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
68 Conditional configuration
70 If-statements can be used to make configuration directives
74 ... regular configuration directives ...
76 ... regular configuration directives ...]
79 The else part is optional. The keywords "if", "else", and "endif"
80 must be typed on their own lines, as if they were regular
81 configuration directives.
83 NOTE: An else-if condition is not supported.
85 These individual conditions types are supported:
88 Always evaluates to true.
90 Always evaluates to false.
92 Equality comparison of two integer numbers.
97 The following SMP-related preprocessor macros can be used.
99 ${process_name} expands to the current Squid process "name"
100 (e.g., squid1, squid2, or cache1).
102 ${process_number} expands to the current Squid process
103 identifier, which is an integer number (e.g., 1, 2, 3) unique
104 across all Squid processes.
107 # options still not yet ported from 2.7 to 3.x
108 NAME: broken_vary_encoding
111 This option is not yet supported by Squid-3.
117 This option is not yet supported by Squid-3.
120 NAME: collapsed_forwarding
123 This option is not yet supported by Squid-3. see http://bugs.squid-cache.org/show_bug.cgi?id=3495
129 This option is not yet supported by Squid-3.
132 NAME: external_refresh_check
135 This option is not yet supported by Squid-3.
138 NAME: ignore_ims_on_miss
141 This option is not yet supported by Squid-3.
144 NAME: location_rewrite_program location_rewrite_access location_rewrite_children location_rewrite_concurrency
147 This option is not yet supported by Squid-3.
150 NAME: refresh_stale_hit
153 This option is not yet supported by Squid-3.
156 # no Options Removed in 3.3
158 # Options Removed in 3.2
159 NAME: ignore_expect_100
162 Remove this line. The HTTP/1.1 feature is now fully supported by default.
165 NAME: dns_v4_fallback
168 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
174 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
177 NAME: maximum_single_addr_tries
180 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
186 Remove this line. The feature is supported by default in storage types where update is implemented.
189 NAME: url_rewrite_concurrency
192 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
195 # Options Removed in 3.1
199 Remove this line. DNS is no longer tested on startup.
202 NAME: extension_methods
205 Remove this line. All valid methods for HTTP are accepted by default.
208 # 2.7 Options Removed/Replaced in 3.2
213 # 2.7 Options Removed/Replaced in 3.1
221 Remove this line. HTTP/1.1 is supported by default.
224 NAME: upgrade_http0.9
227 Remove this line. ICY/1.0 streaming protocol is supported by default.
230 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
233 Alter these entries. Use the qos_flows directive instead.
236 # Options Removed in 3.0
240 Since squid-3.0 replace with request_header_access or reply_header_access
241 depending on whether you wish to match client requests or server replies.
244 NAME: httpd_accel_no_pmtu_disc
247 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
250 NAME: wais_relay_host
253 Replace this line with 'cache_peer' configuration.
256 NAME: wais_relay_port
259 Replace this line with 'cache_peer' configuration.
263 OPTIONS FOR AUTHENTICATION
264 -----------------------------------------------------------------------------
273 This is used to define parameters for the various authentication
274 schemes supported by Squid.
276 format: auth_param scheme parameter [setting]
278 The order in which authentication schemes are presented to the client is
279 dependent on the order the scheme first appears in config file. IE
280 has a bug (it's not RFC 2617 compliant) in that it will use the basic
281 scheme if basic is the first entry presented, even if more secure
282 schemes are presented. For now use the order in the recommended
283 settings section below. If other browsers have difficulties (don't
284 recognize the schemes offered even if you are using basic) either
285 put basic first, or disable the other schemes (by commenting out their
288 Once an authentication scheme is fully configured, it can only be
289 shutdown by shutting squid down and restarting. Changes can be made on
290 the fly and activated with a reconfigure. I.E. You can change to a
291 different helper, but not unconfigure the helper completely.
293 Please note that while this directive defines how Squid processes
294 authentication it does not automatically activate authentication.
295 To use authentication you must in addition make use of ACLs based
296 on login name in http_access (proxy_auth, proxy_auth_regex or
297 external with %LOGIN used in the format tag). The browser will be
298 challenged for authentication on the first such acl encountered
299 in http_access processing and will also be re-challenged for new
300 login credentials if the request is being denied by a proxy_auth
303 WARNING: authentication can't be used in a transparently intercepting
304 proxy as the client then thinks it is talking to an origin server and
305 not the proxy. This is a limitation of bending the TCP/IP protocol to
306 transparently intercepting port 80, not a limitation in Squid.
307 Ports flagged 'transparent', 'intercept', or 'tproxy' have
308 authentication disabled.
310 === Parameters for the basic scheme follow. ===
313 Specify the command for the external authenticator. Such a program
314 reads a line containing "username password" and replies with one of
321 the user does not exist.
324 An internal error occurred in the helper, preventing
325 a result being identified.
327 "ERR" and "BH" results may optionally be followed by message="..."
328 containing a description available as %m in the returned error page.
330 If you use an authenticator, make sure you have 1 acl of type
333 By default, the basic authentication scheme is not used unless a
334 program is specified.
336 If you want to use the traditional NCSA proxy authentication, set
337 this line to something like
339 auth_param basic program @DEFAULT_PREFIX@/libexec/basic_ncsa_auth @DEFAULT_PREFIX@/etc/passwd
342 HTTP uses iso-latin-1 as character set, while some authentication
343 backends such as LDAP expects UTF-8. If this is set to on Squid will
344 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
345 username & password to the helper.
347 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
348 The maximum number of authenticator processes to spawn. If you start too few
349 Squid will have to wait for them to process a backlog of credential
350 verifications, slowing it down. When password verifications are
351 done via a (slow) network you are likely to need lots of
352 authenticator processes.
354 The startup= and idle= options permit some skew in the exact amount
355 run. A minimum of startup=N will begin during startup and reconfigure.
356 Squid will start more in groups of up to idle=N in an attempt to meet
357 traffic needs and to keep idle=N free above those traffic needs up to
360 The concurrency= option sets the number of concurrent requests the
361 helper can process. The default of 0 is used for helpers who only
362 supports one request at a time. Setting this to a number greater than
363 0 changes the protocol used to include a channel number first on the
364 request/response line, allowing multiple requests to be sent to the
365 same helper in parallel without waiting for the response.
366 Must not be set unless it's known the helper supports this.
368 auth_param basic children 20 startup=0 idle=1
371 Specifies the realm name which is to be reported to the
372 client for the basic proxy authentication scheme (part of
373 the text the user will see when prompted their username and
374 password). There is no default.
375 auth_param basic realm Squid proxy-caching web server
377 "credentialsttl" timetolive
378 Specifies how long squid assumes an externally validated
379 username:password pair is valid for - in other words how
380 often the helper program is called for that user. Set this
381 low to force revalidation with short lived passwords. Note
382 setting this high does not impact your susceptibility
383 to replay attacks unless you are using an one-time password
384 system (such as SecureID). If you are using such a system,
385 you will be vulnerable to replay attacks unless you also
386 use the max_user_ip ACL in an http_access rule.
388 "casesensitive" on|off
389 Specifies if usernames are case sensitive. Most user databases are
390 case insensitive allowing the same username to be spelled using both
391 lower and upper case letters, but some are case sensitive. This
392 makes a big difference for user_max_ip ACL processing and similar.
393 auth_param basic casesensitive off
395 === Parameters for the digest scheme follow ===
398 Specify the command for the external authenticator. Such
399 a program reads a line containing "username":"realm" and
400 replies with one of three results:
403 the user exists. The ha1= key is mandatory and
404 contains the appropriate H(A1) value, hex encoded.
405 See rfc 2616 for the definition of H(A1).
408 the user does not exist.
411 An internal error occurred in the helper, preventing
412 a result being identified.
414 "ERR" and "BH" results may optionally be followed by message="..."
415 containing a description available as %m in the returned error page.
417 By default, the digest authentication scheme is not used unless a
418 program is specified.
420 If you want to use a digest authenticator, set this line to
423 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
426 HTTP uses iso-latin-1 as character set, while some authentication
427 backends such as LDAP expects UTF-8. If this is set to on Squid will
428 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
429 username & password to the helper.
431 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
432 The maximum number of authenticator processes to spawn (default 5).
433 If you start too few Squid will have to wait for them to
434 process a backlog of H(A1) calculations, slowing it down.
435 When the H(A1) calculations are done via a (slow) network
436 you are likely to need lots of authenticator processes.
438 The startup= and idle= options permit some skew in the exact amount
439 run. A minimum of startup=N will begin during startup and reconfigure.
440 Squid will start more in groups of up to idle=N in an attempt to meet
441 traffic needs and to keep idle=N free above those traffic needs up to
444 The concurrency= option sets the number of concurrent requests the
445 helper can process. The default of 0 is used for helpers who only
446 supports one request at a time. Setting this to a number greater than
447 0 changes the protocol used to include a channel number first on the
448 request/response line, allowing multiple requests to be sent to the
449 same helper in parallel without waiting for the response.
450 Must not be set unless it's known the helper supports this.
452 auth_param digest children 20 startup=0 idle=1
455 Specifies the realm name which is to be reported to the
456 client for the digest proxy authentication scheme (part of
457 the text the user will see when prompted their username and
458 password). There is no default.
459 auth_param digest realm Squid proxy-caching web server
461 "nonce_garbage_interval" timeinterval
462 Specifies the interval that nonces that have been issued
463 to client_agent's are checked for validity.
465 "nonce_max_duration" timeinterval
466 Specifies the maximum length of time a given nonce will be
469 "nonce_max_count" number
470 Specifies the maximum number of times a given nonce can be
473 "nonce_strictness" on|off
474 Determines if squid requires strict increment-by-1 behavior
475 for nonce counts, or just incrementing (off - for use when
476 user agents generate nonce counts that occasionally miss 1
477 (ie, 1,2,4,6)). Default off.
479 "check_nonce_count" on|off
480 This directive if set to off can disable the nonce count check
481 completely to work around buggy digest qop implementations in
482 certain mainstream browser versions. Default on to check the
483 nonce count to protect from authentication replay attacks.
485 "post_workaround" on|off
486 This is a workaround to certain buggy browsers who sends
487 an incorrect request digest in POST requests when reusing
488 the same nonce as acquired earlier on a GET request.
490 === NTLM scheme options follow ===
493 Specify the command for the external NTLM authenticator.
494 Such a program reads exchanged NTLMSSP packets with
495 the browser via Squid until authentication is completed.
496 If you use an NTLM authenticator, make sure you have 1 acl
497 of type proxy_auth. By default, the NTLM authenticator program
500 auth_param ntlm program /usr/bin/ntlm_auth
502 "children" numberofchildren [startup=N] [idle=N]
503 The maximum number of authenticator processes to spawn (default 5).
504 If you start too few Squid will have to wait for them to
505 process a backlog of credential verifications, slowing it
506 down. When credential verifications are done via a (slow)
507 network you are likely to need lots of authenticator
510 The startup= and idle= options permit some skew in the exact amount
511 run. A minimum of startup=N will begin during startup and reconfigure.
512 Squid will start more in groups of up to idle=N in an attempt to meet
513 traffic needs and to keep idle=N free above those traffic needs up to
516 auth_param ntlm children 20 startup=0 idle=1
519 If you experience problems with PUT/POST requests when using the
520 Negotiate authentication scheme then you can try setting this to
521 off. This will cause Squid to forcibly close the connection on
522 the initial requests where the browser asks which schemes are
523 supported by the proxy.
525 auth_param ntlm keep_alive on
527 === Options for configuring the NEGOTIATE auth-scheme follow ===
530 Specify the command for the external Negotiate authenticator.
531 This protocol is used in Microsoft Active-Directory enabled setups with
532 the Microsoft Internet Explorer or Mozilla Firefox browsers.
533 Its main purpose is to exchange credentials with the Squid proxy
534 using the Kerberos mechanisms.
535 If you use a Negotiate authenticator, make sure you have at least
536 one acl of type proxy_auth active. By default, the negotiate
537 authenticator program is not used.
538 The only supported program for this role is the ntlm_auth
539 program distributed as part of Samba, version 4 or later.
541 auth_param negotiate program /usr/bin/ntlm_auth --helper-protocol=gss-spnego
543 "children" numberofchildren [startup=N] [idle=N]
544 The maximum number of authenticator processes to spawn (default 5).
545 If you start too few Squid will have to wait for them to
546 process a backlog of credential verifications, slowing it
547 down. When credential verifications are done via a (slow)
548 network you are likely to need lots of authenticator
551 The startup= and idle= options permit some skew in the exact amount
552 run. A minimum of startup=N will begin during startup and reconfigure.
553 Squid will start more in groups of up to idle=N in an attempt to meet
554 traffic needs and to keep idle=N free above those traffic needs up to
557 auth_param negotiate children 20 startup=0 idle=1
560 If you experience problems with PUT/POST requests when using the
561 Negotiate authentication scheme then you can try setting this to
562 off. This will cause Squid to forcibly close the connection on
563 the initial requests where the browser asks which schemes are
564 supported by the proxy.
566 auth_param negotiate keep_alive on
571 #Recommended minimum configuration per scheme:
572 #auth_param negotiate program <uncomment and complete this line to activate>
573 #auth_param negotiate children 20 startup=0 idle=1
574 #auth_param negotiate keep_alive on
576 #auth_param ntlm program <uncomment and complete this line to activate>
577 #auth_param ntlm children 20 startup=0 idle=1
578 #auth_param ntlm keep_alive on
580 #auth_param digest program <uncomment and complete this line>
581 #auth_param digest children 20 startup=0 idle=1
582 #auth_param digest realm Squid proxy-caching web server
583 #auth_param digest nonce_garbage_interval 5 minutes
584 #auth_param digest nonce_max_duration 30 minutes
585 #auth_param digest nonce_max_count 50
587 #auth_param basic program <uncomment and complete this line>
588 #auth_param basic children 5 startup=5 idle=1
589 #auth_param basic realm Squid proxy-caching web server
590 #auth_param basic credentialsttl 2 hours
593 NAME: authenticate_cache_garbage_interval
596 LOC: Config.authenticateGCInterval
598 The time period between garbage collection across the username cache.
599 This is a trade-off between memory utilization (long intervals - say
600 2 days) and CPU (short intervals - say 1 minute). Only change if you
604 NAME: authenticate_ttl
607 LOC: Config.authenticateTTL
609 The time a user & their credentials stay in the logged in
610 user cache since their last request. When the garbage
611 interval passes, all user credentials that have passed their
612 TTL are removed from memory.
615 NAME: authenticate_ip_ttl
617 LOC: Config.authenticateIpTTL
620 If you use proxy authentication and the 'max_user_ip' ACL,
621 this directive controls how long Squid remembers the IP
622 addresses associated with each user. Use a small value
623 (e.g., 60 seconds) if your users might change addresses
624 quickly, as is the case with dialup. You might be safe
625 using a larger value (e.g., 2 hours) in a corporate LAN
626 environment with relatively static address assignments.
631 -----------------------------------------------------------------------------
634 NAME: external_acl_type
635 TYPE: externalAclHelper
636 LOC: Config.externalAclHelperList
639 This option defines external acl classes using a helper program
640 to look up the status
642 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
646 ttl=n TTL in seconds for cached results (defaults to 3600
649 TTL for cached negative lookups (default same
652 Maximum number of acl helper processes spawned to service
653 external acl lookups of this type. (default 20)
655 Minimum number of acl helper processes to spawn during
656 startup and reconfigure to service external acl lookups
657 of this type. (default 0)
659 Number of acl helper processes to keep ahead of traffic
660 loads. Squid will spawn this many at once whenever load
661 rises above the capabilities of existing processes.
662 Up to the value of children-max. (default 1)
663 concurrency=n concurrency level per process. Only used with helpers
664 capable of processing more than one query at a time.
665 cache=n limit the result cache size, default is unbounded.
666 grace=n Percentage remaining of TTL where a refresh of a
667 cached entry should be initiated without needing to
668 wait for a new reply. (default is for no grace period)
669 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
670 ipv4 / ipv6 IP protocol used to communicate with this helper.
671 The default is to auto-detect IPv6 and use it when available.
673 FORMAT specifications
675 %LOGIN Authenticated user login name
676 %EXT_USER Username from previous external acl
677 %EXT_LOG Log details from previous external acl
678 %EXT_TAG Tag from previous external acl
679 %IDENT Ident user name
681 %SRCPORT Client source port
684 %PROTO Requested protocol
686 %PATH Requested URL path
687 %METHOD Request method
688 %MYADDR Squid interface address
689 %MYPORT Squid http_port number
690 %PATH Requested URL-path (including query-string if any)
691 %USER_CERT SSL User certificate in PEM format
692 %USER_CERTCHAIN SSL User certificate chain in PEM format
693 %USER_CERT_xx SSL User certificate subject attribute xx
694 %USER_CA_xx SSL User certificate issuer attribute xx
696 %>{Header} HTTP request header "Header"
698 HTTP request header "Hdr" list member "member"
700 HTTP request header list member using ; as
701 list separator. ; can be any non-alphanumeric
704 %<{Header} HTTP reply header "Header"
706 HTTP reply header "Hdr" list member "member"
708 HTTP reply header list member using ; as
709 list separator. ; can be any non-alphanumeric
712 %ACL The name of the ACL being tested.
713 %DATA The ACL arguments. If not used then any arguments
714 is automatically added at the end of the line
716 NOTE: this will encode the arguments as one token,
717 whereas the default will pass each separately.
719 %% The percent sign. Useful for helpers which need
720 an unchanging input format.
723 General request syntax:
725 [channel-ID] FORMAT-values [acl-values ...]
728 FORMAT-values consists of transaction details expanded with
729 whitespace separation per the config file FORMAT specification
730 using the FORMAT macros listed above.
732 acl-values consists of any string specified in the referencing
733 config 'acl ... external' line. see the "acl external" directive.
735 Request values sent to the helper are URL escaped to protect
736 each value in requests against whitespaces.
738 If using protocol=2.5 then the request sent to the helper is not
739 URL escaped to protect against whitespace.
741 NOTE: protocol=3.0 is deprecated as no longer necessary.
743 When using the concurrency= option the protocol is changed by
744 introducing a query channel tag in front of the request/response.
745 The query channel tag is a number between 0 and concurrency-1.
746 This value must be echoed back unchanged to Squid as the first part
747 of the response relating to its request.
750 The helper receives lines expanded per the above format specification
751 and for each input line returns 1 line starting with OK/ERR/BH result
752 code and optionally followed by additional keywords with more details.
755 General result syntax:
757 [channel-ID] result keyword=value ...
759 Result consists of one of the codes:
762 the ACL test produced a match.
765 the ACL test does not produce a match.
768 An internal error occurred in the helper, preventing
769 a result being identified.
771 The meaning of 'a match' is determined by your squid.conf
772 access control configuration. See the Squid wiki for details.
776 user= The users name (login)
778 password= The users password (for login= cache_peer option)
780 message= Message describing the reason for this response.
781 Available as %o in error pages.
782 Useful on (ERR and BH results).
784 tag= Apply a tag to a request. Only sets a tag once,
785 does not alter existing tags.
787 log= String to be logged in access.log. Available as
788 %ea in logformat specifications.
790 Any keywords may be sent on any response whether OK, ERR or BH.
792 All response keyword values need to be a single token with URL
793 escaping, or enclosed in double quotes (") and escaped using \ on
794 any double quotes or \ characters within the value. The wrapping
795 double quotes are removed before the value is interpreted by Squid.
796 \r and \n are also replace by CR and LF.
798 Some example key values:
802 user="J. \"Bob\" Smith"
809 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
810 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
811 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
812 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
813 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
816 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
817 DEFAULT: localhost src 127.0.0.1/32 ::1
818 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
819 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
821 Defining an Access List
823 Every access list definition must begin with an aclname and acltype,
824 followed by either type-specific arguments or a quoted filename that
827 acl aclname acltype argument ...
828 acl aclname acltype "file" ...
830 When using "file", the file should contain one item per line.
832 Some acl types supports options which changes their default behaviour.
833 The available options are:
835 -i,+i By default, regular expressions are CASE-SENSITIVE. To make them
836 case-insensitive, use the -i option. To return case-sensitive
837 use the +i option between patterns, or make a new ACL line
840 -n Disable lookups and address type conversions. If lookup or
841 conversion is required because the parameter type (IP or
842 domain name) does not match the message address type (domain
843 name or IP), then the ACL would immediately declare a mismatch
844 without any warnings or lookups.
846 -- Used to stop processing all options, in the case the first acl
847 value has '-' character as first character (for example the '-'
848 is a valid domain name)
850 Some acl types require suspending the current request in order
851 to access some external data source.
852 Those which do are marked with the tag [slow], those which
853 don't are marked as [fast].
854 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
855 for further information
857 ***** ACL TYPES AVAILABLE *****
859 acl aclname src ip-address/mask ... # clients IP address [fast]
860 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
861 acl aclname dst [-n] ip-address/mask ... # URL host's IP address [slow]
862 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
864 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
865 # The arp ACL requires the special configure option --enable-arp-acl.
866 # Furthermore, the ARP ACL code is not portable to all operating systems.
867 # It works on Linux, Solaris, Windows, FreeBSD, and some
868 # other *BSD variants.
871 # NOTE: Squid can only determine the MAC address for clients that are on
872 # the same subnet. If the client is on a different subnet,
873 # then Squid cannot find out its MAC address.
875 acl aclname srcdomain .foo.com ...
876 # reverse lookup, from client IP [slow]
877 acl aclname dstdomain [-n] .foo.com ...
878 # Destination server from URL [fast]
879 acl aclname srcdom_regex [-i] \.foo\.com ...
880 # regex matching client name [slow]
881 acl aclname dstdom_regex [-n] [-i] \.foo\.com ...
882 # regex matching server [fast]
884 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
885 # based URL is used and no match is found. The name "none" is used
886 # if the reverse lookup fails.
888 acl aclname src_as number ...
889 acl aclname dst_as number ...
891 # Except for access control, AS numbers can be used for
892 # routing of requests to specific caches. Here's an
893 # example for routing all requests for AS#1241 and only
894 # those to mycache.mydomain.net:
895 # acl asexample dst_as 1241
896 # cache_peer_access mycache.mydomain.net allow asexample
897 # cache_peer_access mycache_mydomain.net deny all
899 acl aclname peername myPeer ...
901 # match against a named cache_peer entry
902 # set unique name= on cache_peer lines for reliable use.
904 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
914 # h1:m1 must be less than h2:m2
916 acl aclname url_regex [-i] ^http:// ...
917 # regex matching on whole URL [fast]
918 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
919 # regex matching on URL login field
920 acl aclname urlpath_regex [-i] \.gif$ ...
921 # regex matching on URL path [fast]
923 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
925 acl aclname localport 3128 ... # TCP port the client connected to [fast]
926 # NP: for interception mode this is usually '80'
928 acl aclname myportname 3128 ... # http(s)_port name [fast]
930 acl aclname proto HTTP FTP ... # request protocol [fast]
932 acl aclname method GET POST ... # HTTP request method [fast]
934 acl aclname http_status 200 301 500- 400-403 ...
935 # status code in reply [fast]
937 acl aclname browser [-i] regexp ...
938 # pattern match on User-Agent header (see also req_header below) [fast]
940 acl aclname referer_regex [-i] regexp ...
941 # pattern match on Referer header [fast]
942 # Referer is highly unreliable, so use with care
944 acl aclname ident username ...
945 acl aclname ident_regex [-i] pattern ...
946 # string match on ident output [slow]
947 # use REQUIRED to accept any non-null ident.
949 acl aclname proxy_auth [-i] username ...
950 acl aclname proxy_auth_regex [-i] pattern ...
951 # perform http authentication challenge to the client and match against
952 # supplied credentials [slow]
954 # takes a list of allowed usernames.
955 # use REQUIRED to accept any valid username.
957 # Will use proxy authentication in forward-proxy scenarios, and plain
958 # http authenticaiton in reverse-proxy scenarios
960 # NOTE: when a Proxy-Authentication header is sent but it is not
961 # needed during ACL checking the username is NOT logged
964 # NOTE: proxy_auth requires a EXTERNAL authentication program
965 # to check username/password combinations (see
966 # auth_param directive).
968 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
969 # as the browser needs to be configured for using a proxy in order
970 # to respond to proxy authentication.
972 acl aclname snmp_community string ...
973 # A community string to limit access to your SNMP Agent [fast]
976 # acl snmppublic snmp_community public
978 acl aclname maxconn number
979 # This will be matched when the client's IP address has
980 # more than <number> TCP connections established. [fast]
981 # NOTE: This only measures direct TCP links so X-Forwarded-For
982 # indirect clients are not counted.
984 acl aclname max_user_ip [-s] number
985 # This will be matched when the user attempts to log in from more
986 # than <number> different ip addresses. The authenticate_ip_ttl
987 # parameter controls the timeout on the ip entries. [fast]
988 # If -s is specified the limit is strict, denying browsing
989 # from any further IP addresses until the ttl has expired. Without
990 # -s Squid will just annoy the user by "randomly" denying requests.
991 # (the counter is reset each time the limit is reached and a
993 # NOTE: in acceleration mode or where there is mesh of child proxies,
994 # clients may appear to come from multiple addresses if they are
995 # going through proxy farms, so a limit of 1 may cause user problems.
997 acl aclname random probability
998 # Pseudo-randomly match requests. Based on the probability given.
999 # Probability may be written as a decimal (0.333), fraction (1/3)
1000 # or ratio of matches:non-matches (3:5).
1002 acl aclname req_mime_type [-i] mime-type ...
1003 # regex match against the mime type of the request generated
1004 # by the client. Can be used to detect file upload or some
1005 # types HTTP tunneling requests [fast]
1006 # NOTE: This does NOT match the reply. You cannot use this
1007 # to match the returned file type.
1009 acl aclname req_header header-name [-i] any\.regex\.here
1010 # regex match against any of the known request headers. May be
1011 # thought of as a superset of "browser", "referer" and "mime-type"
1014 acl aclname rep_mime_type [-i] mime-type ...
1015 # regex match against the mime type of the reply received by
1016 # squid. Can be used to detect file download or some
1017 # types HTTP tunneling requests. [fast]
1018 # NOTE: This has no effect in http_access rules. It only has
1019 # effect in rules that affect the reply data stream such as
1020 # http_reply_access.
1022 acl aclname rep_header header-name [-i] any\.regex\.here
1023 # regex match against any of the known reply headers. May be
1024 # thought of as a superset of "browser", "referer" and "mime-type"
1027 acl aclname external class_name [arguments...]
1028 # external ACL lookup via a helper class defined by the
1029 # external_acl_type directive [slow]
1031 acl aclname user_cert attribute values...
1032 # match against attributes in a user SSL certificate
1033 # attribute is one of DN/C/O/CN/L/ST [fast]
1035 acl aclname ca_cert attribute values...
1036 # match against attributes a users issuing CA SSL certificate
1037 # attribute is one of DN/C/O/CN/L/ST [fast]
1039 acl aclname ext_user username ...
1040 acl aclname ext_user_regex [-i] pattern ...
1041 # string match on username returned by external acl helper [slow]
1042 # use REQUIRED to accept any non-null user name.
1044 acl aclname tag tagvalue ...
1045 # string match on tag returned by external acl helper [slow]
1047 acl aclname hier_code codename ...
1048 # string match against squid hierarchy code(s); [fast]
1049 # e.g., DIRECT, PARENT_HIT, NONE, etc.
1051 # NOTE: This has no effect in http_access rules. It only has
1052 # effect in rules that affect the reply data stream such as
1053 # http_reply_access.
1055 acl aclname note name [value ...]
1056 # match transaction annotation [fast]
1057 # Without values, matches any annotation with a given name.
1058 # With value(s), matches any annotation with a given name that
1059 # also has one of the given values.
1060 # Names and values are compared using a string equality test.
1061 # Annotation sources include note and adaptation_meta directives
1062 # as well as helper and eCAP responses.
1065 acl aclname ssl_error errorname
1066 # match against SSL certificate validation error [fast]
1068 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
1071 # The following can be used as shortcuts for certificate properties:
1072 # [ssl::]certHasExpired: the "not after" field is in the past
1073 # [ssl::]certNotYetValid: the "not before" field is in the future
1074 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
1075 # [ssl::]certSelfSigned: The certificate is self signed.
1076 # [ssl::]certDomainMismatch: The certificate CN domain does not
1077 # match the name the name of the host we are connecting to.
1079 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1080 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1081 # predefined ACLs, just like the 'all' ACL.
1083 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1084 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1086 acl aclname server_cert_fingerprint [-sha1] fingerprint
1087 # match against server SSL certificate fingerprint [fast]
1089 # The fingerprint is the digest of the DER encoded version
1090 # of the whole certificate. The user should use the form: XX:XX:...
1091 # Optional argument specifies the digest algorithm to use.
1092 # The SHA1 digest algorithm is the default and is currently
1093 # the only algorithm supported (-sha1).
1095 acl aclname any-of acl1 acl2 ...
1096 # match any one of the acls [fast or slow]
1097 # The first matching ACL stops further ACL evaluation.
1099 # ACLs from multiple any-of lines with the same name are ORed.
1100 # For example, A = (a1 or a2) or (a3 or a4) can be written as
1101 # acl A any-of a1 a2
1102 # acl A any-of a3 a4
1104 # This group ACL is fast if all evaluated ACLs in the group are fast
1105 # and slow otherwise.
1107 acl aclname all-of acl1 acl2 ...
1108 # match all of the acls [fast or slow]
1109 # The first mismatching ACL stops further ACL evaluation.
1111 # ACLs from multiple all-of lines with the same name are ORed.
1112 # For example, B = (b1 and b2) or (b3 and b4) can be written as
1113 # acl B all-of b1 b2
1114 # acl B all-of b3 b4
1116 # This group ACL is fast if all evaluated ACLs in the group are fast
1117 # and slow otherwise.
1120 acl macaddress arp 09:00:2b:23:45:67
1121 acl myexample dst_as 1241
1122 acl password proxy_auth REQUIRED
1123 acl fileupload req_mime_type -i ^multipart/form-data$
1124 acl javascript rep_mime_type -i ^application/x-javascript$
1128 # Recommended minimum configuration:
1131 # Example rule allowing access from your local networks.
1132 # Adapt to list your (internal) IP networks from where browsing
1134 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1135 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1136 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1137 acl localnet src fc00::/7 # RFC 4193 local private network range
1138 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1140 acl SSL_ports port 443
1141 acl Safe_ports port 80 # http
1142 acl Safe_ports port 21 # ftp
1143 acl Safe_ports port 443 # https
1144 acl Safe_ports port 70 # gopher
1145 acl Safe_ports port 210 # wais
1146 acl Safe_ports port 1025-65535 # unregistered ports
1147 acl Safe_ports port 280 # http-mgmt
1148 acl Safe_ports port 488 # gss-http
1149 acl Safe_ports port 591 # filemaker
1150 acl Safe_ports port 777 # multiling http
1151 acl CONNECT method CONNECT
1155 NAME: follow_x_forwarded_for
1157 IFDEF: FOLLOW_X_FORWARDED_FOR
1158 LOC: Config.accessList.followXFF
1159 DEFAULT_IF_NONE: deny all
1160 DEFAULT_DOC: X-Forwarded-For header will be ignored.
1162 Allowing or Denying the X-Forwarded-For header to be followed to
1163 find the original source of a request.
1165 Requests may pass through a chain of several other proxies
1166 before reaching us. The X-Forwarded-For header will contain a
1167 comma-separated list of the IP addresses in the chain, with the
1168 rightmost address being the most recent.
1170 If a request reaches us from a source that is allowed by this
1171 configuration item, then we consult the X-Forwarded-For header
1172 to see where that host received the request from. If the
1173 X-Forwarded-For header contains multiple addresses, we continue
1174 backtracking until we reach an address for which we are not allowed
1175 to follow the X-Forwarded-For header, or until we reach the first
1176 address in the list. For the purpose of ACL used in the
1177 follow_x_forwarded_for directive the src ACL type always matches
1178 the address we are testing and srcdomain matches its rDNS.
1180 The end result of this process is an IP address that we will
1181 refer to as the indirect client address. This address may
1182 be treated as the client address for access control, ICAP, delay
1183 pools and logging, depending on the acl_uses_indirect_client,
1184 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1185 log_uses_indirect_client and tproxy_uses_indirect_client options.
1187 This clause only supports fast acl types.
1188 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1190 SECURITY CONSIDERATIONS:
1192 Any host for which we follow the X-Forwarded-For header
1193 can place incorrect information in the header, and Squid
1194 will use the incorrect information as if it were the
1195 source address of the request. This may enable remote
1196 hosts to bypass any access control restrictions that are
1197 based on the client's source addresses.
1201 acl localhost src 127.0.0.1
1202 acl my_other_proxy srcdomain .proxy.example.com
1203 follow_x_forwarded_for allow localhost
1204 follow_x_forwarded_for allow my_other_proxy
1207 NAME: acl_uses_indirect_client
1210 IFDEF: FOLLOW_X_FORWARDED_FOR
1212 LOC: Config.onoff.acl_uses_indirect_client
1214 Controls whether the indirect client address
1215 (see follow_x_forwarded_for) is used instead of the
1216 direct client address in acl matching.
1218 NOTE: maxconn ACL considers direct TCP links and indirect
1219 clients will always have zero. So no match.
1222 NAME: delay_pool_uses_indirect_client
1225 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1227 LOC: Config.onoff.delay_pool_uses_indirect_client
1229 Controls whether the indirect client address
1230 (see follow_x_forwarded_for) is used instead of the
1231 direct client address in delay pools.
1234 NAME: log_uses_indirect_client
1237 IFDEF: FOLLOW_X_FORWARDED_FOR
1239 LOC: Config.onoff.log_uses_indirect_client
1241 Controls whether the indirect client address
1242 (see follow_x_forwarded_for) is used instead of the
1243 direct client address in the access log.
1246 NAME: tproxy_uses_indirect_client
1249 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1251 LOC: Config.onoff.tproxy_uses_indirect_client
1253 Controls whether the indirect client address
1254 (see follow_x_forwarded_for) is used instead of the
1255 direct client address when spoofing the outgoing client.
1257 This has no effect on requests arriving in non-tproxy
1260 SECURITY WARNING: Usage of this option is dangerous
1261 and should not be used trivially. Correct configuration
1262 of follow_x_forewarded_for with a limited set of trusted
1263 sources is required to prevent abuse of your proxy.
1266 NAME: spoof_client_ip
1268 LOC: Config.accessList.spoof_client_ip
1270 DEFAULT_DOC: Allow spoofing on all TPROXY traffic.
1272 Control client IP address spoofing of TPROXY traffic based on
1273 defined access lists.
1275 spoof_client_ip allow|deny [!]aclname ...
1277 If there are no "spoof_client_ip" lines present, the default
1278 is to "allow" spoofing of any suitable request.
1280 Note that the cache_peer "no-tproxy" option overrides this ACL.
1282 This clause supports fast acl types.
1283 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1288 LOC: Config.accessList.http
1289 DEFAULT_IF_NONE: deny all
1290 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1292 Allowing or Denying access based on defined access lists
1294 Access to the HTTP port:
1295 http_access allow|deny [!]aclname ...
1297 NOTE on default values:
1299 If there are no "access" lines present, the default is to deny
1302 If none of the "access" lines cause a match, the default is the
1303 opposite of the last line in the list. If the last line was
1304 deny, the default is allow. Conversely, if the last line
1305 is allow, the default will be deny. For these reasons, it is a
1306 good idea to have an "deny all" entry at the end of your access
1307 lists to avoid potential confusion.
1309 This clause supports both fast and slow acl types.
1310 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1315 # Recommended minimum Access Permission configuration:
1317 # Deny requests to certain unsafe ports
1318 http_access deny !Safe_ports
1320 # Deny CONNECT to other than secure SSL ports
1321 http_access deny CONNECT !SSL_ports
1323 # Only allow cachemgr access from localhost
1324 http_access allow localhost manager
1325 http_access deny manager
1327 # We strongly recommend the following be uncommented to protect innocent
1328 # web applications running on the proxy server who think the only
1329 # one who can access services on "localhost" is a local user
1330 #http_access deny to_localhost
1333 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1336 # Example rule allowing access from your local networks.
1337 # Adapt localnet in the ACL section to list your (internal) IP networks
1338 # from where browsing should be allowed
1339 http_access allow localnet
1340 http_access allow localhost
1342 # And finally deny all other access to this proxy
1343 http_access deny all
1347 NAME: adapted_http_access http_access2
1349 LOC: Config.accessList.adapted_http
1351 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1353 Allowing or Denying access based on defined access lists
1355 Essentially identical to http_access, but runs after redirectors
1356 and ICAP/eCAP adaptation. Allowing access control based on their
1359 If not set then only http_access is used.
1362 NAME: http_reply_access
1364 LOC: Config.accessList.reply
1366 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1368 Allow replies to client requests. This is complementary to http_access.
1370 http_reply_access allow|deny [!] aclname ...
1372 NOTE: if there are no access lines present, the default is to allow
1375 If none of the access lines cause a match the opposite of the
1376 last line will apply. Thus it is good practice to end the rules
1377 with an "allow all" or "deny all" entry.
1379 This clause supports both fast and slow acl types.
1380 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1385 LOC: Config.accessList.icp
1387 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1389 Allowing or Denying access to the ICP port based on defined
1392 icp_access allow|deny [!]aclname ...
1394 NOTE: The default if no icp_access lines are present is to
1395 deny all traffic. This default may cause problems with peers
1398 This clause only supports fast acl types.
1399 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1401 # Allow ICP queries from local networks only
1402 #icp_access allow localnet
1403 #icp_access deny all
1409 LOC: Config.accessList.htcp
1411 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1413 Allowing or Denying access to the HTCP port based on defined
1416 htcp_access allow|deny [!]aclname ...
1418 See also htcp_clr_access for details on access control for
1419 cache purge (CLR) HTCP messages.
1421 NOTE: The default if no htcp_access lines are present is to
1422 deny all traffic. This default may cause problems with peers
1423 using the htcp option.
1425 This clause only supports fast acl types.
1426 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1428 # Allow HTCP queries from local networks only
1429 #htcp_access allow localnet
1430 #htcp_access deny all
1433 NAME: htcp_clr_access
1436 LOC: Config.accessList.htcp_clr
1438 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
1440 Allowing or Denying access to purge content using HTCP based
1441 on defined access lists.
1442 See htcp_access for details on general HTCP access control.
1444 htcp_clr_access allow|deny [!]aclname ...
1446 This clause only supports fast acl types.
1447 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1449 # Allow HTCP CLR requests from trusted peers
1450 acl htcp_clr_peer src 192.0.2.2 2001:DB8::2
1451 htcp_clr_access allow htcp_clr_peer
1452 htcp_clr_access deny all
1457 LOC: Config.accessList.miss
1459 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
1461 Determins whether network access is permitted when satisfying a request.
1464 to force your neighbors to use you as a sibling instead of
1467 acl localclients src 192.0.2.0/24 2001:DB8::a:0/64
1468 miss_access deny !localclients
1469 miss_access allow all
1471 This means only your local clients are allowed to fetch relayed/MISS
1472 replies from the network and all other clients can only fetch cached
1475 The default for this setting allows all clients who passed the
1476 http_access rules to relay via this proxy.
1478 This clause only supports fast acl types.
1479 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1482 NAME: ident_lookup_access
1486 DEFAULT_DOC: Unless rules exist in squid.conf, IDENT is not fetched.
1487 LOC: Ident::TheConfig.identLookup
1489 A list of ACL elements which, if matched, cause an ident
1490 (RFC 931) lookup to be performed for this request. For
1491 example, you might choose to always perform ident lookups
1492 for your main multi-user Unix boxes, but not for your Macs
1493 and PCs. By default, ident lookups are not performed for
1496 To enable ident lookups for specific client addresses, you
1497 can follow this example:
1499 acl ident_aware_hosts src 198.168.1.0/24
1500 ident_lookup_access allow ident_aware_hosts
1501 ident_lookup_access deny all
1503 Only src type ACL checks are fully supported. A srcdomain
1504 ACL might work at times, but it will not always provide
1507 This clause only supports fast acl types.
1508 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1511 NAME: reply_body_max_size
1512 COMMENT: size [acl acl...]
1515 DEFAULT_DOC: No limit is applied.
1516 LOC: Config.ReplyBodySize
1518 This option specifies the maximum size of a reply body. It can be
1519 used to prevent users from downloading very large files, such as
1520 MP3's and movies. When the reply headers are received, the
1521 reply_body_max_size lines are processed, and the first line where
1522 all (if any) listed ACLs are true is used as the maximum body size
1525 This size is checked twice. First when we get the reply headers,
1526 we check the content-length value. If the content length value exists
1527 and is larger than the allowed size, the request is denied and the
1528 user receives an error message that says "the request or reply
1529 is too large." If there is no content-length, and the reply
1530 size exceeds this limit, the client's connection is just closed
1531 and they will receive a partial reply.
1533 WARNING: downstream caches probably can not detect a partial reply
1534 if there is no content-length header, so they will cache
1535 partial responses and give them out as hits. You should NOT
1536 use this option if you have downstream caches.
1538 WARNING: A maximum size smaller than the size of squid's error messages
1539 will cause an infinite loop and crash squid. Ensure that the smallest
1540 non-zero value you use is greater that the maximum header size plus
1541 the size of your largest error page.
1543 If you set this parameter none (the default), there will be
1546 Configuration Format is:
1547 reply_body_max_size SIZE UNITS [acl ...]
1549 reply_body_max_size 10 MB
1555 -----------------------------------------------------------------------------
1558 NAME: http_port ascii_port
1561 LOC: Config.Sockaddr.http
1563 Usage: port [mode] [options]
1564 hostname:port [mode] [options]
1565 1.2.3.4:port [mode] [options]
1567 The socket addresses where Squid will listen for HTTP client
1568 requests. You may specify multiple socket addresses.
1569 There are three forms: port alone, hostname with port, and
1570 IP address with port. If you specify a hostname or IP
1571 address, Squid binds the socket to that specific
1572 address. Most likely, you do not need to bind to a specific
1573 address, so you can use the port number alone.
1575 If you are running Squid in accelerator mode, you
1576 probably want to listen on port 80 also, or instead.
1578 The -a command line option may be used to specify additional
1579 port(s) where Squid listens for proxy request. Such ports will
1580 be plain proxy ports with no options.
1582 You may specify multiple socket addresses on multiple lines.
1586 intercept Support for IP-Layer interception of
1587 outgoing requests without browser settings.
1588 NP: disables authentication and IPv6 on the port.
1590 tproxy Support Linux TPROXY for spoofing outgoing
1591 connections using the client IP address.
1592 NP: disables authentication and maybe IPv6 on the port.
1594 accel Accelerator / reverse proxy mode
1596 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1597 establish secure connection with the client and with
1598 the server, decrypt HTTPS messages as they pass through
1599 Squid, and treat them as unencrypted HTTP messages,
1600 becoming the man-in-the-middle.
1602 The ssl_bump option is required to fully enable
1603 bumping of CONNECT requests.
1605 Omitting the mode flag causes default forward proxy mode to be used.
1608 Accelerator Mode Options:
1610 defaultsite=domainname
1611 What to use for the Host: header if it is not present
1612 in a request. Determines what site (not origin server)
1613 accelerators should consider the default.
1615 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1617 protocol= Protocol to reconstruct accelerated requests with.
1618 Defaults to http for http_port and https for
1621 vport Virtual host port support. Using the http_port number
1622 instead of the port passed on Host: headers.
1624 vport=NN Virtual host port support. Using the specified port
1625 number instead of the port passed on Host: headers.
1628 Act as if this Squid is the origin server.
1629 This currently means generate new Date: and Expires:
1630 headers on HIT instead of adding Age:.
1632 ignore-cc Ignore request Cache-Control headers.
1634 WARNING: This option violates HTTP specifications if
1635 used in non-accelerator setups.
1637 allow-direct Allow direct forwarding in accelerator mode. Normally
1638 accelerated requests are denied direct forwarding as if
1639 never_direct was used.
1641 WARNING: this option opens accelerator mode to security
1642 vulnerabilities usually only affecting in interception
1643 mode. Make sure to protect forwarding with suitable
1644 http_access rules when using this.
1647 SSL Bump Mode Options:
1648 In addition to these options ssl-bump requires TLS/SSL options.
1650 generate-host-certificates[=<on|off>]
1651 Dynamically create SSL server certificates for the
1652 destination hosts of bumped CONNECT requests.When
1653 enabled, the cert and key options are used to sign
1654 generated certificates. Otherwise generated
1655 certificate will be selfsigned.
1656 If there is a CA certificate lifetime of the generated
1657 certificate equals lifetime of the CA certificate. If
1658 generated certificate is selfsigned lifetime is three
1660 This option is enabled by default when ssl-bump is used.
1661 See the ssl-bump option above for more information.
1663 dynamic_cert_mem_cache_size=SIZE
1664 Approximate total RAM size spent on cached generated
1665 certificates. If set to zero, caching is disabled. The
1666 default value is 4MB.
1670 cert= Path to SSL certificate (PEM format).
1672 key= Path to SSL private key file (PEM format)
1673 if not specified, the certificate file is
1674 assumed to be a combined certificate and
1677 version= The version of SSL/TLS supported
1678 1 automatic (default)
1685 cipher= Colon separated list of supported ciphers.
1686 NOTE: some ciphers such as EDH ciphers depend on
1687 additional settings. If those settings are
1688 omitted the ciphers may be silently ignored
1689 by the OpenSSL library.
1691 options= Various SSL implementation options. The most important
1693 NO_SSLv2 Disallow the use of SSLv2
1694 NO_SSLv3 Disallow the use of SSLv3
1695 NO_TLSv1 Disallow the use of TLSv1.0
1696 NO_TLSv1_1 Disallow the use of TLSv1.1
1697 NO_TLSv1_2 Disallow the use of TLSv1.2
1698 SINGLE_DH_USE Always create a new key when using
1699 temporary/ephemeral DH key exchanges
1700 ALL Enable various bug workarounds
1701 suggested as "harmless" by OpenSSL
1702 Be warned that this reduces SSL/TLS
1703 strength to some attacks.
1704 See OpenSSL SSL_CTX_set_options documentation for a
1705 complete list of options.
1707 clientca= File containing the list of CAs to use when
1708 requesting a client certificate.
1710 cafile= File containing additional CA certificates to
1711 use when verifying client certificates. If unset
1712 clientca will be used.
1714 capath= Directory containing additional CA certificates
1715 and CRL lists to use when verifying client certificates.
1717 crlfile= File of additional CRL lists to use when verifying
1718 the client certificate, in addition to CRLs stored in
1719 the capath. Implies VERIFY_CRL flag below.
1721 dhparams= File containing DH parameters for temporary/ephemeral
1722 DH key exchanges. See OpenSSL documentation for details
1723 on how to create this file.
1724 WARNING: EDH ciphers will be silently disabled if this
1727 sslflags= Various flags modifying the use of SSL:
1729 Don't request client certificates
1730 immediately, but wait until acl processing
1731 requires a certificate (not yet implemented).
1733 Don't use the default CA lists built in
1736 Don't allow for session reuse. Each connection
1737 will result in a new SSL session.
1739 Verify CRL lists when accepting client
1742 Verify CRL lists for all certificates in the
1743 client certificate chain.
1745 sslcontext= SSL session ID context identifier.
1749 connection-auth[=on|off]
1750 use connection-auth=off to tell Squid to prevent
1751 forwarding Microsoft connection oriented authentication
1752 (NTLM, Negotiate and Kerberos)
1754 disable-pmtu-discovery=
1755 Control Path-MTU discovery usage:
1756 off lets OS decide on what to do (default).
1757 transparent disable PMTU discovery when transparent
1759 always disable always PMTU discovery.
1761 In many setups of transparently intercepting proxies
1762 Path-MTU discovery can not work on traffic towards the
1763 clients. This is the case when the intercepting device
1764 does not fully track connections and fails to forward
1765 ICMP must fragment messages to the cache server. If you
1766 have such setup and experience that certain clients
1767 sporadically hang or never complete requests set
1768 disable-pmtu-discovery option to 'transparent'.
1770 name= Specifies a internal name for the port. Defaults to
1771 the port specification (port or addr:port)
1773 tcpkeepalive[=idle,interval,timeout]
1774 Enable TCP keepalive probes of idle connections.
1775 In seconds; idle is the initial time before TCP starts
1776 probing the connection, interval how often to probe, and
1777 timeout the time before giving up.
1779 If you run Squid on a dual-homed machine with an internal
1780 and an external interface we recommend you to specify the
1781 internal address:port in http_port. This way Squid will only be
1782 visible on the internal address.
1786 # Squid normally listens to port 3128
1787 http_port @DEFAULT_HTTP_PORT@
1795 LOC: Config.Sockaddr.https
1797 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1799 The socket address where Squid will listen for client requests made
1800 over TLS or SSL connections. Commonly referred to as HTTPS.
1802 This is most useful for situations where you are running squid in
1803 accelerator mode and you want to do the SSL work at the accelerator level.
1805 You may specify multiple socket addresses on multiple lines,
1806 each with their own SSL certificate and/or options.
1810 accel Accelerator / reverse proxy mode
1812 intercept Support for IP-Layer interception of
1813 outgoing requests without browser settings.
1814 NP: disables authentication and IPv6 on the port.
1816 tproxy Support Linux TPROXY for spoofing outgoing
1817 connections using the client IP address.
1818 NP: disables authentication and maybe IPv6 on the port.
1820 ssl-bump For each intercepted connection allowed by ssl_bump
1821 ACLs, establish a secure connection with the client and with
1822 the server, decrypt HTTPS messages as they pass through
1823 Squid, and treat them as unencrypted HTTP messages,
1824 becoming the man-in-the-middle.
1826 An "ssl_bump server-first" match is required to
1827 fully enable bumping of intercepted SSL connections.
1829 Requires tproxy or intercept.
1831 Omitting the mode flag causes default forward proxy mode to be used.
1834 See http_port for a list of generic options
1839 cert= Path to SSL certificate (PEM format).
1841 key= Path to SSL private key file (PEM format)
1842 if not specified, the certificate file is
1843 assumed to be a combined certificate and
1846 version= The version of SSL/TLS supported
1847 1 automatic (default)
1852 cipher= Colon separated list of supported ciphers.
1854 options= Various SSL engine options. The most important
1856 NO_SSLv2 Disallow the use of SSLv2
1857 NO_SSLv3 Disallow the use of SSLv3
1858 NO_TLSv1 Disallow the use of TLSv1
1859 SINGLE_DH_USE Always create a new key when using
1860 temporary/ephemeral DH key exchanges
1861 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1862 documentation for a complete list of options.
1864 clientca= File containing the list of CAs to use when
1865 requesting a client certificate.
1867 cafile= File containing additional CA certificates to
1868 use when verifying client certificates. If unset
1869 clientca will be used.
1871 capath= Directory containing additional CA certificates
1872 and CRL lists to use when verifying client certificates.
1874 crlfile= File of additional CRL lists to use when verifying
1875 the client certificate, in addition to CRLs stored in
1876 the capath. Implies VERIFY_CRL flag below.
1878 dhparams= File containing DH parameters for temporary/ephemeral
1881 sslflags= Various flags modifying the use of SSL:
1883 Don't request client certificates
1884 immediately, but wait until acl processing
1885 requires a certificate (not yet implemented).
1887 Don't use the default CA lists built in
1890 Don't allow for session reuse. Each connection
1891 will result in a new SSL session.
1893 Verify CRL lists when accepting client
1896 Verify CRL lists for all certificates in the
1897 client certificate chain.
1899 sslcontext= SSL session ID context identifier.
1901 generate-host-certificates[=<on|off>]
1902 Dynamically create SSL server certificates for the
1903 destination hosts of bumped SSL requests.When
1904 enabled, the cert and key options are used to sign
1905 generated certificates. Otherwise generated
1906 certificate will be selfsigned.
1907 If there is CA certificate life time of generated
1908 certificate equals lifetime of CA certificate. If
1909 generated certificate is selfsigned lifetime is three
1911 This option is enabled by default when SslBump is used.
1912 See the sslBump option above for more information.
1914 dynamic_cert_mem_cache_size=SIZE
1915 Approximate total RAM size spent on cached generated
1916 certificates. If set to zero, caching is disabled. The
1917 default value is 4MB.
1919 See http_port for a list of available options.
1922 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1925 LOC: Ip::Qos::TheConfig.tosToServer
1927 Allows you to select a TOS/Diffserv value for packets outgoing
1928 on the server side, based on an ACL.
1930 tcp_outgoing_tos ds-field [!]aclname ...
1932 Example where normal_service_net uses the TOS value 0x00
1933 and good_service_net uses 0x20
1935 acl normal_service_net src 10.0.0.0/24
1936 acl good_service_net src 10.0.1.0/24
1937 tcp_outgoing_tos 0x00 normal_service_net
1938 tcp_outgoing_tos 0x20 good_service_net
1940 TOS/DSCP values really only have local significance - so you should
1941 know what you're specifying. For more information, see RFC2474,
1942 RFC2475, and RFC3260.
1944 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1945 "default" to use whatever default your host has. Note that in
1946 practice often only multiples of 4 is usable as the two rightmost bits
1947 have been redefined for use by ECN (RFC 3168 section 23.1).
1949 Processing proceeds in the order specified, and stops at first fully
1953 NAME: clientside_tos
1956 LOC: Ip::Qos::TheConfig.tosToClient
1958 Allows you to select a TOS/Diffserv value for packets being transmitted
1959 on the client-side, based on an ACL.
1961 clientside_tos ds-field [!]aclname ...
1963 Example where normal_service_net uses the TOS value 0x00
1964 and good_service_net uses 0x20
1966 acl normal_service_net src 10.0.0.0/24
1967 acl good_service_net src 10.0.1.0/24
1968 clientside_tos 0x00 normal_service_net
1969 clientside_tos 0x20 good_service_net
1971 Note: This feature is incompatible with qos_flows. Any TOS values set here
1972 will be overwritten by TOS values in qos_flows.
1975 NAME: tcp_outgoing_mark
1977 IFDEF: SO_MARK&&USE_LIBCAP
1979 LOC: Ip::Qos::TheConfig.nfmarkToServer
1981 Allows you to apply a Netfilter mark value to outgoing packets
1982 on the server side, based on an ACL.
1984 tcp_outgoing_mark mark-value [!]aclname ...
1986 Example where normal_service_net uses the mark value 0x00
1987 and good_service_net uses 0x20
1989 acl normal_service_net src 10.0.0.0/24
1990 acl good_service_net src 10.0.1.0/24
1991 tcp_outgoing_mark 0x00 normal_service_net
1992 tcp_outgoing_mark 0x20 good_service_net
1995 NAME: clientside_mark
1997 IFDEF: SO_MARK&&USE_LIBCAP
1999 LOC: Ip::Qos::TheConfig.nfmarkToClient
2001 Allows you to apply a Netfilter mark value to packets being transmitted
2002 on the client-side, based on an ACL.
2004 clientside_mark mark-value [!]aclname ...
2006 Example where normal_service_net uses the mark value 0x00
2007 and good_service_net uses 0x20
2009 acl normal_service_net src 10.0.0.0/24
2010 acl good_service_net src 10.0.1.0/24
2011 clientside_mark 0x00 normal_service_net
2012 clientside_mark 0x20 good_service_net
2014 Note: This feature is incompatible with qos_flows. Any mark values set here
2015 will be overwritten by mark values in qos_flows.
2022 LOC: Ip::Qos::TheConfig
2024 Allows you to select a TOS/DSCP value to mark outgoing
2025 connections with, based on where the reply was sourced. For
2026 platforms using netfilter, allows you to set a netfilter mark
2027 value instead of, or in addition to, a TOS value.
2029 TOS values really only have local significance - so you should
2030 know what you're specifying. For more information, see RFC2474,
2031 RFC2475, and RFC3260.
2033 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
2034 in practice often only multiples of 4 is usable as the two rightmost bits
2035 have been redefined for use by ECN (RFC 3168 section 23.1).
2037 Mark values can be any unsigned 32-bit integer value.
2039 This setting is configured by setting the following values:
2041 tos|mark Whether to set TOS or netfilter mark values
2043 local-hit=0xFF Value to mark local cache hits.
2045 sibling-hit=0xFF Value to mark hits from sibling peers.
2047 parent-hit=0xFF Value to mark hits from parent peers.
2049 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
2050 over the preserve-miss feature (see below), unless
2051 mask is specified, in which case only the bits
2052 specified in the mask are written.
2054 The TOS variant of the following features are only possible on Linux
2055 and require your kernel to be patched with the TOS preserving ZPH
2056 patch, available from http://zph.bratcheda.org
2057 No patch is needed to preserve the netfilter mark, which will work
2058 with all variants of netfilter.
2060 disable-preserve-miss
2061 This option disables the preservation of the TOS or netfilter
2062 mark. By default, the existing TOS or netfilter mark value of
2063 the response coming from the remote server will be retained
2064 and masked with miss-mark.
2065 NOTE: in the case of a netfilter mark, the mark must be set on
2066 the connection (using the CONNMARK target) not on the packet
2070 Allows you to mask certain bits in the TOS or mark value
2071 received from the remote server, before copying the value to
2072 the TOS sent towards clients.
2073 Default for tos: 0xFF (TOS from server is not changed).
2074 Default for mark: 0xFFFFFFFF (mark from server is not changed).
2076 All of these features require the --enable-zph-qos compilation flag
2077 (enabled by default). Netfilter marking also requires the
2078 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
2079 libcap 2.09+ (--with-libcap).
2083 NAME: tcp_outgoing_address
2086 DEFAULT_DOC: Address selection is performed by the operating system.
2087 LOC: Config.accessList.outgoing_address
2089 Allows you to map requests to different outgoing IP addresses
2090 based on the username or source address of the user making
2093 tcp_outgoing_address ipaddr [[!]aclname] ...
2096 Forwarding clients with dedicated IPs for certain subnets.
2098 acl normal_service_net src 10.0.0.0/24
2099 acl good_service_net src 10.0.2.0/24
2101 tcp_outgoing_address 2001:db8::c001 good_service_net
2102 tcp_outgoing_address 10.1.0.2 good_service_net
2104 tcp_outgoing_address 2001:db8::beef normal_service_net
2105 tcp_outgoing_address 10.1.0.1 normal_service_net
2107 tcp_outgoing_address 2001:db8::1
2108 tcp_outgoing_address 10.1.0.3
2110 Processing proceeds in the order specified, and stops at first fully
2113 Squid will add an implicit IP version test to each line.
2114 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
2115 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
2118 NOTE: The use of this directive using client dependent ACLs is
2119 incompatible with the use of server side persistent connections. To
2120 ensure correct results it is best to set server_persistent_connections
2121 to off when using this directive in such configurations.
2123 NOTE: The use of this directive to set a local IP on outgoing TCP links
2124 is incompatible with using TPROXY to set client IP out outbound TCP links.
2125 When needing to contact peers use the no-tproxy cache_peer option and the
2126 client_dst_passthru directive re-enable normal forwarding such as this.
2130 NAME: host_verify_strict
2133 LOC: Config.onoff.hostStrictVerify
2135 Regardless of this option setting, when dealing with intercepted
2136 traffic, Squid always verifies that the destination IP address matches
2137 the Host header domain or IP (called 'authority form URL').
2139 This enforcement is performed to satisfy a MUST-level requirement in
2140 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2141 authority of the origin server or gateway given by the original URL".
2144 Squid always responds with an HTTP 409 (Conflict) error
2145 page and logs a security warning if there is no match.
2147 Squid verifies that the destination IP address matches
2148 the Host header for forward-proxy and reverse-proxy traffic
2149 as well. For those traffic types, Squid also enables the
2150 following checks, comparing the corresponding Host header
2151 and Request-URI components:
2153 * The host names (domain or IP) must be identical,
2154 but valueless or missing Host header disables all checks.
2155 For the two host names to match, both must be either IP
2158 * Port numbers must be identical, but if a port is missing
2159 the scheme-default port is assumed.
2162 When set to OFF (the default):
2163 Squid allows suspicious requests to continue but logs a
2164 security warning and blocks caching of the response.
2166 * Forward-proxy traffic is not checked at all.
2168 * Reverse-proxy traffic is not checked at all.
2170 * Intercepted traffic which passes verification is handled
2171 according to client_dst_passthru.
2173 * Intercepted requests which fail verification are sent
2174 to the client original destination instead of DIRECT.
2175 This overrides 'client_dst_passthru off'.
2177 For now suspicious intercepted CONNECT requests are always
2178 responded to with an HTTP 409 (Conflict) error page.
2183 As described in CVE-2009-0801 when the Host: header alone is used
2184 to determine the destination of a request it becomes trivial for
2185 malicious scripts on remote websites to bypass browser same-origin
2186 security policy and sandboxing protections.
2188 The cause of this is that such applets are allowed to perform their
2189 own HTTP stack, in which case the same-origin policy of the browser
2190 sandbox only verifies that the applet tries to contact the same IP
2191 as from where it was loaded at the IP level. The Host: header may
2192 be different from the connected IP and approved origin.
2196 NAME: client_dst_passthru
2199 LOC: Config.onoff.client_dst_passthru
2201 With NAT or TPROXY intercepted traffic Squid may pass the request
2202 directly to the original client destination IP or seek a faster
2203 source using the HTTP Host header.
2205 Using Host to locate alternative servers can provide faster
2206 connectivity with a range of failure recovery options.
2207 But can also lead to connectivity trouble when the client and
2208 server are attempting stateful interactions unaware of the proxy.
2210 This option (on by default) prevents alternative DNS entries being
2211 located to send intercepted traffic DIRECT to an origin server.
2212 The clients original destination IP and port will be used instead.
2214 Regardless of this option setting, when dealing with intercepted
2215 traffic Squid will verify the Host: header and any traffic which
2216 fails Host verification will be treated as if this option were ON.
2218 see host_verify_strict for details on the verification process.
2223 -----------------------------------------------------------------------------
2226 NAME: ssl_unclean_shutdown
2230 LOC: Config.SSL.unclean_shutdown
2232 Some browsers (especially MSIE) bugs out on SSL shutdown
2239 LOC: Config.SSL.ssl_engine
2242 The OpenSSL engine to use. You will need to set this if you
2243 would like to use hardware SSL acceleration for example.
2246 NAME: sslproxy_client_certificate
2249 LOC: Config.ssl_client.cert
2252 Client SSL Certificate to use when proxying https:// URLs
2255 NAME: sslproxy_client_key
2258 LOC: Config.ssl_client.key
2261 Client SSL Key to use when proxying https:// URLs
2264 NAME: sslproxy_version
2267 DEFAULT_DOC: automatic SSL/TLS version negotiation
2268 LOC: Config.ssl_client.version
2271 SSL version level to use when proxying https:// URLs
2273 The versions of SSL/TLS supported:
2275 1 automatic (default)
2283 NAME: sslproxy_options
2286 LOC: Config.ssl_client.options
2289 SSL implementation options to use when proxying https:// URLs
2291 The most important being:
2293 NO_SSLv2 Disallow the use of SSLv2
2294 NO_SSLv3 Disallow the use of SSLv3
2295 NO_TLSv1 Disallow the use of TLSv1.0
2296 NO_TLSv1_1 Disallow the use of TLSv1.1
2297 NO_TLSv1_2 Disallow the use of TLSv1.2
2299 Always create a new key when using temporary/ephemeral
2302 Disable use of RFC5077 session tickets. Some servers
2303 may have problems understanding the TLS extension due
2304 to ambiguous specification in RFC4507.
2305 ALL Enable various bug workarounds suggested as "harmless"
2306 by OpenSSL. Be warned that this may reduce SSL/TLS
2307 strength to some attacks.
2309 See the OpenSSL SSL_CTX_set_options documentation for a
2310 complete list of possible options.
2313 NAME: sslproxy_cipher
2316 LOC: Config.ssl_client.cipher
2319 SSL cipher list to use when proxying https:// URLs
2321 Colon separated list of supported ciphers.
2324 NAME: sslproxy_cafile
2327 LOC: Config.ssl_client.cafile
2330 file containing CA certificates to use when verifying server
2331 certificates while proxying https:// URLs
2334 NAME: sslproxy_capath
2337 LOC: Config.ssl_client.capath
2340 directory containing CA certificates to use when verifying
2341 server certificates while proxying https:// URLs
2346 TYPE: sslproxy_ssl_bump
2347 LOC: Config.accessList.ssl_bump
2348 DEFAULT_DOC: Does not bump unless rules are present in squid.conf
2351 This option is consulted when a CONNECT request is received on
2352 an http_port (or a new connection is intercepted at an
2353 https_port), provided that port was configured with an ssl-bump
2354 flag. The subsequent data on the connection is either treated as
2355 HTTPS and decrypted OR tunneled at TCP level without decryption,
2356 depending on the first bumping "mode" which ACLs match.
2358 ssl_bump <mode> [!]acl ...
2360 The following bumping modes are supported:
2363 Allow bumping of the connection. Establish a secure connection
2364 with the client first, then connect to the server. This old mode
2365 does not allow Squid to mimic server SSL certificate and does
2366 not work with intercepted SSL connections.
2369 Allow bumping of the connection. Establish a secure connection
2370 with the server first, then establish a secure connection with
2371 the client, using a mimicked server certificate. Works with both
2372 CONNECT requests and intercepted SSL connections.
2375 Become a TCP tunnel without decoding the connection.
2376 Works with both CONNECT requests and intercepted SSL
2377 connections. This is the default behavior when no
2378 ssl_bump option is given or no ssl_bump ACLs match.
2380 By default, no connections are bumped.
2382 The first matching ssl_bump option wins. If no ACLs match, the
2383 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2384 does not have an implicit "negate the last given option" rule. You
2385 must make that rule explicit if you convert old ssl_bump allow/deny
2386 rules that rely on such an implicit rule.
2388 This clause supports both fast and slow acl types.
2389 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2391 See also: http_port ssl-bump, https_port ssl-bump
2394 # Example: Bump all requests except those originating from
2395 # localhost or those going to example.com.
2397 acl broken_sites dstdomain .example.com
2398 ssl_bump none localhost
2399 ssl_bump none broken_sites
2400 ssl_bump server-first all
2403 NAME: sslproxy_flags
2406 LOC: Config.ssl_client.flags
2409 Various flags modifying the use of SSL while proxying https:// URLs:
2410 DONT_VERIFY_PEER Accept certificates that fail verification.
2411 For refined control, see sslproxy_cert_error.
2412 NO_DEFAULT_CA Don't use the default CA list built in
2416 NAME: sslproxy_cert_error
2419 DEFAULT_DOC: Server certificate errors terminate the transaction.
2420 LOC: Config.ssl_client.cert_error
2423 Use this ACL to bypass server certificate validation errors.
2425 For example, the following lines will bypass all validation errors
2426 when talking to servers for example.com. All other
2427 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2429 acl BrokenButTrustedServers dstdomain example.com
2430 sslproxy_cert_error allow BrokenButTrustedServers
2431 sslproxy_cert_error deny all
2433 This clause only supports fast acl types.
2434 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2435 Using slow acl types may result in server crashes
2437 Without this option, all server certificate validation errors
2438 terminate the transaction to protect Squid and the client.
2441 Bypassing validation errors is dangerous because an
2442 error usually implies that the server cannot be trusted
2443 and the connection may be insecure.
2445 See also: sslproxy_flags and DONT_VERIFY_PEER.
2448 NAME: sslproxy_cert_sign
2451 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2452 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2453 POSTSCRIPTUM: signTrusted all
2454 TYPE: sslproxy_cert_sign
2455 LOC: Config.ssl_client.cert_sign
2458 sslproxy_cert_sign <signing algorithm> acl ...
2460 The following certificate signing algorithms are supported:
2463 Sign using the configured CA certificate which is usually
2464 placed in and trusted by end-user browsers. This is the
2465 default for trusted origin server certificates.
2468 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2469 This is the default for untrusted origin server certificates
2470 that are not self-signed (see ssl::certUntrusted).
2473 Sign using a self-signed certificate with the right CN to
2474 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2475 browser. This is the default for self-signed origin server
2476 certificates (see ssl::certSelfSigned).
2478 This clause only supports fast acl types.
2480 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2481 signing algorithm to generate the certificate and ignores all
2482 subsequent sslproxy_cert_sign options (the first match wins). If no
2483 acl(s) match, the default signing algorithm is determined by errors
2484 detected when obtaining and validating the origin server certificate.
2486 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2487 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2488 CONNECT request that carries a domain name. In all other cases (CONNECT
2489 to an IP address or an intercepted SSL connection), Squid cannot detect
2490 the domain mismatch at certificate generation time when
2491 bump-server-first is used.
2494 NAME: sslproxy_cert_adapt
2497 TYPE: sslproxy_cert_adapt
2498 LOC: Config.ssl_client.cert_adapt
2501 sslproxy_cert_adapt <adaptation algorithm> acl ...
2503 The following certificate adaptation algorithms are supported:
2506 Sets the "Not After" property to the "Not After" property of
2507 the CA certificate used to sign generated certificates.
2510 Sets the "Not Before" property to the "Not Before" property of
2511 the CA certificate used to sign generated certificates.
2513 setCommonName or setCommonName{CN}
2514 Sets Subject.CN property to the host name specified as a
2515 CN parameter or, if no explicit CN parameter was specified,
2516 extracted from the CONNECT request. It is a misconfiguration
2517 to use setCommonName without an explicit parameter for
2518 intercepted or tproxied SSL connections.
2520 This clause only supports fast acl types.
2522 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2523 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2524 corresponding adaptation algorithm to generate the certificate and
2525 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2526 group (i.e., the first match wins within each algorithm group). If no
2527 acl(s) match, the default mimicking action takes place.
2529 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2530 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2531 CONNECT request that carries a domain name. In all other cases (CONNECT
2532 to an IP address or an intercepted SSL connection), Squid cannot detect
2533 the domain mismatch at certificate generation time when
2534 bump-server-first is used.
2537 NAME: sslpassword_program
2540 LOC: Config.Program.ssl_password
2543 Specify a program used for entering SSL key passphrases
2544 when using encrypted SSL certificate keys. If not specified
2545 keys must either be unencrypted, or Squid started with the -N
2546 option to allow it to query interactively for the passphrase.
2548 The key file name is given as argument to the program allowing
2549 selection of the right password if you have multiple encrypted
2554 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2555 -----------------------------------------------------------------------------
2558 NAME: sslcrtd_program
2561 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2562 LOC: Ssl::TheConfig.ssl_crtd
2564 Specify the location and options of the executable for ssl_crtd process.
2565 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2566 For more information use:
2567 @DEFAULT_SSL_CRTD@ -h
2570 NAME: sslcrtd_children
2571 TYPE: HelperChildConfig
2573 DEFAULT: 32 startup=5 idle=1
2574 LOC: Ssl::TheConfig.ssl_crtdChildren
2576 The maximum number of processes spawn to service ssl server.
2577 The maximum this may be safely set to is 32.
2579 The startup= and idle= options allow some measure of skew in your
2584 Sets the minimum number of processes to spawn when Squid
2585 starts or reconfigures. When set to zero the first request will
2586 cause spawning of the first child process to handle it.
2588 Starting too few children temporary slows Squid under load while it
2589 tries to spawn enough additional processes to cope with traffic.
2593 Sets a minimum of how many processes Squid is to try and keep available
2594 at all times. When traffic begins to rise above what the existing
2595 processes can handle this many more will be spawned up to the maximum
2596 configured. A minimum setting of 1 is required.
2598 You must have at least one ssl_crtd process.
2601 NAME: sslcrtvalidator_program
2605 LOC: Ssl::TheConfig.ssl_crt_validator
2607 Specify the location and options of the executable for ssl_crt_validator
2610 Usage: sslcrtvalidator_program [ttl=n] [cache=n] path ...
2613 ttl=n TTL in seconds for cached results. The default is 60 secs
2614 cache=n limit the result cache size. The default value is 2048
2617 NAME: sslcrtvalidator_children
2618 TYPE: HelperChildConfig
2620 DEFAULT: 32 startup=5 idle=1 concurrency=1
2621 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2623 The maximum number of processes spawn to service SSL server.
2624 The maximum this may be safely set to is 32.
2626 The startup= and idle= options allow some measure of skew in your
2631 Sets the minimum number of processes to spawn when Squid
2632 starts or reconfigures. When set to zero the first request will
2633 cause spawning of the first child process to handle it.
2635 Starting too few children temporary slows Squid under load while it
2636 tries to spawn enough additional processes to cope with traffic.
2640 Sets a minimum of how many processes Squid is to try and keep available
2641 at all times. When traffic begins to rise above what the existing
2642 processes can handle this many more will be spawned up to the maximum
2643 configured. A minimum setting of 1 is required.
2647 The number of requests each certificate validator helper can handle in
2648 parallel. Defaults to 0 which indicates the certficate validator
2649 is a old-style single threaded redirector.
2651 When this directive is set to a value >= 1 then the protocol
2652 used to communicate with the helper is modified to include
2653 a request ID in front of the request/response. The request
2654 ID from the request must be echoed back with the response
2657 You must have at least one ssl_crt_validator process.
2661 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2662 -----------------------------------------------------------------------------
2670 To specify other caches in a hierarchy, use the format:
2672 cache_peer hostname type http-port icp-port [options]
2677 # hostname type port port options
2678 # -------------------- -------- ----- ----- -----------
2679 cache_peer parent.foo.net parent 3128 3130 default
2680 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2681 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2682 cache_peer example.com parent 80 0 default
2683 cache_peer cdn.example.com sibling 3128 0
2685 type: either 'parent', 'sibling', or 'multicast'.
2687 proxy-port: The port number where the peer accept HTTP requests.
2688 For other Squid proxies this is usually 3128
2689 For web servers this is usually 80
2691 icp-port: Used for querying neighbor caches about objects.
2692 Set to 0 if the peer does not support ICP or HTCP.
2693 See ICP and HTCP options below for additional details.
2696 ==== ICP OPTIONS ====
2698 You MUST also set icp_port and icp_access explicitly when using these options.
2699 The defaults will prevent peer traffic using ICP.
2702 no-query Disable ICP queries to this neighbor.
2705 Indicates the named peer is a member of a multicast group.
2706 ICP queries will not be sent directly to the peer, but ICP
2707 replies will be accepted from it.
2709 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2710 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2713 To only send ICP queries to this neighbor infrequently.
2714 This is used to keep the neighbor round trip time updated
2715 and is usually used in conjunction with weighted-round-robin.
2718 ==== HTCP OPTIONS ====
2720 You MUST also set htcp_port and htcp_access explicitly when using these options.
2721 The defaults will prevent peer traffic using HTCP.
2724 htcp Send HTCP, instead of ICP, queries to the neighbor.
2725 You probably also want to set the "icp-port" to 4827
2726 instead of 3130. This directive accepts a comma separated
2727 list of options described below.
2729 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2731 htcp=no-clr Send HTCP to the neighbor but without
2732 sending any CLR requests. This cannot be used with
2735 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2736 This cannot be used with no-clr.
2739 Send HTCP to the neighbor including CLRs but only when
2740 they do not result from PURGE requests.
2743 Forward any HTCP CLR requests this proxy receives to the peer.
2746 ==== PEER SELECTION METHODS ====
2748 The default peer selection method is ICP, with the first responding peer
2749 being used as source. These options can be used for better load balancing.
2752 default This is a parent cache which can be used as a "last-resort"
2753 if a peer cannot be located by any of the peer-selection methods.
2754 If specified more than once, only the first is used.
2756 round-robin Load-Balance parents which should be used in a round-robin
2757 fashion in the absence of any ICP queries.
2758 weight=N can be used to add bias.
2760 weighted-round-robin
2761 Load-Balance parents which should be used in a round-robin
2762 fashion with the frequency of each parent being based on the
2763 round trip time. Closer parents are used more often.
2764 Usually used for background-ping parents.
2765 weight=N can be used to add bias.
2767 carp Load-Balance parents which should be used as a CARP array.
2768 The requests will be distributed among the parents based on the
2769 CARP load balancing hash function based on their weight.
2771 userhash Load-balance parents based on the client proxy_auth or ident username.
2773 sourcehash Load-balance parents based on the client source IP.
2776 To be used only for cache peers of type "multicast".
2777 ALL members of this multicast group have "sibling"
2778 relationship with it, not "parent". This is to a multicast
2779 group when the requested object would be fetched only from
2780 a "parent" cache, anyway. It's useful, e.g., when
2781 configuring a pool of redundant Squid proxies, being
2782 members of the same multicast group.
2785 ==== PEER SELECTION OPTIONS ====
2787 weight=N use to affect the selection of a peer during any weighted
2788 peer-selection mechanisms.
2789 The weight must be an integer; default is 1,
2790 larger weights are favored more.
2791 This option does not affect parent selection if a peering
2792 protocol is not in use.
2794 basetime=N Specify a base amount to be subtracted from round trip
2796 It is subtracted before division by weight in calculating
2797 which parent to fectch from. If the rtt is less than the
2798 base time the rtt is set to a minimal value.
2800 ttl=N Specify a TTL to use when sending multicast ICP queries
2802 Only useful when sending to a multicast group.
2803 Because we don't accept ICP replies from random
2804 hosts, you must configure other group members as
2805 peers with the 'multicast-responder' option.
2807 no-delay To prevent access to this neighbor from influencing the
2810 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2811 enabled) for this host from the specified URL rather
2812 than the Squid default location.
2815 ==== CARP OPTIONS ====
2817 carp-key=key-specification
2818 use a different key than the full URL to hash against the peer.
2819 the key-specification is a comma-separated list of the keywords
2820 scheme, host, port, path, params
2821 Order is not important.
2823 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2825 originserver Causes this parent to be contacted as an origin server.
2826 Meant to be used in accelerator setups when the peer
2830 Set the Host header of requests forwarded to this peer.
2831 Useful in accelerator setups where the server (peer)
2832 expects a certain domain name but clients may request
2833 others. ie example.com or www.example.com
2835 no-digest Disable request of cache digests.
2838 Disables requesting ICMP RTT database (NetDB).
2841 ==== AUTHENTICATION OPTIONS ====
2844 If this is a personal/workgroup proxy and your parent
2845 requires proxy authentication.
2847 Note: The string can include URL escapes (i.e. %20 for
2848 spaces). This also means % must be written as %%.
2851 Send login details received from client to this peer.
2852 Both Proxy- and WWW-Authorization headers are passed
2853 without alteration to the peer.
2854 Authentication is not required by Squid for this to work.
2856 Note: This will pass any form of authentication but
2857 only Basic auth will work through a proxy unless the
2858 connection-auth options are also used.
2860 login=PASS Send login details received from client to this peer.
2861 Authentication is not required by this option.
2863 If there are no client-provided authentication headers
2864 to pass on, but username and password are available
2865 from an external ACL user= and password= result tags
2866 they may be sent instead.
2868 Note: To combine this with proxy_auth both proxies must
2869 share the same user database as HTTP only allows for
2870 a single login (one for proxy, one for origin server).
2871 Also be warned this will expose your users proxy
2872 password to the peer. USE WITH CAUTION
2875 Send the username to the upstream cache, but with a
2876 fixed password. This is meant to be used when the peer
2877 is in another administrative domain, but it is still
2878 needed to identify each user.
2879 The star can optionally be followed by some extra
2880 information which is added to the username. This can
2881 be used to identify this proxy to the peer, similar to
2882 the login=username:password option above.
2885 If this is a personal/workgroup proxy and your parent
2886 requires a secure proxy authentication.
2887 The first principal from the default keytab or defined by
2888 the environment variable KRB5_KTNAME will be used.
2890 WARNING: The connection may transmit requests from multiple
2891 clients. Negotiate often assumes end-to-end authentication
2892 and a single-client. Which is not strictly true here.
2894 login=NEGOTIATE:principal_name
2895 If this is a personal/workgroup proxy and your parent
2896 requires a secure proxy authentication.
2897 The principal principal_name from the default keytab or
2898 defined by the environment variable KRB5_KTNAME will be
2901 WARNING: The connection may transmit requests from multiple
2902 clients. Negotiate often assumes end-to-end authentication
2903 and a single-client. Which is not strictly true here.
2905 connection-auth=on|off
2906 Tell Squid that this peer does or not support Microsoft
2907 connection oriented authentication, and any such
2908 challenges received from there should be ignored.
2909 Default is auto to automatically determine the status
2913 ==== SSL / HTTPS / TLS OPTIONS ====
2915 ssl Encrypt connections to this peer with SSL/TLS.
2917 sslcert=/path/to/ssl/certificate
2918 A client SSL certificate to use when connecting to
2921 sslkey=/path/to/ssl/key
2922 The private SSL key corresponding to sslcert above.
2923 If 'sslkey' is not specified 'sslcert' is assumed to
2924 reference a combined file containing both the
2925 certificate and the key.
2927 sslversion=1|2|3|4|5|6
2928 The SSL version to use when connecting to this peer
2929 1 = automatic (default)
2936 sslcipher=... The list of valid SSL ciphers to use when connecting
2939 ssloptions=... Specify various SSL implementation options:
2941 NO_SSLv2 Disallow the use of SSLv2
2942 NO_SSLv3 Disallow the use of SSLv3
2943 NO_TLSv1 Disallow the use of TLSv1.0
2944 NO_TLSv1_1 Disallow the use of TLSv1.1
2945 NO_TLSv1_2 Disallow the use of TLSv1.2
2947 Always create a new key when using
2948 temporary/ephemeral DH key exchanges
2949 ALL Enable various bug workarounds
2950 suggested as "harmless" by OpenSSL
2951 Be warned that this reduces SSL/TLS
2952 strength to some attacks.
2954 See the OpenSSL SSL_CTX_set_options documentation for a
2957 sslcafile=... A file containing additional CA certificates to use
2958 when verifying the peer certificate.
2960 sslcapath=... A directory containing additional CA certificates to
2961 use when verifying the peer certificate.
2963 sslcrlfile=... A certificate revocation list file to use when
2964 verifying the peer certificate.
2966 sslflags=... Specify various flags modifying the SSL implementation:
2969 Accept certificates even if they fail to
2972 Don't use the default CA list built in
2975 Don't verify the peer certificate
2976 matches the server name
2978 ssldomain= The peer name as advertised in it's certificate.
2979 Used for verifying the correctness of the received peer
2980 certificate. If not specified the peer hostname will be
2984 Enable the "Front-End-Https: On" header needed when
2985 using Squid as a SSL frontend in front of Microsoft OWA.
2986 See MS KB document Q307347 for details on this header.
2987 If set to auto the header will only be added if the
2988 request is forwarded as a https:// URL.
2991 ==== GENERAL OPTIONS ====
2994 A peer-specific connect timeout.
2995 Also see the peer_connect_timeout directive.
2997 connect-fail-limit=N
2998 How many times connecting to a peer must fail before
2999 it is marked as down. Default is 10.
3001 allow-miss Disable Squid's use of only-if-cached when forwarding
3002 requests to siblings. This is primarily useful when
3003 icp_hit_stale is used by the sibling. To extensive use
3004 of this option may result in forwarding loops, and you
3005 should avoid having two-way peerings with this option.
3006 For example to deny peer usage on requests from peer
3007 by denying cache_peer_access if the source is a peer.
3009 max-conn=N Limit the amount of connections Squid may open to this
3012 name=xxx Unique name for the peer.
3013 Required if you have multiple peers on the same host
3014 but different ports.
3015 This name can be used in cache_peer_access and similar
3016 directives to dentify the peer.
3017 Can be used by outgoing access controls through the
3020 no-tproxy Do not use the client-spoof TPROXY support when forwarding
3021 requests to this peer. Use normal address selection instead.
3022 This overrides the spoof_client_ip ACL.
3024 proxy-only objects fetched from the peer will not be stored locally.
3028 NAME: cache_peer_domain cache_host_domain
3033 Use to limit the domains for which a neighbor cache will be
3037 cache_peer_domain cache-host domain [domain ...]
3038 cache_peer_domain cache-host !domain
3040 For example, specifying
3042 cache_peer_domain parent.foo.net .edu
3044 has the effect such that UDP query packets are sent to
3045 'bigserver' only when the requested object exists on a
3046 server in the .edu domain. Prefixing the domainname
3047 with '!' means the cache will be queried for objects
3050 NOTE: * Any number of domains may be given for a cache-host,
3051 either on the same or separate lines.
3052 * When multiple domains are given for a particular
3053 cache-host, the first matched domain is applied.
3054 * Cache hosts with no domain restrictions are queried
3056 * There are no defaults.
3057 * There is also a 'cache_peer_access' tag in the ACL
3061 NAME: cache_peer_access
3066 Similar to 'cache_peer_domain' but provides more flexibility by
3070 cache_peer_access cache-host allow|deny [!]aclname ...
3072 The syntax is identical to 'http_access' and the other lists of
3073 ACL elements. See the comments for 'http_access' below, or
3074 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
3077 NAME: neighbor_type_domain
3078 TYPE: hostdomaintype
3080 DEFAULT_DOC: The peer type from cache_peer directive is used for all requests to that peer.
3083 Modify the cache_peer neighbor type when passing requests
3084 about specific domains to the peer.
3087 neighbor_type_domain neighbor parent|sibling domain domain ...
3090 cache_peer foo.example.com parent 3128 3130
3091 neighbor_type_domain foo.example.com sibling .au .de
3093 The above configuration treats all requests to foo.example.com as a
3094 parent proxy unless the request is for a .au or .de ccTLD domain name.
3097 NAME: dead_peer_timeout
3101 LOC: Config.Timeout.deadPeer
3103 This controls how long Squid waits to declare a peer cache
3104 as "dead." If there are no ICP replies received in this
3105 amount of time, Squid will declare the peer dead and not
3106 expect to receive any further ICP replies. However, it
3107 continues to send ICP queries, and will mark the peer as
3108 alive upon receipt of the first subsequent ICP reply.
3110 This timeout also affects when Squid expects to receive ICP
3111 replies from peers. If more than 'dead_peer' seconds have
3112 passed since the last ICP reply was received, Squid will not
3113 expect to receive an ICP reply on the next query. Thus, if
3114 your time between requests is greater than this timeout, you
3115 will see a lot of requests sent DIRECT to origin servers
3116 instead of to your parents.
3119 NAME: forward_max_tries
3122 LOC: Config.forward_max_tries
3124 Controls how many different forward paths Squid will try
3125 before giving up. See also forward_timeout.
3127 NOTE: connect_retries (default: none) can make each of these
3128 possible forwarding paths be tried multiple times.
3131 NAME: hierarchy_stoplist
3134 LOC: Config.hierarchy_stoplist
3136 A list of words which, if found in a URL, cause the object to
3137 be handled directly by this cache. In other words, use this
3138 to not query neighbor caches for certain objects. You may
3139 list this option multiple times.
3142 hierarchy_stoplist cgi-bin ?
3144 Note: never_direct overrides this option.
3148 MEMORY CACHE OPTIONS
3149 -----------------------------------------------------------------------------
3156 LOC: Config.memMaxSize
3158 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3159 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3160 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3161 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3163 'cache_mem' specifies the ideal amount of memory to be used
3165 * In-Transit objects
3167 * Negative-Cached objects
3169 Data for these objects are stored in 4 KB blocks. This
3170 parameter specifies the ideal upper limit on the total size of
3171 4 KB blocks allocated. In-Transit objects take the highest
3174 In-transit objects have priority over the others. When
3175 additional space is needed for incoming data, negative-cached
3176 and hot objects will be released. In other words, the
3177 negative-cached and hot objects will fill up any unused space
3178 not needed for in-transit objects.
3180 If circumstances require, this limit will be exceeded.
3181 Specifically, if your incoming request rate requires more than
3182 'cache_mem' of memory to hold in-transit objects, Squid will
3183 exceed this limit to satisfy the new requests. When the load
3184 decreases, blocks will be freed until the high-water mark is
3185 reached. Thereafter, blocks will be used to store hot
3188 If shared memory caching is enabled, Squid does not use the shared
3189 cache space for in-transit objects, but they still consume as much
3190 local memory as they need. For more details about the shared memory
3191 cache, see memory_cache_shared.
3194 NAME: maximum_object_size_in_memory
3198 LOC: Config.Store.maxInMemObjSize
3200 Objects greater than this size will not be attempted to kept in
3201 the memory cache. This should be set high enough to keep objects
3202 accessed frequently in memory to improve performance whilst low
3203 enough to keep larger objects from hoarding cache_mem.
3206 NAME: memory_cache_shared
3209 LOC: Config.memShared
3211 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3213 Controls whether the memory cache is shared among SMP workers.
3215 The shared memory cache is meant to occupy cache_mem bytes and replace
3216 the non-shared memory cache, although some entities may still be
3217 cached locally by workers for now (e.g., internal and in-transit
3218 objects may be served from a local memory cache even if shared memory
3219 caching is enabled).
3221 By default, the memory cache is shared if and only if all of the
3222 following conditions are satisfied: Squid runs in SMP mode with
3223 multiple workers, cache_mem is positive, and Squid environment
3224 supports required IPC primitives (e.g., POSIX shared memory segments
3225 and GCC-style atomic operations).
3227 To avoid blocking locks, shared memory uses opportunistic algorithms
3228 that do not guarantee that every cachable entity that could have been
3229 shared among SMP workers will actually be shared.
3231 Currently, entities exceeding 32KB in size cannot be shared.
3234 NAME: memory_cache_mode
3238 DEFAULT_DOC: Keep the most recently fetched objects in memory
3240 Controls which objects to keep in the memory cache (cache_mem)
3242 always Keep most recently fetched objects in memory (default)
3244 disk Only disk cache hits are kept in memory, which means
3245 an object must first be cached on disk and then hit
3246 a second time before cached in memory.
3248 network Only objects fetched from network is kept in memory
3251 NAME: memory_replacement_policy
3253 LOC: Config.memPolicy
3256 The memory replacement policy parameter determines which
3257 objects are purged from memory when memory space is needed.
3259 See cache_replacement_policy for details on algorithms.
3264 -----------------------------------------------------------------------------
3267 NAME: cache_replacement_policy
3269 LOC: Config.replPolicy
3272 The cache replacement policy parameter determines which
3273 objects are evicted (replaced) when disk space is needed.
3275 lru : Squid's original list based LRU policy
3276 heap GDSF : Greedy-Dual Size Frequency
3277 heap LFUDA: Least Frequently Used with Dynamic Aging
3278 heap LRU : LRU policy implemented using a heap
3280 Applies to any cache_dir lines listed below this directive.
3282 The LRU policies keeps recently referenced objects.
3284 The heap GDSF policy optimizes object hit rate by keeping smaller
3285 popular objects in cache so it has a better chance of getting a
3286 hit. It achieves a lower byte hit rate than LFUDA though since
3287 it evicts larger (possibly popular) objects.
3289 The heap LFUDA policy keeps popular objects in cache regardless of
3290 their size and thus optimizes byte hit rate at the expense of
3291 hit rate since one large, popular object will prevent many
3292 smaller, slightly less popular objects from being cached.
3294 Both policies utilize a dynamic aging mechanism that prevents
3295 cache pollution that can otherwise occur with frequency-based
3296 replacement policies.
3298 NOTE: if using the LFUDA replacement policy you should increase
3299 the value of maximum_object_size above its default of 4 MB to
3300 to maximize the potential byte hit rate improvement of LFUDA.
3302 For more information about the GDSF and LFUDA cache replacement
3303 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3304 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3310 DEFAULT_DOC: No disk cache. Store cache ojects only in memory.
3311 LOC: Config.cacheSwap
3314 cache_dir Type Directory-Name Fs-specific-data [options]
3316 You can specify multiple cache_dir lines to spread the
3317 cache among different disk partitions.
3319 Type specifies the kind of storage system to use. Only "ufs"
3320 is built by default. To enable any of the other storage systems
3321 see the --enable-storeio configure option.
3323 'Directory' is a top-level directory where cache swap
3324 files will be stored. If you want to use an entire disk
3325 for caching, this can be the mount-point directory.
3326 The directory must exist and be writable by the Squid
3327 process. Squid will NOT create this directory for you.
3329 In SMP configurations, cache_dir must not precede the workers option
3330 and should use configuration macros or conditionals to give each
3331 worker interested in disk caching a dedicated cache directory.
3334 ==== The ufs store type ====
3336 "ufs" is the old well-known Squid storage format that has always
3340 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3342 'Mbytes' is the amount of disk space (MB) to use under this
3343 directory. The default is 100 MB. Change this to suit your
3344 configuration. Do NOT put the size of your disk drive here.
3345 Instead, if you want Squid to use the entire disk drive,
3346 subtract 20% and use that value.
3348 'L1' is the number of first-level subdirectories which
3349 will be created under the 'Directory'. The default is 16.
3351 'L2' is the number of second-level subdirectories which
3352 will be created under each first-level directory. The default
3356 ==== The aufs store type ====
3358 "aufs" uses the same storage format as "ufs", utilizing
3359 POSIX-threads to avoid blocking the main Squid process on
3360 disk-I/O. This was formerly known in Squid as async-io.
3363 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3365 see argument descriptions under ufs above
3368 ==== The diskd store type ====
3370 "diskd" uses the same storage format as "ufs", utilizing a
3371 separate process to avoid blocking the main Squid process on
3375 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3377 see argument descriptions under ufs above
3379 Q1 specifies the number of unacknowledged I/O requests when Squid
3380 stops opening new files. If this many messages are in the queues,
3381 Squid won't open new files. Default is 64
3383 Q2 specifies the number of unacknowledged messages when Squid
3384 starts blocking. If this many messages are in the queues,
3385 Squid blocks until it receives some replies. Default is 72
3387 When Q1 < Q2 (the default), the cache directory is optimized
3388 for lower response time at the expense of a decrease in hit
3389 ratio. If Q1 > Q2, the cache directory is optimized for
3390 higher hit ratio at the expense of an increase in response
3394 ==== The rock store type ====
3397 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3399 The Rock Store type is a database-style storage. All cached
3400 entries are stored in a "database" file, using fixed-size slots,
3401 one entry per slot. The database size is specified in MB. The
3402 slot size is specified in bytes using the max-size option. See
3403 below for more info on the max-size option.
3405 If possible, Squid using Rock Store creates a dedicated kid
3406 process called "disker" to avoid blocking Squid worker(s) on disk
3407 I/O. One disker kid is created for each rock cache_dir. Diskers
3408 are created only when Squid, running in daemon mode, has support
3409 for the IpcIo disk I/O module.
3411 swap-timeout=msec: Squid will not start writing a miss to or
3412 reading a hit from disk if it estimates that the swap operation
3413 will take more than the specified number of milliseconds. By
3414 default and when set to zero, disables the disk I/O time limit
3415 enforcement. Ignored when using blocking I/O module because
3416 blocking synchronous I/O does not allow Squid to estimate the
3417 expected swap wait time.
3419 max-swap-rate=swaps/sec: Artificially limits disk access using
3420 the specified I/O rate limit. Swap out requests that
3421 would cause the average I/O rate to exceed the limit are
3422 delayed. Individual swap in requests (i.e., hits or reads) are
3423 not delayed, but they do contribute to measured swap rate and
3424 since they are placed in the same FIFO queue as swap out
3425 requests, they may wait longer if max-swap-rate is smaller.
3426 This is necessary on file systems that buffer "too
3427 many" writes and then start blocking Squid and other processes
3428 while committing those writes to disk. Usually used together
3429 with swap-timeout to avoid excessive delays and queue overflows
3430 when disk demand exceeds available disk "bandwidth". By default
3431 and when set to zero, disables the disk I/O rate limit
3432 enforcement. Currently supported by IpcIo module only.
3435 ==== The coss store type ====
3437 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3438 production use and has thus been removed from this release.
3439 We hope that it can be made usable again soon.
3441 block-size=n defines the "block size" for COSS cache_dir's.
3442 Squid uses file numbers as block numbers. Since file numbers
3443 are limited to 24 bits, the block size determines the maximum
3444 size of the COSS partition. The default is 512 bytes, which
3445 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3446 you should not change the coss block size after Squid
3447 has written some objects to the cache_dir.
3449 The coss file store has changed from 2.5. Now it uses a file
3450 called 'stripe' in the directory names in the config - and
3451 this will be created by squid -z.
3454 ==== COMMON OPTIONS ====
3456 no-store no new objects should be stored to this cache_dir.
3458 min-size=n the minimum object size in bytes this cache_dir
3459 will accept. It's used to restrict a cache_dir
3460 to only store large objects (e.g. AUFS) while
3461 other stores are optimized for smaller objects
3465 max-size=n the maximum object size in bytes this cache_dir
3467 The value in maximum_object_size directive, sets
3468 a default unless more specific details are available
3469 about the cache_dir (ie a small store capacity).
3471 Note: To make optimal use of the max-size limits you should order
3472 the cache_dir lines with the smallest max-size value first.
3474 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3475 which can be changed with the --with-coss-membuf-size=N configure
3480 # Uncomment and adjust the following to add a disk cache directory.
3481 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3485 NAME: store_dir_select_algorithm
3487 LOC: Config.store_dir_select_algorithm
3490 How Squid selects which cache_dir to use when the response
3491 object will fit into more than one.
3493 Regardless of which algorithm is used the cache_dir min-size
3494 and max-size parameters are obeyed. As such they can affect
3495 the selection algorithm by limiting the set of considered
3502 This algorithm is suited to caches with similar cache_dir
3503 sizes and disk speeds.
3505 The disk with the least I/O pending is selected.
3506 When there are multiple disks with the same I/O load ranking
3507 the cache_dir with most available capacity is selected.
3509 When a mix of cache_dir sizes are configured the faster disks
3510 have a naturally lower I/O loading and larger disks have more
3511 capacity. So space used to store objects and data throughput
3512 may be very unbalanced towards larger disks.
3517 This algorithm is suited to caches with unequal cache_dir
3520 Each cache_dir is selected in a rotation. The next suitable
3523 Available cache_dir capacity is only considered in relation
3524 to whether the object will fit and meets the min-size and
3525 max-size parameters.
3527 Disk I/O loading is only considered to prevent overload on slow
3528 disks. This algorithm does not spread objects by size, so any
3529 I/O loading per-disk may appear very unbalanced and volatile.
3533 NAME: max_open_disk_fds
3535 LOC: Config.max_open_disk_fds
3537 DEFAULT_DOC: no limit
3539 To avoid having disk as the I/O bottleneck Squid can optionally
3540 bypass the on-disk cache if more than this amount of disk file
3541 descriptors are open.
3543 A value of 0 indicates no limit.
3546 NAME: minimum_object_size
3550 DEFAULT_DOC: no limit
3551 LOC: Config.Store.minObjectSize
3553 Objects smaller than this size will NOT be saved on disk. The
3554 value is specified in bytes, and the default is 0 KB, which
3555 means all responses can be stored.
3558 NAME: maximum_object_size
3562 LOC: Config.Store.maxObjectSize
3564 The default limit on size of objects stored to disk.
3565 This size is used for cache_dir where max-size is not set.
3566 The value is specified in bytes, and the default is 4 MB.
3568 If you wish to get a high BYTES hit ratio, you should probably
3569 increase this (one 32 MB object hit counts for 3200 10KB
3572 If you wish to increase hit ratio more than you want to
3573 save bandwidth you should leave this low.
3575 NOTE: if using the LFUDA replacement policy you should increase
3576 this value to maximize the byte hit rate improvement of LFUDA!
3577 See replacement_policy below for a discussion of this policy.
3580 NAME: cache_swap_low
3581 COMMENT: (percent, 0-100)
3584 LOC: Config.Swap.lowWaterMark
3586 The low-water mark for cache object replacement.
3587 Replacement begins when the swap (disk) usage is above the
3588 low-water mark and attempts to maintain utilization near the
3589 low-water mark. As swap utilization gets close to high-water
3590 mark object eviction becomes more aggressive. If utilization is
3591 close to the low-water mark less replacement is done each time.
3593 Defaults are 90% and 95%. If you have a large cache, 5% could be
3594 hundreds of MB. If this is the case you may wish to set these
3595 numbers closer together.
3597 See also cache_swap_high
3600 NAME: cache_swap_high
3601 COMMENT: (percent, 0-100)
3604 LOC: Config.Swap.highWaterMark
3606 The high-water mark for cache object replacement.
3607 Replacement begins when the swap (disk) usage is above the
3608 low-water mark and attempts to maintain utilization near the
3609 low-water mark. As swap utilization gets close to high-water
3610 mark object eviction becomes more aggressive. If utilization is
3611 close to the low-water mark less replacement is done each time.
3613 Defaults are 90% and 95%. If you have a large cache, 5% could be
3614 hundreds of MB. If this is the case you may wish to set these
3615 numbers closer together.
3617 See also cache_swap_low
3622 -----------------------------------------------------------------------------
3629 DEFAULT_DOC: The format definitions squid, common, combined, referrer, useragent are built in.
3633 logformat <name> <format specification>
3635 Defines an access log format.
3637 The <format specification> is a string with embedded % format codes
3639 % format codes all follow the same basic structure where all but
3640 the formatcode is optional. Output strings are automatically escaped
3641 as required according to their context and the output format
3642 modifiers are usually not needed, but can be specified if an explicit
3643 output format is desired.
3645 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3647 " output in quoted string format
3648 [ output in squid text log format as used by log_mime_hdrs
3649 # output in URL quoted format
3654 width minimum and/or maximum field width:
3655 [width_min][.width_max]
3656 When minimum starts with 0, the field is zero-padded.
3657 String values exceeding maximum width are truncated.
3659 {arg} argument such as header name etc
3663 % a literal % character
3664 sn Unique sequence number per log line entry
3665 err_code The ID of an error response served by Squid or
3666 a similar internal error identifier.
3667 err_detail Additional err_code-dependent error information.
3668 note The meta header specified by the argument. Also
3669 logs the adaptation meta headers set by the
3670 adaptation_meta configuration parameter.
3671 If no argument given all meta headers logged.
3673 Connection related format codes:
3675 >a Client source IP address
3677 >p Client source port
3678 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3679 >la Local IP address the client connected to
3680 >lp Local port number the client connected to
3682 la Local listening IP address the client connection was connected to.
3683 lp Local listening port number the client connection was connected to.
3685 <a Server IP address of the last server or peer connection
3686 <A Server FQDN or peer name
3687 <p Server port number of the last server or peer connection
3688 <la Local IP address of the last server or peer connection
3689 <lp Local port number of the last server or peer connection
3691 Time related format codes:
3693 ts Seconds since epoch
3694 tu subsecond time (milliseconds)
3695 tl Local time. Optional strftime format argument
3696 default %d/%b/%Y:%H:%M:%S %z
3697 tg GMT time. Optional strftime format argument
3698 default %d/%b/%Y:%H:%M:%S %z
3699 tr Response time (milliseconds)
3700 dt Total time spent making DNS lookups (milliseconds)
3702 Access Control related format codes:
3704 et Tag returned by external acl
3705 ea Log string returned by external acl
3706 un User name (any available)
3707 ul User name from authentication
3708 ue User name from external acl helper
3709 ui User name from ident
3710 us User name from SSL
3712 HTTP related format codes:
3714 [http::]>h Original received request header.
3715 Usually differs from the request header sent by
3716 Squid, although most fields are often preserved.
3717 Accepts optional header field name/value filter
3718 argument using name[:[separator]element] format.
3719 [http::]>ha Received request header after adaptation and
3720 redirection (pre-cache REQMOD vectoring point).
3721 Usually differs from the request header sent by
3722 Squid, although most fields are often preserved.
3723 Optional header name argument as for >h
3724 [http::]<h Reply header. Optional header name argument
3726 [http::]>Hs HTTP status code sent to the client
3727 [http::]<Hs HTTP status code received from the next hop
3728 [http::]<bs Number of HTTP-equivalent message body bytes
3729 received from the next hop, excluding chunked
3730 transfer encoding and control messages.
3731 Generated FTP/Gopher listings are treated as
3733 [http::]mt MIME content type
3734 [http::]rm Request method (GET/POST etc)
3735 [http::]>rm Request method from client
3736 [http::]<rm Request method sent to server or peer
3737 [http::]ru Request URL from client (historic, filtered for logging)
3738 [http::]>ru Request URL from client
3739 [http::]<ru Request URL sent to server or peer
3740 [http::]rp Request URL-Path excluding hostname
3741 [http::]>rp Request URL-Path excluding hostname from client
3742 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3743 [http::]rv Request protocol version
3744 [http::]>rv Request protocol version from client
3745 [http::]<rv Request protocol version sent to server or peer
3746 [http::]<st Sent reply size including HTTP headers
3747 [http::]>st Received request size including HTTP headers. In the
3748 case of chunked requests the chunked encoding metadata
3750 [http::]>sh Received HTTP request headers size
3751 [http::]<sh Sent HTTP reply headers size
3752 [http::]st Request+Reply size including HTTP headers
3753 [http::]<sH Reply high offset sent
3754 [http::]<sS Upstream object size
3755 [http::]<pt Peer response time in milliseconds. The timer starts
3756 when the last request byte is sent to the next hop
3757 and stops when the last response byte is received.
3758 [http::]<tt Total server-side time in milliseconds. The timer
3759 starts with the first connect request (or write I/O)
3760 sent to the first selected peer. The timer stops
3761 with the last I/O with the last peer.
3763 Squid handling related format codes:
3765 Ss Squid request status (TCP_MISS etc)
3766 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3768 SSL-related format codes:
3770 ssl::bump_mode SslBump decision for the transaction:
3772 For CONNECT requests that initiated bumping of
3773 a connection and for any request received on
3774 an already bumped connection, Squid logs the
3775 corresponding SslBump mode ("server-first" or
3776 "client-first"). See the ssl_bump option for
3777 more information about these modes.
3779 A "none" token is logged for requests that
3780 triggered "ssl_bump" ACL evaluation matching
3781 either a "none" rule or no rules at all.
3783 In all other cases, a single dash ("-") is
3786 If ICAP is enabled, the following code becomes available (as
3787 well as ICAP log codes documented with the icap_log option):
3789 icap::tt Total ICAP processing time for the HTTP
3790 transaction. The timer ticks when ICAP
3791 ACLs are checked and when ICAP
3792 transaction is in progress.
3794 If adaptation is enabled the following three codes become available:
3796 adapt::<last_h The header of the last ICAP response or
3797 meta-information from the last eCAP
3798 transaction related to the HTTP transaction.
3799 Like <h, accepts an optional header name
3802 adapt::sum_trs Summed adaptation transaction response
3803 times recorded as a comma-separated list in
3804 the order of transaction start time. Each time
3805 value is recorded as an integer number,
3806 representing response time of one or more
3807 adaptation (ICAP or eCAP) transaction in
3808 milliseconds. When a failed transaction is
3809 being retried or repeated, its time is not
3810 logged individually but added to the
3811 replacement (next) transaction. See also:
3814 adapt::all_trs All adaptation transaction response times.
3815 Same as adaptation_strs but response times of
3816 individual transactions are never added
3817 together. Instead, all transaction response
3818 times are recorded individually.
3820 You can prefix adapt::*_trs format codes with adaptation
3821 service name in curly braces to record response time(s) specific
3822 to that service. For example: %{my_service}adapt::sum_trs
3824 If SSL is enabled, the following formating codes become available:
3826 %ssl::>cert_subject The Subject field of the received client
3827 SSL certificate or a dash ('-') if Squid has
3828 received an invalid/malformed certificate or
3829 no certificate at all. Consider encoding the
3830 logged value because Subject often has spaces.
3832 %ssl::>cert_issuer The Issuer field of the received client
3833 SSL certificate or a dash ('-') if Squid has
3834 received an invalid/malformed certificate or
3835 no certificate at all. Consider encoding the
3836 logged value because Issuer often has spaces.
3838 The default formats available (which do not need re-defining) are:
3840 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3841 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3842 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3843 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3844 logformat useragent %>a [%tl] "%{User-Agent}>h"
3846 NOTE: When the log_mime_hdrs directive is set to ON.
3847 The squid, common and combined formats have a safely encoded copy
3848 of the mime headers appended to each line within a pair of brackets.
3850 NOTE: The common and combined formats are not quite true to the Apache definition.
3851 The logs from Squid contain an extra status and hierarchy code appended.
3855 NAME: access_log cache_access_log
3857 LOC: Config.Log.accesslogs
3858 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3860 Configures whether and how Squid logs HTTP and ICP transactions.
3861 If access logging is enabled, a single line is logged for every
3862 matching HTTP or ICP request. The recommended directive formats are:
3864 access_log <module>:<place> [option ...] [acl acl ...]
3865 access_log none [acl acl ...]
3867 The following directive format is accepted but may be deprecated:
3868 access_log <module>:<place> [<logformat name> [acl acl ...]]
3870 In most cases, the first ACL name must not contain the '=' character
3871 and should not be equal to an existing logformat name. You can always
3872 start with an 'all' ACL to work around those restrictions.
3874 Will log to the specified module:place using the specified format (which
3875 must be defined in a logformat directive) those entries which match
3876 ALL the acl's specified (which must be defined in acl clauses).
3877 If no acl is specified, all requests will be logged to this destination.
3879 ===== Available options for the recommended directive format =====
3881 logformat=name Names log line format (either built-in or
3882 defined by a logformat directive). Defaults
3885 buffer-size=64KB Defines approximate buffering limit for log
3886 records (see buffered_logs). Squid should not
3887 keep more than the specified size and, hence,
3888 should flush records before the buffer becomes
3889 full to avoid overflows under normal
3890 conditions (the exact flushing algorithm is
3891 module-dependent though). The on-error option
3892 controls overflow handling.
3894 on-error=die|drop Defines action on unrecoverable errors. The
3895 'drop' action ignores (i.e., does not log)
3896 affected log records. The default 'die' action
3897 kills the affected worker. The drop action
3898 support has not been tested for modules other
3901 ===== Modules Currently available =====
3903 none Do not log any requests matching these ACL.
3904 Do not specify Place or logformat name.
3906 stdio Write each log line to disk immediately at the completion of
3908 Place: the filename and path to be written.
3910 daemon Very similar to stdio. But instead of writing to disk the log
3911 line is passed to a daemon helper for asychronous handling instead.
3912 Place: varies depending on the daemon.
3914 log_file_daemon Place: the file name and path to be written.
3916 syslog To log each request via syslog facility.
3917 Place: The syslog facility and priority level for these entries.
3918 Place Format: facility.priority
3920 where facility could be any of:
3921 authpriv, daemon, local0 ... local7 or user.
3923 And priority could be any of:
3924 err, warning, notice, info, debug.
3926 udp To send each log line as text data to a UDP receiver.
3927 Place: The destination host name or IP and port.
3928 Place Format: //host:port
3930 tcp To send each log line as text data to a TCP receiver.
3931 Lines may be accumulated before sending (see buffered_logs).
3932 Place: The destination host name or IP and port.
3933 Place Format: //host:port
3936 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3942 LOC: Config.Log.icaplogs
3945 ICAP log files record ICAP transaction summaries, one line per
3948 The icap_log option format is:
3949 icap_log <filepath> [<logformat name> [acl acl ...]]
3950 icap_log none [acl acl ...]]
3952 Please see access_log option documentation for details. The two
3953 kinds of logs share the overall configuration approach and many
3956 ICAP processing of a single HTTP message or transaction may
3957 require multiple ICAP transactions. In such cases, multiple
3958 ICAP transaction log lines will correspond to a single access
3961 ICAP log uses logformat codes that make sense for an ICAP
3962 transaction. Header-related codes are applied to the HTTP header
3963 embedded in an ICAP server response, with the following caveats:
3964 For REQMOD, there is no HTTP response header unless the ICAP
3965 server performed request satisfaction. For RESPMOD, the HTTP
3966 request header is the header sent to the ICAP server. For
3967 OPTIONS, there are no HTTP headers.
3969 The following format codes are also available for ICAP logs:
3971 icap::<A ICAP server IP address. Similar to <A.
3973 icap::<service_name ICAP service name from the icap_service
3974 option in Squid configuration file.
3976 icap::ru ICAP Request-URI. Similar to ru.
3978 icap::rm ICAP request method (REQMOD, RESPMOD, or
3979 OPTIONS). Similar to existing rm.
3981 icap::>st Bytes sent to the ICAP server (TCP payload
3982 only; i.e., what Squid writes to the socket).
3984 icap::<st Bytes received from the ICAP server (TCP
3985 payload only; i.e., what Squid reads from
3988 icap::<bs Number of message body bytes received from the
3989 ICAP server. ICAP message body, if any, usually
3990 includes encapsulated HTTP message headers and
3991 possibly encapsulated HTTP message body. The
3992 HTTP body part is dechunked before its size is
3995 icap::tr Transaction response time (in
3996 milliseconds). The timer starts when
3997 the ICAP transaction is created and
3998 stops when the transaction is completed.
4001 icap::tio Transaction I/O time (in milliseconds). The
4002 timer starts when the first ICAP request
4003 byte is scheduled for sending. The timers
4004 stops when the last byte of the ICAP response
4007 icap::to Transaction outcome: ICAP_ERR* for all
4008 transaction errors, ICAP_OPT for OPTION
4009 transactions, ICAP_ECHO for 204
4010 responses, ICAP_MOD for message
4011 modification, and ICAP_SAT for request
4012 satisfaction. Similar to Ss.
4014 icap::Hs ICAP response status code. Similar to Hs.
4016 icap::>h ICAP request header(s). Similar to >h.
4018 icap::<h ICAP response header(s). Similar to <h.
4020 The default ICAP log format, which can be used without an explicit
4021 definition, is called icap_squid:
4023 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
4025 See also: logformat, log_icap, and %adapt::<last_h
4028 NAME: logfile_daemon
4030 DEFAULT: @DEFAULT_LOGFILED@
4031 LOC: Log::TheConfig.logfile_daemon
4033 Specify the path to the logfile-writing daemon. This daemon is
4034 used to write the access and store logs, if configured.
4036 Squid sends a number of commands to the log daemon:
4037 L<data>\n - logfile data
4042 r<n>\n - set rotate count to <n>
4043 b<n>\n - 1 = buffer output, 0 = don't buffer output
4045 No responses is expected.
4051 Remove this line. Use acls with access_log directives to control access logging
4057 Remove this line. Use acls with icap_log directives to control icap logging
4060 NAME: stats_collection
4062 LOC: Config.accessList.stats_collection
4064 DEFAULT_DOC: Allow logging for all transactions.
4065 COMMENT: allow|deny acl acl...
4067 This options allows you to control which requests gets accounted
4068 in performance counters.
4070 This clause only supports fast acl types.
4071 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4074 NAME: cache_store_log
4077 LOC: Config.Log.store
4079 Logs the activities of the storage manager. Shows which
4080 objects are ejected from the cache, and which objects are
4081 saved and for how long.
4082 There are not really utilities to analyze this data, so you can safely
4083 disable it (the default).
4085 Store log uses modular logging outputs. See access_log for the list
4086 of modules supported.
4089 cache_store_log stdio:@DEFAULT_STORE_LOG@
4090 cache_store_log daemon:@DEFAULT_STORE_LOG@
4093 NAME: cache_swap_state cache_swap_log
4095 LOC: Config.Log.swap
4097 DEFAULT_DOC: Store the journal inside its cache_dir
4099 Location for the cache "swap.state" file. This index file holds
4100 the metadata of objects saved on disk. It is used to rebuild
4101 the cache during startup. Normally this file resides in each
4102 'cache_dir' directory, but you may specify an alternate
4103 pathname here. Note you must give a full filename, not just
4104 a directory. Since this is the index for the whole object
4105 list you CANNOT periodically rotate it!
4107 If %s can be used in the file name it will be replaced with a
4108 a representation of the cache_dir name where each / is replaced
4109 with '.'. This is needed to allow adding/removing cache_dir
4110 lines when cache_swap_log is being used.
4112 If have more than one 'cache_dir', and %s is not used in the name
4113 these swap logs will have names such as:
4119 The numbered extension (which is added automatically)
4120 corresponds to the order of the 'cache_dir' lines in this
4121 configuration file. If you change the order of the 'cache_dir'
4122 lines in this file, these index files will NOT correspond to
4123 the correct 'cache_dir' entry (unless you manually rename
4124 them). We recommend you do NOT use this option. It is
4125 better to keep these index files in each 'cache_dir' directory.
4128 NAME: logfile_rotate
4131 LOC: Config.Log.rotateNumber
4133 Specifies the number of logfile rotations to make when you
4134 type 'squid -k rotate'. The default is 10, which will rotate
4135 with extensions 0 through 9. Setting logfile_rotate to 0 will
4136 disable the file name rotation, but the logfiles are still closed
4137 and re-opened. This will enable you to rename the logfiles
4138 yourself just before sending the rotate signal.
4140 Note, the 'squid -k rotate' command normally sends a USR1
4141 signal to the running squid process. In certain situations
4142 (e.g. on Linux with Async I/O), USR1 is used for other
4143 purposes, so -k rotate uses another signal. It is best to get
4144 in the habit of using 'squid -k rotate' instead of 'kill -USR1
4147 Note, from Squid-3.1 this option is only a default for cache.log,
4148 that log can be rotated separately by using debug_options.
4151 NAME: emulate_httpd_log
4154 Replace this with an access_log directive using the format 'common' or 'combined'.
4157 NAME: log_ip_on_direct
4160 Remove this option from your config. To log server or peer names use %<A in the log format.
4165 DEFAULT: @DEFAULT_MIME_TABLE@
4166 LOC: Config.mimeTablePathname
4168 Path to Squid's icon configuration file.
4170 You shouldn't need to change this, but the default file contains
4171 examples and formatting information if you do.
4177 LOC: Config.onoff.log_mime_hdrs
4180 The Cache can record both the request and the response MIME
4181 headers for each HTTP transaction. The headers are encoded
4182 safely and will appear as two bracketed fields at the end of
4183 the access log (for either the native or httpd-emulated log
4184 formats). To enable this logging set log_mime_hdrs to 'on'.
4190 Replace this with an access_log directive using the format 'useragent'.
4193 NAME: referer_log referrer_log
4196 Replace this with an access_log directive using the format 'referrer'.
4201 DEFAULT: @DEFAULT_PID_FILE@
4202 LOC: Config.pidFilename
4204 A filename to write the process-id to. To disable, enter "none".
4210 Remove this option from your config. To log FQDN use %>A in the log format.
4213 NAME: client_netmask
4215 LOC: Config.Addrs.client_netmask
4217 DEFAULT_DOC: Log full client IP address
4219 A netmask for client addresses in logfiles and cachemgr output.
4220 Change this to protect the privacy of your cache clients.
4221 A netmask of 255.255.255.0 will log all IP's in that range with
4222 the last digit set to '0'.
4228 Use a regular access.log with ACL limiting it to MISS events.
4231 NAME: strip_query_terms
4233 LOC: Config.onoff.strip_query_terms
4236 By default, Squid strips query terms from requested URLs before
4237 logging. This protects your user's privacy and reduces log size.
4239 When investigating HIT/MISS or other caching behaviour you
4240 will need to disable this to see the full URL used by Squid.
4247 LOC: Config.onoff.buffered_logs
4249 Whether to write/send access_log records ASAP or accumulate them and
4250 then write/send them in larger chunks. Buffering may improve
4251 performance because it decreases the number of I/Os. However,
4252 buffering increases the delay before log records become available to
4253 the final recipient (e.g., a disk file or logging daemon) and,
4254 hence, increases the risk of log records loss.
4256 Note that even when buffered_logs are off, Squid may have to buffer
4257 records if it cannot write/send them immediately due to pending I/Os
4258 (e.g., the I/O writing the previous log record) or connectivity loss.
4260 Currently honored by 'daemon' and 'tcp' access_log modules only.
4263 NAME: netdb_filename
4265 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
4266 LOC: Config.netdbFilename
4269 Where Squid stores it's netdb journal.
4270 When enabled this journal preserves netdb state between restarts.
4272 To disable, enter "none".
4276 OPTIONS FOR TROUBLESHOOTING
4277 -----------------------------------------------------------------------------
4282 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
4283 LOC: Debug::cache_log
4285 Squid administrative logging file.
4287 This is where general information about Squid behavior goes. You can
4288 increase the amount of data logged to this file and how often it is
4289 rotated with "debug_options"
4295 DEFAULT_DOC: Log all critical and important messages.
4296 LOC: Debug::debugOptions
4298 Logging options are set as section,level where each source file
4299 is assigned a unique section. Lower levels result in less
4300 output, Full debugging (level 9) can result in a very large
4301 log file, so be careful.
4303 The magic word "ALL" sets debugging levels for all sections.
4304 The default is to run with "ALL,1" to record important warnings.
4306 The rotate=N option can be used to keep more or less of these logs
4307 than would otherwise be kept by logfile_rotate.
4308 For most uses a single log should be enough to monitor current
4309 events affecting Squid.
4314 LOC: Config.coredump_dir
4315 DEFAULT_IF_NONE: none
4316 DEFAULT_DOC: Use the directory from where Squid was started.
4318 By default Squid leaves core files in the directory from where
4319 it was started. If you set 'coredump_dir' to a directory
4320 that exists, Squid will chdir() to that directory at startup
4321 and coredump files will be left there.
4325 # Leave coredumps in the first cache dir
4326 coredump_dir @DEFAULT_SWAP_DIR@
4332 OPTIONS FOR FTP GATEWAYING
4333 -----------------------------------------------------------------------------
4339 LOC: Config.Ftp.anon_user
4341 If you want the anonymous login password to be more informative
4342 (and enable the use of picky FTP servers), set this to something
4343 reasonable for your domain, like wwwuser@somewhere.net
4345 The reason why this is domainless by default is the
4346 request can be made on the behalf of a user in any domain,
4347 depending on how the cache is used.
4348 Some FTP server also validate the email address is valid
4349 (for example perl.com).
4355 LOC: Config.Ftp.passive
4357 If your firewall does not allow Squid to use passive
4358 connections, turn off this option.
4360 Use of ftp_epsv_all option requires this to be ON.
4366 LOC: Config.Ftp.epsv_all
4368 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4370 NATs may be able to put the connection on a "fast path" through the
4371 translator, as the EPRT command will never be used and therefore,
4372 translation of the data portion of the segments will never be needed.
4374 When a client only expects to do two-way FTP transfers this may be
4376 If squid finds that it must do a three-way FTP transfer after issuing
4377 an EPSV ALL command, the FTP session will fail.
4379 If you have any doubts about this option do not use it.
4380 Squid will nicely attempt all other connection methods.
4382 Requires ftp_passive to be ON (default) for any effect.
4388 LOC: Config.Ftp.epsv
4390 FTP Protocol extensions permit the use of a special "EPSV" command.
4392 NATs may be able to put the connection on a "fast path" through the
4393 translator using EPSV, as the EPRT command will never be used
4394 and therefore, translation of the data portion of the segments
4395 will never be needed.
4397 Turning this OFF will prevent EPSV being attempted.
4398 WARNING: Doing so will convert Squid back to the old behavior with all
4399 the related problems with external NAT devices/layers.
4401 Requires ftp_passive to be ON (default) for any effect.
4407 LOC: Config.Ftp.eprt
4409 FTP Protocol extensions permit the use of a special "EPRT" command.
4411 This extension provides a protocol neutral alternative to the
4412 IPv4-only PORT command. When supported it enables active FTP data
4413 channels over IPv6 and efficient NAT handling.
4415 Turning this OFF will prevent EPRT being attempted and will skip
4416 straight to using PORT for IPv4 servers.
4418 Some devices are known to not handle this extension correctly and
4419 may result in crashes. Devices which suport EPRT enough to fail
4420 cleanly will result in Squid attempting PORT anyway. This directive
4421 should only be disabled when EPRT results in device failures.
4423 WARNING: Doing so will convert Squid back to the old behavior with all
4424 the related problems with external NAT devices/layers and IPv4-only FTP.
4427 NAME: ftp_sanitycheck
4430 LOC: Config.Ftp.sanitycheck
4432 For security and data integrity reasons Squid by default performs
4433 sanity checks of the addresses of FTP data connections ensure the
4434 data connection is to the requested server. If you need to allow
4435 FTP connections to servers using another IP address for the data
4436 connection turn this off.
4439 NAME: ftp_telnet_protocol
4442 LOC: Config.Ftp.telnet
4444 The FTP protocol is officially defined to use the telnet protocol
4445 as transport channel for the control connection. However, many
4446 implementations are broken and does not respect this aspect of
4449 If you have trouble accessing files with ASCII code 255 in the
4450 path or similar problems involving this ASCII code you can
4451 try setting this directive to off. If that helps, report to the
4452 operator of the FTP server in question that their FTP server
4453 is broken and does not follow the FTP standard.
4457 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4458 -----------------------------------------------------------------------------
4463 DEFAULT: @DEFAULT_DISKD@
4464 LOC: Config.Program.diskd
4466 Specify the location of the diskd executable.
4467 Note this is only useful if you have compiled in
4468 diskd as one of the store io modules.
4471 NAME: unlinkd_program
4474 DEFAULT: @DEFAULT_UNLINKD@
4475 LOC: Config.Program.unlinkd
4477 Specify the location of the executable for file deletion process.
4480 NAME: pinger_program
4482 DEFAULT: @DEFAULT_PINGER@
4483 LOC: Config.pinger.program
4486 Specify the location of the executable for the pinger process.
4492 LOC: Config.pinger.enable
4495 Control whether the pinger is active at run-time.
4496 Enables turning ICMP pinger on and off with a simple
4497 squid -k reconfigure.
4502 OPTIONS FOR URL REWRITING
4503 -----------------------------------------------------------------------------
4506 NAME: url_rewrite_program redirect_program
4508 LOC: Config.Program.redirect
4511 Specify the location of the executable URL rewriter to use.
4512 Since they can perform almost any function there isn't one included.
4514 For each requested URL, the rewriter will receive on line with the format
4516 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4519 After processing the request the helper must reply using the following format:
4521 [channel-ID <SP>] result [<SP> kv-pairs]
4523 The result code can be:
4525 OK status=30N url="..."
4526 Redirect the URL to the one supplied in 'url='.
4527 'status=' is optional and contains the status code to send
4528 the client in Squids HTTP response. It must be one of the
4529 HTTP redirect status codes: 301, 302, 303, 307, 308.
4530 When no status is given Squid will use 302.
4532 OK rewrite-url="..."
4533 Rewrite the URL to the one supplied in 'rewrite-url='.
4534 The new URL is fetched directly by Squid and returned to
4535 the client as the response to its request.
4538 Do not change the URL.
4541 An internal error occurred in the helper, preventing
4542 a result being identified.
4545 In the future, the interface protocol will be extended with
4546 key=value pairs ("kv-pairs" shown above). Helper programs
4547 should be prepared to receive and possibly ignore additional
4548 whitespace-separated tokens on each input line.
4550 When using the concurrency= option the protocol is changed by
4551 introducing a query channel tag in front of the request/response.
4552 The query channel tag is a number between 0 and concurrency-1.
4553 This value must be echoed back unchanged to Squid as the first part
4554 of the response relating to its request.
4556 WARNING: URL re-writing ability should be avoided whenever possible.
4557 Use the URL redirect form of response instead.
4559 Re-write creates a difference in the state held by the client
4560 and server. Possibly causing confusion when the server response
4561 contains snippets of its view state. Embeded URLs, response
4562 and content Location headers, etc. are not re-written by this
4565 By default, a URL rewriter is not used.
4568 NAME: url_rewrite_children redirect_children
4569 TYPE: HelperChildConfig
4570 DEFAULT: 20 startup=0 idle=1 concurrency=0
4571 LOC: Config.redirectChildren
4573 The maximum number of redirector processes to spawn. If you limit
4574 it too few Squid will have to wait for them to process a backlog of
4575 URLs, slowing it down. If you allow too many they will use RAM
4576 and other system resources noticably.
4578 The startup= and idle= options allow some measure of skew in your
4583 Sets a minimum of how many processes are to be spawned when Squid
4584 starts or reconfigures. When set to zero the first request will
4585 cause spawning of the first child process to handle it.
4587 Starting too few will cause an initial slowdown in traffic as Squid
4588 attempts to simultaneously spawn enough processes to cope.
4592 Sets a minimum of how many processes Squid is to try and keep available
4593 at all times. When traffic begins to rise above what the existing
4594 processes can handle this many more will be spawned up to the maximum
4595 configured. A minimum setting of 1 is required.
4599 The number of requests each redirector helper can handle in
4600 parallel. Defaults to 0 which indicates the redirector
4601 is a old-style single threaded redirector.
4603 When this directive is set to a value >= 1 then the protocol
4604 used to communicate with the helper is modified to include
4605 an ID in front of the request/response. The ID from the request
4606 must be echoed back with the response to that request.
4609 NAME: url_rewrite_host_header redirect_rewrites_host_header
4612 LOC: Config.onoff.redir_rewrites_host
4614 To preserve same-origin security policies in browsers and
4615 prevent Host: header forgery by redirectors Squid rewrites
4616 any Host: header in redirected requests.
4618 If you are running an accelerator this may not be a wanted
4619 effect of a redirector. This directive enables you disable
4620 Host: alteration in reverse-proxy traffic.
4622 WARNING: Entries are cached on the result of the URL rewriting
4623 process, so be careful if you have domain-virtual hosts.
4625 WARNING: Squid and other software verifies the URL and Host
4626 are matching, so be careful not to relay through other proxies
4627 or inspecting firewalls with this disabled.
4630 NAME: url_rewrite_access redirector_access
4633 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4634 LOC: Config.accessList.redirector
4636 If defined, this access list specifies which requests are
4637 sent to the redirector processes.
4639 This clause supports both fast and slow acl types.
4640 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4643 NAME: url_rewrite_bypass redirector_bypass
4645 LOC: Config.onoff.redirector_bypass
4648 When this is 'on', a request will not go through the
4649 redirector if all the helpers are busy. If this is 'off'
4650 and the redirector queue grows too large, Squid will exit
4651 with a FATAL error and ask you to increase the number of
4652 redirectors. You should only enable this if the redirectors
4653 are not critical to your caching system. If you use
4654 redirectors for access control, and you enable this option,
4655 users may have access to pages they should not
4656 be allowed to request.
4660 OPTIONS FOR STORE ID
4661 -----------------------------------------------------------------------------
4664 NAME: store_id_program storeurl_rewrite_program
4666 LOC: Config.Program.store_id
4669 Specify the location of the executable StoreID helper to use.
4670 Since they can perform almost any function there isn't one included.
4672 For each requested URL, the helper will receive one line with the format
4674 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4677 After processing the request the helper must reply using the following format:
4679 [channel-ID <SP>] result [<SP> kv-pairs]
4681 The result code can be:
4684 Use the StoreID supplied in 'store-id='.
4687 The default is to use HTTP request URL as the store ID.
4690 An internal error occured in the helper, preventing
4691 a result being identified.
4694 Helper programs should be prepared to receive and possibly ignore additional
4695 kv-pairs with keys they do not support.
4697 When using the concurrency= option the protocol is changed by
4698 introducing a query channel tag in front of the request/response.
4699 The query channel tag is a number between 0 and concurrency-1.
4700 This value must be echoed back unchanged to Squid as the first part
4701 of the response relating to its request.
4703 NOTE: when using StoreID refresh_pattern will apply to the StoreID
4704 returned from the helper and not the URL.
4706 WARNING: Wrong StoreID value returned by a careless helper may result
4707 in the wrong cached response returned to the user.
4709 By default, a StoreID helper is not used.
4712 NAME: store_id_children storeurl_rewrite_children
4713 TYPE: HelperChildConfig
4714 DEFAULT: 20 startup=0 idle=1 concurrency=0
4715 LOC: Config.storeIdChildren
4717 The maximum number of StoreID helper processes to spawn. If you limit
4718 it too few Squid will have to wait for them to process a backlog of
4719 requests, slowing it down. If you allow too many they will use RAM
4720 and other system resources noticably.
4722 The startup= and idle= options allow some measure of skew in your
4727 Sets a minimum of how many processes are to be spawned when Squid
4728 starts or reconfigures. When set to zero the first request will
4729 cause spawning of the first child process to handle it.
4731 Starting too few will cause an initial slowdown in traffic as Squid
4732 attempts to simultaneously spawn enough processes to cope.
4736 Sets a minimum of how many processes Squid is to try and keep available
4737 at all times. When traffic begins to rise above what the existing
4738 processes can handle this many more will be spawned up to the maximum
4739 configured. A minimum setting of 1 is required.
4743 The number of requests each storeID helper can handle in
4744 parallel. Defaults to 0 which indicates the helper
4745 is a old-style single threaded program.
4747 When this directive is set to a value >= 1 then the protocol
4748 used to communicate with the helper is modified to include
4749 an ID in front of the request/response. The ID from the request
4750 must be echoed back with the response to that request.
4753 NAME: store_id_access storeurl_rewrite_access
4756 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
4757 LOC: Config.accessList.store_id
4759 If defined, this access list specifies which requests are
4760 sent to the StoreID processes. By default all requests
4763 This clause supports both fast and slow acl types.
4764 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4767 NAME: store_id_bypass storeurl_rewrite_bypass
4769 LOC: Config.onoff.store_id_bypass
4772 When this is 'on', a request will not go through the
4773 helper if all helpers are busy. If this is 'off'
4774 and the helper queue grows too large, Squid will exit
4775 with a FATAL error and ask you to increase the number of
4776 helpers. You should only enable this if the helperss
4777 are not critical to your caching system. If you use
4778 helpers for critical caching components, and you enable this
4779 option, users may not get objects from cache.
4783 OPTIONS FOR TUNING THE CACHE
4784 -----------------------------------------------------------------------------
4787 NAME: cache no_cache
4790 DEFAULT_DOC: Allow caching, unless rules exist in squid.conf.
4791 LOC: Config.accessList.noCache
4793 A list of ACL elements which, if matched and denied, cause the request to
4794 not be satisfied from the cache and the reply to not be cached.
4795 In other words, use this to force certain objects to never be cached.
4797 You must use the words 'allow' or 'deny' to indicate whether items
4798 matching the ACL should be allowed or denied into the cache.
4800 This clause supports both fast and slow acl types.
4801 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4807 LOC: Config.maxStale
4810 This option puts an upper limit on how stale content Squid
4811 will serve from the cache if cache validation fails.
4812 Can be overriden by the refresh_pattern max-stale option.
4815 NAME: refresh_pattern
4816 TYPE: refreshpattern
4820 usage: refresh_pattern [-i] regex min percent max [options]
4822 By default, regular expressions are CASE-SENSITIVE. To make
4823 them case-insensitive, use the -i option.
4825 'Min' is the time (in minutes) an object without an explicit
4826 expiry time should be considered fresh. The recommended
4827 value is 0, any higher values may cause dynamic applications
4828 to be erroneously cached unless the application designer
4829 has taken the appropriate actions.
4831 'Percent' is a percentage of the objects age (time since last
4832 modification age) an object without explicit expiry time
4833 will be considered fresh.
4835 'Max' is an upper limit on how long objects without an explicit
4836 expiry time will be considered fresh.
4838 options: override-expire
4843 ignore-must-revalidate
4850 override-expire enforces min age even if the server
4851 sent an explicit expiry time (e.g., with the
4852 Expires: header or Cache-Control: max-age). Doing this
4853 VIOLATES the HTTP standard. Enabling this feature
4854 could make you liable for problems which it causes.
4856 Note: override-expire does not enforce staleness - it only extends
4857 freshness / min. If the server returns a Expires time which
4858 is longer than your max time, Squid will still consider
4859 the object fresh for that period of time.
4861 override-lastmod enforces min age even on objects
4862 that were modified recently.
4864 reload-into-ims changes client no-cache or ``reload''
4865 to If-Modified-Since requests. Doing this VIOLATES the
4866 HTTP standard. Enabling this feature could make you
4867 liable for problems which it causes.
4869 ignore-reload ignores a client no-cache or ``reload''
4870 header. Doing this VIOLATES the HTTP standard. Enabling
4871 this feature could make you liable for problems which
4874 ignore-no-store ignores any ``Cache-control: no-store''
4875 headers received from a server. Doing this VIOLATES
4876 the HTTP standard. Enabling this feature could make you
4877 liable for problems which it causes.
4879 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4880 headers received from a server. Doing this VIOLATES
4881 the HTTP standard. Enabling this feature could make you
4882 liable for problems which it causes.
4884 ignore-private ignores any ``Cache-control: private''
4885 headers received from a server. Doing this VIOLATES
4886 the HTTP standard. Enabling this feature could make you
4887 liable for problems which it causes.
4889 ignore-auth caches responses to requests with authorization,
4890 as if the originserver had sent ``Cache-control: public''
4891 in the response header. Doing this VIOLATES the HTTP standard.
4892 Enabling this feature could make you liable for problems which
4895 refresh-ims causes squid to contact the origin server
4896 when a client issues an If-Modified-Since request. This
4897 ensures that the client will receive an updated version
4898 if one is available.
4900 store-stale stores responses even if they don't have explicit
4901 freshness or a validator (i.e., Last-Modified or an ETag)
4902 present, or if they're already stale. By default, Squid will
4903 not cache such responses because they usually can't be
4904 reused. Note that such responses will be stale by default.
4906 max-stale=NN provide a maximum staleness factor. Squid won't
4907 serve objects more stale than this even if it failed to
4908 validate the object. Default: use the max_stale global limit.
4910 Basically a cached object is:
4912 FRESH if expires < now, else STALE
4914 FRESH if lm-factor < percent, else STALE
4918 The refresh_pattern lines are checked in the order listed here.
4919 The first entry which matches is used. If none of the entries
4920 match the default will be used.
4922 Note, you must uncomment all the default lines if you want
4923 to change one. The default setting is only active if none is
4929 # Add any of your own refresh_pattern entries above these.
4931 refresh_pattern ^ftp: 1440 20% 10080
4932 refresh_pattern ^gopher: 1440 0% 1440
4933 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4934 refresh_pattern . 0 20% 4320
4938 NAME: quick_abort_min
4942 LOC: Config.quickAbort.min
4945 NAME: quick_abort_max
4949 LOC: Config.quickAbort.max
4952 NAME: quick_abort_pct
4956 LOC: Config.quickAbort.pct
4958 The cache by default continues downloading aborted requests
4959 which are almost completed (less than 16 KB remaining). This
4960 may be undesirable on slow (e.g. SLIP) links and/or very busy
4961 caches. Impatient users may tie up file descriptors and
4962 bandwidth by repeatedly requesting and immediately aborting
4965 When the user aborts a request, Squid will check the
4966 quick_abort values to the amount of data transfered until
4969 If the transfer has less than 'quick_abort_min' KB remaining,
4970 it will finish the retrieval.
4972 If the transfer has more than 'quick_abort_max' KB remaining,
4973 it will abort the retrieval.
4975 If more than 'quick_abort_pct' of the transfer has completed,
4976 it will finish the retrieval.
4978 If you do not want any retrieval to continue after the client
4979 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4982 If you want retrievals to always continue if they are being
4983 cached set 'quick_abort_min' to '-1 KB'.
4986 NAME: read_ahead_gap
4987 COMMENT: buffer-size
4989 LOC: Config.readAheadGap
4992 The amount of data the cache will buffer ahead of what has been
4993 sent to the client when retrieving an object from another server.
4997 IFDEF: USE_HTTP_VIOLATIONS
5000 LOC: Config.negativeTtl
5003 Set the Default Time-to-Live (TTL) for failed requests.
5004 Certain types of failures (such as "connection refused" and
5005 "404 Not Found") are able to be negatively-cached for a short time.
5006 Modern web servers should provide Expires: header, however if they
5007 do not this can provide a minimum TTL.
5008 The default is not to cache errors with unknown expiry details.
5010 Note that this is different from negative caching of DNS lookups.
5012 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5013 this feature could make you liable for problems which it
5017 NAME: positive_dns_ttl
5020 LOC: Config.positiveDnsTtl
5023 Upper limit on how long Squid will cache positive DNS responses.
5024 Default is 6 hours (360 minutes). This directive must be set
5025 larger than negative_dns_ttl.
5028 NAME: negative_dns_ttl
5031 LOC: Config.negativeDnsTtl
5034 Time-to-Live (TTL) for negative caching of failed DNS lookups.
5035 This also sets the lower cache limit on positive lookups.
5036 Minimum value is 1 second, and it is not recommendable to go
5037 much below 10 seconds.
5040 NAME: range_offset_limit
5041 COMMENT: size [acl acl...]
5043 LOC: Config.rangeOffsetLimit
5046 usage: (size) [units] [[!]aclname]
5048 Sets an upper limit on how far (number of bytes) into the file
5049 a Range request may be to cause Squid to prefetch the whole file.
5050 If beyond this limit, Squid forwards the Range request as it is and
5051 the result is NOT cached.
5053 This is to stop a far ahead range request (lets say start at 17MB)
5054 from making Squid fetch the whole object up to that point before
5055 sending anything to the client.
5057 Multiple range_offset_limit lines may be specified, and they will
5058 be searched from top to bottom on each request until a match is found.
5059 The first match found will be used. If no line matches a request, the
5060 default limit of 0 bytes will be used.
5062 'size' is the limit specified as a number of units.
5064 'units' specifies whether to use bytes, KB, MB, etc.
5065 If no units are specified bytes are assumed.
5067 A size of 0 causes Squid to never fetch more than the
5068 client requested. (default)
5070 A size of 'none' causes Squid to always fetch the object from the
5071 beginning so it may cache the result. (2.0 style)
5073 'aclname' is the name of a defined ACL.
5075 NP: Using 'none' as the byte value here will override any quick_abort settings
5076 that may otherwise apply to the range request. The range request will
5077 be fully fetched from start to finish regardless of the client
5078 actions. This affects bandwidth usage.
5081 NAME: minimum_expiry_time
5084 LOC: Config.minimum_expiry_time
5087 The minimum caching time according to (Expires - Date)
5088 headers Squid honors if the object can't be revalidated.
5089 The default is 60 seconds.
5091 In reverse proxy environments it might be desirable to honor
5092 shorter object lifetimes. It is most likely better to make
5093 your server return a meaningful Last-Modified header however.
5095 In ESI environments where page fragments often have short
5096 lifetimes, this will often be best set to 0.
5099 NAME: store_avg_object_size
5103 LOC: Config.Store.avgObjectSize
5105 Average object size, used to estimate number of objects your
5106 cache can hold. The default is 13 KB.
5108 This is used to pre-seed the cache index memory allocation to
5109 reduce expensive reallocate operations while handling clients
5110 traffic. Too-large values may result in memory allocation during
5111 peak traffic, too-small values will result in wasted memory.
5113 Check the cache manager 'info' report metrics for the real
5114 object sizes seen by your Squid before tuning this.
5117 NAME: store_objects_per_bucket
5120 LOC: Config.Store.objectsPerBucket
5122 Target number of objects per bucket in the store hash table.
5123 Lowering this value increases the total number of buckets and
5124 also the storage maintenance rate. The default is 20.
5129 -----------------------------------------------------------------------------
5132 NAME: request_header_max_size
5136 LOC: Config.maxRequestHeaderSize
5138 This specifies the maximum size for HTTP headers in a request.
5139 Request headers are usually relatively small (about 512 bytes).
5140 Placing a limit on the request header size will catch certain
5141 bugs (for example with persistent connections) and possibly
5142 buffer-overflow or denial-of-service attacks.
5145 NAME: reply_header_max_size
5149 LOC: Config.maxReplyHeaderSize
5151 This specifies the maximum size for HTTP headers in a reply.
5152 Reply headers are usually relatively small (about 512 bytes).
5153 Placing a limit on the reply header size will catch certain
5154 bugs (for example with persistent connections) and possibly
5155 buffer-overflow or denial-of-service attacks.
5158 NAME: request_body_max_size
5162 DEFAULT_DOC: No limit.
5163 LOC: Config.maxRequestBodySize
5165 This specifies the maximum size for an HTTP request body.
5166 In other words, the maximum size of a PUT/POST request.
5167 A user who attempts to send a request with a body larger
5168 than this limit receives an "Invalid Request" error message.
5169 If you set this parameter to a zero (the default), there will
5170 be no limit imposed.
5172 See also client_request_buffer_max_size for an alternative
5173 limitation on client uploads which can be configured.
5176 NAME: client_request_buffer_max_size
5180 LOC: Config.maxRequestBufferSize
5182 This specifies the maximum buffer size of a client request.
5183 It prevents squid eating too much memory when somebody uploads
5187 NAME: chunked_request_body_max_size
5191 LOC: Config.maxChunkedRequestBodySize
5193 A broken or confused HTTP/1.1 client may send a chunked HTTP
5194 request to Squid. Squid does not have full support for that
5195 feature yet. To cope with such requests, Squid buffers the
5196 entire request and then dechunks request body to create a
5197 plain HTTP/1.0 request with a known content length. The plain
5198 request is then used by the rest of Squid code as usual.
5200 The option value specifies the maximum size of the buffer used
5201 to hold the request before the conversion. If the chunked
5202 request size exceeds the specified limit, the conversion
5203 fails, and the client receives an "unsupported request" error,
5204 as if dechunking was disabled.
5206 Dechunking is enabled by default. To disable conversion of
5207 chunked requests, set the maximum to zero.
5209 Request dechunking feature and this option in particular are a
5210 temporary hack. When chunking requests and responses are fully
5211 supported, there will be no need to buffer a chunked request.
5215 IFDEF: USE_HTTP_VIOLATIONS
5218 DEFAULT_DOC: Obey RFC 2616.
5219 LOC: Config.accessList.brokenPosts
5221 A list of ACL elements which, if matched, causes Squid to send
5222 an extra CRLF pair after the body of a PUT/POST request.
5224 Some HTTP servers has broken implementations of PUT/POST,
5225 and rely on an extra CRLF pair sent by some WWW clients.
5227 Quote from RFC2616 section 4.1 on this matter:
5229 Note: certain buggy HTTP/1.0 client implementations generate an
5230 extra CRLF's after a POST request. To restate what is explicitly
5231 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
5232 a request with an extra CRLF.
5234 This clause only supports fast acl types.
5235 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5238 acl buggy_server url_regex ^http://....
5239 broken_posts allow buggy_server
5242 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
5245 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
5247 LOC: Adaptation::Config::use_indirect_client
5249 Controls whether the indirect client IP address (instead of the direct
5250 client IP address) is passed to adaptation services.
5252 See also: follow_x_forwarded_for adaptation_send_client_ip
5256 IFDEF: USE_HTTP_VIOLATIONS
5260 LOC: Config.onoff.via
5262 If set (default), Squid will include a Via header in requests and
5263 replies as required by RFC2616.
5269 LOC: Config.onoff.ie_refresh
5272 Microsoft Internet Explorer up until version 5.5 Service
5273 Pack 1 has an issue with transparent proxies, wherein it
5274 is impossible to force a refresh. Turning this on provides
5275 a partial fix to the problem, by causing all IMS-REFRESH
5276 requests from older IE versions to check the origin server
5277 for fresh content. This reduces hit ratio by some amount
5278 (~10% in my experience), but allows users to actually get
5279 fresh content when they want it. Note because Squid
5280 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
5281 of 5.5 is unchanged from old versions of Squid (i.e. a
5282 forced refresh is impossible). Newer versions of IE will,
5283 hopefully, continue to have the new behavior and will be
5284 handled based on that assumption. This option defaults to
5285 the old Squid behavior, which is better for hit ratios but
5286 worse for clients using IE, if they need to be able to
5287 force fresh content.
5290 NAME: vary_ignore_expire
5293 LOC: Config.onoff.vary_ignore_expire
5296 Many HTTP servers supporting Vary gives such objects
5297 immediate expiry time with no cache-control header
5298 when requested by a HTTP/1.0 client. This option
5299 enables Squid to ignore such expiry times until
5300 HTTP/1.1 is fully implemented.
5302 WARNING: If turned on this may eventually cause some
5303 varying objects not intended for caching to get cached.
5306 NAME: request_entities
5308 LOC: Config.onoff.request_entities
5311 Squid defaults to deny GET and HEAD requests with request entities,
5312 as the meaning of such requests are undefined in the HTTP standard
5313 even if not explicitly forbidden.
5315 Set this directive to on if you have clients which insists
5316 on sending request entities in GET or HEAD requests. But be warned
5317 that there is server software (both proxies and web servers) which
5318 can fail to properly process this kind of request which may make you
5319 vulnerable to cache pollution attacks if enabled.
5322 NAME: request_header_access
5323 IFDEF: USE_HTTP_VIOLATIONS
5324 TYPE: http_header_access
5325 LOC: Config.request_header_access
5327 DEFAULT_DOC: No limits.
5329 Usage: request_header_access header_name allow|deny [!]aclname ...
5331 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5332 this feature could make you liable for problems which it
5335 This option replaces the old 'anonymize_headers' and the
5336 older 'http_anonymizer' option with something that is much
5337 more configurable. A list of ACLs for each header name allows
5338 removal of specific header fields under specific conditions.
5340 This option only applies to outgoing HTTP request headers (i.e.,
5341 headers sent by Squid to the next HTTP hop such as a cache peer
5342 or an origin server). The option has no effect during cache hit
5343 detection. The equivalent adaptation vectoring point in ICAP
5344 terminology is post-cache REQMOD.
5346 The option is applied to individual outgoing request header
5347 fields. For each request header field F, Squid uses the first
5348 qualifying sets of request_header_access rules:
5350 1. Rules with header_name equal to F's name.
5351 2. Rules with header_name 'Other', provided F's name is not
5352 on the hard-coded list of commonly used HTTP header names.
5353 3. Rules with header_name 'All'.
5355 Within that qualifying rule set, rule ACLs are checked as usual.
5356 If ACLs of an "allow" rule match, the header field is allowed to
5357 go through as is. If ACLs of a "deny" rule match, the header is
5358 removed and request_header_replace is then checked to identify
5359 if the removed header has a replacement. If no rules within the
5360 set have matching ACLs, the header field is left as is.
5362 For example, to achieve the same behavior as the old
5363 'http_anonymizer standard' option, you should use:
5365 request_header_access From deny all
5366 request_header_access Referer deny all
5367 request_header_access User-Agent deny all
5369 Or, to reproduce the old 'http_anonymizer paranoid' feature
5372 request_header_access Authorization allow all
5373 request_header_access Proxy-Authorization allow all
5374 request_header_access Cache-Control allow all
5375 request_header_access Content-Length allow all
5376 request_header_access Content-Type allow all
5377 request_header_access Date allow all
5378 request_header_access Host allow all
5379 request_header_access If-Modified-Since allow all
5380 request_header_access Pragma allow all
5381 request_header_access Accept allow all
5382 request_header_access Accept-Charset allow all
5383 request_header_access Accept-Encoding allow all
5384 request_header_access Accept-Language allow all
5385 request_header_access Connection allow all
5386 request_header_access All deny all
5388 HTTP reply headers are controlled with the reply_header_access directive.
5390 By default, all headers are allowed (no anonymizing is performed).
5393 NAME: reply_header_access
5394 IFDEF: USE_HTTP_VIOLATIONS
5395 TYPE: http_header_access
5396 LOC: Config.reply_header_access
5398 DEFAULT_DOC: No limits.
5400 Usage: reply_header_access header_name allow|deny [!]aclname ...
5402 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5403 this feature could make you liable for problems which it
5406 This option only applies to reply headers, i.e., from the
5407 server to the client.
5409 This is the same as request_header_access, but in the other
5410 direction. Please see request_header_access for detailed
5413 For example, to achieve the same behavior as the old
5414 'http_anonymizer standard' option, you should use:
5416 reply_header_access Server deny all
5417 reply_header_access WWW-Authenticate deny all
5418 reply_header_access Link deny all
5420 Or, to reproduce the old 'http_anonymizer paranoid' feature
5423 reply_header_access Allow allow all
5424 reply_header_access WWW-Authenticate allow all
5425 reply_header_access Proxy-Authenticate allow all
5426 reply_header_access Cache-Control allow all
5427 reply_header_access Content-Encoding allow all
5428 reply_header_access Content-Length allow all
5429 reply_header_access Content-Type allow all
5430 reply_header_access Date allow all
5431 reply_header_access Expires allow all
5432 reply_header_access Last-Modified allow all
5433 reply_header_access Location allow all
5434 reply_header_access Pragma allow all
5435 reply_header_access Content-Language allow all
5436 reply_header_access Retry-After allow all
5437 reply_header_access Title allow all
5438 reply_header_access Content-Disposition allow all
5439 reply_header_access Connection allow all
5440 reply_header_access All deny all
5442 HTTP request headers are controlled with the request_header_access directive.
5444 By default, all headers are allowed (no anonymizing is
5448 NAME: request_header_replace header_replace
5449 IFDEF: USE_HTTP_VIOLATIONS
5450 TYPE: http_header_replace
5451 LOC: Config.request_header_access
5454 Usage: request_header_replace header_name message
5455 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5457 This option allows you to change the contents of headers
5458 denied with request_header_access above, by replacing them
5459 with some fixed string.
5461 This only applies to request headers, not reply headers.
5463 By default, headers are removed if denied.
5466 NAME: reply_header_replace
5467 IFDEF: USE_HTTP_VIOLATIONS
5468 TYPE: http_header_replace
5469 LOC: Config.reply_header_access
5472 Usage: reply_header_replace header_name message
5473 Example: reply_header_replace Server Foo/1.0
5475 This option allows you to change the contents of headers
5476 denied with reply_header_access above, by replacing them
5477 with some fixed string.
5479 This only applies to reply headers, not request headers.
5481 By default, headers are removed if denied.
5484 NAME: request_header_add
5485 TYPE: HeaderWithAclList
5486 LOC: Config.request_header_add
5489 Usage: request_header_add field-name field-value acl1 [acl2] ...
5490 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5492 This option adds header fields to outgoing HTTP requests (i.e.,
5493 request headers sent by Squid to the next HTTP hop such as a
5494 cache peer or an origin server). The option has no effect during
5495 cache hit detection. The equivalent adaptation vectoring point
5496 in ICAP terminology is post-cache REQMOD.
5498 Field-name is a token specifying an HTTP header name. If a
5499 standard HTTP header name is used, Squid does not check whether
5500 the new header conflicts with any existing headers or violates
5501 HTTP rules. If the request to be modified already contains a
5502 field with the same name, the old field is preserved but the
5503 header field values are not merged.
5505 Field-value is either a token or a quoted string. If quoted
5506 string format is used, then the surrounding quotes are removed
5507 while escape sequences and %macros are processed.
5509 In theory, all of the logformat codes can be used as %macros.
5510 However, unlike logging (which happens at the very end of
5511 transaction lifetime), the transaction may not yet have enough
5512 information to expand a macro when the new header value is needed.
5513 And some information may already be available to Squid but not yet
5514 committed where the macro expansion code can access it (report
5515 such instances!). The macro will be expanded into a single dash
5516 ('-') in such cases. Not all macros have been tested.
5518 One or more Squid ACLs may be specified to restrict header
5519 injection to matching requests. As always in squid.conf, all
5520 ACLs in an option ACL list must be satisfied for the insertion
5521 to happen. The request_header_add option supports fast ACLs
5530 This option used to log custom information about the master
5531 transaction. For example, an admin may configure Squid to log
5532 which "user group" the transaction belongs to, where "user group"
5533 will be determined based on a set of ACLs and not [just]
5534 authentication information.
5535 Values of key/value pairs can be logged using %{key}note macros:
5537 note key value acl ...
5538 logformat myFormat ... %{key}note ...
5541 NAME: relaxed_header_parser
5542 COMMENT: on|off|warn
5544 LOC: Config.onoff.relaxed_header_parser
5547 In the default "on" setting Squid accepts certain forms
5548 of non-compliant HTTP messages where it is unambiguous
5549 what the sending application intended even if the message
5550 is not correctly formatted. The messages is then normalized
5551 to the correct form when forwarded by Squid.
5553 If set to "warn" then a warning will be emitted in cache.log
5554 each time such HTTP error is encountered.
5556 If set to "off" then such HTTP errors will cause the request
5557 or response to be rejected.
5562 -----------------------------------------------------------------------------
5565 NAME: forward_timeout
5568 LOC: Config.Timeout.forward
5571 This parameter specifies how long Squid should at most attempt in
5572 finding a forwarding path for the request before giving up.
5575 NAME: connect_timeout
5578 LOC: Config.Timeout.connect
5581 This parameter specifies how long to wait for the TCP connect to
5582 the requested server or peer to complete before Squid should
5583 attempt to find another path where to forward the request.
5586 NAME: peer_connect_timeout
5589 LOC: Config.Timeout.peer_connect
5592 This parameter specifies how long to wait for a pending TCP
5593 connection to a peer cache. The default is 30 seconds. You
5594 may also set different timeout values for individual neighbors
5595 with the 'connect-timeout' option on a 'cache_peer' line.
5601 LOC: Config.Timeout.read
5604 The read_timeout is applied on server-side connections. After
5605 each successful read(), the timeout will be extended by this
5606 amount. If no data is read again after this amount of time,
5607 the request is aborted and logged with ERR_READ_TIMEOUT. The
5608 default is 15 minutes.
5614 LOC: Config.Timeout.write
5617 This timeout is tracked for all connections that have data
5618 available for writing and are waiting for the socket to become
5619 ready. After each successful write, the timeout is extended by
5620 the configured amount. If Squid has data to write but the
5621 connection is not ready for the configured duration, the
5622 transaction associated with the connection is terminated. The
5623 default is 15 minutes.
5626 NAME: request_timeout
5628 LOC: Config.Timeout.request
5631 How long to wait for complete HTTP request headers after initial
5632 connection establishment.
5635 NAME: client_idle_pconn_timeout persistent_request_timeout
5637 LOC: Config.Timeout.clientIdlePconn
5640 How long to wait for the next HTTP request on a persistent
5641 client connection after the previous request completes.
5644 NAME: client_lifetime
5647 LOC: Config.Timeout.lifetime
5650 The maximum amount of time a client (browser) is allowed to
5651 remain connected to the cache process. This protects the Cache
5652 from having a lot of sockets (and hence file descriptors) tied up
5653 in a CLOSE_WAIT state from remote clients that go away without
5654 properly shutting down (either because of a network failure or
5655 because of a poor client implementation). The default is one
5658 NOTE: The default value is intended to be much larger than any
5659 client would ever need to be connected to your cache. You
5660 should probably change client_lifetime only as a last resort.
5661 If you seem to have many client connections tying up
5662 filedescriptors, we recommend first tuning the read_timeout,
5663 request_timeout, persistent_request_timeout and quick_abort values.
5666 NAME: half_closed_clients
5668 LOC: Config.onoff.half_closed_clients
5671 Some clients may shutdown the sending side of their TCP
5672 connections, while leaving their receiving sides open. Sometimes,
5673 Squid can not tell the difference between a half-closed and a
5674 fully-closed TCP connection.
5676 By default, Squid will immediately close client connections when
5677 read(2) returns "no more data to read."
5679 Change this option to 'on' and Squid will keep open connections
5680 until a read(2) or write(2) on the socket returns an error.
5681 This may show some benefits for reverse proxies. But if not
5682 it is recommended to leave OFF.
5685 NAME: server_idle_pconn_timeout pconn_timeout
5687 LOC: Config.Timeout.serverIdlePconn
5690 Timeout for idle persistent connections to servers and other
5697 LOC: Ident::TheConfig.timeout
5700 Maximum time to wait for IDENT lookups to complete.
5702 If this is too high, and you enabled IDENT lookups from untrusted
5703 users, you might be susceptible to denial-of-service by having
5704 many ident requests going at once.
5707 NAME: shutdown_lifetime
5710 LOC: Config.shutdownLifetime
5713 When SIGTERM or SIGHUP is received, the cache is put into
5714 "shutdown pending" mode until all active sockets are closed.
5715 This value is the lifetime to set for all open descriptors
5716 during shutdown mode. Any active clients after this many
5717 seconds will receive a 'timeout' message.
5721 ADMINISTRATIVE PARAMETERS
5722 -----------------------------------------------------------------------------
5728 LOC: Config.adminEmail
5730 Email-address of local cache manager who will receive
5731 mail if the cache dies. The default is "webmaster".
5737 LOC: Config.EmailFrom
5739 From: email-address for mail sent when the cache dies.
5740 The default is to use 'squid@unique_hostname'.
5742 See also: unique_hostname directive.
5748 LOC: Config.EmailProgram
5750 Email program used to send mail if the cache dies.
5751 The default is "mail". The specified program must comply
5752 with the standard Unix mail syntax:
5753 mail-program recipient < mailfile
5755 Optional command line options can be specified.
5758 NAME: cache_effective_user
5760 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5761 LOC: Config.effectiveUser
5763 If you start Squid as root, it will change its effective/real
5764 UID/GID to the user specified below. The default is to change
5765 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5766 see also; cache_effective_group
5769 NAME: cache_effective_group
5772 DEFAULT_DOC: Use system group memberships of the cache_effective_user account
5773 LOC: Config.effectiveGroup
5775 Squid sets the GID to the effective user's default group ID
5776 (taken from the password file) and supplementary group list
5777 from the groups membership.
5779 If you want Squid to run with a specific GID regardless of
5780 the group memberships of the effective user then set this
5781 to the group (or GID) you want Squid to run as. When set
5782 all other group privileges of the effective user are ignored
5783 and only this GID is effective. If Squid is not started as
5784 root the user starting Squid MUST be member of the specified
5787 This option is not recommended by the Squid Team.
5788 Our preference is for administrators to configure a secure
5789 user account for squid with UID/GID matching system policies.
5792 NAME: httpd_suppress_version_string
5796 LOC: Config.onoff.httpd_suppress_version_string
5798 Suppress Squid version string info in HTTP headers and HTML error pages.
5801 NAME: visible_hostname
5803 LOC: Config.visibleHostname
5805 DEFAULT_DOC: Automatically detect the system host name
5807 If you want to present a special hostname in error messages, etc,
5808 define this. Otherwise, the return value of gethostname()
5809 will be used. If you have multiple caches in a cluster and
5810 get errors about IP-forwarding you must set them to have individual
5811 names with this setting.
5814 NAME: unique_hostname
5816 LOC: Config.uniqueHostname
5818 DEFAULT_DOC: Copy the value from visible_hostname
5820 If you want to have multiple machines with the same
5821 'visible_hostname' you must give each machine a different
5822 'unique_hostname' so forwarding loops can be detected.
5825 NAME: hostname_aliases
5827 LOC: Config.hostnameAliases
5830 A list of other DNS names your cache has.
5838 Minimum umask which should be enforced while the proxy
5839 is running, in addition to the umask set at startup.
5841 For a traditional octal representation of umasks, start
5846 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5847 -----------------------------------------------------------------------------
5849 This section contains parameters for the (optional) cache
5850 announcement service. This service is provided to help
5851 cache administrators locate one another in order to join or
5852 create cache hierarchies.
5854 An 'announcement' message is sent (via UDP) to the registration
5855 service by Squid. By default, the announcement message is NOT
5856 SENT unless you enable it with 'announce_period' below.
5858 The announcement message includes your hostname, plus the
5859 following information from this configuration file:
5865 All current information is processed regularly and made
5866 available on the Web at http://www.ircache.net/Cache/Tracker/.
5869 NAME: announce_period
5871 LOC: Config.Announce.period
5873 DEFAULT_DOC: Announcement messages disabled.
5875 This is how frequently to send cache announcements.
5877 To enable announcing your cache, just set an announce period.
5880 announce_period 1 day
5885 DEFAULT: tracker.ircache.net
5886 LOC: Config.Announce.host
5888 Set the hostname where announce registration messages will be sent.
5890 See also announce_port and announce_file
5896 LOC: Config.Announce.file
5898 The contents of this file will be included in the announce
5899 registration messages.
5905 LOC: Config.Announce.port
5907 Set the port where announce registration messages will be sent.
5909 See also announce_host and announce_file
5913 HTTPD-ACCELERATOR OPTIONS
5914 -----------------------------------------------------------------------------
5917 NAME: httpd_accel_surrogate_id
5920 DEFAULT_DOC: visible_hostname is used if no specific ID is set.
5921 LOC: Config.Accel.surrogate_id
5923 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5924 need an identification token to allow control targeting. Because
5925 a farm of surrogates may all perform the same tasks, they may share
5926 an identification token.
5929 NAME: http_accel_surrogate_remote
5933 LOC: Config.onoff.surrogate_is_remote
5935 Remote surrogates (such as those in a CDN) honour the header
5936 "Surrogate-Control: no-store-remote".
5938 Set this to on to have squid behave as a remote surrogate.
5942 IFDEF: USE_SQUID_ESI
5943 COMMENT: libxml2|expat|custom
5945 LOC: ESIParser::Type
5948 ESI markup is not strictly XML compatible. The custom ESI parser
5949 will give higher performance, but cannot handle non ASCII character
5954 DELAY POOL PARAMETERS
5955 -----------------------------------------------------------------------------
5959 TYPE: delay_pool_count
5961 IFDEF: USE_DELAY_POOLS
5964 This represents the number of delay pools to be used. For example,
5965 if you have one class 2 delay pool and one class 3 delays pool, you
5966 have a total of 2 delay pools.
5968 See also delay_parameters, delay_class, delay_access for pool
5969 configuration details.
5973 TYPE: delay_pool_class
5975 IFDEF: USE_DELAY_POOLS
5978 This defines the class of each delay pool. There must be exactly one
5979 delay_class line for each delay pool. For example, to define two
5980 delay pools, one of class 2 and one of class 3, the settings above
5984 delay_pools 4 # 4 delay pools
5985 delay_class 1 2 # pool 1 is a class 2 pool
5986 delay_class 2 3 # pool 2 is a class 3 pool
5987 delay_class 3 4 # pool 3 is a class 4 pool
5988 delay_class 4 5 # pool 4 is a class 5 pool
5990 The delay pool classes are:
5992 class 1 Everything is limited by a single aggregate
5995 class 2 Everything is limited by a single aggregate
5996 bucket as well as an "individual" bucket chosen
5997 from bits 25 through 32 of the IPv4 address.
5999 class 3 Everything is limited by a single aggregate
6000 bucket as well as a "network" bucket chosen
6001 from bits 17 through 24 of the IP address and a
6002 "individual" bucket chosen from bits 17 through
6003 32 of the IPv4 address.
6005 class 4 Everything in a class 3 delay pool, with an
6006 additional limit on a per user basis. This
6007 only takes effect if the username is established
6008 in advance - by forcing authentication in your
6011 class 5 Requests are grouped according their tag (see
6012 external_acl's tag= reply).
6015 Each pool also requires a delay_parameters directive to configure the pool size
6016 and speed limits used whenever the pool is applied to a request. Along with
6017 a set of delay_access directives to determine when it is used.
6019 NOTE: If an IP address is a.b.c.d
6020 -> bits 25 through 32 are "d"
6021 -> bits 17 through 24 are "c"
6022 -> bits 17 through 32 are "c * 256 + d"
6024 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
6025 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
6027 This clause only supports fast acl types.
6028 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6030 See also delay_parameters and delay_access.
6034 TYPE: delay_pool_access
6036 DEFAULT_DOC: Deny using the pool, unless allow rules exist in squid.conf for the pool.
6037 IFDEF: USE_DELAY_POOLS
6040 This is used to determine which delay pool a request falls into.
6042 delay_access is sorted per pool and the matching starts with pool 1,
6043 then pool 2, ..., and finally pool N. The first delay pool where the
6044 request is allowed is selected for the request. If it does not allow
6045 the request to any pool then the request is not delayed (default).
6047 For example, if you want some_big_clients in delay
6048 pool 1 and lotsa_little_clients in delay pool 2:
6050 delay_access 1 allow some_big_clients
6051 delay_access 1 deny all
6052 delay_access 2 allow lotsa_little_clients
6053 delay_access 2 deny all
6054 delay_access 3 allow authenticated_clients
6056 See also delay_parameters and delay_class.
6060 NAME: delay_parameters
6061 TYPE: delay_pool_rates
6063 IFDEF: USE_DELAY_POOLS
6066 This defines the parameters for a delay pool. Each delay pool has
6067 a number of "buckets" associated with it, as explained in the
6068 description of delay_class.
6070 For a class 1 delay pool, the syntax is:
6072 delay_parameters pool aggregate
6074 For a class 2 delay pool:
6076 delay_parameters pool aggregate individual
6078 For a class 3 delay pool:
6080 delay_parameters pool aggregate network individual
6082 For a class 4 delay pool:
6084 delay_parameters pool aggregate network individual user
6086 For a class 5 delay pool:
6088 delay_parameters pool tagrate
6090 The option variables are:
6092 pool a pool number - ie, a number between 1 and the
6093 number specified in delay_pools as used in
6096 aggregate the speed limit parameters for the aggregate bucket
6099 individual the speed limit parameters for the individual
6100 buckets (class 2, 3).
6102 network the speed limit parameters for the network buckets
6105 user the speed limit parameters for the user buckets
6108 tagrate the speed limit parameters for the tag buckets
6111 A pair of delay parameters is written restore/maximum, where restore is
6112 the number of bytes (not bits - modem and network speeds are usually
6113 quoted in bits) per second placed into the bucket, and maximum is the
6114 maximum number of bytes which can be in the bucket at any time.
6116 There must be one delay_parameters line for each delay pool.
6119 For example, if delay pool number 1 is a class 2 delay pool as in the
6120 above example, and is being used to strictly limit each host to 64Kbit/sec
6121 (plus overheads), with no overall limit, the line is:
6123 delay_parameters 1 -1/-1 8000/8000
6125 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
6127 Note that the figure -1 is used to represent "unlimited".
6130 And, if delay pool number 2 is a class 3 delay pool as in the above
6131 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
6132 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
6133 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
6134 to permit a decent web page to be downloaded at a decent speed
6135 (if the network is not being limited due to overuse) but slow down
6136 large downloads more significantly:
6138 delay_parameters 2 32000/32000 8000/8000 600/8000
6140 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
6141 8 x 8000 KByte/sec -> 64Kbit/sec.
6142 8 x 600 Byte/sec -> 4800bit/sec.
6145 Finally, for a class 4 delay pool as in the example - each user will
6146 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
6148 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
6151 See also delay_class and delay_access.
6155 NAME: delay_initial_bucket_level
6156 COMMENT: (percent, 0-100)
6159 IFDEF: USE_DELAY_POOLS
6160 LOC: Config.Delay.initial
6162 The initial bucket percentage is used to determine how much is put
6163 in each bucket when squid starts, is reconfigured, or first notices
6164 a host accessing it (in class 2 and class 3, individual hosts and
6165 networks only have buckets associated with them once they have been
6170 CLIENT DELAY POOL PARAMETERS
6171 -----------------------------------------------------------------------------
6174 NAME: client_delay_pools
6175 TYPE: client_delay_pool_count
6177 IFDEF: USE_DELAY_POOLS
6178 LOC: Config.ClientDelay
6180 This option specifies the number of client delay pools used. It must
6181 preceed other client_delay_* options.
6184 client_delay_pools 2
6186 See also client_delay_parameters and client_delay_access.
6189 NAME: client_delay_initial_bucket_level
6190 COMMENT: (percent, 0-no_limit)
6193 IFDEF: USE_DELAY_POOLS
6194 LOC: Config.ClientDelay.initial
6196 This option determines the initial bucket size as a percentage of
6197 max_bucket_size from client_delay_parameters. Buckets are created
6198 at the time of the "first" connection from the matching IP. Idle
6199 buckets are periodically deleted up.
6201 You can specify more than 100 percent but note that such "oversized"
6202 buckets are not refilled until their size goes down to max_bucket_size
6203 from client_delay_parameters.
6206 client_delay_initial_bucket_level 50
6209 NAME: client_delay_parameters
6210 TYPE: client_delay_pool_rates
6212 IFDEF: USE_DELAY_POOLS
6213 LOC: Config.ClientDelay
6216 This option configures client-side bandwidth limits using the
6219 client_delay_parameters pool speed_limit max_bucket_size
6221 pool is an integer ID used for client_delay_access matching.
6223 speed_limit is bytes added to the bucket per second.
6225 max_bucket_size is the maximum size of a bucket, enforced after any
6226 speed_limit additions.
6228 Please see the delay_parameters option for more information and
6232 client_delay_parameters 1 1024 2048
6233 client_delay_parameters 2 51200 16384
6235 See also client_delay_access.
6239 NAME: client_delay_access
6240 TYPE: client_delay_pool_access
6242 DEFAULT_DOC: Deny use of the pool, unless allow rules exist in squid.conf for the pool.
6243 IFDEF: USE_DELAY_POOLS
6244 LOC: Config.ClientDelay
6246 This option determines the client-side delay pool for the
6249 client_delay_access pool_ID allow|deny acl_name
6251 All client_delay_access options are checked in their pool ID
6252 order, starting with pool 1. The first checked pool with allowed
6253 request is selected for the request. If no ACL matches or there
6254 are no client_delay_access options, the request bandwidth is not
6257 The ACL-selected pool is then used to find the
6258 client_delay_parameters for the request. Client-side pools are
6259 not used to aggregate clients. Clients are always aggregated
6260 based on their source IP addresses (one bucket per source IP).
6262 This clause only supports fast acl types.
6263 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6264 Additionally, only the client TCP connection details are available.
6265 ACLs testing HTTP properties will not work.
6267 Please see delay_access for more examples.
6270 client_delay_access 1 allow low_rate_network
6271 client_delay_access 2 allow vips_network
6274 See also client_delay_parameters and client_delay_pools.
6278 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
6279 -----------------------------------------------------------------------------
6284 LOC: Config.Wccp.router
6286 DEFAULT_DOC: WCCP disabled.
6289 Use this option to define your WCCP ``home'' router for
6292 wccp_router supports a single WCCP(v1) router
6294 wccp2_router supports multiple WCCPv2 routers
6296 only one of the two may be used at the same time and defines
6297 which version of WCCP to use.
6301 TYPE: IpAddress_list
6302 LOC: Config.Wccp2.router
6304 DEFAULT_DOC: WCCPv2 disabled.
6307 Use this option to define your WCCP ``home'' router for
6310 wccp_router supports a single WCCP(v1) router
6312 wccp2_router supports multiple WCCPv2 routers
6314 only one of the two may be used at the same time and defines
6315 which version of WCCP to use.
6320 LOC: Config.Wccp.version
6324 This directive is only relevant if you need to set up WCCP(v1)
6325 to some very old and end-of-life Cisco routers. In all other
6326 setups it must be left unset or at the default setting.
6327 It defines an internal version in the WCCP(v1) protocol,
6328 with version 4 being the officially documented protocol.
6330 According to some users, Cisco IOS 11.2 and earlier only
6331 support WCCP version 3. If you're using that or an earlier
6332 version of IOS, you may need to change this value to 3, otherwise
6333 do not specify this parameter.
6336 NAME: wccp2_rebuild_wait
6338 LOC: Config.Wccp2.rebuildwait
6342 If this is enabled Squid will wait for the cache dir rebuild to finish
6343 before sending the first wccp2 HereIAm packet
6346 NAME: wccp2_forwarding_method
6348 LOC: Config.Wccp2.forwarding_method
6352 WCCP2 allows the setting of forwarding methods between the
6353 router/switch and the cache. Valid values are as follows:
6355 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6356 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6358 Currently (as of IOS 12.4) cisco routers only support GRE.
6359 Cisco switches only support the L2 redirect assignment method.
6362 NAME: wccp2_return_method
6364 LOC: Config.Wccp2.return_method
6368 WCCP2 allows the setting of return methods between the
6369 router/switch and the cache for packets that the cache
6370 decides not to handle. Valid values are as follows:
6372 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6373 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6375 Currently (as of IOS 12.4) cisco routers only support GRE.
6376 Cisco switches only support the L2 redirect assignment.
6378 If the "ip wccp redirect exclude in" command has been
6379 enabled on the cache interface, then it is still safe for
6380 the proxy server to use a l2 redirect method even if this
6381 option is set to GRE.
6384 NAME: wccp2_assignment_method
6386 LOC: Config.Wccp2.assignment_method
6390 WCCP2 allows the setting of methods to assign the WCCP hash
6391 Valid values are as follows:
6393 hash - Hash assignment
6394 mask - Mask assignment
6396 As a general rule, cisco routers support the hash assignment method
6397 and cisco switches support the mask assignment method.
6402 LOC: Config.Wccp2.info
6403 DEFAULT_IF_NONE: standard 0
6404 DEFAULT_DOC: Use the 'web-cache' standard service.
6407 WCCP2 allows for multiple traffic services. There are two
6408 types: "standard" and "dynamic". The standard type defines
6409 one service id - http (id 0). The dynamic service ids can be from
6410 51 to 255 inclusive. In order to use a dynamic service id
6411 one must define the type of traffic to be redirected; this is done
6412 using the wccp2_service_info option.
6414 The "standard" type does not require a wccp2_service_info option,
6415 just specifying the service id will suffice.
6417 MD5 service authentication can be enabled by adding
6418 "password=<password>" to the end of this service declaration.
6422 wccp2_service standard 0 # for the 'web-cache' standard service
6423 wccp2_service dynamic 80 # a dynamic service type which will be
6424 # fleshed out with subsequent options.
6425 wccp2_service standard 0 password=foo
6428 NAME: wccp2_service_info
6429 TYPE: wccp2_service_info
6430 LOC: Config.Wccp2.info
6434 Dynamic WCCPv2 services require further information to define the
6435 traffic you wish to have diverted.
6439 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6440 priority=<priority> ports=<port>,<port>..
6442 The relevant WCCPv2 flags:
6443 + src_ip_hash, dst_ip_hash
6444 + source_port_hash, dst_port_hash
6445 + src_ip_alt_hash, dst_ip_alt_hash
6446 + src_port_alt_hash, dst_port_alt_hash
6449 The port list can be one to eight entries.
6453 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6454 priority=240 ports=80
6456 Note: the service id must have been defined by a previous
6457 'wccp2_service dynamic <id>' entry.
6462 LOC: Config.Wccp2.weight
6466 Each cache server gets assigned a set of the destination
6467 hash proportional to their weight.
6472 LOC: Config.Wccp.address
6474 DEFAULT_DOC: Address selected by the operating system.
6477 Use this option if you require WCCPv2 to use a specific
6480 The default behavior is to not bind to any specific address.
6485 LOC: Config.Wccp2.address
6487 DEFAULT_DOC: Address selected by the operating system.
6490 Use this option if you require WCCP to use a specific
6493 The default behavior is to not bind to any specific address.
6497 PERSISTENT CONNECTION HANDLING
6498 -----------------------------------------------------------------------------
6500 Also see "pconn_timeout" in the TIMEOUTS section
6503 NAME: client_persistent_connections
6505 LOC: Config.onoff.client_pconns
6508 Persistent connection support for clients.
6509 Squid uses persistent connections (when allowed). You can use
6510 this option to disable persistent connections with clients.
6513 NAME: server_persistent_connections
6515 LOC: Config.onoff.server_pconns
6518 Persistent connection support for servers.
6519 Squid uses persistent connections (when allowed). You can use
6520 this option to disable persistent connections with servers.
6523 NAME: persistent_connection_after_error
6525 LOC: Config.onoff.error_pconns
6528 With this directive the use of persistent connections after
6529 HTTP errors can be disabled. Useful if you have clients
6530 who fail to handle errors on persistent connections proper.
6533 NAME: detect_broken_pconn
6535 LOC: Config.onoff.detect_broken_server_pconns
6538 Some servers have been found to incorrectly signal the use
6539 of HTTP/1.0 persistent connections even on replies not
6540 compatible, causing significant delays. This server problem
6541 has mostly been seen on redirects.
6543 By enabling this directive Squid attempts to detect such
6544 broken replies and automatically assume the reply is finished
6545 after 10 seconds timeout.
6549 CACHE DIGEST OPTIONS
6550 -----------------------------------------------------------------------------
6553 NAME: digest_generation
6554 IFDEF: USE_CACHE_DIGESTS
6556 LOC: Config.onoff.digest_generation
6559 This controls whether the server will generate a Cache Digest
6560 of its contents. By default, Cache Digest generation is
6561 enabled if Squid is compiled with --enable-cache-digests defined.
6564 NAME: digest_bits_per_entry
6565 IFDEF: USE_CACHE_DIGESTS
6567 LOC: Config.digest.bits_per_entry
6570 This is the number of bits of the server's Cache Digest which
6571 will be associated with the Digest entry for a given HTTP
6572 Method and URL (public key) combination. The default is 5.
6575 NAME: digest_rebuild_period
6576 IFDEF: USE_CACHE_DIGESTS
6579 LOC: Config.digest.rebuild_period
6582 This is the wait time between Cache Digest rebuilds.
6585 NAME: digest_rewrite_period
6587 IFDEF: USE_CACHE_DIGESTS
6589 LOC: Config.digest.rewrite_period
6592 This is the wait time between Cache Digest writes to
6596 NAME: digest_swapout_chunk_size
6599 IFDEF: USE_CACHE_DIGESTS
6600 LOC: Config.digest.swapout_chunk_size
6603 This is the number of bytes of the Cache Digest to write to
6604 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6608 NAME: digest_rebuild_chunk_percentage
6609 COMMENT: (percent, 0-100)
6610 IFDEF: USE_CACHE_DIGESTS
6612 LOC: Config.digest.rebuild_chunk_percentage
6615 This is the percentage of the Cache Digest to be scanned at a
6616 time. By default it is set to 10% of the Cache Digest.
6621 -----------------------------------------------------------------------------
6626 LOC: Config.Port.snmp
6628 DEFAULT_DOC: SNMP disabled.
6631 The port number where Squid listens for SNMP requests. To enable
6632 SNMP support set this to a suitable port number. Port number
6633 3401 is often used for the Squid SNMP agent. By default it's
6634 set to "0" (disabled)
6642 LOC: Config.accessList.snmp
6644 DEFAULT_DOC: Deny, unless rules exist in squid.conf.
6647 Allowing or denying access to the SNMP port.
6649 All access to the agent is denied by default.
6652 snmp_access allow|deny [!]aclname ...
6654 This clause only supports fast acl types.
6655 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6658 snmp_access allow snmppublic localhost
6659 snmp_access deny all
6662 NAME: snmp_incoming_address
6664 LOC: Config.Addrs.snmp_incoming
6666 DEFAULT_DOC: Accept SNMP packets from all machine interfaces.
6669 Just like 'udp_incoming_address', but for the SNMP port.
6671 snmp_incoming_address is used for the SNMP socket receiving
6672 messages from SNMP agents.
6674 The default snmp_incoming_address is to listen on all
6675 available network interfaces.
6678 NAME: snmp_outgoing_address
6680 LOC: Config.Addrs.snmp_outgoing
6682 DEFAULT_DOC: Use snmp_incoming_address or an address selected by the operating system.
6685 Just like 'udp_outgoing_address', but for the SNMP port.
6687 snmp_outgoing_address is used for SNMP packets returned to SNMP
6690 If snmp_outgoing_address is not set it will use the same socket
6691 as snmp_incoming_address. Only change this if you want to have
6692 SNMP replies sent using another address than where this Squid
6693 listens for SNMP queries.
6695 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6696 the same value since they both use the same port.
6701 -----------------------------------------------------------------------------
6704 NAME: icp_port udp_port
6707 DEFAULT_DOC: ICP disabled.
6708 LOC: Config.Port.icp
6710 The port number where Squid sends and receives ICP queries to
6711 and from neighbor caches. The standard UDP port for ICP is 3130.
6714 icp_port @DEFAULT_ICP_PORT@
6721 DEFAULT_DOC: HTCP disabled.
6722 LOC: Config.Port.htcp
6724 The port number where Squid sends and receives HTCP queries to
6725 and from neighbor caches. To turn it on you want to set it to
6732 NAME: log_icp_queries
6736 LOC: Config.onoff.log_udp
6738 If set, ICP queries are logged to access.log. You may wish
6739 do disable this if your ICP load is VERY high to speed things
6740 up or to simplify log analysis.
6743 NAME: udp_incoming_address
6745 LOC:Config.Addrs.udp_incoming
6747 DEFAULT_DOC: Accept packets from all machine interfaces.
6749 udp_incoming_address is used for UDP packets received from other
6752 The default behavior is to not bind to any specific address.
6754 Only change this if you want to have all UDP queries received on
6755 a specific interface/address.
6757 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6758 modules. Altering it will affect all of them in the same manner.
6760 see also; udp_outgoing_address
6762 NOTE, udp_incoming_address and udp_outgoing_address can not
6763 have the same value since they both use the same port.
6766 NAME: udp_outgoing_address
6768 LOC: Config.Addrs.udp_outgoing
6770 DEFAULT_DOC: Use udp_incoming_address or an address selected by the operating system.
6772 udp_outgoing_address is used for UDP packets sent out to other
6775 The default behavior is to not bind to any specific address.
6777 Instead it will use the same socket as udp_incoming_address.
6778 Only change this if you want to have UDP queries sent using another
6779 address than where this Squid listens for UDP queries from other
6782 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6783 modules. Altering it will affect all of them in the same manner.
6785 see also; udp_incoming_address
6787 NOTE, udp_incoming_address and udp_outgoing_address can not
6788 have the same value since they both use the same port.
6795 LOC: Config.onoff.icp_hit_stale
6797 If you want to return ICP_HIT for stale cache objects, set this
6798 option to 'on'. If you have sibling relationships with caches
6799 in other administrative domains, this should be 'off'. If you only
6800 have sibling relationships with caches under your control,
6801 it is probably okay to set this to 'on'.
6802 If set to 'on', your siblings should use the option "allow-miss"
6803 on their cache_peer lines for connecting to you.
6806 NAME: minimum_direct_hops
6809 LOC: Config.minDirectHops
6811 If using the ICMP pinging stuff, do direct fetches for sites
6812 which are no more than this many hops away.
6815 NAME: minimum_direct_rtt
6819 LOC: Config.minDirectRtt
6821 If using the ICMP pinging stuff, do direct fetches for sites
6822 which are no more than this many rtt milliseconds away.
6828 LOC: Config.Netdb.low
6830 The low water mark for the ICMP measurement database.
6832 Note: high watermark controlled by netdb_high directive.
6834 These watermarks are counts, not percents. The defaults are
6835 (low) 900 and (high) 1000. When the high water mark is
6836 reached, database entries will be deleted until the low
6843 LOC: Config.Netdb.high
6845 The high water mark for the ICMP measurement database.
6847 Note: low watermark controlled by netdb_low directive.
6849 These watermarks are counts, not percents. The defaults are
6850 (low) 900 and (high) 1000. When the high water mark is
6851 reached, database entries will be deleted until the low
6855 NAME: netdb_ping_period
6857 LOC: Config.Netdb.period
6860 The minimum period for measuring a site. There will be at
6861 least this much delay between successive pings to the same
6862 network. The default is five minutes.
6869 LOC: Config.onoff.query_icmp
6871 If you want to ask your peers to include ICMP data in their ICP
6872 replies, enable this option.
6874 If your peer has configured Squid (during compilation) with
6875 '--enable-icmp' that peer will send ICMP pings to origin server
6876 sites of the URLs it receives. If you enable this option the
6877 ICP replies from that peer will include the ICMP data (if available).
6878 Then, when choosing a parent cache, Squid will choose the parent with
6879 the minimal RTT to the origin server. When this happens, the
6880 hierarchy field of the access.log will be
6881 "CLOSEST_PARENT_MISS". This option is off by default.
6884 NAME: test_reachability
6888 LOC: Config.onoff.test_reachability
6890 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6891 instead of ICP_MISS if the target host is NOT in the ICMP
6892 database, or has a zero RTT.
6895 NAME: icp_query_timeout
6898 DEFAULT_DOC: Dynamic detection.
6900 LOC: Config.Timeout.icp_query
6902 Normally Squid will automatically determine an optimal ICP
6903 query timeout value based on the round-trip-time of recent ICP
6904 queries. If you want to override the value determined by
6905 Squid, set this 'icp_query_timeout' to a non-zero value. This
6906 value is specified in MILLISECONDS, so, to use a 2-second
6907 timeout (the old default), you would write:
6909 icp_query_timeout 2000
6912 NAME: maximum_icp_query_timeout
6916 LOC: Config.Timeout.icp_query_max
6918 Normally the ICP query timeout is determined dynamically. But
6919 sometimes it can lead to very large values (say 5 seconds).
6920 Use this option to put an upper limit on the dynamic timeout
6921 value. Do NOT use this option to always use a fixed (instead
6922 of a dynamic) timeout value. To set a fixed timeout see the
6923 'icp_query_timeout' directive.
6926 NAME: minimum_icp_query_timeout
6930 LOC: Config.Timeout.icp_query_min
6932 Normally the ICP query timeout is determined dynamically. But
6933 sometimes it can lead to very small timeouts, even lower than
6934 the normal latency variance on your link due to traffic.
6935 Use this option to put an lower limit on the dynamic timeout
6936 value. Do NOT use this option to always use a fixed (instead
6937 of a dynamic) timeout value. To set a fixed timeout see the
6938 'icp_query_timeout' directive.
6941 NAME: background_ping_rate
6945 LOC: Config.backgroundPingRate
6947 Controls how often the ICP pings are sent to siblings that
6948 have background-ping set.
6952 MULTICAST ICP OPTIONS
6953 -----------------------------------------------------------------------------
6958 LOC: Config.mcast_group_list
6961 This tag specifies a list of multicast groups which your server
6962 should join to receive multicasted ICP queries.
6964 NOTE! Be very careful what you put here! Be sure you
6965 understand the difference between an ICP _query_ and an ICP
6966 _reply_. This option is to be set only if you want to RECEIVE
6967 multicast queries. Do NOT set this option to SEND multicast
6968 ICP (use cache_peer for that). ICP replies are always sent via
6969 unicast, so this option does not affect whether or not you will
6970 receive replies from multicast group members.
6972 You must be very careful to NOT use a multicast address which
6973 is already in use by another group of caches.
6975 If you are unsure about multicast, please read the Multicast
6976 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6978 Usage: mcast_groups 239.128.16.128 224.0.1.20
6980 By default, Squid doesn't listen on any multicast groups.
6983 NAME: mcast_miss_addr
6984 IFDEF: MULTICAST_MISS_STREAM
6986 LOC: Config.mcast_miss.addr
6988 DEFAULT_DOC: disabled.
6990 If you enable this option, every "cache miss" URL will
6991 be sent out on the specified multicast address.
6993 Do not enable this option unless you are are absolutely
6994 certain you understand what you are doing.
6997 NAME: mcast_miss_ttl
6998 IFDEF: MULTICAST_MISS_STREAM
7000 LOC: Config.mcast_miss.ttl
7003 This is the time-to-live value for packets multicasted
7004 when multicasting off cache miss URLs is enabled. By
7005 default this is set to 'site scope', i.e. 16.
7008 NAME: mcast_miss_port
7009 IFDEF: MULTICAST_MISS_STREAM
7011 LOC: Config.mcast_miss.port
7014 This is the port number to be used in conjunction with
7018 NAME: mcast_miss_encode_key
7019 IFDEF: MULTICAST_MISS_STREAM
7021 LOC: Config.mcast_miss.encode_key
7022 DEFAULT: XXXXXXXXXXXXXXXX
7024 The URLs that are sent in the multicast miss stream are
7025 encrypted. This is the encryption key.
7028 NAME: mcast_icp_query_timeout
7032 LOC: Config.Timeout.mcast_icp_query
7034 For multicast peers, Squid regularly sends out ICP "probes" to
7035 count how many other peers are listening on the given multicast
7036 address. This value specifies how long Squid should wait to
7037 count all the replies. The default is 2000 msec, or 2
7042 INTERNAL ICON OPTIONS
7043 -----------------------------------------------------------------------------
7046 NAME: icon_directory
7048 LOC: Config.icons.directory
7049 DEFAULT: @DEFAULT_ICON_DIR@
7051 Where the icons are stored. These are normally kept in
7055 NAME: global_internal_static
7057 LOC: Config.onoff.global_internal_static
7060 This directive controls is Squid should intercept all requests for
7061 /squid-internal-static/ no matter which host the URL is requesting
7062 (default on setting), or if nothing special should be done for
7063 such URLs (off setting). The purpose of this directive is to make
7064 icons etc work better in complex cache hierarchies where it may
7065 not always be possible for all corners in the cache mesh to reach
7066 the server generating a directory listing.
7069 NAME: short_icon_urls
7071 LOC: Config.icons.use_short_names
7074 If this is enabled Squid will use short URLs for icons.
7075 If disabled it will revert to the old behavior of including
7076 it's own name and port in the URL.
7078 If you run a complex cache hierarchy with a mix of Squid and
7079 other proxies you may need to disable this directive.
7084 -----------------------------------------------------------------------------
7087 NAME: error_directory
7089 LOC: Config.errorDirectory
7091 DEFAULT_DOC: Send error pages in the clients preferred language
7093 If you wish to create your own versions of the default
7094 error files to customize them to suit your company copy
7095 the error/template files to another directory and point
7098 WARNING: This option will disable multi-language support
7099 on error pages if used.
7101 The squid developers are interested in making squid available in
7102 a wide variety of languages. If you are making translations for a
7103 language that Squid does not currently provide please consider
7104 contributing your translation back to the project.
7105 http://wiki.squid-cache.org/Translations
7107 The squid developers working on translations are happy to supply drop-in
7108 translated error files in exchange for any new language contributions.
7111 NAME: error_default_language
7112 IFDEF: USE_ERR_LOCALES
7114 LOC: Config.errorDefaultLanguage
7116 DEFAULT_DOC: Generate English language pages.
7118 Set the default language which squid will send error pages in
7119 if no existing translation matches the clients language
7122 If unset (default) generic English will be used.
7124 The squid developers are interested in making squid available in
7125 a wide variety of languages. If you are interested in making
7126 translations for any language see the squid wiki for details.
7127 http://wiki.squid-cache.org/Translations
7130 NAME: error_log_languages
7131 IFDEF: USE_ERR_LOCALES
7133 LOC: Config.errorLogMissingLanguages
7136 Log to cache.log what languages users are attempting to
7137 auto-negotiate for translations.
7139 Successful negotiations are not logged. Only failures
7140 have meaning to indicate that Squid may need an upgrade
7141 of its error page translations.
7144 NAME: err_page_stylesheet
7146 LOC: Config.errorStylesheet
7147 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
7149 CSS Stylesheet to pattern the display of Squid default error pages.
7151 For information on CSS see http://www.w3.org/Style/CSS/
7156 LOC: Config.errHtmlText
7159 HTML text to include in error messages. Make this a "mailto"
7160 URL to your admin address, or maybe just a link to your
7161 organizations Web page.
7163 To include this in your error messages, you must rewrite
7164 the error template files (found in the "errors" directory).
7165 Wherever you want the 'err_html_text' line to appear,
7166 insert a %L tag in the error template file.
7169 NAME: email_err_data
7172 LOC: Config.onoff.emailErrData
7175 If enabled, information about the occurred error will be
7176 included in the mailto links of the ERR pages (if %W is set)
7177 so that the email body contains the data.
7178 Syntax is <A HREF="mailto:%w%W">%w</A>
7183 LOC: Config.denyInfoList
7186 Usage: deny_info err_page_name acl
7187 or deny_info http://... acl
7188 or deny_info TCP_RESET acl
7190 This can be used to return a ERR_ page for requests which
7191 do not pass the 'http_access' rules. Squid remembers the last
7192 acl it evaluated in http_access, and if a 'deny_info' line exists
7193 for that ACL Squid returns a corresponding error page.
7195 The acl is typically the last acl on the http_access deny line which
7196 denied access. The exceptions to this rule are:
7197 - When Squid needs to request authentication credentials. It's then
7198 the first authentication related acl encountered
7199 - When none of the http_access lines matches. It's then the last
7200 acl processed on the last http_access line.
7201 - When the decision to deny access was made by an adaptation service,
7202 the acl name is the corresponding eCAP or ICAP service_name.
7204 NP: If providing your own custom error pages with error_directory
7205 you may also specify them by your custom file name:
7206 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
7208 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
7209 may be specified by prefixing the file name with the code and a colon.
7210 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
7212 Alternatively you can tell Squid to reset the TCP connection
7213 by specifying TCP_RESET.
7215 Or you can specify an error URL or URL pattern. The browsers will
7216 get redirected to the specified URL after formatting tags have
7217 been replaced. Redirect will be done with 302 or 307 according to
7218 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
7219 the URL. e.g. 303:http://example.com/
7222 %a - username (if available. Password NOT included)
7225 %E - Error description
7227 %H - Request domain name
7228 %i - Client IP Address
7230 %o - Message result from external ACL helper
7231 %p - Request Port number
7232 %P - Request Protocol name
7233 %R - Request URL path
7234 %T - Timestamp in RFC 1123 format
7235 %U - Full canonical URL from client
7236 (HTTPS URLs terminate with *)
7237 %u - Full canonical URL from client
7238 %w - Admin email from squid.conf
7240 %% - Literal percent (%) code
7245 OPTIONS INFLUENCING REQUEST FORWARDING
7246 -----------------------------------------------------------------------------
7249 NAME: nonhierarchical_direct
7251 LOC: Config.onoff.nonhierarchical_direct
7254 By default, Squid will send any non-hierarchical requests
7255 (matching hierarchy_stoplist or not cacheable request type) direct
7258 When this is set to "off", Squid will prefer to send these
7259 requests to parents.
7261 Note that in most configurations, by turning this off you will only
7262 add latency to these request without any improvement in global hit
7265 This option only sets a preference. If the parent is unavailable a
7266 direct connection to the origin server may still be attempted. To
7267 completely prevent direct connections use never_direct.
7272 LOC: Config.onoff.prefer_direct
7275 Normally Squid tries to use parents for most requests. If you for some
7276 reason like it to first try going direct and only use a parent if
7277 going direct fails set this to on.
7279 By combining nonhierarchical_direct off and prefer_direct on you
7280 can set up Squid to use a parent as a backup path if going direct
7283 Note: If you want Squid to use parents for all requests see
7284 the never_direct directive. prefer_direct only modifies how Squid
7285 acts on cacheable requests.
7290 LOC: Config.accessList.AlwaysDirect
7292 DEFAULT_DOC: Prevent any cache_peer being used for this request.
7294 Usage: always_direct allow|deny [!]aclname ...
7296 Here you can use ACL elements to specify requests which should
7297 ALWAYS be forwarded by Squid to the origin servers without using
7298 any peers. For example, to always directly forward requests for
7299 local servers ignoring any parents or siblings you may have use
7302 acl local-servers dstdomain my.domain.net
7303 always_direct allow local-servers
7305 To always forward FTP requests directly, use
7308 always_direct allow FTP
7310 NOTE: There is a similar, but opposite option named
7311 'never_direct'. You need to be aware that "always_direct deny
7312 foo" is NOT the same thing as "never_direct allow foo". You
7313 may need to use a deny rule to exclude a more-specific case of
7314 some other rule. Example:
7316 acl local-external dstdomain external.foo.net
7317 acl local-servers dstdomain .foo.net
7318 always_direct deny local-external
7319 always_direct allow local-servers
7321 NOTE: If your goal is to make the client forward the request
7322 directly to the origin server bypassing Squid then this needs
7323 to be done in the client configuration. Squid configuration
7324 can only tell Squid how Squid should fetch the object.
7326 NOTE: This directive is not related to caching. The replies
7327 is cached as usual even if you use always_direct. To not cache
7328 the replies see the 'cache' directive.
7330 This clause supports both fast and slow acl types.
7331 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7336 LOC: Config.accessList.NeverDirect
7338 DEFAULT_DOC: Allow DNS results to be used for this request.
7340 Usage: never_direct allow|deny [!]aclname ...
7342 never_direct is the opposite of always_direct. Please read
7343 the description for always_direct if you have not already.
7345 With 'never_direct' you can use ACL elements to specify
7346 requests which should NEVER be forwarded directly to origin
7347 servers. For example, to force the use of a proxy for all
7348 requests, except those in your local domain use something like:
7350 acl local-servers dstdomain .foo.net
7351 never_direct deny local-servers
7352 never_direct allow all
7354 or if Squid is inside a firewall and there are local intranet
7355 servers inside the firewall use something like:
7357 acl local-intranet dstdomain .foo.net
7358 acl local-external dstdomain external.foo.net
7359 always_direct deny local-external
7360 always_direct allow local-intranet
7361 never_direct allow all
7363 This clause supports both fast and slow acl types.
7364 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7368 ADVANCED NETWORKING OPTIONS
7369 -----------------------------------------------------------------------------
7372 NAME: incoming_udp_average incoming_icp_average
7375 LOC: Config.comm_incoming.udp.average
7377 Heavy voodoo here. I can't even believe you are reading this.
7378 Are you crazy? Don't even think about adjusting these unless
7379 you understand the algorithms in comm_select.c first!
7382 NAME: incoming_tcp_average incoming_http_average
7385 LOC: Config.comm_incoming.tcp.average
7387 Heavy voodoo here. I can't even believe you are reading this.
7388 Are you crazy? Don't even think about adjusting these unless
7389 you understand the algorithms in comm_select.c first!
7392 NAME: incoming_dns_average
7395 LOC: Config.comm_incoming.dns.average
7397 Heavy voodoo here. I can't even believe you are reading this.
7398 Are you crazy? Don't even think about adjusting these unless
7399 you understand the algorithms in comm_select.c first!
7402 NAME: min_udp_poll_cnt min_icp_poll_cnt
7405 LOC: Config.comm_incoming.udp.min_poll
7407 Heavy voodoo here. I can't even believe you are reading this.
7408 Are you crazy? Don't even think about adjusting these unless
7409 you understand the algorithms in comm_select.c first!
7412 NAME: min_dns_poll_cnt
7415 LOC: Config.comm_incoming.dns.min_poll
7417 Heavy voodoo here. I can't even believe you are reading this.
7418 Are you crazy? Don't even think about adjusting these unless
7419 you understand the algorithms in comm_select.c first!
7422 NAME: min_tcp_poll_cnt min_http_poll_cnt
7425 LOC: Config.comm_incoming.tcp.min_poll
7427 Heavy voodoo here. I can't even believe you are reading this.
7428 Are you crazy? Don't even think about adjusting these unless
7429 you understand the algorithms in comm_select.c first!
7435 LOC: Config.accept_filter
7439 The name of an accept(2) filter to install on Squid's
7440 listen socket(s). This feature is perhaps specific to
7441 FreeBSD and requires support in the kernel.
7443 The 'httpready' filter delays delivering new connections
7444 to Squid until a full HTTP request has been received.
7445 See the accf_http(9) man page for details.
7447 The 'dataready' filter delays delivering new connections
7448 to Squid until there is some data to process.
7449 See the accf_dataready(9) man page for details.
7453 The 'data' filter delays delivering of new connections
7454 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
7455 You may optionally specify a number of seconds to wait by
7456 'data=N' where N is the number of seconds. Defaults to 30
7457 if not specified. See the tcp(7) man page for details.
7460 accept_filter httpready
7465 NAME: client_ip_max_connections
7467 LOC: Config.client_ip_max_connections
7469 DEFAULT_DOC: No limit.
7471 Set an absolute limit on the number of connections a single
7472 client IP can use. Any more than this and Squid will begin to drop
7473 new connections from the client until it closes some links.
7475 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
7476 connections from the client. For finer control use the ACL access controls.
7478 Requires client_db to be enabled (the default).
7480 WARNING: This may noticably slow down traffic received via external proxies
7481 or NAT devices and cause them to rebound error messages back to their clients.
7484 NAME: tcp_recv_bufsize
7488 DEFAULT_DOC: Use operating system TCP defaults.
7489 LOC: Config.tcpRcvBufsz
7491 Size of receive buffer to set for TCP sockets. Probably just
7492 as easy to change your kernel's default.
7493 Omit from squid.conf to use the default buffer size.
7498 -----------------------------------------------------------------------------
7505 LOC: Adaptation::Icap::TheConfig.onoff
7508 If you want to enable the ICAP module support, set this to on.
7511 NAME: icap_connect_timeout
7514 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7517 This parameter specifies how long to wait for the TCP connect to
7518 the requested ICAP server to complete before giving up and either
7519 terminating the HTTP transaction or bypassing the failure.
7521 The default for optional services is peer_connect_timeout.
7522 The default for essential services is connect_timeout.
7523 If this option is explicitly set, its value applies to all services.
7526 NAME: icap_io_timeout
7530 DEFAULT_DOC: Use read_timeout.
7531 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7534 This parameter specifies how long to wait for an I/O activity on
7535 an established, active ICAP connection before giving up and
7536 either terminating the HTTP transaction or bypassing the
7540 NAME: icap_service_failure_limit
7541 COMMENT: limit [in memory-depth time-units]
7542 TYPE: icap_service_failure_limit
7544 LOC: Adaptation::Icap::TheConfig
7547 The limit specifies the number of failures that Squid tolerates
7548 when establishing a new TCP connection with an ICAP service. If
7549 the number of failures exceeds the limit, the ICAP service is
7550 not used for new ICAP requests until it is time to refresh its
7553 A negative value disables the limit. Without the limit, an ICAP
7554 service will not be considered down due to connectivity failures
7555 between ICAP OPTIONS requests.
7557 Squid forgets ICAP service failures older than the specified
7558 value of memory-depth. The memory fading algorithm
7559 is approximate because Squid does not remember individual
7560 errors but groups them instead, splitting the option
7561 value into ten time slots of equal length.
7563 When memory-depth is 0 and by default this option has no
7564 effect on service failure expiration.
7566 Squid always forgets failures when updating service settings
7567 using an ICAP OPTIONS transaction, regardless of this option
7571 # suspend service usage after 10 failures in 5 seconds:
7572 icap_service_failure_limit 10 in 5 seconds
7575 NAME: icap_service_revival_delay
7578 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7581 The delay specifies the number of seconds to wait after an ICAP
7582 OPTIONS request failure before requesting the options again. The
7583 failed ICAP service is considered "down" until fresh OPTIONS are
7586 The actual delay cannot be smaller than the hardcoded minimum
7587 delay of 30 seconds.
7590 NAME: icap_preview_enable
7594 LOC: Adaptation::Icap::TheConfig.preview_enable
7597 The ICAP Preview feature allows the ICAP server to handle the
7598 HTTP message by looking only at the beginning of the message body
7599 or even without receiving the body at all. In some environments,
7600 previews greatly speedup ICAP processing.
7602 During an ICAP OPTIONS transaction, the server may tell Squid what
7603 HTTP messages should be previewed and how big the preview should be.
7604 Squid will not use Preview if the server did not request one.
7606 To disable ICAP Preview for all ICAP services, regardless of
7607 individual ICAP server OPTIONS responses, set this option to "off".
7609 icap_preview_enable off
7612 NAME: icap_preview_size
7615 LOC: Adaptation::Icap::TheConfig.preview_size
7617 DEFAULT_DOC: No preview sent.
7619 The default size of preview data to be sent to the ICAP server.
7620 This value might be overwritten on a per server basis by OPTIONS requests.
7623 NAME: icap_206_enable
7627 LOC: Adaptation::Icap::TheConfig.allow206_enable
7630 206 (Partial Content) responses is an ICAP extension that allows the
7631 ICAP agents to optionally combine adapted and original HTTP message
7632 content. The decision to combine is postponed until the end of the
7633 ICAP response. Squid supports Partial Content extension by default.
7635 Activation of the Partial Content extension is negotiated with each
7636 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7637 negotation correctly even if they do not support the extension, but
7638 some might fail. To disable Partial Content support for all ICAP
7639 services and to avoid any negotiation, set this option to "off".
7645 NAME: icap_default_options_ttl
7648 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7651 The default TTL value for ICAP OPTIONS responses that don't have
7652 an Options-TTL header.
7655 NAME: icap_persistent_connections
7659 LOC: Adaptation::Icap::TheConfig.reuse_connections
7662 Whether or not Squid should use persistent connections to
7666 NAME: adaptation_send_client_ip icap_send_client_ip
7668 IFDEF: USE_ADAPTATION
7670 LOC: Adaptation::Config::send_client_ip
7673 If enabled, Squid shares HTTP client IP information with adaptation
7674 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7675 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7677 See also: adaptation_uses_indirect_client
7680 NAME: adaptation_send_username icap_send_client_username
7682 IFDEF: USE_ADAPTATION
7684 LOC: Adaptation::Config::send_username
7687 This sends authenticated HTTP client username (if available) to
7688 the adaptation service.
7690 For ICAP, the username value is encoded based on the
7691 icap_client_username_encode option and is sent using the header
7692 specified by the icap_client_username_header option.
7695 NAME: icap_client_username_header
7698 LOC: Adaptation::Icap::TheConfig.client_username_header
7699 DEFAULT: X-Client-Username
7701 ICAP request header name to use for adaptation_send_username.
7704 NAME: icap_client_username_encode
7708 LOC: Adaptation::Icap::TheConfig.client_username_encode
7711 Whether to base64 encode the authenticated client username.
7715 TYPE: icap_service_type
7717 LOC: Adaptation::Icap::TheConfig
7720 Defines a single ICAP service using the following format:
7722 icap_service id vectoring_point uri [option ...]
7725 an opaque identifier or name which is used to direct traffic to
7726 this specific service. Must be unique among all adaptation
7727 services in squid.conf.
7729 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7730 This specifies at which point of transaction processing the
7731 ICAP service should be activated. *_postcache vectoring points
7732 are not yet supported.
7734 uri: icap://servername:port/servicepath
7735 ICAP server and service location.
7737 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7738 transactions. Squid does not enforce that requirement. You can specify
7739 services with the same service_url and different vectoring_points. You
7740 can even specify multiple identical services as long as their
7741 service_names differ.
7743 To activate a service, use the adaptation_access directive. To group
7744 services, use adaptation_service_chain and adaptation_service_set.
7746 Service options are separated by white space. ICAP services support
7747 the following name=value options:
7750 If set to 'on' or '1', the ICAP service is treated as
7751 optional. If the service cannot be reached or malfunctions,
7752 Squid will try to ignore any errors and process the message as
7753 if the service was not enabled. No all ICAP errors can be
7754 bypassed. If set to 0, the ICAP service is treated as
7755 essential and all ICAP errors will result in an error page
7756 returned to the HTTP client.
7758 Bypass is off by default: services are treated as essential.
7761 If set to 'on' or '1', the ICAP service is allowed to
7762 dynamically change the current message adaptation plan by
7763 returning a chain of services to be used next. The services
7764 are specified using the X-Next-Services ICAP response header
7765 value, formatted as a comma-separated list of service names.
7766 Each named service should be configured in squid.conf. Other
7767 services are ignored. An empty X-Next-Services value results
7768 in an empty plan which ends the current adaptation.
7770 Dynamic adaptation plan may cross or cover multiple supported
7771 vectoring points in their natural processing order.
7773 Routing is not allowed by default: the ICAP X-Next-Services
7774 response header is ignored.
7777 Only has effect on split-stack systems. The default on those systems
7778 is to use IPv4-only connections. When set to 'on' this option will
7779 make Squid use IPv6-only connections to contact this ICAP service.
7781 on-overload=block|bypass|wait|force
7782 If the service Max-Connections limit has been reached, do
7783 one of the following for each new ICAP transaction:
7784 * block: send an HTTP error response to the client
7785 * bypass: ignore the "over-connected" ICAP service
7786 * wait: wait (in a FIFO queue) for an ICAP connection slot
7787 * force: proceed, ignoring the Max-Connections limit
7789 In SMP mode with N workers, each worker assumes the service
7790 connection limit is Max-Connections/N, even though not all
7791 workers may use a given service.
7793 The default value is "bypass" if service is bypassable,
7794 otherwise it is set to "wait".
7798 Use the given number as the Max-Connections limit, regardless
7799 of the Max-Connections value given by the service, if any.
7801 Older icap_service format without optional named parameters is
7802 deprecated but supported for backward compatibility.
7805 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7806 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7810 TYPE: icap_class_type
7815 This deprecated option was documented to define an ICAP service
7816 chain, even though it actually defined a set of similar, redundant
7817 services, and the chains were not supported.
7819 To define a set of redundant services, please use the
7820 adaptation_service_set directive. For service chains, use
7821 adaptation_service_chain.
7825 TYPE: icap_access_type
7830 This option is deprecated. Please use adaptation_access, which
7831 has the same ICAP functionality, but comes with better
7832 documentation, and eCAP support.
7837 -----------------------------------------------------------------------------
7844 LOC: Adaptation::Ecap::TheConfig.onoff
7847 Controls whether eCAP support is enabled.
7851 TYPE: ecap_service_type
7853 LOC: Adaptation::Ecap::TheConfig
7856 Defines a single eCAP service
7858 ecap_service id vectoring_point uri [option ...]
7861 an opaque identifier or name which is used to direct traffic to
7862 this specific service. Must be unique among all adaptation
7863 services in squid.conf.
7865 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7866 This specifies at which point of transaction processing the
7867 eCAP service should be activated. *_postcache vectoring points
7868 are not yet supported.
7870 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7871 Squid uses the eCAP service URI to match this configuration
7872 line with one of the dynamically loaded services. Each loaded
7873 eCAP service must have a unique URI. Obtain the right URI from
7874 the service provider.
7876 To activate a service, use the adaptation_access directive. To group
7877 services, use adaptation_service_chain and adaptation_service_set.
7879 Service options are separated by white space. eCAP services support
7880 the following name=value options:
7883 If set to 'on' or '1', the eCAP service is treated as optional.
7884 If the service cannot be reached or malfunctions, Squid will try
7885 to ignore any errors and process the message as if the service
7886 was not enabled. No all eCAP errors can be bypassed.
7887 If set to 'off' or '0', the eCAP service is treated as essential
7888 and all eCAP errors will result in an error page returned to the
7891 Bypass is off by default: services are treated as essential.
7894 If set to 'on' or '1', the eCAP service is allowed to
7895 dynamically change the current message adaptation plan by
7896 returning a chain of services to be used next.
7898 Dynamic adaptation plan may cross or cover multiple supported
7899 vectoring points in their natural processing order.
7901 Routing is not allowed by default.
7903 Older ecap_service format without optional named parameters is
7904 deprecated but supported for backward compatibility.
7908 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7909 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7912 NAME: loadable_modules
7914 IFDEF: USE_LOADABLE_MODULES
7915 LOC: Config.loadable_module_names
7918 Instructs Squid to load the specified dynamic module(s) or activate
7919 preloaded module(s).
7921 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7925 MESSAGE ADAPTATION OPTIONS
7926 -----------------------------------------------------------------------------
7929 NAME: adaptation_service_set
7930 TYPE: adaptation_service_set_type
7931 IFDEF: USE_ADAPTATION
7936 Configures an ordered set of similar, redundant services. This is
7937 useful when hot standby or backup adaptation servers are available.
7939 adaptation_service_set set_name service_name1 service_name2 ...
7941 The named services are used in the set declaration order. The first
7942 applicable adaptation service from the set is used first. The next
7943 applicable service is tried if and only if the transaction with the
7944 previous service fails and the message waiting to be adapted is still
7947 When adaptation starts, broken services are ignored as if they were
7948 not a part of the set. A broken service is a down optional service.
7950 The services in a set must be attached to the same vectoring point
7951 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7953 If all services in a set are optional then adaptation failures are
7954 bypassable. If all services in the set are essential, then a
7955 transaction failure with one service may still be retried using
7956 another service from the set, but when all services fail, the master
7957 transaction fails as well.
7959 A set may contain a mix of optional and essential services, but that
7960 is likely to lead to surprising results because broken services become
7961 ignored (see above), making previously bypassable failures fatal.
7962 Technically, it is the bypassability of the last failed service that
7965 See also: adaptation_access adaptation_service_chain
7968 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7969 adaptation service_set svcLogger loggerLocal loggerRemote
7972 NAME: adaptation_service_chain
7973 TYPE: adaptation_service_chain_type
7974 IFDEF: USE_ADAPTATION
7979 Configures a list of complementary services that will be applied
7980 one-by-one, forming an adaptation chain or pipeline. This is useful
7981 when Squid must perform different adaptations on the same message.
7983 adaptation_service_chain chain_name service_name1 svc_name2 ...
7985 The named services are used in the chain declaration order. The first
7986 applicable adaptation service from the chain is used first. The next
7987 applicable service is applied to the successful adaptation results of
7988 the previous service in the chain.
7990 When adaptation starts, broken services are ignored as if they were
7991 not a part of the chain. A broken service is a down optional service.
7993 Request satisfaction terminates the adaptation chain because Squid
7994 does not currently allow declaration of RESPMOD services at the
7995 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7997 The services in a chain must be attached to the same vectoring point
7998 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
8000 A chain may contain a mix of optional and essential services. If an
8001 essential adaptation fails (or the failure cannot be bypassed for
8002 other reasons), the master transaction fails. Otherwise, the failure
8003 is bypassed as if the failed adaptation service was not in the chain.
8005 See also: adaptation_access adaptation_service_set
8008 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
8011 NAME: adaptation_access
8012 TYPE: adaptation_access_type
8013 IFDEF: USE_ADAPTATION
8016 DEFAULT_DOC: Allow, unless rules exist in squid.conf.
8018 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
8020 adaptation_access service_name allow|deny [!]aclname...
8021 adaptation_access set_name allow|deny [!]aclname...
8023 At each supported vectoring point, the adaptation_access
8024 statements are processed in the order they appear in this
8025 configuration file. Statements pointing to the following services
8026 are ignored (i.e., skipped without checking their ACL):
8028 - services serving different vectoring points
8029 - "broken-but-bypassable" services
8030 - "up" services configured to ignore such transactions
8031 (e.g., based on the ICAP Transfer-Ignore header).
8033 When a set_name is used, all services in the set are checked
8034 using the same rules, to find the first applicable one. See
8035 adaptation_service_set for details.
8037 If an access list is checked and there is a match, the
8038 processing stops: For an "allow" rule, the corresponding
8039 adaptation service is used for the transaction. For a "deny"
8040 rule, no adaptation service is activated.
8042 It is currently not possible to apply more than one adaptation
8043 service at the same vectoring point to the same HTTP transaction.
8045 See also: icap_service and ecap_service
8048 adaptation_access service_1 allow all
8051 NAME: adaptation_service_iteration_limit
8053 IFDEF: USE_ADAPTATION
8054 LOC: Adaptation::Config::service_iteration_limit
8057 Limits the number of iterations allowed when applying adaptation
8058 services to a message. If your longest adaptation set or chain
8059 may have more than 16 services, increase the limit beyond its
8060 default value of 16. If detecting infinite iteration loops sooner
8061 is critical, make the iteration limit match the actual number
8062 of services in your longest adaptation set or chain.
8064 Infinite adaptation loops are most likely with routing services.
8066 See also: icap_service routing=1
8069 NAME: adaptation_masterx_shared_names
8071 IFDEF: USE_ADAPTATION
8072 LOC: Adaptation::Config::masterx_shared_name
8075 For each master transaction (i.e., the HTTP request and response
8076 sequence, including all related ICAP and eCAP exchanges), Squid
8077 maintains a table of metadata. The table entries are (name, value)
8078 pairs shared among eCAP and ICAP exchanges. The table is destroyed
8079 with the master transaction.
8081 This option specifies the table entry names that Squid must accept
8082 from and forward to the adaptation transactions.
8084 An ICAP REQMOD or RESPMOD transaction may set an entry in the
8085 shared table by returning an ICAP header field with a name
8086 specified in adaptation_masterx_shared_names.
8088 An eCAP REQMOD or RESPMOD transaction may set an entry in the
8089 shared table by implementing the libecap::visitEachOption() API
8090 to provide an option with a name specified in
8091 adaptation_masterx_shared_names.
8093 Squid will store and forward the set entry to subsequent adaptation
8094 transactions within the same master transaction scope.
8096 Only one shared entry name is supported at this time.
8099 # share authentication information among ICAP services
8100 adaptation_masterx_shared_names X-Subscriber-ID
8103 NAME: adaptation_meta
8105 IFDEF: USE_ADAPTATION
8106 LOC: Adaptation::Config::metaHeaders
8109 This option allows Squid administrator to add custom ICAP request
8110 headers or eCAP options to Squid ICAP requests or eCAP transactions.
8111 Use it to pass custom authentication tokens and other
8112 transaction-state related meta information to an ICAP/eCAP service.
8114 The addition of a meta header is ACL-driven:
8115 adaptation_meta name value [!]aclname ...
8117 Processing for a given header name stops after the first ACL list match.
8118 Thus, it is impossible to add two headers with the same name. If no ACL
8119 lists match for a given header name, no such header is added. For
8122 # do not debug transactions except for those that need debugging
8123 adaptation_meta X-Debug 1 needs_debugging
8125 # log all transactions except for those that must remain secret
8126 adaptation_meta X-Log 1 !keep_secret
8128 # mark transactions from users in the "G 1" group
8129 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
8131 The "value" parameter may be a regular squid.conf token or a "double
8132 quoted string". Within the quoted string, use backslash (\) to escape
8133 any character, which is currently only useful for escaping backslashes
8134 and double quotes. For example,
8135 "this string has one backslash (\\) and two \"quotes\""
8137 Used adaptation_meta header values may be logged via %note
8138 logformat code. If multiple adaptation_meta headers with the same name
8139 are used during master transaction lifetime, the header values are
8140 logged in the order they were used and duplicate values are ignored
8141 (only the first repeated value will be logged).
8147 LOC: Adaptation::Icap::TheConfig.repeat
8148 DEFAULT_IF_NONE: deny all
8150 This ACL determines which retriable ICAP transactions are
8151 retried. Transactions that received a complete ICAP response
8152 and did not have to consume or produce HTTP bodies to receive
8153 that response are usually retriable.
8155 icap_retry allow|deny [!]aclname ...
8157 Squid automatically retries some ICAP I/O timeouts and errors
8158 due to persistent connection race conditions.
8160 See also: icap_retry_limit
8163 NAME: icap_retry_limit
8166 LOC: Adaptation::Icap::TheConfig.repeat_limit
8168 DEFAULT_DOC: No retries are allowed.
8170 Limits the number of retries allowed.
8172 Communication errors due to persistent connection race
8173 conditions are unavoidable, automatically retried, and do not
8174 count against this limit.
8176 See also: icap_retry
8182 -----------------------------------------------------------------------------
8185 NAME: check_hostnames
8188 LOC: Config.onoff.check_hostnames
8190 For security and stability reasons Squid can check
8191 hostnames for Internet standard RFC compliance. If you want
8192 Squid to perform these checks turn this directive on.
8195 NAME: allow_underscore
8198 LOC: Config.onoff.allow_underscore
8200 Underscore characters is not strictly allowed in Internet hostnames
8201 but nevertheless used by many sites. Set this to off if you want
8202 Squid to be strict about the standard.
8203 This check is performed only when check_hostnames is set to on.
8206 NAME: cache_dns_program
8208 IFDEF: USE_DNSHELPER
8209 DEFAULT: @DEFAULT_DNSSERVER@
8210 LOC: Config.Program.dnsserver
8212 Specify the location of the executable for dnslookup process.
8216 TYPE: HelperChildConfig
8217 IFDEF: USE_DNSHELPER
8218 DEFAULT: 32 startup=1 idle=1
8219 LOC: Config.dnsChildren
8221 The maximum number of processes spawn to service DNS name lookups.
8222 If you limit it too few Squid will have to wait for them to process
8223 a backlog of requests, slowing it down. If you allow too many they
8224 will use RAM and other system resources noticably.
8225 The maximum this may be safely set to is 32.
8227 The startup= and idle= options allow some measure of skew in your
8232 Sets a minimum of how many processes are to be spawned when Squid
8233 starts or reconfigures. When set to zero the first request will
8234 cause spawning of the first child process to handle it.
8236 Starting too few will cause an initial slowdown in traffic as Squid
8237 attempts to simultaneously spawn enough processes to cope.
8241 Sets a minimum of how many processes Squid is to try and keep available
8242 at all times. When traffic begins to rise above what the existing
8243 processes can handle this many more will be spawned up to the maximum
8244 configured. A minimum setting of 1 is required.
8247 NAME: dns_retransmit_interval
8250 LOC: Config.Timeout.idns_retransmit
8251 IFDEF: !USE_DNSHELPER
8253 Initial retransmit interval for DNS queries. The interval is
8254 doubled each time all configured DNS servers have been tried.
8260 LOC: Config.Timeout.idns_query
8261 IFDEF: !USE_DNSHELPER
8263 DNS Query timeout. If no response is received to a DNS query
8264 within this time all DNS servers for the queried domain
8265 are assumed to be unavailable.
8268 NAME: dns_packet_max
8270 DEFAULT_DOC: EDNS disabled
8272 LOC: Config.dns.packet_max
8273 IFDEF: !USE_DNSHELPER
8275 Maximum number of bytes packet size to advertise via EDNS.
8276 Set to "none" to disable EDNS large packet support.
8278 For legacy reasons DNS UDP replies will default to 512 bytes which
8279 is too small for many responses. EDNS provides a means for Squid to
8280 negotiate receiving larger responses back immediately without having
8281 to failover with repeat requests. Responses larger than this limit
8282 will retain the old behaviour of failover to TCP DNS.
8284 Squid has no real fixed limit internally, but allowing packet sizes
8285 over 1500 bytes requires network jumbogram support and is usually not
8288 WARNING: The RFC also indicates that some older resolvers will reply
8289 with failure of the whole request if the extension is added. Some
8290 resolvers have already been identified which will reply with mangled
8291 EDNS response on occasion. Usually in response to many-KB jumbogram
8292 sizes being advertised by Squid.
8293 Squid will currently treat these both as an unable-to-resolve domain
8294 even if it would be resolvable without EDNS.
8301 DEFAULT_DOC: Search for single-label domain names is disabled.
8302 LOC: Config.onoff.res_defnames
8304 Normally the RES_DEFNAMES resolver option is disabled
8305 (see res_init(3)). This prevents caches in a hierarchy
8306 from interpreting single-component hostnames locally. To allow
8307 Squid to handle single-component names, enable this option.
8310 NAME: dns_nameservers
8313 DEFAULT_DOC: Use operating system definitions
8314 LOC: Config.dns_nameservers
8316 Use this if you want to specify a list of DNS name servers
8317 (IP addresses) to use instead of those given in your
8318 /etc/resolv.conf file.
8320 On Windows platforms, if no value is specified here or in
8321 the /etc/resolv.conf file, the list of DNS name servers are
8322 taken from the Windows registry, both static and dynamic DHCP
8323 configurations are supported.
8325 Example: dns_nameservers 10.0.0.1 192.172.0.4
8330 DEFAULT: @DEFAULT_HOSTS@
8331 LOC: Config.etcHostsPath
8333 Location of the host-local IP name-address associations
8334 database. Most Operating Systems have such a file on different
8336 - Un*X & Linux: /etc/hosts
8337 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
8338 (%SystemRoot% value install default is c:\winnt)
8339 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
8340 (%SystemRoot% value install default is c:\windows)
8341 - Windows 9x/Me: %windir%\hosts
8342 (%windir% value is usually c:\windows)
8343 - Cygwin: /etc/hosts
8345 The file contains newline-separated definitions, in the
8346 form ip_address_in_dotted_form name [name ...] names are
8347 whitespace-separated. Lines beginning with an hash (#)
8348 character are comments.
8350 The file is checked at startup and upon configuration.
8351 If set to 'none', it won't be checked.
8352 If append_domain is used, that domain will be added to
8353 domain-local (i.e. not containing any dot character) host
8359 LOC: Config.appendDomain
8361 DEFAULT_DOC: Use operating system definitions
8363 Appends local domain name to hostnames without any dots in
8364 them. append_domain must begin with a period.
8366 Be warned there are now Internet names with no dots in
8367 them using only top-domain names, so setting this may
8368 cause some Internet sites to become unavailable.
8371 append_domain .yourdomain.com
8374 NAME: ignore_unknown_nameservers
8376 LOC: Config.onoff.ignore_unknown_nameservers
8378 IFDEF: !USE_DNSHELPER
8380 By default Squid checks that DNS responses are received
8381 from the same IP addresses they are sent to. If they
8382 don't match, Squid ignores the response and writes a warning
8383 message to cache.log. You can allow responses from unknown
8384 nameservers by setting this option to 'off'.
8390 LOC: Config.dns.v4_first
8391 IFDEF: !USE_DNSHELPER
8393 With the IPv6 Internet being as fast or faster than IPv4 Internet
8394 for most networks Squid prefers to contact websites over IPv6.
8396 This option reverses the order of preference to make Squid contact
8397 dual-stack websites over IPv4 first. Squid will still perform both
8398 IPv6 and IPv4 DNS lookups before connecting.
8401 This option will restrict the situations under which IPv6
8402 connectivity is used (and tested). Hiding network problems
8403 which would otherwise be detected and warned about.
8407 COMMENT: (number of entries)
8410 LOC: Config.ipcache.size
8412 Maximum number of DNS IP cache entries.
8419 LOC: Config.ipcache.low
8426 LOC: Config.ipcache.high
8428 The size, low-, and high-water marks for the IP cache.
8431 NAME: fqdncache_size
8432 COMMENT: (number of entries)
8435 LOC: Config.fqdncache.size
8437 Maximum number of FQDN cache entries.
8442 -----------------------------------------------------------------------------
8449 LOC: Config.onoff.mem_pools
8451 If set, Squid will keep pools of allocated (but unused) memory
8452 available for future use. If memory is a premium on your
8453 system and you believe your malloc library outperforms Squid
8454 routines, disable this.
8457 NAME: memory_pools_limit
8461 LOC: Config.MemPools.limit
8463 Used only with memory_pools on:
8464 memory_pools_limit 50 MB
8466 If set to a non-zero value, Squid will keep at most the specified
8467 limit of allocated (but unused) memory in memory pools. All free()
8468 requests that exceed this limit will be handled by your malloc
8469 library. Squid does not pre-allocate any memory, just safe-keeps
8470 objects that otherwise would be free()d. Thus, it is safe to set
8471 memory_pools_limit to a reasonably high value even if your
8472 configuration will use less memory.
8474 If set to none, Squid will keep all memory it can. That is, there
8475 will be no limit on the total amount of memory used for safe-keeping.
8477 To disable memory allocation optimization, do not set
8478 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
8480 An overhead for maintaining memory pools is not taken into account
8481 when the limit is checked. This overhead is close to four bytes per
8482 object kept. However, pools may actually _save_ memory because of
8483 reduced memory thrashing in your malloc library.
8487 COMMENT: on|off|transparent|truncate|delete
8490 LOC: opt_forwarded_for
8492 If set to "on", Squid will append your client's IP address
8493 in the HTTP requests it forwards. By default it looks like:
8495 X-Forwarded-For: 192.1.2.3
8497 If set to "off", it will appear as
8499 X-Forwarded-For: unknown
8501 If set to "transparent", Squid will not alter the
8502 X-Forwarded-For header in any way.
8504 If set to "delete", Squid will delete the entire
8505 X-Forwarded-For header.
8507 If set to "truncate", Squid will remove all existing
8508 X-Forwarded-For entries, and place the client IP as the sole entry.
8511 NAME: cachemgr_passwd
8512 TYPE: cachemgrpasswd
8514 DEFAULT_DOC: No password. Actions which require password are denied.
8515 LOC: Config.passwd_list
8517 Specify passwords for cachemgr operations.
8519 Usage: cachemgr_passwd password action action ...
8521 Some valid actions are (see cache manager menu for a full list):
8561 * Indicates actions which will not be performed without a
8562 valid password, others can be performed if not listed here.
8564 To disable an action, set the password to "disable".
8565 To allow performing an action without a password, set the
8568 Use the keyword "all" to set the same password for all actions.
8571 cachemgr_passwd secret shutdown
8572 cachemgr_passwd lesssssssecret info stats/objects
8573 cachemgr_passwd disable all
8580 LOC: Config.onoff.client_db
8582 If you want to disable collecting per-client statistics,
8583 turn off client_db here.
8586 NAME: refresh_all_ims
8590 LOC: Config.onoff.refresh_all_ims
8592 When you enable this option, squid will always check
8593 the origin server for an update when a client sends an
8594 If-Modified-Since request. Many browsers use IMS
8595 requests when the user requests a reload, and this
8596 ensures those clients receive the latest version.
8598 By default (off), squid may return a Not Modified response
8599 based on the age of the cached version.
8602 NAME: reload_into_ims
8603 IFDEF: USE_HTTP_VIOLATIONS
8607 LOC: Config.onoff.reload_into_ims
8609 When you enable this option, client no-cache or ``reload''
8610 requests will be changed to If-Modified-Since requests.
8611 Doing this VIOLATES the HTTP standard. Enabling this
8612 feature could make you liable for problems which it
8615 see also refresh_pattern for a more selective approach.
8618 NAME: connect_retries
8620 LOC: Config.connect_retries
8622 DEFAULT_DOC: Do not retry failed connections.
8624 This sets the maximum number of connection attempts made for each
8625 TCP connection. The connect_retries attempts must all still
8626 complete within the connection timeout period.
8628 The default is not to re-try if the first connection attempt fails.
8629 The (not recommended) maximum is 10 tries.
8631 A warning message will be generated if it is set to a too-high
8632 value and the configured value will be over-ridden.
8634 Note: These re-tries are in addition to forward_max_tries
8635 which limit how many different addresses may be tried to find
8639 NAME: retry_on_error
8641 LOC: Config.retry.onerror
8644 If set to ON Squid will automatically retry requests when
8645 receiving an error response with status 403 (Forbidden),
8646 500 (Internal Error), 501 or 503 (Service not available).
8647 Status 502 and 504 (Gateway errors) are always retried.
8649 This is mainly useful if you are in a complex cache hierarchy to
8650 work around access control errors.
8652 NOTE: This retry will attempt to find another working destination.
8653 Which is different from the server which just failed.
8656 NAME: as_whois_server
8658 LOC: Config.as_whois_server
8659 DEFAULT: whois.ra.net
8661 WHOIS server to query for AS numbers. NOTE: AS numbers are
8662 queried only when Squid starts up, not for every request.
8667 LOC: Config.onoff.offline
8670 Enable this option and Squid will never try to validate cached
8674 NAME: uri_whitespace
8675 TYPE: uri_whitespace
8676 LOC: Config.uri_whitespace
8679 What to do with requests that have whitespace characters in the
8682 strip: The whitespace characters are stripped out of the URL.
8683 This is the behavior recommended by RFC2396 and RFC3986
8684 for tolerant handling of generic URI.
8685 NOTE: This is one difference between generic URI and HTTP URLs.
8687 deny: The request is denied. The user receives an "Invalid
8689 This is the behaviour recommended by RFC2616 for safe
8690 handling of HTTP request URL.
8692 allow: The request is allowed and the URI is not changed. The
8693 whitespace characters remain in the URI. Note the
8694 whitespace is passed to redirector processes if they
8696 Note this may be considered a violation of RFC2616
8697 request parsing where whitespace is prohibited in the
8700 encode: The request is allowed and the whitespace characters are
8701 encoded according to RFC1738.
8703 chop: The request is allowed and the URI is chopped at the
8707 NOTE the current Squid implementation of encode and chop violates
8708 RFC2616 by not using a 301 redirect after altering the URL.
8713 LOC: Config.chroot_dir
8716 Specifies a directory where Squid should do a chroot() while
8717 initializing. This also causes Squid to fully drop root
8718 privileges after initializing. This means, for example, if you
8719 use a HTTP port less than 1024 and try to reconfigure, you may
8720 get an error saying that Squid can not open the port.
8723 NAME: balance_on_multiple_ip
8725 LOC: Config.onoff.balance_on_multiple_ip
8728 Modern IP resolvers in squid sort lookup results by preferred access.
8729 By default squid will use these IP in order and only rotates to
8730 the next listed when the most preffered fails.
8732 Some load balancing servers based on round robin DNS have been
8733 found not to preserve user session state across requests
8734 to different IP addresses.
8736 Enabling this directive Squid rotates IP's per request.
8739 NAME: pipeline_prefetch
8740 TYPE: pipelinePrefetch
8741 LOC: Config.pipeline_max_prefetch
8743 DEFAULT_DOC: Do not pre-parse pipelined requests.
8745 HTTP clients may send a pipeline of 1+N requests to Squid using a
8746 single connection, without waiting for Squid to respond to the first
8747 of those requests. This option limits the number of concurrent
8748 requests Squid will try to handle in parallel. If set to N, Squid
8749 will try to receive and process up to 1+N requests on the same
8750 connection concurrently.
8752 Defaults to 0 (off) for bandwidth management and access logging
8755 NOTE: pipelining requires persistent connections to clients.
8757 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8760 NAME: high_response_time_warning
8763 LOC: Config.warnings.high_rptm
8765 DEFAULT_DOC: disabled.
8767 If the one-minute median response time exceeds this value,
8768 Squid prints a WARNING with debug level 0 to get the
8769 administrators attention. The value is in milliseconds.
8772 NAME: high_page_fault_warning
8774 LOC: Config.warnings.high_pf
8776 DEFAULT_DOC: disabled.
8778 If the one-minute average page fault rate exceeds this
8779 value, Squid prints a WARNING with debug level 0 to get
8780 the administrators attention. The value is in page faults
8784 NAME: high_memory_warning
8786 LOC: Config.warnings.high_memory
8788 DEFAULT_DOC: disabled.
8790 If the memory usage (as determined by mallinfo) exceeds
8791 this amount, Squid prints a WARNING with debug level 0 to get
8792 the administrators attention.
8795 NAME: sleep_after_fork
8796 COMMENT: (microseconds)
8798 LOC: Config.sleep_after_fork
8801 When this is set to a non-zero value, the main Squid process
8802 sleeps the specified number of microseconds after a fork()
8803 system call. This sleep may help the situation where your
8804 system reports fork() failures due to lack of (virtual)
8805 memory. Note, however, if you have a lot of child
8806 processes, these sleep delays will add up and your
8807 Squid will not service requests for some amount of time
8808 until all the child processes have been started.
8809 On Windows value less then 1000 (1 milliseconds) are
8813 NAME: windows_ipaddrchangemonitor
8814 IFDEF: _SQUID_WINDOWS_
8818 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8820 On Windows Squid by default will monitor IP address changes and will
8821 reconfigure itself after any detected event. This is very useful for
8822 proxies connected to internet with dial-up interfaces.
8823 In some cases (a Proxy server acting as VPN gateway is one) it could be
8824 desiderable to disable this behaviour setting this to 'off'.
8825 Note: after changing this, Squid service must be restarted.
8830 IFDEF: USE_SQUID_EUI
8832 LOC: Eui::TheConfig.euiLookup
8834 Whether to lookup the EUI or MAC address of a connected client.
8837 NAME: max_filedescriptors max_filedesc
8840 DEFAULT_DOC: Use operating system limits set by ulimit.
8841 LOC: Config.max_filedescriptors
8843 Reduce the maximum number of filedescriptors supported below
8844 the usual operating system defaults.
8846 Remove from squid.conf to inherit the current ulimit setting.
8848 Note: Changing this requires a restart of Squid. Also
8849 not all I/O types supports large values (eg on Windows).
8856 DEFAULT_DOC: SMP support disabled.
8858 Number of main Squid processes or "workers" to fork and maintain.
8859 0: "no daemon" mode, like running "squid -N ..."
8860 1: "no SMP" mode, start one main Squid process daemon (default)
8861 N: start N main Squid process daemons (i.e., SMP mode)
8863 In SMP mode, each worker does nearly all what a single Squid daemon
8864 does (e.g., listen on http_port and forward HTTP requests).
8867 NAME: cpu_affinity_map
8868 TYPE: CpuAffinityMap
8869 LOC: Config.cpuAffinityMap
8871 DEFAULT_DOC: Let operating system decide.
8873 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8875 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8877 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8879 affects processes 1 through 4 only and places them on the first
8880 four even cores, starting with core #1.
8882 CPU cores are numbered starting from 1. Requires support for
8883 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8885 Multiple cpu_affinity_map options are merged.