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 3.5
173 NAME: hierarchy_stoplist
176 Remove this line. Use always_direct or cache_peer_access ACLs instead if you need to prevent cache_peer use.
179 # Options Removed in 3.3
180 NAME: ignore_ims_on_miss
183 Remove this line. The HTTP/1.1 feature is now configured by 'cache_miss_revalidate'.
186 # Options Removed in 3.2
187 NAME: ignore_expect_100
190 Remove this line. The HTTP/1.1 feature is now fully supported by default.
193 NAME: dns_v4_fallback
196 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
202 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
205 NAME: maximum_single_addr_tries
208 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
214 Remove this line. The feature is supported by default in storage types where update is implemented.
217 NAME: url_rewrite_concurrency
220 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
223 # Options Removed in 3.1
227 Remove this line. DNS is no longer tested on startup.
230 NAME: extension_methods
233 Remove this line. All valid methods for HTTP are accepted by default.
236 # 2.7 Options Removed/Replaced in 3.2
241 # 2.7 Options Removed/Replaced in 3.1
249 Remove this line. HTTP/1.1 is supported by default.
252 NAME: upgrade_http0.9
255 Remove this line. ICY/1.0 streaming protocol is supported by default.
258 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
261 Alter these entries. Use the qos_flows directive instead.
264 # Options Removed in 3.0
268 Since squid-3.0 replace with request_header_access or reply_header_access
269 depending on whether you wish to match client requests or server replies.
272 NAME: httpd_accel_no_pmtu_disc
275 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
278 NAME: wais_relay_host
281 Replace this line with 'cache_peer' configuration.
284 NAME: wais_relay_port
287 Replace this line with 'cache_peer' configuration.
291 OPTIONS FOR AUTHENTICATION
292 -----------------------------------------------------------------------------
301 This is used to define parameters for the various authentication
302 schemes supported by Squid.
304 format: auth_param scheme parameter [setting]
306 The order in which authentication schemes are presented to the client is
307 dependent on the order the scheme first appears in config file. IE
308 has a bug (it's not RFC 2617 compliant) in that it will use the basic
309 scheme if basic is the first entry presented, even if more secure
310 schemes are presented. For now use the order in the recommended
311 settings section below. If other browsers have difficulties (don't
312 recognize the schemes offered even if you are using basic) either
313 put basic first, or disable the other schemes (by commenting out their
316 Once an authentication scheme is fully configured, it can only be
317 shutdown by shutting squid down and restarting. Changes can be made on
318 the fly and activated with a reconfigure. I.E. You can change to a
319 different helper, but not unconfigure the helper completely.
321 Please note that while this directive defines how Squid processes
322 authentication it does not automatically activate authentication.
323 To use authentication you must in addition make use of ACLs based
324 on login name in http_access (proxy_auth, proxy_auth_regex or
325 external with %LOGIN used in the format tag). The browser will be
326 challenged for authentication on the first such acl encountered
327 in http_access processing and will also be re-challenged for new
328 login credentials if the request is being denied by a proxy_auth
331 WARNING: authentication can't be used in a transparently intercepting
332 proxy as the client then thinks it is talking to an origin server and
333 not the proxy. This is a limitation of bending the TCP/IP protocol to
334 transparently intercepting port 80, not a limitation in Squid.
335 Ports flagged 'transparent', 'intercept', or 'tproxy' have
336 authentication disabled.
338 === Parameters common to all schemes. ===
341 Specifies the command for the external authenticator.
343 By default, each authentication scheme is not used unless a
344 program is specified.
346 See http://wiki.squid-cache.org/Features/AddonHelpers for
347 more details on helper operations and creating your own.
350 Specifies a string to be append to request line format for
351 the authentication helper. "Quoted" format values may contain
352 spaces and logformat %macros. In theory, any logformat %macro
353 can be used. In practice, a %macro expands as a dash (-) if
354 the helper request is sent before the required macro
355 information is available to Squid.
357 By default, Squid uses request formats provided in
358 scheme-specific examples below (search for %credentials).
360 The expanded key_extras value is added to the Squid credentials
361 cache and, hence, will affect authentication. It can be used to
362 autenticate different users with identical user names (e.g.,
363 when user authentication depends on http_port).
365 Avoid adding frequently changing information to key_extras. For
366 example, if you add user source IP, and it changes frequently
367 in your environment, then max_user_ip ACL is going to treat
368 every user+IP combination as a unique "user", breaking the ACL
369 and wasting a lot of memory on those user records. It will also
370 force users to authenticate from scratch whenever their IP
374 Specifies the protection scope (aka realm name) which is to be
375 reported to the client for the authentication scheme. It is
376 commonly part of the text the user will see when prompted for
377 their username and password.
379 For Basic the default is "Squid proxy-caching web server".
380 For Digest there is no default, this parameter is mandatory.
381 For NTLM and Negotiate this parameter is ignored.
383 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
385 The maximum number of authenticator processes to spawn. If
386 you start too few Squid will have to wait for them to process
387 a backlog of credential verifications, slowing it down. When
388 password verifications are done via a (slow) network you are
389 likely to need lots of authenticator processes.
391 The startup= and idle= options permit some skew in the exact
392 amount run. A minimum of startup=N will begin during startup
393 and reconfigure. Squid will start more in groups of up to
394 idle=N in an attempt to meet traffic needs and to keep idle=N
395 free above those traffic needs up to the maximum.
397 The concurrency= option sets the number of concurrent requests
398 the helper can process. The default of 0 is used for helpers
399 who only supports one request at a time. Setting this to a
400 number greater than 0 changes the protocol used to include a
401 channel ID field first on the request/response line, allowing
402 multiple requests to be sent to the same helper in parallel
403 without waiting for the response.
405 Concurrency must not be set unless it's known the helper
406 supports the input format with channel-ID fields.
408 NOTE: NTLM and Negotiate schemes do not support concurrency
409 in the Squid code module even though some helpers can.
412 IF HAVE_AUTH_MODULE_BASIC
413 === Basic authentication parameters ===
416 HTTP uses iso-latin-1 as character set, while some
417 authentication backends such as LDAP expects UTF-8. If this is
418 set to on Squid will translate the HTTP iso-latin-1 charset to
419 UTF-8 before sending the username and password to the helper.
421 "credentialsttl" timetolive
422 Specifies how long squid assumes an externally validated
423 username:password pair is valid for - in other words how
424 often the helper program is called for that user. Set this
425 low to force revalidation with short lived passwords.
427 NOTE: setting this high does not impact your susceptibility
428 to replay attacks unless you are using an one-time password
429 system (such as SecureID). If you are using such a system,
430 you will be vulnerable to replay attacks unless you also
431 use the max_user_ip ACL in an http_access rule.
433 "casesensitive" on|off
434 Specifies if usernames are case sensitive. Most user databases
435 are case insensitive allowing the same username to be spelled
436 using both lower and upper case letters, but some are case
437 sensitive. This makes a big difference for user_max_ip ACL
438 processing and similar.
441 IF HAVE_AUTH_MODULE_DIGEST
442 === Digest authentication parameters ===
445 HTTP uses iso-latin-1 as character set, while some
446 authentication backends such as LDAP expects UTF-8. If this is
447 set to on Squid will translate the HTTP iso-latin-1 charset to
448 UTF-8 before sending the username and password to the helper.
450 "nonce_garbage_interval" timeinterval
451 Specifies the interval that nonces that have been issued
452 to client_agent's are checked for validity.
454 "nonce_max_duration" timeinterval
455 Specifies the maximum length of time a given nonce will be
458 "nonce_max_count" number
459 Specifies the maximum number of times a given nonce can be
462 "nonce_strictness" on|off
463 Determines if squid requires strict increment-by-1 behavior
464 for nonce counts, or just incrementing (off - for use when
465 user agents generate nonce counts that occasionally miss 1
466 (ie, 1,2,4,6)). Default off.
468 "check_nonce_count" on|off
469 This directive if set to off can disable the nonce count check
470 completely to work around buggy digest qop implementations in
471 certain mainstream browser versions. Default on to check the
472 nonce count to protect from authentication replay attacks.
474 "post_workaround" on|off
475 This is a workaround to certain buggy browsers who send an
476 incorrect request digest in POST requests when reusing the
477 same nonce as acquired earlier on a GET request.
480 IF HAVE_AUTH_MODULE_NEGOTIATE
481 === Negotiate authentication parameters ===
484 If you experience problems with PUT/POST requests when using
485 the this authentication scheme then you can try setting this
486 to off. This will cause Squid to forcibly close the connection
487 on the initial request where the browser asks which schemes
488 are supported by the proxy.
491 IF HAVE_AUTH_MODULE_NTLM
492 === NTLM authentication parameters ===
495 If you experience problems with PUT/POST requests when using
496 the this authentication scheme then you can try setting this
497 to off. This will cause Squid to forcibly close the connection
498 on the initial request where the browser asks which schemes
499 are supported by the proxy.
502 === Example Configuration ===
504 This configuration displays the recommended authentication scheme
505 order from most to least secure with recommended minimum configuration
506 settings for each scheme:
508 #auth_param negotiate program <uncomment and complete this line to activate>
509 #auth_param negotiate children 20 startup=0 idle=1
510 #auth_param negotiate keep_alive on
512 #auth_param digest program <uncomment and complete this line to activate>
513 #auth_param digest children 20 startup=0 idle=1
514 #auth_param digest realm Squid proxy-caching web server
515 #auth_param digest nonce_garbage_interval 5 minutes
516 #auth_param digest nonce_max_duration 30 minutes
517 #auth_param digest nonce_max_count 50
519 #auth_param ntlm program <uncomment and complete this line to activate>
520 #auth_param ntlm children 20 startup=0 idle=1
521 #auth_param ntlm keep_alive on
523 #auth_param basic program <uncomment and complete this line>
524 #auth_param basic children 5 startup=5 idle=1
525 #auth_param basic realm Squid proxy-caching web server
526 #auth_param basic credentialsttl 2 hours
529 NAME: authenticate_cache_garbage_interval
532 LOC: Config.authenticateGCInterval
534 The time period between garbage collection across the username cache.
535 This is a trade-off between memory utilization (long intervals - say
536 2 days) and CPU (short intervals - say 1 minute). Only change if you
540 NAME: authenticate_ttl
543 LOC: Config.authenticateTTL
545 The time a user & their credentials stay in the logged in
546 user cache since their last request. When the garbage
547 interval passes, all user credentials that have passed their
548 TTL are removed from memory.
551 NAME: authenticate_ip_ttl
553 LOC: Config.authenticateIpTTL
556 If you use proxy authentication and the 'max_user_ip' ACL,
557 this directive controls how long Squid remembers the IP
558 addresses associated with each user. Use a small value
559 (e.g., 60 seconds) if your users might change addresses
560 quickly, as is the case with dialup. You might be safe
561 using a larger value (e.g., 2 hours) in a corporate LAN
562 environment with relatively static address assignments.
567 -----------------------------------------------------------------------------
570 NAME: external_acl_type
571 TYPE: externalAclHelper
572 LOC: Config.externalAclHelperList
575 This option defines external acl classes using a helper program
576 to look up the status
578 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
582 ttl=n TTL in seconds for cached results (defaults to 3600
585 TTL for cached negative lookups (default same
588 Maximum number of acl helper processes spawned to service
589 external acl lookups of this type. (default 20)
591 Minimum number of acl helper processes to spawn during
592 startup and reconfigure to service external acl lookups
593 of this type. (default 0)
595 Number of acl helper processes to keep ahead of traffic
596 loads. Squid will spawn this many at once whenever load
597 rises above the capabilities of existing processes.
598 Up to the value of children-max. (default 1)
599 concurrency=n concurrency level per process. Only used with helpers
600 capable of processing more than one query at a time.
601 cache=n limit the result cache size, default is 262144.
602 grace=n Percentage remaining of TTL where a refresh of a
603 cached entry should be initiated without needing to
604 wait for a new reply. (default is for no grace period)
605 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
606 ipv4 / ipv6 IP protocol used to communicate with this helper.
607 The default is to auto-detect IPv6 and use it when available.
609 FORMAT specifications
611 %LOGIN Authenticated user login name
612 %EXT_USER Username from previous external acl
613 %EXT_LOG Log details from previous external acl
614 %EXT_TAG Tag from previous external acl
615 %IDENT Ident user name
617 %SRCPORT Client source port
620 %PROTO Requested URL scheme
622 %PATH Requested URL path
623 %METHOD Request method
624 %MYADDR Squid interface address
625 %MYPORT Squid http_port number
626 %PATH Requested URL-path (including query-string if any)
627 %USER_CERT SSL User certificate in PEM format
628 %USER_CERTCHAIN SSL User certificate chain in PEM format
629 %USER_CERT_xx SSL User certificate subject attribute xx
630 %USER_CA_CERT_xx SSL User certificate issuer attribute xx
632 %>{Header} HTTP request header "Header"
634 HTTP request header "Hdr" list member "member"
636 HTTP request header list member using ; as
637 list separator. ; can be any non-alphanumeric
640 %<{Header} HTTP reply header "Header"
642 HTTP reply header "Hdr" list member "member"
644 HTTP reply header list member using ; as
645 list separator. ; can be any non-alphanumeric
648 %ACL The name of the ACL being tested.
649 %DATA The ACL arguments. If not used then any arguments
650 is automatically added at the end of the line
652 NOTE: this will encode the arguments as one token,
653 whereas the default will pass each separately.
655 %% The percent sign. Useful for helpers which need
656 an unchanging input format.
659 General request syntax:
661 [channel-ID] FORMAT-values [acl-values ...]
664 FORMAT-values consists of transaction details expanded with
665 whitespace separation per the config file FORMAT specification
666 using the FORMAT macros listed above.
668 acl-values consists of any string specified in the referencing
669 config 'acl ... external' line. see the "acl external" directive.
671 Request values sent to the helper are URL escaped to protect
672 each value in requests against whitespaces.
674 If using protocol=2.5 then the request sent to the helper is not
675 URL escaped to protect against whitespace.
677 NOTE: protocol=3.0 is deprecated as no longer necessary.
679 When using the concurrency= option the protocol is changed by
680 introducing a query channel tag in front of the request/response.
681 The query channel tag is a number between 0 and concurrency-1.
682 This value must be echoed back unchanged to Squid as the first part
683 of the response relating to its request.
686 The helper receives lines expanded per the above format specification
687 and for each input line returns 1 line starting with OK/ERR/BH result
688 code and optionally followed by additional keywords with more details.
691 General result syntax:
693 [channel-ID] result keyword=value ...
695 Result consists of one of the codes:
698 the ACL test produced a match.
701 the ACL test does not produce a match.
704 An internal error occurred in the helper, preventing
705 a result being identified.
707 The meaning of 'a match' is determined by your squid.conf
708 access control configuration. See the Squid wiki for details.
712 user= The users name (login)
714 password= The users password (for login= cache_peer option)
716 message= Message describing the reason for this response.
717 Available as %o in error pages.
718 Useful on (ERR and BH results).
720 tag= Apply a tag to a request. Only sets a tag once,
721 does not alter existing tags.
723 log= String to be logged in access.log. Available as
724 %ea in logformat specifications.
726 clt_conn_tag= Associates a TAG with the client TCP connection.
727 Please see url_rewrite_program related documentation for
730 Any keywords may be sent on any response whether OK, ERR or BH.
732 All response keyword values need to be a single token with URL
733 escaping, or enclosed in double quotes (") and escaped using \ on
734 any double quotes or \ characters within the value. The wrapping
735 double quotes are removed before the value is interpreted by Squid.
736 \r and \n are also replace by CR and LF.
738 Some example key values:
742 user="J. \"Bob\" Smith"
749 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
750 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
751 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
752 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
753 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
756 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
757 DEFAULT: localhost src 127.0.0.1/32 ::1
758 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
759 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
761 Defining an Access List
763 Every access list definition must begin with an aclname and acltype,
764 followed by either type-specific arguments or a quoted filename that
767 acl aclname acltype argument ...
768 acl aclname acltype "file" ...
770 When using "file", the file should contain one item per line.
772 Some acl types supports options which changes their default behaviour.
773 The available options are:
775 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
776 case-insensitive, use the -i option. To return case-sensitive
777 use the +i option between patterns, or make a new ACL line
780 -n Disable lookups and address type conversions. If lookup or
781 conversion is required because the parameter type (IP or
782 domain name) does not match the message address type (domain
783 name or IP), then the ACL would immediately declare a mismatch
784 without any warnings or lookups.
786 -- Used to stop processing all options, in the case the first acl
787 value has '-' character as first character (for example the '-'
788 is a valid domain name)
790 Some acl types require suspending the current request in order
791 to access some external data source.
792 Those which do are marked with the tag [slow], those which
793 don't are marked as [fast].
794 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
795 for further information
797 ***** ACL TYPES AVAILABLE *****
799 acl aclname src ip-address/mask ... # clients IP address [fast]
800 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
801 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
802 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
804 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
805 # The arp ACL requires the special configure option --enable-arp-acl.
806 # Furthermore, the ARP ACL code is not portable to all operating systems.
807 # It works on Linux, Solaris, Windows, FreeBSD, and some
808 # other *BSD variants.
811 # NOTE: Squid can only determine the MAC address for clients that are on
812 # the same subnet. If the client is on a different subnet,
813 # then Squid cannot find out its MAC address.
815 acl aclname srcdomain .foo.com ...
816 # reverse lookup, from client IP [slow]
817 acl aclname dstdomain [-n] .foo.com ...
818 # Destination server from URL [fast]
819 acl aclname srcdom_regex [-i] \.foo\.com ...
820 # regex matching client name [slow]
821 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
822 # regex matching server [fast]
824 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
825 # based URL is used and no match is found. The name "none" is used
826 # if the reverse lookup fails.
828 acl aclname src_as number ...
829 acl aclname dst_as number ...
831 # Except for access control, AS numbers can be used for
832 # routing of requests to specific caches. Here's an
833 # example for routing all requests for AS#1241 and only
834 # those to mycache.mydomain.net:
835 # acl asexample dst_as 1241
836 # cache_peer_access mycache.mydomain.net allow asexample
837 # cache_peer_access mycache_mydomain.net deny all
839 acl aclname peername myPeer ...
841 # match against a named cache_peer entry
842 # set unique name= on cache_peer lines for reliable use.
844 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
854 # h1:m1 must be less than h2:m2
856 acl aclname url_regex [-i] ^http:// ...
857 # regex matching on whole URL [fast]
858 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
859 # regex matching on URL login field
860 acl aclname urlpath_regex [-i] \.gif$ ...
861 # regex matching on URL path [fast]
863 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
865 acl aclname localport 3128 ... # TCP port the client connected to [fast]
866 # NP: for interception mode this is usually '80'
868 acl aclname myportname 3128 ... # http(s)_port name [fast]
870 acl aclname proto HTTP FTP ... # request protocol [fast]
872 acl aclname method GET POST ... # HTTP request method [fast]
874 acl aclname http_status 200 301 500- 400-403 ...
875 # status code in reply [fast]
877 acl aclname browser [-i] regexp ...
878 # pattern match on User-Agent header (see also req_header below) [fast]
880 acl aclname referer_regex [-i] regexp ...
881 # pattern match on Referer header [fast]
882 # Referer is highly unreliable, so use with care
884 acl aclname ident username ...
885 acl aclname ident_regex [-i] pattern ...
886 # string match on ident output [slow]
887 # use REQUIRED to accept any non-null ident.
889 acl aclname proxy_auth [-i] username ...
890 acl aclname proxy_auth_regex [-i] pattern ...
891 # perform http authentication challenge to the client and match against
892 # supplied credentials [slow]
894 # takes a list of allowed usernames.
895 # use REQUIRED to accept any valid username.
897 # Will use proxy authentication in forward-proxy scenarios, and plain
898 # http authenticaiton in reverse-proxy scenarios
900 # NOTE: when a Proxy-Authentication header is sent but it is not
901 # needed during ACL checking the username is NOT logged
904 # NOTE: proxy_auth requires a EXTERNAL authentication program
905 # to check username/password combinations (see
906 # auth_param directive).
908 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
909 # as the browser needs to be configured for using a proxy in order
910 # to respond to proxy authentication.
912 acl aclname snmp_community string ...
913 # A community string to limit access to your SNMP Agent [fast]
916 # acl snmppublic snmp_community public
918 acl aclname maxconn number
919 # This will be matched when the client's IP address has
920 # more than <number> TCP connections established. [fast]
921 # NOTE: This only measures direct TCP links so X-Forwarded-For
922 # indirect clients are not counted.
924 acl aclname max_user_ip [-s] number
925 # This will be matched when the user attempts to log in from more
926 # than <number> different ip addresses. The authenticate_ip_ttl
927 # parameter controls the timeout on the ip entries. [fast]
928 # If -s is specified the limit is strict, denying browsing
929 # from any further IP addresses until the ttl has expired. Without
930 # -s Squid will just annoy the user by "randomly" denying requests.
931 # (the counter is reset each time the limit is reached and a
933 # NOTE: in acceleration mode or where there is mesh of child proxies,
934 # clients may appear to come from multiple addresses if they are
935 # going through proxy farms, so a limit of 1 may cause user problems.
937 acl aclname random probability
938 # Pseudo-randomly match requests. Based on the probability given.
939 # Probability may be written as a decimal (0.333), fraction (1/3)
940 # or ratio of matches:non-matches (3:5).
942 acl aclname req_mime_type [-i] mime-type ...
943 # regex match against the mime type of the request generated
944 # by the client. Can be used to detect file upload or some
945 # types HTTP tunneling requests [fast]
946 # NOTE: This does NOT match the reply. You cannot use this
947 # to match the returned file type.
949 acl aclname req_header header-name [-i] any\.regex\.here
950 # regex match against any of the known request headers. May be
951 # thought of as a superset of "browser", "referer" and "mime-type"
954 acl aclname rep_mime_type [-i] mime-type ...
955 # regex match against the mime type of the reply received by
956 # squid. Can be used to detect file download or some
957 # types HTTP tunneling requests. [fast]
958 # NOTE: This has no effect in http_access rules. It only has
959 # effect in rules that affect the reply data stream such as
962 acl aclname rep_header header-name [-i] any\.regex\.here
963 # regex match against any of the known reply headers. May be
964 # thought of as a superset of "browser", "referer" and "mime-type"
967 acl aclname external class_name [arguments...]
968 # external ACL lookup via a helper class defined by the
969 # external_acl_type directive [slow]
971 acl aclname user_cert attribute values...
972 # match against attributes in a user SSL certificate
973 # attribute is one of DN/C/O/CN/L/ST [fast]
975 acl aclname ca_cert attribute values...
976 # match against attributes a users issuing CA SSL certificate
977 # attribute is one of DN/C/O/CN/L/ST [fast]
979 acl aclname ext_user username ...
980 acl aclname ext_user_regex [-i] pattern ...
981 # string match on username returned by external acl helper [slow]
982 # use REQUIRED to accept any non-null user name.
984 acl aclname tag tagvalue ...
985 # string match on tag returned by external acl helper [fast]
986 # DEPRECATED. Only the first tag will match with this ACL.
987 # Use the 'note' ACL instead for handling multiple tag values.
989 acl aclname hier_code codename ...
990 # string match against squid hierarchy code(s); [fast]
991 # e.g., DIRECT, PARENT_HIT, NONE, etc.
993 # NOTE: This has no effect in http_access rules. It only has
994 # effect in rules that affect the reply data stream such as
997 acl aclname note name [value ...]
998 # match transaction annotation [fast]
999 # Without values, matches any annotation with a given name.
1000 # With value(s), matches any annotation with a given name that
1001 # also has one of the given values.
1002 # Names and values are compared using a string equality test.
1003 # Annotation sources include note and adaptation_meta directives
1004 # as well as helper and eCAP responses.
1006 acl aclname adaptation_service service ...
1007 # Matches the name of any icap_service, ecap_service,
1008 # adaptation_service_set, or adaptation_service_chain that Squid
1009 # has used (or attempted to use) for the master transaction.
1010 # This ACL must be defined after the corresponding adaptation
1011 # service is named in squid.conf. This ACL is usable with
1012 # adaptation_meta because it starts matching immediately after
1013 # the service has been selected for adaptation.
1016 acl aclname ssl_error errorname
1017 # match against SSL certificate validation error [fast]
1019 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1022 # The following can be used as shortcuts for certificate properties:
1023 # [ssl::]certHasExpired: the "not after" field is in the past
1024 # [ssl::]certNotYetValid: the "not before" field is in the future
1025 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1026 # [ssl::]certSelfSigned: The certificate is self signed.
1027 # [ssl::]certDomainMismatch: The certificate CN domain does not
1028 # match the name the name of the host we are connecting to.
1030 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1031 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1032 # predefined ACLs, just like the 'all' ACL.
1034 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1035 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1037 acl aclname server_cert_fingerprint [-sha1] fingerprint
1038 # match against server SSL certificate fingerprint [fast]
1040 # The fingerprint is the digest of the DER encoded version
1041 # of the whole certificate. The user should use the form: XX:XX:...
1042 # Optional argument specifies the digest algorithm to use.
1043 # The SHA1 digest algorithm is the default and is currently
1044 # the only algorithm supported (-sha1).
1046 acl aclname any-of acl1 acl2 ...
1047 # match any one of the acls [fast or slow]
1048 # The first matching ACL stops further ACL evaluation.
1050 # ACLs from multiple any-of lines with the same name are ORed.
1051 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1052 # acl A any-of a1 a2
1053 # acl A any-of a3 a4
1055 # This group ACL is fast if all evaluated ACLs in the group are fast
1056 # and slow otherwise.
1058 acl aclname all-of acl1 acl2 ...
1059 # match all of the acls [fast or slow]
1060 # The first mismatching ACL stops further ACL evaluation.
1062 # ACLs from multiple all-of lines with the same name are ORed.
1063 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1064 # acl B all-of b1 b2
1065 # acl B all-of b3 b4
1067 # This group ACL is fast if all evaluated ACLs in the group are fast
1068 # and slow otherwise.
1071 acl macaddress arp 09:00:2b:23:45:67
1072 acl myexample dst_as 1241
1073 acl password proxy_auth REQUIRED
1074 acl fileupload req_mime_type -i ^multipart/form-data$
1075 acl javascript rep_mime_type -i ^application/x-javascript$
1079 # Recommended minimum configuration:
1082 # Example rule allowing access from your local networks.
1083 # Adapt to list your (internal) IP networks from where browsing
1085 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1086 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1087 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1088 acl localnet src fc00::/7 # RFC 4193 local private network range
1089 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1091 acl SSL_ports port 443
1092 acl Safe_ports port 80 # http
1093 acl Safe_ports port 21 # ftp
1094 acl Safe_ports port 443 # https
1095 acl Safe_ports port 70 # gopher
1096 acl Safe_ports port 210 # wais
1097 acl Safe_ports port 1025-65535 # unregistered ports
1098 acl Safe_ports port 280 # http-mgmt
1099 acl Safe_ports port 488 # gss-http
1100 acl Safe_ports port 591 # filemaker
1101 acl Safe_ports port 777 # multiling http
1102 acl CONNECT method CONNECT
1106 NAME: follow_x_forwarded_for
1108 IFDEF: FOLLOW_X_FORWARDED_FOR
1109 LOC: Config.accessList.followXFF
1110 DEFAULT_IF_NONE: deny all
1111 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1113 Allowing or Denying the X-Forwarded-For header to be followed to
1114 find the original source of a request.
1116 Requests may pass through a chain of several other proxies
1117 before reaching us. The X-Forwarded-For header will contain a
1118 comma-separated list of the IP addresses in the chain, with the
1119 rightmost address being the most recent.
1121 If a request reaches us from a source that is allowed by this
1122 configuration item, then we consult the X-Forwarded-For header
1123 to see where that host received the request from. If the
1124 X-Forwarded-For header contains multiple addresses, we continue
1125 backtracking until we reach an address for which we are not allowed
1126 to follow the X-Forwarded-For header, or until we reach the first
1127 address in the list. For the purpose of ACL used in the
1128 follow_x_forwarded_for directive the src ACL type always matches
1129 the address we are testing and srcdomain matches its rDNS.
1131 The end result of this process is an IP address that we will
1132 refer to as the indirect client address. This address may
1133 be treated as the client address for access control, ICAP, delay
1134 pools and logging, depending on the acl_uses_indirect_client,
1135 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1136 log_uses_indirect_client and tproxy_uses_indirect_client options.
1138 This clause only supports fast acl types.
1139 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1141 SECURITY CONSIDERATIONS:
1143 Any host for which we follow the X-Forwarded-For header
1144 can place incorrect information in the header, and Squid
1145 will use the incorrect information as if it were the
1146 source address of the request. This may enable remote
1147 hosts to bypass any access control restrictions that are
1148 based on the client's source addresses.
1152 acl localhost src 127.0.0.1
1153 acl my_other_proxy srcdomain .proxy.example.com
1154 follow_x_forwarded_for allow localhost
1155 follow_x_forwarded_for allow my_other_proxy
1158 NAME: acl_uses_indirect_client
1161 IFDEF: FOLLOW_X_FORWARDED_FOR
1163 LOC: Config.onoff.acl_uses_indirect_client
1165 Controls whether the indirect client address
1166 (see follow_x_forwarded_for) is used instead of the
1167 direct client address in acl matching.
1169 NOTE: maxconn ACL considers direct TCP links and indirect
1170 clients will always have zero. So no match.
1173 NAME: delay_pool_uses_indirect_client
1176 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1178 LOC: Config.onoff.delay_pool_uses_indirect_client
1180 Controls whether the indirect client address
1181 (see follow_x_forwarded_for) is used instead of the
1182 direct client address in delay pools.
1185 NAME: log_uses_indirect_client
1188 IFDEF: FOLLOW_X_FORWARDED_FOR
1190 LOC: Config.onoff.log_uses_indirect_client
1192 Controls whether the indirect client address
1193 (see follow_x_forwarded_for) is used instead of the
1194 direct client address in the access log.
1197 NAME: tproxy_uses_indirect_client
1200 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1202 LOC: Config.onoff.tproxy_uses_indirect_client
1204 Controls whether the indirect client address
1205 (see follow_x_forwarded_for) is used instead of the
1206 direct client address when spoofing the outgoing client.
1208 This has no effect on requests arriving in non-tproxy
1211 SECURITY WARNING: Usage of this option is dangerous
1212 and should not be used trivially. Correct configuration
1213 of follow_x_forewarded_for with a limited set of trusted
1214 sources is required to prevent abuse of your proxy.
1217 NAME: spoof_client_ip
1219 LOC: Config.accessList.spoof_client_ip
1221 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1223 Control client IP address spoofing of TPROXY traffic based on
1224 defined access lists.
1226 spoof_client_ip allow|deny [!]aclname ...
1228 If there are no "spoof_client_ip" lines present, the default
1229 is to "allow" spoofing of any suitable request.
1231 Note that the cache_peer "no-tproxy" option overrides this ACL.
1233 This clause supports fast acl types.
1234 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1239 LOC: Config.accessList.http
1240 DEFAULT_IF_NONE: deny all
1241 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1243 Allowing or Denying access based on defined access lists
1245 Access to the HTTP port:
1246 http_access allow|deny [!]aclname ...
1248 NOTE on default values:
1250 If there are no "access" lines present, the default is to deny
1253 If none of the "access" lines cause a match, the default is the
1254 opposite of the last line in the list. If the last line was
1255 deny, the default is allow. Conversely, if the last line
1256 is allow, the default will be deny. For these reasons, it is a
1257 good idea to have an "deny all" entry at the end of your access
1258 lists to avoid potential confusion.
1260 This clause supports both fast and slow acl types.
1261 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1266 # Recommended minimum Access Permission configuration:
1268 # Deny requests to certain unsafe ports
1269 http_access deny !Safe_ports
1271 # Deny CONNECT to other than secure SSL ports
1272 http_access deny CONNECT !SSL_ports
1274 # Only allow cachemgr access from localhost
1275 http_access allow localhost manager
1276 http_access deny manager
1278 # We strongly recommend the following be uncommented to protect innocent
1279 # web applications running on the proxy server who think the only
1280 # one who can access services on "localhost" is a local user
1281 #http_access deny to_localhost
1284 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1287 # Example rule allowing access from your local networks.
1288 # Adapt localnet in the ACL section to list your (internal) IP networks
1289 # from where browsing should be allowed
1290 http_access allow localnet
1291 http_access allow localhost
1293 # And finally deny all other access to this proxy
1294 http_access deny all
1298 NAME: adapted_http_access http_access2
1300 LOC: Config.accessList.adapted_http
1302 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1304 Allowing or Denying access based on defined access lists
1306 Essentially identical to http_access, but runs after redirectors
1307 and ICAP/eCAP adaptation. Allowing access control based on their
1310 If not set then only http_access is used.
1313 NAME: http_reply_access
1315 LOC: Config.accessList.reply
1317 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1319 Allow replies to client requests. This is complementary to http_access.
1321 http_reply_access allow|deny [!] aclname ...
1323 NOTE: if there are no access lines present, the default is to allow
1326 If none of the access lines cause a match the opposite of the
1327 last line will apply. Thus it is good practice to end the rules
1328 with an "allow all" or "deny all" entry.
1330 This clause supports both fast and slow acl types.
1331 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1336 LOC: Config.accessList.icp
1338 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1340 Allowing or Denying access to the ICP port based on defined
1343 icp_access allow|deny [!]aclname ...
1345 NOTE: The default if no icp_access lines are present is to
1346 deny all traffic. This default may cause problems with peers
1349 This clause only supports fast acl types.
1350 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1352 # Allow ICP queries from local networks only
1353 #icp_access allow localnet
1354 #icp_access deny all
1360 LOC: Config.accessList.htcp
1362 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1364 Allowing or Denying access to the HTCP port based on defined
1367 htcp_access allow|deny [!]aclname ...
1369 See also htcp_clr_access for details on access control for
1370 cache purge (CLR) HTCP messages.
1372 NOTE: The default if no htcp_access lines are present is to
1373 deny all traffic. This default may cause problems with peers
1374 using the htcp option.
1376 This clause only supports fast acl types.
1377 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1379 # Allow HTCP queries from local networks only
1380 #htcp_access allow localnet
1381 #htcp_access deny all
1384 NAME: htcp_clr_access
1387 LOC: Config.accessList.htcp_clr
1389 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1391 Allowing or Denying access to purge content using HTCP based
1392 on defined access lists.
1393 See htcp_access for details on general HTCP access control.
1395 htcp_clr_access allow|deny [!]aclname ...
1397 This clause only supports fast acl types.
1398 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1400 # Allow HTCP CLR requests from trusted peers
1401 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1402 htcp_clr_access allow htcp_clr_peer
1403 htcp_clr_access deny all
1408 LOC: Config.accessList.miss
1410 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1412 Determins whether network access is permitted when satisfying a request.
1415 to force your neighbors to use you as a sibling instead of
1418 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1419 miss_access deny !localclients
1420 miss_access allow all
1422 This means only your local clients are allowed to fetch relayed/MISS
1423 replies from the network and all other clients can only fetch cached
1426 The default for this setting allows all clients who passed the
1427 http_access rules to relay via this proxy.
1429 This clause only supports fast acl types.
1430 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1433 NAME: ident_lookup_access
1437 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1438 LOC: Ident::TheConfig.identLookup
1440 A list of ACL elements which, if matched, cause an ident
1441 (RFC 931) lookup to be performed for this request. For
1442 example, you might choose to always perform ident lookups
1443 for your main multi-user Unix boxes, but not for your Macs
1444 and PCs. By default, ident lookups are not performed for
1447 To enable ident lookups for specific client addresses, you
1448 can follow this example:
1450 acl ident_aware_hosts src 198.168.1.0/24
1451 ident_lookup_access allow ident_aware_hosts
1452 ident_lookup_access deny all
1454 Only src type ACL checks are fully supported. A srcdomain
1455 ACL might work at times, but it will not always provide
1458 This clause only supports fast acl types.
1459 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1462 NAME: reply_body_max_size
1463 COMMENT: size [acl acl...]
1466 DEFAULT_DOC: No limit is applied.
1467 LOC: Config.ReplyBodySize
1469 This option specifies the maximum size of a reply body. It can be
1470 used to prevent users from downloading very large files, such as
1471 MP3's and movies. When the reply headers are received, the
1472 reply_body_max_size lines are processed, and the first line where
1473 all (if any) listed ACLs are true is used as the maximum body size
1476 This size is checked twice. First when we get the reply headers,
1477 we check the content-length value. If the content length value exists
1478 and is larger than the allowed size, the request is denied and the
1479 user receives an error message that says "the request or reply
1480 is too large." If there is no content-length, and the reply
1481 size exceeds this limit, the client's connection is just closed
1482 and they will receive a partial reply.
1484 WARNING: downstream caches probably can not detect a partial reply
1485 if there is no content-length header, so they will cache
1486 partial responses and give them out as hits. You should NOT
1487 use this option if you have downstream caches.
1489 WARNING: A maximum size smaller than the size of squid's error messages
1490 will cause an infinite loop and crash squid. Ensure that the smallest
1491 non-zero value you use is greater that the maximum header size plus
1492 the size of your largest error page.
1494 If you set this parameter none (the default), there will be
1497 Configuration Format is:
1498 reply_body_max_size SIZE UNITS [acl ...]
1500 reply_body_max_size 10 MB
1506 -----------------------------------------------------------------------------
1509 NAME: http_port ascii_port
1514 Usage: port [mode] [options]
1515 hostname:port [mode] [options]
1516 1.2.3.4:port [mode] [options]
1518 The socket addresses where Squid will listen for HTTP client
1519 requests. You may specify multiple socket addresses.
1520 There are three forms: port alone, hostname with port, and
1521 IP address with port. If you specify a hostname or IP
1522 address, Squid binds the socket to that specific
1523 address. Most likely, you do not need to bind to a specific
1524 address, so you can use the port number alone.
1526 If you are running Squid in accelerator mode, you
1527 probably want to listen on port 80 also, or instead.
1529 The -a command line option may be used to specify additional
1530 port(s) where Squid listens for proxy request. Such ports will
1531 be plain proxy ports with no options.
1533 You may specify multiple socket addresses on multiple lines.
1537 intercept Support for IP-Layer interception of
1538 outgoing requests without browser settings.
1539 NP: disables authentication and IPv6 on the port.
1541 tproxy Support Linux TPROXY for spoofing outgoing
1542 connections using the client IP address.
1543 NP: disables authentication and maybe IPv6 on the port.
1545 accel Accelerator / reverse proxy mode
1547 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1548 establish secure connection with the client and with
1549 the server, decrypt HTTPS messages as they pass through
1550 Squid, and treat them as unencrypted HTTP messages,
1551 becoming the man-in-the-middle.
1553 The ssl_bump option is required to fully enable
1554 bumping of CONNECT requests.
1556 Omitting the mode flag causes default forward proxy mode to be used.
1559 Accelerator Mode Options:
1561 defaultsite=domainname
1562 What to use for the Host: header if it is not present
1563 in a request. Determines what site (not origin server)
1564 accelerators should consider the default.
1566 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1568 protocol= Protocol to reconstruct accelerated and intercepted
1569 requests with. Defaults to HTTP/1.1 for http_port and
1570 HTTPS/1.1 for https_port.
1571 When an unsupported value is configured Squid will
1572 produce a FATAL error.
1573 Values: HTTP or HTTP/1.1, HTTPS or HTTPS/1.1
1575 vport Virtual host port support. Using the http_port number
1576 instead of the port passed on Host: headers.
1578 vport=NN Virtual host port support. Using the specified port
1579 number instead of the port passed on Host: headers.
1582 Act as if this Squid is the origin server.
1583 This currently means generate new Date: and Expires:
1584 headers on HIT instead of adding Age:.
1586 ignore-cc Ignore request Cache-Control headers.
1588 WARNING: This option violates HTTP specifications if
1589 used in non-accelerator setups.
1591 allow-direct Allow direct forwarding in accelerator mode. Normally
1592 accelerated requests are denied direct forwarding as if
1593 never_direct was used.
1595 WARNING: this option opens accelerator mode to security
1596 vulnerabilities usually only affecting in interception
1597 mode. Make sure to protect forwarding with suitable
1598 http_access rules when using this.
1601 SSL Bump Mode Options:
1602 In addition to these options ssl-bump requires TLS/SSL options.
1604 generate-host-certificates[=<on|off>]
1605 Dynamically create SSL server certificates for the
1606 destination hosts of bumped CONNECT requests.When
1607 enabled, the cert and key options are used to sign
1608 generated certificates. Otherwise generated
1609 certificate will be selfsigned.
1610 If there is a CA certificate lifetime of the generated
1611 certificate equals lifetime of the CA certificate. If
1612 generated certificate is selfsigned lifetime is three
1614 This option is enabled by default when ssl-bump is used.
1615 See the ssl-bump option above for more information.
1617 dynamic_cert_mem_cache_size=SIZE
1618 Approximate total RAM size spent on cached generated
1619 certificates. If set to zero, caching is disabled. The
1620 default value is 4MB.
1624 cert= Path to SSL certificate (PEM format).
1626 key= Path to SSL private key file (PEM format)
1627 if not specified, the certificate file is
1628 assumed to be a combined certificate and
1631 version= The version of SSL/TLS supported
1632 1 automatic (default)
1639 cipher= Colon separated list of supported ciphers.
1640 NOTE: some ciphers such as EDH ciphers depend on
1641 additional settings. If those settings are
1642 omitted the ciphers may be silently ignored
1643 by the OpenSSL library.
1645 options= Various SSL implementation options. The most important
1647 NO_SSLv2 Disallow the use of SSLv2
1648 NO_SSLv3 Disallow the use of SSLv3
1649 NO_TLSv1 Disallow the use of TLSv1.0
1650 NO_TLSv1_1 Disallow the use of TLSv1.1
1651 NO_TLSv1_2 Disallow the use of TLSv1.2
1652 SINGLE_DH_USE Always create a new key when using
1653 temporary/ephemeral DH key exchanges
1654 ALL Enable various bug workarounds
1655 suggested as "harmless" by OpenSSL
1656 Be warned that this reduces SSL/TLS
1657 strength to some attacks.
1658 See OpenSSL SSL_CTX_set_options documentation for a
1659 complete list of options.
1661 clientca= File containing the list of CAs to use when
1662 requesting a client certificate.
1664 cafile= File containing additional CA certificates to
1665 use when verifying client certificates. If unset
1666 clientca will be used.
1668 capath= Directory containing additional CA certificates
1669 and CRL lists to use when verifying client certificates.
1671 crlfile= File of additional CRL lists to use when verifying
1672 the client certificate, in addition to CRLs stored in
1673 the capath. Implies VERIFY_CRL flag below.
1675 dhparams= File containing DH parameters for temporary/ephemeral
1676 DH key exchanges. See OpenSSL documentation for details
1677 on how to create this file.
1678 WARNING: EDH ciphers will be silently disabled if this
1681 sslflags= Various flags modifying the use of SSL:
1683 Don't request client certificates
1684 immediately, but wait until acl processing
1685 requires a certificate (not yet implemented).
1687 Don't use the default CA lists built in
1690 Don't allow for session reuse. Each connection
1691 will result in a new SSL session.
1693 Verify CRL lists when accepting client
1696 Verify CRL lists for all certificates in the
1697 client certificate chain.
1699 sslcontext= SSL session ID context identifier.
1703 connection-auth[=on|off]
1704 use connection-auth=off to tell Squid to prevent
1705 forwarding Microsoft connection oriented authentication
1706 (NTLM, Negotiate and Kerberos)
1708 disable-pmtu-discovery=
1709 Control Path-MTU discovery usage:
1710 off lets OS decide on what to do (default).
1711 transparent disable PMTU discovery when transparent
1713 always disable always PMTU discovery.
1715 In many setups of transparently intercepting proxies
1716 Path-MTU discovery can not work on traffic towards the
1717 clients. This is the case when the intercepting device
1718 does not fully track connections and fails to forward
1719 ICMP must fragment messages to the cache server. If you
1720 have such setup and experience that certain clients
1721 sporadically hang or never complete requests set
1722 disable-pmtu-discovery option to 'transparent'.
1724 name= Specifies a internal name for the port. Defaults to
1725 the port specification (port or addr:port)
1727 tcpkeepalive[=idle,interval,timeout]
1728 Enable TCP keepalive probes of idle connections.
1729 In seconds; idle is the initial time before TCP starts
1730 probing the connection, interval how often to probe, and
1731 timeout the time before giving up.
1733 If you run Squid on a dual-homed machine with an internal
1734 and an external interface we recommend you to specify the
1735 internal address:port in http_port. This way Squid will only be
1736 visible on the internal address.
1740 # Squid normally listens to port 3128
1741 http_port @DEFAULT_HTTP_PORT@
1751 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1753 The socket address where Squid will listen for client requests made
1754 over TLS or SSL connections. Commonly referred to as HTTPS.
1756 This is most useful for situations where you are running squid in
1757 accelerator mode and you want to do the SSL work at the accelerator level.
1759 You may specify multiple socket addresses on multiple lines,
1760 each with their own SSL certificate and/or options.
1764 accel Accelerator / reverse proxy mode
1766 intercept Support for IP-Layer interception of
1767 outgoing requests without browser settings.
1768 NP: disables authentication and IPv6 on the port.
1770 tproxy Support Linux TPROXY for spoofing outgoing
1771 connections using the client IP address.
1772 NP: disables authentication and maybe IPv6 on the port.
1774 ssl-bump For each intercepted connection allowed by ssl_bump
1775 ACLs, establish a secure connection with the client and with
1776 the server, decrypt HTTPS messages as they pass through
1777 Squid, and treat them as unencrypted HTTP messages,
1778 becoming the man-in-the-middle.
1780 An "ssl_bump server-first" match is required to
1781 fully enable bumping of intercepted SSL connections.
1783 Requires tproxy or intercept.
1785 Omitting the mode flag causes default forward proxy mode to be used.
1788 See http_port for a list of generic options
1793 cert= Path to SSL certificate (PEM format).
1795 key= Path to SSL private key file (PEM format)
1796 if not specified, the certificate file is
1797 assumed to be a combined certificate and
1800 version= The version of SSL/TLS supported
1801 1 automatic (default)
1806 cipher= Colon separated list of supported ciphers.
1808 options= Various SSL engine options. The most important
1810 NO_SSLv2 Disallow the use of SSLv2
1811 NO_SSLv3 Disallow the use of SSLv3
1812 NO_TLSv1 Disallow the use of TLSv1
1813 SINGLE_DH_USE Always create a new key when using
1814 temporary/ephemeral DH key exchanges
1815 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1816 documentation for a complete list of options.
1818 clientca= File containing the list of CAs to use when
1819 requesting a client certificate.
1821 cafile= File containing additional CA certificates to
1822 use when verifying client certificates. If unset
1823 clientca will be used.
1825 capath= Directory containing additional CA certificates
1826 and CRL lists to use when verifying client certificates.
1828 crlfile= File of additional CRL lists to use when verifying
1829 the client certificate, in addition to CRLs stored in
1830 the capath. Implies VERIFY_CRL flag below.
1832 dhparams= File containing DH parameters for temporary/ephemeral
1835 sslflags= Various flags modifying the use of SSL:
1837 Don't request client certificates
1838 immediately, but wait until acl processing
1839 requires a certificate (not yet implemented).
1841 Don't use the default CA lists built in
1844 Don't allow for session reuse. Each connection
1845 will result in a new SSL session.
1847 Verify CRL lists when accepting client
1850 Verify CRL lists for all certificates in the
1851 client certificate chain.
1853 sslcontext= SSL session ID context identifier.
1855 generate-host-certificates[=<on|off>]
1856 Dynamically create SSL server certificates for the
1857 destination hosts of bumped SSL requests.When
1858 enabled, the cert and key options are used to sign
1859 generated certificates. Otherwise generated
1860 certificate will be selfsigned.
1861 If there is CA certificate life time of generated
1862 certificate equals lifetime of CA certificate. If
1863 generated certificate is selfsigned lifetime is three
1865 This option is enabled by default when SslBump is used.
1866 See the sslBump option above for more information.
1868 dynamic_cert_mem_cache_size=SIZE
1869 Approximate total RAM size spent on cached generated
1870 certificates. If set to zero, caching is disabled. The
1871 default value is 4MB.
1873 See http_port for a list of available options.
1876 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1879 LOC: Ip::Qos::TheConfig.tosToServer
1881 Allows you to select a TOS/Diffserv value for packets outgoing
1882 on the server side, based on an ACL.
1884 tcp_outgoing_tos ds-field [!]aclname ...
1886 Example where normal_service_net uses the TOS value 0x00
1887 and good_service_net uses 0x20
1889 acl normal_service_net src 10.0.0.0/24
1890 acl good_service_net src 10.0.1.0/24
1891 tcp_outgoing_tos 0x00 normal_service_net
1892 tcp_outgoing_tos 0x20 good_service_net
1894 TOS/DSCP values really only have local significance - so you should
1895 know what you're specifying. For more information, see RFC2474,
1896 RFC2475, and RFC3260.
1898 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1899 "default" to use whatever default your host has. Note that in
1900 practice often only multiples of 4 is usable as the two rightmost bits
1901 have been redefined for use by ECN (RFC 3168 section 23.1).
1903 Processing proceeds in the order specified, and stops at first fully
1906 Only fast ACLs are supported.
1909 NAME: clientside_tos
1912 LOC: Ip::Qos::TheConfig.tosToClient
1914 Allows you to select a TOS/Diffserv value for packets being transmitted
1915 on the client-side, based on an ACL.
1917 clientside_tos ds-field [!]aclname ...
1919 Example where normal_service_net uses the TOS value 0x00
1920 and good_service_net uses 0x20
1922 acl normal_service_net src 10.0.0.0/24
1923 acl good_service_net src 10.0.1.0/24
1924 clientside_tos 0x00 normal_service_net
1925 clientside_tos 0x20 good_service_net
1927 Note: This feature is incompatible with qos_flows. Any TOS values set here
1928 will be overwritten by TOS values in qos_flows.
1931 NAME: tcp_outgoing_mark
1933 IFDEF: SO_MARK&&USE_LIBCAP
1935 LOC: Ip::Qos::TheConfig.nfmarkToServer
1937 Allows you to apply a Netfilter mark value to outgoing packets
1938 on the server side, based on an ACL.
1940 tcp_outgoing_mark mark-value [!]aclname ...
1942 Example where normal_service_net uses the mark value 0x00
1943 and good_service_net uses 0x20
1945 acl normal_service_net src 10.0.0.0/24
1946 acl good_service_net src 10.0.1.0/24
1947 tcp_outgoing_mark 0x00 normal_service_net
1948 tcp_outgoing_mark 0x20 good_service_net
1950 Only fast ACLs are supported.
1953 NAME: clientside_mark
1955 IFDEF: SO_MARK&&USE_LIBCAP
1957 LOC: Ip::Qos::TheConfig.nfmarkToClient
1959 Allows you to apply a Netfilter mark value to packets being transmitted
1960 on the client-side, based on an ACL.
1962 clientside_mark mark-value [!]aclname ...
1964 Example where normal_service_net uses the mark value 0x00
1965 and good_service_net uses 0x20
1967 acl normal_service_net src 10.0.0.0/24
1968 acl good_service_net src 10.0.1.0/24
1969 clientside_mark 0x00 normal_service_net
1970 clientside_mark 0x20 good_service_net
1972 Note: This feature is incompatible with qos_flows. Any mark values set here
1973 will be overwritten by mark values in qos_flows.
1980 LOC: Ip::Qos::TheConfig
1982 Allows you to select a TOS/DSCP value to mark outgoing
1983 connections to the client, based on where the reply was sourced.
1984 For platforms using netfilter, allows you to set a netfilter mark
1985 value instead of, or in addition to, a TOS value.
1987 By default this functionality is disabled. To enable it with the default
1988 settings simply use "qos_flows mark" or "qos_flows tos". Default
1989 settings will result in the netfilter mark or TOS value being copied
1990 from the upstream connection to the client. Note that it is the connection
1991 CONNMARK value not the packet MARK value that is copied.
1993 It is not currently possible to copy the mark or TOS value from the
1994 client to the upstream connection request.
1996 TOS values really only have local significance - so you should
1997 know what you're specifying. For more information, see RFC2474,
1998 RFC2475, and RFC3260.
2000 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
2001 in practice often only multiples of 4 is usable as the two rightmost bits
2002 have been redefined for use by ECN (RFC 3168 section 23.1).
2004 Mark values can be any unsigned 32-bit integer value.
2006 This setting is configured by setting the following values:
2008 tos|mark Whether to set TOS or netfilter mark values
2010 local-hit=0xFF Value to mark local cache hits.
2012 sibling-hit=0xFF Value to mark hits from sibling peers.
2014 parent-hit=0xFF Value to mark hits from parent peers.
2016 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2017 over the preserve-miss feature (see below), unless
2018 mask is specified, in which case only the bits
2019 specified in the mask are written.
2021 The TOS variant of the following features are only possible on Linux
2022 and require your kernel to be patched with the TOS preserving ZPH
2023 patch, available from http://zph.bratcheda.org
2024 No patch is needed to preserve the netfilter mark, which will work
2025 with all variants of netfilter.
2027 disable-preserve-miss
2028 This option disables the preservation of the TOS or netfilter
2029 mark. By default, the existing TOS or netfilter mark value of
2030 the response coming from the remote server will be retained
2031 and masked with miss-mark.
2032 NOTE: in the case of a netfilter mark, the mark must be set on
2033 the connection (using the CONNMARK target) not on the packet
2037 Allows you to mask certain bits in the TOS or mark value
2038 received from the remote server, before copying the value to
2039 the TOS sent towards clients.
2040 Default for tos: 0xFF (TOS from server is not changed).
2041 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2043 All of these features require the --enable-zph-qos compilation flag
2044 (enabled by default). Netfilter marking also requires the
2045 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2046 libcap 2.09+ (--with-libcap).
2050 NAME: tcp_outgoing_address
2053 DEFAULT_DOC: Address selection is performed by the operating system.
2054 LOC: Config.accessList.outgoing_address
2056 Allows you to map requests to different outgoing IP addresses
2057 based on the username or source address of the user making
2060 tcp_outgoing_address ipaddr [[!]aclname] ...
2063 Forwarding clients with dedicated IPs for certain subnets.
2065 acl normal_service_net src 10.0.0.0/24
2066 acl good_service_net src 10.0.2.0/24
2068 tcp_outgoing_address 2001:db8::c001 good_service_net
2069 tcp_outgoing_address 10.1.0.2 good_service_net
2071 tcp_outgoing_address 2001:db8::beef normal_service_net
2072 tcp_outgoing_address 10.1.0.1 normal_service_net
2074 tcp_outgoing_address 2001:db8::1
2075 tcp_outgoing_address 10.1.0.3
2077 Processing proceeds in the order specified, and stops at first fully
2080 Squid will add an implicit IP version test to each line.
2081 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2082 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2085 NOTE: The use of this directive using client dependent ACLs is
2086 incompatible with the use of server side persistent connections. To
2087 ensure correct results it is best to set server_persistent_connections
2088 to off when using this directive in such configurations.
2090 NOTE: The use of this directive to set a local IP on outgoing TCP links
2091 is incompatible with using TPROXY to set client IP out outbound TCP links.
2092 When needing to contact peers use the no-tproxy cache_peer option and the
2093 client_dst_passthru directive re-enable normal forwarding such as this.
2097 NAME: host_verify_strict
2100 LOC: Config.onoff.hostStrictVerify
2102 Regardless of this option setting, when dealing with intercepted
2103 traffic, Squid always verifies that the destination IP address matches
2104 the Host header domain or IP (called 'authority form URL').
2106 This enforcement is performed to satisfy a MUST-level requirement in
2107 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2108 authority of the origin server or gateway given by the original URL".
2111 Squid always responds with an HTTP 409 (Conflict) error
2112 page and logs a security warning if there is no match.
2114 Squid verifies that the destination IP address matches
2115 the Host header for forward-proxy and reverse-proxy traffic
2116 as well. For those traffic types, Squid also enables the
2117 following checks, comparing the corresponding Host header
2118 and Request-URI components:
2120 * The host names (domain or IP) must be identical,
2121 but valueless or missing Host header disables all checks.
2122 For the two host names to match, both must be either IP
2125 * Port numbers must be identical, but if a port is missing
2126 the scheme-default port is assumed.
2129 When set to OFF (the default):
2130 Squid allows suspicious requests to continue but logs a
2131 security warning and blocks caching of the response.
2133 * Forward-proxy traffic is not checked at all.
2135 * Reverse-proxy traffic is not checked at all.
2137 * Intercepted traffic which passes verification is handled
2138 according to client_dst_passthru.
2140 * Intercepted requests which fail verification are sent
2141 to the client original destination instead of DIRECT.
2142 This overrides 'client_dst_passthru off'.
2144 For now suspicious intercepted CONNECT requests are always
2145 responded to with an HTTP 409 (Conflict) error page.
2150 As described in CVE-2009-0801 when the Host: header alone is used
2151 to determine the destination of a request it becomes trivial for
2152 malicious scripts on remote websites to bypass browser same-origin
2153 security policy and sandboxing protections.
2155 The cause of this is that such applets are allowed to perform their
2156 own HTTP stack, in which case the same-origin policy of the browser
2157 sandbox only verifies that the applet tries to contact the same IP
2158 as from where it was loaded at the IP level. The Host: header may
2159 be different from the connected IP and approved origin.
2163 NAME: client_dst_passthru
2166 LOC: Config.onoff.client_dst_passthru
2168 With NAT or TPROXY intercepted traffic Squid may pass the request
2169 directly to the original client destination IP or seek a faster
2170 source using the HTTP Host header.
2172 Using Host to locate alternative servers can provide faster
2173 connectivity with a range of failure recovery options.
2174 But can also lead to connectivity trouble when the client and
2175 server are attempting stateful interactions unaware of the proxy.
2177 This option (on by default) prevents alternative DNS entries being
2178 located to send intercepted traffic DIRECT to an origin server.
2179 The clients original destination IP and port will be used instead.
2181 Regardless of this option setting, when dealing with intercepted
2182 traffic Squid will verify the Host: header and any traffic which
2183 fails Host verification will be treated as if this option were ON.
2185 see host_verify_strict for details on the verification process.
2190 -----------------------------------------------------------------------------
2193 NAME: ssl_unclean_shutdown
2197 LOC: Config.SSL.unclean_shutdown
2199 Some browsers (especially MSIE) bugs out on SSL shutdown
2206 LOC: Config.SSL.ssl_engine
2209 The OpenSSL engine to use. You will need to set this if you
2210 would like to use hardware SSL acceleration for example.
2213 NAME: sslproxy_client_certificate
2216 LOC: Config.ssl_client.cert
2219 Client SSL Certificate to use when proxying https:// URLs
2222 NAME: sslproxy_client_key
2225 LOC: Config.ssl_client.key
2228 Client SSL Key to use when proxying https:// URLs
2231 NAME: sslproxy_version
2234 DEFAULT_DOC: automatic SSL/TLS version negotiation
2235 LOC: Config.ssl_client.version
2238 SSL version level to use when proxying https:// URLs
2240 The versions of SSL/TLS supported:
2242 1 automatic (default)
2250 NAME: sslproxy_options
2253 LOC: Config.ssl_client.options
2256 SSL implementation options to use when proxying https:// URLs
2258 The most important being:
2260 NO_SSLv2 Disallow the use of SSLv2
2261 NO_SSLv3 Disallow the use of SSLv3
2262 NO_TLSv1 Disallow the use of TLSv1.0
2263 NO_TLSv1_1 Disallow the use of TLSv1.1
2264 NO_TLSv1_2 Disallow the use of TLSv1.2
2266 Always create a new key when using temporary/ephemeral
2269 Disable use of RFC5077 session tickets. Some servers
2270 may have problems understanding the TLS extension due
2271 to ambiguous specification in RFC4507.
2272 ALL Enable various bug workarounds suggested as "harmless"
2273 by OpenSSL. Be warned that this may reduce SSL/TLS
2274 strength to some attacks.
2276 See the OpenSSL SSL_CTX_set_options documentation for a
2277 complete list of possible options.
2280 NAME: sslproxy_cipher
2283 LOC: Config.ssl_client.cipher
2286 SSL cipher list to use when proxying https:// URLs
2288 Colon separated list of supported ciphers.
2291 NAME: sslproxy_cafile
2294 LOC: Config.ssl_client.cafile
2297 file containing CA certificates to use when verifying server
2298 certificates while proxying https:// URLs
2301 NAME: sslproxy_capath
2304 LOC: Config.ssl_client.capath
2307 directory containing CA certificates to use when verifying
2308 server certificates while proxying https:// URLs
2311 NAME: sslproxy_session_ttl
2314 LOC: Config.SSL.session_ttl
2317 Sets the timeout value for SSL sessions
2320 NAME: sslproxy_session_cache_size
2323 LOC: Config.SSL.sessionCacheSize
2326 Sets the cache size to use for ssl session
2331 TYPE: sslproxy_ssl_bump
2332 LOC: Config.accessList.ssl_bump
2333 DEFAULT_DOC: Does not bump unless rules are present in squid.conf
2336 This option is consulted when a CONNECT request is received on
2337 an http_port (or a new connection is intercepted at an
2338 https_port), provided that port was configured with an ssl-bump
2339 flag. The subsequent data on the connection is either treated as
2340 HTTPS and decrypted OR tunneled at TCP level without decryption,
2341 depending on the first bumping "mode" which ACLs match.
2343 ssl_bump <mode> [!]acl ...
2345 The following bumping modes are supported:
2348 Allow bumping of the connection. Establish a secure connection
2349 with the client first, then connect to the server. This old mode
2350 does not allow Squid to mimic server SSL certificate and does
2351 not work with intercepted SSL connections.
2354 Allow bumping of the connection. Establish a secure connection
2355 with the server first, then establish a secure connection with
2356 the client, using a mimicked server certificate. Works with both
2357 CONNECT requests and intercepted SSL connections.
2360 Become a TCP tunnel without decoding the connection.
2361 Works with both CONNECT requests and intercepted SSL
2362 connections. This is the default behavior when no
2363 ssl_bump option is given or no ssl_bump ACLs match.
2365 By default, no connections are bumped.
2367 The first matching ssl_bump option wins. If no ACLs match, the
2368 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2369 does not have an implicit "negate the last given option" rule. You
2370 must make that rule explicit if you convert old ssl_bump allow/deny
2371 rules that rely on such an implicit rule.
2373 This clause supports both fast and slow acl types.
2374 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2376 See also: http_port ssl-bump, https_port ssl-bump
2379 # Example: Bump all requests except those originating from
2380 # localhost or those going to example.com.
2382 acl broken_sites dstdomain .example.com
2383 ssl_bump none localhost
2384 ssl_bump none broken_sites
2385 ssl_bump server-first all
2388 NAME: sslproxy_flags
2391 LOC: Config.ssl_client.flags
2394 Various flags modifying the use of SSL while proxying https:// URLs:
2395 DONT_VERIFY_PEER Accept certificates that fail verification.
2396 For refined control, see sslproxy_cert_error.
2397 NO_DEFAULT_CA Don't use the default CA list built in
2401 NAME: sslproxy_cert_error
2404 DEFAULT_DOC: Server certificate errors terminate the transaction.
2405 LOC: Config.ssl_client.cert_error
2408 Use this ACL to bypass server certificate validation errors.
2410 For example, the following lines will bypass all validation errors
2411 when talking to servers for example.com. All other
2412 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2414 acl BrokenButTrustedServers dstdomain example.com
2415 sslproxy_cert_error allow BrokenButTrustedServers
2416 sslproxy_cert_error deny all
2418 This clause only supports fast acl types.
2419 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2420 Using slow acl types may result in server crashes
2422 Without this option, all server certificate validation errors
2423 terminate the transaction to protect Squid and the client.
2425 SQUID_X509_V_ERR_INFINITE_VALIDATION error cannot be bypassed
2426 but should not happen unless your OpenSSL library is buggy.
2429 Bypassing validation errors is dangerous because an
2430 error usually implies that the server cannot be trusted
2431 and the connection may be insecure.
2433 See also: sslproxy_flags and DONT_VERIFY_PEER.
2436 NAME: sslproxy_cert_sign
2439 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2440 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2441 POSTSCRIPTUM: signTrusted all
2442 TYPE: sslproxy_cert_sign
2443 LOC: Config.ssl_client.cert_sign
2446 sslproxy_cert_sign <signing algorithm> acl ...
2448 The following certificate signing algorithms are supported:
2451 Sign using the configured CA certificate which is usually
2452 placed in and trusted by end-user browsers. This is the
2453 default for trusted origin server certificates.
2456 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2457 This is the default for untrusted origin server certificates
2458 that are not self-signed (see ssl::certUntrusted).
2461 Sign using a self-signed certificate with the right CN to
2462 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2463 browser. This is the default for self-signed origin server
2464 certificates (see ssl::certSelfSigned).
2466 This clause only supports fast acl types.
2468 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2469 signing algorithm to generate the certificate and ignores all
2470 subsequent sslproxy_cert_sign options (the first match wins). If no
2471 acl(s) match, the default signing algorithm is determined by errors
2472 detected when obtaining and validating the origin server certificate.
2474 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2475 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2476 CONNECT request that carries a domain name. In all other cases (CONNECT
2477 to an IP address or an intercepted SSL connection), Squid cannot detect
2478 the domain mismatch at certificate generation time when
2479 bump-server-first is used.
2482 NAME: sslproxy_cert_adapt
2485 TYPE: sslproxy_cert_adapt
2486 LOC: Config.ssl_client.cert_adapt
2489 sslproxy_cert_adapt <adaptation algorithm> acl ...
2491 The following certificate adaptation algorithms are supported:
2494 Sets the "Not After" property to the "Not After" property of
2495 the CA certificate used to sign generated certificates.
2498 Sets the "Not Before" property to the "Not Before" property of
2499 the CA certificate used to sign generated certificates.
2501 setCommonName or setCommonName{CN}
2502 Sets Subject.CN property to the host name specified as a
2503 CN parameter or, if no explicit CN parameter was specified,
2504 extracted from the CONNECT request. It is a misconfiguration
2505 to use setCommonName without an explicit parameter for
2506 intercepted or tproxied SSL connections.
2508 This clause only supports fast acl types.
2510 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2511 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2512 corresponding adaptation algorithm to generate the certificate and
2513 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2514 group (i.e., the first match wins within each algorithm group). If no
2515 acl(s) match, the default mimicking action takes place.
2517 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2518 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2519 CONNECT request that carries a domain name. In all other cases (CONNECT
2520 to an IP address or an intercepted SSL connection), Squid cannot detect
2521 the domain mismatch at certificate generation time when
2522 bump-server-first is used.
2525 NAME: sslpassword_program
2528 LOC: Config.Program.ssl_password
2531 Specify a program used for entering SSL key passphrases
2532 when using encrypted SSL certificate keys. If not specified
2533 keys must either be unencrypted, or Squid started with the -N
2534 option to allow it to query interactively for the passphrase.
2536 The key file name is given as argument to the program allowing
2537 selection of the right password if you have multiple encrypted
2542 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2543 -----------------------------------------------------------------------------
2546 NAME: sslcrtd_program
2549 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2550 LOC: Ssl::TheConfig.ssl_crtd
2552 Specify the location and options of the executable for ssl_crtd process.
2553 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2554 For more information use:
2555 @DEFAULT_SSL_CRTD@ -h
2558 NAME: sslcrtd_children
2559 TYPE: HelperChildConfig
2561 DEFAULT: 32 startup=5 idle=1
2562 LOC: Ssl::TheConfig.ssl_crtdChildren
2564 The maximum number of processes spawn to service ssl server.
2565 The maximum this may be safely set to is 32.
2567 The startup= and idle= options allow some measure of skew in your
2572 Sets the minimum number of processes to spawn when Squid
2573 starts or reconfigures. When set to zero the first request will
2574 cause spawning of the first child process to handle it.
2576 Starting too few children temporary slows Squid under load while it
2577 tries to spawn enough additional processes to cope with traffic.
2581 Sets a minimum of how many processes Squid is to try and keep available
2582 at all times. When traffic begins to rise above what the existing
2583 processes can handle this many more will be spawned up to the maximum
2584 configured. A minimum setting of 1 is required.
2586 You must have at least one ssl_crtd process.
2589 NAME: sslcrtvalidator_program
2593 LOC: Ssl::TheConfig.ssl_crt_validator
2595 Specify the location and options of the executable for ssl_crt_validator
2598 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2601 ttl=n TTL in seconds for cached results. The default is 60 secs
2602 cache=n limit the result cache size. The default value is 2048
2605 NAME: sslcrtvalidator_children
2606 TYPE: HelperChildConfig
2608 DEFAULT: 32 startup=5 idle=1 concurrency=1
2609 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2611 The maximum number of processes spawn to service SSL server.
2612 The maximum this may be safely set to is 32.
2614 The startup= and idle= options allow some measure of skew in your
2619 Sets the minimum number of processes to spawn when Squid
2620 starts or reconfigures. When set to zero the first request will
2621 cause spawning of the first child process to handle it.
2623 Starting too few children temporary slows Squid under load while it
2624 tries to spawn enough additional processes to cope with traffic.
2628 Sets a minimum of how many processes Squid is to try and keep available
2629 at all times. When traffic begins to rise above what the existing
2630 processes can handle this many more will be spawned up to the maximum
2631 configured. A minimum setting of 1 is required.
2635 The number of requests each certificate validator helper can handle in
2636 parallel. A value of 0 indicates the certficate validator does not
2637 support concurrency. Defaults to 1.
2639 When this directive is set to a value >= 1 then the protocol
2640 used to communicate with the helper is modified to include
2641 a request ID in front of the request/response. The request
2642 ID from the request must be echoed back with the response
2645 You must have at least one ssl_crt_validator process.
2649 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2650 -----------------------------------------------------------------------------
2658 To specify other caches in a hierarchy, use the format:
2660 cache_peer hostname type http-port icp-port [options]
2665 # hostname type port port options
2666 # -------------------- -------- ----- ----- -----------
2667 cache_peer parent.foo.net parent 3128 3130 default
2668 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2669 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2670 cache_peer example.com parent 80 0 default
2671 cache_peer cdn.example.com sibling 3128 0
2673 type: either 'parent', 'sibling', or 'multicast'.
2675 proxy-port: The port number where the peer accept HTTP requests.
2676 For other Squid proxies this is usually 3128
2677 For web servers this is usually 80
2679 icp-port: Used for querying neighbor caches about objects.
2680 Set to 0 if the peer does not support ICP or HTCP.
2681 See ICP and HTCP options below for additional details.
2684 ==== ICP OPTIONS ====
2686 You MUST also set icp_port and icp_access explicitly when using these options.
2687 The defaults will prevent peer traffic using ICP.
2690 no-query Disable ICP queries to this neighbor.
2693 Indicates the named peer is a member of a multicast group.
2694 ICP queries will not be sent directly to the peer, but ICP
2695 replies will be accepted from it.
2697 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2698 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2701 To only send ICP queries to this neighbor infrequently.
2702 This is used to keep the neighbor round trip time updated
2703 and is usually used in conjunction with weighted-round-robin.
2706 ==== HTCP OPTIONS ====
2708 You MUST also set htcp_port and htcp_access explicitly when using these options.
2709 The defaults will prevent peer traffic using HTCP.
2712 htcp Send HTCP, instead of ICP, queries to the neighbor.
2713 You probably also want to set the "icp-port" to 4827
2714 instead of 3130. This directive accepts a comma separated
2715 list of options described below.
2717 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2719 htcp=no-clr Send HTCP to the neighbor but without
2720 sending any CLR requests. This cannot be used with
2723 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2724 This cannot be used with no-clr.
2727 Send HTCP to the neighbor including CLRs but only when
2728 they do not result from PURGE requests.
2731 Forward any HTCP CLR requests this proxy receives to the peer.
2734 ==== PEER SELECTION METHODS ====
2736 The default peer selection method is ICP, with the first responding peer
2737 being used as source. These options can be used for better load balancing.
2740 default This is a parent cache which can be used as a "last-resort"
2741 if a peer cannot be located by any of the peer-selection methods.
2742 If specified more than once, only the first is used.
2744 round-robin Load-Balance parents which should be used in a round-robin
2745 fashion in the absence of any ICP queries.
2746 weight=N can be used to add bias.
2748 weighted-round-robin
2749 Load-Balance parents which should be used in a round-robin
2750 fashion with the frequency of each parent being based on the
2751 round trip time. Closer parents are used more often.
2752 Usually used for background-ping parents.
2753 weight=N can be used to add bias.
2755 carp Load-Balance parents which should be used as a CARP array.
2756 The requests will be distributed among the parents based on the
2757 CARP load balancing hash function based on their weight.
2759 userhash Load-balance parents based on the client proxy_auth or ident username.
2761 sourcehash Load-balance parents based on the client source IP.
2764 To be used only for cache peers of type "multicast".
2765 ALL members of this multicast group have "sibling"
2766 relationship with it, not "parent". This is to a multicast
2767 group when the requested object would be fetched only from
2768 a "parent" cache, anyway. It's useful, e.g., when
2769 configuring a pool of redundant Squid proxies, being
2770 members of the same multicast group.
2773 ==== PEER SELECTION OPTIONS ====
2775 weight=N use to affect the selection of a peer during any weighted
2776 peer-selection mechanisms.
2777 The weight must be an integer; default is 1,
2778 larger weights are favored more.
2779 This option does not affect parent selection if a peering
2780 protocol is not in use.
2782 basetime=N Specify a base amount to be subtracted from round trip
2784 It is subtracted before division by weight in calculating
2785 which parent to fectch from. If the rtt is less than the
2786 base time the rtt is set to a minimal value.
2788 ttl=N Specify a TTL to use when sending multicast ICP queries
2790 Only useful when sending to a multicast group.
2791 Because we don't accept ICP replies from random
2792 hosts, you must configure other group members as
2793 peers with the 'multicast-responder' option.
2795 no-delay To prevent access to this neighbor from influencing the
2798 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2799 enabled) for this host from the specified URL rather
2800 than the Squid default location.
2803 ==== CARP OPTIONS ====
2805 carp-key=key-specification
2806 use a different key than the full URL to hash against the peer.
2807 the key-specification is a comma-separated list of the keywords
2808 scheme, host, port, path, params
2809 Order is not important.
2811 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2813 originserver Causes this parent to be contacted as an origin server.
2814 Meant to be used in accelerator setups when the peer
2818 Set the Host header of requests forwarded to this peer.
2819 Useful in accelerator setups where the server (peer)
2820 expects a certain domain name but clients may request
2821 others. ie example.com or www.example.com
2823 no-digest Disable request of cache digests.
2826 Disables requesting ICMP RTT database (NetDB).
2829 ==== AUTHENTICATION OPTIONS ====
2832 If this is a personal/workgroup proxy and your parent
2833 requires proxy authentication.
2835 Note: The string can include URL escapes (i.e. %20 for
2836 spaces). This also means % must be written as %%.
2839 Send login details received from client to this peer.
2840 Both Proxy- and WWW-Authorization headers are passed
2841 without alteration to the peer.
2842 Authentication is not required by Squid for this to work.
2844 Note: This will pass any form of authentication but
2845 only Basic auth will work through a proxy unless the
2846 connection-auth options are also used.
2848 login=PASS Send login details received from client to this peer.
2849 Authentication is not required by this option.
2851 If there are no client-provided authentication headers
2852 to pass on, but username and password are available
2853 from an external ACL user= and password= result tags
2854 they may be sent instead.
2856 Note: To combine this with proxy_auth both proxies must
2857 share the same user database as HTTP only allows for
2858 a single login (one for proxy, one for origin server).
2859 Also be warned this will expose your users proxy
2860 password to the peer. USE WITH CAUTION
2863 Send the username to the upstream cache, but with a
2864 fixed password. This is meant to be used when the peer
2865 is in another administrative domain, but it is still
2866 needed to identify each user.
2867 The star can optionally be followed by some extra
2868 information which is added to the username. This can
2869 be used to identify this proxy to the peer, similar to
2870 the login=username:password option above.
2873 If this is a personal/workgroup proxy and your parent
2874 requires a secure proxy authentication.
2875 The first principal from the default keytab or defined by
2876 the environment variable KRB5_KTNAME will be used.
2878 WARNING: The connection may transmit requests from multiple
2879 clients. Negotiate often assumes end-to-end authentication
2880 and a single-client. Which is not strictly true here.
2882 login=NEGOTIATE:principal_name
2883 If this is a personal/workgroup proxy and your parent
2884 requires a secure proxy authentication.
2885 The principal principal_name from the default keytab or
2886 defined by the environment variable KRB5_KTNAME will be
2889 WARNING: The connection may transmit requests from multiple
2890 clients. Negotiate often assumes end-to-end authentication
2891 and a single-client. Which is not strictly true here.
2893 connection-auth=on|off
2894 Tell Squid that this peer does or not support Microsoft
2895 connection oriented authentication, and any such
2896 challenges received from there should be ignored.
2897 Default is auto to automatically determine the status
2901 ==== SSL / HTTPS / TLS OPTIONS ====
2903 ssl Encrypt connections to this peer with SSL/TLS.
2905 sslcert=/path/to/ssl/certificate
2906 A client SSL certificate to use when connecting to
2909 sslkey=/path/to/ssl/key
2910 The private SSL key corresponding to sslcert above.
2911 If 'sslkey' is not specified 'sslcert' is assumed to
2912 reference a combined file containing both the
2913 certificate and the key.
2915 sslversion=1|2|3|4|5|6
2916 The SSL version to use when connecting to this peer
2917 1 = automatic (default)
2924 sslcipher=... The list of valid SSL ciphers to use when connecting
2927 ssloptions=... Specify various SSL implementation options:
2929 NO_SSLv2 Disallow the use of SSLv2
2930 NO_SSLv3 Disallow the use of SSLv3
2931 NO_TLSv1 Disallow the use of TLSv1.0
2932 NO_TLSv1_1 Disallow the use of TLSv1.1
2933 NO_TLSv1_2 Disallow the use of TLSv1.2
2935 Always create a new key when using
2936 temporary/ephemeral DH key exchanges
2937 ALL Enable various bug workarounds
2938 suggested as "harmless" by OpenSSL
2939 Be warned that this reduces SSL/TLS
2940 strength to some attacks.
2942 See the OpenSSL SSL_CTX_set_options documentation for a
2945 sslcafile=... A file containing additional CA certificates to use
2946 when verifying the peer certificate.
2948 sslcapath=... A directory containing additional CA certificates to
2949 use when verifying the peer certificate.
2951 sslcrlfile=... A certificate revocation list file to use when
2952 verifying the peer certificate.
2954 sslflags=... Specify various flags modifying the SSL implementation:
2957 Accept certificates even if they fail to
2960 Don't use the default CA list built in
2963 Don't verify the peer certificate
2964 matches the server name
2966 ssldomain= The peer name as advertised in it's certificate.
2967 Used for verifying the correctness of the received peer
2968 certificate. If not specified the peer hostname will be
2972 Enable the "Front-End-Https: On" header needed when
2973 using Squid as a SSL frontend in front of Microsoft OWA.
2974 See MS KB document Q307347 for details on this header.
2975 If set to auto the header will only be added if the
2976 request is forwarded as a https:// URL.
2979 ==== GENERAL OPTIONS ====
2982 A peer-specific connect timeout.
2983 Also see the peer_connect_timeout directive.
2985 connect-fail-limit=N
2986 How many times connecting to a peer must fail before
2987 it is marked as down. Standby connection failures
2988 count towards this limit. Default is 10.
2990 allow-miss Disable Squid's use of only-if-cached when forwarding
2991 requests to siblings. This is primarily useful when
2992 icp_hit_stale is used by the sibling. To extensive use
2993 of this option may result in forwarding loops, and you
2994 should avoid having two-way peerings with this option.
2995 For example to deny peer usage on requests from peer
2996 by denying cache_peer_access if the source is a peer.
2998 max-conn=N Limit the number of concurrent connections the Squid
2999 may open to this peer, including already opened idle
3000 and standby connections. There is no peer-specific
3001 connection limit by default.
3003 A peer exceeding the limit is not used for new
3004 requests unless a standby connection is available.
3006 max-conn currently works poorly with idle persistent
3007 connections: When a peer reaches its max-conn limit,
3008 and there are idle persistent connections to the peer,
3009 the peer may not be selected because the limiting code
3010 does not know whether Squid can reuse those idle
3013 standby=N Maintain a pool of N "hot standby" connections to an
3014 UP peer, available for requests when no idle
3015 persistent connection is available (or safe) to use.
3016 By default and with zero N, no such pool is maintained.
3017 N must not exceed the max-conn limit (if any).
3019 At start or after reconfiguration, Squid opens new TCP
3020 standby connections until there are N connections
3021 available and then replenishes the standby pool as
3022 opened connections are used up for requests. A used
3023 connection never goes back to the standby pool, but
3024 may go to the regular idle persistent connection pool
3025 shared by all peers and origin servers.
3027 Squid never opens multiple new standby connections
3028 concurrently. This one-at-a-time approach minimizes
3029 flooding-like effect on peers. Furthermore, just a few
3030 standby connections should be sufficient in most cases
3031 to supply most new requests with a ready-to-use
3034 Standby connections obey server_idle_pconn_timeout.
3035 For the feature to work as intended, the peer must be
3036 configured to accept and keep them open longer than
3037 the idle timeout at the connecting Squid, to minimize
3038 race conditions typical to idle used persistent
3039 connections. Default request_timeout and
3040 server_idle_pconn_timeout values ensure such a
3043 name=xxx Unique name for the peer.
3044 Required if you have multiple peers on the same host
3045 but different ports.
3046 This name can be used in cache_peer_access and similar
3047 directives to dentify the peer.
3048 Can be used by outgoing access controls through the
3051 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3052 requests to this peer. Use normal address selection instead.
3053 This overrides the spoof_client_ip ACL.
3055 proxy-only objects fetched from the peer will not be stored locally.
3059 NAME: cache_peer_domain cache_host_domain
3064 Use to limit the domains for which a neighbor cache will be
3068 cache_peer_domain cache-host domain [domain ...]
3069 cache_peer_domain cache-host !domain
3071 For example, specifying
3073 cache_peer_domain parent.foo.net .edu
3075 has the effect such that UDP query packets are sent to
3076 'bigserver' only when the requested object exists on a
3077 server in the .edu domain. Prefixing the domainname
3078 with '!' means the cache will be queried for objects
3081 NOTE: * Any number of domains may be given for a cache-host,
3082 either on the same or separate lines.
3083 * When multiple domains are given for a particular
3084 cache-host, the first matched domain is applied.
3085 * Cache hosts with no domain restrictions are queried
3087 * There are no defaults.
3088 * There is also a 'cache_peer_access' tag in the ACL
3092 NAME: cache_peer_access
3097 Similar to 'cache_peer_domain' but provides more flexibility by
3101 cache_peer_access cache-host allow|deny [!]aclname ...
3103 The syntax is identical to 'http_access' and the other lists of
3104 ACL elements. See the comments for 'http_access' below, or
3105 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3108 NAME: neighbor_type_domain
3109 TYPE: hostdomaintype
3111 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3114 Modify the cache_peer neighbor type when passing requests
3115 about specific domains to the peer.
3118 neighbor_type_domain neighbor parent|sibling domain domain ...
3121 cache_peer foo.example.com parent 3128 3130
3122 neighbor_type_domain foo.example.com sibling .au .de
3124 The above configuration treats all requests to foo.example.com as a
3125 parent proxy unless the request is for a .au or .de ccTLD domain name.
3128 NAME: dead_peer_timeout
3132 LOC: Config.Timeout.deadPeer
3134 This controls how long Squid waits to declare a peer cache
3135 as "dead." If there are no ICP replies received in this
3136 amount of time, Squid will declare the peer dead and not
3137 expect to receive any further ICP replies. However, it
3138 continues to send ICP queries, and will mark the peer as
3139 alive upon receipt of the first subsequent ICP reply.
3141 This timeout also affects when Squid expects to receive ICP
3142 replies from peers. If more than 'dead_peer' seconds have
3143 passed since the last ICP reply was received, Squid will not
3144 expect to receive an ICP reply on the next query. Thus, if
3145 your time between requests is greater than this timeout, you
3146 will see a lot of requests sent DIRECT to origin servers
3147 instead of to your parents.
3150 NAME: forward_max_tries
3153 LOC: Config.forward_max_tries
3155 Controls how many different forward paths Squid will try
3156 before giving up. See also forward_timeout.
3158 NOTE: connect_retries (default: none) can make each of these
3159 possible forwarding paths be tried multiple times.
3163 MEMORY CACHE OPTIONS
3164 -----------------------------------------------------------------------------
3171 LOC: Config.memMaxSize
3173 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3174 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3175 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3176 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3178 'cache_mem' specifies the ideal amount of memory to be used
3180 * In-Transit objects
3182 * Negative-Cached objects
3184 Data for these objects are stored in 4 KB blocks. This
3185 parameter specifies the ideal upper limit on the total size of
3186 4 KB blocks allocated. In-Transit objects take the highest
3189 In-transit objects have priority over the others. When
3190 additional space is needed for incoming data, negative-cached
3191 and hot objects will be released. In other words, the
3192 negative-cached and hot objects will fill up any unused space
3193 not needed for in-transit objects.
3195 If circumstances require, this limit will be exceeded.
3196 Specifically, if your incoming request rate requires more than
3197 'cache_mem' of memory to hold in-transit objects, Squid will
3198 exceed this limit to satisfy the new requests. When the load
3199 decreases, blocks will be freed until the high-water mark is
3200 reached. Thereafter, blocks will be used to store hot
3203 If shared memory caching is enabled, Squid does not use the shared
3204 cache space for in-transit objects, but they still consume as much
3205 local memory as they need. For more details about the shared memory
3206 cache, see memory_cache_shared.
3209 NAME: maximum_object_size_in_memory
3213 LOC: Config.Store.maxInMemObjSize
3215 Objects greater than this size will not be attempted to kept in
3216 the memory cache. This should be set high enough to keep objects
3217 accessed frequently in memory to improve performance whilst low
3218 enough to keep larger objects from hoarding cache_mem.
3221 NAME: memory_cache_shared
3224 LOC: Config.memShared
3226 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3228 Controls whether the memory cache is shared among SMP workers.
3230 The shared memory cache is meant to occupy cache_mem bytes and replace
3231 the non-shared memory cache, although some entities may still be
3232 cached locally by workers for now (e.g., internal and in-transit
3233 objects may be served from a local memory cache even if shared memory
3234 caching is enabled).
3236 By default, the memory cache is shared if and only if all of the
3237 following conditions are satisfied: Squid runs in SMP mode with
3238 multiple workers, cache_mem is positive, and Squid environment
3239 supports required IPC primitives (e.g., POSIX shared memory segments
3240 and GCC-style atomic operations).
3242 To avoid blocking locks, shared memory uses opportunistic algorithms
3243 that do not guarantee that every cachable entity that could have been
3244 shared among SMP workers will actually be shared.
3246 Currently, entities exceeding 32KB in size cannot be shared.
3249 NAME: memory_cache_mode
3253 DEFAULT_DOC: Keep the most recently fetched objects in memory
3255 Controls which objects to keep in the memory cache (cache_mem)
3257 always Keep most recently fetched objects in memory (default)
3259 disk Only disk cache hits are kept in memory, which means
3260 an object must first be cached on disk and then hit
3261 a second time before cached in memory.
3263 network Only objects fetched from network is kept in memory
3266 NAME: memory_replacement_policy
3268 LOC: Config.memPolicy
3271 The memory replacement policy parameter determines which
3272 objects are purged from memory when memory space is needed.
3274 See cache_replacement_policy for details on algorithms.
3279 -----------------------------------------------------------------------------
3282 NAME: cache_replacement_policy
3284 LOC: Config.replPolicy
3287 The cache replacement policy parameter determines which
3288 objects are evicted (replaced) when disk space is needed.
3290 lru : Squid's original list based LRU policy
3291 heap GDSF : Greedy-Dual Size Frequency
3292 heap LFUDA: Least Frequently Used with Dynamic Aging
3293 heap LRU : LRU policy implemented using a heap
3295 Applies to any cache_dir lines listed below this directive.
3297 The LRU policies keeps recently referenced objects.
3299 The heap GDSF policy optimizes object hit rate by keeping smaller
3300 popular objects in cache so it has a better chance of getting a
3301 hit. It achieves a lower byte hit rate than LFUDA though since
3302 it evicts larger (possibly popular) objects.
3304 The heap LFUDA policy keeps popular objects in cache regardless of
3305 their size and thus optimizes byte hit rate at the expense of
3306 hit rate since one large, popular object will prevent many
3307 smaller, slightly less popular objects from being cached.
3309 Both policies utilize a dynamic aging mechanism that prevents
3310 cache pollution that can otherwise occur with frequency-based
3311 replacement policies.
3313 NOTE: if using the LFUDA replacement policy you should increase
3314 the value of maximum_object_size above its default of 4 MB to
3315 to maximize the potential byte hit rate improvement of LFUDA.
3317 For more information about the GDSF and LFUDA cache replacement
3318 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3319 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3322 NAME: minimum_object_size
3326 DEFAULT_DOC: no limit
3327 LOC: Config.Store.minObjectSize
3329 Objects smaller than this size will NOT be saved on disk. The
3330 value is specified in bytes, and the default is 0 KB, which
3331 means all responses can be stored.
3334 NAME: maximum_object_size
3338 LOC: Config.Store.maxObjectSize
3340 Set the default value for max-size parameter on any cache_dir.
3341 The value is specified in bytes, and the default is 4 MB.
3343 If you wish to get a high BYTES hit ratio, you should probably
3344 increase this (one 32 MB object hit counts for 3200 10KB
3347 If you wish to increase hit ratio more than you want to
3348 save bandwidth you should leave this low.
3350 NOTE: if using the LFUDA replacement policy you should increase
3351 this value to maximize the byte hit rate improvement of LFUDA!
3352 See cache_replacement_policy for a discussion of this policy.
3358 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3359 LOC: Config.cacheSwap
3362 cache_dir Type Directory-Name Fs-specific-data [options]
3364 You can specify multiple cache_dir lines to spread the
3365 cache among different disk partitions.
3367 Type specifies the kind of storage system to use. Only "ufs"
3368 is built by default. To enable any of the other storage systems
3369 see the --enable-storeio configure option.
3371 'Directory' is a top-level directory where cache swap
3372 files will be stored. If you want to use an entire disk
3373 for caching, this can be the mount-point directory.
3374 The directory must exist and be writable by the Squid
3375 process. Squid will NOT create this directory for you.
3377 In SMP configurations, cache_dir must not precede the workers option
3378 and should use configuration macros or conditionals to give each
3379 worker interested in disk caching a dedicated cache directory.
3382 ==== The ufs store type ====
3384 "ufs" is the old well-known Squid storage format that has always
3388 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3390 'Mbytes' is the amount of disk space (MB) to use under this
3391 directory. The default is 100 MB. Change this to suit your
3392 configuration. Do NOT put the size of your disk drive here.
3393 Instead, if you want Squid to use the entire disk drive,
3394 subtract 20% and use that value.
3396 'L1' is the number of first-level subdirectories which
3397 will be created under the 'Directory'. The default is 16.
3399 'L2' is the number of second-level subdirectories which
3400 will be created under each first-level directory. The default
3404 ==== The aufs store type ====
3406 "aufs" uses the same storage format as "ufs", utilizing
3407 POSIX-threads to avoid blocking the main Squid process on
3408 disk-I/O. This was formerly known in Squid as async-io.
3411 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3413 see argument descriptions under ufs above
3416 ==== The diskd store type ====
3418 "diskd" uses the same storage format as "ufs", utilizing a
3419 separate process to avoid blocking the main Squid process on
3423 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3425 see argument descriptions under ufs above
3427 Q1 specifies the number of unacknowledged I/O requests when Squid
3428 stops opening new files. If this many messages are in the queues,
3429 Squid won't open new files. Default is 64
3431 Q2 specifies the number of unacknowledged messages when Squid
3432 starts blocking. If this many messages are in the queues,
3433 Squid blocks until it receives some replies. Default is 72
3435 When Q1 < Q2 (the default), the cache directory is optimized
3436 for lower response time at the expense of a decrease in hit
3437 ratio. If Q1 > Q2, the cache directory is optimized for
3438 higher hit ratio at the expense of an increase in response
3442 ==== The rock store type ====
3445 cache_dir rock Directory-Name Mbytes [options]
3447 The Rock Store type is a database-style storage. All cached
3448 entries are stored in a "database" file, using fixed-size slots.
3449 A single entry occupies one or more slots.
3451 If possible, Squid using Rock Store creates a dedicated kid
3452 process called "disker" to avoid blocking Squid worker(s) on disk
3453 I/O. One disker kid is created for each rock cache_dir. Diskers
3454 are created only when Squid, running in daemon mode, has support
3455 for the IpcIo disk I/O module.
3457 swap-timeout=msec: Squid will not start writing a miss to or
3458 reading a hit from disk if it estimates that the swap operation
3459 will take more than the specified number of milliseconds. By
3460 default and when set to zero, disables the disk I/O time limit
3461 enforcement. Ignored when using blocking I/O module because
3462 blocking synchronous I/O does not allow Squid to estimate the
3463 expected swap wait time.
3465 max-swap-rate=swaps/sec: Artificially limits disk access using
3466 the specified I/O rate limit. Swap out requests that
3467 would cause the average I/O rate to exceed the limit are
3468 delayed. Individual swap in requests (i.e., hits or reads) are
3469 not delayed, but they do contribute to measured swap rate and
3470 since they are placed in the same FIFO queue as swap out
3471 requests, they may wait longer if max-swap-rate is smaller.
3472 This is necessary on file systems that buffer "too
3473 many" writes and then start blocking Squid and other processes
3474 while committing those writes to disk. Usually used together
3475 with swap-timeout to avoid excessive delays and queue overflows
3476 when disk demand exceeds available disk "bandwidth". By default
3477 and when set to zero, disables the disk I/O rate limit
3478 enforcement. Currently supported by IpcIo module only.
3480 slot-size=bytes: The size of a database "record" used for
3481 storing cached responses. A cached response occupies at least
3482 one slot and all database I/O is done using individual slots so
3483 increasing this parameter leads to more disk space waste while
3484 decreasing it leads to more disk I/O overheads. Should be a
3485 multiple of your operating system I/O page size. Defaults to
3486 16KBytes. A housekeeping header is stored with each slot and
3487 smaller slot-sizes will be rejected. The header is smaller than
3491 ==== COMMON OPTIONS ====
3493 no-store no new objects should be stored to this cache_dir.
3495 min-size=n the minimum object size in bytes this cache_dir
3496 will accept. It's used to restrict a cache_dir
3497 to only store large objects (e.g. AUFS) while
3498 other stores are optimized for smaller objects
3502 max-size=n the maximum object size in bytes this cache_dir
3504 The value in maximum_object_size directive sets
3505 the default unless more specific details are
3506 available (ie a small store capacity).
3508 Note: To make optimal use of the max-size limits you should order
3509 the cache_dir lines with the smallest max-size value first.
3513 # Uncomment and adjust the following to add a disk cache directory.
3514 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3518 NAME: store_dir_select_algorithm
3520 LOC: Config.store_dir_select_algorithm
3523 How Squid selects which cache_dir to use when the response
3524 object will fit into more than one.
3526 Regardless of which algorithm is used the cache_dir min-size
3527 and max-size parameters are obeyed. As such they can affect
3528 the selection algorithm by limiting the set of considered
3535 This algorithm is suited to caches with similar cache_dir
3536 sizes and disk speeds.
3538 The disk with the least I/O pending is selected.
3539 When there are multiple disks with the same I/O load ranking
3540 the cache_dir with most available capacity is selected.
3542 When a mix of cache_dir sizes are configured the faster disks
3543 have a naturally lower I/O loading and larger disks have more
3544 capacity. So space used to store objects and data throughput
3545 may be very unbalanced towards larger disks.
3550 This algorithm is suited to caches with unequal cache_dir
3553 Each cache_dir is selected in a rotation. The next suitable
3556 Available cache_dir capacity is only considered in relation
3557 to whether the object will fit and meets the min-size and
3558 max-size parameters.
3560 Disk I/O loading is only considered to prevent overload on slow
3561 disks. This algorithm does not spread objects by size, so any
3562 I/O loading per-disk may appear very unbalanced and volatile.
3564 If several cache_dirs use similar min-size, max-size, or other
3565 limits to to reject certain responses, then do not group such
3566 cache_dir lines together, to avoid round-robin selection bias
3567 towards the first cache_dir after the group. Instead, interleave
3568 cache_dir lines from different groups. For example:
3570 store_dir_select_algorithm round-robin
3571 cache_dir rock /hdd1 ... min-size=100000
3572 cache_dir rock /ssd1 ... max-size=99999
3573 cache_dir rock /hdd2 ... min-size=100000
3574 cache_dir rock /ssd2 ... max-size=99999
3575 cache_dir rock /hdd3 ... min-size=100000
3576 cache_dir rock /ssd3 ... max-size=99999
3579 NAME: max_open_disk_fds
3581 LOC: Config.max_open_disk_fds
3583 DEFAULT_DOC: no limit
3585 To avoid having disk as the I/O bottleneck Squid can optionally
3586 bypass the on-disk cache if more than this amount of disk file
3587 descriptors are open.
3589 A value of 0 indicates no limit.
3592 NAME: cache_swap_low
3593 COMMENT: (percent, 0-100)
3596 LOC: Config.Swap.lowWaterMark
3598 The low-water mark for cache object replacement.
3599 Replacement begins when the swap (disk) usage is above the
3600 low-water mark and attempts to maintain utilization near the
3601 low-water mark. As swap utilization gets close to high-water
3602 mark object eviction becomes more aggressive. If utilization is
3603 close to the low-water mark less replacement is done each time.
3605 Defaults are 90% and 95%. If you have a large cache, 5% could be
3606 hundreds of MB. If this is the case you may wish to set these
3607 numbers closer together.
3609 See also cache_swap_high
3612 NAME: cache_swap_high
3613 COMMENT: (percent, 0-100)
3616 LOC: Config.Swap.highWaterMark
3618 The high-water mark for cache object replacement.
3619 Replacement begins when the swap (disk) usage is above the
3620 low-water mark and attempts to maintain utilization near the
3621 low-water mark. As swap utilization gets close to high-water
3622 mark object eviction becomes more aggressive. If utilization is
3623 close to the low-water mark less replacement is done each time.
3625 Defaults are 90% and 95%. If you have a large cache, 5% could be
3626 hundreds of MB. If this is the case you may wish to set these
3627 numbers closer together.
3629 See also cache_swap_low
3634 -----------------------------------------------------------------------------
3641 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
3645 logformat <name> <format specification>
3647 Defines an access log format.
3649 The <format specification> is a string with embedded % format codes
3651 % format codes all follow the same basic structure where all but
3652 the formatcode is optional. Output strings are automatically escaped
3653 as required according to their context and the output format
3654 modifiers are usually not needed, but can be specified if an explicit
3655 output format is desired.
3657 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3659 " output in quoted string format
3660 [ output in squid text log format as used by log_mime_hdrs
3661 # output in URL quoted format
3666 width minimum and/or maximum field width:
3667 [width_min][.width_max]
3668 When minimum starts with 0, the field is zero-padded.
3669 String values exceeding maximum width are truncated.
3671 {arg} argument such as header name etc
3675 % a literal % character
3676 sn Unique sequence number per log line entry
3677 err_code The ID of an error response served by Squid or
3678 a similar internal error identifier.
3679 err_detail Additional err_code-dependent error information.
3680 note The annotation specified by the argument. Also
3681 logs the adaptation meta headers set by the
3682 adaptation_meta configuration parameter.
3683 If no argument given all annotations logged.
3684 The argument may include a separator to use with
3687 By default, multiple note values are separated with ","
3688 and multiple notes are separated with "\r\n".
3689 When logging named notes with %{name}note, the
3690 explicitly configured separator is used between note
3691 values. When logging all notes with %note, the
3692 explicitly configured separator is used between
3693 individual notes. There is currently no way to
3694 specify both value and notes separators when logging
3695 all notes with %note.
3697 Connection related format codes:
3699 >a Client source IP address
3701 >p Client source port
3702 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3703 >la Local IP address the client connected to
3704 >lp Local port number the client connected to
3705 >qos Client connection TOS/DSCP value set by Squid
3706 >nfmark Client connection netfilter mark set by Squid
3708 la Local listening IP address the client connection was connected to.
3709 lp Local listening port number the client connection was connected to.
3711 <a Server IP address of the last server or peer connection
3712 <A Server FQDN or peer name
3713 <p Server port number of the last server or peer connection
3714 <la Local IP address of the last server or peer connection
3715 <lp Local port number of the last server or peer connection
3716 <qos Server connection TOS/DSCP value set by Squid
3717 <nfmark Server connection netfilter mark set by Squid
3719 Time related format codes:
3721 ts Seconds since epoch
3722 tu subsecond time (milliseconds)
3723 tl Local time. Optional strftime format argument
3724 default %d/%b/%Y:%H:%M:%S %z
3725 tg GMT time. Optional strftime format argument
3726 default %d/%b/%Y:%H:%M:%S %z
3727 tr Response time (milliseconds)
3728 dt Total time spent making DNS lookups (milliseconds)
3729 tS Approximate master transaction start time in
3730 <full seconds since epoch>.<fractional seconds> format.
3731 Currently, Squid considers the master transaction
3732 started when a complete HTTP request header initiating
3733 the transaction is received from the client. This is
3734 the same value that Squid uses to calculate transaction
3735 response time when logging %tr to access.log. Currently,
3736 Squid uses millisecond resolution for %tS values,
3737 similar to the default access.log "current time" field
3740 Access Control related format codes:
3742 et Tag returned by external acl
3743 ea Log string returned by external acl
3744 un User name (any available)
3745 ul User name from authentication
3746 ue User name from external acl helper
3747 ui User name from ident
3748 us User name from SSL
3749 credentials Client credentials. The exact meaning depends on
3750 the authentication scheme: For Basic authentication,
3751 it is the password; for Digest, the realm sent by the
3752 client; for NTLM and Negotiate, the client challenge
3753 or client credentials prefixed with "YR " or "KK ".
3755 HTTP related format codes:
3759 [http::]rm Request method (GET/POST etc)
3760 [http::]>rm Request method from client
3761 [http::]<rm Request method sent to server or peer
3762 [http::]ru Request URL from client (historic, filtered for logging)
3763 [http::]>ru Request URL from client
3764 [http::]<ru Request URL sent to server or peer
3765 [http::]>rs Request URL scheme from client
3766 [http::]<rs Request URL scheme sent to server or peer
3767 [http::]>rd Request URL domain from client
3768 [http::]>rd Request URL domain sent to server or peer
3769 [http::]>rP Request URL port from client
3770 [http::]<rP Request URL port sent to server or peer
3771 [http::]rp Request URL path excluding hostname
3772 [http::]>rp Request URL path excluding hostname from client
3773 [http::]<rp Request URL path excluding hostname sent to server or peer
3774 [http::]rv Request protocol version
3775 [http::]>rv Request protocol version from client
3776 [http::]<rv Request protocol version sent to server or peer
3778 [http::]>h Original received request header.
3779 Usually differs from the request header sent by
3780 Squid, although most fields are often preserved.
3781 Accepts optional header field name/value filter
3782 argument using name[:[separator]element] format.
3783 [http::]>ha Received request header after adaptation and
3784 redirection (pre-cache REQMOD vectoring point).
3785 Usually differs from the request header sent by
3786 Squid, although most fields are often preserved.
3787 Optional header name argument as for >h
3792 [http::]<Hs HTTP status code received from the next hop
3793 [http::]>Hs HTTP status code sent to the client
3795 [http::]<h Reply header. Optional header name argument
3798 [http::]mt MIME content type
3803 [http::]st Total size of request + reply traffic with client
3804 [http::]>st Total size of request received from client.
3805 Excluding chunked encoding bytes.
3806 [http::]<st Total size of reply sent to client (after adaptation)
3808 [http::]>sh Size of request headers received from client
3809 [http::]<sh Size of reply headers sent to client (after adaptation)
3811 [http::]<sH Reply high offset sent
3812 [http::]<sS Upstream object size
3814 [http::]<bs Number of HTTP-equivalent message body bytes
3815 received from the next hop, excluding chunked
3816 transfer encoding and control messages.
3817 Generated FTP/Gopher listings are treated as
3823 [http::]<pt Peer response time in milliseconds. The timer starts
3824 when the last request byte is sent to the next hop
3825 and stops when the last response byte is received.
3826 [http::]<tt Total server-side time in milliseconds. The timer
3827 starts with the first connect request (or write I/O)
3828 sent to the first selected peer. The timer stops
3829 with the last I/O with the last peer.
3831 Squid handling related format codes:
3833 Ss Squid request status (TCP_MISS etc)
3834 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3836 SSL-related format codes:
3838 ssl::bump_mode SslBump decision for the transaction:
3840 For CONNECT requests that initiated bumping of
3841 a connection and for any request received on
3842 an already bumped connection, Squid logs the
3843 corresponding SslBump mode ("server-first" or
3844 "client-first"). See the ssl_bump option for
3845 more information about these modes.
3847 A "none" token is logged for requests that
3848 triggered "ssl_bump" ACL evaluation matching
3849 either a "none" rule or no rules at all.
3851 In all other cases, a single dash ("-") is
3854 If ICAP is enabled, the following code becomes available (as
3855 well as ICAP log codes documented with the icap_log option):
3857 icap::tt Total ICAP processing time for the HTTP
3858 transaction. The timer ticks when ICAP
3859 ACLs are checked and when ICAP
3860 transaction is in progress.
3862 If adaptation is enabled the following three codes become available:
3864 adapt::<last_h The header of the last ICAP response or
3865 meta-information from the last eCAP
3866 transaction related to the HTTP transaction.
3867 Like <h, accepts an optional header name
3870 adapt::sum_trs Summed adaptation transaction response
3871 times recorded as a comma-separated list in
3872 the order of transaction start time. Each time
3873 value is recorded as an integer number,
3874 representing response time of one or more
3875 adaptation (ICAP or eCAP) transaction in
3876 milliseconds. When a failed transaction is
3877 being retried or repeated, its time is not
3878 logged individually but added to the
3879 replacement (next) transaction. See also:
3882 adapt::all_trs All adaptation transaction response times.
3883 Same as adaptation_strs but response times of
3884 individual transactions are never added
3885 together. Instead, all transaction response
3886 times are recorded individually.
3888 You can prefix adapt::*_trs format codes with adaptation
3889 service name in curly braces to record response time(s) specific
3890 to that service. For example: %{my_service}adapt::sum_trs
3892 If SSL is enabled, the following formating codes become available:
3894 %ssl::>cert_subject The Subject field of the received client
3895 SSL certificate or a dash ('-') if Squid has
3896 received an invalid/malformed certificate or
3897 no certificate at all. Consider encoding the
3898 logged value because Subject often has spaces.
3900 %ssl::>cert_issuer The Issuer field of the received client
3901 SSL certificate or a dash ('-') if Squid has
3902 received an invalid/malformed certificate or
3903 no certificate at all. Consider encoding the
3904 logged value because Issuer often has spaces.
3906 The default formats available (which do not need re-defining) are:
3908 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3909 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3910 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3911 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3912 logformat useragent %>a [%tl] "%{User-Agent}>h"
3914 NOTE: When the log_mime_hdrs directive is set to ON.
3915 The squid, common and combined formats have a safely encoded copy
3916 of the mime headers appended to each line within a pair of brackets.
3918 NOTE: The common and combined formats are not quite true to the Apache definition.
3919 The logs from Squid contain an extra status and hierarchy code appended.
3923 NAME: access_log cache_access_log
3925 LOC: Config.Log.accesslogs
3926 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3928 Configures whether and how Squid logs HTTP and ICP transactions.
3929 If access logging is enabled, a single line is logged for every
3930 matching HTTP or ICP request. The recommended directive formats are:
3932 access_log <module>:<place> [option ...] [acl acl ...]
3933 access_log none [acl acl ...]
3935 The following directive format is accepted but may be deprecated:
3936 access_log <module>:<place> [<logformat name> [acl acl ...]]
3938 In most cases, the first ACL name must not contain the '=' character
3939 and should not be equal to an existing logformat name. You can always
3940 start with an 'all' ACL to work around those restrictions.
3942 Will log to the specified module:place using the specified format (which
3943 must be defined in a logformat directive) those entries which match
3944 ALL the acl's specified (which must be defined in acl clauses).
3945 If no acl is specified, all requests will be logged to this destination.
3947 ===== Available options for the recommended directive format =====
3949 logformat=name Names log line format (either built-in or
3950 defined by a logformat directive). Defaults
3953 buffer-size=64KB Defines approximate buffering limit for log
3954 records (see buffered_logs). Squid should not
3955 keep more than the specified size and, hence,
3956 should flush records before the buffer becomes
3957 full to avoid overflows under normal
3958 conditions (the exact flushing algorithm is
3959 module-dependent though). The on-error option
3960 controls overflow handling.
3962 on-error=die|drop Defines action on unrecoverable errors. The
3963 'drop' action ignores (i.e., does not log)
3964 affected log records. The default 'die' action
3965 kills the affected worker. The drop action
3966 support has not been tested for modules other
3969 ===== Modules Currently available =====
3971 none Do not log any requests matching these ACL.
3972 Do not specify Place or logformat name.
3974 stdio Write each log line to disk immediately at the completion of
3976 Place: the filename and path to be written.
3978 daemon Very similar to stdio. But instead of writing to disk the log
3979 line is passed to a daemon helper for asychronous handling instead.
3980 Place: varies depending on the daemon.
3982 log_file_daemon Place: the file name and path to be written.
3984 syslog To log each request via syslog facility.
3985 Place: The syslog facility and priority level for these entries.
3986 Place Format: facility.priority
3988 where facility could be any of:
3989 authpriv, daemon, local0 ... local7 or user.
3991 And priority could be any of:
3992 err, warning, notice, info, debug.
3994 udp To send each log line as text data to a UDP receiver.
3995 Place: The destination host name or IP and port.
3996 Place Format: //host:port
3998 tcp To send each log line as text data to a TCP receiver.
3999 Lines may be accumulated before sending (see buffered_logs).
4000 Place: The destination host name or IP and port.
4001 Place Format: //host:port
4004 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
4010 LOC: Config.Log.icaplogs
4013 ICAP log files record ICAP transaction summaries, one line per
4016 The icap_log option format is:
4017 icap_log <filepath> [<logformat name> [acl acl ...]]
4018 icap_log none [acl acl ...]]
4020 Please see access_log option documentation for details. The two
4021 kinds of logs share the overall configuration approach and many
4024 ICAP processing of a single HTTP message or transaction may
4025 require multiple ICAP transactions. In such cases, multiple
4026 ICAP transaction log lines will correspond to a single access
4029 ICAP log uses logformat codes that make sense for an ICAP
4030 transaction. Header-related codes are applied to the HTTP header
4031 embedded in an ICAP server response, with the following caveats:
4032 For REQMOD, there is no HTTP response header unless the ICAP
4033 server performed request satisfaction. For RESPMOD, the HTTP
4034 request header is the header sent to the ICAP server. For
4035 OPTIONS, there are no HTTP headers.
4037 The following format codes are also available for ICAP logs:
4039 icap::<A ICAP server IP address. Similar to <A.
4041 icap::<service_name ICAP service name from the icap_service
4042 option in Squid configuration file.
4044 icap::ru ICAP Request-URI. Similar to ru.
4046 icap::rm ICAP request method (REQMOD, RESPMOD, or
4047 OPTIONS). Similar to existing rm.
4049 icap::>st Bytes sent to the ICAP server (TCP payload
4050 only; i.e., what Squid writes to the socket).
4052 icap::<st Bytes received from the ICAP server (TCP
4053 payload only; i.e., what Squid reads from
4056 icap::<bs Number of message body bytes received from the
4057 ICAP server. ICAP message body, if any, usually
4058 includes encapsulated HTTP message headers and
4059 possibly encapsulated HTTP message body. The
4060 HTTP body part is dechunked before its size is
4063 icap::tr Transaction response time (in
4064 milliseconds). The timer starts when
4065 the ICAP transaction is created and
4066 stops when the transaction is completed.
4069 icap::tio Transaction I/O time (in milliseconds). The
4070 timer starts when the first ICAP request
4071 byte is scheduled for sending. The timers
4072 stops when the last byte of the ICAP response
4075 icap::to Transaction outcome: ICAP_ERR* for all
4076 transaction errors, ICAP_OPT for OPTION
4077 transactions, ICAP_ECHO for 204
4078 responses, ICAP_MOD for message
4079 modification, and ICAP_SAT for request
4080 satisfaction. Similar to Ss.
4082 icap::Hs ICAP response status code. Similar to Hs.
4084 icap::>h ICAP request header(s). Similar to >h.
4086 icap::<h ICAP response header(s). Similar to <h.
4088 The default ICAP log format, which can be used without an explicit
4089 definition, is called icap_squid:
4091 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4093 See also: logformat, log_icap, and %adapt::<last_h
4096 NAME: logfile_daemon
4098 DEFAULT: @DEFAULT_LOGFILED@
4099 LOC: Log::TheConfig.logfile_daemon
4101 Specify the path to the logfile-writing daemon. This daemon is
4102 used to write the access and store logs, if configured.
4104 Squid sends a number of commands to the log daemon:
4105 L<data>\n - logfile data
4110 r<n>\n - set rotate count to <n>
4111 b<n>\n - 1 = buffer output, 0 = don't buffer output
4113 No responses is expected.
4119 Remove this line. Use acls with access_log directives to control access logging
4125 Remove this line. Use acls with icap_log directives to control icap logging
4128 NAME: stats_collection
4130 LOC: Config.accessList.stats_collection
4132 DEFAULT_DOC: Allow logging for all transactions.
4133 COMMENT: allow|deny acl acl...
4135 This options allows you to control which requests gets accounted
4136 in performance counters.
4138 This clause only supports fast acl types.
4139 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4142 NAME: cache_store_log
4145 LOC: Config.Log.store
4147 Logs the activities of the storage manager. Shows which
4148 objects are ejected from the cache, and which objects are
4149 saved and for how long.
4150 There are not really utilities to analyze this data, so you can safely
4151 disable it (the default).
4153 Store log uses modular logging outputs. See access_log for the list
4154 of modules supported.
4157 cache_store_log stdio:@DEFAULT_STORE_LOG@
4158 cache_store_log daemon:@DEFAULT_STORE_LOG@
4161 NAME: cache_swap_state cache_swap_log
4163 LOC: Config.Log.swap
4165 DEFAULT_DOC: Store the journal inside its cache_dir
4167 Location for the cache "swap.state" file. This index file holds
4168 the metadata of objects saved on disk. It is used to rebuild
4169 the cache during startup. Normally this file resides in each
4170 'cache_dir' directory, but you may specify an alternate
4171 pathname here. Note you must give a full filename, not just
4172 a directory. Since this is the index for the whole object
4173 list you CANNOT periodically rotate it!
4175 If %s can be used in the file name it will be replaced with a
4176 a representation of the cache_dir name where each / is replaced
4177 with '.'. This is needed to allow adding/removing cache_dir
4178 lines when cache_swap_log is being used.
4180 If have more than one 'cache_dir', and %s is not used in the name
4181 these swap logs will have names such as:
4187 The numbered extension (which is added automatically)
4188 corresponds to the order of the 'cache_dir' lines in this
4189 configuration file. If you change the order of the 'cache_dir'
4190 lines in this file, these index files will NOT correspond to
4191 the correct 'cache_dir' entry (unless you manually rename
4192 them). We recommend you do NOT use this option. It is
4193 better to keep these index files in each 'cache_dir' directory.
4196 NAME: logfile_rotate
4199 LOC: Config.Log.rotateNumber
4201 Specifies the number of logfile rotations to make when you
4202 type 'squid -k rotate'. The default is 10, which will rotate
4203 with extensions 0 through 9. Setting logfile_rotate to 0 will
4204 disable the file name rotation, but the logfiles are still closed
4205 and re-opened. This will enable you to rename the logfiles
4206 yourself just before sending the rotate signal.
4208 Note, the 'squid -k rotate' command normally sends a USR1
4209 signal to the running squid process. In certain situations
4210 (e.g. on Linux with Async I/O), USR1 is used for other
4211 purposes, so -k rotate uses another signal. It is best to get
4212 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4215 Note, from Squid-3.1 this option is only a default for cache.log,
4216 that log can be rotated separately by using debug_options.
4219 NAME: emulate_httpd_log
4222 Replace this with an access_log directive using the format 'common' or 'combined'.
4225 NAME: log_ip_on_direct
4228 Remove this option from your config. To log server or peer names use %<A in the log format.
4233 DEFAULT: @DEFAULT_MIME_TABLE@
4234 LOC: Config.mimeTablePathname
4236 Path to Squid's icon configuration file.
4238 You shouldn't need to change this, but the default file contains
4239 examples and formatting information if you do.
4245 LOC: Config.onoff.log_mime_hdrs
4248 The Cache can record both the request and the response MIME
4249 headers for each HTTP transaction. The headers are encoded
4250 safely and will appear as two bracketed fields at the end of
4251 the access log (for either the native or httpd-emulated log
4252 formats). To enable this logging set log_mime_hdrs to 'on'.
4258 Replace this with an access_log directive using the format 'useragent'.
4261 NAME: referer_log referrer_log
4264 Replace this with an access_log directive using the format 'referrer'.
4269 DEFAULT: @DEFAULT_PID_FILE@
4270 LOC: Config.pidFilename
4272 A filename to write the process-id to. To disable, enter "none".
4278 Remove this option from your config. To log FQDN use %>A in the log format.
4281 NAME: client_netmask
4283 LOC: Config.Addrs.client_netmask
4285 DEFAULT_DOC: Log full client IP address
4287 A netmask for client addresses in logfiles and cachemgr output.
4288 Change this to protect the privacy of your cache clients.
4289 A netmask of 255.255.255.0 will log all IP's in that range with
4290 the last digit set to '0'.
4296 Use a regular access.log with ACL limiting it to MISS events.
4299 NAME: strip_query_terms
4301 LOC: Config.onoff.strip_query_terms
4304 By default, Squid strips query terms from requested URLs before
4305 logging. This protects your user's privacy and reduces log size.
4307 When investigating HIT/MISS or other caching behaviour you
4308 will need to disable this to see the full URL used by Squid.
4315 LOC: Config.onoff.buffered_logs
4317 Whether to write/send access_log records ASAP or accumulate them and
4318 then write/send them in larger chunks. Buffering may improve
4319 performance because it decreases the number of I/Os. However,
4320 buffering increases the delay before log records become available to
4321 the final recipient (e.g., a disk file or logging daemon) and,
4322 hence, increases the risk of log records loss.
4324 Note that even when buffered_logs are off, Squid may have to buffer
4325 records if it cannot write/send them immediately due to pending I/Os
4326 (e.g., the I/O writing the previous log record) or connectivity loss.
4328 Currently honored by 'daemon' and 'tcp' access_log modules only.
4331 NAME: netdb_filename
4333 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4334 LOC: Config.netdbFilename
4337 Where Squid stores it's netdb journal.
4338 When enabled this journal preserves netdb state between restarts.
4340 To disable, enter "none".
4344 OPTIONS FOR TROUBLESHOOTING
4345 -----------------------------------------------------------------------------
4350 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4351 LOC: Debug::cache_log
4353 Squid administrative logging file.
4355 This is where general information about Squid behavior goes. You can
4356 increase the amount of data logged to this file and how often it is
4357 rotated with "debug_options"
4363 DEFAULT_DOC: Log all critical and important messages.
4364 LOC: Debug::debugOptions
4366 Logging options are set as section,level where each source file
4367 is assigned a unique section. Lower levels result in less
4368 output, Full debugging (level 9) can result in a very large
4369 log file, so be careful.
4371 The magic word "ALL" sets debugging levels for all sections.
4372 The default is to run with "ALL,1" to record important warnings.
4374 The rotate=N option can be used to keep more or less of these logs
4375 than would otherwise be kept by logfile_rotate.
4376 For most uses a single log should be enough to monitor current
4377 events affecting Squid.
4382 LOC: Config.coredump_dir
4383 DEFAULT_IF_NONE: none
4384 DEFAULT_DOC: Use the directory from where Squid was started.
4386 By default Squid leaves core files in the directory from where
4387 it was started. If you set 'coredump_dir' to a directory
4388 that exists, Squid will chdir() to that directory at startup
4389 and coredump files will be left there.
4393 # Leave coredumps in the first cache dir
4394 coredump_dir @DEFAULT_SWAP_DIR@
4400 OPTIONS FOR FTP GATEWAYING
4401 -----------------------------------------------------------------------------
4407 LOC: Config.Ftp.anon_user
4409 If you want the anonymous login password to be more informative
4410 (and enable the use of picky FTP servers), set this to something
4411 reasonable for your domain, like wwwuser@somewhere.net
4413 The reason why this is domainless by default is the
4414 request can be made on the behalf of a user in any domain,
4415 depending on how the cache is used.
4416 Some FTP server also validate the email address is valid
4417 (for example perl.com).
4423 LOC: Config.Ftp.passive
4425 If your firewall does not allow Squid to use passive
4426 connections, turn off this option.
4428 Use of ftp_epsv_all option requires this to be ON.
4434 LOC: Config.Ftp.epsv_all
4436 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4438 NATs may be able to put the connection on a "fast path" through the
4439 translator, as the EPRT command will never be used and therefore,
4440 translation of the data portion of the segments will never be needed.
4442 When a client only expects to do two-way FTP transfers this may be
4444 If squid finds that it must do a three-way FTP transfer after issuing
4445 an EPSV ALL command, the FTP session will fail.
4447 If you have any doubts about this option do not use it.
4448 Squid will nicely attempt all other connection methods.
4450 Requires ftp_passive to be ON (default) for any effect.
4456 LOC: Config.accessList.ftp_epsv
4458 FTP Protocol extensions permit the use of a special "EPSV" command.
4460 NATs may be able to put the connection on a "fast path" through the
4461 translator using EPSV, as the EPRT command will never be used
4462 and therefore, translation of the data portion of the segments
4463 will never be needed.
4465 EPSV is often required to interoperate with FTP servers on IPv6
4466 networks. On the other hand, it may break some IPv4 servers.
4468 By default, EPSV may try EPSV with any FTP server. To fine tune
4469 that decision, you may restrict EPSV to certain clients or servers
4472 ftp_epsv allow|deny al1 acl2 ...
4474 WARNING: Disabling EPSV may cause problems with external NAT and IPv6.
4476 Only fast ACLs are supported.
4477 Requires ftp_passive to be ON (default) for any effect.
4483 LOC: Config.Ftp.eprt
4485 FTP Protocol extensions permit the use of a special "EPRT" command.
4487 This extension provides a protocol neutral alternative to the
4488 IPv4-only PORT command. When supported it enables active FTP data
4489 channels over IPv6 and efficient NAT handling.
4491 Turning this OFF will prevent EPRT being attempted and will skip
4492 straight to using PORT for IPv4 servers.
4494 Some devices are known to not handle this extension correctly and
4495 may result in crashes. Devices which suport EPRT enough to fail
4496 cleanly will result in Squid attempting PORT anyway. This directive
4497 should only be disabled when EPRT results in device failures.
4499 WARNING: Doing so will convert Squid back to the old behavior with all
4500 the related problems with external NAT devices/layers and IPv4-only FTP.
4503 NAME: ftp_sanitycheck
4506 LOC: Config.Ftp.sanitycheck
4508 For security and data integrity reasons Squid by default performs
4509 sanity checks of the addresses of FTP data connections ensure the
4510 data connection is to the requested server. If you need to allow
4511 FTP connections to servers using another IP address for the data
4512 connection turn this off.
4515 NAME: ftp_telnet_protocol
4518 LOC: Config.Ftp.telnet
4520 The FTP protocol is officially defined to use the telnet protocol
4521 as transport channel for the control connection. However, many
4522 implementations are broken and does not respect this aspect of
4525 If you have trouble accessing files with ASCII code 255 in the
4526 path or similar problems involving this ASCII code you can
4527 try setting this directive to off. If that helps, report to the
4528 operator of the FTP server in question that their FTP server
4529 is broken and does not follow the FTP standard.
4533 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4534 -----------------------------------------------------------------------------
4539 DEFAULT: @DEFAULT_DISKD@
4540 LOC: Config.Program.diskd
4542 Specify the location of the diskd executable.
4543 Note this is only useful if you have compiled in
4544 diskd as one of the store io modules.
4547 NAME: unlinkd_program
4550 DEFAULT: @DEFAULT_UNLINKD@
4551 LOC: Config.Program.unlinkd
4553 Specify the location of the executable for file deletion process.
4556 NAME: pinger_program
4558 DEFAULT: @DEFAULT_PINGER@
4559 LOC: Config.pinger.program
4562 Specify the location of the executable for the pinger process.
4568 LOC: Config.pinger.enable
4571 Control whether the pinger is active at run-time.
4572 Enables turning ICMP pinger on and off with a simple
4573 squid -k reconfigure.
4578 OPTIONS FOR URL REWRITING
4579 -----------------------------------------------------------------------------
4582 NAME: url_rewrite_program redirect_program
4584 LOC: Config.Program.redirect
4587 Specify the location of the executable URL rewriter to use.
4588 Since they can perform almost any function there isn't one included.
4590 For each requested URL, the rewriter will receive on line with the format
4592 [channel-ID <SP>] URL [<SP> extras]<NL>
4594 See url_rewrite_extras on how to send "extras" with optional values to
4596 After processing the request the helper must reply using the following format:
4598 [channel-ID <SP>] result [<SP> kv-pairs]
4600 The result code can be:
4602 OK status=30N url="..."
4603 Redirect the URL to the one supplied in 'url='.
4604 'status=' is optional and contains the status code to send
4605 the client in Squids HTTP response. It must be one of the
4606 HTTP redirect status codes: 301, 302, 303, 307, 308.
4607 When no status is given Squid will use 302.
4609 OK rewrite-url="..."
4610 Rewrite the URL to the one supplied in 'rewrite-url='.
4611 The new URL is fetched directly by Squid and returned to
4612 the client as the response to its request.
4615 When neither of url= and rewrite-url= are sent Squid does
4619 Do not change the URL.
4622 An internal error occurred in the helper, preventing
4623 a result being identified. The 'message=' key name is
4624 reserved for delivering a log message.
4627 In addition to the above kv-pairs Squid also understands the following
4628 optional kv-pairs received from URL rewriters:
4630 Associates a TAG with the client TCP connection.
4631 The TAG is treated as a regular annotation but persists across
4632 future requests on the client connection rather than just the
4633 current request. A helper may update the TAG during subsequent
4634 requests be returning a new kv-pair.
4636 When using the concurrency= option the protocol is changed by
4637 introducing a query channel tag in front of the request/response.
4638 The query channel tag is a number between 0 and concurrency-1.
4639 This value must be echoed back unchanged to Squid as the first part
4640 of the response relating to its request.
4642 WARNING: URL re-writing ability should be avoided whenever possible.
4643 Use the URL redirect form of response instead.
4645 Re-write creates a difference in the state held by the client
4646 and server. Possibly causing confusion when the server response
4647 contains snippets of its view state. Embeded URLs, response
4648 and content Location headers, etc. are not re-written by this
4651 By default, a URL rewriter is not used.
4654 NAME: url_rewrite_children redirect_children
4655 TYPE: HelperChildConfig
4656 DEFAULT: 20 startup=0 idle=1 concurrency=0
4657 LOC: Config.redirectChildren
4659 The maximum number of redirector processes to spawn. If you limit
4660 it too few Squid will have to wait for them to process a backlog of
4661 URLs, slowing it down. If you allow too many they will use RAM
4662 and other system resources noticably.
4664 The startup= and idle= options allow some measure of skew in your
4669 Sets a minimum of how many processes are to be spawned when Squid
4670 starts or reconfigures. When set to zero the first request will
4671 cause spawning of the first child process to handle it.
4673 Starting too few will cause an initial slowdown in traffic as Squid
4674 attempts to simultaneously spawn enough processes to cope.
4678 Sets a minimum of how many processes Squid is to try and keep available
4679 at all times. When traffic begins to rise above what the existing
4680 processes can handle this many more will be spawned up to the maximum
4681 configured. A minimum setting of 1 is required.
4685 The number of requests each redirector helper can handle in
4686 parallel. Defaults to 0 which indicates the redirector
4687 is a old-style single threaded redirector.
4689 When this directive is set to a value >= 1 then the protocol
4690 used to communicate with the helper is modified to include
4691 an ID in front of the request/response. The ID from the request
4692 must be echoed back with the response to that request.
4695 NAME: url_rewrite_host_header redirect_rewrites_host_header
4698 LOC: Config.onoff.redir_rewrites_host
4700 To preserve same-origin security policies in browsers and
4701 prevent Host: header forgery by redirectors Squid rewrites
4702 any Host: header in redirected requests.
4704 If you are running an accelerator this may not be a wanted
4705 effect of a redirector. This directive enables you disable
4706 Host: alteration in reverse-proxy traffic.
4708 WARNING: Entries are cached on the result of the URL rewriting
4709 process, so be careful if you have domain-virtual hosts.
4711 WARNING: Squid and other software verifies the URL and Host
4712 are matching, so be careful not to relay through other proxies
4713 or inspecting firewalls with this disabled.
4716 NAME: url_rewrite_access redirector_access
4719 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4720 LOC: Config.accessList.redirector
4722 If defined, this access list specifies which requests are
4723 sent to the redirector processes.
4725 This clause supports both fast and slow acl types.
4726 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4729 NAME: url_rewrite_bypass redirector_bypass
4731 LOC: Config.onoff.redirector_bypass
4734 When this is 'on', a request will not go through the
4735 redirector if all the helpers are busy. If this is 'off'
4736 and the redirector queue grows too large, Squid will exit
4737 with a FATAL error and ask you to increase the number of
4738 redirectors. You should only enable this if the redirectors
4739 are not critical to your caching system. If you use
4740 redirectors for access control, and you enable this option,
4741 users may have access to pages they should not
4742 be allowed to request.
4745 NAME: url_rewrite_extras
4746 TYPE: TokenOrQuotedString
4747 LOC: Config.redirector_extras
4748 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4750 Specifies a string to be append to request line format for the
4751 rewriter helper. "Quoted" format values may contain spaces and
4752 logformat %macros. In theory, any logformat %macro can be used.
4753 In practice, a %macro expands as a dash (-) if the helper request is
4754 sent before the required macro information is available to Squid.
4758 OPTIONS FOR STORE ID
4759 -----------------------------------------------------------------------------
4762 NAME: store_id_program storeurl_rewrite_program
4764 LOC: Config.Program.store_id
4767 Specify the location of the executable StoreID helper to use.
4768 Since they can perform almost any function there isn't one included.
4770 For each requested URL, the helper will receive one line with the format
4772 [channel-ID <SP>] URL [<SP> extras]<NL>
4775 After processing the request the helper must reply using the following format:
4777 [channel-ID <SP>] result [<SP> kv-pairs]
4779 The result code can be:
4782 Use the StoreID supplied in 'store-id='.
4785 The default is to use HTTP request URL as the store ID.
4788 An internal error occured in the helper, preventing
4789 a result being identified.
4791 In addition to the above kv-pairs Squid also understands the following
4792 optional kv-pairs received from URL rewriters:
4794 Associates a TAG with the client TCP connection.
4795 Please see url_rewrite_program related documentation for this
4798 Helper programs should be prepared to receive and possibly ignore
4799 additional whitespace-separated tokens on each input line.
4801 When using the concurrency= option the protocol is changed by
4802 introducing a query channel tag in front of the request/response.
4803 The query channel tag is a number between 0 and concurrency-1.
4804 This value must be echoed back unchanged to Squid as the first part
4805 of the response relating to its request.
4807 NOTE: when using StoreID refresh_pattern will apply to the StoreID
4808 returned from the helper and not the URL.
4810 WARNING: Wrong StoreID value returned by a careless helper may result
4811 in the wrong cached response returned to the user.
4813 By default, a StoreID helper is not used.
4816 NAME: store_id_extras
4817 TYPE: TokenOrQuotedString
4818 LOC: Config.storeId_extras
4819 DEFAULT: "%>a/%>A %un %>rm myip=%la myport=%lp"
4821 Specifies a string to be append to request line format for the
4822 StoreId helper. "Quoted" format values may contain spaces and
4823 logformat %macros. In theory, any logformat %macro can be used.
4824 In practice, a %macro expands as a dash (-) if the helper request is
4825 sent before the required macro information is available to Squid.
4828 NAME: store_id_children storeurl_rewrite_children
4829 TYPE: HelperChildConfig
4830 DEFAULT: 20 startup=0 idle=1 concurrency=0
4831 LOC: Config.storeIdChildren
4833 The maximum number of StoreID helper processes to spawn. If you limit
4834 it too few Squid will have to wait for them to process a backlog of
4835 requests, slowing it down. If you allow too many they will use RAM
4836 and other system resources noticably.
4838 The startup= and idle= options allow some measure of skew in your
4843 Sets a minimum of how many processes are to be spawned when Squid
4844 starts or reconfigures. When set to zero the first request will
4845 cause spawning of the first child process to handle it.
4847 Starting too few will cause an initial slowdown in traffic as Squid
4848 attempts to simultaneously spawn enough processes to cope.
4852 Sets a minimum of how many processes Squid is to try and keep available
4853 at all times. When traffic begins to rise above what the existing
4854 processes can handle this many more will be spawned up to the maximum
4855 configured. A minimum setting of 1 is required.
4859 The number of requests each storeID helper can handle in
4860 parallel. Defaults to 0 which indicates the helper
4861 is a old-style single threaded program.
4863 When this directive is set to a value >= 1 then the protocol
4864 used to communicate with the helper is modified to include
4865 an ID in front of the request/response. The ID from the request
4866 must be echoed back with the response to that request.
4869 NAME: store_id_access storeurl_rewrite_access
4872 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4873 LOC: Config.accessList.store_id
4875 If defined, this access list specifies which requests are
4876 sent to the StoreID processes. By default all requests
4879 This clause supports both fast and slow acl types.
4880 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4883 NAME: store_id_bypass storeurl_rewrite_bypass
4885 LOC: Config.onoff.store_id_bypass
4888 When this is 'on', a request will not go through the
4889 helper if all helpers are busy. If this is 'off'
4890 and the helper queue grows too large, Squid will exit
4891 with a FATAL error and ask you to increase the number of
4892 helpers. You should only enable this if the helperss
4893 are not critical to your caching system. If you use
4894 helpers for critical caching components, and you enable this
4895 option, users may not get objects from cache.
4899 OPTIONS FOR TUNING THE CACHE
4900 -----------------------------------------------------------------------------
4903 NAME: cache no_cache
4906 DEFAULT_DOC: By default, this directive is unused and has no effect.
4907 LOC: Config.accessList.noCache
4909 Requests denied by this directive will not be served from the cache
4910 and their responses will not be stored in the cache. This directive
4911 has no effect on other transactions and on already cached responses.
4913 This clause supports both fast and slow acl types.
4914 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4916 This and the two other similar caching directives listed below are
4917 checked at different transaction processing stages, have different
4918 access to response information, affect different cache operations,
4919 and differ in slow ACLs support:
4921 * cache: Checked before Squid makes a hit/miss determination.
4922 No access to reply information!
4923 Denies both serving a hit and storing a miss.
4924 Supports both fast and slow ACLs.
4925 * send_hit: Checked after a hit was detected.
4926 Has access to reply (hit) information.
4927 Denies serving a hit only.
4928 Supports fast ACLs only.
4929 * store_miss: Checked before storing a cachable miss.
4930 Has access to reply (miss) information.
4931 Denies storing a miss only.
4932 Supports fast ACLs only.
4934 If you are not sure which of the three directives to use, apply the
4935 following decision logic:
4937 * If your ACL(s) are of slow type _and_ need response info, redesign.
4938 Squid does not support that particular combination at this time.
4940 * If your directive ACL(s) are of slow type, use "cache"; and/or
4941 * if your directive ACL(s) need no response info, use "cache".
4943 * If you do not want the response cached, use store_miss; and/or
4944 * if you do not want a hit on a cached response, use send_hit.
4950 DEFAULT_DOC: By default, this directive is unused and has no effect.
4951 LOC: Config.accessList.sendHit
4953 Responses denied by this directive will not be served from the cache
4954 (but may still be cached, see store_miss). This directive has no
4955 effect on the responses it allows and on the cached objects.
4957 Please see the "cache" directive for a summary of differences among
4958 store_miss, send_hit, and cache directives.
4960 Unlike the "cache" directive, send_hit only supports fast acl
4961 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4965 # apply custom Store ID mapping to some URLs
4966 acl MapMe dstdomain .c.example.com
4967 store_id_program ...
4968 store_id_access allow MapMe
4970 # but prevent caching of special responses
4971 # such as 302 redirects that cause StoreID loops
4972 acl Ordinary http_status 200-299
4973 store_miss deny MapMe !Ordinary
4975 # and do not serve any previously stored special responses
4976 # from the cache (in case they were already cached before
4977 # the above store_miss rule was in effect).
4978 send_hit deny MapMe !Ordinary
4984 DEFAULT_DOC: By default, this directive is unused and has no effect.
4985 LOC: Config.accessList.storeMiss
4987 Responses denied by this directive will not be cached (but may still
4988 be served from the cache, see send_hit). This directive has no
4989 effect on the responses it allows and on the already cached responses.
4991 Please see the "cache" directive for a summary of differences among
4992 store_miss, send_hit, and cache directives. See the
4993 send_hit directive for a usage example.
4995 Unlike the "cache" directive, store_miss only supports fast acl
4996 types. See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5002 LOC: Config.maxStale
5005 This option puts an upper limit on how stale content Squid
5006 will serve from the cache if cache validation fails.
5007 Can be overriden by the refresh_pattern max-stale option.
5010 NAME: refresh_pattern
5011 TYPE: refreshpattern
5015 usage: refresh_pattern [-i] regex min percent max [options]
5017 By default, regular expressions are CASE-SENSITIVE. To make
5018 them case-insensitive, use the -i option.
5020 'Min' is the time (in minutes) an object without an explicit
5021 expiry time should be considered fresh. The recommended
5022 value is 0, any higher values may cause dynamic applications
5023 to be erroneously cached unless the application designer
5024 has taken the appropriate actions.
5026 'Percent' is a percentage of the objects age (time since last
5027 modification age) an object without explicit expiry time
5028 will be considered fresh.
5030 'Max' is an upper limit on how long objects without an explicit
5031 expiry time will be considered fresh.
5033 options: override-expire
5038 ignore-must-revalidate
5045 override-expire enforces min age even if the server
5046 sent an explicit expiry time (e.g., with the
5047 Expires: header or Cache-Control: max-age). Doing this
5048 VIOLATES the HTTP standard. Enabling this feature
5049 could make you liable for problems which it causes.
5051 Note: override-expire does not enforce staleness - it only extends
5052 freshness / min. If the server returns a Expires time which
5053 is longer than your max time, Squid will still consider
5054 the object fresh for that period of time.
5056 override-lastmod enforces min age even on objects
5057 that were modified recently.
5059 reload-into-ims changes a client no-cache or ``reload''
5060 request for a cached entry into a conditional request using
5061 If-Modified-Since and/or If-None-Match headers, provided the
5062 cached entry has a Last-Modified and/or a strong ETag header.
5063 Doing this VIOLATES the HTTP standard. Enabling this feature
5064 could make you liable for problems which it causes.
5066 ignore-reload ignores a client no-cache or ``reload''
5067 header. Doing this VIOLATES the HTTP standard. Enabling
5068 this feature could make you liable for problems which
5071 ignore-no-store ignores any ``Cache-control: no-store''
5072 headers received from a server. Doing this VIOLATES
5073 the HTTP standard. Enabling this feature could make you
5074 liable for problems which it causes.
5076 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
5077 headers received from a server. Doing this VIOLATES
5078 the HTTP standard. Enabling this feature could make you
5079 liable for problems which it causes.
5081 ignore-private ignores any ``Cache-control: private''
5082 headers received from a server. Doing this VIOLATES
5083 the HTTP standard. Enabling this feature could make you
5084 liable for problems which it causes.
5086 ignore-auth caches responses to requests with authorization,
5087 as if the originserver had sent ``Cache-control: public''
5088 in the response header. Doing this VIOLATES the HTTP standard.
5089 Enabling this feature could make you liable for problems which
5092 refresh-ims causes squid to contact the origin server
5093 when a client issues an If-Modified-Since request. This
5094 ensures that the client will receive an updated version
5095 if one is available.
5097 store-stale stores responses even if they don't have explicit
5098 freshness or a validator (i.e., Last-Modified or an ETag)
5099 present, or if they're already stale. By default, Squid will
5100 not cache such responses because they usually can't be
5101 reused. Note that such responses will be stale by default.
5103 max-stale=NN provide a maximum staleness factor. Squid won't
5104 serve objects more stale than this even if it failed to
5105 validate the object. Default: use the max_stale global limit.
5107 Basically a cached object is:
5109 FRESH if expires < now, else STALE
5111 FRESH if lm-factor < percent, else STALE
5115 The refresh_pattern lines are checked in the order listed here.
5116 The first entry which matches is used. If none of the entries
5117 match the default will be used.
5119 Note, you must uncomment all the default lines if you want
5120 to change one. The default setting is only active if none is
5126 # Add any of your own refresh_pattern entries above these.
5128 refresh_pattern ^ftp: 1440 20% 10080
5129 refresh_pattern ^gopher: 1440 0% 1440
5130 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
5131 refresh_pattern . 0 20% 4320
5135 NAME: quick_abort_min
5139 LOC: Config.quickAbort.min
5142 NAME: quick_abort_max
5146 LOC: Config.quickAbort.max
5149 NAME: quick_abort_pct
5153 LOC: Config.quickAbort.pct
5155 The cache by default continues downloading aborted requests
5156 which are almost completed (less than 16 KB remaining). This
5157 may be undesirable on slow (e.g. SLIP) links and/or very busy
5158 caches. Impatient users may tie up file descriptors and
5159 bandwidth by repeatedly requesting and immediately aborting
5162 When the user aborts a request, Squid will check the
5163 quick_abort values to the amount of data transferred until
5166 If the transfer has less than 'quick_abort_min' KB remaining,
5167 it will finish the retrieval.
5169 If the transfer has more than 'quick_abort_max' KB remaining,
5170 it will abort the retrieval.
5172 If more than 'quick_abort_pct' of the transfer has completed,
5173 it will finish the retrieval.
5175 If you do not want any retrieval to continue after the client
5176 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
5179 If you want retrievals to always continue if they are being
5180 cached set 'quick_abort_min' to '-1 KB'.
5183 NAME: read_ahead_gap
5184 COMMENT: buffer-size
5186 LOC: Config.readAheadGap
5189 The amount of data the cache will buffer ahead of what has been
5190 sent to the client when retrieving an object from another server.
5194 IFDEF: USE_HTTP_VIOLATIONS
5197 LOC: Config.negativeTtl
5200 Set the Default Time-to-Live (TTL) for failed requests.
5201 Certain types of failures (such as "connection refused" and
5202 "404 Not Found") are able to be negatively-cached for a short time.
5203 Modern web servers should provide Expires: header, however if they
5204 do not this can provide a minimum TTL.
5205 The default is not to cache errors with unknown expiry details.
5207 Note that this is different from negative caching of DNS lookups.
5209 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5210 this feature could make you liable for problems which it
5214 NAME: positive_dns_ttl
5217 LOC: Config.positiveDnsTtl
5220 Upper limit on how long Squid will cache positive DNS responses.
5221 Default is 6 hours (360 minutes). This directive must be set
5222 larger than negative_dns_ttl.
5225 NAME: negative_dns_ttl
5228 LOC: Config.negativeDnsTtl
5231 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5232 This also sets the lower cache limit on positive lookups.
5233 Minimum value is 1 second, and it is not recommendable to go
5234 much below 10 seconds.
5237 NAME: range_offset_limit
5238 COMMENT: size [acl acl...]
5240 LOC: Config.rangeOffsetLimit
5243 usage: (size) [units] [[!]aclname]
5245 Sets an upper limit on how far (number of bytes) into the file
5246 a Range request may be to cause Squid to prefetch the whole file.
5247 If beyond this limit, Squid forwards the Range request as it is and
5248 the result is NOT cached.
5250 This is to stop a far ahead range request (lets say start at 17MB)
5251 from making Squid fetch the whole object up to that point before
5252 sending anything to the client.
5254 Multiple range_offset_limit lines may be specified, and they will
5255 be searched from top to bottom on each request until a match is found.
5256 The first match found will be used. If no line matches a request, the
5257 default limit of 0 bytes will be used.
5259 'size' is the limit specified as a number of units.
5261 'units' specifies whether to use bytes, KB, MB, etc.
5262 If no units are specified bytes are assumed.
5264 A size of 0 causes Squid to never fetch more than the
5265 client requested. (default)
5267 A size of 'none' causes Squid to always fetch the object from the
5268 beginning so it may cache the result. (2.0 style)
5270 'aclname' is the name of a defined ACL.
5272 NP: Using 'none' as the byte value here will override any quick_abort settings
5273 that may otherwise apply to the range request. The range request will
5274 be fully fetched from start to finish regardless of the client
5275 actions. This affects bandwidth usage.
5278 NAME: minimum_expiry_time
5281 LOC: Config.minimum_expiry_time
5284 The minimum caching time according to (Expires - Date)
5285 headers Squid honors if the object can't be revalidated.
5286 The default is 60 seconds.
5288 In reverse proxy environments it might be desirable to honor
5289 shorter object lifetimes. It is most likely better to make
5290 your server return a meaningful Last-Modified header however.
5292 In ESI environments where page fragments often have short
5293 lifetimes, this will often be best set to 0.
5296 NAME: store_avg_object_size
5300 LOC: Config.Store.avgObjectSize
5302 Average object size, used to estimate number of objects your
5303 cache can hold. The default is 13 KB.
5305 This is used to pre-seed the cache index memory allocation to
5306 reduce expensive reallocate operations while handling clients
5307 traffic. Too-large values may result in memory allocation during
5308 peak traffic, too-small values will result in wasted memory.
5310 Check the cache manager 'info' report metrics for the real
5311 object sizes seen by your Squid before tuning this.
5314 NAME: store_objects_per_bucket
5317 LOC: Config.Store.objectsPerBucket
5319 Target number of objects per bucket in the store hash table.
5320 Lowering this value increases the total number of buckets and
5321 also the storage maintenance rate. The default is 20.
5326 -----------------------------------------------------------------------------
5329 NAME: request_header_max_size
5333 LOC: Config.maxRequestHeaderSize
5335 This specifies the maximum size for HTTP headers in a request.
5336 Request headers are usually relatively small (about 512 bytes).
5337 Placing a limit on the request header size will catch certain
5338 bugs (for example with persistent connections) and possibly
5339 buffer-overflow or denial-of-service attacks.
5342 NAME: reply_header_max_size
5346 LOC: Config.maxReplyHeaderSize
5348 This specifies the maximum size for HTTP headers in a reply.
5349 Reply headers are usually relatively small (about 512 bytes).
5350 Placing a limit on the reply header size will catch certain
5351 bugs (for example with persistent connections) and possibly
5352 buffer-overflow or denial-of-service attacks.
5355 NAME: request_body_max_size
5359 DEFAULT_DOC: No limit.
5360 LOC: Config.maxRequestBodySize
5362 This specifies the maximum size for an HTTP request body.
5363 In other words, the maximum size of a PUT/POST request.
5364 A user who attempts to send a request with a body larger
5365 than this limit receives an "Invalid Request" error message.
5366 If you set this parameter to a zero (the default), there will
5367 be no limit imposed.
5369 See also client_request_buffer_max_size for an alternative
5370 limitation on client uploads which can be configured.
5373 NAME: client_request_buffer_max_size
5377 LOC: Config.maxRequestBufferSize
5379 This specifies the maximum buffer size of a client request.
5380 It prevents squid eating too much memory when somebody uploads
5384 NAME: chunked_request_body_max_size
5388 LOC: Config.maxChunkedRequestBodySize
5390 A broken or confused HTTP/1.1 client may send a chunked HTTP
5391 request to Squid. Squid does not have full support for that
5392 feature yet. To cope with such requests, Squid buffers the
5393 entire request and then dechunks request body to create a
5394 plain HTTP/1.0 request with a known content length. The plain
5395 request is then used by the rest of Squid code as usual.
5397 The option value specifies the maximum size of the buffer used
5398 to hold the request before the conversion. If the chunked
5399 request size exceeds the specified limit, the conversion
5400 fails, and the client receives an "unsupported request" error,
5401 as if dechunking was disabled.
5403 Dechunking is enabled by default. To disable conversion of
5404 chunked requests, set the maximum to zero.
5406 Request dechunking feature and this option in particular are a
5407 temporary hack. When chunking requests and responses are fully
5408 supported, there will be no need to buffer a chunked request.
5412 IFDEF: USE_HTTP_VIOLATIONS
5415 DEFAULT_DOC: Obey RFC 2616.
5416 LOC: Config.accessList.brokenPosts
5418 A list of ACL elements which, if matched, causes Squid to send
5419 an extra CRLF pair after the body of a PUT/POST request.
5421 Some HTTP servers has broken implementations of PUT/POST,
5422 and rely on an extra CRLF pair sent by some WWW clients.
5424 Quote from RFC2616 section 4.1 on this matter:
5426 Note: certain buggy HTTP/1.0 client implementations generate an
5427 extra CRLF's after a POST request. To restate what is explicitly
5428 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5429 a request with an extra CRLF.
5431 This clause only supports fast acl types.
5432 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5435 acl buggy_server url_regex ^http://....
5436 broken_posts allow buggy_server
5439 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5442 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5444 LOC: Adaptation::Config::use_indirect_client
5446 Controls whether the indirect client IP address (instead of the direct
5447 client IP address) is passed to adaptation services.
5449 See also: follow_x_forwarded_for adaptation_send_client_ip
5453 IFDEF: USE_HTTP_VIOLATIONS
5457 LOC: Config.onoff.via
5459 If set (default), Squid will include a Via header in requests and
5460 replies as required by RFC2616.
5466 LOC: Config.onoff.ie_refresh
5469 Microsoft Internet Explorer up until version 5.5 Service
5470 Pack 1 has an issue with transparent proxies, wherein it
5471 is impossible to force a refresh. Turning this on provides
5472 a partial fix to the problem, by causing all IMS-REFRESH
5473 requests from older IE versions to check the origin server
5474 for fresh content. This reduces hit ratio by some amount
5475 (~10% in my experience), but allows users to actually get
5476 fresh content when they want it. Note because Squid
5477 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5478 of 5.5 is unchanged from old versions of Squid (i.e. a
5479 forced refresh is impossible). Newer versions of IE will,
5480 hopefully, continue to have the new behavior and will be
5481 handled based on that assumption. This option defaults to
5482 the old Squid behavior, which is better for hit ratios but
5483 worse for clients using IE, if they need to be able to
5484 force fresh content.
5487 NAME: vary_ignore_expire
5490 LOC: Config.onoff.vary_ignore_expire
5493 Many HTTP servers supporting Vary gives such objects
5494 immediate expiry time with no cache-control header
5495 when requested by a HTTP/1.0 client. This option
5496 enables Squid to ignore such expiry times until
5497 HTTP/1.1 is fully implemented.
5499 WARNING: If turned on this may eventually cause some
5500 varying objects not intended for caching to get cached.
5503 NAME: request_entities
5505 LOC: Config.onoff.request_entities
5508 Squid defaults to deny GET and HEAD requests with request entities,
5509 as the meaning of such requests are undefined in the HTTP standard
5510 even if not explicitly forbidden.
5512 Set this directive to on if you have clients which insists
5513 on sending request entities in GET or HEAD requests. But be warned
5514 that there is server software (both proxies and web servers) which
5515 can fail to properly process this kind of request which may make you
5516 vulnerable to cache pollution attacks if enabled.
5519 NAME: request_header_access
5520 IFDEF: USE_HTTP_VIOLATIONS
5521 TYPE: http_header_access
5522 LOC: Config.request_header_access
5524 DEFAULT_DOC: No limits.
5526 Usage: request_header_access header_name allow|deny [!]aclname ...
5528 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5529 this feature could make you liable for problems which it
5532 This option replaces the old 'anonymize_headers' and the
5533 older 'http_anonymizer' option with something that is much
5534 more configurable. A list of ACLs for each header name allows
5535 removal of specific header fields under specific conditions.
5537 This option only applies to outgoing HTTP request headers (i.e.,
5538 headers sent by Squid to the next HTTP hop such as a cache peer
5539 or an origin server). The option has no effect during cache hit
5540 detection. The equivalent adaptation vectoring point in ICAP
5541 terminology is post-cache REQMOD.
5543 The option is applied to individual outgoing request header
5544 fields. For each request header field F, Squid uses the first
5545 qualifying sets of request_header_access rules:
5547 1. Rules with header_name equal to F's name.
5548 2. Rules with header_name 'Other', provided F's name is not
5549 on the hard-coded list of commonly used HTTP header names.
5550 3. Rules with header_name 'All'.
5552 Within that qualifying rule set, rule ACLs are checked as usual.
5553 If ACLs of an "allow" rule match, the header field is allowed to
5554 go through as is. If ACLs of a "deny" rule match, the header is
5555 removed and request_header_replace is then checked to identify
5556 if the removed header has a replacement. If no rules within the
5557 set have matching ACLs, the header field is left as is.
5559 For example, to achieve the same behavior as the old
5560 'http_anonymizer standard' option, you should use:
5562 request_header_access From deny all
5563 request_header_access Referer deny all
5564 request_header_access User-Agent deny all
5566 Or, to reproduce the old 'http_anonymizer paranoid' feature
5569 request_header_access Authorization allow all
5570 request_header_access Proxy-Authorization allow all
5571 request_header_access Cache-Control allow all
5572 request_header_access Content-Length allow all
5573 request_header_access Content-Type allow all
5574 request_header_access Date allow all
5575 request_header_access Host allow all
5576 request_header_access If-Modified-Since allow all
5577 request_header_access Pragma allow all
5578 request_header_access Accept allow all
5579 request_header_access Accept-Charset allow all
5580 request_header_access Accept-Encoding allow all
5581 request_header_access Accept-Language allow all
5582 request_header_access Connection allow all
5583 request_header_access All deny all
5585 HTTP reply headers are controlled with the reply_header_access directive.
5587 By default, all headers are allowed (no anonymizing is performed).
5590 NAME: reply_header_access
5591 IFDEF: USE_HTTP_VIOLATIONS
5592 TYPE: http_header_access
5593 LOC: Config.reply_header_access
5595 DEFAULT_DOC: No limits.
5597 Usage: reply_header_access header_name allow|deny [!]aclname ...
5599 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5600 this feature could make you liable for problems which it
5603 This option only applies to reply headers, i.e., from the
5604 server to the client.
5606 This is the same as request_header_access, but in the other
5607 direction. Please see request_header_access for detailed
5610 For example, to achieve the same behavior as the old
5611 'http_anonymizer standard' option, you should use:
5613 reply_header_access Server deny all
5614 reply_header_access WWW-Authenticate deny all
5615 reply_header_access Link deny all
5617 Or, to reproduce the old 'http_anonymizer paranoid' feature
5620 reply_header_access Allow allow all
5621 reply_header_access WWW-Authenticate allow all
5622 reply_header_access Proxy-Authenticate allow all
5623 reply_header_access Cache-Control allow all
5624 reply_header_access Content-Encoding allow all
5625 reply_header_access Content-Length allow all
5626 reply_header_access Content-Type allow all
5627 reply_header_access Date allow all
5628 reply_header_access Expires allow all
5629 reply_header_access Last-Modified allow all
5630 reply_header_access Location allow all
5631 reply_header_access Pragma allow all
5632 reply_header_access Content-Language allow all
5633 reply_header_access Retry-After allow all
5634 reply_header_access Title allow all
5635 reply_header_access Content-Disposition allow all
5636 reply_header_access Connection allow all
5637 reply_header_access All deny all
5639 HTTP request headers are controlled with the request_header_access directive.
5641 By default, all headers are allowed (no anonymizing is
5645 NAME: request_header_replace header_replace
5646 IFDEF: USE_HTTP_VIOLATIONS
5647 TYPE: http_header_replace
5648 LOC: Config.request_header_access
5651 Usage: request_header_replace header_name message
5652 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5654 This option allows you to change the contents of headers
5655 denied with request_header_access above, by replacing them
5656 with some fixed string.
5658 This only applies to request headers, not reply headers.
5660 By default, headers are removed if denied.
5663 NAME: reply_header_replace
5664 IFDEF: USE_HTTP_VIOLATIONS
5665 TYPE: http_header_replace
5666 LOC: Config.reply_header_access
5669 Usage: reply_header_replace header_name message
5670 Example: reply_header_replace Server Foo/1.0
5672 This option allows you to change the contents of headers
5673 denied with reply_header_access above, by replacing them
5674 with some fixed string.
5676 This only applies to reply headers, not request headers.
5678 By default, headers are removed if denied.
5681 NAME: request_header_add
5682 TYPE: HeaderWithAclList
5683 LOC: Config.request_header_add
5686 Usage: request_header_add field-name field-value acl1 [acl2] ...
5687 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5689 This option adds header fields to outgoing HTTP requests (i.e.,
5690 request headers sent by Squid to the next HTTP hop such as a
5691 cache peer or an origin server). The option has no effect during
5692 cache hit detection. The equivalent adaptation vectoring point
5693 in ICAP terminology is post-cache REQMOD.
5695 Field-name is a token specifying an HTTP header name. If a
5696 standard HTTP header name is used, Squid does not check whether
5697 the new header conflicts with any existing headers or violates
5698 HTTP rules. If the request to be modified already contains a
5699 field with the same name, the old field is preserved but the
5700 header field values are not merged.
5702 Field-value is either a token or a quoted string. If quoted
5703 string format is used, then the surrounding quotes are removed
5704 while escape sequences and %macros are processed.
5706 In theory, all of the logformat codes can be used as %macros.
5707 However, unlike logging (which happens at the very end of
5708 transaction lifetime), the transaction may not yet have enough
5709 information to expand a macro when the new header value is needed.
5710 And some information may already be available to Squid but not yet
5711 committed where the macro expansion code can access it (report
5712 such instances!). The macro will be expanded into a single dash
5713 ('-') in such cases. Not all macros have been tested.
5715 One or more Squid ACLs may be specified to restrict header
5716 injection to matching requests. As always in squid.conf, all
5717 ACLs in an option ACL list must be satisfied for the insertion
5718 to happen. The request_header_add option supports fast ACLs
5727 This option used to log custom information about the master
5728 transaction. For example, an admin may configure Squid to log
5729 which "user group" the transaction belongs to, where "user group"
5730 will be determined based on a set of ACLs and not [just]
5731 authentication information.
5732 Values of key/value pairs can be logged using %{key}note macros:
5734 note key value acl ...
5735 logformat myFormat ... %{key}note ...
5738 NAME: relaxed_header_parser
5739 COMMENT: on|off|warn
5741 LOC: Config.onoff.relaxed_header_parser
5744 In the default "on" setting Squid accepts certain forms
5745 of non-compliant HTTP messages where it is unambiguous
5746 what the sending application intended even if the message
5747 is not correctly formatted. The messages is then normalized
5748 to the correct form when forwarded by Squid.
5750 If set to "warn" then a warning will be emitted in cache.log
5751 each time such HTTP error is encountered.
5753 If set to "off" then such HTTP errors will cause the request
5754 or response to be rejected.
5757 NAME: collapsed_forwarding
5760 LOC: Config.onoff.collapsed_forwarding
5763 This option controls whether Squid is allowed to merge multiple
5764 potentially cachable requests for the same URI before Squid knows
5765 whether the response is going to be cachable.
5767 This feature is disabled by default: Enabling collapsed forwarding
5768 needlessly delays forwarding requests that look cachable (when they are
5769 collapsed) but then need to be forwarded individually anyway because
5770 they end up being for uncachable content. However, in some cases, such
5771 as accelleration of highly cachable content with periodic or groupped
5772 expiration times, the gains from collapsing [large volumes of
5773 simultenous refresh requests] outweigh losses from such delays.
5778 -----------------------------------------------------------------------------
5781 NAME: forward_timeout
5784 LOC: Config.Timeout.forward
5787 This parameter specifies how long Squid should at most attempt in
5788 finding a forwarding path for the request before giving up.
5791 NAME: connect_timeout
5794 LOC: Config.Timeout.connect
5797 This parameter specifies how long to wait for the TCP connect to
5798 the requested server or peer to complete before Squid should
5799 attempt to find another path where to forward the request.
5802 NAME: peer_connect_timeout
5805 LOC: Config.Timeout.peer_connect
5808 This parameter specifies how long to wait for a pending TCP
5809 connection to a peer cache. The default is 30 seconds. You
5810 may also set different timeout values for individual neighbors
5811 with the 'connect-timeout' option on a 'cache_peer' line.
5817 LOC: Config.Timeout.read
5820 The read_timeout is applied on server-side connections. After
5821 each successful read(), the timeout will be extended by this
5822 amount. If no data is read again after this amount of time,
5823 the request is aborted and logged with ERR_READ_TIMEOUT. The
5824 default is 15 minutes.
5830 LOC: Config.Timeout.write
5833 This timeout is tracked for all connections that have data
5834 available for writing and are waiting for the socket to become
5835 ready. After each successful write, the timeout is extended by
5836 the configured amount. If Squid has data to write but the
5837 connection is not ready for the configured duration, the
5838 transaction associated with the connection is terminated. The
5839 default is 15 minutes.
5842 NAME: request_timeout
5844 LOC: Config.Timeout.request
5847 How long to wait for complete HTTP request headers after initial
5848 connection establishment.
5851 NAME: client_idle_pconn_timeout persistent_request_timeout
5853 LOC: Config.Timeout.clientIdlePconn
5856 How long to wait for the next HTTP request on a persistent
5857 client connection after the previous request completes.
5860 NAME: client_lifetime
5863 LOC: Config.Timeout.lifetime
5866 The maximum amount of time a client (browser) is allowed to
5867 remain connected to the cache process. This protects the Cache
5868 from having a lot of sockets (and hence file descriptors) tied up
5869 in a CLOSE_WAIT state from remote clients that go away without
5870 properly shutting down (either because of a network failure or
5871 because of a poor client implementation). The default is one
5874 NOTE: The default value is intended to be much larger than any
5875 client would ever need to be connected to your cache. You
5876 should probably change client_lifetime only as a last resort.
5877 If you seem to have many client connections tying up
5878 filedescriptors, we recommend first tuning the read_timeout,
5879 request_timeout, persistent_request_timeout and quick_abort values.
5882 NAME: half_closed_clients
5884 LOC: Config.onoff.half_closed_clients
5887 Some clients may shutdown the sending side of their TCP
5888 connections, while leaving their receiving sides open. Sometimes,
5889 Squid can not tell the difference between a half-closed and a
5890 fully-closed TCP connection.
5892 By default, Squid will immediately close client connections when
5893 read(2) returns "no more data to read."
5895 Change this option to 'on' and Squid will keep open connections
5896 until a read(2) or write(2) on the socket returns an error.
5897 This may show some benefits for reverse proxies. But if not
5898 it is recommended to leave OFF.
5901 NAME: server_idle_pconn_timeout pconn_timeout
5903 LOC: Config.Timeout.serverIdlePconn
5906 Timeout for idle persistent connections to servers and other
5913 LOC: Ident::TheConfig.timeout
5916 Maximum time to wait for IDENT lookups to complete.
5918 If this is too high, and you enabled IDENT lookups from untrusted
5919 users, you might be susceptible to denial-of-service by having
5920 many ident requests going at once.
5923 NAME: shutdown_lifetime
5926 LOC: Config.shutdownLifetime
5929 When SIGTERM or SIGHUP is received, the cache is put into
5930 "shutdown pending" mode until all active sockets are closed.
5931 This value is the lifetime to set for all open descriptors
5932 during shutdown mode. Any active clients after this many
5933 seconds will receive a 'timeout' message.
5937 ADMINISTRATIVE PARAMETERS
5938 -----------------------------------------------------------------------------
5944 LOC: Config.adminEmail
5946 Email-address of local cache manager who will receive
5947 mail if the cache dies. The default is "webmaster".
5953 LOC: Config.EmailFrom
5955 From: email-address for mail sent when the cache dies.
5956 The default is to use 'squid@unique_hostname'.
5958 See also: unique_hostname directive.
5964 LOC: Config.EmailProgram
5966 Email program used to send mail if the cache dies.
5967 The default is "mail". The specified program must comply
5968 with the standard Unix mail syntax:
5969 mail-program recipient < mailfile
5971 Optional command line options can be specified.
5974 NAME: cache_effective_user
5976 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5977 LOC: Config.effectiveUser
5979 If you start Squid as root, it will change its effective/real
5980 UID/GID to the user specified below. The default is to change
5981 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5982 see also; cache_effective_group
5985 NAME: cache_effective_group
5988 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
5989 LOC: Config.effectiveGroup
5991 Squid sets the GID to the effective user's default group ID
5992 (taken from the password file) and supplementary group list
5993 from the groups membership.
5995 If you want Squid to run with a specific GID regardless of
5996 the group memberships of the effective user then set this
5997 to the group (or GID) you want Squid to run as. When set
5998 all other group privileges of the effective user are ignored
5999 and only this GID is effective. If Squid is not started as
6000 root the user starting Squid MUST be member of the specified
6003 This option is not recommended by the Squid Team.
6004 Our preference is for administrators to configure a secure
6005 user account for squid with UID/GID matching system policies.
6008 NAME: httpd_suppress_version_string
6012 LOC: Config.onoff.httpd_suppress_version_string
6014 Suppress Squid version string info in HTTP headers and HTML error pages.
6017 NAME: visible_hostname
6019 LOC: Config.visibleHostname
6021 DEFAULT_DOC: Automatically detect the system host name
6023 If you want to present a special hostname in error messages, etc,
6024 define this. Otherwise, the return value of gethostname()
6025 will be used. If you have multiple caches in a cluster and
6026 get errors about IP-forwarding you must set them to have individual
6027 names with this setting.
6030 NAME: unique_hostname
6032 LOC: Config.uniqueHostname
6034 DEFAULT_DOC: Copy the value from visible_hostname
6036 If you want to have multiple machines with the same
6037 'visible_hostname' you must give each machine a different
6038 'unique_hostname' so forwarding loops can be detected.
6041 NAME: hostname_aliases
6043 LOC: Config.hostnameAliases
6046 A list of other DNS names your cache has.
6054 Minimum umask which should be enforced while the proxy
6055 is running, in addition to the umask set at startup.
6057 For a traditional octal representation of umasks, start
6062 OPTIONS FOR THE CACHE REGISTRATION SERVICE
6063 -----------------------------------------------------------------------------
6065 This section contains parameters for the (optional) cache
6066 announcement service. This service is provided to help
6067 cache administrators locate one another in order to join or
6068 create cache hierarchies.
6070 An 'announcement' message is sent (via UDP) to the registration
6071 service by Squid. By default, the announcement message is NOT
6072 SENT unless you enable it with 'announce_period' below.
6074 The announcement message includes your hostname, plus the
6075 following information from this configuration file:
6081 All current information is processed regularly and made
6082 available on the Web at http://www.ircache.net/Cache/Tracker/.
6085 NAME: announce_period
6087 LOC: Config.Announce.period
6089 DEFAULT_DOC: Announcement messages disabled.
6091 This is how frequently to send cache announcements.
6093 To enable announcing your cache, just set an announce period.
6096 announce_period 1 day
6101 DEFAULT: tracker.ircache.net
6102 LOC: Config.Announce.host
6104 Set the hostname where announce registration messages will be sent.
6106 See also announce_port and announce_file
6112 LOC: Config.Announce.file
6114 The contents of this file will be included in the announce
6115 registration messages.
6121 LOC: Config.Announce.port
6123 Set the port where announce registration messages will be sent.
6125 See also announce_host and announce_file
6129 HTTPD-ACCELERATOR OPTIONS
6130 -----------------------------------------------------------------------------
6133 NAME: httpd_accel_surrogate_id
6136 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
6137 LOC: Config.Accel.surrogate_id
6139 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
6140 need an identification token to allow control targeting. Because
6141 a farm of surrogates may all perform the same tasks, they may share
6142 an identification token.
6145 NAME: http_accel_surrogate_remote
6149 LOC: Config.onoff.surrogate_is_remote
6151 Remote surrogates (such as those in a CDN) honour the header
6152 "Surrogate-Control: no-store-remote".
6154 Set this to on to have squid behave as a remote surrogate.
6158 IFDEF: USE_SQUID_ESI
6159 COMMENT: libxml2|expat|custom
6161 LOC: ESIParser::Type
6164 ESI markup is not strictly XML compatible. The custom ESI parser
6165 will give higher performance, but cannot handle non ASCII character
6170 DELAY POOL PARAMETERS
6171 -----------------------------------------------------------------------------
6175 TYPE: delay_pool_count
6177 IFDEF: USE_DELAY_POOLS
6180 This represents the number of delay pools to be used. For example,
6181 if you have one class 2 delay pool and one class 3 delays pool, you
6182 have a total of 2 delay pools.
6184 See also delay_parameters, delay_class, delay_access for pool
6185 configuration details.
6189 TYPE: delay_pool_class
6191 IFDEF: USE_DELAY_POOLS
6194 This defines the class of each delay pool. There must be exactly one
6195 delay_class line for each delay pool. For example, to define two
6196 delay pools, one of class 2 and one of class 3, the settings above
6200 delay_pools 4 # 4 delay pools
6201 delay_class 1 2 # pool 1 is a class 2 pool
6202 delay_class 2 3 # pool 2 is a class 3 pool
6203 delay_class 3 4 # pool 3 is a class 4 pool
6204 delay_class 4 5 # pool 4 is a class 5 pool
6206 The delay pool classes are:
6208 class 1 Everything is limited by a single aggregate
6211 class 2 Everything is limited by a single aggregate
6212 bucket as well as an "individual" bucket chosen
6213 from bits 25 through 32 of the IPv4 address.
6215 class 3 Everything is limited by a single aggregate
6216 bucket as well as a "network" bucket chosen
6217 from bits 17 through 24 of the IP address and a
6218 "individual" bucket chosen from bits 17 through
6219 32 of the IPv4 address.
6221 class 4 Everything in a class 3 delay pool, with an
6222 additional limit on a per user basis. This
6223 only takes effect if the username is established
6224 in advance - by forcing authentication in your
6227 class 5 Requests are grouped according their tag (see
6228 external_acl's tag= reply).
6231 Each pool also requires a delay_parameters directive to configure the pool size
6232 and speed limits used whenever the pool is applied to a request. Along with
6233 a set of delay_access directives to determine when it is used.
6235 NOTE: If an IP address is a.b.c.d
6236 -> bits 25 through 32 are "d"
6237 -> bits 17 through 24 are "c"
6238 -> bits 17 through 32 are "c * 256 + d"
6240 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6241 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6243 This clause only supports fast acl types.
6244 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6246 See also delay_parameters and delay_access.
6250 TYPE: delay_pool_access
6252 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6253 IFDEF: USE_DELAY_POOLS
6256 This is used to determine which delay pool a request falls into.
6258 delay_access is sorted per pool and the matching starts with pool 1,
6259 then pool 2, ..., and finally pool N. The first delay pool where the
6260 request is allowed is selected for the request. If it does not allow
6261 the request to any pool then the request is not delayed (default).
6263 For example, if you want some_big_clients in delay
6264 pool 1 and lotsa_little_clients in delay pool 2:
6266 delay_access 1 allow some_big_clients
6267 delay_access 1 deny all
6268 delay_access 2 allow lotsa_little_clients
6269 delay_access 2 deny all
6270 delay_access 3 allow authenticated_clients
6272 See also delay_parameters and delay_class.
6276 NAME: delay_parameters
6277 TYPE: delay_pool_rates
6279 IFDEF: USE_DELAY_POOLS
6282 This defines the parameters for a delay pool. Each delay pool has
6283 a number of "buckets" associated with it, as explained in the
6284 description of delay_class.
6286 For a class 1 delay pool, the syntax is:
6288 delay_parameters pool aggregate
6290 For a class 2 delay pool:
6292 delay_parameters pool aggregate individual
6294 For a class 3 delay pool:
6296 delay_parameters pool aggregate network individual
6298 For a class 4 delay pool:
6300 delay_parameters pool aggregate network individual user
6302 For a class 5 delay pool:
6304 delay_parameters pool tagrate
6306 The option variables are:
6308 pool a pool number - ie, a number between 1 and the
6309 number specified in delay_pools as used in
6312 aggregate the speed limit parameters for the aggregate bucket
6315 individual the speed limit parameters for the individual
6316 buckets (class 2, 3).
6318 network the speed limit parameters for the network buckets
6321 user the speed limit parameters for the user buckets
6324 tagrate the speed limit parameters for the tag buckets
6327 A pair of delay parameters is written restore/maximum, where restore is
6328 the number of bytes (not bits - modem and network speeds are usually
6329 quoted in bits) per second placed into the bucket, and maximum is the
6330 maximum number of bytes which can be in the bucket at any time.
6332 There must be one delay_parameters line for each delay pool.
6335 For example, if delay pool number 1 is a class 2 delay pool as in the
6336 above example, and is being used to strictly limit each host to 64Kbit/sec
6337 (plus overheads), with no overall limit, the line is:
6339 delay_parameters 1 -1/-1 8000/8000
6341 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6343 Note that the figure -1 is used to represent "unlimited".
6346 And, if delay pool number 2 is a class 3 delay pool as in the above
6347 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6348 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6349 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6350 to permit a decent web page to be downloaded at a decent speed
6351 (if the network is not being limited due to overuse) but slow down
6352 large downloads more significantly:
6354 delay_parameters 2 32000/32000 8000/8000 600/8000
6356 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6357 8 x 8000 KByte/sec -> 64Kbit/sec.
6358 8 x 600 Byte/sec -> 4800bit/sec.
6361 Finally, for a class 4 delay pool as in the example - each user will
6362 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6364 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6367 See also delay_class and delay_access.
6371 NAME: delay_initial_bucket_level
6372 COMMENT: (percent, 0-100)
6375 IFDEF: USE_DELAY_POOLS
6376 LOC: Config.Delay.initial
6378 The initial bucket percentage is used to determine how much is put
6379 in each bucket when squid starts, is reconfigured, or first notices
6380 a host accessing it (in class 2 and class 3, individual hosts and
6381 networks only have buckets associated with them once they have been
6386 CLIENT DELAY POOL PARAMETERS
6387 -----------------------------------------------------------------------------
6390 NAME: client_delay_pools
6391 TYPE: client_delay_pool_count
6393 IFDEF: USE_DELAY_POOLS
6394 LOC: Config.ClientDelay
6396 This option specifies the number of client delay pools used. It must
6397 preceed other client_delay_* options.
6400 client_delay_pools 2
6402 See also client_delay_parameters and client_delay_access.
6405 NAME: client_delay_initial_bucket_level
6406 COMMENT: (percent, 0-no_limit)
6409 IFDEF: USE_DELAY_POOLS
6410 LOC: Config.ClientDelay.initial
6412 This option determines the initial bucket size as a percentage of
6413 max_bucket_size from client_delay_parameters. Buckets are created
6414 at the time of the "first" connection from the matching IP. Idle
6415 buckets are periodically deleted up.
6417 You can specify more than 100 percent but note that such "oversized"
6418 buckets are not refilled until their size goes down to max_bucket_size
6419 from client_delay_parameters.
6422 client_delay_initial_bucket_level 50
6425 NAME: client_delay_parameters
6426 TYPE: client_delay_pool_rates
6428 IFDEF: USE_DELAY_POOLS
6429 LOC: Config.ClientDelay
6432 This option configures client-side bandwidth limits using the
6435 client_delay_parameters pool speed_limit max_bucket_size
6437 pool is an integer ID used for client_delay_access matching.
6439 speed_limit is bytes added to the bucket per second.
6441 max_bucket_size is the maximum size of a bucket, enforced after any
6442 speed_limit additions.
6444 Please see the delay_parameters option for more information and
6448 client_delay_parameters 1 1024 2048
6449 client_delay_parameters 2 51200 16384
6451 See also client_delay_access.
6455 NAME: client_delay_access
6456 TYPE: client_delay_pool_access
6458 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6459 IFDEF: USE_DELAY_POOLS
6460 LOC: Config.ClientDelay
6462 This option determines the client-side delay pool for the
6465 client_delay_access pool_ID allow|deny acl_name
6467 All client_delay_access options are checked in their pool ID
6468 order, starting with pool 1. The first checked pool with allowed
6469 request is selected for the request. If no ACL matches or there
6470 are no client_delay_access options, the request bandwidth is not
6473 The ACL-selected pool is then used to find the
6474 client_delay_parameters for the request. Client-side pools are
6475 not used to aggregate clients. Clients are always aggregated
6476 based on their source IP addresses (one bucket per source IP).
6478 This clause only supports fast acl types.
6479 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6480 Additionally, only the client TCP connection details are available.
6481 ACLs testing HTTP properties will not work.
6483 Please see delay_access for more examples.
6486 client_delay_access 1 allow low_rate_network
6487 client_delay_access 2 allow vips_network
6490 See also client_delay_parameters and client_delay_pools.
6494 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6495 -----------------------------------------------------------------------------
6500 LOC: Config.Wccp.router
6502 DEFAULT_DOC: WCCP disabled.
6505 Use this option to define your WCCP ``home'' router for
6508 wccp_router supports a single WCCP(v1) router
6510 wccp2_router supports multiple WCCPv2 routers
6512 only one of the two may be used at the same time and defines
6513 which version of WCCP to use.
6517 TYPE: IpAddress_list
6518 LOC: Config.Wccp2.router
6520 DEFAULT_DOC: WCCPv2 disabled.
6523 Use this option to define your WCCP ``home'' router for
6526 wccp_router supports a single WCCP(v1) router
6528 wccp2_router supports multiple WCCPv2 routers
6530 only one of the two may be used at the same time and defines
6531 which version of WCCP to use.
6536 LOC: Config.Wccp.version
6540 This directive is only relevant if you need to set up WCCP(v1)
6541 to some very old and end-of-life Cisco routers. In all other
6542 setups it must be left unset or at the default setting.
6543 It defines an internal version in the WCCP(v1) protocol,
6544 with version 4 being the officially documented protocol.
6546 According to some users, Cisco IOS 11.2 and earlier only
6547 support WCCP version 3. If you're using that or an earlier
6548 version of IOS, you may need to change this value to 3, otherwise
6549 do not specify this parameter.
6552 NAME: wccp2_rebuild_wait
6554 LOC: Config.Wccp2.rebuildwait
6558 If this is enabled Squid will wait for the cache dir rebuild to finish
6559 before sending the first wccp2 HereIAm packet
6562 NAME: wccp2_forwarding_method
6564 LOC: Config.Wccp2.forwarding_method
6568 WCCP2 allows the setting of forwarding methods between the
6569 router/switch and the cache. Valid values are as follows:
6571 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6572 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6574 Currently (as of IOS 12.4) cisco routers only support GRE.
6575 Cisco switches only support the L2 redirect assignment method.
6578 NAME: wccp2_return_method
6580 LOC: Config.Wccp2.return_method
6584 WCCP2 allows the setting of return methods between the
6585 router/switch and the cache for packets that the cache
6586 decides not to handle. Valid values are as follows:
6588 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6589 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6591 Currently (as of IOS 12.4) cisco routers only support GRE.
6592 Cisco switches only support the L2 redirect assignment.
6594 If the "ip wccp redirect exclude in" command has been
6595 enabled on the cache interface, then it is still safe for
6596 the proxy server to use a l2 redirect method even if this
6597 option is set to GRE.
6600 NAME: wccp2_assignment_method
6602 LOC: Config.Wccp2.assignment_method
6606 WCCP2 allows the setting of methods to assign the WCCP hash
6607 Valid values are as follows:
6609 hash - Hash assignment
6610 mask - Mask assignment
6612 As a general rule, cisco routers support the hash assignment method
6613 and cisco switches support the mask assignment method.
6618 LOC: Config.Wccp2.info
6619 DEFAULT_IF_NONE: standard 0
6620 DEFAULT_DOC: Use the 'web-cache' standard service.
6623 WCCP2 allows for multiple traffic services. There are two
6624 types: "standard" and "dynamic". The standard type defines
6625 one service id - http (id 0). The dynamic service ids can be from
6626 51 to 255 inclusive. In order to use a dynamic service id
6627 one must define the type of traffic to be redirected; this is done
6628 using the wccp2_service_info option.
6630 The "standard" type does not require a wccp2_service_info option,
6631 just specifying the service id will suffice.
6633 MD5 service authentication can be enabled by adding
6634 "password=<password>" to the end of this service declaration.
6638 wccp2_service standard 0 # for the 'web-cache' standard service
6639 wccp2_service dynamic 80 # a dynamic service type which will be
6640 # fleshed out with subsequent options.
6641 wccp2_service standard 0 password=foo
6644 NAME: wccp2_service_info
6645 TYPE: wccp2_service_info
6646 LOC: Config.Wccp2.info
6650 Dynamic WCCPv2 services require further information to define the
6651 traffic you wish to have diverted.
6655 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6656 priority=<priority> ports=<port>,<port>..
6658 The relevant WCCPv2 flags:
6659 + src_ip_hash, dst_ip_hash
6660 + source_port_hash, dst_port_hash
6661 + src_ip_alt_hash, dst_ip_alt_hash
6662 + src_port_alt_hash, dst_port_alt_hash
6665 The port list can be one to eight entries.
6669 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6670 priority=240 ports=80
6672 Note: the service id must have been defined by a previous
6673 'wccp2_service dynamic <id>' entry.
6678 LOC: Config.Wccp2.weight
6682 Each cache server gets assigned a set of the destination
6683 hash proportional to their weight.
6688 LOC: Config.Wccp.address
6690 DEFAULT_DOC: Address selected by the operating system.
6693 Use this option if you require WCCPv2 to use a specific
6696 The default behavior is to not bind to any specific address.
6701 LOC: Config.Wccp2.address
6703 DEFAULT_DOC: Address selected by the operating system.
6706 Use this option if you require WCCP to use a specific
6709 The default behavior is to not bind to any specific address.
6713 PERSISTENT CONNECTION HANDLING
6714 -----------------------------------------------------------------------------
6716 Also see "pconn_timeout" in the TIMEOUTS section
6719 NAME: client_persistent_connections
6721 LOC: Config.onoff.client_pconns
6724 Persistent connection support for clients.
6725 Squid uses persistent connections (when allowed). You can use
6726 this option to disable persistent connections with clients.
6729 NAME: server_persistent_connections
6731 LOC: Config.onoff.server_pconns
6734 Persistent connection support for servers.
6735 Squid uses persistent connections (when allowed). You can use
6736 this option to disable persistent connections with servers.
6739 NAME: persistent_connection_after_error
6741 LOC: Config.onoff.error_pconns
6744 With this directive the use of persistent connections after
6745 HTTP errors can be disabled. Useful if you have clients
6746 who fail to handle errors on persistent connections proper.
6749 NAME: detect_broken_pconn
6751 LOC: Config.onoff.detect_broken_server_pconns
6754 Some servers have been found to incorrectly signal the use
6755 of HTTP/1.0 persistent connections even on replies not
6756 compatible, causing significant delays. This server problem
6757 has mostly been seen on redirects.
6759 By enabling this directive Squid attempts to detect such
6760 broken replies and automatically assume the reply is finished
6761 after 10 seconds timeout.
6765 CACHE DIGEST OPTIONS
6766 -----------------------------------------------------------------------------
6769 NAME: digest_generation
6770 IFDEF: USE_CACHE_DIGESTS
6772 LOC: Config.onoff.digest_generation
6775 This controls whether the server will generate a Cache Digest
6776 of its contents. By default, Cache Digest generation is
6777 enabled if Squid is compiled with --enable-cache-digests defined.
6780 NAME: digest_bits_per_entry
6781 IFDEF: USE_CACHE_DIGESTS
6783 LOC: Config.digest.bits_per_entry
6786 This is the number of bits of the server's Cache Digest which
6787 will be associated with the Digest entry for a given HTTP
6788 Method and URL (public key) combination. The default is 5.
6791 NAME: digest_rebuild_period
6792 IFDEF: USE_CACHE_DIGESTS
6795 LOC: Config.digest.rebuild_period
6798 This is the wait time between Cache Digest rebuilds.
6801 NAME: digest_rewrite_period
6803 IFDEF: USE_CACHE_DIGESTS
6805 LOC: Config.digest.rewrite_period
6808 This is the wait time between Cache Digest writes to
6812 NAME: digest_swapout_chunk_size
6815 IFDEF: USE_CACHE_DIGESTS
6816 LOC: Config.digest.swapout_chunk_size
6819 This is the number of bytes of the Cache Digest to write to
6820 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6824 NAME: digest_rebuild_chunk_percentage
6825 COMMENT: (percent, 0-100)
6826 IFDEF: USE_CACHE_DIGESTS
6828 LOC: Config.digest.rebuild_chunk_percentage
6831 This is the percentage of the Cache Digest to be scanned at a
6832 time. By default it is set to 10% of the Cache Digest.
6837 -----------------------------------------------------------------------------
6842 LOC: Config.Port.snmp
6844 DEFAULT_DOC: SNMP disabled.
6847 The port number where Squid listens for SNMP requests. To enable
6848 SNMP support set this to a suitable port number. Port number
6849 3401 is often used for the Squid SNMP agent. By default it's
6850 set to "0" (disabled)
6858 LOC: Config.accessList.snmp
6860 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
6863 Allowing or denying access to the SNMP port.
6865 All access to the agent is denied by default.
6868 snmp_access allow|deny [!]aclname ...
6870 This clause only supports fast acl types.
6871 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6874 snmp_access allow snmppublic localhost
6875 snmp_access deny all
6878 NAME: snmp_incoming_address
6880 LOC: Config.Addrs.snmp_incoming
6882 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
6885 Just like 'udp_incoming_address', but for the SNMP port.
6887 snmp_incoming_address is used for the SNMP socket receiving
6888 messages from SNMP agents.
6890 The default snmp_incoming_address is to listen on all
6891 available network interfaces.
6894 NAME: snmp_outgoing_address
6896 LOC: Config.Addrs.snmp_outgoing
6898 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
6901 Just like 'udp_outgoing_address', but for the SNMP port.
6903 snmp_outgoing_address is used for SNMP packets returned to SNMP
6906 If snmp_outgoing_address is not set it will use the same socket
6907 as snmp_incoming_address. Only change this if you want to have
6908 SNMP replies sent using another address than where this Squid
6909 listens for SNMP queries.
6911 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6912 the same value since they both use the same port.
6917 -----------------------------------------------------------------------------
6920 NAME: icp_port udp_port
6923 DEFAULT_DOC: ICP disabled.
6924 LOC: Config.Port.icp
6926 The port number where Squid sends and receives ICP queries to
6927 and from neighbor caches. The standard UDP port for ICP is 3130.
6930 icp_port @DEFAULT_ICP_PORT@
6937 DEFAULT_DOC: HTCP disabled.
6938 LOC: Config.Port.htcp
6940 The port number where Squid sends and receives HTCP queries to
6941 and from neighbor caches. To turn it on you want to set it to
6948 NAME: log_icp_queries
6952 LOC: Config.onoff.log_udp
6954 If set, ICP queries are logged to access.log. You may wish
6955 do disable this if your ICP load is VERY high to speed things
6956 up or to simplify log analysis.
6959 NAME: udp_incoming_address
6961 LOC:Config.Addrs.udp_incoming
6963 DEFAULT_DOC: Accept packets from all machine interfaces.
6965 udp_incoming_address is used for UDP packets received from other
6968 The default behavior is to not bind to any specific address.
6970 Only change this if you want to have all UDP queries received on
6971 a specific interface/address.
6973 NOTE: udp_incoming_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_outgoing_address
6978 NOTE, udp_incoming_address and udp_outgoing_address can not
6979 have the same value since they both use the same port.
6982 NAME: udp_outgoing_address
6984 LOC: Config.Addrs.udp_outgoing
6986 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
6988 udp_outgoing_address is used for UDP packets sent out to other
6991 The default behavior is to not bind to any specific address.
6993 Instead it will use the same socket as udp_incoming_address.
6994 Only change this if you want to have UDP queries sent using another
6995 address than where this Squid listens for UDP queries from other
6998 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6999 modules. Altering it will affect all of them in the same manner.
7001 see also; udp_incoming_address
7003 NOTE, udp_incoming_address and udp_outgoing_address can not
7004 have the same value since they both use the same port.
7011 LOC: Config.onoff.icp_hit_stale
7013 If you want to return ICP_HIT for stale cache objects, set this
7014 option to 'on'. If you have sibling relationships with caches
7015 in other administrative domains, this should be 'off'. If you only
7016 have sibling relationships with caches under your control,
7017 it is probably okay to set this to 'on'.
7018 If set to 'on', your siblings should use the option "allow-miss"
7019 on their cache_peer lines for connecting to you.
7022 NAME: minimum_direct_hops
7025 LOC: Config.minDirectHops
7027 If using the ICMP pinging stuff, do direct fetches for sites
7028 which are no more than this many hops away.
7031 NAME: minimum_direct_rtt
7035 LOC: Config.minDirectRtt
7037 If using the ICMP pinging stuff, do direct fetches for sites
7038 which are no more than this many rtt milliseconds away.
7044 LOC: Config.Netdb.low
7046 The low water mark for the ICMP measurement database.
7048 Note: high watermark controlled by netdb_high directive.
7050 These watermarks are counts, not percents. The defaults are
7051 (low) 900 and (high) 1000. When the high water mark is
7052 reached, database entries will be deleted until the low
7059 LOC: Config.Netdb.high
7061 The high water mark for the ICMP measurement database.
7063 Note: low watermark controlled by netdb_low directive.
7065 These watermarks are counts, not percents. The defaults are
7066 (low) 900 and (high) 1000. When the high water mark is
7067 reached, database entries will be deleted until the low
7071 NAME: netdb_ping_period
7073 LOC: Config.Netdb.period
7076 The minimum period for measuring a site. There will be at
7077 least this much delay between successive pings to the same
7078 network. The default is five minutes.
7085 LOC: Config.onoff.query_icmp
7087 If you want to ask your peers to include ICMP data in their ICP
7088 replies, enable this option.
7090 If your peer has configured Squid (during compilation) with
7091 '--enable-icmp' that peer will send ICMP pings to origin server
7092 sites of the URLs it receives. If you enable this option the
7093 ICP replies from that peer will include the ICMP data (if available).
7094 Then, when choosing a parent cache, Squid will choose the parent with
7095 the minimal RTT to the origin server. When this happens, the
7096 hierarchy field of the access.log will be
7097 "CLOSEST_PARENT_MISS". This option is off by default.
7100 NAME: test_reachability
7104 LOC: Config.onoff.test_reachability
7106 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
7107 instead of ICP_MISS if the target host is NOT in the ICMP
7108 database, or has a zero RTT.
7111 NAME: icp_query_timeout
7114 DEFAULT_DOC: Dynamic detection.
7116 LOC: Config.Timeout.icp_query
7118 Normally Squid will automatically determine an optimal ICP
7119 query timeout value based on the round-trip-time of recent ICP
7120 queries. If you want to override the value determined by
7121 Squid, set this 'icp_query_timeout' to a non-zero value. This
7122 value is specified in MILLISECONDS, so, to use a 2-second
7123 timeout (the old default), you would write:
7125 icp_query_timeout 2000
7128 NAME: maximum_icp_query_timeout
7132 LOC: Config.Timeout.icp_query_max
7134 Normally the ICP query timeout is determined dynamically. But
7135 sometimes it can lead to very large values (say 5 seconds).
7136 Use this option to put an upper limit on the dynamic timeout
7137 value. Do NOT use this option to always use a fixed (instead
7138 of a dynamic) timeout value. To set a fixed timeout see the
7139 'icp_query_timeout' directive.
7142 NAME: minimum_icp_query_timeout
7146 LOC: Config.Timeout.icp_query_min
7148 Normally the ICP query timeout is determined dynamically. But
7149 sometimes it can lead to very small timeouts, even lower than
7150 the normal latency variance on your link due to traffic.
7151 Use this option to put an lower limit on the dynamic timeout
7152 value. Do NOT use this option to always use a fixed (instead
7153 of a dynamic) timeout value. To set a fixed timeout see the
7154 'icp_query_timeout' directive.
7157 NAME: background_ping_rate
7161 LOC: Config.backgroundPingRate
7163 Controls how often the ICP pings are sent to siblings that
7164 have background-ping set.
7168 MULTICAST ICP OPTIONS
7169 -----------------------------------------------------------------------------
7174 LOC: Config.mcast_group_list
7177 This tag specifies a list of multicast groups which your server
7178 should join to receive multicasted ICP queries.
7180 NOTE! Be very careful what you put here! Be sure you
7181 understand the difference between an ICP _query_ and an ICP
7182 _reply_. This option is to be set only if you want to RECEIVE
7183 multicast queries. Do NOT set this option to SEND multicast
7184 ICP (use cache_peer for that). ICP replies are always sent via
7185 unicast, so this option does not affect whether or not you will
7186 receive replies from multicast group members.
7188 You must be very careful to NOT use a multicast address which
7189 is already in use by another group of caches.
7191 If you are unsure about multicast, please read the Multicast
7192 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
7194 Usage: mcast_groups 239.128.16.128 224.0.1.20
7196 By default, Squid doesn't listen on any multicast groups.
7199 NAME: mcast_miss_addr
7200 IFDEF: MULTICAST_MISS_STREAM
7202 LOC: Config.mcast_miss.addr
7204 DEFAULT_DOC: disabled.
7206 If you enable this option, every "cache miss" URL will
7207 be sent out on the specified multicast address.
7209 Do not enable this option unless you are are absolutely
7210 certain you understand what you are doing.
7213 NAME: mcast_miss_ttl
7214 IFDEF: MULTICAST_MISS_STREAM
7216 LOC: Config.mcast_miss.ttl
7219 This is the time-to-live value for packets multicasted
7220 when multicasting off cache miss URLs is enabled. By
7221 default this is set to 'site scope', i.e. 16.
7224 NAME: mcast_miss_port
7225 IFDEF: MULTICAST_MISS_STREAM
7227 LOC: Config.mcast_miss.port
7230 This is the port number to be used in conjunction with
7234 NAME: mcast_miss_encode_key
7235 IFDEF: MULTICAST_MISS_STREAM
7237 LOC: Config.mcast_miss.encode_key
7238 DEFAULT: XXXXXXXXXXXXXXXX
7240 The URLs that are sent in the multicast miss stream are
7241 encrypted. This is the encryption key.
7244 NAME: mcast_icp_query_timeout
7248 LOC: Config.Timeout.mcast_icp_query
7250 For multicast peers, Squid regularly sends out ICP "probes" to
7251 count how many other peers are listening on the given multicast
7252 address. This value specifies how long Squid should wait to
7253 count all the replies. The default is 2000 msec, or 2
7258 INTERNAL ICON OPTIONS
7259 -----------------------------------------------------------------------------
7262 NAME: icon_directory
7264 LOC: Config.icons.directory
7265 DEFAULT: @DEFAULT_ICON_DIR@
7267 Where the icons are stored. These are normally kept in
7271 NAME: global_internal_static
7273 LOC: Config.onoff.global_internal_static
7276 This directive controls is Squid should intercept all requests for
7277 /squid-internal-static/ no matter which host the URL is requesting
7278 (default on setting), or if nothing special should be done for
7279 such URLs (off setting). The purpose of this directive is to make
7280 icons etc work better in complex cache hierarchies where it may
7281 not always be possible for all corners in the cache mesh to reach
7282 the server generating a directory listing.
7285 NAME: short_icon_urls
7287 LOC: Config.icons.use_short_names
7290 If this is enabled Squid will use short URLs for icons.
7291 If disabled it will revert to the old behavior of including
7292 it's own name and port in the URL.
7294 If you run a complex cache hierarchy with a mix of Squid and
7295 other proxies you may need to disable this directive.
7300 -----------------------------------------------------------------------------
7303 NAME: error_directory
7305 LOC: Config.errorDirectory
7307 DEFAULT_DOC: Send error pages in the clients preferred language
7309 If you wish to create your own versions of the default
7310 error files to customize them to suit your company copy
7311 the error/template files to another directory and point
7314 WARNING: This option will disable multi-language support
7315 on error pages if used.
7317 The squid developers are interested in making squid available in
7318 a wide variety of languages. If you are making translations for a
7319 language that Squid does not currently provide please consider
7320 contributing your translation back to the project.
7321 http://wiki.squid-cache.org/Translations
7323 The squid developers working on translations are happy to supply drop-in
7324 translated error files in exchange for any new language contributions.
7327 NAME: error_default_language
7328 IFDEF: USE_ERR_LOCALES
7330 LOC: Config.errorDefaultLanguage
7332 DEFAULT_DOC: Generate English language pages.
7334 Set the default language which squid will send error pages in
7335 if no existing translation matches the clients language
7338 If unset (default) generic English will be used.
7340 The squid developers are interested in making squid available in
7341 a wide variety of languages. If you are interested in making
7342 translations for any language see the squid wiki for details.
7343 http://wiki.squid-cache.org/Translations
7346 NAME: error_log_languages
7347 IFDEF: USE_ERR_LOCALES
7349 LOC: Config.errorLogMissingLanguages
7352 Log to cache.log what languages users are attempting to
7353 auto-negotiate for translations.
7355 Successful negotiations are not logged. Only failures
7356 have meaning to indicate that Squid may need an upgrade
7357 of its error page translations.
7360 NAME: err_page_stylesheet
7362 LOC: Config.errorStylesheet
7363 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7365 CSS Stylesheet to pattern the display of Squid default error pages.
7367 For information on CSS see http://www.w3.org/Style/CSS/
7372 LOC: Config.errHtmlText
7375 HTML text to include in error messages. Make this a "mailto"
7376 URL to your admin address, or maybe just a link to your
7377 organizations Web page.
7379 To include this in your error messages, you must rewrite
7380 the error template files (found in the "errors" directory).
7381 Wherever you want the 'err_html_text' line to appear,
7382 insert a %L tag in the error template file.
7385 NAME: email_err_data
7388 LOC: Config.onoff.emailErrData
7391 If enabled, information about the occurred error will be
7392 included in the mailto links of the ERR pages (if %W is set)
7393 so that the email body contains the data.
7394 Syntax is <A HREF="mailto:%w%W">%w</A>
7399 LOC: Config.denyInfoList
7402 Usage: deny_info err_page_name acl
7403 or deny_info http://... acl
7404 or deny_info TCP_RESET acl
7406 This can be used to return a ERR_ page for requests which
7407 do not pass the 'http_access' rules. Squid remembers the last
7408 acl it evaluated in http_access, and if a 'deny_info' line exists
7409 for that ACL Squid returns a corresponding error page.
7411 The acl is typically the last acl on the http_access deny line which
7412 denied access. The exceptions to this rule are:
7413 - When Squid needs to request authentication credentials. It's then
7414 the first authentication related acl encountered
7415 - When none of the http_access lines matches. It's then the last
7416 acl processed on the last http_access line.
7417 - When the decision to deny access was made by an adaptation service,
7418 the acl name is the corresponding eCAP or ICAP service_name.
7420 NP: If providing your own custom error pages with error_directory
7421 you may also specify them by your custom file name:
7422 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7424 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7425 may be specified by prefixing the file name with the code and a colon.
7426 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7428 Alternatively you can tell Squid to reset the TCP connection
7429 by specifying TCP_RESET.
7431 Or you can specify an error URL or URL pattern. The browsers will
7432 get redirected to the specified URL after formatting tags have
7433 been replaced. Redirect will be done with 302 or 307 according to
7434 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7435 the URL. e.g. 303:http://example.com/
7438 %a - username (if available. Password NOT included)
7441 %E - Error description
7443 %H - Request domain name
7444 %i - Client IP Address
7446 %o - Message result from external ACL helper
7447 %p - Request Port number
7448 %P - Request Protocol name
7449 %R - Request URL path
7450 %T - Timestamp in RFC 1123 format
7451 %U - Full canonical URL from client
7452 (HTTPS URLs terminate with *)
7453 %u - Full canonical URL from client
7454 %w - Admin email from squid.conf
7456 %% - Literal percent (%) code
7461 OPTIONS INFLUENCING REQUEST FORWARDING
7462 -----------------------------------------------------------------------------
7465 NAME: nonhierarchical_direct
7467 LOC: Config.onoff.nonhierarchical_direct
7470 By default, Squid will send any non-hierarchical requests
7471 (not cacheable request type) direct to origin servers.
7473 When this is set to "off", Squid will prefer to send these
7474 requests to parents.
7476 Note that in most configurations, by turning this off you will only
7477 add latency to these request without any improvement in global hit
7480 This option only sets a preference. If the parent is unavailable a
7481 direct connection to the origin server may still be attempted. To
7482 completely prevent direct connections use never_direct.
7487 LOC: Config.onoff.prefer_direct
7490 Normally Squid tries to use parents for most requests. If you for some
7491 reason like it to first try going direct and only use a parent if
7492 going direct fails set this to on.
7494 By combining nonhierarchical_direct off and prefer_direct on you
7495 can set up Squid to use a parent as a backup path if going direct
7498 Note: If you want Squid to use parents for all requests see
7499 the never_direct directive. prefer_direct only modifies how Squid
7500 acts on cacheable requests.
7503 NAME: cache_miss_revalidate
7507 LOC: Config.onoff.cache_miss_revalidate
7509 RFC 7232 defines a conditional request mechanism to prevent
7510 response objects being unnecessarily transferred over the network.
7511 If that mechanism is used by the client and a cache MISS occurs
7512 it can prevent new cache entries being created.
7514 This option determines whether Squid on cache MISS will pass the
7515 client revalidation request to the server or tries to fetch new
7516 content for caching. It can be useful while the cache is mostly
7517 empty to more quickly have the cache populated by generating
7518 non-conditional GETs.
7520 When set to 'on' (default), Squid will pass all client If-* headers
7521 to the server. This permits server responses without a cacheable
7522 payload to be delivered and on MISS no new cache entry is created.
7524 When set to 'off' and if the request is cacheable, Squid will
7525 remove the clients If-Modified-Since and If-None-Match headers from
7526 the request sent to the server. This requests a 200 status response
7527 from the server to create a new cache entry with.
7532 LOC: Config.accessList.AlwaysDirect
7534 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7536 Usage: always_direct allow|deny [!]aclname ...
7538 Here you can use ACL elements to specify requests which should
7539 ALWAYS be forwarded by Squid to the origin servers without using
7540 any peers. For example, to always directly forward requests for
7541 local servers ignoring any parents or siblings you may have use
7544 acl local-servers dstdomain my.domain.net
7545 always_direct allow local-servers
7547 To always forward FTP requests directly, use
7550 always_direct allow FTP
7552 NOTE: There is a similar, but opposite option named
7553 'never_direct'. You need to be aware that "always_direct deny
7554 foo" is NOT the same thing as "never_direct allow foo". You
7555 may need to use a deny rule to exclude a more-specific case of
7556 some other rule. Example:
7558 acl local-external dstdomain external.foo.net
7559 acl local-servers dstdomain .foo.net
7560 always_direct deny local-external
7561 always_direct allow local-servers
7563 NOTE: If your goal is to make the client forward the request
7564 directly to the origin server bypassing Squid then this needs
7565 to be done in the client configuration. Squid configuration
7566 can only tell Squid how Squid should fetch the object.
7568 NOTE: This directive is not related to caching. The replies
7569 is cached as usual even if you use always_direct. To not cache
7570 the replies see the 'cache' directive.
7572 This clause supports both fast and slow acl types.
7573 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7578 LOC: Config.accessList.NeverDirect
7580 DEFAULT_DOC: Allow DNS results to be used for this request.
7582 Usage: never_direct allow|deny [!]aclname ...
7584 never_direct is the opposite of always_direct. Please read
7585 the description for always_direct if you have not already.
7587 With 'never_direct' you can use ACL elements to specify
7588 requests which should NEVER be forwarded directly to origin
7589 servers. For example, to force the use of a proxy for all
7590 requests, except those in your local domain use something like:
7592 acl local-servers dstdomain .foo.net
7593 never_direct deny local-servers
7594 never_direct allow all
7596 or if Squid is inside a firewall and there are local intranet
7597 servers inside the firewall use something like:
7599 acl local-intranet dstdomain .foo.net
7600 acl local-external dstdomain external.foo.net
7601 always_direct deny local-external
7602 always_direct allow local-intranet
7603 never_direct allow all
7605 This clause supports both fast and slow acl types.
7606 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7610 ADVANCED NETWORKING OPTIONS
7611 -----------------------------------------------------------------------------
7614 NAME: incoming_udp_average incoming_icp_average
7617 LOC: Config.comm_incoming.udp.average
7619 Heavy voodoo here. I can't even believe you are reading this.
7620 Are you crazy? Don't even think about adjusting these unless
7621 you understand the algorithms in comm_select.c first!
7624 NAME: incoming_tcp_average incoming_http_average
7627 LOC: Config.comm_incoming.tcp.average
7629 Heavy voodoo here. I can't even believe you are reading this.
7630 Are you crazy? Don't even think about adjusting these unless
7631 you understand the algorithms in comm_select.c first!
7634 NAME: incoming_dns_average
7637 LOC: Config.comm_incoming.dns.average
7639 Heavy voodoo here. I can't even believe you are reading this.
7640 Are you crazy? Don't even think about adjusting these unless
7641 you understand the algorithms in comm_select.c first!
7644 NAME: min_udp_poll_cnt min_icp_poll_cnt
7647 LOC: Config.comm_incoming.udp.min_poll
7649 Heavy voodoo here. I can't even believe you are reading this.
7650 Are you crazy? Don't even think about adjusting these unless
7651 you understand the algorithms in comm_select.c first!
7654 NAME: min_dns_poll_cnt
7657 LOC: Config.comm_incoming.dns.min_poll
7659 Heavy voodoo here. I can't even believe you are reading this.
7660 Are you crazy? Don't even think about adjusting these unless
7661 you understand the algorithms in comm_select.c first!
7664 NAME: min_tcp_poll_cnt min_http_poll_cnt
7667 LOC: Config.comm_incoming.tcp.min_poll
7669 Heavy voodoo here. I can't even believe you are reading this.
7670 Are you crazy? Don't even think about adjusting these unless
7671 you understand the algorithms in comm_select.c first!
7677 LOC: Config.accept_filter
7681 The name of an accept(2) filter to install on Squid's
7682 listen socket(s). This feature is perhaps specific to
7683 FreeBSD and requires support in the kernel.
7685 The 'httpready' filter delays delivering new connections
7686 to Squid until a full HTTP request has been received.
7687 See the accf_http(9) man page for details.
7689 The 'dataready' filter delays delivering new connections
7690 to Squid until there is some data to process.
7691 See the accf_dataready(9) man page for details.
7695 The 'data' filter delays delivering of new connections
7696 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
7697 You may optionally specify a number of seconds to wait by
7698 'data=N' where N is the number of seconds. Defaults to 30
7699 if not specified. See the tcp(7) man page for details.
7702 accept_filter httpready
7707 NAME: client_ip_max_connections
7709 LOC: Config.client_ip_max_connections
7711 DEFAULT_DOC: No limit.
7713 Set an absolute limit on the number of connections a single
7714 client IP can use. Any more than this and Squid will begin to drop
7715 new connections from the client until it closes some links.
7717 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
7718 connections from the client. For finer control use the ACL access controls.
7720 Requires client_db to be enabled (the default).
7722 WARNING: This may noticably slow down traffic received via external proxies
7723 or NAT devices and cause them to rebound error messages back to their clients.
7726 NAME: tcp_recv_bufsize
7730 DEFAULT_DOC: Use operating system TCP defaults.
7731 LOC: Config.tcpRcvBufsz
7733 Size of receive buffer to set for TCP sockets. Probably just
7734 as easy to change your kernel's default.
7735 Omit from squid.conf to use the default buffer size.
7740 -----------------------------------------------------------------------------
7747 LOC: Adaptation::Icap::TheConfig.onoff
7750 If you want to enable the ICAP module support, set this to on.
7753 NAME: icap_connect_timeout
7756 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7759 This parameter specifies how long to wait for the TCP connect to
7760 the requested ICAP server to complete before giving up and either
7761 terminating the HTTP transaction or bypassing the failure.
7763 The default for optional services is peer_connect_timeout.
7764 The default for essential services is connect_timeout.
7765 If this option is explicitly set, its value applies to all services.
7768 NAME: icap_io_timeout
7772 DEFAULT_DOC: Use read_timeout.
7773 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7776 This parameter specifies how long to wait for an I/O activity on
7777 an established, active ICAP connection before giving up and
7778 either terminating the HTTP transaction or bypassing the
7782 NAME: icap_service_failure_limit
7783 COMMENT: limit [in memory-depth time-units]
7784 TYPE: icap_service_failure_limit
7786 LOC: Adaptation::Icap::TheConfig
7789 The limit specifies the number of failures that Squid tolerates
7790 when establishing a new TCP connection with an ICAP service. If
7791 the number of failures exceeds the limit, the ICAP service is
7792 not used for new ICAP requests until it is time to refresh its
7795 A negative value disables the limit. Without the limit, an ICAP
7796 service will not be considered down due to connectivity failures
7797 between ICAP OPTIONS requests.
7799 Squid forgets ICAP service failures older than the specified
7800 value of memory-depth. The memory fading algorithm
7801 is approximate because Squid does not remember individual
7802 errors but groups them instead, splitting the option
7803 value into ten time slots of equal length.
7805 When memory-depth is 0 and by default this option has no
7806 effect on service failure expiration.
7808 Squid always forgets failures when updating service settings
7809 using an ICAP OPTIONS transaction, regardless of this option
7813 # suspend service usage after 10 failures in 5 seconds:
7814 icap_service_failure_limit 10 in 5 seconds
7817 NAME: icap_service_revival_delay
7820 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7823 The delay specifies the number of seconds to wait after an ICAP
7824 OPTIONS request failure before requesting the options again. The
7825 failed ICAP service is considered "down" until fresh OPTIONS are
7828 The actual delay cannot be smaller than the hardcoded minimum
7829 delay of 30 seconds.
7832 NAME: icap_preview_enable
7836 LOC: Adaptation::Icap::TheConfig.preview_enable
7839 The ICAP Preview feature allows the ICAP server to handle the
7840 HTTP message by looking only at the beginning of the message body
7841 or even without receiving the body at all. In some environments,
7842 previews greatly speedup ICAP processing.
7844 During an ICAP OPTIONS transaction, the server may tell Squid what
7845 HTTP messages should be previewed and how big the preview should be.
7846 Squid will not use Preview if the server did not request one.
7848 To disable ICAP Preview for all ICAP services, regardless of
7849 individual ICAP server OPTIONS responses, set this option to "off".
7851 icap_preview_enable off
7854 NAME: icap_preview_size
7857 LOC: Adaptation::Icap::TheConfig.preview_size
7859 DEFAULT_DOC: No preview sent.
7861 The default size of preview data to be sent to the ICAP server.
7862 This value might be overwritten on a per server basis by OPTIONS requests.
7865 NAME: icap_206_enable
7869 LOC: Adaptation::Icap::TheConfig.allow206_enable
7872 206 (Partial Content) responses is an ICAP extension that allows the
7873 ICAP agents to optionally combine adapted and original HTTP message
7874 content. The decision to combine is postponed until the end of the
7875 ICAP response. Squid supports Partial Content extension by default.
7877 Activation of the Partial Content extension is negotiated with each
7878 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7879 negotation correctly even if they do not support the extension, but
7880 some might fail. To disable Partial Content support for all ICAP
7881 services and to avoid any negotiation, set this option to "off".
7887 NAME: icap_default_options_ttl
7890 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7893 The default TTL value for ICAP OPTIONS responses that don't have
7894 an Options-TTL header.
7897 NAME: icap_persistent_connections
7901 LOC: Adaptation::Icap::TheConfig.reuse_connections
7904 Whether or not Squid should use persistent connections to
7908 NAME: adaptation_send_client_ip icap_send_client_ip
7910 IFDEF: USE_ADAPTATION
7912 LOC: Adaptation::Config::send_client_ip
7915 If enabled, Squid shares HTTP client IP information with adaptation
7916 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7917 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7919 See also: adaptation_uses_indirect_client
7922 NAME: adaptation_send_username icap_send_client_username
7924 IFDEF: USE_ADAPTATION
7926 LOC: Adaptation::Config::send_username
7929 This sends authenticated HTTP client username (if available) to
7930 the adaptation service.
7932 For ICAP, the username value is encoded based on the
7933 icap_client_username_encode option and is sent using the header
7934 specified by the icap_client_username_header option.
7937 NAME: icap_client_username_header
7940 LOC: Adaptation::Icap::TheConfig.client_username_header
7941 DEFAULT: X-Client-Username
7943 ICAP request header name to use for adaptation_send_username.
7946 NAME: icap_client_username_encode
7950 LOC: Adaptation::Icap::TheConfig.client_username_encode
7953 Whether to base64 encode the authenticated client username.
7957 TYPE: icap_service_type
7959 LOC: Adaptation::Icap::TheConfig
7962 Defines a single ICAP service using the following format:
7964 icap_service id vectoring_point uri [option ...]
7967 an opaque identifier or name which is used to direct traffic to
7968 this specific service. Must be unique among all adaptation
7969 services in squid.conf.
7971 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7972 This specifies at which point of transaction processing the
7973 ICAP service should be activated. *_postcache vectoring points
7974 are not yet supported.
7976 uri: icap://servername:port/servicepath
7977 ICAP server and service location.
7979 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7980 transactions. Squid does not enforce that requirement. You can specify
7981 services with the same service_url and different vectoring_points. You
7982 can even specify multiple identical services as long as their
7983 service_names differ.
7985 To activate a service, use the adaptation_access directive. To group
7986 services, use adaptation_service_chain and adaptation_service_set.
7988 Service options are separated by white space. ICAP services support
7989 the following name=value options:
7992 If set to 'on' or '1', the ICAP service is treated as
7993 optional. If the service cannot be reached or malfunctions,
7994 Squid will try to ignore any errors and process the message as
7995 if the service was not enabled. No all ICAP errors can be
7996 bypassed. If set to 0, the ICAP service is treated as
7997 essential and all ICAP errors will result in an error page
7998 returned to the HTTP client.
8000 Bypass is off by default: services are treated as essential.
8003 If set to 'on' or '1', the ICAP service is allowed to
8004 dynamically change the current message adaptation plan by
8005 returning a chain of services to be used next. The services
8006 are specified using the X-Next-Services ICAP response header
8007 value, formatted as a comma-separated list of service names.
8008 Each named service should be configured in squid.conf. Other
8009 services are ignored. An empty X-Next-Services value results
8010 in an empty plan which ends the current adaptation.
8012 Dynamic adaptation plan may cross or cover multiple supported
8013 vectoring points in their natural processing order.
8015 Routing is not allowed by default: the ICAP X-Next-Services
8016 response header is ignored.
8019 Only has effect on split-stack systems. The default on those systems
8020 is to use IPv4-only connections. When set to 'on' this option will
8021 make Squid use IPv6-only connections to contact this ICAP service.
8023 on-overload=block|bypass|wait|force
8024 If the service Max-Connections limit has been reached, do
8025 one of the following for each new ICAP transaction:
8026 * block: send an HTTP error response to the client
8027 * bypass: ignore the "over-connected" ICAP service
8028 * wait: wait (in a FIFO queue) for an ICAP connection slot
8029 * force: proceed, ignoring the Max-Connections limit
8031 In SMP mode with N workers, each worker assumes the service
8032 connection limit is Max-Connections/N, even though not all
8033 workers may use a given service.
8035 The default value is "bypass" if service is bypassable,
8036 otherwise it is set to "wait".
8040 Use the given number as the Max-Connections limit, regardless
8041 of the Max-Connections value given by the service, if any.
8043 Older icap_service format without optional named parameters is
8044 deprecated but supported for backward compatibility.
8047 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
8048 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
8052 TYPE: icap_class_type
8057 This deprecated option was documented to define an ICAP service
8058 chain, even though it actually defined a set of similar, redundant
8059 services, and the chains were not supported.
8061 To define a set of redundant services, please use the
8062 adaptation_service_set directive. For service chains, use
8063 adaptation_service_chain.
8067 TYPE: icap_access_type
8072 This option is deprecated. Please use adaptation_access, which
8073 has the same ICAP functionality, but comes with better
8074 documentation, and eCAP support.
8079 -----------------------------------------------------------------------------
8086 LOC: Adaptation::Ecap::TheConfig.onoff
8089 Controls whether eCAP support is enabled.
8093 TYPE: ecap_service_type
8095 LOC: Adaptation::Ecap::TheConfig
8098 Defines a single eCAP service
8100 ecap_service id vectoring_point uri [option ...]
8103 an opaque identifier or name which is used to direct traffic to
8104 this specific service. Must be unique among all adaptation
8105 services in squid.conf.
8107 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
8108 This specifies at which point of transaction processing the
8109 eCAP service should be activated. *_postcache vectoring points
8110 are not yet supported.
8112 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
8113 Squid uses the eCAP service URI to match this configuration
8114 line with one of the dynamically loaded services. Each loaded
8115 eCAP service must have a unique URI. Obtain the right URI from
8116 the service provider.
8118 To activate a service, use the adaptation_access directive. To group
8119 services, use adaptation_service_chain and adaptation_service_set.
8121 Service options are separated by white space. eCAP services support
8122 the following name=value options:
8125 If set to 'on' or '1', the eCAP service is treated as optional.
8126 If the service cannot be reached or malfunctions, Squid will try
8127 to ignore any errors and process the message as if the service
8128 was not enabled. No all eCAP errors can be bypassed.
8129 If set to 'off' or '0', the eCAP service is treated as essential
8130 and all eCAP errors will result in an error page returned to the
8133 Bypass is off by default: services are treated as essential.
8136 If set to 'on' or '1', the eCAP service is allowed to
8137 dynamically change the current message adaptation plan by
8138 returning a chain of services to be used next.
8140 Dynamic adaptation plan may cross or cover multiple supported
8141 vectoring points in their natural processing order.
8143 Routing is not allowed by default.
8145 Older ecap_service format without optional named parameters is
8146 deprecated but supported for backward compatibility.
8150 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
8151 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
8154 NAME: loadable_modules
8156 IFDEF: USE_LOADABLE_MODULES
8157 LOC: Config.loadable_module_names
8160 Instructs Squid to load the specified dynamic module(s) or activate
8161 preloaded module(s).
8163 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
8167 MESSAGE ADAPTATION OPTIONS
8168 -----------------------------------------------------------------------------
8171 NAME: adaptation_service_set
8172 TYPE: adaptation_service_set_type
8173 IFDEF: USE_ADAPTATION
8178 Configures an ordered set of similar, redundant services. This is
8179 useful when hot standby or backup adaptation servers are available.
8181 adaptation_service_set set_name service_name1 service_name2 ...
8183 The named services are used in the set declaration order. The first
8184 applicable adaptation service from the set is used first. The next
8185 applicable service is tried if and only if the transaction with the
8186 previous service fails and the message waiting to be adapted is still
8189 When adaptation starts, broken services are ignored as if they were
8190 not a part of the set. A broken service is a down optional service.
8192 The services in a set must be attached to the same vectoring point
8193 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8195 If all services in a set are optional then adaptation failures are
8196 bypassable. If all services in the set are essential, then a
8197 transaction failure with one service may still be retried using
8198 another service from the set, but when all services fail, the master
8199 transaction fails as well.
8201 A set may contain a mix of optional and essential services, but that
8202 is likely to lead to surprising results because broken services become
8203 ignored (see above), making previously bypassable failures fatal.
8204 Technically, it is the bypassability of the last failed service that
8207 See also: adaptation_access adaptation_service_chain
8210 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
8211 adaptation service_set svcLogger loggerLocal loggerRemote
8214 NAME: adaptation_service_chain
8215 TYPE: adaptation_service_chain_type
8216 IFDEF: USE_ADAPTATION
8221 Configures a list of complementary services that will be applied
8222 one-by-one, forming an adaptation chain or pipeline. This is useful
8223 when Squid must perform different adaptations on the same message.
8225 adaptation_service_chain chain_name service_name1 svc_name2 ...
8227 The named services are used in the chain declaration order. The first
8228 applicable adaptation service from the chain is used first. The next
8229 applicable service is applied to the successful adaptation results of
8230 the previous service in the chain.
8232 When adaptation starts, broken services are ignored as if they were
8233 not a part of the chain. A broken service is a down optional service.
8235 Request satisfaction terminates the adaptation chain because Squid
8236 does not currently allow declaration of RESPMOD services at the
8237 "reqmod_precache" vectoring point (see icap_service or ecap_service).
8239 The services in a chain must be attached to the same vectoring point
8240 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8242 A chain may contain a mix of optional and essential services. If an
8243 essential adaptation fails (or the failure cannot be bypassed for
8244 other reasons), the master transaction fails. Otherwise, the failure
8245 is bypassed as if the failed adaptation service was not in the chain.
8247 See also: adaptation_access adaptation_service_set
8250 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8253 NAME: adaptation_access
8254 TYPE: adaptation_access_type
8255 IFDEF: USE_ADAPTATION
8258 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8260 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8262 adaptation_access service_name allow|deny [!]aclname...
8263 adaptation_access set_name allow|deny [!]aclname...
8265 At each supported vectoring point, the adaptation_access
8266 statements are processed in the order they appear in this
8267 configuration file. Statements pointing to the following services
8268 are ignored (i.e., skipped without checking their ACL):
8270 - services serving different vectoring points
8271 - "broken-but-bypassable" services
8272 - "up" services configured to ignore such transactions
8273 (e.g., based on the ICAP Transfer-Ignore header).
8275 When a set_name is used, all services in the set are checked
8276 using the same rules, to find the first applicable one. See
8277 adaptation_service_set for details.
8279 If an access list is checked and there is a match, the
8280 processing stops: For an "allow" rule, the corresponding
8281 adaptation service is used for the transaction. For a "deny"
8282 rule, no adaptation service is activated.
8284 It is currently not possible to apply more than one adaptation
8285 service at the same vectoring point to the same HTTP transaction.
8287 See also: icap_service and ecap_service
8290 adaptation_access service_1 allow all
8293 NAME: adaptation_service_iteration_limit
8295 IFDEF: USE_ADAPTATION
8296 LOC: Adaptation::Config::service_iteration_limit
8299 Limits the number of iterations allowed when applying adaptation
8300 services to a message. If your longest adaptation set or chain
8301 may have more than 16 services, increase the limit beyond its
8302 default value of 16. If detecting infinite iteration loops sooner
8303 is critical, make the iteration limit match the actual number
8304 of services in your longest adaptation set or chain.
8306 Infinite adaptation loops are most likely with routing services.
8308 See also: icap_service routing=1
8311 NAME: adaptation_masterx_shared_names
8313 IFDEF: USE_ADAPTATION
8314 LOC: Adaptation::Config::masterx_shared_name
8317 For each master transaction (i.e., the HTTP request and response
8318 sequence, including all related ICAP and eCAP exchanges), Squid
8319 maintains a table of metadata. The table entries are (name, value)
8320 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8321 with the master transaction.
8323 This option specifies the table entry names that Squid must accept
8324 from and forward to the adaptation transactions.
8326 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8327 shared table by returning an ICAP header field with a name
8328 specified in adaptation_masterx_shared_names.
8330 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8331 shared table by implementing the libecap::visitEachOption() API
8332 to provide an option with a name specified in
8333 adaptation_masterx_shared_names.
8335 Squid will store and forward the set entry to subsequent adaptation
8336 transactions within the same master transaction scope.
8338 Only one shared entry name is supported at this time.
8341 # share authentication information among ICAP services
8342 adaptation_masterx_shared_names X-Subscriber-ID
8345 NAME: adaptation_meta
8347 IFDEF: USE_ADAPTATION
8348 LOC: Adaptation::Config::metaHeaders
8351 This option allows Squid administrator to add custom ICAP request
8352 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8353 Use it to pass custom authentication tokens and other
8354 transaction-state related meta information to an ICAP/eCAP service.
8356 The addition of a meta header is ACL-driven:
8357 adaptation_meta name value [!]aclname ...
8359 Processing for a given header name stops after the first ACL list match.
8360 Thus, it is impossible to add two headers with the same name. If no ACL
8361 lists match for a given header name, no such header is added. For
8364 # do not debug transactions except for those that need debugging
8365 adaptation_meta X-Debug 1 needs_debugging
8367 # log all transactions except for those that must remain secret
8368 adaptation_meta X-Log 1 !keep_secret
8370 # mark transactions from users in the "G 1" group
8371 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8373 The "value" parameter may be a regular squid.conf token or a "double
8374 quoted string". Within the quoted string, use backslash (\) to escape
8375 any character, which is currently only useful for escaping backslashes
8376 and double quotes. For example,
8377 "this string has one backslash (\\) and two \"quotes\""
8379 Used adaptation_meta header values may be logged via %note
8380 logformat code. If multiple adaptation_meta headers with the same name
8381 are used during master transaction lifetime, the header values are
8382 logged in the order they were used and duplicate values are ignored
8383 (only the first repeated value will be logged).
8389 LOC: Adaptation::Icap::TheConfig.repeat
8390 DEFAULT_IF_NONE: deny all
8392 This ACL determines which retriable ICAP transactions are
8393 retried. Transactions that received a complete ICAP response
8394 and did not have to consume or produce HTTP bodies to receive
8395 that response are usually retriable.
8397 icap_retry allow|deny [!]aclname ...
8399 Squid automatically retries some ICAP I/O timeouts and errors
8400 due to persistent connection race conditions.
8402 See also: icap_retry_limit
8405 NAME: icap_retry_limit
8408 LOC: Adaptation::Icap::TheConfig.repeat_limit
8410 DEFAULT_DOC: No retries are allowed.
8412 Limits the number of retries allowed.
8414 Communication errors due to persistent connection race
8415 conditions are unavoidable, automatically retried, and do not
8416 count against this limit.
8418 See also: icap_retry
8424 -----------------------------------------------------------------------------
8427 NAME: check_hostnames
8430 LOC: Config.onoff.check_hostnames
8432 For security and stability reasons Squid can check
8433 hostnames for Internet standard RFC compliance. If you want
8434 Squid to perform these checks turn this directive on.
8437 NAME: allow_underscore
8440 LOC: Config.onoff.allow_underscore
8442 Underscore characters is not strictly allowed in Internet hostnames
8443 but nevertheless used by many sites. Set this to off if you want
8444 Squid to be strict about the standard.
8445 This check is performed only when check_hostnames is set to on.
8448 NAME: dns_retransmit_interval
8451 LOC: Config.Timeout.idns_retransmit
8453 Initial retransmit interval for DNS queries. The interval is
8454 doubled each time all configured DNS servers have been tried.
8460 LOC: Config.Timeout.idns_query
8462 DNS Query timeout. If no response is received to a DNS query
8463 within this time all DNS servers for the queried domain
8464 are assumed to be unavailable.
8467 NAME: dns_packet_max
8469 DEFAULT_DOC: EDNS disabled
8471 LOC: Config.dns.packet_max
8473 Maximum number of bytes packet size to advertise via EDNS.
8474 Set to "none" to disable EDNS large packet support.
8476 For legacy reasons DNS UDP replies will default to 512 bytes which
8477 is too small for many responses. EDNS provides a means for Squid to
8478 negotiate receiving larger responses back immediately without having
8479 to failover with repeat requests. Responses larger than this limit
8480 will retain the old behaviour of failover to TCP DNS.
8482 Squid has no real fixed limit internally, but allowing packet sizes
8483 over 1500 bytes requires network jumbogram support and is usually not
8486 WARNING: The RFC also indicates that some older resolvers will reply
8487 with failure of the whole request if the extension is added. Some
8488 resolvers have already been identified which will reply with mangled
8489 EDNS response on occasion. Usually in response to many-KB jumbogram
8490 sizes being advertised by Squid.
8491 Squid will currently treat these both as an unable-to-resolve domain
8492 even if it would be resolvable without EDNS.
8499 DEFAULT_DOC: Search for single-label domain names is disabled.
8500 LOC: Config.onoff.res_defnames
8502 Normally the RES_DEFNAMES resolver option is disabled
8503 (see res_init(3)). This prevents caches in a hierarchy
8504 from interpreting single-component hostnames locally. To allow
8505 Squid to handle single-component names, enable this option.
8508 NAME: dns_multicast_local
8512 DEFAULT_DOC: Search for .local and .arpa names is disabled.
8513 LOC: Config.onoff.dns_mdns
8515 When set to on, Squid sends multicast DNS lookups on the local
8516 network for domains ending in .local and .arpa.
8517 This enables local servers and devices to be contacted in an
8518 ad-hoc or zero-configuration network environment.
8521 NAME: dns_nameservers
8524 DEFAULT_DOC: Use operating system definitions
8525 LOC: Config.dns_nameservers
8527 Use this if you want to specify a list of DNS name servers
8528 (IP addresses) to use instead of those given in your
8529 /etc/resolv.conf file.
8531 On Windows platforms, if no value is specified here or in
8532 the /etc/resolv.conf file, the list of DNS name servers are
8533 taken from the Windows registry, both static and dynamic DHCP
8534 configurations are supported.
8536 Example: dns_nameservers 10.0.0.1 192.172.0.4
8541 DEFAULT: @DEFAULT_HOSTS@
8542 LOC: Config.etcHostsPath
8544 Location of the host-local IP name-address associations
8545 database. Most Operating Systems have such a file on different
8547 - Un*X & Linux: /etc/hosts
8548 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
8549 (%SystemRoot% value install default is c:\winnt)
8550 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
8551 (%SystemRoot% value install default is c:\windows)
8552 - Windows 9x/Me: %windir%\hosts
8553 (%windir% value is usually c:\windows)
8554 - Cygwin: /etc/hosts
8556 The file contains newline-separated definitions, in the
8557 form ip_address_in_dotted_form name [name ...] names are
8558 whitespace-separated. Lines beginning with an hash (#)
8559 character are comments.
8561 The file is checked at startup and upon configuration.
8562 If set to 'none', it won't be checked.
8563 If append_domain is used, that domain will be added to
8564 domain-local (i.e. not containing any dot character) host
8570 LOC: Config.appendDomain
8572 DEFAULT_DOC: Use operating system definitions
8574 Appends local domain name to hostnames without any dots in
8575 them. append_domain must begin with a period.
8577 Be warned there are now Internet names with no dots in
8578 them using only top-domain names, so setting this may
8579 cause some Internet sites to become unavailable.
8582 append_domain .yourdomain.com
8585 NAME: ignore_unknown_nameservers
8587 LOC: Config.onoff.ignore_unknown_nameservers
8590 By default Squid checks that DNS responses are received
8591 from the same IP addresses they are sent to. If they
8592 don't match, Squid ignores the response and writes a warning
8593 message to cache.log. You can allow responses from unknown
8594 nameservers by setting this option to 'off'.
8600 LOC: Config.dns.v4_first
8602 With the IPv6 Internet being as fast or faster than IPv4 Internet
8603 for most networks Squid prefers to contact websites over IPv6.
8605 This option reverses the order of preference to make Squid contact
8606 dual-stack websites over IPv4 first. Squid will still perform both
8607 IPv6 and IPv4 DNS lookups before connecting.
8610 This option will restrict the situations under which IPv6
8611 connectivity is used (and tested). Hiding network problems
8612 which would otherwise be detected and warned about.
8616 COMMENT: (number of entries)
8619 LOC: Config.ipcache.size
8621 Maximum number of DNS IP cache entries.
8628 LOC: Config.ipcache.low
8635 LOC: Config.ipcache.high
8637 The size, low-, and high-water marks for the IP cache.
8640 NAME: fqdncache_size
8641 COMMENT: (number of entries)
8644 LOC: Config.fqdncache.size
8646 Maximum number of FQDN cache entries.
8651 -----------------------------------------------------------------------------
8654 NAME: configuration_includes_quoted_values
8656 TYPE: configuration_includes_quoted_values
8658 LOC: ConfigParser::RecognizeQuotedValues
8660 If set, Squid will recognize each "quoted string" after a configuration
8661 directive as a single parameter. The quotes are stripped before the
8662 parameter value is interpreted or used.
8663 See "Values with spaces, quotes, and other special characters"
8664 section for more details.
8671 LOC: Config.onoff.mem_pools
8673 If set, Squid will keep pools of allocated (but unused) memory
8674 available for future use. If memory is a premium on your
8675 system and you believe your malloc library outperforms Squid
8676 routines, disable this.
8679 NAME: memory_pools_limit
8683 LOC: Config.MemPools.limit
8685 Used only with memory_pools on:
8686 memory_pools_limit 50 MB
8688 If set to a non-zero value, Squid will keep at most the specified
8689 limit of allocated (but unused) memory in memory pools. All free()
8690 requests that exceed this limit will be handled by your malloc
8691 library. Squid does not pre-allocate any memory, just safe-keeps
8692 objects that otherwise would be free()d. Thus, it is safe to set
8693 memory_pools_limit to a reasonably high value even if your
8694 configuration will use less memory.
8696 If set to none, Squid will keep all memory it can. That is, there
8697 will be no limit on the total amount of memory used for safe-keeping.
8699 To disable memory allocation optimization, do not set
8700 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
8702 An overhead for maintaining memory pools is not taken into account
8703 when the limit is checked. This overhead is close to four bytes per
8704 object kept. However, pools may actually _save_ memory because of
8705 reduced memory thrashing in your malloc library.
8709 COMMENT: on|off|transparent|truncate|delete
8712 LOC: opt_forwarded_for
8714 If set to "on", Squid will append your client's IP address
8715 in the HTTP requests it forwards. By default it looks like:
8717 X-Forwarded-For: 192.1.2.3
8719 If set to "off", it will appear as
8721 X-Forwarded-For: unknown
8723 If set to "transparent", Squid will not alter the
8724 X-Forwarded-For header in any way.
8726 If set to "delete", Squid will delete the entire
8727 X-Forwarded-For header.
8729 If set to "truncate", Squid will remove all existing
8730 X-Forwarded-For entries, and place the client IP as the sole entry.
8733 NAME: cachemgr_passwd
8734 TYPE: cachemgrpasswd
8736 DEFAULT_DOC: No password. Actions which require password are denied.
8737 LOC: Config.passwd_list
8739 Specify passwords for cachemgr operations.
8741 Usage: cachemgr_passwd password action action ...
8743 Some valid actions are (see cache manager menu for a full list):
8783 * Indicates actions which will not be performed without a
8784 valid password, others can be performed if not listed here.
8786 To disable an action, set the password to "disable".
8787 To allow performing an action without a password, set the
8790 Use the keyword "all" to set the same password for all actions.
8793 cachemgr_passwd secret shutdown
8794 cachemgr_passwd lesssssssecret info stats/objects
8795 cachemgr_passwd disable all
8802 LOC: Config.onoff.client_db
8804 If you want to disable collecting per-client statistics,
8805 turn off client_db here.
8808 NAME: refresh_all_ims
8812 LOC: Config.onoff.refresh_all_ims
8814 When you enable this option, squid will always check
8815 the origin server for an update when a client sends an
8816 If-Modified-Since request. Many browsers use IMS
8817 requests when the user requests a reload, and this
8818 ensures those clients receive the latest version.
8820 By default (off), squid may return a Not Modified response
8821 based on the age of the cached version.
8824 NAME: reload_into_ims
8825 IFDEF: USE_HTTP_VIOLATIONS
8829 LOC: Config.onoff.reload_into_ims
8831 When you enable this option, client no-cache or ``reload''
8832 requests will be changed to If-Modified-Since requests.
8833 Doing this VIOLATES the HTTP standard. Enabling this
8834 feature could make you liable for problems which it
8837 see also refresh_pattern for a more selective approach.
8840 NAME: connect_retries
8842 LOC: Config.connect_retries
8844 DEFAULT_DOC: Do not retry failed connections.
8846 This sets the maximum number of connection attempts made for each
8847 TCP connection. The connect_retries attempts must all still
8848 complete within the connection timeout period.
8850 The default is not to re-try if the first connection attempt fails.
8851 The (not recommended) maximum is 10 tries.
8853 A warning message will be generated if it is set to a too-high
8854 value and the configured value will be over-ridden.
8856 Note: These re-tries are in addition to forward_max_tries
8857 which limit how many different addresses may be tried to find
8861 NAME: retry_on_error
8863 LOC: Config.retry.onerror
8866 If set to ON Squid will automatically retry requests when
8867 receiving an error response with status 403 (Forbidden),
8868 500 (Internal Error), 501 or 503 (Service not available).
8869 Status 502 and 504 (Gateway errors) are always retried.
8871 This is mainly useful if you are in a complex cache hierarchy to
8872 work around access control errors.
8874 NOTE: This retry will attempt to find another working destination.
8875 Which is different from the server which just failed.
8878 NAME: as_whois_server
8880 LOC: Config.as_whois_server
8881 DEFAULT: whois.ra.net
8883 WHOIS server to query for AS numbers. NOTE: AS numbers are
8884 queried only when Squid starts up, not for every request.
8889 LOC: Config.onoff.offline
8892 Enable this option and Squid will never try to validate cached
8896 NAME: uri_whitespace
8897 TYPE: uri_whitespace
8898 LOC: Config.uri_whitespace
8901 What to do with requests that have whitespace characters in the
8904 strip: The whitespace characters are stripped out of the URL.
8905 This is the behavior recommended by RFC2396 and RFC3986
8906 for tolerant handling of generic URI.
8907 NOTE: This is one difference between generic URI and HTTP URLs.
8909 deny: The request is denied. The user receives an "Invalid
8911 This is the behaviour recommended by RFC2616 for safe
8912 handling of HTTP request URL.
8914 allow: The request is allowed and the URI is not changed. The
8915 whitespace characters remain in the URI. Note the
8916 whitespace is passed to redirector processes if they
8918 Note this may be considered a violation of RFC2616
8919 request parsing where whitespace is prohibited in the
8922 encode: The request is allowed and the whitespace characters are
8923 encoded according to RFC1738.
8925 chop: The request is allowed and the URI is chopped at the
8929 NOTE the current Squid implementation of encode and chop violates
8930 RFC2616 by not using a 301 redirect after altering the URL.
8935 LOC: Config.chroot_dir
8938 Specifies a directory where Squid should do a chroot() while
8939 initializing. This also causes Squid to fully drop root
8940 privileges after initializing. This means, for example, if you
8941 use a HTTP port less than 1024 and try to reconfigure, you may
8942 get an error saying that Squid can not open the port.
8945 NAME: balance_on_multiple_ip
8947 LOC: Config.onoff.balance_on_multiple_ip
8950 Modern IP resolvers in squid sort lookup results by preferred access.
8951 By default squid will use these IP in order and only rotates to
8952 the next listed when the most preffered fails.
8954 Some load balancing servers based on round robin DNS have been
8955 found not to preserve user session state across requests
8956 to different IP addresses.
8958 Enabling this directive Squid rotates IP's per request.
8961 NAME: pipeline_prefetch
8962 TYPE: pipelinePrefetch
8963 LOC: Config.pipeline_max_prefetch
8965 DEFAULT_DOC: Do not pre-parse pipelined requests.
8967 HTTP clients may send a pipeline of 1+N requests to Squid using a
8968 single connection, without waiting for Squid to respond to the first
8969 of those requests. This option limits the number of concurrent
8970 requests Squid will try to handle in parallel. If set to N, Squid
8971 will try to receive and process up to 1+N requests on the same
8972 connection concurrently.
8974 Defaults to 0 (off) for bandwidth management and access logging
8977 NOTE: pipelining requires persistent connections to clients.
8979 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8982 NAME: high_response_time_warning
8985 LOC: Config.warnings.high_rptm
8987 DEFAULT_DOC: disabled.
8989 If the one-minute median response time exceeds this value,
8990 Squid prints a WARNING with debug level 0 to get the
8991 administrators attention. The value is in milliseconds.
8994 NAME: high_page_fault_warning
8996 LOC: Config.warnings.high_pf
8998 DEFAULT_DOC: disabled.
9000 If the one-minute average page fault rate exceeds this
9001 value, Squid prints a WARNING with debug level 0 to get
9002 the administrators attention. The value is in page faults
9006 NAME: high_memory_warning
9008 LOC: Config.warnings.high_memory
9009 IFDEF: HAVE_MSTATS&&HAVE_GNUMALLOC_H
9011 DEFAULT_DOC: disabled.
9013 If the memory usage (as determined by gnumalloc, if available and used)
9014 exceeds this amount, Squid prints a WARNING with debug level 0 to get
9015 the administrators attention.
9017 # TODO: link high_memory_warning to mempools?
9019 NAME: sleep_after_fork
9020 COMMENT: (microseconds)
9022 LOC: Config.sleep_after_fork
9025 When this is set to a non-zero value, the main Squid process
9026 sleeps the specified number of microseconds after a fork()
9027 system call. This sleep may help the situation where your
9028 system reports fork() failures due to lack of (virtual)
9029 memory. Note, however, if you have a lot of child
9030 processes, these sleep delays will add up and your
9031 Squid will not service requests for some amount of time
9032 until all the child processes have been started.
9033 On Windows value less then 1000 (1 milliseconds) are
9037 NAME: windows_ipaddrchangemonitor
9038 IFDEF: _SQUID_WINDOWS_
9042 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
9044 On Windows Squid by default will monitor IP address changes and will
9045 reconfigure itself after any detected event. This is very useful for
9046 proxies connected to internet with dial-up interfaces.
9047 In some cases (a Proxy server acting as VPN gateway is one) it could be
9048 desiderable to disable this behaviour setting this to 'off'.
9049 Note: after changing this, Squid service must be restarted.
9054 IFDEF: USE_SQUID_EUI
9056 LOC: Eui::TheConfig.euiLookup
9058 Whether to lookup the EUI or MAC address of a connected client.
9061 NAME: max_filedescriptors max_filedesc
9064 DEFAULT_DOC: Use operating system limits set by ulimit.
9065 LOC: Config.max_filedescriptors
9067 Reduce the maximum number of filedescriptors supported below
9068 the usual operating system defaults.
9070 Remove from squid.conf to inherit the current ulimit setting.
9072 Note: Changing this requires a restart of Squid. Also
9073 not all I/O types supports large values (eg on Windows).
9080 DEFAULT_DOC: SMP support disabled.
9082 Number of main Squid processes or "workers" to fork and maintain.
9083 0: "no daemon" mode, like running "squid -N ..."
9084 1: "no SMP" mode, start one main Squid process daemon (default)
9085 N: start N main Squid process daemons (i.e., SMP mode)
9087 In SMP mode, each worker does nearly all what a single Squid daemon
9088 does (e.g., listen on http_port and forward HTTP requests).
9091 NAME: cpu_affinity_map
9092 TYPE: CpuAffinityMap
9093 LOC: Config.cpuAffinityMap
9095 DEFAULT_DOC: Let operating system decide.
9097 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
9099 Sets 1:1 mapping between Squid processes and CPU cores. For example,
9101 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
9103 affects processes 1 through 4 only and places them on the first
9104 four even cores, starting with core #1.
9106 CPU cores are numbered starting from 1. Requires support for
9107 sched_getaffinity(2) and sched_setaffinity(2) system calls.
9109 Multiple cpu_affinity_map options are merged.