMAT macros listed above.
661 acl-values consists of any string specified in the referencing
662 config 'acl ... external' line. see the "acl external" directive.
664 Request values sent to the helper are URL escaped to protect
665 each value in requests against whitespaces.
667 If using protocol=2.5 then the request sent to the helper is not
668 URL escaped to protect against whitespace.
670 NOTE: protocol=3.0 is deprecated as no longer necessary.
672 When using the concurrency= option the protocol is changed by
673 introducing a query channel tag in front of the request/response.
674 The query channel tag is a number between 0 and concurrency-1.
675 This value must be echoed back unchanged to Squid as the first part
676 of the response relating to its request.
679 The helper receives lines expanded per the above format specification
680 and for each input line returns 1 line starting with OK/ERR/BH result
681 code and optionally followed by additional keywords with more details.
684 General result syntax:
686 [channel-ID] result keyword=value ...
688 Result consists of one of the codes:
691 the ACL test produced a match.
694 the ACL test does not produce a match.
697 An internal error occurred in the helper, preventing
698 a result being identified.
700 The meaning of 'a match' is determined by your squid.conf
701 access control configuration. See the Squid wiki for details.
705 user= The users name (login)
707 password= The users password (for login= cache_peer option)
709 message= Message describing the reason for this response.
710 Available as %o in error pages.
711 Useful on (ERR and BH results).
713 tag= Apply a tag to a request. Only sets a tag once,
714 does not alter existing tags.
716 log= String to be logged in access.log. Available as
717 %ea in logformat specifications.
719 Any keywords may be sent on any response whether OK, ERR or BH.
721 All response keyword values need to be a single token with URL
722 escaping, or enclosed in double quotes (") and escaped using \ on
723 any double quotes or \ characters within the value. The wrapping
724 double quotes are removed before the value is interpreted by Squid.
725 \r and \n are also replace by CR and LF.
727 Some example key values:
731 user="J. \"Bob\" Smith"
738 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
739 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
740 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
741 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
742 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
745 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
746 DEFAULT: localhost src 127.0.0.1/32 ::1
747 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
748 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
750 Defining an Access List
752 Every access list definition must begin with an aclname and acltype,
753 followed by either type-specific arguments or a quoted filename that
756 acl aclname acltype argument ...
757 acl aclname acltype "file" ...
759 When using "file", the file should contain one item per line.
761 Some acl types supports options which changes their default behaviour.
762 The available options are:
764 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
765 case-insensitive, use the -i option. To return case-sensitive
766 use the +i option between patterns, or make a new ACL line
769 -n Disable lookups and address type conversions. If lookup or
770 conversion is required because the parameter type (IP or
771 domain name) does not match the message address type (domain
772 name or IP), then the ACL would immediately declare a mismatch
773 without any warnings or lookups.
775 -- Used to stop processing all options, in the case the first acl
776 value has '-' character as first character (for example the '-'
777 is a valid domain name)
779 Some acl types require suspending the current request in order
780 to access some external data source.
781 Those which do are marked with the tag [slow], those which
782 don't are marked as [fast].
783 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
784 for further information
786 ***** ACL TYPES AVAILABLE *****
788 acl aclname src ip-address/mask ... # clients IP address [fast]
789 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
790 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
791 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
793 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
794 # The arp ACL requires the special configure option --enable-arp-acl.
795 # Furthermore, the ARP ACL code is not portable to all operating systems.
796 # It works on Linux, Solaris, Windows, FreeBSD, and some
797 # other *BSD variants.
800 # NOTE: Squid can only determine the MAC address for clients that are on
801 # the same subnet. If the client is on a different subnet,
802 # then Squid cannot find out its MAC address.
804 acl aclname srcdomain .foo.com ...
805 # reverse lookup, from client IP [slow]
806 acl aclname dstdomain [-n] .foo.com ...
807 # Destination server from URL [fast]
808 acl aclname srcdom_regex [-i] \.foo\.com ...
809 # regex matching client name [slow]
810 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
811 # regex matching server [fast]
813 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
814 # based URL is used and no match is found. The name "none" is used
815 # if the reverse lookup fails.
817 acl aclname src_as number ...
818 acl aclname dst_as number ...
820 # Except for access control, AS numbers can be used for
821 # routing of requests to specific caches. Here's an
822 # example for routing all requests for AS#1241 and only
823 # those to mycache.mydomain.net:
824 # acl asexample dst_as 1241
825 # cache_peer_access mycache.mydomain.net allow asexample
826 # cache_peer_access mycache_mydomain.net deny all
828 acl aclname peername myPeer ...
830 # match against a named cache_peer entry
831 # set unique name= on cache_peer lines for reliable use.
833 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
843 # h1:m1 must be less than h2:m2
845 acl aclname url_regex [-i] ^http:// ...
846 # regex matching on whole URL [fast]
847 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
848 # regex matching on URL login field
849 acl aclname urlpath_regex [-i] \.gif$ ...
850 # regex matching on URL path [fast]
852 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
854 acl aclname localport 3128 ... # TCP port the client connected to [fast]
855 # NP: for interception mode this is usually '80'
857 acl aclname myportname 3128 ... # http(s)_port name [fast]
859 acl aclname proto HTTP FTP ... # request protocol [fast]
861 acl aclname method GET POST ... # HTTP request method [fast]
863 acl aclname http_status 200 301 500- 400-403 ...
864 # status code in reply [fast]
866 acl aclname browser [-i] regexp ...
867 # pattern match on User-Agent header (see also req_header below) [fast]
869 acl aclname referer_regex [-i] regexp ...
870 # pattern match on Referer header [fast]
871 # Referer is highly unreliable, so use with care
873 acl aclname ident username ...
874 acl aclname ident_regex [-i] pattern ...
875 # string match on ident output [slow]
876 # use REQUIRED to accept any non-null ident.
878 acl aclname proxy_auth [-i] username ...
879 acl aclname proxy_auth_regex [-i] pattern ...
880 # perform http authentication challenge to the client and match against
881 # supplied credentials [slow]
883 # takes a list of allowed usernames.
884 # use REQUIRED to accept any valid username.
886 # Will use proxy authentication in forward-proxy scenarios, and plain
887 # http authenticaiton in reverse-proxy scenarios
889 # NOTE: when a Proxy-Authentication header is sent but it is not
890 # needed during ACL checking the username is NOT logged
893 # NOTE: proxy_auth requires a EXTERNAL authentication program
894 # to check username/password combinations (see
895 # auth_param directive).
897 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
898 # as the browser needs to be configured for using a proxy in order
899 # to respond to proxy authentication.
901 acl aclname snmp_community string ...
902 # A community string to limit access to your SNMP Agent [fast]
905 # acl snmppublic snmp_community public
907 acl aclname maxconn number
908 # This will be matched when the client's IP address has
909 # more than <number> TCP connections established. [fast]
910 # NOTE: This only measures direct TCP links so X-Forwarded-For
911 # indirect clients are not counted.
913 acl aclname max_user_ip [-s] number
914 # This will be matched when the user attempts to log in from more
915 # than <number> different ip addresses. The authenticate_ip_ttl
916 # parameter controls the timeout on the ip entries. [fast]
917 # If -s is specified the limit is strict, denying browsing
918 # from any further IP addresses until the ttl has expired. Without
919 # -s Squid will just annoy the user by "randomly" denying requests.
920 # (the counter is reset each time the limit is reached and a
922 # NOTE: in acceleration mode or where there is mesh of child proxies,
923 # clients may appear to come from multiple addresses if they are
924 # going through proxy farms, so a limit of 1 may cause user problems.
926 acl aclname random probability
927 # Pseudo-randomly match requests. Based on the probability given.
928 # Probability may be written as a decimal (0.333), fraction (1/3)
929 # or ratio of matches:non-matches (3:5).
931 acl aclname req_mime_type [-i] mime-type ...
932 # regex match against the mime type of the request generated
933 # by the client. Can be used to detect file upload or some
934 # types HTTP tunneling requests [fast]
935 # NOTE: This does NOT match the reply. You cannot use this
936 # to match the returned file type.
938 acl aclname req_header header-name [-i] any\.regex\.here
939 # regex match against any of the known request headers. May be
940 # thought of as a superset of "browser", "referer" and "mime-type"
943 acl aclname rep_mime_type [-i] mime-type ...
944 # regex match against the mime type of the reply received by
945 # squid. Can be used to detect file download or some
946 # types HTTP tunneling requests. [fast]
947 # NOTE: This has no effect in http_access rules. It only has
948 # effect in rules that affect the reply data stream such as
951 acl aclname rep_header header-name [-i] any\.regex\.here
952 # regex match against any of the known reply headers. May be
953 # thought of as a superset of "browser", "referer" and "mime-type"
956 acl aclname external class_name [arguments...]
957 # external ACL lookup via a helper class defined by the
958 # external_acl_type directive [slow]
960 acl aclname user_cert attribute values...
961 # match against attributes in a user SSL certificate
962 # attribute is one of DN/C/O/CN/L/ST [fast]
964 acl aclname ca_cert attribute values...
965 # match against attributes a users issuing CA SSL certificate
966 # attribute is one of DN/C/O/CN/L/ST [fast]
968 acl aclname ext_user username ...
969 acl aclname ext_user_regex [-i] pattern ...
970 # string match on username returned by external acl helper [slow]
971 # use REQUIRED to accept any non-null user name.
973 acl aclname tag tagvalue ...
974 # string match on tag returned by external acl helper [slow]
976 acl aclname hier_code codename ...
977 # string match against squid hierarchy code(s); [fast]
978 # e.g., DIRECT, PARENT_HIT, NONE, etc.
980 # NOTE: This has no effect in http_access rules. It only has
981 # effect in rules that affect the reply data stream such as
984 acl aclname note name [value ...]
985 # match transaction annotation [fast]
986 # Without values, matches any annotation with a given name.
987 # With value(s), matches any annotation with a given name that
988 # also has one of the given values.
989 # Names and values are compared using a string equality test.
990 # Annotation sources include note and adaptation_meta directives
991 # as well as helper and eCAP responses.
993 acl aclname adaptation_service service ...
994 # Matches the name of any icap_service, ecap_service,
995 # adaptation_service_set, or adaptation_service_chain that Squid
996 # has used (or attempted to use) for the master transaction.
997 # This ACL must be defined after the corresponding adaptation
998 # service is named in squid.conf. This ACL is usable with
999 # adaptation_meta because it starts matching immediately after
1000 # the service has been selected for adaptation.
1003 acl aclname ssl_error errorname
1004 # match against SSL certificate validation error [fast]
1006 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1009 # The following can be used as shortcuts for certificate properties:
1010 # [ssl::]certHasExpired: the "not after" field is in the past
1011 # [ssl::]certNotYetValid: the "not before" field is in the future
1012 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1013 # [ssl::]certSelfSigned: The certificate is self signed.
1014 # [ssl::]certDomainMismatch: The certificate CN domain does not
1015 # match the name the name of the host we are connecting to.
1017 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1018 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1019 # predefined ACLs, just like the 'all' ACL.
1021 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1022 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1024 acl aclname server_cert_fingerprint [-sha1] fingerprint
1025 # match against server SSL certificate fingerprint [fast]
1027 # The fingerprint is the digest of the DER encoded version
1028 # of the whole certificate. The user should use the form: XX:XX:...
1029 # Optional argument specifies the digest algorithm to use.
1030 # The SHA1 digest algorithm is the default and is currently
1031 # the only algorithm supported (-sha1).
1033 acl aclname any-of acl1 acl2 ...
1034 # match any one of the acls [fast or slow]
1035 # The first matching ACL stops further ACL evaluation.
1037 # ACLs from multiple any-of lines with the same name are ORed.
1038 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1039 # acl A any-of a1 a2
1040 # acl A any-of a3 a4
1042 # This group ACL is fast if all evaluated ACLs in the group are fast
1043 # and slow otherwise.
1045 acl aclname all-of acl1 acl2 ...
1046 # match all of the acls [fast or slow]
1047 # The first mismatching ACL stops further ACL evaluation.
1049 # ACLs from multiple all-of lines with the same name are ORed.
1050 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1051 # acl B all-of b1 b2
1052 # acl B all-of b3 b4
1054 # This group ACL is fast if all evaluated ACLs in the group are fast
1055 # and slow otherwise.
1058 acl macaddress arp 09:00:2b:23:45:67
1059 acl myexample dst_as 1241
1060 acl password proxy_auth REQUIRED
1061 acl fileupload req_mime_type -i ^multipart/form-data$
1062 acl javascript rep_mime_type -i ^application/x-javascript$
1066 # Recommended minimum configuration:
1069 # Example rule allowing access from your local networks.
1070 # Adapt to list your (internal) IP networks from where browsing
1072 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1073 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1074 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1075 acl localnet src fc00::/7 # RFC 4193 local private network range
1076 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1078 acl SSL_ports port 443
1079 acl Safe_ports port 80 # http
1080 acl Safe_ports port 21 # ftp
1081 acl Safe_ports port 443 # https
1082 acl Safe_ports port 70 # gopher
1083 acl Safe_ports port 210 # wais
1084 acl Safe_ports port 1025-65535 # unregistered ports
1085 acl Safe_ports port 280 # http-mgmt
1086 acl Safe_ports port 488 # gss-http
1087 acl Safe_ports port 591 # filemaker
1088 acl Safe_ports port 777 # multiling http
1089 acl CONNECT method CONNECT
1093 NAME: follow_x_forwarded_for
1095 IFDEF: FOLLOW_X_FORWARDED_FOR
1096 LOC: Config.accessList.followXFF
1097 DEFAULT_IF_NONE: deny all
1098 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1100 Allowing or Denying the X-Forwarded-For header to be followed to
1101 find the original source of a request.
1103 Requests may pass through a chain of several other proxies
1104 before reaching us. The X-Forwarded-For header will contain a
1105 comma-separated list of the IP addresses in the chain, with the
1106 rightmost address being the most recent.
1108 If a request reaches us from a source that is allowed by this
1109 configuration item, then we consult the X-Forwarded-For header
1110 to see where that host received the request from. If the
1111 X-Forwarded-For header contains multiple addresses, we continue
1112 backtracking until we reach an address for which we are not allowed
1113 to follow the X-Forwarded-For header, or until we reach the first
1114 address in the list. For the purpose of ACL used in the
1115 follow_x_forwarded_for directive the src ACL type always matches
1116 the address we are testing and srcdomain matches its rDNS.
1118 The end result of this process is an IP address that we will
1119 refer to as the indirect client address. This address may
1120 be treated as the client address for access control, ICAP, delay
1121 pools and logging, depending on the acl_uses_indirect_client,
1122 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1123 log_uses_indirect_client and tproxy_uses_indirect_client options.
1125 This clause only supports fast acl types.
1126 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1128 SECURITY CONSIDERATIONS:
1130 Any host for which we follow the X-Forwarded-For header
1131 can place incorrect information in the header, and Squid
1132 will use the incorrect information as if it were the
1133 source address of the request. This may enable remote
1134 hosts to bypass any access control restrictions that are
1135 based on the client's source addresses.
1139 acl localhost src 127.0.0.1
1140 acl my_other_proxy srcdomain .proxy.example.com
1141 follow_x_forwarded_for allow localhost
1142 follow_x_forwarded_for allow my_other_proxy
1145 NAME: acl_uses_indirect_client
1148 IFDEF: FOLLOW_X_FORWARDED_FOR
1150 LOC: Config.onoff.acl_uses_indirect_client
1152 Controls whether the indirect client address
1153 (see follow_x_forwarded_for) is used instead of the
1154 direct client address in acl matching.
1156 NOTE: maxconn ACL considers direct TCP links and indirect
1157 clients will always have zero. So no match.
1160 NAME: delay_pool_uses_indirect_client
1163 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1165 LOC: Config.onoff.delay_pool_uses_indirect_client
1167 Controls whether the indirect client address
1168 (see follow_x_forwarded_for) is used instead of the
1169 direct client address in delay pools.
1172 NAME: log_uses_indirect_client
1175 IFDEF: FOLLOW_X_FORWARDED_FOR
1177 LOC: Config.onoff.log_uses_indirect_client
1179 Controls whether the indirect client address
1180 (see follow_x_forwarded_for) is used instead of the
1181 direct client address in the access log.
1184 NAME: tproxy_uses_indirect_client
1187 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1189 LOC: Config.onoff.tproxy_uses_indirect_client
1191 Controls whether the indirect client address
1192 (see follow_x_forwarded_for) is used instead of the
1193 direct client address when spoofing the outgoing client.
1195 This has no effect on requests arriving in non-tproxy
1198 SECURITY WARNING: Usage of this option is dangerous
1199 and should not be used trivially. Correct configuration
1200 of follow_x_forewarded_for with a limited set of trusted
1201 sources is required to prevent abuse of your proxy.
1204 NAME: spoof_client_ip
1206 LOC: Config.accessList.spoof_client_ip
1208 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1210 Control client IP address spoofing of TPROXY traffic based on
1211 defined access lists.
1213 spoof_client_ip allow|deny [!]aclname ...
1215 If there are no "spoof_client_ip" lines present, the default
1216 is to "allow" spoofing of any suitable request.
1218 Note that the cache_peer "no-tproxy" option overrides this ACL.
1220 This clause supports fast acl types.
1221 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1226 LOC: Config.accessList.http
1227 DEFAULT_IF_NONE: deny all
1228 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1230 Allowing or Denying access based on defined access lists
1232 Access to the HTTP port:
1233 http_access allow|deny [!]aclname ...
1235 NOTE on default values:
1237 If there are no "access" lines present, the default is to deny
1240 If none of the "access" lines cause a match, the default is the
1241 opposite of the last line in the list. If the last line was
1242 deny, the default is allow. Conversely, if the last line
1243 is allow, the default will be deny. For these reasons, it is a
1244 good idea to have an "deny all" entry at the end of your access
1245 lists to avoid potential confusion.
1247 This clause supports both fast and slow acl types.
1248 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1253 # Recommended minimum Access Permission configuration:
1255 # Deny requests to certain unsafe ports
1256 http_access deny !Safe_ports
1258 # Deny CONNECT to other than secure SSL ports
1259 http_access deny CONNECT !SSL_ports
1261 # Only allow cachemgr access from localhost
1262 http_access allow localhost manager
1263 http_access deny manager
1265 # We strongly recommend the following be uncommented to protect innocent
1266 # web applications running on the proxy server who think the only
1267 # one who can access services on "localhost" is a local user
1268 #http_access deny to_localhost
1271 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1274 # Example rule allowing access from your local networks.
1275 # Adapt localnet in the ACL section to list your (internal) IP networks
1276 # from where browsing should be allowed
1277 http_access allow localnet
1278 http_access allow localhost
1280 # And finally deny all other access to this proxy
1281 http_access deny all
1285 NAME: adapted_http_access http_access2
1287 LOC: Config.accessList.adapted_http
1289 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1291 Allowing or Denying access based on defined access lists
1293 Essentially identical to http_access, but runs after redirectors
1294 and ICAP/eCAP adaptation. Allowing access control based on their
1297 If not set then only http_access is used.
1300 NAME: http_reply_access
1302 LOC: Config.accessList.reply
1304 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1306 Allow replies to client requests. This is complementary to http_access.
1308 http_reply_access allow|deny [!] aclname ...
1310 NOTE: if there are no access lines present, the default is to allow
1313 If none of the access lines cause a match the opposite of the
1314 last line will apply. Thus it is good practice to end the rules
1315 with an "allow all" or "deny all" entry.
1317 This clause supports both fast and slow acl types.
1318 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1323 LOC: Config.accessList.icp
1325 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1327 Allowing or Denying access to the ICP port based on defined
1330 icp_access allow|deny [!]aclname ...
1332 NOTE: The default if no icp_access lines are present is to
1333 deny all traffic. This default may cause problems with peers
1336 This clause only supports fast acl types.
1337 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1339 # Allow ICP queries from local networks only
1340 #icp_access allow localnet
1341 #icp_access deny all
1347 LOC: Config.accessList.htcp
1349 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1351 Allowing or Denying access to the HTCP port based on defined
1354 htcp_access allow|deny [!]aclname ...
1356 See also htcp_clr_access for details on access control for
1357 cache purge (CLR) HTCP messages.
1359 NOTE: The default if no htcp_access lines are present is to
1360 deny all traffic. This default may cause problems with peers
1361 using the htcp option.
1363 This clause only supports fast acl types.
1364 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1366 # Allow HTCP queries from local networks only
1367 #htcp_access allow localnet
1368 #htcp_access deny all
1371 NAME: htcp_clr_access
1374 LOC: Config.accessList.htcp_clr
1376 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1378 Allowing or Denying access to purge content using HTCP based
1379 on defined access lists.
1380 See htcp_access for details on general HTCP access control.
1382 htcp_clr_access allow|deny [!]aclname ...
1384 This clause only supports fast acl types.
1385 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1387 # Allow HTCP CLR requests from trusted peers
1388 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1389 htcp_clr_access allow htcp_clr_peer
1390 htcp_clr_access deny all
1395 LOC: Config.accessList.miss
1397 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1399 Determins whether network access is permitted when satisfying a request.
1402 to force your neighbors to use you as a sibling instead of
1405 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1406 miss_access deny !localclients
1407 miss_access allow all
1409 This means only your local clients are allowed to fetch relayed/MISS
1410 replies from the network and all other clients can only fetch cached
1413 The default for this setting allows all clients who passed the
1414 http_access rules to relay via this proxy.
1416 This clause only supports fast acl types.
1417 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1420 NAME: ident_lookup_access
1424 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1425 LOC: Ident::TheConfig.identLookup
1427 A list of ACL elements which, if matched, cause an ident
1428 (RFC 931) lookup to be performed for this request. For
1429 example, you might choose to always perform ident lookups
1430 for your main multi-user Unix boxes, but not for your Macs
1431 and PCs. By default, ident lookups are not performed for
1434 To enable ident lookups for specific client addresses, you
1435 can follow this example:
1437 acl ident_aware_hosts src 198.168.1.0/24
1438 ident_lookup_access allow ident_aware_hosts
1439 ident_lookup_access deny all
1441 Only src type ACL checks are fully supported. A srcdomain
1442 ACL might work at times, but it will not always provide
1445 This clause only supports fast acl types.
1446 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1449 NAME: reply_body_max_size
1450 COMMENT: size [acl acl...]
1453 DEFAULT_DOC: No limit is applied.
1454 LOC: Config.ReplyBodySize
1456 This option specifies the maximum size of a reply body. It can be
1457 used to prevent users from downloading very large files, such as
1458 MP3's and movies. When the reply headers are received, the
1459 reply_body_max_size lines are processed, and the first line where
1460 all (if any) listed ACLs are true is used as the maximum body size
1463 This size is checked twice. First when we get the reply headers,
1464 we check the content-length value. If the content length value exists
1465 and is larger than the allowed size, the request is denied and the
1466 user receives an error message that says "the request or reply
1467 is too large." If there is no content-length, and the reply
1468 size exceeds this limit, the client's connection is just closed
1469 and they will receive a partial reply.
1471 WARNING: downstream caches probably can not detect a partial reply
1472 if there is no content-length header, so they will cache
1473 partial responses and give them out as hits. You should NOT
1474 use this option if you have downstream caches.
1476 WARNING: A maximum size smaller than the size of squid's error messages
1477 will cause an infinite loop and crash squid. Ensure that the smallest
1478 non-zero value you use is greater that the maximum header size plus
1479 the size of your largest error page.
1481 If you set this parameter none (the default), there will be
1484 Configuration Format is:
1485 reply_body_max_size SIZE UNITS [acl ...]
1487 reply_body_max_size 10 MB
1493 -----------------------------------------------------------------------------
1496 NAME: http_port ascii_port
1499 LOC: Config.Sockaddr.http
1501 Usage: port [mode] [options]
1502 hostname:port [mode] [options]
1503 1.2.3.4:port [mode] [options]
1505 The socket addresses where Squid will listen for HTTP client
1506 requests. You may specify multiple socket addresses.
1507 There are three forms: port alone, hostname with port, and
1508 IP address with port. If you specify a hostname or IP
1509 address, Squid binds the socket to that specific
1510 address. Most likely, you do not need to bind to a specific
1511 address, so you can use the port number alone.
1513 If you are running Squid in accelerator mode, you
1514 probably want to listen on port 80 also, or instead.
1516 The -a command line option may be used to specify additional
1517 port(s) where Squid listens for proxy request. Such ports will
1518 be plain proxy ports with no options.
1520 You may specify multiple socket addresses on multiple lines.
1524 intercept Support for IP-Layer interception of
1525 outgoing requests without browser settings.
1526 NP: disables authentication and IPv6 on the port.
1528 tproxy Support Linux TPROXY for spoofing outgoing
1529 connections using the client IP address.
1530 NP: disables authentication and maybe IPv6 on the port.
1532 accel Accelerator / reverse proxy mode
1534 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1535 establish secure connection with the client and with
1536 the server, decrypt HTTPS messages as they pass through
1537 Squid, and treat them as unencrypted HTTP messages,
1538 becoming the man-in-the-middle.
1540 The ssl_bump option is required to fully enable
1541 bumping of CONNECT requests.
1543 Omitting the mode flag causes default forward proxy mode to be used.
1546 Accelerator Mode Options:
1548 defaultsite=domainname
1549 What to use for the Host: header if it is not present
1550 in a request. Determines what site (not origin server)
1551 accelerators should consider the default.
1553 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1555 protocol= Protocol to reconstruct accelerated and intercepted
1556 requests with. Defaults to HTTP/1.1 for http_port and
1557 HTTPS/1.1 for https_port.
1558 When an unsupported value is configured Squid will
1559 produce a FATAL error.
1560 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1562 vport Virtual host port support. Using the http_port number
1563 instead of the port passed on Host: headers.
1565 vport=NN Virtual host port support. Using the specified port
1566 number instead of the port passed on Host: headers.
1569 Act as if this Squid is the origin server.
1570 This currently means generate new Date: and Expires:
1571 headers on HIT instead of adding Age:.
1573 ignore-cc Ignore request Cache-Control headers.
1575 WARNING: This option violates HTTP specifications if
1576 used in non-accelerator setups.
1578 allow-direct Allow direct forwarding in accelerator mode. Normally
1579 accelerated requests are denied direct forwarding as if
1580 never_direct was used.
1582 WARNING: this option opens accelerator mode to security
1583 vulnerabilities usually only affecting in interception
1584 mode. Make sure to protect forwarding with suitable
1585 http_access rules when using this.
1588 SSL Bump Mode Options:
1589 In addition to these options ssl-bump requires TLS/SSL options.
1591 generate-host-certificates[=<on|off>]
1592 Dynamically create SSL server certificates for the
1593 destination hosts of bumped CONNECT requests.When
1594 enabled, the cert and key options are used to sign
1595 generated certificates. Otherwise generated
1596 certificate will be selfsigned.
1597 If there is a CA certificate lifetime of the generated
1598 certificate equals lifetime of the CA certificate. If
1599 generated certificate is selfsigned lifetime is three
1601 This option is enabled by default when ssl-bump is used.
1602 See the ssl-bump option above for more information.
1604 dynamic_cert_mem_cache_size=SIZE
1605 Approximate total RAM size spent on cached generated
1606 certificates. If set to zero, caching is disabled. The
1607 default value is 4MB.
1611 cert= Path to SSL certificate (PEM format).
1613 key= Path to SSL private key file (PEM format)
1614 if not specified, the certificate file is
1615 assumed to be a combined certificate and
1618 version= The version of SSL/TLS supported
1619 1 automatic (default)
1626 cipher= Colon separated list of supported ciphers.
1627 NOTE: some ciphers such as EDH ciphers depend on
1628 additional settings. If those settings are
1629 omitted the ciphers may be silently ignored
1630 by the OpenSSL library.
1632 options= Various SSL implementation options. The most important
1634 NO_SSLv2 Disallow the use of SSLv2
1635 NO_SSLv3 Disallow the use of SSLv3
1636 NO_TLSv1 Disallow the use of TLSv1.0
1637 NO_TLSv1_1 Disallow the use of TLSv1.1
1638 NO_TLSv1_2 Disallow the use of TLSv1.2
1639 SINGLE_DH_USE Always create a new key when using
1640 temporary/ephemeral DH key exchanges
1641 ALL Enable various bug workarounds
1642 suggested as "harmless" by OpenSSL
1643 Be warned that this reduces SSL/TLS
1644 strength to some attacks.
1645 See OpenSSL SSL_CTX_set_options documentation for a
1646 complete list of options.
1648 clientca= File containing the list of CAs to use when
1649 requesting a client certificate.
1651 cafile= File containing additional CA certificates to
1652 use when verifying client certificates. If unset
1653 clientca will be used.
1655 capath= Directory containing additional CA certificates
1656 and CRL lists to use when verifying client certificates.
1658 crlfile= File of additional CRL lists to use when verifying
1659 the client certificate, in addition to CRLs stored in
1660 the capath. Implies VERIFY_CRL flag below.
1662 dhparams= File containing DH parameters for temporary/ephemeral
1663 DH key exchanges. See OpenSSL documentation for details
1664 on how to create this file.
1665 WARNING: EDH ciphers will be silently disabled if this
1668 sslflags= Various flags modifying the use of SSL:
1670 Don't request client certificates
1671 immediately, but wait until acl processing
1672 requires a certificate (not yet implemented).
1674 Don't use the default CA lists built in
1677 Don't allow for session reuse. Each connection
1678 will result in a new SSL session.
1680 Verify CRL lists when accepting client
1683 Verify CRL lists for all certificates in the
1684 client certificate chain.
1686 sslcontext= SSL session ID context identifier.
1690 connection-auth[=on|off]
1691 use connection-auth=off to tell Squid to prevent
1692 forwarding Microsoft connection oriented authentication
1693 (NTLM, Negotiate and Kerberos)
1695 disable-pmtu-discovery=
1696 Control Path-MTU discovery usage:
1697 off lets OS decide on what to do (default).
1698 transparent disable PMTU discovery when transparent
1700 always disable always PMTU discovery.
1702 In many setups of transparently intercepting proxies
1703 Path-MTU discovery can not work on traffic towards the
1704 clients. This is the case when the intercepting device
1705 does not fully track connections and fails to forward
1706 ICMP must fragment messages to the cache server. If you
1707 have such setup and experience that certain clients
1708 sporadically hang or never complete requests set
1709 disable-pmtu-discovery option to 'transparent'.
1711 name= Specifies a internal name for the port. Defaults to
1712 the port specification (port or addr:port)
1714 tcpkeepalive[=idle,interval,timeout]
1715 Enable TCP keepalive probes of idle connections.
1716 In seconds; idle is the initial time before TCP starts
1717 probing the connection, interval how often to probe, and
1718 timeout the time before giving up.
1720 If you run Squid on a dual-homed machine with an internal
1721 and an external interface we recommend you to specify the
1722 internal address:port in http_port. This way Squid will only be
1723 visible on the internal address.
1727 # Squid normally listens to port 3128
1728 http_port @DEFAULT_HTTP_PORT@
1736 LOC: Config.Sockaddr.https
1738 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1740 The socket address where Squid will listen for client requests made
1741 over TLS or SSL connections. Commonly referred to as HTTPS.
1743 This is most useful for situations where you are running squid in
1744 accelerator mode and you want to do the SSL work at the accelerator level.
1746 You may specify multiple socket addresses on multiple lines,
1747 each with their own SSL certificate and/or options.
1751 accel Accelerator / reverse proxy mode
1753 intercept Support for IP-Layer interception of
1754 outgoing requests without browser settings.
1755 NP: disables authentication and IPv6 on the port.
1757 tproxy Support Linux TPROXY for spoofing outgoing
1758 connections using the client IP address.
1759 NP: disables authentication and maybe IPv6 on the port.
1761 ssl-bump For each intercepted connection allowed by ssl_bump
1762 ACLs, establish a secure connection with the client and with
1763 the server, decrypt HTTPS messages as they pass through
1764 Squid, and treat them as unencrypted HTTP messages,
1765 becoming the man-in-the-middle.
1767 An "ssl_bump server-first" match is required to
1768 fully enable bumping of intercepted SSL connections.
1770 Requires tproxy or intercept.
1772 Omitting the mode flag causes default forward proxy mode to be used.
1775 See http_port for a list of generic options
1780 cert= Path to SSL certificate (PEM format).
1782 key= Path to SSL private key file (PEM format)
1783 if not specified, the certificate file is
1784 assumed to be a combined certificate and
1787 version= The version of SSL/TLS supported
1788 1 automatic (default)
1793 cipher= Colon separated list of supported ciphers.
1795 options= Various SSL engine options. The most important
1797 NO_SSLv2 Disallow the use of SSLv2
1798 NO_SSLv3 Disallow the use of SSLv3
1799 NO_TLSv1 Disallow the use of TLSv1
1800 SINGLE_DH_USE Always create a new key when using
1801 temporary/ephemeral DH key exchanges
1802 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1803 documentation for a complete list of options.
1805 clientca= File containing the list of CAs to use when
1806 requesting a client certificate.
1808 cafile= File containing additional CA certificates to
1809 use when verifying client certificates. If unset
1810 clientca will be used.
1812 capath= Directory containing additional CA certificates
1813 and CRL lists to use when verifying client certificates.
1815 crlfile= File of additional CRL lists to use when verifying
1816 the client certificate, in addition to CRLs stored in
1817 the capath. Implies VERIFY_CRL flag below.
1819 dhparams= File containing DH parameters for temporary/ephemeral
1822 sslflags= Various flags modifying the use of SSL:
1824 Don't request client certificates
1825 immediately, but wait until acl processing
1826 requires a certificate (not yet implemented).
1828 Don't use the default CA lists built in
1831 Don't allow for session reuse. Each connection
1832 will result in a new SSL session.
1834 Verify CRL lists when accepting client
1837 Verify CRL lists for all certificates in the
1838 client certificate chain.
1840 sslcontext= SSL session ID context identifier.
1842 generate-host-certificates[=<on|off>]
1843 Dynamically create SSL server certificates for the
1844 destination hosts of bumped SSL requests.When
1845 enabled, the cert and key options are used to sign
1846 generated certificates. Otherwise generated
1847 certificate will be selfsigned.
1848 If there is CA certificate life time of generated
1849 certificate equals lifetime of CA certificate. If
1850 generated certificate is selfsigned lifetime is three
1852 This option is enabled by default when SslBump is used.
1853 See the sslBump option above for more information.
1855 dynamic_cert_mem_cache_size=SIZE
1856 Approximate total RAM size spent on cached generated
1857 certificates. If set to zero, caching is disabled. The
1858 default value is 4MB.
1860 See http_port for a list of available options.
1863 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1866 LOC: Ip::Qos::TheConfig.tosToServer
1868 Allows you to select a TOS/Diffserv value for packets outgoing
1869 on the server side, based on an ACL.
1871 tcp_outgoing_tos ds-field [!]aclname ...
1873 Example where normal_service_net uses the TOS value 0x00
1874 and good_service_net uses 0x20
1876 acl normal_service_net src 10.0.0.0/24
1877 acl good_service_net src 10.0.1.0/24
1878 tcp_outgoing_tos 0x00 normal_service_net
1879 tcp_outgoing_tos 0x20 good_service_net
1881 TOS/DSCP values really only have local significance - so you should
1882 know what you're specifying. For more information, see RFC2474,
1883 RFC2475, and RFC3260.
1885 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1886 "default" to use whatever default your host has. Note that in
1887 practice often only multiples of 4 is usable as the two rightmost bits
1888 have been redefined for use by ECN (RFC 3168 section 23.1).
1890 Processing proceeds in the order specified, and stops at first fully
1894 NAME: clientside_tos
1897 LOC: Ip::Qos::TheConfig.tosToClient
1899 Allows you to select a TOS/Diffserv value for packets being transmitted
1900 on the client-side, based on an ACL.
1902 clientside_tos ds-field [!]aclname ...
1904 Example where normal_service_net uses the TOS value 0x00
1905 and good_service_net uses 0x20
1907 acl normal_service_net src 10.0.0.0/24
1908 acl good_service_net src 10.0.1.0/24
1909 clientside_tos 0x00 normal_service_net
1910 clientside_tos 0x20 good_service_net
1912 Note: This feature is incompatible with qos_flows. Any TOS values set here
1913 will be overwritten by TOS values in qos_flows.
1916 NAME: tcp_outgoing_mark
1918 IFDEF: SO_MARK&&USE_LIBCAP
1920 LOC: Ip::Qos::TheConfig.nfmarkToServer
1922 Allows you to apply a Netfilter mark value to outgoing packets
1923 on the server side, based on an ACL.
1925 tcp_outgoing_mark mark-value [!]aclname ...
1927 Example where normal_service_net uses the mark value 0x00
1928 and good_service_net uses 0x20
1930 acl normal_service_net src 10.0.0.0/24
1931 acl good_service_net src 10.0.1.0/24
1932 tcp_outgoing_mark 0x00 normal_service_net
1933 tcp_outgoing_mark 0x20 good_service_net
1936 NAME: clientside_mark
1938 IFDEF: SO_MARK&&USE_LIBCAP
1940 LOC: Ip::Qos::TheConfig.nfmarkToClient
1942 Allows you to apply a Netfilter mark value to packets being transmitted
1943 on the client-side, based on an ACL.
1945 clientside_mark mark-value [!]aclname ...
1947 Example where normal_service_net uses the mark value 0x00
1948 and good_service_net uses 0x20
1950 acl normal_service_net src 10.0.0.0/24
1951 acl good_service_net src 10.0.1.0/24
1952 clientside_mark 0x00 normal_service_net
1953 clientside_mark 0x20 good_service_net
1955 Note: This feature is incompatible with qos_flows. Any mark values set here
1956 will be overwritten by mark values in qos_flows.
1963 LOC: Ip::Qos::TheConfig
1965 Allows you to select a TOS/DSCP value to mark outgoing
1966 connections to the client, based on where the reply was sourced.
1967 For platforms using netfilter, allows you to set a netfilter mark
1968 value instead of, or in addition to, a TOS value.
1970 By default this functionality is disabled. To enable it with the default
1971 settings simply use "qos_flows mark" or "qos_flows tos". Default
1972 settings will result in the netfilter mark or TOS value being copied
1973 from the upstream connection to the client. Note that it is the connection
1974 CONNMARK value not the packet MARK value that is copied.
1976 It is not currently possible to copy the mark or TOS value from the
1977 client to the upstream connection request.
1979 TOS values really only have local significance - so you should
1980 know what you're specifying. For more information, see RFC2474,
1981 RFC2475, and RFC3260.
1983 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1984 in practice often only multiples of 4 is usable as the two rightmost bits
1985 have been redefined for use by ECN (RFC 3168 section 23.1).
1987 Mark values can be any unsigned 32-bit integer value.
1989 This setting is configured by setting the following values:
1991 tos|mark Whether to set TOS or netfilter mark values
1993 local-hit=0xFF Value to mark local cache hits.
1995 sibling-hit=0xFF Value to mark hits from sibling peers.
1997 parent-hit=0xFF Value to mark hits from parent peers.
1999 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2000 over the preserve-miss feature (see below), unless
2001 mask is specified, in which case only the bits
2002 specified in the mask are written.
2004 The TOS variant of the following features are only possible on Linux
2005 and require your kernel to be patched with the TOS preserving ZPH
2006 patch, available from http://zph.bratcheda.org
2007 No patch is needed to preserve the netfilter mark, which will work
2008 with all variants of netfilter.
2010 disable-preserve-miss
2011 This option disables the preservation of the TOS or netfilter
2012 mark. By default, the existing TOS or netfilter mark value of
2013 the response coming from the remote server will be retained
2014 and masked with miss-mark.
2015 NOTE: in the case of a netfilter mark, the mark must be set on
2016 the connection (using the CONNMARK target) not on the packet
2020 Allows you to mask certain bits in the TOS or mark value
2021 received from the remote server, before copying the value to
2022 the TOS sent towards clients.
2023 Default for tos: 0xFF (TOS from server is not changed).
2024 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2026 All of these features require the --enable-zph-qos compilation flag
2027 (enabled by default). Netfilter marking also requires the
2028 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2029 libcap 2.09+ (--with-libcap).
2033 NAME: tcp_outgoing_address
2036 DEFAULT_DOC: Address selection is performed by the operating system.
2037 LOC: Config.accessList.outgoing_address
2039 Allows you to map requests to different outgoing IP addresses
2040 based on the username or source address of the user making
2043 tcp_outgoing_address ipaddr [[!]aclname] ...
2046 Forwarding clients with dedicated IPs for certain subnets.
2048 acl normal_service_net src 10.0.0.0/24
2049 acl good_service_net src 10.0.2.0/24
2051 tcp_outgoing_address 2001:db8::c001 good_service_net
2052 tcp_outgoing_address 10.1.0.2 good_service_net
2054 tcp_outgoing_address 2001:db8::beef normal_service_net
2055 tcp_outgoing_address 10.1.0.1 normal_service_net
2057 tcp_outgoing_address 2001:db8::1
2058 tcp_outgoing_address 10.1.0.3
2060 Processing proceeds in the order specified, and stops at first fully
2063 Squid will add an implicit IP version test to each line.
2064 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2065 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2068 NOTE: The use of this directive using client dependent ACLs is
2069 incompatible with the use of server side persistent connections. To
2070 ensure correct results it is best to set server_persistent_connections
2071 to off when using this directive in such configurations.
2073 NOTE: The use of this directive to set a local IP on outgoing TCP links
2074 is incompatible with using TPROXY to set client IP out outbound TCP links.
2075 When needing to contact peers use the no-tproxy cache_peer option and the
2076 client_dst_passthru directive re-enable normal forwarding such as this.
2080 NAME: host_verify_strict
2083 LOC: Config.onoff.hostStrictVerify
2085 Regardless of this option setting, when dealing with intercepted
2086 traffic, Squid always verifies that the destination IP address matches
2087 the Host header domain or IP (called 'authority form URL').
2089 This enforcement is performed to satisfy a MUST-level requirement in
2090 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2091 authority of the origin server or gateway given by the original URL".
2094 Squid always responds with an HTTP 409 (Conflict) error
2095 page and logs a security warning if there is no match.
2097 Squid verifies that the destination IP address matches
2098 the Host header for forward-proxy and reverse-proxy traffic
2099 as well. For those traffic types, Squid also enables the
2100 following checks, comparing the corresponding Host header
2101 and Request-URI components:
2103 * The host names (domain or IP) must be identical,
2104 but valueless or missing Host header disables all checks.
2105 For the two host names to match, both must be either IP
2108 * Port numbers must be identical, but if a port is missing
2109 the scheme-default port is assumed.
2112 When set to OFF (the default):
2113 Squid allows suspicious requests to continue but logs a
2114 security warning and blocks caching of the response.
2116 * Forward-proxy traffic is not checked at all.
2118 * Reverse-proxy traffic is not checked at all.
2120 * Intercepted traffic which passes verification is handled
2121 according to client_dst_passthru.
2123 * Intercepted requests which fail verification are sent
2124 to the client original destination instead of DIRECT.
2125 This overrides 'client_dst_passthru off'.
2127 For now suspicious intercepted CONNECT requests are always
2128 responded to with an HTTP 409 (Conflict) error page.
2133 As described in CVE-2009-0801 when the Host: header alone is used
2134 to determine the destination of a request it becomes trivial for
2135 malicious scripts on remote websites to bypass browser same-origin
2136 security policy and sandboxing protections.
2138 The cause of this is that such applets are allowed to perform their
2139 own HTTP stack, in which case the same-origin policy of the browser
2140 sandbox only verifies that the applet tries to contact the same IP
2141 as from where it was loaded at the IP level. The Host: header may
2142 be different from the connected IP and approved origin.
2146 NAME: client_dst_passthru
2149 LOC: Config.onoff.client_dst_passthru
2151 With NAT or TPROXY intercepted traffic Squid may pass the request
2152 directly to the original client destination IP or seek a faster
2153 source using the HTTP Host header.
2155 Using Host to locate alternative servers can provide faster
2156 connectivity with a range of failure recovery options.
2157 But can also lead to connectivity trouble when the client and
2158 server are attempting stateful interactions unaware of the proxy.
2160 This option (on by default) prevents alternative DNS entries being
2161 located to send intercepted traffic DIRECT to an origin server.
2162 The clients original destination IP and port will be used instead.
2164 Regardless of this option setting, when dealing with intercepted
2165 traffic Squid will verify the Host: header and any traffic which
2166 fails Host verification will be treated as if this option were ON.
2168 see host_verify_strict for details on the verification process.
2173 -----------------------------------------------------------------------------
2176 NAME: ssl_unclean_shutdown
2180 LOC: Config.SSL.unclean_shutdown
2182 Some browsers (especially MSIE) bugs out on SSL shutdown
2189 LOC: Config.SSL.ssl_engine
2192 The OpenSSL engine to use. You will need to set this if you
2193 would like to use hardware SSL acceleration for example.
2196 NAME: sslproxy_client_certificate
2199 LOC: Config.ssl_client.cert
2202 Client SSL Certificate to use when proxying https:// URLs
2205 NAME: sslproxy_client_key
2208 LOC: Config.ssl_client.key
2211 Client SSL Key to use when proxying https:// URLs
2214 NAME: sslproxy_version
2217 DEFAULT_DOC: automatic SSL/TLS version negotiation
2218 LOC: Config.ssl_client.version
2221 SSL version level to use when proxying https:// URLs
2223 The versions of SSL/TLS supported:
2225 1 automatic (default)
2233 NAME: sslproxy_options
2236 LOC: Config.ssl_client.options
2239 SSL implementation options to use when proxying https:// URLs
2241 The most important being:
2243 NO_SSLv2 Disallow the use of SSLv2
2244 NO_SSLv3 Disallow the use of SSLv3
2245 NO_TLSv1 Disallow the use of TLSv1.0
2246 NO_TLSv1_1 Disallow the use of TLSv1.1
2247 NO_TLSv1_2 Disallow the use of TLSv1.2
2249 Always create a new key when using temporary/ephemeral
2252 Disable use of RFC5077 session tickets. Some servers
2253 may have problems understanding the TLS extension due
2254 to ambiguous specification in RFC4507.
2255 ALL Enable various bug workarounds suggested as "harmless"
2256 by OpenSSL. Be warned that this may reduce SSL/TLS
2257 strength to some attacks.
2259 See the OpenSSL SSL_CTX_set_options documentation for a
2260 complete list of possible options.
2263 NAME: sslproxy_cipher
2266 LOC: Config.ssl_client.cipher
2269 SSL cipher list to use when proxying https:// URLs
2271 Colon separated list of supported ciphers.
2274 NAME: sslproxy_cafile
2277 LOC: Config.ssl_client.cafile
2280 file containing CA certificates to use when verifying server
2281 certificates while proxying https:// URLs
2284 NAME: sslproxy_capath
2287 LOC: Config.ssl_client.capath
2290 directory containing CA certificates to use when verifying
2291 server certificates while proxying https:// URLs
2294 NAME: sslproxy_session_ttl
2297 LOC: Config.SSL.session_ttl
2300 Sets the timeout value for SSL sessions
2303 NAME: sslproxy_session_cache_size
2306 LOC: Config.SSL.sessionCacheSize
2309 Sets the cache size to use for ssl session
2314 TYPE: sslproxy_ssl_bump
2315 LOC: Config.accessList.ssl_bump
2316 DEFAULT_DOC: Does not bump unless rules are present in squid.conf
2319 This option is consulted when a CONNECT request is received on
2320 an http_port (or a new connection is intercepted at an
2321 https_port), provided that port was configured with an ssl-bump
2322 flag. The subsequent data on the connection is either treated as
2323 HTTPS and decrypted OR tunneled at TCP level without decryption,
2324 depending on the first bumping "mode" which ACLs match.
2326 ssl_bump <mode> [!]acl ...
2328 The following bumping modes are supported:
2331 Allow bumping of the connection. Establish a secure connection
2332 with the client first, then connect to the server. This old mode
2333 does not allow Squid to mimic server SSL certificate and does
2334 not work with intercepted SSL connections.
2337 Allow bumping of the connection. Establish a secure connection
2338 with the server first, then establish a secure connection with
2339 the client, using a mimicked server certificate. Works with both
2340 CONNECT requests and intercepted SSL connections.
2343 Become a TCP tunnel without decoding the connection.
2344 Works with both CONNECT requests and intercepted SSL
2345 connections. This is the default behavior when no
2346 ssl_bump option is given or no ssl_bump ACLs match.
2348 By default, no connections are bumped.
2350 The first matching ssl_bump option wins. If no ACLs match, the
2351 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2352 does not have an implicit "negate the last given option" rule. You
2353 must make that rule explicit if you convert old ssl_bump allow/deny
2354 rules that rely on such an implicit rule.
2356 This clause supports both fast and slow acl types.
2357 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2359 See also: http_port ssl-bump, https_port ssl-bump
2362 # Example: Bump all requests except those originating from
2363 # localhost or those going to example.com.
2365 acl broken_sites dstdomain .example.com
2366 ssl_bump none localhost
2367 ssl_bump none broken_sites
2368 ssl_bump server-first all
2371 NAME: sslproxy_flags
2374 LOC: Config.ssl_client.flags
2377 Various flags modifying the use of SSL while proxying https:// URLs:
2378 DONT_VERIFY_PEER Accept certificates that fail verification.
2379 For refined control, see sslproxy_cert_error.
2380 NO_DEFAULT_CA Don't use the default CA list built in
2384 NAME: sslproxy_cert_error
2387 DEFAULT_DOC: Server certificate errors terminate the transaction.
2388 LOC: Config.ssl_client.cert_error
2391 Use this ACL to bypass server certificate validation errors.
2393 For example, the following lines will bypass all validation errors
2394 when talking to servers for example.com. All other
2395 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2397 acl BrokenButTrustedServers dstdomain example.com
2398 sslproxy_cert_error allow BrokenButTrustedServers
2399 sslproxy_cert_error deny all
2401 This clause only supports fast acl types.
2402 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2403 Using slow acl types may result in server crashes
2405 Without this option, all server certificate validation errors
2406 terminate the transaction to protect Squid and the client.
2408 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2409 but should not happen unless your OpenSSL library is buggy.
2412 Bypassing validation errors is dangerous because an
2413 error usually implies that the server cannot be trusted
2414 and the connection may be insecure.
2416 See also: sslproxy_flags and DONT_VERIFY_PEER.
2419 NAME: sslproxy_cert_sign
2422 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2423 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2424 POSTSCRIPTUM: signTrusted all
2425 TYPE: sslproxy_cert_sign
2426 LOC: Config.ssl_client.cert_sign
2429 sslproxy_cert_sign <signing algorithm> acl ...
2431 The following certificate signing algorithms are supported:
2434 Sign using the configured CA certificate which is usually
2435 placed in and trusted by end-user browsers. This is the
2436 default for trusted origin server certificates.
2439 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2440 This is the default for untrusted origin server certificates
2441 that are not self-signed (see ssl::certUntrusted).
2444 Sign using a self-signed certificate with the right CN to
2445 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2446 browser. This is the default for self-signed origin server
2447 certificates (see ssl::certSelfSigned).
2449 This clause only supports fast acl types.
2451 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2452 signing algorithm to generate the certificate and ignores all
2453 subsequent sslproxy_cert_sign options (the first match wins). If no
2454 acl(s) match, the default signing algorithm is determined by errors
2455 detected when obtaining and validating the origin server certificate.
2457 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2458 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2459 CONNECT request that carries a domain name. In all other cases (CONNECT
2460 to an IP address or an intercepted SSL connection), Squid cannot detect
2461 the domain mismatch at certificate generation time when
2462 bump-server-first is used.
2465 NAME: sslproxy_cert_adapt
2468 TYPE: sslproxy_cert_adapt
2469 LOC: Config.ssl_client.cert_adapt
2472 sslproxy_cert_adapt <adaptation algorithm> acl ...
2474 The following certificate adaptation algorithms are supported:
2477 Sets the "Not After" property to the "Not After" property of
2478 the CA certificate used to sign generated certificates.
2481 Sets the "Not Before" property to the "Not Before" property of
2482 the CA certificate used to sign generated certificates.
2484 setCommonName or setCommonName{CN}
2485 Sets Subject.CN property to the host name specified as a
2486 CN parameter or, if no explicit CN parameter was specified,
2487 extracted from the CONNECT request. It is a misconfiguration
2488 to use setCommonName without an explicit parameter for
2489 intercepted or tproxied SSL connections.
2491 This clause only supports fast acl types.
2493 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2494 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2495 corresponding adaptation algorithm to generate the certificate and
2496 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2497 group (i.e., the first match wins within each algorithm group). If no
2498 acl(s) match, the default mimicking action takes place.
2500 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2501 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2502 CONNECT request that carries a domain name. In all other cases (CONNECT
2503 to an IP address or an intercepted SSL connection), Squid cannot detect
2504 the domain mismatch at certificate generation time when
2505 bump-server-first is used.
2508 NAME: sslpassword_program
2511 LOC: Config.Program.ssl_password
2514 Specify a program used for entering SSL key passphrases
2515 when using encrypted SSL certificate keys. If not specified
2516 keys must either be unencrypted, or Squid started with the -N
2517 option to allow it to query interactively for the passphrase.
2519 The key file name is given as argument to the program allowing
2520 selection of the right password if you have multiple encrypted
2525 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2526 -----------------------------------------------------------------------------
2529 NAME: sslcrtd_program
2532 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2533 LOC: Ssl::TheConfig.ssl_crtd
2535 Specify the location and options of the executable for ssl_crtd process.
2536 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2537 For more information use:
2538 @DEFAULT_SSL_CRTD@ -h
2541 NAME: sslcrtd_children
2542 TYPE: HelperChildConfig
2544 DEFAULT: 32 startup=5 idle=1
2545 LOC: Ssl::TheConfig.ssl_crtdChildren
2547 The maximum number of processes spawn to service ssl server.
2548 The maximum this may be safely set to is 32.
2550 The startup= and idle= options allow some measure of skew in your
2555 Sets the minimum number of processes to spawn when Squid
2556 starts or reconfigures. When set to zero the first request will
2557 cause spawning of the first child process to handle it.
2559 Starting too few children temporary slows Squid under load while it
2560 tries to spawn enough additional processes to cope with traffic.
2564 Sets a minimum of how many processes Squid is to try and keep available
2565 at all times. When traffic begins to rise above what the existing
2566 processes can handle this many more will be spawned up to the maximum
2567 configured. A minimum setting of 1 is required.
2569 You must have at least one ssl_crtd process.
2572 NAME: sslcrtvalidator_program
2576 LOC: Ssl::TheConfig.ssl_crt_validator
2578 Specify the location and options of the executable for ssl_crt_validator
2581 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2584 ttl=n TTL in seconds for cached results. The default is 60 secs
2585 cache=n limit the result cache size. The default value is 2048
2588 NAME: sslcrtvalidator_children
2589 TYPE: HelperChildConfig
2591 DEFAULT: 32 startup=5 idle=1 concurrency=1
2592 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2594 The maximum number of processes spawn to service SSL server.
2595 The maximum this may be safely set to is 32.
2597 The startup= and idle= options allow some measure of skew in your
2602 Sets the minimum number of processes to spawn when Squid
2603 starts or reconfigures. When set to zero the first request will
2604 cause spawning of the first child process to handle it.
2606 Starting too few children temporary slows Squid under load while it
2607 tries to spawn enough additional processes to cope with traffic.
2611 Sets a minimum of how many processes Squid is to try and keep available
2612 at all times. When traffic begins to rise above what the existing
2613 processes can handle this many more will be spawned up to the maximum
2614 configured. A minimum setting of 1 is required.
2618 The number of requests each certificate validator helper can handle in
2619 parallel. A value of 0 indicates the certficate validator does not
2620 support concurrency. Defaults to 1.
2622 When this directive is set to a value >= 1 then the protocol
2623 used to communicate with the helper is modified to include
2624 a request ID in front of the request/response. The request
2625 ID from the request must be echoed back with the response
2628 You must have at least one ssl_crt_validator process.
2632 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2633 -----------------------------------------------------------------------------
2641 To specify other caches in a hierarchy, use the format:
2643 cache_peer hostname type http-port icp-port [options]
2648 # hostname type port port options
2649 # -------------------- -------- ----- ----- -----------
2650 cache_peer parent.foo.net parent 3128 3130 default
2651 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2652 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2653 cache_peer example.com parent 80 0 default
2654 cache_peer cdn.example.com sibling 3128 0
2656 type: either 'parent', 'sibling', or 'multicast'.
2658 proxy-port: The port number where the peer accept HTTP requests.
2659 For other Squid proxies this is usually 3128
2660 For web servers this is usually 80
2662 icp-port: Used for querying neighbor caches about objects.
2663 Set to 0 if the peer does not support ICP or HTCP.
2664 See ICP and HTCP options below for additional details.
2667 ==== ICP OPTIONS ====
2669 You MUST also set icp_port and icp_access explicitly when using these options.
2670 The defaults will prevent peer traffic using ICP.
2673 no-query Disable ICP queries to this neighbor.
2676 Indicates the named peer is a member of a multicast group.
2677 ICP queries will not be sent directly to the peer, but ICP
2678 replies will be accepted from it.
2680 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2681 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2684 To only send ICP queries to this neighbor infrequently.
2685 This is used to keep the neighbor round trip time updated
2686 and is usually used in conjunction with weighted-round-robin.
2689 ==== HTCP OPTIONS ====
2691 You MUST also set htcp_port and htcp_access explicitly when using these options.
2692 The defaults will prevent peer traffic using HTCP.
2695 htcp Send HTCP, instead of ICP, queries to the neighbor.
2696 You probably also want to set the "icp-port" to 4827
2697 instead of 3130. This directive accepts a comma separated
2698 list of options described below.
2700 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2702 htcp=no-clr Send HTCP to the neighbor but without
2703 sending any CLR requests. This cannot be used with
2706 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2707 This cannot be used with no-clr.
2710 Send HTCP to the neighbor including CLRs but only when
2711 they do not result from PURGE requests.
2714 Forward any HTCP CLR requests this proxy receives to the peer.
2717 ==== PEER SELECTION METHODS ====
2719 The default peer selection method is ICP, with the first responding peer
2720 being used as source. These options can be used for better load balancing.
2723 default This is a parent cache which can be used as a "last-resort"
2724 if a peer cannot be located by any of the peer-selection methods.
2725 If specified more than once, only the first is used.
2727 round-robin Load-Balance parents which should be used in a round-robin
2728 fashion in the absence of any ICP queries.
2729 weight=N can be used to add bias.
2731 weighted-round-robin
2732 Load-Balance parents which should be used in a round-robin
2733 fashion with the frequency of each parent being based on the
2734 round trip time. Closer parents are used more often.
2735 Usually used for background-ping parents.
2736 weight=N can be used to add bias.
2738 carp Load-Balance parents which should be used as a CARP array.
2739 The requests will be distributed among the parents based on the
2740 CARP load balancing hash function based on their weight.
2742 userhash Load-balance parents based on the client proxy_auth or ident username.
2744 sourcehash Load-balance parents based on the client source IP.
2747 To be used only for cache peers of type "multicast".
2748 ALL members of this multicast group have "sibling"
2749 relationship with it, not "parent". This is to a multicast
2750 group when the requested object would be fetched only from
2751 a "parent" cache, anyway. It's useful, e.g., when
2752 configuring a pool of redundant Squid proxies, being
2753 members of the same multicast group.
2756 ==== PEER SELECTION OPTIONS ====
2758 weight=N use to affect the selection of a peer during any weighted
2759 peer-selection mechanisms.
2760 The weight must be an integer; default is 1,
2761 larger weights are favored more.
2762 This option does not affect parent selection if a peering
2763 protocol is not in use.
2765 basetime=N Specify a base amount to be subtracted from round trip
2767 It is subtracted before division by weight in calculating
2768 which parent to fectch from. If the rtt is less than the
2769 base time the rtt is set to a minimal value.
2771 ttl=N Specify a TTL to use when sending multicast ICP queries
2773 Only useful when sending to a multicast group.
2774 Because we don't accept ICP replies from random
2775 hosts, you must configure other group members as
2776 peers with the 'multicast-responder' option.
2778 no-delay To prevent access to this neighbor from influencing the
2781 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2782 enabled) for this host from the specified URL rather
2783 than the Squid default location.
2786 ==== CARP OPTIONS ====
2788 carp-key=key-specification
2789 use a different key than the full URL to hash against the peer.
2790 the key-specification is a comma-separated list of the keywords
2791 scheme, host, port, path, params
2792 Order is not important.
2794 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2796 originserver Causes this parent to be contacted as an origin server.
2797 Meant to be used in accelerator setups when the peer
2801 Set the Host header of requests forwarded to this peer.
2802 Useful in accelerator setups where the server (peer)
2803 expects a certain domain name but clients may request
2804 others. ie example.com or www.example.com
2806 no-digest Disable request of cache digests.
2809 Disables requesting ICMP RTT database (NetDB).
2812 ==== AUTHENTICATION OPTIONS ====
2815 If this is a personal/workgroup proxy and your parent
2816 requires proxy authentication.
2818 Note: The string can include URL escapes (i.e. %20 for
2819 spaces). This also means % must be written as %%.
2822 Send login details received from client to this peer.
2823 Both Proxy- and WWW-Authorization headers are passed
2824 without alteration to the peer.
2825 Authentication is not required by Squid for this to work.
2827 Note: This will pass any form of authentication but
2828 only Basic auth will work through a proxy unless the
2829 connection-auth options are also used.
2831 login=PASS Send login details received from client to this peer.
2832 Authentication is not required by this option.
2834 If there are no client-provided authentication headers
2835 to pass on, but username and password are available
2836 from an external ACL user= and password= result tags
2837 they may be sent instead.
2839 Note: To combine this with proxy_auth both proxies must
2840 share the same user database as HTTP only allows for
2841 a single login (one for proxy, one for origin server).
2842 Also be warned this will expose your users proxy
2843 password to the peer. USE WITH CAUTION
2846 Send the username to the upstream cache, but with a
2847 fixed password. This is meant to be used when the peer
2848 is in another administrative domain, but it is still
2849 needed to identify each user.
2850 The star can optionally be followed by some extra
2851 information which is added to the username. This can
2852 be used to identify this proxy to the peer, similar to
2853 the login=username:password option above.
2856 If this is a personal/workgroup proxy and your parent
2857 requires a secure proxy authentication.
2858 The first principal from the default keytab or defined by
2859 the environment variable KRB5_KTNAME will be used.
2861 WARNING: The connection may transmit requests from multiple
2862 clients. Negotiate often assumes end-to-end authentication
2863 and a single-client. Which is not strictly true here.
2865 login=NEGOTIATE:principal_name
2866 If this is a personal/workgroup proxy and your parent
2867 requires a secure proxy authentication.
2868 The principal principal_name from the default keytab or
2869 defined by the environment variable KRB5_KTNAME will be
2872 WARNING: The connection may transmit requests from multiple
2873 clients. Negotiate often assumes end-to-end authentication
2874 and a single-client. Which is not strictly true here.
2876 connection-auth=on|off
2877 Tell Squid that this peer does or not support Microsoft
2878 connection oriented authentication, and any such
2879 challenges received from there should be ignored.
2880 Default is auto to automatically determine the status
2884 ==== SSL / HTTPS / TLS OPTIONS ====
2886 ssl Encrypt connections to this peer with SSL/TLS.
2888 sslcert=/path/to/ssl/certificate
2889 A client SSL certificate to use when connecting to
2892 sslkey=/path/to/ssl/key
2893 The private SSL key corresponding to sslcert above.
2894 If 'sslkey' is not specified 'sslcert' is assumed to
2895 reference a combined file containing both the
2896 certificate and the key.
2898 sslversion=1|2|3|4|5|6
2899 The SSL version to use when connecting to this peer
2900 1 = automatic (default)
2907 sslcipher=... The list of valid SSL ciphers to use when connecting
2910 ssloptions=... Specify various SSL implementation options:
2912 NO_SSLv2 Disallow the use of SSLv2
2913 NO_SSLv3 Disallow the use of SSLv3
2914 NO_TLSv1 Disallow the use of TLSv1.0
2915 NO_TLSv1_1 Disallow the use of TLSv1.1
2916 NO_TLSv1_2 Disallow the use of TLSv1.2
2918 Always create a new key when using
2919 temporary/ephemeral DH key exchanges
2920 ALL Enable various bug workarounds
2921 suggested as "harmless" by OpenSSL
2922 Be warned that this reduces SSL/TLS
2923 strength to some attacks.
2925 See the OpenSSL SSL_CTX_set_options documentation for a
2928 sslcafile=... A file containing additional CA certificates to use
2929 when verifying the peer certificate.
2931 sslcapath=... A directory containing additional CA certificates to
2932 use when verifying the peer certificate.
2934 sslcrlfile=... A certificate revocation list file to use when
2935 verifying the peer certificate.
2937 sslflags=... Specify various flags modifying the SSL implementation:
2940 Accept certificates even if they fail to
2943 Don't use the default CA list built in
2946 Don't verify the peer certificate
2947 matches the server name
2949 ssldomain= The peer name as advertised in it's certificate.
2950 Used for verifying the correctness of the received peer
2951 certificate. If not specified the peer hostname will be
2955 Enable the "Front-End-Https: On" header needed when
2956 using Squid as a SSL frontend in front of Microsoft OWA.
2957 See MS KB document Q307347 for details on this header.
2958 If set to auto the header will only be added if the
2959 request is forwarded as a https:// URL.
2962 ==== GENERAL OPTIONS ====
2965 A peer-specific connect timeout.
2966 Also see the peer_connect_timeout directive.
2968 connect-fail-limit=N
2969 How many times connecting to a peer must fail before
2970 it is marked as down. Standby connection failures
2971 count towards this limit. Default is 10.
2973 allow-miss Disable Squid's use of only-if-cached when forwarding
2974 requests to siblings. This is primarily useful when
2975 icp_hit_stale is used by the sibling. To extensive use
2976 of this option may result in forwarding loops, and you
2977 should avoid having two-way peerings with this option.
2978 For example to deny peer usage on requests from peer
2979 by denying cache_peer_access if the source is a peer.
2981 max-conn=N Limit the number of concurrent connections the Squid
2982 may open to this peer, including already opened idle
2983 and standby connections. There is no peer-specific
2984 connection limit by default.
2986 A peer exceeding the limit is not used for new
2987 requests unless a standby connection is available.
2989 max-conn currently works poorly with idle persistent
2990 connections: When a peer reaches its max-conn limit,
2991 and there are idle persistent connections to the peer,
2992 the peer may not be selected because the limiting code
2993 does not know whether Squid can reuse those idle
2996 standby=N Maintain a pool of N "hot standby" connections to an
2997 UP peer, available for requests when no idle
2998 persistent connection is available (or safe) to use.
2999 By default and with zero N, no such pool is maintained.
3000 N must not exceed the max-conn limit (if any).
3002 At start or after reconfiguration, Squid opens new TCP
3003 standby connections until there are N connections
3004 available and then replenishes the standby pool as
3005 opened connections are used up for requests. A used
3006 connection never goes back to the standby pool, but
3007 may go to the regular idle persistent connection pool
3008 shared by all peers and origin servers.
3010 Squid never opens multiple new standby connections
3011 concurrently. This one-at-a-time approach minimizes
3012 flooding-like effect on peers. Furthermore, just a few
3013 standby connections should be sufficient in most cases
3014 to supply most new requests with a ready-to-use
3017 Standby connections obey server_idle_pconn_timeout.
3018 For the feature to work as intended, the peer must be
3019 configured to accept and keep them open longer than
3020 the idle timeout at the connecting Squid, to minimize
3021 race conditions typical to idle used persistent
3022 connections. Default request_timeout and
3023 server_idle_pconn_timeout values ensure such a
3026 name=xxx Unique name for the peer.
3027 Required if you have multiple peers on the same host
3028 but different ports.
3029 This name can be used in cache_peer_access and similar
3030 directives to dentify the peer.
3031 Can be used by outgoing access controls through the
3034 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3035 requests to this peer. Use normal address selection instead.
3036 This overrides the spoof_client_ip ACL.
3038 proxy-only objects fetched from the peer will not be stored locally.
3042 NAME: cache_peer_domain cache_host_domain
3047 Use to limit the domains for which a neighbor cache will be
3051 cache_peer_domain cache-host domain [domain ...]
3052 cache_peer_domain cache-host !domain
3054 For example, specifying
3056 cache_peer_domain parent.foo.net .edu
3058 has the effect such that UDP query packets are sent to
3059 'bigserver' only when the requested object exists on a
3060 server in the .edu domain. Prefixing the domainname
3061 with '!' means the cache will be queried for objects
3064 NOTE: * Any number of domains may be given for a cache-host,
3065 either on the same or separate lines.
3066 * When multiple domains are given for a particular
3067 cache-host, the first matched domain is applied.
3068 * Cache hosts with no domain restrictions are queried
3070 * There are no defaults.
3071 * There is also a 'cache_peer_access' tag in the ACL
3075 NAME: cache_peer_access
3080 Similar to 'cache_peer_domain' but provides more flexibility by
3084 cache_peer_access cache-host allow|deny [!]aclname ...
3086 The syntax is identical to 'http_access' and the other lists of
3087 ACL elements. See the comments for 'http_access' below, or
3088 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3091 NAME: neighbor_type_domain
3092 TYPE: hostdomaintype
3094 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3097 Modify the cache_peer neighbor type when passing requests
3098 about specific domains to the peer.
3101 neighbor_type_domain neighbor parent|sibling domain domain ...
3104 cache_peer foo.example.com parent 3128 3130
3105 neighbor_type_domain foo.example.com sibling .au .de
3107 The above configuration treats all requests to foo.example.com as a
3108 parent proxy unless the request is for a .au or .de ccTLD domain name.
3111 NAME: dead_peer_timeout
3115 LOC: Config.Timeout.deadPeer
3117 This controls how long Squid waits to declare a peer cache
3118 as "dead." If there are no ICP replies received in this
3119 amount of time, Squid will declare the peer dead and not
3120 expect to receive any further ICP replies. However, it
3121 continues to send ICP queries, and will mark the peer as
3122 alive upon receipt of the first subsequent ICP reply.
3124 This timeout also affects when Squid expects to receive ICP
3125 replies from peers. If more than 'dead_peer' seconds have
3126 passed since the last ICP reply was received, Squid will not
3127 expect to receive an ICP reply on the next query. Thus, if
3128 your time between requests is greater than this timeout, you
3129 will see a lot of requests sent DIRECT to origin servers
3130 instead of to your parents.
3133 NAME: forward_max_tries
3136 LOC: Config.forward_max_tries
3138 Controls how many different forward paths Squid will try
3139 before giving up. See also forward_timeout.
3141 NOTE: connect_retries (default: none) can make each of these
3142 possible forwarding paths be tried multiple times.
3145 NAME: hierarchy_stoplist
3148 LOC: Config.hierarchy_stoplist
3150 A list of words which, if found in a URL, cause the object to
3151 be handled directly by this cache. In other words, use this
3152 to not query neighbor caches for certain objects. You may
3153 list this option multiple times.
3156 hierarchy_stoplist cgi-bin ?
3158 Note: never_direct overrides this option.
3162 MEMORY CACHE OPTIONS
3163 -----------------------------------------------------------------------------
3170 LOC: Config.memMaxSize
3172 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3173 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3174 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3175 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3177 'cache_mem' specifies the ideal amount of memory to be used
3179 * In-Transit objects
3181 * Negative-Cached objects
3183 Data for these objects are stored in 4 KB blocks. This
3184 parameter specifies the ideal upper limit on the total size of
3185 4 KB blocks allocated. In-Transit objects take the highest
3188 In-transit objects have priority over the others. When
3189 additional space is needed for incoming data, negative-cached
3190 and hot objects will be released. In other words, the
3191 negative-cached and hot objects will fill up any unused space
3192 not needed for in-transit objects.
3194 If circumstances require, this limit will be exceeded.
3195 Specifically, if your incoming request rate requires more than
3196 'cache_mem' of memory to hold in-transit objects, Squid will
3197 exceed this limit to satisfy the new requests. When the load
3198 decreases, blocks will be freed until the high-water mark is
3199 reached. Thereafter, blocks will be used to store hot
3202 If shared memory caching is enabled, Squid does not use the shared
3203 cache space for in-transit objects, but they still consume as much
3204 local memory as they need. For more details about the shared memory
3205 cache, see memory_cache_shared.
3208 NAME: maximum_object_size_in_memory
3212 LOC: Config.Store.maxInMemObjSize
3214 Objects greater than this size will not be attempted to kept in
3215 the memory cache. This should be set high enough to keep objects
3216 accessed frequently in memory to improve performance whilst low
3217 enough to keep larger objects from hoarding cache_mem.
3220 NAME: memory_cache_shared
3223 LOC: Config.memShared
3225 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3227 Controls whether the memory cache is shared among SMP workers.
3229 The shared memory cache is meant to occupy cache_mem bytes and replace
3230 the non-shared memory cache, although some entities may still be
3231 cached locally by workers for now (e.g., internal and in-transit
3232 objects may be served from a local memory cache even if shared memory
3233 caching is enabled).
3235 By default, the memory cache is shared if and only if all of the
3236 following conditions are satisfied: Squid runs in SMP mode with
3237 multiple workers, cache_mem is positive, and Squid environment
3238 supports required IPC primitives (e.g., POSIX shared memory segments
3239 and GCC-style atomic operations).
3241 To avoid blocking locks, shared memory uses opportunistic algorithms
3242 that do not guarantee that every cachable entity that could have been
3243 shared among SMP workers will actually be shared.
3245 Currently, entities exceeding 32KB in size cannot be shared.
3248 NAME: memory_cache_mode
3252 DEFAULT_DOC: Keep the most recently fetched objects in memory
3254 Controls which objects to keep in the memory cache (cache_mem)
3256 always Keep most recently fetched objects in memory (default)
3258 disk Only disk cache hits are kept in memory, which means
3259 an object must first be cached on disk and then hit
3260 a second time before cached in memory.
3262 network Only objects fetched from network is kept in memory
3265 NAME: memory_replacement_policy
3267 LOC: Config.memPolicy
3270 The memory replacement policy parameter determines which
3271 objects are purged from memory when memory space is needed.
3273 See cache_replacement_policy for details on algorithms.
3278 -----------------------------------------------------------------------------
3281 NAME: cache_replacement_policy
3283 LOC: Config.replPolicy
3286 The cache replacement policy parameter determines which
3287 objects are evicted (replaced) when disk space is needed.
3289 lru : Squid's original list based LRU policy
3290 heap GDSF : Greedy-Dual Size Frequency
3291 heap LFUDA: Least Frequently Used with Dynamic Aging
3292 heap LRU : LRU policy implemented using a heap
3294 Applies to any cache_dir lines listed below this directive.
3296 The LRU policies keeps recently referenced objects.
3298 The heap GDSF policy optimizes object hit rate by keeping smaller
3299 popular objects in cache so it has a better chance of getting a
3300 hit. It achieves a lower byte hit rate than LFUDA though since
3301 it evicts larger (possibly popular) objects.
3303 The heap LFUDA policy keeps popular objects in cache regardless of
3304 their size and thus optimizes byte hit rate at the expense of
3305 hit rate since one large, popular object will prevent many
3306 smaller, slightly less popular objects from being cached.
3308 Both policies utilize a dynamic aging mechanism that prevents
3309 cache pollution that can otherwise occur with frequency-based
3310 replacement policies.
3312 NOTE: if using the LFUDA replacement policy you should increase
3313 the value of maximum_object_size above its default of 4 MB to
3314 to maximize the potential byte hit rate improvement of LFUDA.
3316 For more information about the GDSF and LFUDA cache replacement
3317 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3318 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3321 NAME: minimum_object_size
3325 DEFAULT_DOC: no limit
3326 LOC: Config.Store.minObjectSize
3328 Objects smaller than this size will NOT be saved on disk. The
3329 value is specified in bytes, and the default is 0 KB, which
3330 means all responses can be stored.
3333 NAME: maximum_object_size
3337 LOC: Config.Store.maxObjectSize
3339 Set the default value for max-size parameter on any cache_dir.
3340 The value is specified in bytes, and the default is 4 MB.
3342 If you wish to get a high BYTES hit ratio, you should probably
3343 increase this (one 32 MB object hit counts for 3200 10KB
3346 If you wish to increase hit ratio more than you want to
3347 save bandwidth you should leave this low.
3349 NOTE: if using the LFUDA replacement policy you should increase
3350 this value to maximize the byte hit rate improvement of LFUDA!
3351 See cache_replacement_policy for a discussion of this policy.
3357 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3358 LOC: Config.cacheSwap
3361 cache_dir Type Directory-Name Fs-specific-data [options]
3363 You can specify multiple cache_dir lines to spread the
3364 cache among different disk partitions.
3366 Type specifies the kind of storage system to use. Only "ufs"
3367 is built by default. To enable any of the other storage systems
3368 see the --enable-storeio configure option.
3370 'Directory' is a top-level directory where cache swap
3371 files will be stored. If you want to use an entire disk
3372 for caching, this can be the mount-point directory.
3373 The directory must exist and be writable by the Squid
3374 process. Squid will NOT create this directory for you.
3376 In SMP configurations, cache_dir must not precede the workers option
3377 and should use configuration macros or conditionals to give each
3378 worker interested in disk caching a dedicated cache directory.
3381 ==== The ufs store type ====
3383 "ufs" is the old well-known Squid storage format that has always
3387 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3389 'Mbytes' is the amount of disk space (MB) to use under this
3390 directory. The default is 100 MB. Change this to suit your
3391 configuration. Do NOT put the size of your disk drive here.
3392 Instead, if you want Squid to use the entire disk drive,
3393 subtract 20% and use that value.
3395 'L1' is the number of first-level subdirectories which
3396 will be created under the 'Directory'. The default is 16.
3398 'L2' is the number of second-level subdirectories which
3399 will be created under each first-level directory. The default
3403 ==== The aufs store type ====
3405 "aufs" uses the same storage format as "ufs", utilizing
3406 POSIX-threads to avoid blocking the main Squid process on
3407 disk-I/O. This was formerly known in Squid as async-io.
3410 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3412 see argument descriptions under ufs above
3415 ==== The diskd store type ====
3417 "diskd" uses the same storage format as "ufs", utilizing a
3418 separate process to avoid blocking the main Squid process on
3422 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3424 see argument descriptions under ufs above
3426 Q1 specifies the number of unacknowledged I/O requests when Squid
3427 stops opening new files. If this many messages are in the queues,
3428 Squid won't open new files. Default is 64
3430 Q2 specifies the number of unacknowledged messages when Squid
3431 starts blocking. If this many messages are in the queues,
3432 Squid blocks until it receives some replies. Default is 72
3434 When Q1 < Q2 (the default), the cache directory is optimized
3435 for lower response time at the expense of a decrease in hit
3436 ratio. If Q1 > Q2, the cache directory is optimized for
3437 higher hit ratio at the expense of an increase in response
3441 ==== The rock store type ====
3444 cache_dir rock Directory-Name Mbytes [options]
3446 The Rock Store type is a database-style storage. All cached
3447 entries are stored in a "database" file, using fixed-size slots.
3448 A single entry occupies one or more slots.
3450 If possible, Squid using Rock Store creates a dedicated kid
3451 process called "disker" to avoid blocking Squid worker(s) on disk
3452 I/O. One disker kid is created for each rock cache_dir. Diskers
3453 are created only when Squid, running in daemon mode, has support
3454 for the IpcIo disk I/O module.
3456 swap-timeout=msec: Squid will not start writing a miss to or
3457 reading a hit from disk if it estimates that the swap operation
3458 will take more than the specified number of milliseconds. By
3459 default and when set to zero, disables the disk I/O time limit
3460 enforcement. Ignored when using blocking I/O module because
3461 blocking synchronous I/O does not allow Squid to estimate the
3462 expected swap wait time.
3464 max-swap-rate=swaps/sec: Artificially limits disk access using
3465 the specified I/O rate limit. Swap out requests that
3466 would cause the average I/O rate to exceed the limit are
3467 delayed. Individual swap in requests (i.e., hits or reads) are
3468 not delayed, but they do contribute to measured swap rate and
3469 since they are placed in the same FIFO queue as swap out
3470 requests, they may wait longer if max-swap-rate is smaller.
3471 This is necessary on file systems that buffer "too
3472 many" writes and then start blocking Squid and other processes
3473 while committing those writes to disk. Usually used together
3474 with swap-timeout to avoid excessive delays and queue overflows
3475 when disk demand exceeds available disk "bandwidth". By default
3476 and when set to zero, disables the disk I/O rate limit
3477 enforcement. Currently supported by IpcIo module only.
3479 slot-size=bytes: The size of a database "record" used for
3480 storing cached responses. A cached response occupies at least
3481 one slot and all database I/O is done using individual slots so
3482 increasing this parameter leads to more disk space waste while
3483 decreasing it leads to more disk I/O overheads. Should be a
3484 multiple of your operating system I/O page size. Defaults to
3485 16KBytes. A housekeeping header is stored with each slot and
3486 smaller slot-sizes will be rejected. The header is smaller than
3490 ==== COMMON OPTIONS ====
3492 no-store no new objects should be stored to this cache_dir.
3494 min-size=n the minimum object size in bytes this cache_dir
3495 will accept. It's used to restrict a cache_dir
3496 to only store large objects (e.g. AUFS) while
3497 other stores are optimized for smaller objects
3501 max-size=n the maximum object size in bytes this cache_dir
3503 The value in maximum_object_size directive sets
3504 the default unless more specific details are
3505 available (ie a small store capacity).
3507 Note: To make optimal use of the max-size limits you should order
3508 the cache_dir lines with the smallest max-size value first.
3512 # Uncomment and adjust the following to add a disk cache directory.
3513 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3517 NAME: store_dir_select_algorithm
3519 LOC: Config.store_dir_select_algorithm
3522 How Squid selects which cache_dir to use when the response
3523 object will fit into more than one.
3525 Regardless of which algorithm is used the cache_dir min-size
3526 and max-size parameters are obeyed. As such they can affect
3527 the selection algorithm by limiting the set of considered
3534 This algorithm is suited to caches with similar cache_dir
3535 sizes and disk speeds.
3537 The disk with the least I/O pending is selected.
3538 When there are multiple disks with the same I/O load ranking
3539 the cache_dir with most available capacity is selected.
3541 When a mix of cache_dir sizes are configured the faster disks
3542 have a naturally lower I/O loading and larger disks have more
3543 capacity. So space used to store objects and data throughput
3544 may be very unbalanced towards larger disks.
3549 This algorithm is suited to caches with unequal cache_dir
3552 Each cache_dir is selected in a rotation. The next suitable
3555 Available cache_dir capacity is only considered in relation
3556 to whether the object will fit and meets the min-size and
3557 max-size parameters.
3559 Disk I/O loading is only considered to prevent overload on slow
3560 disks. This algorithm does not spread objects by size, so any
3561 I/O loading per-disk may appear very unbalanced and volatile.
3565 NAME: max_open_disk_fds
3567 LOC: Config.max_open_disk_fds
3569 DEFAULT_DOC: no limit
3571 To avoid having disk as the I/O bottleneck Squid can optionally
3572 bypass the on-disk cache if more than this amount of disk file
3573 descriptors are open.
3575 A value of 0 indicates no limit.
3578 NAME: cache_swap_low
3579 COMMENT: (percent, 0-100)
3582 LOC: Config.Swap.lowWaterMark
3584 The low-water mark for cache object replacement.
3585 Replacement begins when the swap (disk) usage is above the
3586 low-water mark and attempts to maintain utilization near the
3587 low-water mark. As swap utilization gets close to high-water
3588 mark object eviction becomes more aggressive. If utilization is
3589 close to the low-water mark less replacement is done each time.
3591 Defaults are 90% and 95%. If you have a large cache, 5% could be
3592 hundreds of MB. If this is the case you may wish to set these
3593 numbers closer together.
3595 See also cache_swap_high
3598 NAME: cache_swap_high
3599 COMMENT: (percent, 0-100)
3602 LOC: Config.Swap.highWaterMark
3604 The high-water mark for cache object replacement.
3605 Replacement begins when the swap (disk) usage is above the
3606 low-water mark and attempts to maintain utilization near the
3607 low-water mark. As swap utilization gets close to high-water
3608 mark object eviction becomes more aggressive. If utilization is
3609 close to the low-water mark less replacement is done each time.
3611 Defaults are 90% and 95%. If you have a large cache, 5% could be
3612 hundreds of MB. If this is the case you may wish to set these
3613 numbers closer together.
3615 See also cache_swap_low
3620 -----------------------------------------------------------------------------
3627 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
3631 logformat <name> <format specification>
3633 Defines an access log format.
3635 The <format specification> is a string with embedded % format codes
3637 % format codes all follow the same basic structure where all but
3638 the formatcode is optional. Output strings are automatically escaped
3639 as required according to their context and the output format
3640 modifiers are usually not needed, but can be specified if an explicit
3641 output format is desired.
3643 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3645 " output in quoted string format
3646 [ output in squid text log format as used by log_mime_hdrs
3647 # output in URL quoted format
3652 width minimum and/or maximum field width:
3653 [width_min][.width_max]
3654 When minimum starts with 0, the field is zero-padded.
3655 String values exceeding maximum width are truncated.
3657 {arg} argument such as header name etc
3661 % a literal % character
3662 sn Unique sequence number per log line entry
3663 err_code The ID of an error response served by Squid or
3664 a similar internal error identifier.
3665 err_detail Additional err_code-dependent error information.
3666 note The annotation specified by the argument. Also
3667 logs the adaptation meta headers set by the
3668 adaptation_meta configuration parameter.
3669 If no argument given all annotations logged.
3670 The argument may include a separator to use with
3673 By default, multiple note values are separated with ","
3674 and multiple notes are separated with "\r\n".
3675 When logging named notes with %{name}note, the
3676 explicitly configured separator is used between note
3677 values. When logging all notes with %note, the
3678 explicitly configured separator is used between
3679 individual notes. There is currently no way to
3680 specify both value and notes separators when logging
3681 all notes with %note.
3683 Connection related format codes:
3685 >a Client source IP address
3687 >p Client source port
3688 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3689 >la Local IP address the client connected to
3690 >lp Local port number the client connected to
3691 >qos Client connection TOS/DSCP value set by Squid
3692 >nfmark Client connection netfilter mark set by Squid
3694 la Local listening IP address the client connection was connected to.
3695 lp Local listening port number the client connection was connected to.
3697 <a Server IP address of the last server or peer connection
3698 <A Server FQDN or peer name
3699 <p Server port number of the last server or peer connection
3700 <la Local IP address of the last server or peer connection
3701 <lp Local port number of the last server or peer connection
3702 <qos Server connection TOS/DSCP value set by Squid
3703 <nfmark Server connection netfilter mark set by Squid
3705 Time related format codes:
3707 ts Seconds since epoch
3708 tu subsecond time (milliseconds)
3709 tl Local time. Optional strftime format argument
3710 default %d/%b/%Y:%H:%M:%S %z
3711 tg GMT time. Optional strftime format argument
3712 default %d/%b/%Y:%H:%M:%S %z
3713 tr Response time (milliseconds)
3714 dt Total time spent making DNS lookups (milliseconds)
3715 tS Approximate master transaction start time in
3716 <full seconds since epoch>.<fractional seconds> format.
3717 Currently, Squid considers the master transaction
3718 started when a complete HTTP request header initiating
3719 the transaction is received from the client. This is
3720 the same value that Squid uses to calculate transaction
3721 response time when logging %tr to access.log. Currently,
3722 Squid uses millisecond resolution for %tS values,
3723 similar to the default access.log "current time" field
3726 Access Control related format codes:
3728 et Tag returned by external acl
3729 ea Log string returned by external acl
3730 un User name (any available)
3731 ul User name from authentication
3732 ue User name from external acl helper
3733 ui User name from ident
3734 us User name from SSL
3735 credentials Client credentials. The exact meaning depends on
3736 the authentication scheme: For Basic authentication,
3737 it is the password; for Digest, the realm sent by the
3738 client; for NTLM and Negotiate, the client challenge
3739 or client credentials prefixed with "YR " or "KK ".
3741 HTTP related format codes:
3745 [http::]rm Request method (GET/POST etc)
3746 [http::]>rm Request method from client
3747 [http::]<rm Request method sent to server or peer
3748 [http::]ru Request URL from client (historic, filtered for logging)
3749 [http::]>ru Request URL from client
3750 [http::]<ru Request URL sent to server or peer
3751 [http::]>rs Request URL scheme from client
3752 [http::]<rs Request URL scheme sent to server or peer
3753 [http::]>rd Request URL domain from client
3754 [http::]>rd Request URL domain sent to server or peer
3755 [http::]>rP Request URL port from client
3756 [http::]<rP Request URL port sent to server or peer
3757 [http::]rp Request URL path excluding hostname
3758 [http::]>rp Request URL path excluding hostname from client
3759 [http::]<rp Request URL path excluding hostname sent to server or peer
3760 [http::]rv Request protocol version
3761 [http::]>rv Request protocol version from client
3762 [http::]<rv Request protocol version sent to server or peer
3764 [http::]>h Original received request header.
3765 Usually differs from the request header sent by
3766 Squid, although most fields are often preserved.
3767 Accepts optional header field name/value filter
3768 argument using name[:[separator]element] format.
3769 [http::]>ha Received request header after adaptation and
3770 redirection (pre-cache REQMOD vectoring point).
3771 Usually differs from the request header sent by
3772 Squid, although most fields are often preserved.
3773 Optional header name argument as for >h
3778 [http::]<Hs HTTP status code received from the next hop
3779 [http::]>Hs HTTP status code sent to the client
3781 [http::]<h Reply header. Optional header name argument
3784 [http::]mt MIME content type
3789 [http::]st Total size of request + reply traffic with client
3790 [http::]>st Total size of request received from client.
3791 Excluding chunked encoding bytes.
3792 [http::]<st Total size of reply sent to client (after adaptation)
3794 [http::]>sh Size of request headers received from client
3795 [http::]<sh Size of reply headers sent to client (after adaptation)
3797 [http::]<sH Reply high offset sent
3798 [http::]<sS Upstream object size
3800 [http::]<bs Number of HTTP-equivalent message body bytes
3801 received from the next hop, excluding chunked
3802 transfer encoding and control messages.
3803 Generated FTP/Gopher listings are treated as
3809 [http::]<pt Peer response time in milliseconds. The timer starts
3810 when the last request byte is sent to the next hop
3811 and stops when the last response byte is received.
3812 [http::]<tt Total server-side time in milliseconds. The timer
3813 starts with the first connect request (or write I/O)
3814 sent to the first selected peer. The timer stops
3815 with the last I/O with the last peer.
3817 Squid handling related format codes:
3819 Ss Squid request status (TCP_MISS etc)
3820 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3822 SSL-related format codes:
3824 ssl::bump_mode SslBump decision for the transaction:
3826 For CONNECT requests that initiated bumping of
3827 a connection and for any request received on
3828 an already bumped connection, Squid logs the
3829 corresponding SslBump mode ("server-first" or
3830 "client-first"). See the ssl_bump option for
3831 more information about these modes.
3833 A "none" token is logged for requests that
3834 triggered "ssl_bump" ACL evaluation matching
3835 either a "none" rule or no rules at all.
3837 In all other cases, a single dash ("-") is
3840 If ICAP is enabled, the following code becomes available (as
3841 well as ICAP log codes documented with the icap_log option):
3843 icap::tt Total ICAP processing time for the HTTP
3844 transaction. The timer ticks when ICAP
3845 ACLs are checked and when ICAP
3846 transaction is in progress.
3848 If adaptation is enabled the following three codes become available:
3850 adapt::<last_h The header of the last ICAP response or
3851 meta-information from the last eCAP
3852 transaction related to the HTTP transaction.
3853 Like <h, accepts an optional header name
3856 adapt::sum_trs Summed adaptation transaction response
3857 times recorded as a comma-separated list in
3858 the order of transaction start time. Each time
3859 value is recorded as an integer number,
3860 representing response time of one or more
3861 adaptation (ICAP or eCAP) transaction in
3862 milliseconds. When a failed transaction is
3863 being retried or repeated, its time is not
3864 logged individually but added to the
3865 replacement (next) transaction. See also:
3868 adapt::all_trs All adaptation transaction response times.
3869 Same as adaptation_strs but response times of
3870 individual transactions are never added
3871 together. Instead, all transaction response
3872 times are recorded individually.
3874 You can prefix adapt::*_trs format codes with adaptation
3875 service name in curly braces to record response time(s) specific
3876 to that service. For example: %{my_service}adapt::sum_trs
3878 If SSL is enabled, the following formating codes become available:
3880 %ssl::>cert_subject The Subject field of the received client
3881 SSL certificate or a dash ('-') if Squid has
3882 received an invalid/malformed certificate or
3883 no certificate at all. Consider encoding the
3884 logged value because Subject often has spaces.
3886 %ssl::>cert_issuer The Issuer field of the received client
3887 SSL certificate or a dash ('-') if Squid has
3888 received an invalid/malformed certificate or
3889 no certificate at all. Consider encoding the
3890 logged value because Issuer often has spaces.
3892 The default formats available (which do not need re-defining) are:
3894 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3895 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3896 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3897 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3898 logformat useragent %>a [%tl] "%{User-Agent}>h"
3900 NOTE: When the log_mime_hdrs directive is set to ON.
3901 The squid, common and combined formats have a safely encoded copy
3902 of the mime headers appended to each line within a pair of brackets.
3904 NOTE: The common and combined formats are not quite true to the Apache definition.
3905 The logs from Squid contain an extra status and hierarchy code appended.
3909 NAME: access_log cache_access_log
3911 LOC: Config.Log.accesslogs
3912 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3914 Configures whether and how Squid logs HTTP and ICP transactions.
3915 If access logging is enabled, a single line is logged for every
3916 matching HTTP or ICP request. The recommended directive formats are:
3918 access_log <module>:<place> [option ...] [acl acl ...]
3919 access_log none [acl acl ...]
3921 The following directive format is accepted but may be deprecated:
3922 access_log <module>:<place> [<logformat name> [acl acl ...]]
3924 In most cases, the first ACL name must not contain the '=' character
3925 and should not be equal to an existing logformat name. You can always
3926 start with an 'all' ACL to work around those restrictions.
3928 Will log to the specified module:place using the specified format (which
3929 must be defined in a logformat directive) those entries which match
3930 ALL the acl's specified (which must be defined in acl clauses).
3931 If no acl is specified, all requests will be logged to this destination.
3933 ===== Available options for the recommended directive format =====
3935 logformat=name Names log line format (either built-in or
3936 defined by a logformat directive). Defaults
3939 buffer-size=64KB Defines approximate buffering limit for log
3940 records (see buffered_logs). Squid should not
3941 keep more than the specified size and, hence,
3942 should flush records before the buffer becomes
3943 full to avoid overflows under normal
3944 conditions (the exact flushing algorithm is
3945 module-dependent though). The on-error option
3946 controls overflow handling.
3948 on-error=die|drop Defines action on unrecoverable errors. The
3949 'drop' action ignores (i.e., does not log)
3950 affected log records. The default 'die' action
3951 kills the affected worker. The drop action
3952 support has not been tested for modules other
3955 ===== Modules Currently available =====
3957 none Do not log any requests matching these ACL.
3958 Do not specify Place or logformat name.
3960 stdio Write each log line to disk immediately at the completion of
3962 Place: the filename and path to be written.
3964 daemon Very similar to stdio. But instead of writing to disk the log
3965 line is passed to a daemon helper for asychronous handling instead.
3966 Place: varies depending on the daemon.
3968 log_file_daemon Place: the file name and path to be written.
3970 syslog To log each request via syslog facility.
3971 Place: The syslog facility and priority level for these entries.
3972 Place Format: facility.priority
3974 where facility could be any of:
3975 authpriv, daemon, local0 ... local7 or user.
3977 And priority could be any of:
3978 err, warning, notice, info, debug.
3980 udp To send each log line as text data to a UDP receiver.
3981 Place: The destination host name or IP and port.
3982 Place Format: //host:port
3984 tcp To send each log line as text data to a TCP receiver.
3985 Lines may be accumulated before sending (see buffered_logs).
3986 Place: The destination host name or IP and port.
3987 Place Format: //host:port
3990 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3996 LOC: Config.Log.icaplogs
3999 ICAP log files record ICAP transaction summaries, one line per
4002 The icap_log option format is:
4003 icap_log <filepath> [<logformat name> [acl acl ...]]
4004 icap_log none [acl acl ...]]
4006 Please see access_log option documentation for details. The two
4007 kinds of logs share the overall configuration approach and many
4010 ICAP processing of a single HTTP message or transaction may
4011 require multiple ICAP transactions. In such cases, multiple
4012 ICAP transaction log lines will correspond to a single access
4015 ICAP log uses logformat codes that make sense for an ICAP
4016 transaction. Header-related codes are applied to the HTTP header
4017 embedded in an ICAP server response, with the following caveats:
4018 For REQMOD, there is no HTTP response header unless the ICAP
4019 server performed request satisfaction. For RESPMOD, the HTTP
4020 request header is the header sent to the ICAP server. For
4021 OPTIONS, there are no HTTP headers.
4023 The following format codes are also available for ICAP logs:
4025 icap::<A ICAP server IP address. Similar to <A.
4027 icap::<service_name ICAP service name from the icap_service
4028 option in Squid configuration file.
4030 icap::ru ICAP Request-URI. Similar to ru.
4032 icap::rm ICAP request method (REQMOD, RESPMOD, or
4033 OPTIONS). Similar to existing rm.
4035 icap::>st Bytes sent to the ICAP server (TCP payload
4036 only; i.e., what Squid writes to the socket).
4038 icap::<st Bytes received from the ICAP server (TCP
4039 payload only; i.e., what Squid reads from
4042 icap::<bs Number of message body bytes received from the
4043 ICAP server. ICAP message body, if any, usually
4044 includes encapsulated HTTP message headers and
4045 possibly encapsulated HTTP message body. The
4046 HTTP body part is dechunked before its size is
4049 icap::tr Transaction response time (in
4050 milliseconds). The timer starts when
4051 the ICAP transaction is created and
4052 stops when the transaction is completed.
4055 icap::tio Transaction I/O time (in milliseconds). The
4056 timer starts when the first ICAP request
4057 byte is scheduled for sending. The timers
4058 stops when the last byte of the ICAP response
4061 icap::to Transaction outcome: ICAP_ERR* for all
4062 transaction errors, ICAP_OPT for OPTION
4063 transactions, ICAP_ECHO for 204
4064 responses, ICAP_MOD for message
4065 modification, and ICAP_SAT for request
4066 satisfaction. Similar to Ss.
4068 icap::Hs ICAP response status code. Similar to Hs.
4070 icap::>h ICAP request header(s). Similar to >h.
4072 icap::<h ICAP response header(s). Similar to <h.
4074 The default ICAP log format, which can be used without an explicit
4075 definition, is called icap_squid:
4077 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4079 See also: logformat, log_icap, and %adapt::<last_h
4082 NAME: logfile_daemon
4084 DEFAULT: @DEFAULT_LOGFILED@
4085 LOC: Log::TheConfig.logfile_daemon
4087 Specify the path to the logfile-writing daemon. This daemon is
4088 used to write the access and store logs, if configured.
4090 Squid sends a number of commands to the log daemon:
4091 L<data>\n - logfile data
4096 r<n>\n - set rotate count to <n>
4097 b<n>\n - 1 = buffer output, 0 = don't buffer output
4099 No responses is expected.
4105 Remove this line. Use acls with access_log directives to control access logging
4111 Remove this line. Use acls with icap_log directives to control icap logging
4114 NAME: stats_collection
4116 LOC: Config.accessList.stats_collection
4118 DEFAULT_DOC: Allow logging for all transactions.
4119 COMMENT: allow|deny acl acl...
4121 This options allows you to control which requests gets accounted
4122 in performance counters.
4124 This clause only supports fast acl types.
4125 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4128 NAME: cache_store_log
4131 LOC: Config.Log.store
4133 Logs the activities of the storage manager. Shows which
4134 objects are ejected from the cache, and which objects are
4135 saved and for how long.
4136 There are not really utilities to analyze this data, so you can safely
4137 disable it (the default).
4139 Store log uses modular logging outputs. See access_log for the list
4140 of modules supported.
4143 cache_store_log stdio:@DEFAULT_STORE_LOG@
4144 cache_store_log daemon:@DEFAULT_STORE_LOG@
4147 NAME: cache_swap_state cache_swap_log
4149 LOC: Config.Log.swap
4151 DEFAULT_DOC: Store the journal inside its cache_dir
4153 Location for the cache "swap.state" file. This index file holds
4154 the metadata of objects saved on disk. It is used to rebuild
4155 the cache during startup. Normally this file resides in each
4156 'cache_dir' directory, but you may specify an alternate
4157 pathname here. Note you must give a full filename, not just
4158 a directory. Since this is the index for the whole object
4159 list you CANNOT periodically rotate it!
4161 If %s can be used in the file name it will be replaced with a
4162 a representation of the cache_dir name where each / is replaced
4163 with '.'. This is needed to allow adding/removing cache_dir
4164 lines when cache_swap_log is being used.
4166 If have more than one 'cache_dir', and %s is not used in the name
4167 these swap logs will have names such as:
4173 The numbered extension (which is added automatically)
4174 corresponds to the order of the 'cache_dir' lines in this
4175 configuration file. If you change the order of the 'cache_dir'
4176 lines in this file, these index files will NOT correspond to
4177 the correct 'cache_dir' entry (unless you manually rename
4178 them). We recommend you do NOT use this option. It is
4179 better to keep these index files in each 'cache_dir' directory.
4182 NAME: logfile_rotate
4185 LOC: Config.Log.rotateNumber
4187 Specifies the number of logfile rotations to make when you
4188 type 'squid -k rotate'. The default is 10, which will rotate
4189 with extensions 0 through 9. Setting logfile_rotate to 0 will
4190 disable the file name rotation, but the logfiles are still closed
4191 and re-opened. This will enable you to rename the logfiles
4192 yourself just before sending the rotate signal.
4194 Note, the 'squid -k rotate' command normally sends a USR1
4195 signal to the running squid process. In certain situations
4196 (e.g. on Linux with Async I/O), USR1 is used for other
4197 purposes, so -k rotate uses another signal. It is best to get
4198 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4201 Note, from Squid-3.1 this option is only a default for cache.log,
4202 that log can be rotated separately by using debug_options.
4205 NAME: emulate_httpd_log
4208 Replace this with an access_log directive using the format 'common' or 'combined'.
4211 NAME: log_ip_on_direct
4214 Remove this option from your config. To log server or peer names use %<A in the log format.
4219 DEFAULT: @DEFAULT_MIME_TABLE@
4220 LOC: Config.mimeTablePathname
4222 Path to Squid's icon configuration file.
4224 You shouldn't need to change this, but the default file contains
4225 examples and formatting information if you do.
4231 LOC: Config.onoff.log_mime_hdrs
4234 The Cache can record both the request and the response MIME
4235 headers for each HTTP transaction. The headers are encoded
4236 safely and will appear as two bracketed fields at the end of
4237 the access log (for either the native or httpd-emulated log
4238 formats). To enable this logging set log_mime_hdrs to 'on'.
4244 Replace this with an access_log directive using the format 'useragent'.
4247 NAME: referer_log referrer_log
4250 Replace this with an access_log directive using the format 'referrer'.
4255 DEFAULT: @DEFAULT_PID_FILE@
4256 LOC: Config.pidFilename
4258 A filename to write the process-id to. To disable, enter "none".
4264 Remove this option from your config. To log FQDN use %>A in the log format.
4267 NAME: client_netmask
4269 LOC: Config.Addrs.client_netmask
4271 DEFAULT_DOC: Log full client IP address
4273 A netmask for client addresses in logfiles and cachemgr output.
4274 Change this to protect the privacy of your cache clients.
4275 A netmask of 255.255.255.0 will log all IP's in that range with
4276 the last digit set to '0'.
4282 Use a regular access.log with ACL limiting it to MISS events.
4285 NAME: strip_query_terms
4287 LOC: Config.onoff.strip_query_terms
4290 By default, Squid strips query terms from requested URLs before
4291 logging. This protects your user's privacy and reduces log size.
4293 When investigating HIT/MISS or other caching behaviour you
4294 will need to disable this to see the full URL used by Squid.
4301 LOC: Config.onoff.buffered_logs
4303 Whether to write/send access_log records ASAP or accumulate them and
4304 then write/send them in larger chunks. Buffering may improve
4305 performance because it decreases the number of I/Os. However,
4306 buffering increases the delay before log records become available to
4307 the final recipient (e.g., a disk file or logging daemon) and,
4308 hence, increases the risk of log records loss.
4310 Note that even when buffered_logs are off, Squid may have to buffer
4311 records if it cannot write/send them immediately due to pending I/Os
4312 (e.g., the I/O writing the previous log record) or connectivity loss.
4314 Currently honored by 'daemon' and 'tcp' access_log modules only.
4317 NAME: netdb_filename
4319 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4320 LOC: Config.netdbFilename
4323 Where Squid stores it's netdb journal.
4324 When enabled this journal preserves netdb state between restarts.
4326 To disable, enter "none".
4330 OPTIONS FOR TROUBLESHOOTING
4331 -----------------------------------------------------------------------------
4336 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4337 LOC: Debug::cache_log
4339 Squid administrative logging file.
4341 This is where general information about Squid behavior goes. You can
4342 increase the amount of data logged to this file and how often it is
4343 rotated with "debug_options"
4349 DEFAULT_DOC: Log all critical and important messages.
4350 LOC: Debug::debugOptions
4352 Logging options are set as section,level where each source file
4353 is assigned a unique section. Lower levels result in less
4354 output, Full debugging (level 9) can result in a very large
4355 log file, so be careful.
4357 The magic word "ALL" sets debugging levels for all sections.
4358 The default is to run with "ALL,1" to record important warnings.
4360 The rotate=N option can be used to keep more or less of these logs
4361 than would otherwise be kept by logfile_rotate.
4362 For most uses a single log should be enough to monitor current
4363 events affecting Squid.
4368 LOC: Config.coredump_dir
4369 DEFAULT_IF_NONE: none
4370 DEFAULT_DOC: Use the directory from where Squid was started.
4372 By default Squid leaves core files in the directory from where
4373 it was started. If you set 'coredump_dir' to a directory
4374 that exists, Squid will chdir() to that directory at startup
4375 and coredump files will be left there.
4379 # Leave coredumps in the first cache dir
4380 coredump_dir @DEFAULT_SWAP_DIR@
4386 OPTIONS FOR FTP GATEWAYING
4387 -----------------------------------------------------------------------------
4393 LOC: Config.Ftp.anon_user
4395 If you want the anonymous login password to be more informative
4396 (and enable the use of picky FTP servers), set this to something
4397 reasonable for your domain, like wwwuser@somewhere.net
4399 The reason why this is domainless by default is the
4400 request can be made on the behalf of a user in any domain,
4401 depending on how the cache is used.
4402 Some FTP server also validate the email address is valid
4403 (for example perl.com).
4409 LOC: Config.Ftp.passive
4411 If your firewall does not allow Squid to use passive
4412 connections, turn off this option.
4414 Use of ftp_epsv_all option requires this to be ON.
4420 LOC: Config.Ftp.epsv_all
4422 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4424 NATs may be able to put the connection on a "fast path" through the
4425 translator, as the EPRT command will never be used and therefore,
4426 translation of the data portion of the segments will never be needed.
4428 When a client only expects to do two-way FTP transfers this may be
4430 If squid finds that it must do a three-way FTP transfer after issuing
4431 an EPSV ALL command, the FTP session will fail.
4433 If you have any doubts about this option do not use it.
4434 Squid will nicely attempt all other connection methods.
4436 Requires ftp_passive to be ON (default) for any effect.
4442 LOC: Config.accessList.ftp_epsv
4444 FTP Protocol extensions permit the use of a special "EPSV" command.
4446 NATs may be able to put the connection on a "fast path" through the
4447 translator using EPSV, as the EPRT command will never be used
4448 and therefore, translation of the data portion of the segments
4449 will never be needed.
4451 EPSV is often required to interoperate with FTP servers on IPv6
4452 networks. On the other hand, it may break some IPv4 servers.
4454 By default, EPSV may try EPSV with any FTP server. To fine tune
4455 that decision, you may restrict EPSV to certain clients or servers
4458 ftp_epsv allow|deny al1 acl2 ...
4460 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4462 Only fast ACLs are supported.
4463 Requires ftp_passive to be ON (default) for any effect.
4469 LOC: Config.Ftp.eprt
4471 FTP Protocol extensions permit the use of a special "EPRT" command.
4473 This extension provides a protocol neutral alternative to the
4474 IPv4-only PORT command. When supported it enables active FTP data
4475 channels over IPv6 and efficient NAT handling.
4477 Turning this OFF will prevent EPRT being attempted and will skip
4478 straight to using PORT for IPv4 servers.
4480 Some devices are known to not handle this extension correctly and
4481 may result in crashes. Devices which suport EPRT enough to fail
4482 cleanly will result in Squid attempting PORT anyway. This directive
4483 should only be disabled when EPRT results in device failures.
4485 WARNING: Doing so will convert Squid back to the old behavior with all
4486 the related problems with external NAT devices/layers and IPv4-only FTP.
4489 NAME: ftp_sanitycheck
4492 LOC: Config.Ftp.sanitycheck
4494 For security and data integrity reasons Squid by default performs
4495 sanity checks of the addresses of FTP data connections ensure the
4496 data connection is to the requested server. If you need to allow
4497 FTP connections to servers using another IP address for the data
4498 connection turn this off.
4501 NAME: ftp_telnet_protocol
4504 LOC: Config.Ftp.telnet
4506 The FTP protocol is officially defined to use the telnet protocol
4507 as transport channel for the control connection. However, many
4508 implementations are broken and does not respect this aspect of
4511 If you have trouble accessing files with ASCII code 255 in the
4512 path or similar problems involving this ASCII code you can
4513 try setting this directive to off. If that helps, report to the
4514 operator of the FTP server in question that their FTP server
4515 is broken and does not follow the FTP standard.
4519 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4520 -----------------------------------------------------------------------------
4525 DEFAULT: @DEFAULT_DISKD@
4526 LOC: Config.Program.diskd
4528 Specify the location of the diskd executable.
4529 Note this is only useful if you have compiled in
4530 diskd as one of the store io modules.
4533 NAME: unlinkd_program
4536 DEFAULT: @DEFAULT_UNLINKD@
4537 LOC: Config.Program.unlinkd
4539 Specify the location of the executable for file deletion process.
4542 NAME: pinger_program
4544 DEFAULT: @DEFAULT_PINGER@
4545 LOC: Config.pinger.program
4548 Specify the location of the executable for the pinger process.
4554 LOC: Config.pinger.enable
4557 Control whether the pinger is active at run-time.
4558 Enables turning ICMP pinger on and off with a simple
4559 squid -k reconfigure.
4564 OPTIONS FOR URL REWRITING
4565 -----------------------------------------------------------------------------
4568 NAME: url_rewrite_program redirect_program
4570 LOC: Config.Program.redirect
4573 Specify the location of the executable URL rewriter to use.
4574 Since they can perform almost any function there isn't one included.
4576 For each requested URL, the rewriter will receive on line with the format
4578 [channel-ID <SP>] URL [<SP> extras]<NL>
4581 After processing the request the helper must reply using the following format:
4583 [channel-ID <SP>] result [<SP> kv-pairs]
4585 The result code can be:
4587 OK status=30N url="..."
4588 Redirect the URL to the one supplied in 'url='.
4589 'status=' is optional and contains the status code to send
4590 the client in Squids HTTP response. It must be one of the
4591 HTTP redirect status codes: 301, 302, 303, 307, 308.
4592 When no status is given Squid will use 302.
4594 OK rewrite-url="..."
4595 Rewrite the URL to the one supplied in 'rewrite-url='.
4596 The new URL is fetched directly by Squid and returned to
4597 the client as the response to its request.
4600 When neither of url= and rewrite-url= are sent Squid does
4604 Do not change the URL.
4607 An internal error occurred in the helper, preventing
4608 a result being identified. The 'message=' key name is
4609 reserved for delivering a log message.
4612 In the future, the interface protocol will be extended with
4613 key=value pairs ("kv-pairs" shown above). Helper programs
4614 should be prepared to receive and possibly ignore additional
4615 whitespace-separated tokens on each input line.
4617 When using the concurrency= option the protocol is changed by
4618 introducing a query channel tag in front of the request/response.
4619 The query channel tag is a number between 0 and concurrency-1.
4620 This value must be echoed back unchanged to Squid as the first part
4621 of the response relating to its request.
4623 WARNING: URL re-writing ability should be avoided whenever possible.
4624 Use the URL redirect form of response instead.
4626 Re-write creates a difference in the state held by the client
4627 and server. Possibly causing confusion when the server response
4628 contains snippets of its view state. Embeded URLs, response
4629 and content Location headers, etc. are not re-written by this
4632 By default, a URL rewriter is not used.
4635 NAME: url_rewrite_children redirect_children
4636 TYPE: HelperChildConfig
4637 DEFAULT: 20 startup=0 idle=1 concurrency=0
4638 LOC: Config.redirectChildren
4640 The maximum number of redirector processes to spawn. If you limit
4641 it too few Squid will have to wait for them to process a backlog of
4642 URLs, slowing it down. If you allow too many they will use RAM
4643 and other system resources noticably.
4645 The startup= and idle= options allow some measure of skew in your
4650 Sets a minimum of how many processes are to be spawned when Squid
4651 starts or reconfigures. When set to zero the first request will
4652 cause spawning of the first child process to handle it.
4654 Starting too few will cause an initial slowdown in traffic as Squid
4655 attempts to simultaneously spawn enough processes to cope.
4659 Sets a minimum of how many processes Squid is to try and keep available
4660 at all times. When traffic begins to rise above what the existing
4661 processes can handle this many more will be spawned up to the maximum
4662 configured. A minimum setting of 1 is required.
4666 The number of requests each redirector helper can handle in
4667 parallel. Defaults to 0 which indicates the redirector
4668 is a old-style single threaded redirector.
4670 When this directive is set to a value >= 1 then the protocol
4671 used to communicate with the helper is modified to include
4672 an ID in front of the request/response. The ID from the request
4673 must be echoed back with the response to that request.
4676 NAME: url_rewrite_host_header redirect_rewrites_host_header
4679 LOC: Config.onoff.redir_rewrites_host
4681 To preserve same-origin security policies in browsers and
4682 prevent Host: header forgery by redirectors Squid rewrites
4683 any Host: header in redirected requests.
4685 If you are running an accelerator this may not be a wanted
4686 effect of a redirector. This directive enables you disable
4687 Host: alteration in reverse-proxy traffic.
4689 WARNING: Entries are cached on the result of the URL rewriting
4690 process, so be careful if you have domain-virtual hosts.
4692 WARNING: Squid and other software verifies the URL and Host
4693 are matching, so be careful not to relay through other proxies
4694 or inspecting firewalls with this disabled.
4697 NAME: url_rewrite_access redirector_access
4700 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4701 LOC: Config.accessList.redirector
4703 If defined, this access list specifies which requests are
4704 sent to the redirector processes.
4706 This clause supports both fast and slow acl types.
4707 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4710 NAME: url_rewrite_bypass redirector_bypass
4712 LOC: Config.onoff.redirector_bypass
4715 When this is 'on', a request will not go through the
4716 redirector if all the helpers are busy. If this is 'off'
4717 and the redirector queue grows too large, Squid will exit
4718 with a FATAL error and ask you to increase the number of
4719 redirectors. You should only enable this if the redirectors
4720 are not critical to your caching system. If you use
4721 redirectors for access control, and you enable this option,
4722 users may have access to pages they should not
4723 be allowed to request.
4726 NAME: url_rewrite_extras
4727 TYPE: TokenOrQuotedString
4728 LOC: Config.redirector_extras
4729 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4731 Specifies a string to be append to request line format for the
4732 rewriter helper. "Quoted" format values may contain spaces and
4733 logformat %macros. In theory, any logformat %macro can be used.
4734 In practice, a %macro expands as a dash (-) if the helper request is
4735 sent before the required macro information is available to Squid.
4739 OPTIONS FOR STORE ID
4740 -----------------------------------------------------------------------------
4743 NAME: store_id_program storeurl_rewrite_program
4745 LOC: Config.Program.store_id
4748 Specify the location of the executable StoreID helper to use.
4749 Since they can perform almost any function there isn't one included.
4751 For each requested URL, the helper will receive one line with the format
4753 [channel-ID <SP>] URL [<SP> extras]<NL>
4756 After processing the request the helper must reply using the following format:
4758 [channel-ID <SP>] result [<SP> kv-pairs]
4760 The result code can be:
4763 Use the StoreID supplied in 'store-id='.
4766 The default is to use HTTP request URL as the store ID.
4769 An internal error occured in the helper, preventing
4770 a result being identified.
4773 Helper programs should be prepared to receive and possibly ignore
4774 additional whitespace-separated tokens on each input line.
4776 When using the concurrency= option the protocol is changed by
4777 introducing a query channel tag in front of the request/response.
4778 The query channel tag is a number between 0 and concurrency-1.
4779 This value must be echoed back unchanged to Squid as the first part
4780 of the response relating to its request.
4782 NOTE: when using StoreID refresh_pattern will apply to the StoreID
4783 returned from the helper and not the URL.
4785 WARNING: Wrong StoreID value returned by a careless helper may result
4786 in the wrong cached response returned to the user.
4788 By default, a StoreID helper is not used.
4791 NAME: store_id_extras
4792 TYPE: TokenOrQuotedString
4793 LOC: Config.storeId_extras
4794 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4796 Specifies a string to be append to request line format for the
4797 StoreId helper. "Quoted" format values may contain spaces and
4798 logformat %macros. In theory, any logformat %macro can be used.
4799 In practice, a %macro expands as a dash (-) if the helper request is
4800 sent before the required macro information is available to Squid.
4803 NAME: store_id_children storeurl_rewrite_children
4804 TYPE: HelperChildConfig
4805 DEFAULT: 20 startup=0 idle=1 concurrency=0
4806 LOC: Config.storeIdChildren
4808 The maximum number of StoreID helper processes to spawn. If you limit
4809 it too few Squid will have to wait for them to process a backlog of
4810 requests, slowing it down. If you allow too many they will use RAM
4811 and other system resources noticably.
4813 The startup= and idle= options allow some measure of skew in your
4818 Sets a minimum of how many processes are to be spawned when Squid
4819 starts or reconfigures. When set to zero the first request will
4820 cause spawning of the first child process to handle it.
4822 Starting too few will cause an initial slowdown in traffic as Squid
4823 attempts to simultaneously spawn enough processes to cope.
4827 Sets a minimum of how many processes Squid is to try and keep available
4828 at all times. When traffic begins to rise above what the existing
4829 processes can handle this many more will be spawned up to the maximum
4830 configured. A minimum setting of 1 is required.
4834 The number of requests each storeID helper can handle in
4835 parallel. Defaults to 0 which indicates the helper
4836 is a old-style single threaded program.
4838 When this directive is set to a value >= 1 then the protocol
4839 used to communicate with the helper is modified to include
4840 an ID in front of the request/response. The ID from the request
4841 must be echoed back with the response to that request.
4844 NAME: store_id_access storeurl_rewrite_access
4847 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4848 LOC: Config.accessList.store_id
4850 If defined, this access list specifies which requests are
4851 sent to the StoreID processes. By default all requests
4854 This clause supports both fast and slow acl types.
4855 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4858 NAME: store_id_bypass storeurl_rewrite_bypass
4860 LOC: Config.onoff.store_id_bypass
4863 When this is 'on', a request will not go through the
4864 helper if all helpers are busy. If this is 'off'
4865 and the helper queue grows too large, Squid will exit
4866 with a FATAL error and ask you to increase the number of
4867 helpers. You should only enable this if the helperss
4868 are not critical to your caching system. If you use
4869 helpers for critical caching components, and you enable this
4870 option, users may not get objects from cache.
4874 OPTIONS FOR TUNING THE CACHE
4875 -----------------------------------------------------------------------------
4878 NAME: cache no_cache
4881 DEFAULT_DOC: By default, this directive is unused and has no effect.
4882 LOC: Config.accessList.noCache
4884 Requests denied by this directive will not be served from the cache
4885 and their responses will not be stored in the cache. This directive
4886 has no effect on other transactions and on already cached responses.
4888 This clause supports both fast and slow acl types.
4889 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4891 This and the two other similar caching directives listed below are
4892 checked at different transaction processing stages, have different
4893 access to response information, affect different cache operations,
4894 and differ in slow ACLs support:
4896 * cache: Checked before Squid makes a hit/miss determination.
4897 No access to reply information!
4898 Denies both serving a hit and storing a miss.
4899 Supports both fast and slow ACLs.
4900 * send_hit: Checked after a hit was detected.
4901 Has access to reply (hit) information.
4902 Denies serving a hit only.
4903 Supports fast ACLs only.
4904 * store_miss: Checked before storing a cachable miss.
4905 Has access to reply (miss) information.
4906 Denies storing a miss only.
4907 Supports fast ACLs only.
4909 If you are not sure which of the three directives to use, apply the
4910 following decision logic:
4912 * If your ACL(s) are of slow type _and_ need response info, redesign.
4913 Squid does not support that particular combination at this time.
4915 * If your directive ACL(s) are of slow type, use "cache"; and/or
4916 * if your directive ACL(s) need no response info, use "cache".
4918 * If you do not want the response cached, use store_miss; and/or
4919 * if you do not want a hit on a cached response, use send_hit.
4925 DEFAULT_DOC: By default, this directive is unused and has no effect.
4926 LOC: Config.accessList.sendHit
4928 Responses denied by this directive will not be served from the cache
4929 (but may still be cached, see store_miss). This directive has no
4930 effect on the responses it allows and on the cached objects.
4932 Please see the "cache" directive for a summary of differences among
4933 store_miss, send_hit, and cache directives.
4935 Unlike the "cache" directive, send_hit only supports fast acl
4936 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4940 # apply custom Store ID mapping to some URLs
4941 acl MapMe dstdomain .c.example.com
4942 store_id_program ...
4943 store_id_access allow MapMe
4945 # but prevent caching of special responses
4946 # such as 302 redirects that cause StoreID loops
4947 acl Ordinary http_status 200-299
4948 store_miss deny MapMe !Ordinary
4950 # and do not serve any previously stored special responses
4951 # from the cache (in case they were already cached before
4952 # the above store_miss rule was in effect).
4953 send_hit deny MapMe !Ordinary
4959 DEFAULT_DOC: By default, this directive is unused and has no effect.
4960 LOC: Config.accessList.storeMiss
4962 Responses denied by this directive will not be cached (but may still
4963 be served from the cache, see send_hit). This directive has no
4964 effect on the responses it allows and on the already cached responses.
4966 Please see the "cache" directive for a summary of differences among
4967 store_miss, send_hit, and cache directives. See the
4968 send_hit directive for a usage example.
4970 Unlike the "cache" directive, store_miss only supports fast acl
4971 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4977 LOC: Config.maxStale
4980 This option puts an upper limit on how stale content Squid
4981 will serve from the cache if cache validation fails.
4982 Can be overriden by the refresh_pattern max-stale option.
4985 NAME: refresh_pattern
4986 TYPE: refreshpattern
4990 usage: refresh_pattern [-i] regex min percent max [options]
4992 By default, regular expressions are CASE-SENSITIVE. To make
4993 them case-insensitive, use the -i option.
4995 'Min' is the time (in minutes) an object without an explicit
4996 expiry time should be considered fresh. The recommended
4997 value is 0, any higher values may cause dynamic applications
4998 to be erroneously cached unless the application designer
4999 has taken the appropriate actions.
5001 'Percent' is a percentage of the objects age (time since last
5002 modification age) an object without explicit expiry time
5003 will be considered fresh.
5005 'Max' is an upper limit on how long objects without an explicit
5006 expiry time will be considered fresh.
5008 options: override-expire
5013 ignore-must-revalidate
5020 override-expire enforces min age even if the server
5021 sent an explicit expiry time (e.g., with the
5022 Expires: header or Cache-Control: max-age). Doing this
5023 VIOLATES the HTTP standard. Enabling this feature
5024 could make you liable for problems which it causes.
5026 Note: override-expire does not enforce staleness - it only extends
5027 freshness / min. If the server returns a Expires time which
5028 is longer than your max time, Squid will still consider
5029 the object fresh for that period of time.
5031 override-lastmod enforces min age even on objects
5032 that were modified recently.
5034 reload-into-ims changes a client no-cache or ``reload''
5035 request for a cached entry into a conditional request using
5036 If-Modified-Since and/or If-None-Match headers, provided the
5037 cached entry has a Last-Modified and/or a strong ETag header.
5038 Doing this VIOLATES the HTTP standard. Enabling this feature
5039 could make you liable for problems which it causes.
5041 ignore-reload ignores a client no-cache or ``reload''
5042 header. Doing this VIOLATES the HTTP standard. Enabling
5043 this feature could make you liable for problems which
5046 ignore-no-store ignores any ``Cache-control: no-store''
5047 headers received from a server. Doing this VIOLATES
5048 the HTTP standard. Enabling this feature could make you
5049 liable for problems which it causes.
5051 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
5052 headers received from a server. Doing this VIOLATES
5053 the HTTP standard. Enabling this feature could make you
5054 liable for problems which it causes.
5056 ignore-private ignores any ``Cache-control: private''
5057 headers received from a server. Doing this VIOLATES
5058 the HTTP standard. Enabling this feature could make you
5059 liable for problems which it causes.
5061 ignore-auth caches responses to requests with authorization,
5062 as if the originserver had sent ``Cache-control: public''
5063 in the response header. Doing this VIOLATES the HTTP standard.
5064 Enabling this feature could make you liable for problems which
5067 refresh-ims causes squid to contact the origin server
5068 when a client issues an If-Modified-Since request. This
5069 ensures that the client will receive an updated version
5070 if one is available.
5072 store-stale stores responses even if they don't have explicit
5073 freshness or a validator (i.e., Last-Modified or an ETag)
5074 present, or if they're already stale. By default, Squid will
5075 not cache such responses because they usually can't be
5076 reused. Note that such responses will be stale by default.
5078 max-stale=NN provide a maximum staleness factor. Squid won't
5079 serve objects more stale than this even if it failed to
5080 validate the object. Default: use the max_stale global limit.
5082 Basically a cached object is:
5084 FRESH if expires < now, else STALE
5086 FRESH if lm-factor < percent, else STALE
5090 The refresh_pattern lines are checked in the order listed here.
5091 The first entry which matches is used. If none of the entries
5092 match the default will be used.
5094 Note, you must uncomment all the default lines if you want
5095 to change one. The default setting is only active if none is
5101 # Add any of your own refresh_pattern entries above these.
5103 refresh_pattern ^ftp: 1440 20% 10080
5104 refresh_pattern ^gopher: 1440 0% 1440
5105 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5106 refresh_pattern . 0 20% 4320
5110 NAME: quick_abort_min
5114 LOC: Config.quickAbort.min
5117 NAME: quick_abort_max
5121 LOC: Config.quickAbort.max
5124 NAME: quick_abort_pct
5128 LOC: Config.quickAbort.pct
5130 The cache by default continues downloading aborted requests
5131 which are almost completed (less than 16 KB remaining). This
5132 may be undesirable on slow (e.g. SLIP) links and/or very busy
5133 caches. Impatient users may tie up file descriptors and
5134 bandwidth by repeatedly requesting and immediately aborting
5137 When the user aborts a request, Squid will check the
5138 quick_abort values to the amount of data transfered until
5141 If the transfer has less than 'quick_abort_min' KB remaining,
5142 it will finish the retrieval.
5144 If the transfer has more than 'quick_abort_max' KB remaining,
5145 it will abort the retrieval.
5147 If more than 'quick_abort_pct' of the transfer has completed,
5148 it will finish the retrieval.
5150 If you do not want any retrieval to continue after the client
5151 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5154 If you want retrievals to always continue if they are being
5155 cached set 'quick_abort_min' to '-1 KB'.
5158 NAME: read_ahead_gap
5159 COMMENT: buffer-size
5161 LOC: Config.readAheadGap
5164 The amount of data the cache will buffer ahead of what has been
5165 sent to the client when retrieving an object from another server.
5169 IFDEF: USE_HTTP_VIOLATIONS
5172 LOC: Config.negativeTtl
5175 Set the Default Time-to-Live (TTL) for failed requests.
5176 Certain types of failures (such as "connection refused" and
5177 "404 Not Found") are able to be negatively-cached for a short time.
5178 Modern web servers should provide Expires: header, however if they
5179 do not this can provide a minimum TTL.
5180 The default is not to cache errors with unknown expiry details.
5182 Note that this is different from negative caching of DNS lookups.
5184 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5185 this feature could make you liable for problems which it
5189 NAME: positive_dns_ttl
5192 LOC: Config.positiveDnsTtl
5195 Upper limit on how long Squid will cache positive DNS responses.
5196 Default is 6 hours (360 minutes). This directive must be set
5197 larger than negative_dns_ttl.
5200 NAME: negative_dns_ttl
5203 LOC: Config.negativeDnsTtl
5206 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5207 This also sets the lower cache limit on positive lookups.
5208 Minimum value is 1 second, and it is not recommendable to go
5209 much below 10 seconds.
5212 NAME: range_offset_limit
5213 COMMENT: size [acl acl...]
5215 LOC: Config.rangeOffsetLimit
5218 usage: (size) [units] [[!]aclname]
5220 Sets an upper limit on how far (number of bytes) into the file
5221 a Range request may be to cause Squid to prefetch the whole file.
5222 If beyond this limit, Squid forwards the Range request as it is and
5223 the result is NOT cached.
5225 This is to stop a far ahead range request (lets say start at 17MB)
5226 from making Squid fetch the whole object up to that point before
5227 sending anything to the client.
5229 Multiple range_offset_limit lines may be specified, and they will
5230 be searched from top to bottom on each request until a match is found.
5231 The first match found will be used. If no line matches a request, the
5232 default limit of 0 bytes will be used.
5234 'size' is the limit specified as a number of units.
5236 'units' specifies whether to use bytes, KB, MB, etc.
5237 If no units are specified bytes are assumed.
5239 A size of 0 causes Squid to never fetch more than the
5240 client requested. (default)
5242 A size of 'none' causes Squid to always fetch the object from the
5243 beginning so it may cache the result. (2.0 style)
5245 'aclname' is the name of a defined ACL.
5247 NP: Using 'none' as the byte value here will override any quick_abort settings
5248 that may otherwise apply to the range request. The range request will
5249 be fully fetched from start to finish regardless of the client
5250 actions. This affects bandwidth usage.
5253 NAME: minimum_expiry_time
5256 LOC: Config.minimum_expiry_time
5259 The minimum caching time according to (Expires - Date)
5260 headers Squid honors if the object can't be revalidated.
5261 The default is 60 seconds.
5263 In reverse proxy environments it might be desirable to honor
5264 shorter object lifetimes. It is most likely better to make
5265 your server return a meaningful Last-Modified header however.
5267 In ESI environments where page fragments often have short
5268 lifetimes, this will often be best set to 0.
5271 NAME: store_avg_object_size
5275 LOC: Config.Store.avgObjectSize
5277 Average object size, used to estimate number of objects your
5278 cache can hold. The default is 13 KB.
5280 This is used to pre-seed the cache index memory allocation to
5281 reduce expensive reallocate operations while handling clients
5282 traffic. Too-large values may result in memory allocation during
5283 peak traffic, too-small values will result in wasted memory.
5285 Check the cache manager 'info' report metrics for the real
5286 object sizes seen by your Squid before tuning this.
5289 NAME: store_objects_per_bucket
5292 LOC: Config.Store.objectsPerBucket
5294 Target number of objects per bucket in the store hash table.
5295 Lowering this value increases the total number of buckets and
5296 also the storage maintenance rate. The default is 20.
5301 -----------------------------------------------------------------------------
5304 NAME: request_header_max_size
5308 LOC: Config.maxRequestHeaderSize
5310 This specifies the maximum size for HTTP headers in a request.
5311 Request headers are usually relatively small (about 512 bytes).
5312 Placing a limit on the request header size will catch certain
5313 bugs (for example with persistent connections) and possibly
5314 buffer-overflow or denial-of-service attacks.
5317 NAME: reply_header_max_size
5321 LOC: Config.maxReplyHeaderSize
5323 This specifies the maximum size for HTTP headers in a reply.
5324 Reply headers are usually relatively small (about 512 bytes).
5325 Placing a limit on the reply header size will catch certain
5326 bugs (for example with persistent connections) and possibly
5327 buffer-overflow or denial-of-service attacks.
5330 NAME: request_body_max_size
5334 DEFAULT_DOC: No limit.
5335 LOC: Config.maxRequestBodySize
5337 This specifies the maximum size for an HTTP request body.
5338 In other words, the maximum size of a PUT/POST request.
5339 A user who attempts to send a request with a body larger
5340 than this limit receives an "Invalid Request" error message.
5341 If you set this parameter to a zero (the default), there will
5342 be no limit imposed.
5344 See also client_request_buffer_max_size for an alternative
5345 limitation on client uploads which can be configured.
5348 NAME: client_request_buffer_max_size
5352 LOC: Config.maxRequestBufferSize
5354 This specifies the maximum buffer size of a client request.
5355 It prevents squid eating too much memory when somebody uploads
5359 NAME: chunked_request_body_max_size
5363 LOC: Config.maxChunkedRequestBodySize
5365 A broken or confused HTTP/1.1 client may send a chunked HTTP
5366 request to Squid. Squid does not have full support for that
5367 feature yet. To cope with such requests, Squid buffers the
5368 entire request and then dechunks request body to create a
5369 plain HTTP/1.0 request with a known content length. The plain
5370 request is then used by the rest of Squid code as usual.
5372 The option value specifies the maximum size of the buffer used
5373 to hold the request before the conversion. If the chunked
5374 request size exceeds the specified limit, the conversion
5375 fails, and the client receives an "unsupported request" error,
5376 as if dechunking was disabled.
5378 Dechunking is enabled by default. To disable conversion of
5379 chunked requests, set the maximum to zero.
5381 Request dechunking feature and this option in particular are a
5382 temporary hack. When chunking requests and responses are fully
5383 supported, there will be no need to buffer a chunked request.
5387 IFDEF: USE_HTTP_VIOLATIONS
5390 DEFAULT_DOC: Obey RFC 2616.
5391 LOC: Config.accessList.brokenPosts
5393 A list of ACL elements which, if matched, causes Squid to send
5394 an extra CRLF pair after the body of a PUT/POST request.
5396 Some HTTP servers has broken implementations of PUT/POST,
5397 and rely on an extra CRLF pair sent by some WWW clients.
5399 Quote from RFC2616 section 4.1 on this matter:
5401 Note: certain buggy HTTP/1.0 client implementations generate an
5402 extra CRLF's after a POST request. To restate what is explicitly
5403 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5404 a request with an extra CRLF.
5406 This clause only supports fast acl types.
5407 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5410 acl buggy_server url_regex ^http://....
5411 broken_posts allow buggy_server
5414 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5417 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5419 LOC: Adaptation::Config::use_indirect_client
5421 Controls whether the indirect client IP address (instead of the direct
5422 client IP address) is passed to adaptation services.
5424 See also: follow_x_forwarded_for adaptation_send_client_ip
5428 IFDEF: USE_HTTP_VIOLATIONS
5432 LOC: Config.onoff.via
5434 If set (default), Squid will include a Via header in requests and
5435 replies as required by RFC2616.
5441 LOC: Config.onoff.ie_refresh
5444 Microsoft Internet Explorer up until version 5.5 Service
5445 Pack 1 has an issue with transparent proxies, wherein it
5446 is impossible to force a refresh. Turning this on provides
5447 a partial fix to the problem, by causing all IMS-REFRESH
5448 requests from older IE versions to check the origin server
5449 for fresh content. This reduces hit ratio by some amount
5450 (~10% in my experience), but allows users to actually get
5451 fresh content when they want it. Note because Squid
5452 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5453 of 5.5 is unchanged from old versions of Squid (i.e. a
5454 forced refresh is impossible). Newer versions of IE will,
5455 hopefully, continue to have the new behavior and will be
5456 handled based on that assumption. This option defaults to
5457 the old Squid behavior, which is better for hit ratios but
5458 worse for clients using IE, if they need to be able to
5459 force fresh content.
5462 NAME: vary_ignore_expire
5465 LOC: Config.onoff.vary_ignore_expire
5468 Many HTTP servers supporting Vary gives such objects
5469 immediate expiry time with no cache-control header
5470 when requested by a HTTP/1.0 client. This option
5471 enables Squid to ignore such expiry times until
5472 HTTP/1.1 is fully implemented.
5474 WARNING: If turned on this may eventually cause some
5475 varying objects not intended for caching to get cached.
5478 NAME: request_entities
5480 LOC: Config.onoff.request_entities
5483 Squid defaults to deny GET and HEAD requests with request entities,
5484 as the meaning of such requests are undefined in the HTTP standard
5485 even if not explicitly forbidden.
5487 Set this directive to on if you have clients which insists
5488 on sending request entities in GET or HEAD requests. But be warned
5489 that there is server software (both proxies and web servers) which
5490 can fail to properly process this kind of request which may make you
5491 vulnerable to cache pollution attacks if enabled.
5494 NAME: request_header_access
5495 IFDEF: USE_HTTP_VIOLATIONS
5496 TYPE: http_header_access
5497 LOC: Config.request_header_access
5499 DEFAULT_DOC: No limits.
5501 Usage: request_header_access header_name allow|deny [!]aclname ...
5503 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5504 this feature could make you liable for problems which it
5507 This option replaces the old 'anonymize_headers' and the
5508 older 'http_anonymizer' option with something that is much
5509 more configurable. A list of ACLs for each header name allows
5510 removal of specific header fields under specific conditions.
5512 This option only applies to outgoing HTTP request headers (i.e.,
5513 headers sent by Squid to the next HTTP hop such as a cache peer
5514 or an origin server). The option has no effect during cache hit
5515 detection. The equivalent adaptation vectoring point in ICAP
5516 terminology is post-cache REQMOD.
5518 The option is applied to individual outgoing request header
5519 fields. For each request header field F, Squid uses the first
5520 qualifying sets of request_header_access rules:
5522 1. Rules with header_name equal to F's name.
5523 2. Rules with header_name 'Other', provided F's name is not
5524 on the hard-coded list of commonly used HTTP header names.
5525 3. Rules with header_name 'All'.
5527 Within that qualifying rule set, rule ACLs are checked as usual.
5528 If ACLs of an "allow" rule match, the header field is allowed to
5529 go through as is. If ACLs of a "deny" rule match, the header is
5530 removed and request_header_replace is then checked to identify
5531 if the removed header has a replacement. If no rules within the
5532 set have matching ACLs, the header field is left as is.
5534 For example, to achieve the same behavior as the old
5535 'http_anonymizer standard' option, you should use:
5537 request_header_access From deny all
5538 request_header_access Referer deny all
5539 request_header_access User-Agent deny all
5541 Or, to reproduce the old 'http_anonymizer paranoid' feature
5544 request_header_access Authorization allow all
5545 request_header_access Proxy-Authorization allow all
5546 request_header_access Cache-Control allow all
5547 request_header_access Content-Length allow all
5548 request_header_access Content-Type allow all
5549 request_header_access Date allow all
5550 request_header_access Host allow all
5551 request_header_access If-Modified-Since allow all
5552 request_header_access Pragma allow all
5553 request_header_access Accept allow all
5554 request_header_access Accept-Charset allow all
5555 request_header_access Accept-Encoding allow all
5556 request_header_access Accept-Language allow all
5557 request_header_access Connection allow all
5558 request_header_access All deny all
5560 HTTP reply headers are controlled with the reply_header_access directive.
5562 By default, all headers are allowed (no anonymizing is performed).
5565 NAME: reply_header_access
5566 IFDEF: USE_HTTP_VIOLATIONS
5567 TYPE: http_header_access
5568 LOC: Config.reply_header_access
5570 DEFAULT_DOC: No limits.
5572 Usage: reply_header_access header_name allow|deny [!]aclname ...
5574 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5575 this feature could make you liable for problems which it
5578 This option only applies to reply headers, i.e., from the
5579 server to the client.
5581 This is the same as request_header_access, but in the other
5582 direction. Please see request_header_access for detailed
5585 For example, to achieve the same behavior as the old
5586 'http_anonymizer standard' option, you should use:
5588 reply_header_access Server deny all
5589 reply_header_access WWW-Authenticate deny all
5590 reply_header_access Link deny all
5592 Or, to reproduce the old 'http_anonymizer paranoid' feature
5595 reply_header_access Allow allow all
5596 reply_header_access WWW-Authenticate allow all
5597 reply_header_access Proxy-Authenticate allow all
5598 reply_header_access Cache-Control allow all
5599 reply_header_access Content-Encoding allow all
5600 reply_header_access Content-Length allow all
5601 reply_header_access Content-Type allow all
5602 reply_header_access Date allow all
5603 reply_header_access Expires allow all
5604 reply_header_access Last-Modified allow all
5605 reply_header_access Location allow all
5606 reply_header_access Pragma allow all
5607 reply_header_access Content-Language allow all
5608 reply_header_access Retry-After allow all
5609 reply_header_access Title allow all
5610 reply_header_access Content-Disposition allow all
5611 reply_header_access Connection allow all
5612 reply_header_access All deny all
5614 HTTP request headers are controlled with the request_header_access directive.
5616 By default, all headers are allowed (no anonymizing is
5620 NAME: request_header_replace header_replace
5621 IFDEF: USE_HTTP_VIOLATIONS
5622 TYPE: http_header_replace
5623 LOC: Config.request_header_access
5626 Usage: request_header_replace header_name message
5627 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5629 This option allows you to change the contents of headers
5630 denied with request_header_access above, by replacing them
5631 with some fixed string.
5633 This only applies to request headers, not reply headers.
5635 By default, headers are removed if denied.
5638 NAME: reply_header_replace
5639 IFDEF: USE_HTTP_VIOLATIONS
5640 TYPE: http_header_replace
5641 LOC: Config.reply_header_access
5644 Usage: reply_header_replace header_name message
5645 Example: reply_header_replace Server Foo/1.0
5647 This option allows you to change the contents of headers
5648 denied with reply_header_access above, by replacing them
5649 with some fixed string.
5651 This only applies to reply headers, not request headers.
5653 By default, headers are removed if denied.
5656 NAME: request_header_add
5657 TYPE: HeaderWithAclList
5658 LOC: Config.request_header_add
5661 Usage: request_header_add field-name field-value acl1 [acl2] ...
5662 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5664 This option adds header fields to outgoing HTTP requests (i.e.,
5665 request headers sent by Squid to the next HTTP hop such as a
5666 cache peer or an origin server). The option has no effect during
5667 cache hit detection. The equivalent adaptation vectoring point
5668 in ICAP terminology is post-cache REQMOD.
5670 Field-name is a token specifying an HTTP header name. If a
5671 standard HTTP header name is used, Squid does not check whether
5672 the new header conflicts with any existing headers or violates
5673 HTTP rules. If the request to be modified already contains a
5674 field with the same name, the old field is preserved but the
5675 header field values are not merged.
5677 Field-value is either a token or a quoted string. If quoted
5678 string format is used, then the surrounding quotes are removed
5679 while escape sequences and %macros are processed.
5681 In theory, all of the logformat codes can be used as %macros.
5682 However, unlike logging (which happens at the very end of
5683 transaction lifetime), the transaction may not yet have enough
5684 information to expand a macro when the new header value is needed.
5685 And some information may already be available to Squid but not yet
5686 committed where the macro expansion code can access it (report
5687 such instances!). The macro will be expanded into a single dash
5688 ('-') in such cases. Not all macros have been tested.
5690 One or more Squid ACLs may be specified to restrict header
5691 injection to matching requests. As always in squid.conf, all
5692 ACLs in an option ACL list must be satisfied for the insertion
5693 to happen. The request_header_add option supports fast ACLs
5702 This option used to log custom information about the master
5703 transaction. For example, an admin may configure Squid to log
5704 which "user group" the transaction belongs to, where "user group"
5705 will be determined based on a set of ACLs and not [just]
5706 authentication information.
5707 Values of key/value pairs can be logged using %{key}note macros:
5709 note key value acl ...
5710 logformat myFormat ... %{key}note ...
5713 NAME: relaxed_header_parser
5714 COMMENT: on|off|warn
5716 LOC: Config.onoff.relaxed_header_parser
5719 In the default "on" setting Squid accepts certain forms
5720 of non-compliant HTTP messages where it is unambiguous
5721 what the sending application intended even if the message
5722 is not correctly formatted. The messages is then normalized
5723 to the correct form when forwarded by Squid.
5725 If set to "warn" then a warning will be emitted in cache.log
5726 each time such HTTP error is encountered.
5728 If set to "off" then such HTTP errors will cause the request
5729 or response to be rejected.
5732 NAME: collapsed_forwarding
5735 LOC: Config.onoff.collapsed_forwarding
5738 This option controls whether Squid is allowed to merge multiple
5739 potentially cachable requests for the same URI before Squid knows
5740 whether the response is going to be cachable.
5742 This feature is disabled by default: Enabling collapsed forwarding
5743 needlessly delays forwarding requests that look cachable (when they are
5744 collapsed) but then need to be forwarded individually anyway because
5745 they end up being for uncachable content. However, in some cases, such
5746 as accelleration of highly cachable content with periodic or groupped
5747 expiration times, the gains from collapsing [large volumes of
5748 simultenous refresh requests] outweigh losses from such delays.
5753 -----------------------------------------------------------------------------
5756 NAME: forward_timeout
5759 LOC: Config.Timeout.forward
5762 This parameter specifies how long Squid should at most attempt in
5763 finding a forwarding path for the request before giving up.
5766 NAME: connect_timeout
5769 LOC: Config.Timeout.connect
5772 This parameter specifies how long to wait for the TCP connect to
5773 the requested server or peer to complete before Squid should
5774 attempt to find another path where to forward the request.
5777 NAME: peer_connect_timeout
5780 LOC: Config.Timeout.peer_connect
5783 This parameter specifies how long to wait for a pending TCP
5784 connection to a peer cache. The default is 30 seconds. You
5785 may also set different timeout values for individual neighbors
5786 with the 'connect-timeout' option on a 'cache_peer' line.
5792 LOC: Config.Timeout.read
5795 The read_timeout is applied on server-side connections. After
5796 each successful read(), the timeout will be extended by this
5797 amount. If no data is read again after this amount of time,
5798 the request is aborted and logged with ERR_READ_TIMEOUT. The
5799 default is 15 minutes.
5805 LOC: Config.Timeout.write
5808 This timeout is tracked for all connections that have data
5809 available for writing and are waiting for the socket to become
5810 ready. After each successful write, the timeout is extended by
5811 the configured amount. If Squid has data to write but the
5812 connection is not ready for the configured duration, the
5813 transaction associated with the connection is terminated. The
5814 default is 15 minutes.
5817 NAME: request_timeout
5819 LOC: Config.Timeout.request
5822 How long to wait for complete HTTP request headers after initial
5823 connection establishment.
5826 NAME: client_idle_pconn_timeout persistent_request_timeout
5828 LOC: Config.Timeout.clientIdlePconn
5831 How long to wait for the next HTTP request on a persistent
5832 client connection after the previous request completes.
5835 NAME: client_lifetime
5838 LOC: Config.Timeout.lifetime
5841 The maximum amount of time a client (browser) is allowed to
5842 remain connected to the cache process. This protects the Cache
5843 from having a lot of sockets (and hence file descriptors) tied up
5844 in a CLOSE_WAIT state from remote clients that go away without
5845 properly shutting down (either because of a network failure or
5846 because of a poor client implementation). The default is one
5849 NOTE: The default value is intended to be much larger than any
5850 client would ever need to be connected to your cache. You
5851 should probably change client_lifetime only as a last resort.
5852 If you seem to have many client connections tying up
5853 filedescriptors, we recommend first tuning the read_timeout,
5854 request_timeout, persistent_request_timeout and quick_abort values.
5857 NAME: half_closed_clients
5859 LOC: Config.onoff.half_closed_clients
5862 Some clients may shutdown the sending side of their TCP
5863 connections, while leaving their receiving sides open. Sometimes,
5864 Squid can not tell the difference between a half-closed and a
5865 fully-closed TCP connection.
5867 By default, Squid will immediately close client connections when
5868 read(2) returns "no more data to read."
5870 Change this option to 'on' and Squid will keep open connections
5871 until a read(2) or write(2) on the socket returns an error.
5872 This may show some benefits for reverse proxies. But if not
5873 it is recommended to leave OFF.
5876 NAME: server_idle_pconn_timeout pconn_timeout
5878 LOC: Config.Timeout.serverIdlePconn
5881 Timeout for idle persistent connections to servers and other
5888 LOC: Ident::TheConfig.timeout
5891 Maximum time to wait for IDENT lookups to complete.
5893 If this is too high, and you enabled IDENT lookups from untrusted
5894 users, you might be susceptible to denial-of-service by having
5895 many ident requests going at once.
5898 NAME: shutdown_lifetime
5901 LOC: Config.shutdownLifetime
5904 When SIGTERM or SIGHUP is received, the cache is put into
5905 "shutdown pending" mode until all active sockets are closed.
5906 This value is the lifetime to set for all open descriptors
5907 during shutdown mode. Any active clients after this many
5908 seconds will receive a 'timeout' message.
5912 ADMINISTRATIVE PARAMETERS
5913 -----------------------------------------------------------------------------
5919 LOC: Config.adminEmail
5921 Email-address of local cache manager who will receive
5922 mail if the cache dies. The default is "webmaster".
5928 LOC: Config.EmailFrom
5930 From: email-address for mail sent when the cache dies.
5931 The default is to use 'squid@unique_hostname'.
5933 See also: unique_hostname directive.
5939 LOC: Config.EmailProgram
5941 Email program used to send mail if the cache dies.
5942 The default is "mail". The specified program must comply
5943 with the standard Unix mail syntax:
5944 mail-program recipient < mailfile
5946 Optional command line options can be specified.
5949 NAME: cache_effective_user
5951 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5952 LOC: Config.effectiveUser
5954 If you start Squid as root, it will change its effective/real
5955 UID/GID to the user specified below. The default is to change
5956 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5957 see also; cache_effective_group
5960 NAME: cache_effective_group
5963 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
5964 LOC: Config.effectiveGroup
5966 Squid sets the GID to the effective user's default group ID
5967 (taken from the password file) and supplementary group list
5968 from the groups membership.
5970 If you want Squid to run with a specific GID regardless of
5971 the group memberships of the effective user then set this
5972 to the group (or GID) you want Squid to run as. When set
5973 all other group privileges of the effective user are ignored
5974 and only this GID is effective. If Squid is not started as
5975 root the user starting Squid MUST be member of the specified
5978 This option is not recommended by the Squid Team.
5979 Our preference is for administrators to configure a secure
5980 user account for squid with UID/GID matching system policies.
5983 NAME: httpd_suppress_version_string
5987 LOC: Config.onoff.httpd_suppress_version_string
5989 Suppress Squid version string info in HTTP headers and HTML error pages.
5992 NAME: visible_hostname
5994 LOC: Config.visibleHostname
5996 DEFAULT_DOC: Automatically detect the system host name
5998 If you want to present a special hostname in error messages, etc,
5999 define this. Otherwise, the return value of gethostname()
6000 will be used. If you have multiple caches in a cluster and
6001 get errors about IP-forwarding you must set them to have individual
6002 names with this setting.
6005 NAME: unique_hostname
6007 LOC: Config.uniqueHostname
6009 DEFAULT_DOC: Copy the value from visible_hostname
6011 If you want to have multiple machines with the same
6012 'visible_hostname' you must give each machine a different
6013 'unique_hostname' so forwarding loops can be detected.
6016 NAME: hostname_aliases
6018 LOC: Config.hostnameAliases
6021 A list of other DNS names your cache has.
6029 Minimum umask which should be enforced while the proxy
6030 is running, in addition to the umask set at startup.
6032 For a traditional octal representation of umasks, start
6037 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6038 -----------------------------------------------------------------------------
6040 This section contains parameters for the (optional) cache
6041 announcement service. This service is provided to help
6042 cache administrators locate one another in order to join or
6043 create cache hierarchies.
6045 An 'announcement' message is sent (via UDP) to the registration
6046 service by Squid. By default, the announcement message is NOT
6047 SENT unless you enable it with 'announce_period' below.
6049 The announcement message includes your hostname, plus the
6050 following information from this configuration file:
6056 All current information is processed regularly and made
6057 available on the Web at http://www.ircache.net/Cache/Tracker/.
6060 NAME: announce_period
6062 LOC: Config.Announce.period
6064 DEFAULT_DOC: Announcement messages disabled.
6066 This is how frequently to send cache announcements.
6068 To enable announcing your cache, just set an announce period.
6071 announce_period 1 day
6076 DEFAULT: tracker.ircache.net
6077 LOC: Config.Announce.host
6079 Set the hostname where announce registration messages will be sent.
6081 See also announce_port and announce_file
6087 LOC: Config.Announce.file
6089 The contents of this file will be included in the announce
6090 registration messages.
6096 LOC: Config.Announce.port
6098 Set the port where announce registration messages will be sent.
6100 See also announce_host and announce_file
6104 HTTPD-ACCELERATOR OPTIONS
6105 -----------------------------------------------------------------------------
6108 NAME: httpd_accel_surrogate_id
6111 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6112 LOC: Config.Accel.surrogate_id
6114 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6115 need an identification token to allow control targeting. Because
6116 a farm of surrogates may all perform the same tasks, they may share
6117 an identification token.
6120 NAME: http_accel_surrogate_remote
6124 LOC: Config.onoff.surrogate_is_remote
6126 Remote surrogates (such as those in a CDN) honour the header
6127 "Surrogate-Control: no-store-remote".
6129 Set this to on to have squid behave as a remote surrogate.
6133 IFDEF: USE_SQUID_ESI
6134 COMMENT: libxml2|expat|custom
6136 LOC: ESIParser::Type
6139 ESI markup is not strictly XML compatible. The custom ESI parser
6140 will give higher performance, but cannot handle non ASCII character
6145 DELAY POOL PARAMETERS
6146 -----------------------------------------------------------------------------
6150 TYPE: delay_pool_count
6152 IFDEF: USE_DELAY_POOLS
6155 This represents the number of delay pools to be used. For example,
6156 if you have one class 2 delay pool and one class 3 delays pool, you
6157 have a total of 2 delay pools.
6159 See also delay_parameters, delay_class, delay_access for pool
6160 configuration details.
6164 TYPE: delay_pool_class
6166 IFDEF: USE_DELAY_POOLS
6169 This defines the class of each delay pool. There must be exactly one
6170 delay_class line for each delay pool. For example, to define two
6171 delay pools, one of class 2 and one of class 3, the settings above
6175 delay_pools 4 # 4 delay pools
6176 delay_class 1 2 # pool 1 is a class 2 pool
6177 delay_class 2 3 # pool 2 is a class 3 pool
6178 delay_class 3 4 # pool 3 is a class 4 pool
6179 delay_class 4 5 # pool 4 is a class 5 pool
6181 The delay pool classes are:
6183 class 1 Everything is limited by a single aggregate
6186 class 2 Everything is limited by a single aggregate
6187 bucket as well as an "individual" bucket chosen
6188 from bits 25 through 32 of the IPv4 address.
6190 class 3 Everything is limited by a single aggregate
6191 bucket as well as a "network" bucket chosen
6192 from bits 17 through 24 of the IP address and a
6193 "individual" bucket chosen from bits 17 through
6194 32 of the IPv4 address.
6196 class 4 Everything in a class 3 delay pool, with an
6197 additional limit on a per user basis. This
6198 only takes effect if the username is established
6199 in advance - by forcing authentication in your
6202 class 5 Requests are grouped according their tag (see
6203 external_acl's tag= reply).
6206 Each pool also requires a delay_parameters directive to configure the pool size
6207 and speed limits used whenever the pool is applied to a request. Along with
6208 a set of delay_access directives to determine when it is used.
6210 NOTE: If an IP address is a.b.c.d
6211 -> bits 25 through 32 are "d"
6212 -> bits 17 through 24 are "c"
6213 -> bits 17 through 32 are "c * 256 + d"
6215 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6216 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6218 This clause only supports fast acl types.
6219 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6221 See also delay_parameters and delay_access.
6225 TYPE: delay_pool_access
6227 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6228 IFDEF: USE_DELAY_POOLS
6231 This is used to determine which delay pool a request falls into.
6233 delay_access is sorted per pool and the matching starts with pool 1,
6234 then pool 2, ..., and finally pool N. The first delay pool where the
6235 request is allowed is selected for the request. If it does not allow
6236 the request to any pool then the request is not delayed (default).
6238 For example, if you want some_big_clients in delay
6239 pool 1 and lotsa_little_clients in delay pool 2:
6241 delay_access 1 allow some_big_clients
6242 delay_access 1 deny all
6243 delay_access 2 allow lotsa_little_clients
6244 delay_access 2 deny all
6245 delay_access 3 allow authenticated_clients
6247 See also delay_parameters and delay_class.
6251 NAME: delay_parameters
6252 TYPE: delay_pool_rates
6254 IFDEF: USE_DELAY_POOLS
6257 This defines the parameters for a delay pool. Each delay pool has
6258 a number of "buckets" associated with it, as explained in the
6259 description of delay_class.
6261 For a class 1 delay pool, the syntax is:
6263 delay_parameters pool aggregate
6265 For a class 2 delay pool:
6267 delay_parameters pool aggregate individual
6269 For a class 3 delay pool:
6271 delay_parameters pool aggregate network individual
6273 For a class 4 delay pool:
6275 delay_parameters pool aggregate network individual user
6277 For a class 5 delay pool:
6279 delay_parameters pool tagrate
6281 The option variables are:
6283 pool a pool number - ie, a number between 1 and the
6284 number specified in delay_pools as used in
6287 aggregate the speed limit parameters for the aggregate bucket
6290 individual the speed limit parameters for the individual
6291 buckets (class 2, 3).
6293 network the speed limit parameters for the network buckets
6296 user the speed limit parameters for the user buckets
6299 tagrate the speed limit parameters for the tag buckets
6302 A pair of delay parameters is written restore/maximum, where restore is
6303 the number of bytes (not bits - modem and network speeds are usually
6304 quoted in bits) per second placed into the bucket, and maximum is the
6305 maximum number of bytes which can be in the bucket at any time.
6307 There must be one delay_parameters line for each delay pool.
6310 For example, if delay pool number 1 is a class 2 delay pool as in the
6311 above example, and is being used to strictly limit each host to 64Kbit/sec
6312 (plus overheads), with no overall limit, the line is:
6314 delay_parameters 1 -1/-1 8000/8000
6316 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6318 Note that the figure -1 is used to represent "unlimited".
6321 And, if delay pool number 2 is a class 3 delay pool as in the above
6322 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6323 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6324 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6325 to permit a decent web page to be downloaded at a decent speed
6326 (if the network is not being limited due to overuse) but slow down
6327 large downloads more significantly:
6329 delay_parameters 2 32000/32000 8000/8000 600/8000
6331 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6332 8 x 8000 KByte/sec -> 64Kbit/sec.
6333 8 x 600 Byte/sec -> 4800bit/sec.
6336 Finally, for a class 4 delay pool as in the example - each user will
6337 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6339 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6342 See also delay_class and delay_access.
6346 NAME: delay_initial_bucket_level
6347 COMMENT: (percent, 0-100)
6350 IFDEF: USE_DELAY_POOLS
6351 LOC: Config.Delay.initial
6353 The initial bucket percentage is used to determine how much is put
6354 in each bucket when squid starts, is reconfigured, or first notices
6355 a host accessing it (in class 2 and class 3, individual hosts and
6356 networks only have buckets associated with them once they have been
6361 CLIENT DELAY POOL PARAMETERS
6362 -----------------------------------------------------------------------------
6365 NAME: client_delay_pools
6366 TYPE: client_delay_pool_count
6368 IFDEF: USE_DELAY_POOLS
6369 LOC: Config.ClientDelay
6371 This option specifies the number of client delay pools used. It must
6372 preceed other client_delay_* options.
6375 client_delay_pools 2
6377 See also client_delay_parameters and client_delay_access.
6380 NAME: client_delay_initial_bucket_level
6381 COMMENT: (percent, 0-no_limit)
6384 IFDEF: USE_DELAY_POOLS
6385 LOC: Config.ClientDelay.initial
6387 This option determines the initial bucket size as a percentage of
6388 max_bucket_size from client_delay_parameters. Buckets are created
6389 at the time of the "first" connection from the matching IP. Idle
6390 buckets are periodically deleted up.
6392 You can specify more than 100 percent but note that such "oversized"
6393 buckets are not refilled until their size goes down to max_bucket_size
6394 from client_delay_parameters.
6397 client_delay_initial_bucket_level 50
6400 NAME: client_delay_parameters
6401 TYPE: client_delay_pool_rates
6403 IFDEF: USE_DELAY_POOLS
6404 LOC: Config.ClientDelay
6407 This option configures client-side bandwidth limits using the
6410 client_delay_parameters pool speed_limit max_bucket_size
6412 pool is an integer ID used for client_delay_access matching.
6414 speed_limit is bytes added to the bucket per second.
6416 max_bucket_size is the maximum size of a bucket, enforced after any
6417 speed_limit additions.
6419 Please see the delay_parameters option for more information and
6423 client_delay_parameters 1 1024 2048
6424 client_delay_parameters 2 51200 16384
6426 See also client_delay_access.
6430 NAME: client_delay_access
6431 TYPE: client_delay_pool_access
6433 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6434 IFDEF: USE_DELAY_POOLS
6435 LOC: Config.ClientDelay
6437 This option determines the client-side delay pool for the
6440 client_delay_access pool_ID allow|deny acl_name
6442 All client_delay_access options are checked in their pool ID
6443 order, starting with pool 1. The first checked pool with allowed
6444 request is selected for the request. If no ACL matches or there
6445 are no client_delay_access options, the request bandwidth is not
6448 The ACL-selected pool is then used to find the
6449 client_delay_parameters for the request. Client-side pools are
6450 not used to aggregate clients. Clients are always aggregated
6451 based on their source IP addresses (one bucket per source IP).
6453 This clause only supports fast acl types.
6454 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6455 Additionally, only the client TCP connection details are available.
6456 ACLs testing HTTP properties will not work.
6458 Please see delay_access for more examples.
6461 client_delay_access 1 allow low_rate_network
6462 client_delay_access 2 allow vips_network
6465 See also client_delay_parameters and client_delay_pools.
6469 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6470 -----------------------------------------------------------------------------
6475 LOC: Config.Wccp.router
6477 DEFAULT_DOC: WCCP disabled.
6480 Use this option to define your WCCP ``home'' router for
6483 wccp_router supports a single WCCP(v1) router
6485 wccp2_router supports multiple WCCPv2 routers
6487 only one of the two may be used at the same time and defines
6488 which version of WCCP to use.
6492 TYPE: IpAddress_list
6493 LOC: Config.Wccp2.router
6495 DEFAULT_DOC: WCCPv2 disabled.
6498 Use this option to define your WCCP ``home'' router for
6501 wccp_router supports a single WCCP(v1) router
6503 wccp2_router supports multiple WCCPv2 routers
6505 only one of the two may be used at the same time and defines
6506 which version of WCCP to use.
6511 LOC: Config.Wccp.version
6515 This directive is only relevant if you need to set up WCCP(v1)
6516 to some very old and end-of-life Cisco routers. In all other
6517 setups it must be left unset or at the default setting.
6518 It defines an internal version in the WCCP(v1) protocol,
6519 with version 4 being the officially documented protocol.
6521 According to some users, Cisco IOS 11.2 and earlier only
6522 support WCCP version 3. If you're using that or an earlier
6523 version of IOS, you may need to change this value to 3, otherwise
6524 do not specify this parameter.
6527 NAME: wccp2_rebuild_wait
6529 LOC: Config.Wccp2.rebuildwait
6533 If this is enabled Squid will wait for the cache dir rebuild to finish
6534 before sending the first wccp2 HereIAm packet
6537 NAME: wccp2_forwarding_method
6539 LOC: Config.Wccp2.forwarding_method
6543 WCCP2 allows the setting of forwarding methods between the
6544 router/switch and the cache. Valid values are as follows:
6546 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6547 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6549 Currently (as of IOS 12.4) cisco routers only support GRE.
6550 Cisco switches only support the L2 redirect assignment method.
6553 NAME: wccp2_return_method
6555 LOC: Config.Wccp2.return_method
6559 WCCP2 allows the setting of return methods between the
6560 router/switch and the cache for packets that the cache
6561 decides not to handle. Valid values are as follows:
6563 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6564 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6566 Currently (as of IOS 12.4) cisco routers only support GRE.
6567 Cisco switches only support the L2 redirect assignment.
6569 If the "ip wccp redirect exclude in" command has been
6570 enabled on the cache interface, then it is still safe for
6571 the proxy server to use a l2 redirect method even if this
6572 option is set to GRE.
6575 NAME: wccp2_assignment_method
6577 LOC: Config.Wccp2.assignment_method
6581 WCCP2 allows the setting of methods to assign the WCCP hash
6582 Valid values are as follows:
6584 hash - Hash assignment
6585 mask - Mask assignment
6587 As a general rule, cisco routers support the hash assignment method
6588 and cisco switches support the mask assignment method.
6593 LOC: Config.Wccp2.info
6594 DEFAULT_IF_NONE: standard 0
6595 DEFAULT_DOC: Use the 'web-cache' standard service.
6598 WCCP2 allows for multiple traffic services. There are two
6599 types: "standard" and "dynamic". The standard type defines
6600 one service id - http (id 0). The dynamic service ids can be from
6601 51 to 255 inclusive. In order to use a dynamic service id
6602 one must define the type of traffic to be redirected; this is done
6603 using the wccp2_service_info option.
6605 The "standard" type does not require a wccp2_service_info option,
6606 just specifying the service id will suffice.
6608 MD5 service authentication can be enabled by adding
6609 "password=<password>" to the end of this service declaration.
6613 wccp2_service standard 0 # for the 'web-cache' standard service
6614 wccp2_service dynamic 80 # a dynamic service type which will be
6615 # fleshed out with subsequent options.
6616 wccp2_service standard 0 password=foo
6619 NAME: wccp2_service_info
6620 TYPE: wccp2_service_info
6621 LOC: Config.Wccp2.info
6625 Dynamic WCCPv2 services require further information to define the
6626 traffic you wish to have diverted.
6630 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6631 priority=<priority> ports=<port>,<port>..
6633 The relevant WCCPv2 flags:
6634 + src_ip_hash, dst_ip_hash
6635 + source_port_hash, dst_port_hash
6636 + src_ip_alt_hash, dst_ip_alt_hash
6637 + src_port_alt_hash, dst_port_alt_hash
6640 The port list can be one to eight entries.
6644 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6645 priority=240 ports=80
6647 Note: the service id must have been defined by a previous
6648 'wccp2_service dynamic <id>' entry.
6653 LOC: Config.Wccp2.weight
6657 Each cache server gets assigned a set of the destination
6658 hash proportional to their weight.
6663 LOC: Config.Wccp.address
6665 DEFAULT_DOC: Address selected by the operating system.
6668 Use this option if you require WCCPv2 to use a specific
6671 The default behavior is to not bind to any specific address.
6676 LOC: Config.Wccp2.address
6678 DEFAULT_DOC: Address selected by the operating system.
6681 Use this option if you require WCCP to use a specific
6684 The default behavior is to not bind to any specific address.
6688 PERSISTENT CONNECTION HANDLING
6689 -----------------------------------------------------------------------------
6691 Also see "pconn_timeout" in the TIMEOUTS section
6694 NAME: client_persistent_connections
6696 LOC: Config.onoff.client_pconns
6699 Persistent connection support for clients.
6700 Squid uses persistent connections (when allowed). You can use
6701 this option to disable persistent connections with clients.
6704 NAME: server_persistent_connections
6706 LOC: Config.onoff.server_pconns
6709 Persistent connection support for servers.
6710 Squid uses persistent connections (when allowed). You can use
6711 this option to disable persistent connections with servers.
6714 NAME: persistent_connection_after_error
6716 LOC: Config.onoff.error_pconns
6719 With this directive the use of persistent connections after
6720 HTTP errors can be disabled. Useful if you have clients
6721 who fail to handle errors on persistent connections proper.
6724 NAME: detect_broken_pconn
6726 LOC: Config.onoff.detect_broken_server_pconns
6729 Some servers have been found to incorrectly signal the use
6730 of HTTP/1.0 persistent connections even on replies not
6731 compatible, causing significant delays. This server problem
6732 has mostly been seen on redirects.
6734 By enabling this directive Squid attempts to detect such
6735 broken replies and automatically assume the reply is finished
6736 after 10 seconds timeout.
6740 CACHE DIGEST OPTIONS
6741 -----------------------------------------------------------------------------
6744 NAME: digest_generation
6745 IFDEF: USE_CACHE_DIGESTS
6747 LOC: Config.onoff.digest_generation
6750 This controls whether the server will generate a Cache Digest
6751 of its contents. By default, Cache Digest generation is
6752 enabled if Squid is compiled with --enable-cache-digests defined.
6755 NAME: digest_bits_per_entry
6756 IFDEF: USE_CACHE_DIGESTS
6758 LOC: Config.digest.bits_per_entry
6761 This is the number of bits of the server's Cache Digest which
6762 will be associated with the Digest entry for a given HTTP
6763 Method and URL (public key) combination. The default is 5.
6766 NAME: digest_rebuild_period
6767 IFDEF: USE_CACHE_DIGESTS
6770 LOC: Config.digest.rebuild_period
6773 This is the wait time between Cache Digest rebuilds.
6776 NAME: digest_rewrite_period
6778 IFDEF: USE_CACHE_DIGESTS
6780 LOC: Config.digest.rewrite_period
6783 This is the wait time between Cache Digest writes to
6787 NAME: digest_swapout_chunk_size
6790 IFDEF: USE_CACHE_DIGESTS
6791 LOC: Config.digest.swapout_chunk_size
6794 This is the number of bytes of the Cache Digest to write to
6795 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6799 NAME: digest_rebuild_chunk_percentage
6800 COMMENT: (percent, 0-100)
6801 IFDEF: USE_CACHE_DIGESTS
6803 LOC: Config.digest.rebuild_chunk_percentage
6806 This is the percentage of the Cache Digest to be scanned at a
6807 time. By default it is set to 10% of the Cache Digest.
6812 -----------------------------------------------------------------------------
6817 LOC: Config.Port.snmp
6819 DEFAULT_DOC: SNMP disabled.
6822 The port number where Squid listens for SNMP requests. To enable
6823 SNMP support set this to a suitable port number. Port number
6824 3401 is often used for the Squid SNMP agent. By default it's
6825 set to "0" (disabled)
6833 LOC: Config.accessList.snmp
6835 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
6838 Allowing or denying access to the SNMP port.
6840 All access to the agent is denied by default.
6843 snmp_access allow|deny [!]aclname ...
6845 This clause only supports fast acl types.
6846 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6849 snmp_access allow snmppublic localhost
6850 snmp_access deny all
6853 NAME: snmp_incoming_address
6855 LOC: Config.Addrs.snmp_incoming
6857 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
6860 Just like 'udp_incoming_address', but for the SNMP port.
6862 snmp_incoming_address is used for the SNMP socket receiving
6863 messages from SNMP agents.
6865 The default snmp_incoming_address is to listen on all
6866 available network interfaces.
6869 NAME: snmp_outgoing_address
6871 LOC: Config.Addrs.snmp_outgoing
6873 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
6876 Just like 'udp_outgoing_address', but for the SNMP port.
6878 snmp_outgoing_address is used for SNMP packets returned to SNMP
6881 If snmp_outgoing_address is not set it will use the same socket
6882 as snmp_incoming_address. Only change this if you want to have
6883 SNMP replies sent using another address than where this Squid
6884 listens for SNMP queries.
6886 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6887 the same value since they both use the same port.
6892 -----------------------------------------------------------------------------
6895 NAME: icp_port udp_port
6898 DEFAULT_DOC: ICP disabled.
6899 LOC: Config.Port.icp
6901 The port number where Squid sends and receives ICP queries to
6902 and from neighbor caches. The standard UDP port for ICP is 3130.
6905 icp_port @DEFAULT_ICP_PORT@
6912 DEFAULT_DOC: HTCP disabled.
6913 LOC: Config.Port.htcp
6915 The port number where Squid sends and receives HTCP queries to
6916 and from neighbor caches. To turn it on you want to set it to
6923 NAME: log_icp_queries
6927 LOC: Config.onoff.log_udp
6929 If set, ICP queries are logged to access.log. You may wish
6930 do disable this if your ICP load is VERY high to speed things
6931 up or to simplify log analysis.
6934 NAME: udp_incoming_address
6936 LOC:Config.Addrs.udp_incoming
6938 DEFAULT_DOC: Accept packets from all machine interfaces.
6940 udp_incoming_address is used for UDP packets received from other
6943 The default behavior is to not bind to any specific address.
6945 Only change this if you want to have all UDP queries received on
6946 a specific interface/address.
6948 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6949 modules. Altering it will affect all of them in the same manner.
6951 see also; udp_outgoing_address
6953 NOTE, udp_incoming_address and udp_outgoing_address can not
6954 have the same value since they both use the same port.
6957 NAME: udp_outgoing_address
6959 LOC: Config.Addrs.udp_outgoing
6961 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
6963 udp_outgoing_address is used for UDP packets sent out to other
6966 The default behavior is to not bind to any specific address.
6968 Instead it will use the same socket as udp_incoming_address.
6969 Only change this if you want to have UDP queries sent using another
6970 address than where this Squid listens for UDP queries from other
6973 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6974 modules. Altering it will affect all of them in the same manner.
6976 see also; udp_incoming_address
6978 NOTE, udp_incoming_address and udp_outgoing_address can not
6979 have the same value since they both use the same port.
6986 LOC: Config.onoff.icp_hit_stale
6988 If you want to return ICP_HIT for stale cache objects, set this
6989 option to 'on'. If you have sibling relationships with caches
6990 in other administrative domains, this should be 'off'. If you only
6991 have sibling relationships with caches under your control,
6992 it is probably okay to set this to 'on'.
6993 If set to 'on', your siblings should use the option "allow-miss"
6994 on their cache_peer lines for connecting to you.
6997 NAME: minimum_direct_hops
7000 LOC: Config.minDirectHops
7002 If using the ICMP pinging stuff, do direct fetches for sites
7003 which are no more than this many hops away.
7006 NAME: minimum_direct_rtt
7010 LOC: Config.minDirectRtt
7012 If using the ICMP pinging stuff, do direct fetches for sites
7013 which are no more than this many rtt milliseconds away.
7019 LOC: Config.Netdb.low
7021 The low water mark for the ICMP measurement database.
7023 Note: high watermark controlled by netdb_high directive.
7025 These watermarks are counts, not percents. The defaults are
7026 (low) 900 and (high) 1000. When the high water mark is
7027 reached, database entries will be deleted until the low
7034 LOC: Config.Netdb.high
7036 The high water mark for the ICMP measurement database.
7038 Note: low watermark controlled by netdb_low directive.
7040 These watermarks are counts, not percents. The defaults are
7041 (low) 900 and (high) 1000. When the high water mark is
7042 reached, database entries will be deleted until the low
7046 NAME: netdb_ping_period
7048 LOC: Config.Netdb.period
7051 The minimum period for measuring a site. There will be at
7052 least this much delay between successive pings to the same
7053 network. The default is five minutes.
7060 LOC: Config.onoff.query_icmp
7062 If you want to ask your peers to include ICMP data in their ICP
7063 replies, enable this option.
7065 If your peer has configured Squid (during compilation) with
7066 '--enable-icmp' that peer will send ICMP pings to origin server
7067 sites of the URLs it receives. If you enable this option the
7068 ICP replies from that peer will include the ICMP data (if available).
7069 Then, when choosing a parent cache, Squid will choose the parent with
7070 the minimal RTT to the origin server. When this happens, the
7071 hierarchy field of the access.log will be
7072 "CLOSEST_PARENT_MISS". This option is off by default.
7075 NAME: test_reachability
7079 LOC: Config.onoff.test_reachability
7081 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7082 instead of ICP_MISS if the target host is NOT in the ICMP
7083 database, or has a zero RTT.
7086 NAME: icp_query_timeout
7089 DEFAULT_DOC: Dynamic detection.
7091 LOC: Config.Timeout.icp_query
7093 Normally Squid will automatically determine an optimal ICP
7094 query timeout value based on the round-trip-time of recent ICP
7095 queries. If you want to override the value determined by
7096 Squid, set this 'icp_query_timeout' to a non-zero value. This
7097 value is specified in MILLISECONDS, so, to use a 2-second
7098 timeout (the old default), you would write:
7100 icp_query_timeout 2000
7103 NAME: maximum_icp_query_timeout
7107 LOC: Config.Timeout.icp_query_max
7109 Normally the ICP query timeout is determined dynamically. But
7110 sometimes it can lead to very large values (say 5 seconds).
7111 Use this option to put an upper limit on the dynamic timeout
7112 value. Do NOT use this option to always use a fixed (instead
7113 of a dynamic) timeout value. To set a fixed timeout see the
7114 'icp_query_timeout' directive.
7117 NAME: minimum_icp_query_timeout
7121 LOC: Config.Timeout.icp_query_min
7123 Normally the ICP query timeout is determined dynamically. But
7124 sometimes it can lead to very small timeouts, even lower than
7125 the normal latency variance on your link due to traffic.
7126 Use this option to put an lower limit on the dynamic timeout
7127 value. Do NOT use this option to always use a fixed (instead
7128 of a dynamic) timeout value. To set a fixed timeout see the
7129 'icp_query_timeout' directive.
7132 NAME: background_ping_rate
7136 LOC: Config.backgroundPingRate
7138 Controls how often the ICP pings are sent to siblings that
7139 have background-ping set.
7143 MULTICAST ICP OPTIONS
7144 -----------------------------------------------------------------------------
7149 LOC: Config.mcast_group_list
7152 This tag specifies a list of multicast groups which your server
7153 should join to receive multicasted ICP queries.
7155 NOTE! Be very careful what you put here! Be sure you
7156 understand the difference between an ICP _query_ and an ICP
7157 _reply_. This option is to be set only if you want to RECEIVE
7158 multicast queries. Do NOT set this option to SEND multicast
7159 ICP (use cache_peer for that). ICP replies are always sent via
7160 unicast, so this option does not affect whether or not you will
7161 receive replies from multicast group members.
7163 You must be very careful to NOT use a multicast address which
7164 is already in use by another group of caches.
7166 If you are unsure about multicast, please read the Multicast
7167 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7169 Usage: mcast_groups 239.128.16.128 224.0.1.20
7171 By default, Squid doesn't listen on any multicast groups.
7174 NAME: mcast_miss_addr
7175 IFDEF: MULTICAST_MISS_STREAM
7177 LOC: Config.mcast_miss.addr
7179 DEFAULT_DOC: disabled.
7181 If you enable this option, every "cache miss" URL will
7182 be sent out on the specified multicast address.
7184 Do not enable this option unless you are are absolutely
7185 certain you understand what you are doing.
7188 NAME: mcast_miss_ttl
7189 IFDEF: MULTICAST_MISS_STREAM
7191 LOC: Config.mcast_miss.ttl
7194 This is the time-to-live value for packets multicasted
7195 when multicasting off cache miss URLs is enabled. By
7196 default this is set to 'site scope', i.e. 16.
7199 NAME: mcast_miss_port
7200 IFDEF: MULTICAST_MISS_STREAM
7202 LOC: Config.mcast_miss.port
7205 This is the port number to be used in conjunction with
7209 NAME: mcast_miss_encode_key
7210 IFDEF: MULTICAST_MISS_STREAM
7212 LOC: Config.mcast_miss.encode_key
7213 DEFAULT: XXXXXXXXXXXXXXXX
7215 The URLs that are sent in the multicast miss stream are
7216 encrypted. This is the encryption key.
7219 NAME: mcast_icp_query_timeout
7223 LOC: Config.Timeout.mcast_icp_query
7225 For multicast peers, Squid regularly sends out ICP "probes" to
7226 count how many other peers are listening on the given multicast
7227 address. This value specifies how long Squid should wait to
7228 count all the replies. The default is 2000 msec, or 2
7233 INTERNAL ICON OPTIONS
7234 -----------------------------------------------------------------------------
7237 NAME: icon_directory
7239 LOC: Config.icons.directory
7240 DEFAULT: @DEFAULT_ICON_DIR@
7242 Where the icons are stored. These are normally kept in
7246 NAME: global_internal_static
7248 LOC: Config.onoff.global_internal_static
7251 This directive controls is Squid should intercept all requests for
7252 /squid-internal-static/ no matter which host the URL is requesting
7253 (default on setting), or if nothing special should be done for
7254 such URLs (off setting). The purpose of this directive is to make
7255 icons etc work better in complex cache hierarchies where it may
7256 not always be possible for all corners in the cache mesh to reach
7257 the server generating a directory listing.
7260 NAME: short_icon_urls
7262 LOC: Config.icons.use_short_names
7265 If this is enabled Squid will use short URLs for icons.
7266 If disabled it will revert to the old behavior of including
7267 it's own name and port in the URL.
7269 If you run a complex cache hierarchy with a mix of Squid and
7270 other proxies you may need to disable this directive.
7275 -----------------------------------------------------------------------------
7278 NAME: error_directory
7280 LOC: Config.errorDirectory
7282 DEFAULT_DOC: Send error pages in the clients preferred language
7284 If you wish to create your own versions of the default
7285 error files to customize them to suit your company copy
7286 the error/template files to another directory and point
7289 WARNING: This option will disable multi-language support
7290 on error pages if used.
7292 The squid developers are interested in making squid available in
7293 a wide variety of languages. If you are making translations for a
7294 language that Squid does not currently provide please consider
7295 contributing your translation back to the project.
7296 http://wiki.squid-cache.org/Translations
7298 The squid developers working on translations are happy to supply drop-in
7299 translated error files in exchange for any new language contributions.
7302 NAME: error_default_language
7303 IFDEF: USE_ERR_LOCALES
7305 LOC: Config.errorDefaultLanguage
7307 DEFAULT_DOC: Generate English language pages.
7309 Set the default language which squid will send error pages in
7310 if no existing translation matches the clients language
7313 If unset (default) generic English will be used.
7315 The squid developers are interested in making squid available in
7316 a wide variety of languages. If you are interested in making
7317 translations for any language see the squid wiki for details.
7318 http://wiki.squid-cache.org/Translations
7321 NAME: error_log_languages
7322 IFDEF: USE_ERR_LOCALES
7324 LOC: Config.errorLogMissingLanguages
7327 Log to cache.log what languages users are attempting to
7328 auto-negotiate for translations.
7330 Successful negotiations are not logged. Only failures
7331 have meaning to indicate that Squid may need an upgrade
7332 of its error page translations.
7335 NAME: err_page_stylesheet
7337 LOC: Config.errorStylesheet
7338 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7340 CSS Stylesheet to pattern the display of Squid default error pages.
7342 For information on CSS see http://www.w3.org/Style/CSS/
7347 LOC: Config.errHtmlText
7350 HTML text to include in error messages. Make this a "mailto"
7351 URL to your admin address, or maybe just a link to your
7352 organizations Web page.
7354 To include this in your error messages, you must rewrite
7355 the error template files (found in the "errors" directory).
7356 Wherever you want the 'err_html_text' line to appear,
7357 insert a %L tag in the error template file.
7360 NAME: email_err_data
7363 LOC: Config.onoff.emailErrData
7366 If enabled, information about the occurred error will be
7367 included in the mailto links of the ERR pages (if %W is set)
7368 so that the email body contains the data.
7369 Syntax is <A HREF="mailto:%w%W">%w</A>
7374 LOC: Config.denyInfoList
7377 Usage: deny_info err_page_name acl
7378 or deny_info http://... acl
7379 or deny_info TCP_RESET acl
7381 This can be used to return a ERR_ page for requests which
7382 do not pass the 'http_access' rules. Squid remembers the last
7383 acl it evaluated in http_access, and if a 'deny_info' line exists
7384 for that ACL Squid returns a corresponding error page.
7386 The acl is typically the last acl on the http_access deny line which
7387 denied access. The exceptions to this rule are:
7388 - When Squid needs to request authentication credentials. It's then
7389 the first authentication related acl encountered
7390 - When none of the http_access lines matches. It's then the last
7391 acl processed on the last http_access line.
7392 - When the decision to deny access was made by an adaptation service,
7393 the acl name is the corresponding eCAP or ICAP service_name.
7395 NP: If providing your own custom error pages with error_directory
7396 you may also specify them by your custom file name:
7397 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7399 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7400 may be specified by prefixing the file name with the code and a colon.
7401 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7403 Alternatively you can tell Squid to reset the TCP connection
7404 by specifying TCP_RESET.
7406 Or you can specify an error URL or URL pattern. The browsers will
7407 get redirected to the specified URL after formatting tags have
7408 been replaced. Redirect will be done with 302 or 307 according to
7409 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7410 the URL. e.g. 303:http://example.com/
7413 %a - username (if available. Password NOT included)
7416 %E - Error description
7418 %H - Request domain name
7419 %i - Client IP Address
7421 %o - Message result from external ACL helper
7422 %p - Request Port number
7423 %P - Request Protocol name
7424 %R - Request URL path
7425 %T - Timestamp in RFC 1123 format
7426 %U - Full canonical URL from client
7427 (HTTPS URLs terminate with *)
7428 %u - Full canonical URL from client
7429 %w - Admin email from squid.conf
7431 %% - Literal percent (%) code
7436 OPTIONS INFLUENCING REQUEST FORWARDING
7437 -----------------------------------------------------------------------------
7440 NAME: nonhierarchical_direct
7442 LOC: Config.onoff.nonhierarchical_direct
7445 By default, Squid will send any non-hierarchical requests
7446 (matching hierarchy_stoplist or not cacheable request type) direct
7449 When this is set to "off", Squid will prefer to send these
7450 requests to parents.
7452 Note that in most configurations, by turning this off you will only
7453 add latency to these request without any improvement in global hit
7456 This option only sets a preference. If the parent is unavailable a
7457 direct connection to the origin server may still be attempted. To
7458 completely prevent direct connections use never_direct.
7463 LOC: Config.onoff.prefer_direct
7466 Normally Squid tries to use parents for most requests. If you for some
7467 reason like it to first try going direct and only use a parent if
7468 going direct fails set this to on.
7470 By combining nonhierarchical_direct off and prefer_direct on you
7471 can set up Squid to use a parent as a backup path if going direct
7474 Note: If you want Squid to use parents for all requests see
7475 the never_direct directive. prefer_direct only modifies how Squid
7476 acts on cacheable requests.
7479 NAME: cache_miss_revalidate
7483 LOC: Config.onoff.cache_miss_revalidate
7485 Whether Squid on cache MISS will pass client revalidation requests
7486 to the server or tries to fetch new content for caching.
7487 This is useful while the cache is mostly empty to more quickly
7488 have the cache populated.
7490 When set to 'on' (default), Squid will pass all client If-* headers
7493 When set to 'off' and if the request is cacheable, Squid will
7494 remove the clients If-Modified-Since and If-None-Match headers from
7495 the request sent to the server.
7500 LOC: Config.accessList.AlwaysDirect
7502 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7504 Usage: always_direct allow|deny [!]aclname ...
7506 Here you can use ACL elements to specify requests which should
7507 ALWAYS be forwarded by Squid to the origin servers without using
7508 any peers. For example, to always directly forward requests for
7509 local servers ignoring any parents or siblings you may have use
7512 acl local-servers dstdomain my.domain.net
7513 always_direct allow local-servers
7515 To always forward FTP requests directly, use
7518 always_direct allow FTP
7520 NOTE: There is a similar, but opposite option named
7521 'never_direct'. You need to be aware that "always_direct deny
7522 foo" is NOT the same thing as "never_direct allow foo". You
7523 may need to use a deny rule to exclude a more-specific case of
7524 some other rule. Example:
7526 acl local-external dstdomain external.foo.net
7527 acl local-servers dstdomain .foo.net
7528 always_direct deny local-external
7529 always_direct allow local-servers
7531 NOTE: If your goal is to make the client forward the request
7532 directly to the origin server bypassing Squid then this needs
7533 to be done in the client configuration. Squid configuration
7534 can only tell Squid how Squid should fetch the object.
7536 NOTE: This directive is not related to caching. The replies
7537 is cached as usual even if you use always_direct. To not cache
7538 the replies see the 'cache' directive.
7540 This clause supports both fast and slow acl types.
7541 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7546 LOC: Config.accessList.NeverDirect
7548 DEFAULT_DOC: Allow DNS results to be used for this request.
7550 Usage: never_direct allow|deny [!]aclname ...
7552 never_direct is the opposite of always_direct. Please read
7553 the description for always_direct if you have not already.
7555 With 'never_direct' you can use ACL elements to specify
7556 requests which should NEVER be forwarded directly to origin
7557 servers. For example, to force the use of a proxy for all
7558 requests, except those in your local domain use something like:
7560 acl local-servers dstdomain .foo.net
7561 never_direct deny local-servers
7562 never_direct allow all
7564 or if Squid is inside a firewall and there are local intranet
7565 servers inside the firewall use something like:
7567 acl local-intranet dstdomain .foo.net
7568 acl local-external dstdomain external.foo.net
7569 always_direct deny local-external
7570 always_direct allow local-intranet
7571 never_direct allow all
7573 This clause supports both fast and slow acl types.
7574 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7578 ADVANCED NETWORKING OPTIONS
7579 -----------------------------------------------------------------------------
7582 NAME: incoming_udp_average incoming_icp_average
7585 LOC: Config.comm_incoming.udp.average
7587 Heavy voodoo here. I can't even believe you are reading this.
7588 Are you crazy? Don't even think about adjusting these unless
7589 you understand the algorithms in comm_select.c first!
7592 NAME: incoming_tcp_average incoming_http_average
7595 LOC: Config.comm_incoming.tcp.average
7597 Heavy voodoo here. I can't even believe you are reading this.
7598 Are you crazy? Don't even think about adjusting these unless
7599 you understand the algorithms in comm_select.c first!
7602 NAME: incoming_dns_average
7605 LOC: Config.comm_incoming.dns.average
7607 Heavy voodoo here. I can't even believe you are reading this.
7608 Are you crazy? Don't even think about adjusting these unless
7609 you understand the algorithms in comm_select.c first!
7612 NAME: min_udp_poll_cnt min_icp_poll_cnt
7615 LOC: Config.comm_incoming.udp.min_poll
7617 Heavy voodoo here. I can't even believe you are reading this.
7618 Are you crazy? Don't even think about adjusting these unless
7619 you understand the algorithms in comm_select.c first!
7622 NAME: min_dns_poll_cnt
7625 LOC: Config.comm_incoming.dns.min_poll
7627 Heavy voodoo here. I can't even believe you are reading this.
7628 Are you crazy? Don't even think about adjusting these unless
7629 you understand the algorithms in comm_select.c first!
7632 NAME: min_tcp_poll_cnt min_http_poll_cnt
7635 LOC: Config.comm_incoming.tcp.min_poll
7637 Heavy voodoo here. I can't even believe you are reading this.
7638 Are you crazy? Don't even think about adjusting these unless
7639 you understand the algorithms in comm_select.c first!
7645 LOC: Config.accept_filter
7649 The name of an accept(2) filter to install on Squid's
7650 listen socket(s). This feature is perhaps specific to
7651 FreeBSD and requires support in the kernel.
7653 The 'httpready' filter delays delivering new connections
7654 to Squid until a full HTTP request has been received.
7655 See the accf_http(9) man page for details.
7657 The 'dataready' filter delays delivering new connections
7658 to Squid until there is some data to process.
7659 See the accf_dataready(9) man page for details.
7663 The 'data' filter delays delivering of new connections
7664 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
7665 You may optionally specify a number of seconds to wait by
7666 'data=N' where N is the number of seconds. Defaults to 30
7667 if not specified. See the tcp(7) man page for details.
7670 accept_filter httpready
7675 NAME: client_ip_max_connections
7677 LOC: Config.client_ip_max_connections
7679 DEFAULT_DOC: No limit.
7681 Set an absolute limit on the number of connections a single
7682 client IP can use. Any more than this and Squid will begin to drop
7683 new connections from the client until it closes some links.
7685 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
7686 connections from the client. For finer control use the ACL access controls.
7688 Requires client_db to be enabled (the default).
7690 WARNING: This may noticably slow down traffic received via external proxies
7691 or NAT devices and cause them to rebound error messages back to their clients.
7694 NAME: tcp_recv_bufsize
7698 DEFAULT_DOC: Use operating system TCP defaults.
7699 LOC: Config.tcpRcvBufsz
7701 Size of receive buffer to set for TCP sockets. Probably just
7702 as easy to change your kernel's default.
7703 Omit from squid.conf to use the default buffer size.
7708 -----------------------------------------------------------------------------
7715 LOC: Adaptation::Icap::TheConfig.onoff
7718 If you want to enable the ICAP module support, set this to on.
7721 NAME: icap_connect_timeout
7724 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7727 This parameter specifies how long to wait for the TCP connect to
7728 the requested ICAP server to complete before giving up and either
7729 terminating the HTTP transaction or bypassing the failure.
7731 The default for optional services is peer_connect_timeout.
7732 The default for essential services is connect_timeout.
7733 If this option is explicitly set, its value applies to all services.
7736 NAME: icap_io_timeout
7740 DEFAULT_DOC: Use read_timeout.
7741 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7744 This parameter specifies how long to wait for an I/O activity on
7745 an established, active ICAP connection before giving up and
7746 either terminating the HTTP transaction or bypassing the
7750 NAME: icap_service_failure_limit
7751 COMMENT: limit [in memory-depth time-units]
7752 TYPE: icap_service_failure_limit
7754 LOC: Adaptation::Icap::TheConfig
7757 The limit specifies the number of failures that Squid tolerates
7758 when establishing a new TCP connection with an ICAP service. If
7759 the number of failures exceeds the limit, the ICAP service is
7760 not used for new ICAP requests until it is time to refresh its
7763 A negative value disables the limit. Without the limit, an ICAP
7764 service will not be considered down due to connectivity failures
7765 between ICAP OPTIONS requests.
7767 Squid forgets ICAP service failures older than the specified
7768 value of memory-depth. The memory fading algorithm
7769 is approximate because Squid does not remember individual
7770 errors but groups them instead, splitting the option
7771 value into ten time slots of equal length.
7773 When memory-depth is 0 and by default this option has no
7774 effect on service failure expiration.
7776 Squid always forgets failures when updating service settings
7777 using an ICAP OPTIONS transaction, regardless of this option
7781 # suspend service usage after 10 failures in 5 seconds:
7782 icap_service_failure_limit 10 in 5 seconds
7785 NAME: icap_service_revival_delay
7788 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7791 The delay specifies the number of seconds to wait after an ICAP
7792 OPTIONS request failure before requesting the options again. The
7793 failed ICAP service is considered "down" until fresh OPTIONS are
7796 The actual delay cannot be smaller than the hardcoded minimum
7797 delay of 30 seconds.
7800 NAME: icap_preview_enable
7804 LOC: Adaptation::Icap::TheConfig.preview_enable
7807 The ICAP Preview feature allows the ICAP server to handle the
7808 HTTP message by looking only at the beginning of the message body
7809 or even without receiving the body at all. In some environments,
7810 previews greatly speedup ICAP processing.
7812 During an ICAP OPTIONS transaction, the server may tell Squid what
7813 HTTP messages should be previewed and how big the preview should be.
7814 Squid will not use Preview if the server did not request one.
7816 To disable ICAP Preview for all ICAP services, regardless of
7817 individual ICAP server OPTIONS responses, set this option to "off".
7819 icap_preview_enable off
7822 NAME: icap_preview_size
7825 LOC: Adaptation::Icap::TheConfig.preview_size
7827 DEFAULT_DOC: No preview sent.
7829 The default size of preview data to be sent to the ICAP server.
7830 This value might be overwritten on a per server basis by OPTIONS requests.
7833 NAME: icap_206_enable
7837 LOC: Adaptation::Icap::TheConfig.allow206_enable
7840 206 (Partial Content) responses is an ICAP extension that allows the
7841 ICAP agents to optionally combine adapted and original HTTP message
7842 content. The decision to combine is postponed until the end of the
7843 ICAP response. Squid supports Partial Content extension by default.
7845 Activation of the Partial Content extension is negotiated with each
7846 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7847 negotation correctly even if they do not support the extension, but
7848 some might fail. To disable Partial Content support for all ICAP
7849 services and to avoid any negotiation, set this option to "off".
7855 NAME: icap_default_options_ttl
7858 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7861 The default TTL value for ICAP OPTIONS responses that don't have
7862 an Options-TTL header.
7865 NAME: icap_persistent_connections
7869 LOC: Adaptation::Icap::TheConfig.reuse_connections
7872 Whether or not Squid should use persistent connections to
7876 NAME: adaptation_send_client_ip icap_send_client_ip
7878 IFDEF: USE_ADAPTATION
7880 LOC: Adaptation::Config::send_client_ip
7883 If enabled, Squid shares HTTP client IP information with adaptation
7884 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7885 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7887 See also: adaptation_uses_indirect_client
7890 NAME: adaptation_send_username icap_send_client_username
7892 IFDEF: USE_ADAPTATION
7894 LOC: Adaptation::Config::send_username
7897 This sends authenticated HTTP client username (if available) to
7898 the adaptation service.
7900 For ICAP, the username value is encoded based on the
7901 icap_client_username_encode option and is sent using the header
7902 specified by the icap_client_username_header option.
7905 NAME: icap_client_username_header
7908 LOC: Adaptation::Icap::TheConfig.client_username_header
7909 DEFAULT: X-Client-Username
7911 ICAP request header name to use for adaptation_send_username.
7914 NAME: icap_client_username_encode
7918 LOC: Adaptation::Icap::TheConfig.client_username_encode
7921 Whether to base64 encode the authenticated client username.
7925 TYPE: icap_service_type
7927 LOC: Adaptation::Icap::TheConfig
7930 Defines a single ICAP service using the following format:
7932 icap_service id vectoring_point uri [option ...]
7935 an opaque identifier or name which is used to direct traffic to
7936 this specific service. Must be unique among all adaptation
7937 services in squid.conf.
7939 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7940 This specifies at which point of transaction processing the
7941 ICAP service should be activated. *_postcache vectoring points
7942 are not yet supported.
7944 uri: icap://servername:port/servicepath
7945 ICAP server and service location.
7947 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7948 transactions. Squid does not enforce that requirement. You can specify
7949 services with the same service_url and different vectoring_points. You
7950 can even specify multiple identical services as long as their
7951 service_names differ.
7953 To activate a service, use the adaptation_access directive. To group
7954 services, use adaptation_service_chain and adaptation_service_set.
7956 Service options are separated by white space. ICAP services support
7957 the following name=value options:
7960 If set to 'on' or '1', the ICAP service is treated as
7961 optional. If the service cannot be reached or malfunctions,
7962 Squid will try to ignore any errors and process the message as
7963 if the service was not enabled. No all ICAP errors can be
7964 bypassed. If set to 0, the ICAP service is treated as
7965 essential and all ICAP errors will result in an error page
7966 returned to the HTTP client.
7968 Bypass is off by default: services are treated as essential.
7971 If set to 'on' or '1', the ICAP service is allowed to
7972 dynamically change the current message adaptation plan by
7973 returning a chain of services to be used next. The services
7974 are specified using the X-Next-Services ICAP response header
7975 value, formatted as a comma-separated list of service names.
7976 Each named service should be configured in squid.conf. Other
7977 services are ignored. An empty X-Next-Services value results
7978 in an empty plan which ends the current adaptation.
7980 Dynamic adaptation plan may cross or cover multiple supported
7981 vectoring points in their natural processing order.
7983 Routing is not allowed by default: the ICAP X-Next-Services
7984 response header is ignored.
7987 Only has effect on split-stack systems. The default on those systems
7988 is to use IPv4-only connections. When set to 'on' this option will
7989 make Squid use IPv6-only connections to contact this ICAP service.
7991 on-overload=block|bypass|wait|force
7992 If the service Max-Connections limit has been reached, do
7993 one of the following for each new ICAP transaction:
7994 * block: send an HTTP error response to the client
7995 * bypass: ignore the "over-connected" ICAP service
7996 * wait: wait (in a FIFO queue) for an ICAP connection slot
7997 * force: proceed, ignoring the Max-Connections limit
7999 In SMP mode with N workers, each worker assumes the service
8000 connection limit is Max-Connections/N, even though not all
8001 workers may use a given service.
8003 The default value is "bypass" if service is bypassable,
8004 otherwise it is set to "wait".
8008 Use the given number as the Max-Connections limit, regardless
8009 of the Max-Connections value given by the service, if any.
8011 Older icap_service format without optional named parameters is
8012 deprecated but supported for backward compatibility.
8015 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8016 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
8020 TYPE: icap_class_type
8025 This deprecated option was documented to define an ICAP service
8026 chain, even though it actually defined a set of similar, redundant
8027 services, and the chains were not supported.
8029 To define a set of redundant services, please use the
8030 adaptation_service_set directive. For service chains, use
8031 adaptation_service_chain.
8035 TYPE: icap_access_type
8040 This option is deprecated. Please use adaptation_access, which
8041 has the same ICAP functionality, but comes with better
8042 documentation, and eCAP support.
8047 -----------------------------------------------------------------------------
8054 LOC: Adaptation::Ecap::TheConfig.onoff
8057 Controls whether eCAP support is enabled.
8061 TYPE: ecap_service_type
8063 LOC: Adaptation::Ecap::TheConfig
8066 Defines a single eCAP service
8068 ecap_service id vectoring_point uri [option ...]
8071 an opaque identifier or name which is used to direct traffic to
8072 this specific service. Must be unique among all adaptation
8073 services in squid.conf.
8075 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8076 This specifies at which point of transaction processing the
8077 eCAP service should be activated. *_postcache vectoring points
8078 are not yet supported.
8080 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8081 Squid uses the eCAP service URI to match this configuration
8082 line with one of the dynamically loaded services. Each loaded
8083 eCAP service must have a unique URI. Obtain the right URI from
8084 the service provider.
8086 To activate a service, use the adaptation_access directive. To group
8087 services, use adaptation_service_chain and adaptation_service_set.
8089 Service options are separated by white space. eCAP services support
8090 the following name=value options:
8093 If set to 'on' or '1', the eCAP service is treated as optional.
8094 If the service cannot be reached or malfunctions, Squid will try
8095 to ignore any errors and process the message as if the service
8096 was not enabled. No all eCAP errors can be bypassed.
8097 If set to 'off' or '0', the eCAP service is treated as essential
8098 and all eCAP errors will result in an error page returned to the
8101 Bypass is off by default: services are treated as essential.
8104 If set to 'on' or '1', the eCAP service is allowed to
8105 dynamically change the current message adaptation plan by
8106 returning a chain of services to be used next.
8108 Dynamic adaptation plan may cross or cover multiple supported
8109 vectoring points in their natural processing order.
8111 Routing is not allowed by default.
8113 Older ecap_service format without optional named parameters is
8114 deprecated but supported for backward compatibility.
8118 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8119 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8122 NAME: loadable_modules
8124 IFDEF: USE_LOADABLE_MODULES
8125 LOC: Config.loadable_module_names
8128 Instructs Squid to load the specified dynamic module(s) or activate
8129 preloaded module(s).
8131 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8135 MESSAGE ADAPTATION OPTIONS
8136 -----------------------------------------------------------------------------
8139 NAME: adaptation_service_set
8140 TYPE: adaptation_service_set_type
8141 IFDEF: USE_ADAPTATION
8146 Configures an ordered set of similar, redundant services. This is
8147 useful when hot standby or backup adaptation servers are available.
8149 adaptation_service_set set_name service_name1 service_name2 ...
8151 The named services are used in the set declaration order. The first
8152 applicable adaptation service from the set is used first. The next
8153 applicable service is tried if and only if the transaction with the
8154 previous service fails and the message waiting to be adapted is still
8157 When adaptation starts, broken services are ignored as if they were
8158 not a part of the set. A broken service is a down optional service.
8160 The services in a set must be attached to the same vectoring point
8161 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8163 If all services in a set are optional then adaptation failures are
8164 bypassable. If all services in the set are essential, then a
8165 transaction failure with one service may still be retried using
8166 another service from the set, but when all services fail, the master
8167 transaction fails as well.
8169 A set may contain a mix of optional and essential services, but that
8170 is likely to lead to surprising results because broken services become
8171 ignored (see above), making previously bypassable failures fatal.
8172 Technically, it is the bypassability of the last failed service that
8175 See also: adaptation_access adaptation_service_chain
8178 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8179 adaptation service_set svcLogger loggerLocal loggerRemote
8182 NAME: adaptation_service_chain
8183 TYPE: adaptation_service_chain_type
8184 IFDEF: USE_ADAPTATION
8189 Configures a list of complementary services that will be applied
8190 one-by-one, forming an adaptation chain or pipeline. This is useful
8191 when Squid must perform different adaptations on the same message.
8193 adaptation_service_chain chain_name service_name1 svc_name2 ...
8195 The named services are used in the chain declaration order. The first
8196 applicable adaptation service from the chain is used first. The next
8197 applicable service is applied to the successful adaptation results of
8198 the previous service in the chain.
8200 When adaptation starts, broken services are ignored as if they were
8201 not a part of the chain. A broken service is a down optional service.
8203 Request satisfaction terminates the adaptation chain because Squid
8204 does not currently allow declaration of RESPMOD services at the
8205 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8207 The services in a chain must be attached to the same vectoring point
8208 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8210 A chain may contain a mix of optional and essential services. If an
8211 essential adaptation fails (or the failure cannot be bypassed for
8212 other reasons), the master transaction fails. Otherwise, the failure
8213 is bypassed as if the failed adaptation service was not in the chain.
8215 See also: adaptation_access adaptation_service_set
8218 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8221 NAME: adaptation_access
8222 TYPE: adaptation_access_type
8223 IFDEF: USE_ADAPTATION
8226 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8228 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8230 adaptation_access service_name allow|deny [!]aclname...
8231 adaptation_access set_name allow|deny [!]aclname...
8233 At each supported vectoring point, the adaptation_access
8234 statements are processed in the order they appear in this
8235 configuration file. Statements pointing to the following services
8236 are ignored (i.e., skipped without checking their ACL):
8238 - services serving different vectoring points
8239 - "broken-but-bypassable" services
8240 - "up" services configured to ignore such transactions
8241 (e.g., based on the ICAP Transfer-Ignore header).
8243 When a set_name is used, all services in the set are checked
8244 using the same rules, to find the first applicable one. See
8245 adaptation_service_set for details.
8247 If an access list is checked and there is a match, the
8248 processing stops: For an "allow" rule, the corresponding
8249 adaptation service is used for the transaction. For a "deny"
8250 rule, no adaptation service is activated.
8252 It is currently not possible to apply more than one adaptation
8253 service at the same vectoring point to the same HTTP transaction.
8255 See also: icap_service and ecap_service
8258 adaptation_access service_1 allow all
8261 NAME: adaptation_service_iteration_limit
8263 IFDEF: USE_ADAPTATION
8264 LOC: Adaptation::Config::service_iteration_limit
8267 Limits the number of iterations allowed when applying adaptation
8268 services to a message. If your longest adaptation set or chain
8269 may have more than 16 services, increase the limit beyond its
8270 default value of 16. If detecting infinite iteration loops sooner
8271 is critical, make the iteration limit match the actual number
8272 of services in your longest adaptation set or chain.
8274 Infinite adaptation loops are most likely with routing services.
8276 See also: icap_service routing=1
8279 NAME: adaptation_masterx_shared_names
8281 IFDEF: USE_ADAPTATION
8282 LOC: Adaptation::Config::masterx_shared_name
8285 For each master transaction (i.e., the HTTP request and response
8286 sequence, including all related ICAP and eCAP exchanges), Squid
8287 maintains a table of metadata. The table entries are (name, value)
8288 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8289 with the master transaction.
8291 This option specifies the table entry names that Squid must accept
8292 from and forward to the adaptation transactions.
8294 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8295 shared table by returning an ICAP header field with a name
8296 specified in adaptation_masterx_shared_names.
8298 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8299 shared table by implementing the libecap::visitEachOption() API
8300 to provide an option with a name specified in
8301 adaptation_masterx_shared_names.
8303 Squid will store and forward the set entry to subsequent adaptation
8304 transactions within the same master transaction scope.
8306 Only one shared entry name is supported at this time.
8309 # share authentication information among ICAP services
8310 adaptation_masterx_shared_names X-Subscriber-ID
8313 NAME: adaptation_meta
8315 IFDEF: USE_ADAPTATION
8316 LOC: Adaptation::Config::metaHeaders
8319 This option allows Squid administrator to add custom ICAP request
8320 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8321 Use it to pass custom authentication tokens and other
8322 transaction-state related meta information to an ICAP/eCAP service.
8324 The addition of a meta header is ACL-driven:
8325 adaptation_meta name value [!]aclname ...
8327 Processing for a given header name stops after the first ACL list match.
8328 Thus, it is impossible to add two headers with the same name. If no ACL
8329 lists match for a given header name, no such header is added. For
8332 # do not debug transactions except for those that need debugging
8333 adaptation_meta X-Debug 1 needs_debugging
8335 # log all transactions except for those that must remain secret
8336 adaptation_meta X-Log 1 !keep_secret
8338 # mark transactions from users in the "G 1" group
8339 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8341 The "value" parameter may be a regular squid.conf token or a "double
8342 quoted string". Within the quoted string, use backslash (\) to escape
8343 any character, which is currently only useful for escaping backslashes
8344 and double quotes. For example,
8345 "this string has one backslash (\\) and two \"quotes\""
8347 Used adaptation_meta header values may be logged via %note
8348 logformat code. If multiple adaptation_meta headers with the same name
8349 are used during master transaction lifetime, the header values are
8350 logged in the order they were used and duplicate values are ignored
8351 (only the first repeated value will be logged).
8357 LOC: Adaptation::Icap::TheConfig.repeat
8358 DEFAULT_IF_NONE: deny all
8360 This ACL determines which retriable ICAP transactions are
8361 retried. Transactions that received a complete ICAP response
8362 and did not have to consume or produce HTTP bodies to receive
8363 that response are usually retriable.
8365 icap_retry allow|deny [!]aclname ...
8367 Squid automatically retries some ICAP I/O timeouts and errors
8368 due to persistent connection race conditions.
8370 See also: icap_retry_limit
8373 NAME: icap_retry_limit
8376 LOC: Adaptation::Icap::TheConfig.repeat_limit
8378 DEFAULT_DOC: No retries are allowed.
8380 Limits the number of retries allowed.
8382 Communication errors due to persistent connection race
8383 conditions are unavoidable, automatically retried, and do not
8384 count against this limit.
8386 See also: icap_retry
8392 -----------------------------------------------------------------------------
8395 NAME: check_hostnames
8398 LOC: Config.onoff.check_hostnames
8400 For security and stability reasons Squid can check
8401 hostnames for Internet standard RFC compliance. If you want
8402 Squid to perform these checks turn this directive on.
8405 NAME: allow_underscore
8408 LOC: Config.onoff.allow_underscore
8410 Underscore characters is not strictly allowed in Internet hostnames
8411 but nevertheless used by many sites. Set this to off if you want
8412 Squid to be strict about the standard.
8413 This check is performed only when check_hostnames is set to on.
8416 NAME: dns_retransmit_interval
8419 LOC: Config.Timeout.idns_retransmit
8421 Initial retransmit interval for DNS queries. The interval is
8422 doubled each time all configured DNS servers have been tried.
8428 LOC: Config.Timeout.idns_query
8430 DNS Query timeout. If no response is received to a DNS query
8431 within this time all DNS servers for the queried domain
8432 are assumed to be unavailable.
8435 NAME: dns_packet_max
8437 DEFAULT_DOC: EDNS disabled
8439 LOC: Config.dns.packet_max
8441 Maximum number of bytes packet size to advertise via EDNS.
8442 Set to "none" to disable EDNS large packet support.
8444 For legacy reasons DNS UDP replies will default to 512 bytes which
8445 is too small for many responses. EDNS provides a means for Squid to
8446 negotiate receiving larger responses back immediately without having
8447 to failover with repeat requests. Responses larger than this limit
8448 will retain the old behaviour of failover to TCP DNS.
8450 Squid has no real fixed limit internally, but allowing packet sizes
8451 over 1500 bytes requires network jumbogram support and is usually not
8454 WARNING: The RFC also indicates that some older resolvers will reply
8455 with failure of the whole request if the extension is added. Some
8456 resolvers have already been identified which will reply with mangled
8457 EDNS response on occasion. Usually in response to many-KB jumbogram
8458 sizes being advertised by Squid.
8459 Squid will currently treat these both as an unable-to-resolve domain
8460 even if it would be resolvable without EDNS.
8467 DEFAULT_DOC: Search for single-label domain names is disabled.
8468 LOC: Config.onoff.res_defnames
8470 Normally the RES_DEFNAMES resolver option is disabled
8471 (see res_init(3)). This prevents caches in a hierarchy
8472 from interpreting single-component hostnames locally. To allow
8473 Squid to handle single-component names, enable this option.
8476 NAME: dns_multicast_local
8480 DEFAULT_DOC: Search for .local and .arpa names is disabled.
8481 LOC: Config.onoff.dns_mdns
8483 When set to on, Squid sends multicast DNS lookups on the local
8484 network for domains ending in .local and .arpa.
8485 This enables local servers and devices to be contacted in an
8486 ad-hoc or zero-configuration network environment.
8489 NAME: dns_nameservers
8492 DEFAULT_DOC: Use operating system definitions
8493 LOC: Config.dns_nameservers
8495 Use this if you want to specify a list of DNS name servers
8496 (IP addresses) to use instead of those given in your
8497 /etc/resolv.conf file.
8499 On Windows platforms, if no value is specified here or in
8500 the /etc/resolv.conf file, the list of DNS name servers are
8501 taken from the Windows registry, both static and dynamic DHCP
8502 configurations are supported.
8504 Example: dns_nameservers 10.0.0.1 192.172.0.4
8509 DEFAULT: @DEFAULT_HOSTS@
8510 LOC: Config.etcHostsPath
8512 Location of the host-local IP name-address associations
8513 database. Most Operating Systems have such a file on different
8515 - Un*X & Linux: /etc/hosts
8516 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
8517 (%SystemRoot% value install default is c:\winnt)
8518 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
8519 (%SystemRoot% value install default is c:\windows)
8520 - Windows 9x/Me: %windir%\hosts
8521 (%windir% value is usually c:\windows)
8522 - Cygwin: /etc/hosts
8524 The file contains newline-separated definitions, in the
8525 form ip_address_in_dotted_form name [name ...] names are
8526 whitespace-separated. Lines beginning with an hash (#)
8527 character are comments.
8529 The file is checked at startup and upon configuration.
8530 If set to 'none', it won't be checked.
8531 If append_domain is used, that domain will be added to
8532 domain-local (i.e. not containing any dot character) host
8538 LOC: Config.appendDomain
8540 DEFAULT_DOC: Use operating system definitions
8542 Appends local domain name to hostnames without any dots in
8543 them. append_domain must begin with a period.
8545 Be warned there are now Internet names with no dots in
8546 them using only top-domain names, so setting this may
8547 cause some Internet sites to become unavailable.
8550 append_domain .yourdomain.com
8553 NAME: ignore_unknown_nameservers
8555 LOC: Config.onoff.ignore_unknown_nameservers
8558 By default Squid checks that DNS responses are received
8559 from the same IP addresses they are sent to. If they
8560 don't match, Squid ignores the response and writes a warning
8561 message to cache.log. You can allow responses from unknown
8562 nameservers by setting this option to 'off'.
8568 LOC: Config.dns.v4_first
8570 With the IPv6 Internet being as fast or faster than IPv4 Internet
8571 for most networks Squid prefers to contact websites over IPv6.
8573 This option reverses the order of preference to make Squid contact
8574 dual-stack websites over IPv4 first. Squid will still perform both
8575 IPv6 and IPv4 DNS lookups before connecting.
8578 This option will restrict the situations under which IPv6
8579 connectivity is used (and tested). Hiding network problems
8580 which would otherwise be detected and warned about.
8584 COMMENT: (number of entries)
8587 LOC: Config.ipcache.size
8589 Maximum number of DNS IP cache entries.
8596 LOC: Config.ipcache.low
8603 LOC: Config.ipcache.high
8605 The size, low-, and high-water marks for the IP cache.
8608 NAME: fqdncache_size
8609 COMMENT: (number of entries)
8612 LOC: Config.fqdncache.size
8614 Maximum number of FQDN cache entries.
8619 -----------------------------------------------------------------------------
8622 NAME: configuration_includes_quoted_values
8624 TYPE: configuration_includes_quoted_values
8626 LOC: ConfigParser::RecognizeQuotedValues
8628 If set, Squid will recognize each "quoted string" after a configuration
8629 directive as a single parameter. The quotes are stripped before the
8630 parameter value is interpreted or used.
8631 See "Values with spaces, quotes, and other special characters"
8632 section for more details.
8639 LOC: Config.onoff.mem_pools
8641 If set, Squid will keep pools of allocated (but unused) memory
8642 available for future use. If memory is a premium on your
8643 system and you believe your malloc library outperforms Squid
8644 routines, disable this.
8647 NAME: memory_pools_limit
8651 LOC: Config.MemPools.limit
8653 Used only with memory_pools on:
8654 memory_pools_limit 50 MB
8656 If set to a non-zero value, Squid will keep at most the specified
8657 limit of allocated (but unused) memory in memory pools. All free()
8658 requests that exceed this limit will be handled by your malloc
8659 library. Squid does not pre-allocate any memory, just safe-keeps
8660 objects that otherwise would be free()d. Thus, it is safe to set
8661 memory_pools_limit to a reasonably high value even if your
8662 configuration will use less memory.
8664 If set to none, Squid will keep all memory it can. That is, there
8665 will be no limit on the total amount of memory used for safe-keeping.
8667 To disable memory allocation optimization, do not set
8668 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
8670 An overhead for maintaining memory pools is not taken into account
8671 when the limit is checked. This overhead is close to four bytes per
8672 object kept. However, pools may actually _save_ memory because of
8673 reduced memory thrashing in your malloc library.
8677 COMMENT: on|off|transparent|truncate|delete
8680 LOC: opt_forwarded_for
8682 If set to "on", Squid will append your client's IP address
8683 in the HTTP requests it forwards. By default it looks like:
8685 X-Forwarded-For: 192.1.2.3
8687 If set to "off", it will appear as
8689 X-Forwarded-For: unknown
8691 If set to "transparent", Squid will not alter the
8692 X-Forwarded-For header in any way.
8694 If set to "delete", Squid will delete the entire
8695 X-Forwarded-For header.
8697 If set to "truncate", Squid will remove all existing
8698 X-Forwarded-For entries, and place the client IP as the sole entry.
8701 NAME: cachemgr_passwd
8702 TYPE: cachemgrpasswd
8704 DEFAULT_DOC: No password. Actions which require password are denied.
8705 LOC: Config.passwd_list
8707 Specify passwords for cachemgr operations.
8709 Usage: cachemgr_passwd password action action ...
8711 Some valid actions are (see cache manager menu for a full list):
8751 * Indicates actions which will not be performed without a
8752 valid password, others can be performed if not listed here.
8754 To disable an action, set the password to "disable".
8755 To allow performing an action without a password, set the
8758 Use the keyword "all" to set the same password for all actions.
8761 cachemgr_passwd secret shutdown
8762 cachemgr_passwd lesssssssecret info stats/objects
8763 cachemgr_passwd disable all
8770 LOC: Config.onoff.client_db
8772 If you want to disable collecting per-client statistics,
8773 turn off client_db here.
8776 NAME: refresh_all_ims
8780 LOC: Config.onoff.refresh_all_ims
8782 When you enable this option, squid will always check
8783 the origin server for an update when a client sends an
8784 If-Modified-Since request. Many browsers use IMS
8785 requests when the user requests a reload, and this
8786 ensures those clients receive the latest version.
8788 By default (off), squid may return a Not Modified response
8789 based on the age of the cached version.
8792 NAME: reload_into_ims
8793 IFDEF: USE_HTTP_VIOLATIONS
8797 LOC: Config.onoff.reload_into_ims
8799 When you enable this option, client no-cache or ``reload''
8800 requests will be changed to If-Modified-Since requests.
8801 Doing this VIOLATES the HTTP standard. Enabling this
8802 feature could make you liable for problems which it
8805 see also refresh_pattern for a more selective approach.
8808 NAME: connect_retries
8810 LOC: Config.connect_retries
8812 DEFAULT_DOC: Do not retry failed connections.
8814 This sets the maximum number of connection attempts made for each
8815 TCP connection. The connect_retries attempts must all still
8816 complete within the connection timeout period.
8818 The default is not to re-try if the first connection attempt fails.
8819 The (not recommended) maximum is 10 tries.
8821 A warning message will be generated if it is set to a too-high
8822 value and the configured value will be over-ridden.
8824 Note: These re-tries are in addition to forward_max_tries
8825 which limit how many different addresses may be tried to find
8829 NAME: retry_on_error
8831 LOC: Config.retry.onerror
8834 If set to ON Squid will automatically retry requests when
8835 receiving an error response with status 403 (Forbidden),
8836 500 (Internal Error), 501 or 503 (Service not available).
8837 Status 502 and 504 (Gateway errors) are always retried.
8839 This is mainly useful if you are in a complex cache hierarchy to
8840 work around access control errors.
8842 NOTE: This retry will attempt to find another working destination.
8843 Which is different from the server which just failed.
8846 NAME: as_whois_server
8848 LOC: Config.as_whois_server
8849 DEFAULT: whois.ra.net
8851 WHOIS server to query for AS numbers. NOTE: AS numbers are
8852 queried only when Squid starts up, not for every request.
8857 LOC: Config.onoff.offline
8860 Enable this option and Squid will never try to validate cached
8864 NAME: uri_whitespace
8865 TYPE: uri_whitespace
8866 LOC: Config.uri_whitespace
8869 What to do with requests that have whitespace characters in the
8872 strip: The whitespace characters are stripped out of the URL.
8873 This is the behavior recommended by RFC2396 and RFC3986
8874 for tolerant handling of generic URI.
8875 NOTE: This is one difference between generic URI and HTTP URLs.
8877 deny: The request is denied. The user receives an "Invalid
8879 This is the behaviour recommended by RFC2616 for safe
8880 handling of HTTP request URL.
8882 allow: The request is allowed and the URI is not changed. The
8883 whitespace characters remain in the URI. Note the
8884 whitespace is passed to redirector processes if they
8886 Note this may be considered a violation of RFC2616
8887 request parsing where whitespace is prohibited in the
8890 encode: The request is allowed and the whitespace characters are
8891 encoded according to RFC1738.
8893 chop: The request is allowed and the URI is chopped at the
8897 NOTE the current Squid implementation of encode and chop violates
8898 RFC2616 by not using a 301 redirect after altering the URL.
8903 LOC: Config.chroot_dir
8906 Specifies a directory where Squid should do a chroot() while
8907 initializing. This also causes Squid to fully drop root
8908 privileges after initializing. This means, for example, if you
8909 use a HTTP port less than 1024 and try to reconfigure, you may
8910 get an error saying that Squid can not open the port.
8913 NAME: balance_on_multiple_ip
8915 LOC: Config.onoff.balance_on_multiple_ip
8918 Modern IP resolvers in squid sort lookup results by preferred access.
8919 By default squid will use these IP in order and only rotates to
8920 the next listed when the most preffered fails.
8922 Some load balancing servers based on round robin DNS have been
8923 found not to preserve user session state across requests
8924 to different IP addresses.
8926 Enabling this directive Squid rotates IP's per request.
8929 NAME: pipeline_prefetch
8930 TYPE: pipelinePrefetch
8931 LOC: Config.pipeline_max_prefetch
8933 DEFAULT_DOC: Do not pre-parse pipelined requests.
8935 HTTP clients may send a pipeline of 1+N requests to Squid using a
8936 single connection, without waiting for Squid to respond to the first
8937 of those requests. This option limits the number of concurrent
8938 requests Squid will try to handle in parallel. If set to N, Squid
8939 will try to receive and process up to 1+N requests on the same
8940 connection concurrently.
8942 Defaults to 0 (off) for bandwidth management and access logging
8945 NOTE: pipelining requires persistent connections to clients.
8947 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8950 NAME: high_response_time_warning
8953 LOC: Config.warnings.high_rptm
8955 DEFAULT_DOC: disabled.
8957 If the one-minute median response time exceeds this value,
8958 Squid prints a WARNING with debug level 0 to get the
8959 administrators attention. The value is in milliseconds.
8962 NAME: high_page_fault_warning
8964 LOC: Config.warnings.high_pf
8966 DEFAULT_DOC: disabled.
8968 If the one-minute average page fault rate exceeds this
8969 value, Squid prints a WARNING with debug level 0 to get
8970 the administrators attention. The value is in page faults
8974 NAME: high_memory_warning
8976 LOC: Config.warnings.high_memory
8977 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
8979 DEFAULT_DOC: disabled.
8981 If the memory usage (as determined by gnumalloc, if available and used)
8982 exceeds this amount, Squid prints a WARNING with debug level 0 to get
8983 the administrators attention.
8985 # TODO: link high_memory_warning to mempools?
8987 NAME: sleep_after_fork
8988 COMMENT: (microseconds)
8990 LOC: Config.sleep_after_fork
8993 When this is set to a non-zero value, the main Squid process
8994 sleeps the specified number of microseconds after a fork()
8995 system call. This sleep may help the situation where your
8996 system reports fork() failures due to lack of (virtual)
8997 memory. Note, however, if you have a lot of child
8998 processes, these sleep delays will add up and your
8999 Squid will not service requests for some amount of time
9000 until all the child processes have been started.
9001 On Windows value less then 1000 (1 milliseconds) are
9005 NAME: windows_ipaddrchangemonitor
9006 IFDEF: _SQUID_WINDOWS_
9010 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9012 On Windows Squid by default will monitor IP address changes and will
9013 reconfigure itself after any detected event. This is very useful for
9014 proxies connected to internet with dial-up interfaces.
9015 In some cases (a Proxy server acting as VPN gateway is one) it could be
9016 desiderable to disable this behaviour setting this to 'off'.
9017 Note: after changing this, Squid service must be restarted.
9022 IFDEF: USE_SQUID_EUI
9024 LOC: Eui::TheConfig.euiLookup
9026 Whether to lookup the EUI or MAC address of a connected client.
9029 NAME: max_filedescriptors max_filedesc
9032 DEFAULT_DOC: Use operating system limits set by ulimit.
9033 LOC: Config.max_filedescriptors
9035 Reduce the maximum number of filedescriptors supported below
9036 the usual operating system defaults.
9038 Remove from squid.conf to inherit the current ulimit setting.
9040 Note: Changing this requires a restart of Squid. Also
9041 not all I/O types supports large values (eg on Windows).
9048 DEFAULT_DOC: SMP support disabled.
9050 Number of main Squid processes or "workers" to fork and maintain.
9051 0: "no daemon" mode, like running "squid -N ..."
9052 1: "no SMP" mode, start one main Squid process daemon (default)
9053 N: start N main Squid process daemons (i.e., SMP mode)
9055 In SMP mode, each worker does nearly all what a single Squid daemon
9056 does (e.g., listen on http_port and forward HTTP requests).
9059 NAME: cpu_affinity_map
9060 TYPE: CpuAffinityMap
9061 LOC: Config.cpuAffinityMap
9063 DEFAULT_DOC: Let operating system decide.
9065 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
9067 Sets 1:1 mapping between Squid processes and CPU cores. For example,
9069 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
9071 affects processes 1 through 4 only and places them on the first
9072 four even cores, starting with core #1.
9074 CPU cores are numbered starting from 1. Requires support for
9075 sched_getaffinity(2) and sched_setaffinity(2) system calls.
9077 Multiple cpu_affinity_map options are merged.