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 # no Options Removed in 3.3
109 # Options Removed in 3.2
110 NAME: ignore_expect_100
113 Remove this line. The HTTP/1.1 feature is now fully supported by default.
116 NAME: dns_v4_fallback
119 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
125 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
128 NAME: maximum_single_addr_tries
131 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
137 Remove this line. The feature is supported by default in storage types where update is implemented.
140 NAME: url_rewrite_concurrency
143 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
146 # Options Removed in 3.1
150 Remove this line. DNS is no longer tested on startup.
153 NAME: extension_methods
156 Remove this line. All valid methods for HTTP are accepted by default.
159 # 2.7 Options Removed/Replaced in 3.2
164 # 2.7 Options Removed/Replaced in 3.1
172 Remove this line. HTTP/1.1 is supported by default.
175 NAME: upgrade_http0.9
178 Remove this line. ICY/1.0 streaming protocol is supported by default.
181 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
184 Alter these entries. Use the qos_flows directive instead.
187 # Options Removed in 3.0
191 Since squid-3.0 replace with request_header_access or reply_header_access
192 depending on whether you wish to match client requests or server replies.
195 NAME: httpd_accel_no_pmtu_disc
198 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
201 NAME: wais_relay_host
204 Replace this line with 'cache_peer' configuration.
207 NAME: wais_relay_port
210 Replace this line with 'cache_peer' configuration.
214 OPTIONS FOR AUTHENTICATION
215 -----------------------------------------------------------------------------
224 This is used to define parameters for the various authentication
225 schemes supported by Squid.
227 format: auth_param scheme parameter [setting]
229 The order in which authentication schemes are presented to the client is
230 dependent on the order the scheme first appears in config file. IE
231 has a bug (it's not RFC 2617 compliant) in that it will use the basic
232 scheme if basic is the first entry presented, even if more secure
233 schemes are presented. For now use the order in the recommended
234 settings section below. If other browsers have difficulties (don't
235 recognize the schemes offered even if you are using basic) either
236 put basic first, or disable the other schemes (by commenting out their
239 Once an authentication scheme is fully configured, it can only be
240 shutdown by shutting squid down and restarting. Changes can be made on
241 the fly and activated with a reconfigure. I.E. You can change to a
242 different helper, but not unconfigure the helper completely.
244 Please note that while this directive defines how Squid processes
245 authentication it does not automatically activate authentication.
246 To use authentication you must in addition make use of ACLs based
247 on login name in http_access (proxy_auth, proxy_auth_regex or
248 external with %LOGIN used in the format tag). The browser will be
249 challenged for authentication on the first such acl encountered
250 in http_access processing and will also be re-challenged for new
251 login credentials if the request is being denied by a proxy_auth
254 WARNING: authentication can't be used in a transparently intercepting
255 proxy as the client then thinks it is talking to an origin server and
256 not the proxy. This is a limitation of bending the TCP/IP protocol to
257 transparently intercepting port 80, not a limitation in Squid.
258 Ports flagged 'transparent', 'intercept', or 'tproxy' have
259 authentication disabled.
261 === Parameters for the basic scheme follow. ===
264 Specify the command for the external authenticator. Such a program
265 reads a line containing "username password" and replies with one of
272 the user does not exist.
275 An internal error occurred in the helper, preventing
276 a result being identified.
278 "ERR" and "BH" results may optionally be followed by message="..."
279 containing a description available as %m in the returned error page.
281 If you use an authenticator, make sure you have 1 acl of type
284 By default, the basic authentication scheme is not used unless a
285 program is specified.
287 If you want to use the traditional NCSA proxy authentication, set
288 this line to something like
290 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
293 HTTP uses iso-latin-1 as character set, while some authentication
294 backends such as LDAP expects UTF-8. If this is set to on Squid will
295 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
296 username & password to the helper.
298 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
299 The maximum number of authenticator processes to spawn. If you start too few
300 Squid will have to wait for them to process a backlog of credential
301 verifications, slowing it down. When password verifications are
302 done via a (slow) network you are likely to need lots of
303 authenticator processes.
305 The startup= and idle= options permit some skew in the exact amount
306 run. A minimum of startup=N will begin during startup and reconfigure.
307 Squid will start more in groups of up to idle=N in an attempt to meet
308 traffic needs and to keep idle=N free above those traffic needs up to
311 The concurrency= option sets the number of concurrent requests the
312 helper can process. The default of 0 is used for helpers who only
313 supports one request at a time. Setting this to a number greater than
314 0 changes the protocol used to include a channel number first on the
315 request/response line, allowing multiple requests to be sent to the
316 same helper in parallel without waiting for the response.
317 Must not be set unless it's known the helper supports this.
319 auth_param basic children 20 startup=0 idle=1
322 Specifies the realm name which is to be reported to the
323 client for the basic proxy authentication scheme (part of
324 the text the user will see when prompted their username and
325 password). There is no default.
326 auth_param basic realm Squid proxy-caching web server
328 "credentialsttl" timetolive
329 Specifies how long squid assumes an externally validated
330 username:password pair is valid for - in other words how
331 often the helper program is called for that user. Set this
332 low to force revalidation with short lived passwords. Note
333 setting this high does not impact your susceptibility
334 to replay attacks unless you are using an one-time password
335 system (such as SecureID). If you are using such a system,
336 you will be vulnerable to replay attacks unless you also
337 use the max_user_ip ACL in an http_access rule.
339 "casesensitive" on|off
340 Specifies if usernames are case sensitive. Most user databases are
341 case insensitive allowing the same username to be spelled using both
342 lower and upper case letters, but some are case sensitive. This
343 makes a big difference for user_max_ip ACL processing and similar.
344 auth_param basic casesensitive off
346 === Parameters for the digest scheme follow ===
349 Specify the command for the external authenticator. Such
350 a program reads a line containing "username":"realm" and
351 replies with one of three results:
354 the user exists. The ha1= key is mandatory and
355 contains the appropriate H(A1) value, hex encoded.
356 See rfc 2616 for the definition of H(A1).
359 the user does not exist.
362 An internal error occurred in the helper, preventing
363 a result being identified.
365 "ERR" and "BH" results may optionally be followed by message="..."
366 containing a description available as %m in the returned error page.
368 By default, the digest authentication scheme is not used unless a
369 program is specified.
371 If you want to use a digest authenticator, set this line to
374 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
377 HTTP uses iso-latin-1 as character set, while some authentication
378 backends such as LDAP expects UTF-8. If this is set to on Squid will
379 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
380 username & password to the helper.
382 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
383 The maximum number of authenticator processes to spawn (default 5).
384 If you start too few Squid will have to wait for them to
385 process a backlog of H(A1) calculations, slowing it down.
386 When the H(A1) calculations are done via a (slow) network
387 you are likely to need lots of authenticator processes.
389 The startup= and idle= options permit some skew in the exact amount
390 run. A minimum of startup=N will begin during startup and reconfigure.
391 Squid will start more in groups of up to idle=N in an attempt to meet
392 traffic needs and to keep idle=N free above those traffic needs up to
395 The concurrency= option sets the number of concurrent requests the
396 helper can process. The default of 0 is used for helpers who only
397 supports one request at a time. Setting this to a number greater than
398 0 changes the protocol used to include a channel number first on the
399 request/response line, allowing multiple requests to be sent to the
400 same helper in parallel without waiting for the response.
401 Must not be set unless it's known the helper supports this.
403 auth_param digest children 20 startup=0 idle=1
406 Specifies the realm name which is to be reported to the
407 client for the digest proxy authentication scheme (part of
408 the text the user will see when prompted their username and
409 password). There is no default.
410 auth_param digest realm Squid proxy-caching web server
412 "nonce_garbage_interval" timeinterval
413 Specifies the interval that nonces that have been issued
414 to client_agent's are checked for validity.
416 "nonce_max_duration" timeinterval
417 Specifies the maximum length of time a given nonce will be
420 "nonce_max_count" number
421 Specifies the maximum number of times a given nonce can be
424 "nonce_strictness" on|off
425 Determines if squid requires strict increment-by-1 behavior
426 for nonce counts, or just incrementing (off - for use when
427 user agents generate nonce counts that occasionally miss 1
428 (ie, 1,2,4,6)). Default off.
430 "check_nonce_count" on|off
431 This directive if set to off can disable the nonce count check
432 completely to work around buggy digest qop implementations in
433 certain mainstream browser versions. Default on to check the
434 nonce count to protect from authentication replay attacks.
436 "post_workaround" on|off
437 This is a workaround to certain buggy browsers who sends
438 an incorrect request digest in POST requests when reusing
439 the same nonce as acquired earlier on a GET request.
441 === NTLM scheme options follow ===
444 Specify the command for the external NTLM authenticator.
445 Such a program reads exchanged NTLMSSP packets with
446 the browser via Squid until authentication is completed.
447 If you use an NTLM authenticator, make sure you have 1 acl
448 of type proxy_auth. By default, the NTLM authenticator program
451 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
453 "children" numberofchildren [startup=N] [idle=N]
454 The maximum number of authenticator processes to spawn (default 5).
455 If you start too few Squid will have to wait for them to
456 process a backlog of credential verifications, slowing it
457 down. When credential verifications are done via a (slow)
458 network you are likely to need lots of authenticator
461 The startup= and idle= options permit some skew in the exact amount
462 run. A minimum of startup=N will begin during startup and reconfigure.
463 Squid will start more in groups of up to idle=N in an attempt to meet
464 traffic needs and to keep idle=N free above those traffic needs up to
467 auth_param ntlm children 20 startup=0 idle=1
470 If you experience problems with PUT/POST requests when using the
471 Negotiate authentication scheme then you can try setting this to
472 off. This will cause Squid to forcibly close the connection on
473 the initial requests where the browser asks which schemes are
474 supported by the proxy.
476 auth_param ntlm keep_alive on
478 === Options for configuring the NEGOTIATE auth-scheme follow ===
481 Specify the command for the external Negotiate authenticator.
482 This protocol is used in Microsoft Active-Directory enabled setups with
483 the Microsoft Internet Explorer or Mozilla Firefox browsers.
484 Its main purpose is to exchange credentials with the Squid proxy
485 using the Kerberos mechanisms.
486 If you use a Negotiate authenticator, make sure you have at least
487 one acl of type proxy_auth active. By default, the negotiate
488 authenticator program is not used.
489 The only supported program for this role is the ntlm_auth
490 program distributed as part of Samba, version 4 or later.
492 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
494 "children" numberofchildren [startup=N] [idle=N]
495 The maximum number of authenticator processes to spawn (default 5).
496 If you start too few Squid will have to wait for them to
497 process a backlog of credential verifications, slowing it
498 down. When credential verifications are done via a (slow)
499 network you are likely to need lots of authenticator
502 The startup= and idle= options permit some skew in the exact amount
503 run. A minimum of startup=N will begin during startup and reconfigure.
504 Squid will start more in groups of up to idle=N in an attempt to meet
505 traffic needs and to keep idle=N free above those traffic needs up to
508 auth_param negotiate children 20 startup=0 idle=1
511 If you experience problems with PUT/POST requests when using the
512 Negotiate authentication scheme then you can try setting this to
513 off. This will cause Squid to forcibly close the connection on
514 the initial requests where the browser asks which schemes are
515 supported by the proxy.
517 auth_param negotiate keep_alive on
522 #Recommended minimum configuration per scheme:
523 #auth_param negotiate program <uncomment and complete this line to activate>
524 #auth_param negotiate children 20 startup=0 idle=1
525 #auth_param negotiate keep_alive on
527 #auth_param ntlm program <uncomment and complete this line to activate>
528 #auth_param ntlm children 20 startup=0 idle=1
529 #auth_param ntlm keep_alive on
531 #auth_param digest program <uncomment and complete this line>
532 #auth_param digest children 20 startup=0 idle=1
533 #auth_param digest realm Squid proxy-caching web server
534 #auth_param digest nonce_garbage_interval 5 minutes
535 #auth_param digest nonce_max_duration 30 minutes
536 #auth_param digest nonce_max_count 50
538 #auth_param basic program <uncomment and complete this line>
539 #auth_param basic children 5 startup=5 idle=1
540 #auth_param basic realm Squid proxy-caching web server
541 #auth_param basic credentialsttl 2 hours
544 NAME: authenticate_cache_garbage_interval
547 LOC: Config.authenticateGCInterval
549 The time period between garbage collection across the username cache.
550 This is a trade-off between memory utilization (long intervals - say
551 2 days) and CPU (short intervals - say 1 minute). Only change if you
555 NAME: authenticate_ttl
558 LOC: Config.authenticateTTL
560 The time a user & their credentials stay in the logged in
561 user cache since their last request. When the garbage
562 interval passes, all user credentials that have passed their
563 TTL are removed from memory.
566 NAME: authenticate_ip_ttl
568 LOC: Config.authenticateIpTTL
571 If you use proxy authentication and the 'max_user_ip' ACL,
572 this directive controls how long Squid remembers the IP
573 addresses associated with each user. Use a small value
574 (e.g., 60 seconds) if your users might change addresses
575 quickly, as is the case with dialup. You might be safe
576 using a larger value (e.g., 2 hours) in a corporate LAN
577 environment with relatively static address assignments.
582 -----------------------------------------------------------------------------
585 NAME: external_acl_type
586 TYPE: externalAclHelper
587 LOC: Config.externalAclHelperList
590 This option defines external acl classes using a helper program
591 to look up the status
593 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
597 ttl=n TTL in seconds for cached results (defaults to 3600
600 TTL for cached negative lookups (default same
603 Maximum number of acl helper processes spawned to service
604 external acl lookups of this type. (default 20)
606 Minimum number of acl helper processes to spawn during
607 startup and reconfigure to service external acl lookups
608 of this type. (default 0)
610 Number of acl helper processes to keep ahead of traffic
611 loads. Squid will spawn this many at once whenever load
612 rises above the capabilities of existing processes.
613 Up to the value of children-max. (default 1)
614 concurrency=n concurrency level per process. Only used with helpers
615 capable of processing more than one query at a time.
616 cache=n limit the result cache size, default is unbounded.
617 grace=n Percentage remaining of TTL where a refresh of a
618 cached entry should be initiated without needing to
619 wait for a new reply. (default is for no grace period)
620 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
621 ipv4 / ipv6 IP protocol used to communicate with this helper.
622 The default is to auto-detect IPv6 and use it when available.
624 FORMAT specifications
626 %LOGIN Authenticated user login name
627 %EXT_USER Username from previous external acl
628 %EXT_LOG Log details from previous external acl
629 %EXT_TAG Tag from previous external acl
630 %IDENT Ident user name
632 %SRCPORT Client source port
635 %PROTO Requested protocol
637 %PATH Requested URL path
638 %METHOD Request method
639 %MYADDR Squid interface address
640 %MYPORT Squid http_port number
641 %PATH Requested URL-path (including query-string if any)
642 %USER_CERT SSL User certificate in PEM format
643 %USER_CERTCHAIN SSL User certificate chain in PEM format
644 %USER_CERT_xx SSL User certificate subject attribute xx
645 %USER_CA_xx SSL User certificate issuer attribute xx
647 %>{Header} HTTP request header "Header"
649 HTTP request header "Hdr" list member "member"
651 HTTP request header list member using ; as
652 list separator. ; can be any non-alphanumeric
655 %<{Header} HTTP reply header "Header"
657 HTTP reply header "Hdr" list member "member"
659 HTTP reply header list member using ; as
660 list separator. ; can be any non-alphanumeric
663 %% The percent sign. Useful for helpers which need
664 an unchanging input format.
667 General request syntax:
669 [channel-ID] FORMAT-values [acl-values ...]
672 FORMAT-values consists of transaction details expanded with
673 whitespace separation per the config file FORMAT specification
674 using the FORMAT macros listed above.
676 acl-values consists of any string specified in the referencing
677 config 'acl ... external' line. see the "acl external" directive.
679 Request values sent to the helper are URL escaped to protect
680 each value in requests against whitespaces.
682 If using protocol=2.5 then the request sent to the helper is not
683 URL escaped to protect against whitespace.
685 NOTE: protocol=3.0 is deprecated as no longer necessary.
687 When using the concurrency= option the protocol is changed by
688 introducing a query channel tag in front of the request/response.
689 The query channel tag is a number between 0 and concurrency-1.
690 This value must be echoed back unchanged to Squid as the first part
691 of the response relating to its request.
694 The helper receives lines expanded per the above format specification
695 and for each input line returns 1 line starting with OK/ERR/BH result
696 code and optionally followed by additional keywords with more details.
699 General result syntax:
701 [channel-ID] result keyword=value ...
703 Result consists of one of the codes:
706 the ACL test produced a match.
709 the ACL test does not produce a match.
712 An internal error occurred in the helper, preventing
713 a result being identified.
715 The meaning of 'a match' is determined by your squid.conf
716 access control configuration. See the Squid wiki for details.
720 user= The users name (login)
722 password= The users password (for login= cache_peer option)
724 message= Message describing the reason for this response.
725 Available as %o in error pages.
726 Useful on (ERR and BH results).
728 tag= Apply a tag to a request. Only sets a tag once,
729 does not alter existing tags.
731 log= String to be logged in access.log. Available as
732 %ea in logformat specifications.
734 Any keywords may be sent on any response whether OK, ERR or BH.
736 All response keyword values need to be a single token with URL
737 escaping, or enclosed in double quotes (") and escaped using \ on
738 any double quotes or \ characters within the value. The wrapping
739 double quotes are removed before the value is interpreted by Squid.
740 \r and \n are also replace by CR and LF.
742 Some example key values:
746 user="J. \"Bob\" Smith"
753 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
754 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
755 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
756 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
757 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
760 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
761 DEFAULT: localhost src 127.0.0.1/32 ::1
762 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
763 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
765 Defining an Access List
767 Every access list definition must begin with an aclname and acltype,
768 followed by either type-specific arguments or a quoted filename that
771 acl aclname acltype argument ...
772 acl aclname acltype "file" ...
774 When using "file", the file should contain one item per line.
776 By default, regular expressions are CASE-SENSITIVE.
777 To make them case-insensitive, use the -i option. To return case-sensitive
778 use the +i option between patterns, or make a new ACL line without -i.
780 Some acl types require suspending the current request in order
781 to access some external data source.
782 Those which do are marked with the tag [slow], those which
783 don't are marked as [fast].
784 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
785 for further information
787 ***** ACL TYPES AVAILABLE *****
789 acl aclname src ip-address/mask ... # clients IP address [fast]
790 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
791 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
792 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
794 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
795 # The arp ACL requires the special configure option --enable-arp-acl.
796 # Furthermore, the ARP ACL code is not portable to all operating systems.
797 # It works on Linux, Solaris, Windows, FreeBSD, and some
798 # other *BSD variants.
801 # NOTE: Squid can only determine the MAC address for clients that are on
802 # the same subnet. If the client is on a different subnet,
803 # then Squid cannot find out its MAC address.
805 acl aclname srcdomain .foo.com ...
806 # reverse lookup, from client IP [slow]
807 acl aclname dstdomain .foo.com ...
808 # Destination server from URL [fast]
809 acl aclname srcdom_regex [-i] \.foo\.com ...
810 # regex matching client name [slow]
811 acl aclname dstdom_regex [-i] \.foo\.com ...
812 # regex matching server [fast]
814 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
815 # based URL is used and no match is found. The name "none" is used
816 # if the reverse lookup fails.
818 acl aclname src_as number ...
819 acl aclname dst_as number ...
821 # Except for access control, AS numbers can be used for
822 # routing of requests to specific caches. Here's an
823 # example for routing all requests for AS#1241 and only
824 # those to mycache.mydomain.net:
825 # acl asexample dst_as 1241
826 # cache_peer_access mycache.mydomain.net allow asexample
827 # cache_peer_access mycache_mydomain.net deny all
829 acl aclname peername myPeer ...
831 # match against a named cache_peer entry
832 # set unique name= on cache_peer lines for reliable use.
834 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
844 # h1:m1 must be less than h2:m2
846 acl aclname url_regex [-i] ^http:// ...
847 # regex matching on whole URL [fast]
848 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
849 # regex matching on URL login field
850 acl aclname urlpath_regex [-i] \.gif$ ...
851 # regex matching on URL path [fast]
853 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
855 acl aclname localport 3128 ... # TCP port the client connected to [fast]
856 # NP: for interception mode this is usually '80'
858 acl aclname myportname 3128 ... # http(s)_port name [fast]
860 acl aclname proto HTTP FTP ... # request protocol [fast]
862 acl aclname method GET POST ... # HTTP request method [fast]
864 acl aclname http_status 200 301 500- 400-403 ...
865 # status code in reply [fast]
867 acl aclname browser [-i] regexp ...
868 # pattern match on User-Agent header (see also req_header below) [fast]
870 acl aclname referer_regex [-i] regexp ...
871 # pattern match on Referer header [fast]
872 # Referer is highly unreliable, so use with care
874 acl aclname ident username ...
875 acl aclname ident_regex [-i] pattern ...
876 # string match on ident output [slow]
877 # use REQUIRED to accept any non-null ident.
879 acl aclname proxy_auth [-i] username ...
880 acl aclname proxy_auth_regex [-i] pattern ...
881 # perform http authentication challenge to the client and match against
882 # supplied credentials [slow]
884 # takes a list of allowed usernames.
885 # use REQUIRED to accept any valid username.
887 # Will use proxy authentication in forward-proxy scenarios, and plain
888 # http authenticaiton in reverse-proxy scenarios
890 # NOTE: when a Proxy-Authentication header is sent but it is not
891 # needed during ACL checking the username is NOT logged
894 # NOTE: proxy_auth requires a EXTERNAL authentication program
895 # to check username/password combinations (see
896 # auth_param directive).
898 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
899 # as the browser needs to be configured for using a proxy in order
900 # to respond to proxy authentication.
902 acl aclname snmp_community string ...
903 # A community string to limit access to your SNMP Agent [fast]
906 # acl snmppublic snmp_community public
908 acl aclname maxconn number
909 # This will be matched when the client's IP address has
910 # more than <number> TCP connections established. [fast]
911 # NOTE: This only measures direct TCP links so X-Forwarded-For
912 # indirect clients are not counted.
914 acl aclname max_user_ip [-s] number
915 # This will be matched when the user attempts to log in from more
916 # than <number> different ip addresses. The authenticate_ip_ttl
917 # parameter controls the timeout on the ip entries. [fast]
918 # If -s is specified the limit is strict, denying browsing
919 # from any further IP addresses until the ttl has expired. Without
920 # -s Squid will just annoy the user by "randomly" denying requests.
921 # (the counter is reset each time the limit is reached and a
923 # NOTE: in acceleration mode or where there is mesh of child proxies,
924 # clients may appear to come from multiple addresses if they are
925 # going through proxy farms, so a limit of 1 may cause user problems.
927 acl aclname random probability
928 # Pseudo-randomly match requests. Based on the probability given.
929 # Probability may be written as a decimal (0.333), fraction (1/3)
930 # or ratio of matches:non-matches (3:5).
932 acl aclname req_mime_type [-i] mime-type ...
933 # regex match against the mime type of the request generated
934 # by the client. Can be used to detect file upload or some
935 # types HTTP tunneling requests [fast]
936 # NOTE: This does NOT match the reply. You cannot use this
937 # to match the returned file type.
939 acl aclname req_header header-name [-i] any\.regex\.here
940 # regex match against any of the known request headers. May be
941 # thought of as a superset of "browser", "referer" and "mime-type"
944 acl aclname rep_mime_type [-i] mime-type ...
945 # regex match against the mime type of the reply received by
946 # squid. Can be used to detect file download or some
947 # types HTTP tunneling requests. [fast]
948 # NOTE: This has no effect in http_access rules. It only has
949 # effect in rules that affect the reply data stream such as
952 acl aclname rep_header header-name [-i] any\.regex\.here
953 # regex match against any of the known reply headers. May be
954 # thought of as a superset of "browser", "referer" and "mime-type"
957 acl aclname external class_name [arguments...]
958 # external ACL lookup via a helper class defined by the
959 # external_acl_type directive [slow]
961 acl aclname user_cert attribute values...
962 # match against attributes in a user SSL certificate
963 # attribute is one of DN/C/O/CN/L/ST [fast]
965 acl aclname ca_cert attribute values...
966 # match against attributes a users issuing CA SSL certificate
967 # attribute is one of DN/C/O/CN/L/ST [fast]
969 acl aclname ext_user username ...
970 acl aclname ext_user_regex [-i] pattern ...
971 # string match on username returned by external acl helper [slow]
972 # use REQUIRED to accept any non-null user name.
974 acl aclname tag tagvalue ...
975 # string match on tag returned by external acl helper [slow]
977 acl aclname hier_code codename ...
978 # string match against squid hierarchy code(s); [fast]
979 # e.g., DIRECT, PARENT_HIT, NONE, etc.
981 # NOTE: This has no effect in http_access rules. It only has
982 # effect in rules that affect the reply data stream such as
986 acl aclname ssl_error errorname
987 # match against SSL certificate validation error [fast]
989 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
992 # The following can be used as shortcuts for certificate properties:
993 # [ssl::]certHasExpired: the "not after" field is in the past
994 # [ssl::]certNotYetValid: the "not before" field is in the future
995 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
996 # [ssl::]certSelfSigned: The certificate is self signed.
997 # [ssl::]certDomainMismatch: The certificate CN domain does not
998 # match the name the name of the host we are connecting to.
1000 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
1001 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
1002 # predefined ACLs, just like the 'all' ACL.
1004 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
1005 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
1007 acl aclname server_cert_fingerprint [-sha1] fingerprint
1008 # match against server SSL certificate fingerprint [fast]
1010 # The fingerprint is the digest of the DER encoded version
1011 # of the whole certificate. The user should use the form: XX:XX:...
1012 # Optional argument specifies the digest algorithm to use.
1013 # The SHA1 digest algorithm is the default and is currently
1014 # the only algorithm supported (-sha1).
1018 acl macaddress arp 09:00:2b:23:45:67
1019 acl myexample dst_as 1241
1020 acl password proxy_auth REQUIRED
1021 acl fileupload req_mime_type -i ^multipart/form-data$
1022 acl javascript rep_mime_type -i ^application/x-javascript$
1026 # Recommended minimum configuration:
1029 # Example rule allowing access from your local networks.
1030 # Adapt to list your (internal) IP networks from where browsing
1032 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1033 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1034 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1035 acl localnet src fc00::/7 # RFC 4193 local private network range
1036 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1038 acl SSL_ports port 443
1039 acl Safe_ports port 80 # http
1040 acl Safe_ports port 21 # ftp
1041 acl Safe_ports port 443 # https
1042 acl Safe_ports port 70 # gopher
1043 acl Safe_ports port 210 # wais
1044 acl Safe_ports port 1025-65535 # unregistered ports
1045 acl Safe_ports port 280 # http-mgmt
1046 acl Safe_ports port 488 # gss-http
1047 acl Safe_ports port 591 # filemaker
1048 acl Safe_ports port 777 # multiling http
1049 acl CONNECT method CONNECT
1053 NAME: follow_x_forwarded_for
1055 IFDEF: FOLLOW_X_FORWARDED_FOR
1056 LOC: Config.accessList.followXFF
1057 DEFAULT_IF_NONE: deny all
1059 Allowing or Denying the X-Forwarded-For header to be followed to
1060 find the original source of a request.
1062 Requests may pass through a chain of several other proxies
1063 before reaching us. The X-Forwarded-For header will contain a
1064 comma-separated list of the IP addresses in the chain, with the
1065 rightmost address being the most recent.
1067 If a request reaches us from a source that is allowed by this
1068 configuration item, then we consult the X-Forwarded-For header
1069 to see where that host received the request from. If the
1070 X-Forwarded-For header contains multiple addresses, we continue
1071 backtracking until we reach an address for which we are not allowed
1072 to follow the X-Forwarded-For header, or until we reach the first
1073 address in the list. For the purpose of ACL used in the
1074 follow_x_forwarded_for directive the src ACL type always matches
1075 the address we are testing and srcdomain matches its rDNS.
1077 The end result of this process is an IP address that we will
1078 refer to as the indirect client address. This address may
1079 be treated as the client address for access control, ICAP, delay
1080 pools and logging, depending on the acl_uses_indirect_client,
1081 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1082 log_uses_indirect_client and tproxy_uses_indirect_client options.
1084 This clause only supports fast acl types.
1085 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1087 SECURITY CONSIDERATIONS:
1089 Any host for which we follow the X-Forwarded-For header
1090 can place incorrect information in the header, and Squid
1091 will use the incorrect information as if it were the
1092 source address of the request. This may enable remote
1093 hosts to bypass any access control restrictions that are
1094 based on the client's source addresses.
1098 acl localhost src 127.0.0.1
1099 acl my_other_proxy srcdomain .proxy.example.com
1100 follow_x_forwarded_for allow localhost
1101 follow_x_forwarded_for allow my_other_proxy
1104 NAME: acl_uses_indirect_client
1107 IFDEF: FOLLOW_X_FORWARDED_FOR
1109 LOC: Config.onoff.acl_uses_indirect_client
1111 Controls whether the indirect client address
1112 (see follow_x_forwarded_for) is used instead of the
1113 direct client address in acl matching.
1115 NOTE: maxconn ACL considers direct TCP links and indirect
1116 clients will always have zero. So no match.
1119 NAME: delay_pool_uses_indirect_client
1122 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1124 LOC: Config.onoff.delay_pool_uses_indirect_client
1126 Controls whether the indirect client address
1127 (see follow_x_forwarded_for) is used instead of the
1128 direct client address in delay pools.
1131 NAME: log_uses_indirect_client
1134 IFDEF: FOLLOW_X_FORWARDED_FOR
1136 LOC: Config.onoff.log_uses_indirect_client
1138 Controls whether the indirect client address
1139 (see follow_x_forwarded_for) is used instead of the
1140 direct client address in the access log.
1143 NAME: tproxy_uses_indirect_client
1146 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1148 LOC: Config.onoff.tproxy_uses_indirect_client
1150 Controls whether the indirect client address
1151 (see follow_x_forwarded_for) is used instead of the
1152 direct client address when spoofing the outgoing client.
1154 This has no effect on requests arriving in non-tproxy
1157 SECURITY WARNING: Usage of this option is dangerous
1158 and should not be used trivially. Correct configuration
1159 of follow_x_forewarded_for with a limited set of trusted
1160 sources is required to prevent abuse of your proxy.
1165 LOC: Config.accessList.http
1166 DEFAULT_IF_NONE: deny all
1168 Allowing or Denying access based on defined access lists
1170 Access to the HTTP port:
1171 http_access allow|deny [!]aclname ...
1173 NOTE on default values:
1175 If there are no "access" lines present, the default is to deny
1178 If none of the "access" lines cause a match, the default is the
1179 opposite of the last line in the list. If the last line was
1180 deny, the default is allow. Conversely, if the last line
1181 is allow, the default will be deny. For these reasons, it is a
1182 good idea to have an "deny all" entry at the end of your access
1183 lists to avoid potential confusion.
1185 This clause supports both fast and slow acl types.
1186 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1191 # Recommended minimum Access Permission configuration:
1193 # Only allow cachemgr access from localhost
1194 http_access allow localhost manager
1195 http_access deny manager
1197 # Deny requests to certain unsafe ports
1198 http_access deny !Safe_ports
1200 # Deny CONNECT to other than secure SSL ports
1201 http_access deny CONNECT !SSL_ports
1203 # We strongly recommend the following be uncommented to protect innocent
1204 # web applications running on the proxy server who think the only
1205 # one who can access services on "localhost" is a local user
1206 #http_access deny to_localhost
1209 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1212 # Example rule allowing access from your local networks.
1213 # Adapt localnet in the ACL section to list your (internal) IP networks
1214 # from where browsing should be allowed
1215 http_access allow localnet
1216 http_access allow localhost
1218 # And finally deny all other access to this proxy
1219 http_access deny all
1223 NAME: adapted_http_access http_access2
1225 LOC: Config.accessList.adapted_http
1228 Allowing or Denying access based on defined access lists
1230 Essentially identical to http_access, but runs after redirectors
1231 and ICAP/eCAP adaptation. Allowing access control based on their
1234 If not set then only http_access is used.
1237 NAME: http_reply_access
1239 LOC: Config.accessList.reply
1242 Allow replies to client requests. This is complementary to http_access.
1244 http_reply_access allow|deny [!] aclname ...
1246 NOTE: if there are no access lines present, the default is to allow
1249 If none of the access lines cause a match the opposite of the
1250 last line will apply. Thus it is good practice to end the rules
1251 with an "allow all" or "deny all" entry.
1253 This clause supports both fast and slow acl types.
1254 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1259 LOC: Config.accessList.icp
1260 DEFAULT_IF_NONE: deny all
1262 Allowing or Denying access to the ICP port based on defined
1265 icp_access allow|deny [!]aclname ...
1267 See http_access for details
1269 This clause only supports fast acl types.
1270 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1272 # Allow ICP queries from local networks only
1273 #icp_access allow localnet
1274 #icp_access deny all
1280 LOC: Config.accessList.htcp
1281 DEFAULT_IF_NONE: deny all
1283 Allowing or Denying access to the HTCP port based on defined
1286 htcp_access allow|deny [!]aclname ...
1288 See http_access for details
1290 NOTE: The default if no htcp_access lines are present is to
1291 deny all traffic. This default may cause problems with peers
1292 using the htcp option.
1294 This clause only supports fast acl types.
1295 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1297 # Allow HTCP queries from local networks only
1298 #htcp_access allow localnet
1299 #htcp_access deny all
1302 NAME: htcp_clr_access
1305 LOC: Config.accessList.htcp_clr
1306 DEFAULT_IF_NONE: deny all
1308 Allowing or Denying access to purge content using HTCP based
1309 on defined access lists
1311 htcp_clr_access allow|deny [!]aclname ...
1313 See http_access for details
1315 This clause only supports fast acl types.
1316 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1318 # Allow HTCP CLR requests from trusted peers
1319 acl htcp_clr_peer src 172.16.1.2
1320 htcp_clr_access allow htcp_clr_peer
1325 LOC: Config.accessList.miss
1328 Determins whether network access is permitted when satisfying a request.
1331 to force your neighbors to use you as a sibling instead of
1334 acl localclients src 172.16.0.0/16
1335 miss_access allow localclients
1336 miss_access deny !localclients
1338 This means only your local clients are allowed to fetch relayed/MISS
1339 replies from the network and all other clients can only fetch cached
1343 The default for this setting allows all clients who passed the
1344 http_access rules to relay via this proxy.
1346 This clause only supports fast acl types.
1347 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1350 NAME: ident_lookup_access
1353 DEFAULT_IF_NONE: deny all
1354 LOC: Ident::TheConfig.identLookup
1356 A list of ACL elements which, if matched, cause an ident
1357 (RFC 931) lookup to be performed for this request. For
1358 example, you might choose to always perform ident lookups
1359 for your main multi-user Unix boxes, but not for your Macs
1360 and PCs. By default, ident lookups are not performed for
1363 To enable ident lookups for specific client addresses, you
1364 can follow this example:
1366 acl ident_aware_hosts src 198.168.1.0/24
1367 ident_lookup_access allow ident_aware_hosts
1368 ident_lookup_access deny all
1370 Only src type ACL checks are fully supported. A srcdomain
1371 ACL might work at times, but it will not always provide
1374 This clause only supports fast acl types.
1375 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1378 NAME: reply_body_max_size
1379 COMMENT: size [acl acl...]
1382 LOC: Config.ReplyBodySize
1384 This option specifies the maximum size of a reply body. It can be
1385 used to prevent users from downloading very large files, such as
1386 MP3's and movies. When the reply headers are received, the
1387 reply_body_max_size lines are processed, and the first line where
1388 all (if any) listed ACLs are true is used as the maximum body size
1391 This size is checked twice. First when we get the reply headers,
1392 we check the content-length value. If the content length value exists
1393 and is larger than the allowed size, the request is denied and the
1394 user receives an error message that says "the request or reply
1395 is too large." If there is no content-length, and the reply
1396 size exceeds this limit, the client's connection is just closed
1397 and they will receive a partial reply.
1399 WARNING: downstream caches probably can not detect a partial reply
1400 if there is no content-length header, so they will cache
1401 partial responses and give them out as hits. You should NOT
1402 use this option if you have downstream caches.
1404 WARNING: A maximum size smaller than the size of squid's error messages
1405 will cause an infinite loop and crash squid. Ensure that the smallest
1406 non-zero value you use is greater that the maximum header size plus
1407 the size of your largest error page.
1409 If you set this parameter none (the default), there will be
1412 Configuration Format is:
1413 reply_body_max_size SIZE UNITS [acl ...]
1415 reply_body_max_size 10 MB
1421 -----------------------------------------------------------------------------
1424 NAME: http_port ascii_port
1427 LOC: Config.Sockaddr.http
1429 Usage: port [mode] [options]
1430 hostname:port [mode] [options]
1431 1.2.3.4:port [mode] [options]
1433 The socket addresses where Squid will listen for HTTP client
1434 requests. You may specify multiple socket addresses.
1435 There are three forms: port alone, hostname with port, and
1436 IP address with port. If you specify a hostname or IP
1437 address, Squid binds the socket to that specific
1438 address. Most likely, you do not need to bind to a specific
1439 address, so you can use the port number alone.
1441 If you are running Squid in accelerator mode, you
1442 probably want to listen on port 80 also, or instead.
1444 The -a command line option may be used to specify additional
1445 port(s) where Squid listens for proxy request. Such ports will
1446 be plain proxy ports with no options.
1448 You may specify multiple socket addresses on multiple lines.
1452 intercept Support for IP-Layer interception of
1453 outgoing requests without browser settings.
1454 NP: disables authentication and IPv6 on the port.
1456 tproxy Support Linux TPROXY for spoofing outgoing
1457 connections using the client IP address.
1458 NP: disables authentication and maybe IPv6 on the port.
1460 accel Accelerator / reverse proxy mode
1462 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1463 establish secure connection with the client and with
1464 the server, decrypt HTTPS messages as they pass through
1465 Squid, and treat them as unencrypted HTTP messages,
1466 becoming the man-in-the-middle.
1468 The ssl_bump option is required to fully enable
1469 bumping of CONNECT requests.
1471 Omitting the mode flag causes default forward proxy mode to be used.
1474 Accelerator Mode Options:
1476 defaultsite=domainname
1477 What to use for the Host: header if it is not present
1478 in a request. Determines what site (not origin server)
1479 accelerators should consider the default.
1481 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1483 protocol= Protocol to reconstruct accelerated requests with.
1484 Defaults to http for http_port and https for
1487 vport Virtual host port support. Using the http_port number
1488 instead of the port passed on Host: headers.
1490 vport=NN Virtual host port support. Using the specified port
1491 number instead of the port passed on Host: headers.
1494 Act as if this Squid is the origin server.
1495 This currently means generate new Date: and Expires:
1496 headers on HIT instead of adding Age:.
1498 ignore-cc Ignore request Cache-Control headers.
1500 WARNING: This option violates HTTP specifications if
1501 used in non-accelerator setups.
1503 allow-direct Allow direct forwarding in accelerator mode. Normally
1504 accelerated requests are denied direct forwarding as if
1505 never_direct was used.
1507 WARNING: this option opens accelerator mode to security
1508 vulnerabilities usually only affecting in interception
1509 mode. Make sure to protect forwarding with suitable
1510 http_access rules when using this.
1513 SSL Bump Mode Options:
1514 In addition to these options ssl-bump requires TLS/SSL options.
1516 generate-host-certificates[=<on|off>]
1517 Dynamically create SSL server certificates for the
1518 destination hosts of bumped CONNECT requests.When
1519 enabled, the cert and key options are used to sign
1520 generated certificates. Otherwise generated
1521 certificate will be selfsigned.
1522 If there is a CA certificate lifetime of the generated
1523 certificate equals lifetime of the CA certificate. If
1524 generated certificate is selfsigned lifetime is three
1526 This option is enabled by default when ssl-bump is used.
1527 See the ssl-bump option above for more information.
1529 dynamic_cert_mem_cache_size=SIZE
1530 Approximate total RAM size spent on cached generated
1531 certificates. If set to zero, caching is disabled. The
1532 default value is 4MB.
1536 cert= Path to SSL certificate (PEM format).
1538 key= Path to SSL private key file (PEM format)
1539 if not specified, the certificate file is
1540 assumed to be a combined certificate and
1543 version= The version of SSL/TLS supported
1544 1 automatic (default)
1551 cipher= Colon separated list of supported ciphers.
1552 NOTE: some ciphers such as EDH ciphers depend on
1553 additional settings. If those settings are
1554 omitted the ciphers may be silently ignored
1555 by the OpenSSL library.
1557 options= Various SSL implementation options. The most important
1559 NO_SSLv2 Disallow the use of SSLv2
1560 NO_SSLv3 Disallow the use of SSLv3
1561 NO_TLSv1 Disallow the use of TLSv1.0
1562 NO_TLSv1_1 Disallow the use of TLSv1.1
1563 NO_TLSv1_2 Disallow the use of TLSv1.2
1564 SINGLE_DH_USE Always create a new key when using
1565 temporary/ephemeral DH key exchanges
1566 ALL Enable various bug workarounds
1567 suggested as "harmless" by OpenSSL
1568 Be warned that this reduces SSL/TLS
1569 strength to some attacks.
1570 See OpenSSL SSL_CTX_set_options documentation for a
1571 complete list of options.
1573 clientca= File containing the list of CAs to use when
1574 requesting a client certificate.
1576 cafile= File containing additional CA certificates to
1577 use when verifying client certificates. If unset
1578 clientca will be used.
1580 capath= Directory containing additional CA certificates
1581 and CRL lists to use when verifying client certificates.
1583 crlfile= File of additional CRL lists to use when verifying
1584 the client certificate, in addition to CRLs stored in
1585 the capath. Implies VERIFY_CRL flag below.
1587 dhparams= File containing DH parameters for temporary/ephemeral
1588 DH key exchanges. See OpenSSL documentation for details
1589 on how to create this file.
1590 WARNING: EDH ciphers will be silently disabled if this
1593 sslflags= Various flags modifying the use of SSL:
1595 Don't request client certificates
1596 immediately, but wait until acl processing
1597 requires a certificate (not yet implemented).
1599 Don't use the default CA lists built in
1602 Don't allow for session reuse. Each connection
1603 will result in a new SSL session.
1605 Verify CRL lists when accepting client
1608 Verify CRL lists for all certificates in the
1609 client certificate chain.
1611 sslcontext= SSL session ID context identifier.
1615 connection-auth[=on|off]
1616 use connection-auth=off to tell Squid to prevent
1617 forwarding Microsoft connection oriented authentication
1618 (NTLM, Negotiate and Kerberos)
1620 disable-pmtu-discovery=
1621 Control Path-MTU discovery usage:
1622 off lets OS decide on what to do (default).
1623 transparent disable PMTU discovery when transparent
1625 always disable always PMTU discovery.
1627 In many setups of transparently intercepting proxies
1628 Path-MTU discovery can not work on traffic towards the
1629 clients. This is the case when the intercepting device
1630 does not fully track connections and fails to forward
1631 ICMP must fragment messages to the cache server. If you
1632 have such setup and experience that certain clients
1633 sporadically hang or never complete requests set
1634 disable-pmtu-discovery option to 'transparent'.
1636 name= Specifies a internal name for the port. Defaults to
1637 the port specification (port or addr:port)
1639 tcpkeepalive[=idle,interval,timeout]
1640 Enable TCP keepalive probes of idle connections.
1641 In seconds; idle is the initial time before TCP starts
1642 probing the connection, interval how often to probe, and
1643 timeout the time before giving up.
1645 If you run Squid on a dual-homed machine with an internal
1646 and an external interface we recommend you to specify the
1647 internal address:port in http_port. This way Squid will only be
1648 visible on the internal address.
1652 # Squid normally listens to port 3128
1653 http_port @DEFAULT_HTTP_PORT@
1661 LOC: Config.Sockaddr.https
1663 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1665 The socket address where Squid will listen for client requests made
1666 over TLS or SSL connections. Commonly referred to as HTTPS.
1668 This is most useful for situations where you are running squid in
1669 accelerator mode and you want to do the SSL work at the accelerator level.
1671 You may specify multiple socket addresses on multiple lines,
1672 each with their own SSL certificate and/or options.
1676 accel Accelerator / reverse proxy mode
1678 intercept Support for IP-Layer interception of
1679 outgoing requests without browser settings.
1680 NP: disables authentication and IPv6 on the port.
1682 tproxy Support Linux TPROXY for spoofing outgoing
1683 connections using the client IP address.
1684 NP: disables authentication and maybe IPv6 on the port.
1686 ssl-bump For each intercepted connection allowed by ssl_bump
1687 ACLs, establish a secure connection with the client and with
1688 the server, decrypt HTTPS messages as they pass through
1689 Squid, and treat them as unencrypted HTTP messages,
1690 becoming the man-in-the-middle.
1692 An "ssl_bump server-first" match is required to
1693 fully enable bumping of intercepted SSL connections.
1695 Requires tproxy or intercept.
1697 Omitting the mode flag causes default forward proxy mode to be used.
1700 See http_port for a list of generic options
1705 cert= Path to SSL certificate (PEM format).
1707 key= Path to SSL private key file (PEM format)
1708 if not specified, the certificate file is
1709 assumed to be a combined certificate and
1712 version= The version of SSL/TLS supported
1713 1 automatic (default)
1718 cipher= Colon separated list of supported ciphers.
1720 options= Various SSL engine options. The most important
1722 NO_SSLv2 Disallow the use of SSLv2
1723 NO_SSLv3 Disallow the use of SSLv3
1724 NO_TLSv1 Disallow the use of TLSv1
1725 SINGLE_DH_USE Always create a new key when using
1726 temporary/ephemeral DH key exchanges
1727 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1728 documentation for a complete list of options.
1730 clientca= File containing the list of CAs to use when
1731 requesting a client certificate.
1733 cafile= File containing additional CA certificates to
1734 use when verifying client certificates. If unset
1735 clientca will be used.
1737 capath= Directory containing additional CA certificates
1738 and CRL lists to use when verifying client certificates.
1740 crlfile= File of additional CRL lists to use when verifying
1741 the client certificate, in addition to CRLs stored in
1742 the capath. Implies VERIFY_CRL flag below.
1744 dhparams= File containing DH parameters for temporary/ephemeral
1747 sslflags= Various flags modifying the use of SSL:
1749 Don't request client certificates
1750 immediately, but wait until acl processing
1751 requires a certificate (not yet implemented).
1753 Don't use the default CA lists built in
1756 Don't allow for session reuse. Each connection
1757 will result in a new SSL session.
1759 Verify CRL lists when accepting client
1762 Verify CRL lists for all certificates in the
1763 client certificate chain.
1765 sslcontext= SSL session ID context identifier.
1767 generate-host-certificates[=<on|off>]
1768 Dynamically create SSL server certificates for the
1769 destination hosts of bumped SSL requests.When
1770 enabled, the cert and key options are used to sign
1771 generated certificates. Otherwise generated
1772 certificate will be selfsigned.
1773 If there is CA certificate life time of generated
1774 certificate equals lifetime of CA certificate. If
1775 generated certificate is selfsigned lifetime is three
1777 This option is enabled by default when SslBump is used.
1778 See the sslBump option above for more information.
1780 dynamic_cert_mem_cache_size=SIZE
1781 Approximate total RAM size spent on cached generated
1782 certificates. If set to zero, caching is disabled. The
1783 default value is 4MB.
1785 See http_port for a list of available options.
1788 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1791 LOC: Ip::Qos::TheConfig.tosToServer
1793 Allows you to select a TOS/Diffserv value for packets outgoing
1794 on the server side, based on an ACL.
1796 tcp_outgoing_tos ds-field [!]aclname ...
1798 Example where normal_service_net uses the TOS value 0x00
1799 and good_service_net uses 0x20
1801 acl normal_service_net src 10.0.0.0/24
1802 acl good_service_net src 10.0.1.0/24
1803 tcp_outgoing_tos 0x00 normal_service_net
1804 tcp_outgoing_tos 0x20 good_service_net
1806 TOS/DSCP values really only have local significance - so you should
1807 know what you're specifying. For more information, see RFC2474,
1808 RFC2475, and RFC3260.
1810 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1811 "default" to use whatever default your host has. Note that in
1812 practice often only multiples of 4 is usable as the two rightmost bits
1813 have been redefined for use by ECN (RFC 3168 section 23.1).
1815 Processing proceeds in the order specified, and stops at first fully
1819 NAME: clientside_tos
1822 LOC: Ip::Qos::TheConfig.tosToClient
1824 Allows you to select a TOS/Diffserv value for packets being transmitted
1825 on the client-side, based on an ACL.
1827 clientside_tos ds-field [!]aclname ...
1829 Example where normal_service_net uses the TOS value 0x00
1830 and good_service_net uses 0x20
1832 acl normal_service_net src 10.0.0.0/24
1833 acl good_service_net src 10.0.1.0/24
1834 clientside_tos 0x00 normal_service_net
1835 clientside_tos 0x20 good_service_net
1837 Note: This feature is incompatible with qos_flows. Any TOS values set here
1838 will be overwritten by TOS values in qos_flows.
1841 NAME: tcp_outgoing_mark
1843 IFDEF: SO_MARK&&USE_LIBCAP
1845 LOC: Ip::Qos::TheConfig.nfmarkToServer
1847 Allows you to apply a Netfilter mark value to outgoing packets
1848 on the server side, based on an ACL.
1850 tcp_outgoing_mark mark-value [!]aclname ...
1852 Example where normal_service_net uses the mark value 0x00
1853 and good_service_net uses 0x20
1855 acl normal_service_net src 10.0.0.0/24
1856 acl good_service_net src 10.0.1.0/24
1857 tcp_outgoing_mark 0x00 normal_service_net
1858 tcp_outgoing_mark 0x20 good_service_net
1861 NAME: clientside_mark
1863 IFDEF: SO_MARK&&USE_LIBCAP
1865 LOC: Ip::Qos::TheConfig.nfmarkToClient
1867 Allows you to apply a Netfilter mark value to packets being transmitted
1868 on the client-side, based on an ACL.
1870 clientside_mark mark-value [!]aclname ...
1872 Example where normal_service_net uses the mark value 0x00
1873 and good_service_net uses 0x20
1875 acl normal_service_net src 10.0.0.0/24
1876 acl good_service_net src 10.0.1.0/24
1877 clientside_mark 0x00 normal_service_net
1878 clientside_mark 0x20 good_service_net
1880 Note: This feature is incompatible with qos_flows. Any mark values set here
1881 will be overwritten by mark values in qos_flows.
1888 LOC: Ip::Qos::TheConfig
1890 Allows you to select a TOS/DSCP value to mark outgoing
1891 connections with, based on where the reply was sourced. For
1892 platforms using netfilter, allows you to set a netfilter mark
1893 value instead of, or in addition to, a TOS value.
1895 TOS values really only have local significance - so you should
1896 know what you're specifying. For more information, see RFC2474,
1897 RFC2475, and RFC3260.
1899 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1900 in practice often only multiples of 4 is usable as the two rightmost bits
1901 have been redefined for use by ECN (RFC 3168 section 23.1).
1903 Mark values can be any unsigned 32-bit integer value.
1905 This setting is configured by setting the following values:
1907 tos|mark Whether to set TOS or netfilter mark values
1909 local-hit=0xFF Value to mark local cache hits.
1911 sibling-hit=0xFF Value to mark hits from sibling peers.
1913 parent-hit=0xFF Value to mark hits from parent peers.
1915 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1916 over the preserve-miss feature (see below), unless
1917 mask is specified, in which case only the bits
1918 specified in the mask are written.
1920 The TOS variant of the following features are only possible on Linux
1921 and require your kernel to be patched with the TOS preserving ZPH
1922 patch, available from http://zph.bratcheda.org
1923 No patch is needed to preserve the netfilter mark, which will work
1924 with all variants of netfilter.
1926 disable-preserve-miss
1927 This option disables the preservation of the TOS or netfilter
1928 mark. By default, the existing TOS or netfilter mark value of
1929 the response coming from the remote server will be retained
1930 and masked with miss-mark.
1931 NOTE: in the case of a netfilter mark, the mark must be set on
1932 the connection (using the CONNMARK target) not on the packet
1936 Allows you to mask certain bits in the TOS or mark value
1937 received from the remote server, before copying the value to
1938 the TOS sent towards clients.
1939 Default for tos: 0xFF (TOS from server is not changed).
1940 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1942 All of these features require the --enable-zph-qos compilation flag
1943 (enabled by default). Netfilter marking also requires the
1944 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1945 libcap 2.09+ (--with-libcap).
1949 NAME: tcp_outgoing_address
1952 LOC: Config.accessList.outgoing_address
1954 Allows you to map requests to different outgoing IP addresses
1955 based on the username or source address of the user making
1958 tcp_outgoing_address ipaddr [[!]aclname] ...
1961 Forwarding clients with dedicated IPs for certain subnets.
1963 acl normal_service_net src 10.0.0.0/24
1964 acl good_service_net src 10.0.2.0/24
1966 tcp_outgoing_address 2001:db8::c001 good_service_net
1967 tcp_outgoing_address 10.1.0.2 good_service_net
1969 tcp_outgoing_address 2001:db8::beef normal_service_net
1970 tcp_outgoing_address 10.1.0.1 normal_service_net
1972 tcp_outgoing_address 2001:db8::1
1973 tcp_outgoing_address 10.1.0.3
1975 Processing proceeds in the order specified, and stops at first fully
1978 Squid will add an implicit IP version test to each line.
1979 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1980 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1983 NOTE: The use of this directive using client dependent ACLs is
1984 incompatible with the use of server side persistent connections. To
1985 ensure correct results it is best to set server_persistent_connections
1986 to off when using this directive in such configurations.
1988 NOTE: The use of this directive to set a local IP on outgoing TCP links
1989 is incompatible with using TPROXY to set client IP out outbound TCP links.
1990 When needing to contact peers use the no-tproxy cache_peer option and the
1991 client_dst_passthru directive re-enable normal forwarding such as this.
1995 NAME: host_verify_strict
1998 LOC: Config.onoff.hostStrictVerify
2000 Regardless of this option setting, when dealing with intercepted
2001 traffic, Squid always verifies that the destination IP address matches
2002 the Host header domain or IP (called 'authority form URL').
2004 This enforcement is performed to satisfy a MUST-level requirement in
2005 RFC 2616 section 14.23: "The Host field value MUST represent the naming
2006 authority of the origin server or gateway given by the original URL".
2009 Squid always responds with an HTTP 409 (Conflict) error
2010 page and logs a security warning if there is no match.
2012 Squid verifies that the destination IP address matches
2013 the Host header for forward-proxy and reverse-proxy traffic
2014 as well. For those traffic types, Squid also enables the
2015 following checks, comparing the corresponding Host header
2016 and Request-URI components:
2018 * The host names (domain or IP) must be identical,
2019 but valueless or missing Host header disables all checks.
2020 For the two host names to match, both must be either IP
2023 * Port numbers must be identical, but if a port is missing
2024 the scheme-default port is assumed.
2027 When set to OFF (the default):
2028 Squid allows suspicious requests to continue but logs a
2029 security warning and blocks caching of the response.
2031 * Forward-proxy traffic is not checked at all.
2033 * Reverse-proxy traffic is not checked at all.
2035 * Intercepted traffic which passes verification is handled
2036 according to client_dst_passthru.
2038 * Intercepted requests which fail verification are sent
2039 to the client original destination instead of DIRECT.
2040 This overrides 'client_dst_passthru off'.
2042 For now suspicious intercepted CONNECT requests are always
2043 responded to with an HTTP 409 (Conflict) error page.
2048 As described in CVE-2009-0801 when the Host: header alone is used
2049 to determine the destination of a request it becomes trivial for
2050 malicious scripts on remote websites to bypass browser same-origin
2051 security policy and sandboxing protections.
2053 The cause of this is that such applets are allowed to perform their
2054 own HTTP stack, in which case the same-origin policy of the browser
2055 sandbox only verifies that the applet tries to contact the same IP
2056 as from where it was loaded at the IP level. The Host: header may
2057 be different from the connected IP and approved origin.
2061 NAME: client_dst_passthru
2064 LOC: Config.onoff.client_dst_passthru
2066 With NAT or TPROXY intercepted traffic Squid may pass the request
2067 directly to the original client destination IP or seek a faster
2068 source using the HTTP Host header.
2070 Using Host to locate alternative servers can provide faster
2071 connectivity with a range of failure recovery options.
2072 But can also lead to connectivity trouble when the client and
2073 server are attempting stateful interactions unaware of the proxy.
2075 This option (on by default) prevents alternative DNS entries being
2076 located to send intercepted traffic DIRECT to an origin server.
2077 The clients original destination IP and port will be used instead.
2079 Regardless of this option setting, when dealing with intercepted
2080 traffic Squid will verify the Host: header and any traffic which
2081 fails Host verification will be treated as if this option were ON.
2083 see host_verify_strict for details on the verification process.
2088 -----------------------------------------------------------------------------
2091 NAME: ssl_unclean_shutdown
2095 LOC: Config.SSL.unclean_shutdown
2097 Some browsers (especially MSIE) bugs out on SSL shutdown
2104 LOC: Config.SSL.ssl_engine
2107 The OpenSSL engine to use. You will need to set this if you
2108 would like to use hardware SSL acceleration for example.
2111 NAME: sslproxy_client_certificate
2114 LOC: Config.ssl_client.cert
2117 Client SSL Certificate to use when proxying https:// URLs
2120 NAME: sslproxy_client_key
2123 LOC: Config.ssl_client.key
2126 Client SSL Key to use when proxying https:// URLs
2129 NAME: sslproxy_version
2132 LOC: Config.ssl_client.version
2135 SSL version level to use when proxying https:// URLs
2137 The versions of SSL/TLS supported:
2139 1 automatic (default)
2147 NAME: sslproxy_options
2150 LOC: Config.ssl_client.options
2153 SSL implementation options to use when proxying https:// URLs
2155 The most important being:
2157 NO_SSLv2 Disallow the use of SSLv2
2158 NO_SSLv3 Disallow the use of SSLv3
2159 NO_TLSv1 Disallow the use of TLSv1.0
2160 NO_TLSv1_1 Disallow the use of TLSv1.1
2161 NO_TLSv1_2 Disallow the use of TLSv1.2
2163 Always create a new key when using temporary/ephemeral
2166 Disable use of RFC5077 session tickets. Some servers
2167 may have problems understanding the TLS extension due
2168 to ambiguous specification in RFC4507.
2169 ALL Enable various bug workarounds suggested as "harmless"
2170 by OpenSSL. Be warned that this may reduce SSL/TLS
2171 strength to some attacks.
2173 See the OpenSSL SSL_CTX_set_options documentation for a
2174 complete list of possible options.
2177 NAME: sslproxy_cipher
2180 LOC: Config.ssl_client.cipher
2183 SSL cipher list to use when proxying https:// URLs
2185 Colon separated list of supported ciphers.
2188 NAME: sslproxy_cafile
2191 LOC: Config.ssl_client.cafile
2194 file containing CA certificates to use when verifying server
2195 certificates while proxying https:// URLs
2198 NAME: sslproxy_capath
2201 LOC: Config.ssl_client.capath
2204 directory containing CA certificates to use when verifying
2205 server certificates while proxying https:// URLs
2210 TYPE: sslproxy_ssl_bump
2211 LOC: Config.accessList.ssl_bump
2214 This option is consulted when a CONNECT request is received on
2215 an http_port (or a new connection is intercepted at an
2216 https_port), provided that port was configured with an ssl-bump
2217 flag. The subsequent data on the connection is either treated as
2218 HTTPS and decrypted OR tunneled at TCP level without decryption,
2219 depending on the first bumping "mode" which ACLs match.
2221 ssl_bump <mode> [!]acl ...
2223 The following bumping modes are supported:
2226 Allow bumping of the connection. Establish a secure connection
2227 with the client first, then connect to the server. This old mode
2228 does not allow Squid to mimic server SSL certificate and does
2229 not work with intercepted SSL connections.
2232 Allow bumping of the connection. Establish a secure connection
2233 with the server first, then establish a secure connection with
2234 the client, using a mimicked server certificate. Works with both
2235 CONNECT requests and intercepted SSL connections.
2238 Become a TCP tunnel without decoding the connection.
2239 Works with both CONNECT requests and intercepted SSL
2240 connections. This is the default behavior when no
2241 ssl_bump option is given or no ssl_bump ACLs match.
2243 By default, no connections are bumped.
2245 The first matching ssl_bump option wins. If no ACLs match, the
2246 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2247 does not have an implicit "negate the last given option" rule. You
2248 must make that rule explicit if you convert old ssl_bump allow/deny
2249 rules that rely on such an implicit rule.
2251 This clause supports both fast and slow acl types.
2252 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2254 See also: http_port ssl-bump, https_port ssl-bump
2257 # Example: Bump all requests except those originating from
2258 # localhost and those going to example.com.
2260 acl broken_sites dstdomain .example.com
2261 ssl_bump none localhost
2262 ssl_bump none broken_sites
2263 ssl_bump server-first all
2266 NAME: sslproxy_flags
2269 LOC: Config.ssl_client.flags
2272 Various flags modifying the use of SSL while proxying https:// URLs:
2273 DONT_VERIFY_PEER Accept certificates that fail verification.
2274 For refined control, see sslproxy_cert_error.
2275 NO_DEFAULT_CA Don't use the default CA list built in
2279 NAME: sslproxy_cert_error
2282 LOC: Config.ssl_client.cert_error
2285 Use this ACL to bypass server certificate validation errors.
2287 For example, the following lines will bypass all validation errors
2288 when talking to servers for example.com. All other
2289 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2291 acl BrokenButTrustedServers dstdomain example.com
2292 sslproxy_cert_error allow BrokenButTrustedServers
2293 sslproxy_cert_error deny all
2295 This clause only supports fast acl types.
2296 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2297 Using slow acl types may result in server crashes
2299 Without this option, all server certificate validation errors
2300 terminate the transaction. Bypassing validation errors is dangerous
2301 because an error usually implies that the server cannot be trusted and
2302 the connection may be insecure.
2304 See also: sslproxy_flags and DONT_VERIFY_PEER.
2306 Default setting: sslproxy_cert_error deny all
2309 NAME: sslproxy_cert_sign
2312 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2313 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2314 POSTSCRIPTUM: signTrusted all
2315 TYPE: sslproxy_cert_sign
2316 LOC: Config.ssl_client.cert_sign
2319 sslproxy_cert_sign <signing algorithm> acl ...
2321 The following certificate signing algorithms are supported:
2323 Sign using the configured CA certificate which is usually
2324 placed in and trusted by end-user browsers. This is the
2325 default for trusted origin server certificates.
2327 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2328 This is the default for untrusted origin server certificates
2329 that are not self-signed (see ssl::certUntrusted).
2331 Sign using a self-signed certificate with the right CN to
2332 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2333 browser. This is the default for self-signed origin server
2334 certificates (see ssl::certSelfSigned).
2336 This clause only supports fast acl types.
2338 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2339 signing algorithm to generate the certificate and ignores all
2340 subsequent sslproxy_cert_sign options (the first match wins). If no
2341 acl(s) match, the default signing algorithm is determined by errors
2342 detected when obtaining and validating the origin server certificate.
2344 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2345 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2346 CONNECT request that carries a domain name. In all other cases (CONNECT
2347 to an IP address or an intercepted SSL connection), Squid cannot detect
2348 the domain mismatch at certificate generation time when
2349 bump-server-first is used.
2352 NAME: sslproxy_cert_adapt
2355 TYPE: sslproxy_cert_adapt
2356 LOC: Config.ssl_client.cert_adapt
2359 sslproxy_cert_adapt <adaptation algorithm> acl ...
2361 The following certificate adaptation algorithms are supported:
2363 Sets the "Not After" property to the "Not After" property of
2364 the CA certificate used to sign generated certificates.
2366 Sets the "Not Before" property to the "Not Before" property of
2367 the CA certificate used to sign generated certificates.
2368 setCommonName or setCommonName{CN}
2369 Sets Subject.CN property to the host name specified as a
2370 CN parameter or, if no explicit CN parameter was specified,
2371 extracted from the CONNECT request. It is a misconfiguration
2372 to use setCommonName without an explicit parameter for
2373 intercepted or tproxied SSL connections.
2375 This clause only supports fast acl types.
2377 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2378 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2379 corresponding adaptation algorithm to generate the certificate and
2380 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2381 group (i.e., the first match wins within each algorithm group). If no
2382 acl(s) match, the default mimicking action takes place.
2384 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2385 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2386 CONNECT request that carries a domain name. In all other cases (CONNECT
2387 to an IP address or an intercepted SSL connection), Squid cannot detect
2388 the domain mismatch at certificate generation time when
2389 bump-server-first is used.
2392 NAME: sslpassword_program
2395 LOC: Config.Program.ssl_password
2398 Specify a program used for entering SSL key passphrases
2399 when using encrypted SSL certificate keys. If not specified
2400 keys must either be unencrypted, or Squid started with the -N
2401 option to allow it to query interactively for the passphrase.
2403 The key file name is given as argument to the program allowing
2404 selection of the right password if you have multiple encrypted
2409 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2410 -----------------------------------------------------------------------------
2413 NAME: sslcrtd_program
2416 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2417 LOC: Ssl::TheConfig.ssl_crtd
2419 Specify the location and options of the executable for ssl_crtd process.
2420 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2421 For more information use:
2422 @DEFAULT_SSL_CRTD@ -h
2425 NAME: sslcrtd_children
2426 TYPE: HelperChildConfig
2428 DEFAULT: 32 startup=5 idle=1
2429 LOC: Ssl::TheConfig.ssl_crtdChildren
2431 The maximum number of processes spawn to service ssl server.
2432 The maximum this may be safely set to is 32.
2434 The startup= and idle= options allow some measure of skew in your
2439 Sets the minimum number of processes to spawn when Squid
2440 starts or reconfigures. When set to zero the first request will
2441 cause spawning of the first child process to handle it.
2443 Starting too few children temporary slows Squid under load while it
2444 tries to spawn enough additional processes to cope with traffic.
2448 Sets a minimum of how many processes Squid is to try and keep available
2449 at all times. When traffic begins to rise above what the existing
2450 processes can handle this many more will be spawned up to the maximum
2451 configured. A minimum setting of 1 is required.
2453 You must have at least one ssl_crtd process.
2456 NAME: sslcrtvalidator_program
2460 LOC: Ssl::TheConfig.ssl_crt_validator
2462 Specify the location and options of the executable for ssl_crt_validator
2464 sslcrtvalidator_program [ttl=n] [cache=n] path ...
2467 ttl=n TTL in seconds for cached results.The default is 60 secs
2468 cache=n limit the result cache size. The default value is 2048
2471 NAME: sslcrtvalidator_children
2472 TYPE: HelperChildConfig
2474 DEFAULT: 32 startup=5 idle=1 concurrency=1
2475 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2477 The maximum number of processes spawn to service ssl server.
2478 The maximum this may be safely set to is 32.
2480 The startup= and idle= options allow some measure of skew in your
2485 Sets the minimum number of processes to spawn when Squid
2486 starts or reconfigures. When set to zero the first request will
2487 cause spawning of the first child process to handle it.
2489 Starting too few children temporary slows Squid under load while it
2490 tries to spawn enough additional processes to cope with traffic.
2494 Sets a minimum of how many processes Squid is to try and keep available
2495 at all times. When traffic begins to rise above what the existing
2496 processes can handle this many more will be spawned up to the maximum
2497 configured. A minimum setting of 1 is required.
2501 The number of requests each certificate validator helper can handle in
2502 parallel. Defaults to 0 which indicates the certficate validator
2503 is a old-style single threaded redirector.
2505 When this directive is set to a value >= 1 then the protocol
2506 used to communicate with the helper is modified to include
2507 a request ID in front of the request/response. The request
2508 ID from the request must be echoed back with the response
2511 You must have at least one ssl_crt_validator process.
2515 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2516 -----------------------------------------------------------------------------
2524 To specify other caches in a hierarchy, use the format:
2526 cache_peer hostname type http-port icp-port [options]
2531 # hostname type port port options
2532 # -------------------- -------- ----- ----- -----------
2533 cache_peer parent.foo.net parent 3128 3130 default
2534 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2535 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2536 cache_peer example.com parent 80 0 default
2537 cache_peer cdn.example.com sibling 3128 0
2539 type: either 'parent', 'sibling', or 'multicast'.
2541 proxy-port: The port number where the peer accept HTTP requests.
2542 For other Squid proxies this is usually 3128
2543 For web servers this is usually 80
2545 icp-port: Used for querying neighbor caches about objects.
2546 Set to 0 if the peer does not support ICP or HTCP.
2547 See ICP and HTCP options below for additional details.
2550 ==== ICP OPTIONS ====
2552 You MUST also set icp_port and icp_access explicitly when using these options.
2553 The defaults will prevent peer traffic using ICP.
2556 no-query Disable ICP queries to this neighbor.
2559 Indicates the named peer is a member of a multicast group.
2560 ICP queries will not be sent directly to the peer, but ICP
2561 replies will be accepted from it.
2563 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2564 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2567 To only send ICP queries to this neighbor infrequently.
2568 This is used to keep the neighbor round trip time updated
2569 and is usually used in conjunction with weighted-round-robin.
2572 ==== HTCP OPTIONS ====
2574 You MUST also set htcp_port and htcp_access explicitly when using these options.
2575 The defaults will prevent peer traffic using HTCP.
2578 htcp Send HTCP, instead of ICP, queries to the neighbor.
2579 You probably also want to set the "icp-port" to 4827
2580 instead of 3130. This directive accepts a comma separated
2581 list of options described below.
2583 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2585 htcp=no-clr Send HTCP to the neighbor but without
2586 sending any CLR requests. This cannot be used with
2589 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2590 This cannot be used with no-clr.
2593 Send HTCP to the neighbor including CLRs but only when
2594 they do not result from PURGE requests.
2597 Forward any HTCP CLR requests this proxy receives to the peer.
2600 ==== PEER SELECTION METHODS ====
2602 The default peer selection method is ICP, with the first responding peer
2603 being used as source. These options can be used for better load balancing.
2606 default This is a parent cache which can be used as a "last-resort"
2607 if a peer cannot be located by any of the peer-selection methods.
2608 If specified more than once, only the first is used.
2610 round-robin Load-Balance parents which should be used in a round-robin
2611 fashion in the absence of any ICP queries.
2612 weight=N can be used to add bias.
2614 weighted-round-robin
2615 Load-Balance parents which should be used in a round-robin
2616 fashion with the frequency of each parent being based on the
2617 round trip time. Closer parents are used more often.
2618 Usually used for background-ping parents.
2619 weight=N can be used to add bias.
2621 carp Load-Balance parents which should be used as a CARP array.
2622 The requests will be distributed among the parents based on the
2623 CARP load balancing hash function based on their weight.
2625 userhash Load-balance parents based on the client proxy_auth or ident username.
2627 sourcehash Load-balance parents based on the client source IP.
2630 To be used only for cache peers of type "multicast".
2631 ALL members of this multicast group have "sibling"
2632 relationship with it, not "parent". This is to a multicast
2633 group when the requested object would be fetched only from
2634 a "parent" cache, anyway. It's useful, e.g., when
2635 configuring a pool of redundant Squid proxies, being
2636 members of the same multicast group.
2639 ==== PEER SELECTION OPTIONS ====
2641 weight=N use to affect the selection of a peer during any weighted
2642 peer-selection mechanisms.
2643 The weight must be an integer; default is 1,
2644 larger weights are favored more.
2645 This option does not affect parent selection if a peering
2646 protocol is not in use.
2648 basetime=N Specify a base amount to be subtracted from round trip
2650 It is subtracted before division by weight in calculating
2651 which parent to fectch from. If the rtt is less than the
2652 base time the rtt is set to a minimal value.
2654 ttl=N Specify a TTL to use when sending multicast ICP queries
2656 Only useful when sending to a multicast group.
2657 Because we don't accept ICP replies from random
2658 hosts, you must configure other group members as
2659 peers with the 'multicast-responder' option.
2661 no-delay To prevent access to this neighbor from influencing the
2664 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2665 enabled) for this host from the specified URL rather
2666 than the Squid default location.
2669 ==== CARP OPTIONS ====
2671 carp-key=key-specification
2672 use a different key than the full URL to hash against the peer.
2673 the key-specification is a comma-separated list of the keywords
2674 scheme, host, port, path, params
2675 Order is not important.
2677 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2679 originserver Causes this parent to be contacted as an origin server.
2680 Meant to be used in accelerator setups when the peer
2684 Set the Host header of requests forwarded to this peer.
2685 Useful in accelerator setups where the server (peer)
2686 expects a certain domain name but clients may request
2687 others. ie example.com or www.example.com
2689 no-digest Disable request of cache digests.
2692 Disables requesting ICMP RTT database (NetDB).
2695 ==== AUTHENTICATION OPTIONS ====
2698 If this is a personal/workgroup proxy and your parent
2699 requires proxy authentication.
2701 Note: The string can include URL escapes (i.e. %20 for
2702 spaces). This also means % must be written as %%.
2705 Send login details received from client to this peer.
2706 Both Proxy- and WWW-Authorization headers are passed
2707 without alteration to the peer.
2708 Authentication is not required by Squid for this to work.
2710 Note: This will pass any form of authentication but
2711 only Basic auth will work through a proxy unless the
2712 connection-auth options are also used.
2714 login=PASS Send login details received from client to this peer.
2715 Authentication is not required by this option.
2717 If there are no client-provided authentication headers
2718 to pass on, but username and password are available
2719 from an external ACL user= and password= result tags
2720 they may be sent instead.
2722 Note: To combine this with proxy_auth both proxies must
2723 share the same user database as HTTP only allows for
2724 a single login (one for proxy, one for origin server).
2725 Also be warned this will expose your users proxy
2726 password to the peer. USE WITH CAUTION
2729 Send the username to the upstream cache, but with a
2730 fixed password. This is meant to be used when the peer
2731 is in another administrative domain, but it is still
2732 needed to identify each user.
2733 The star can optionally be followed by some extra
2734 information which is added to the username. This can
2735 be used to identify this proxy to the peer, similar to
2736 the login=username:password option above.
2739 If this is a personal/workgroup proxy and your parent
2740 requires a secure proxy authentication.
2741 The first principal from the default keytab or defined by
2742 the environment variable KRB5_KTNAME will be used.
2744 WARNING: The connection may transmit requests from multiple
2745 clients. Negotiate often assumes end-to-end authentication
2746 and a single-client. Which is not strictly true here.
2748 login=NEGOTIATE:principal_name
2749 If this is a personal/workgroup proxy and your parent
2750 requires a secure proxy authentication.
2751 The principal principal_name from the default keytab or
2752 defined by the environment variable KRB5_KTNAME will be
2755 WARNING: The connection may transmit requests from multiple
2756 clients. Negotiate often assumes end-to-end authentication
2757 and a single-client. Which is not strictly true here.
2759 connection-auth=on|off
2760 Tell Squid that this peer does or not support Microsoft
2761 connection oriented authentication, and any such
2762 challenges received from there should be ignored.
2763 Default is auto to automatically determine the status
2767 ==== SSL / HTTPS / TLS OPTIONS ====
2769 ssl Encrypt connections to this peer with SSL/TLS.
2771 sslcert=/path/to/ssl/certificate
2772 A client SSL certificate to use when connecting to
2775 sslkey=/path/to/ssl/key
2776 The private SSL key corresponding to sslcert above.
2777 If 'sslkey' is not specified 'sslcert' is assumed to
2778 reference a combined file containing both the
2779 certificate and the key.
2781 sslversion=1|2|3|4|5|6
2782 The SSL version to use when connecting to this peer
2783 1 = automatic (default)
2790 sslcipher=... The list of valid SSL ciphers to use when connecting
2793 ssloptions=... Specify various SSL implementation options:
2795 NO_SSLv2 Disallow the use of SSLv2
2796 NO_SSLv3 Disallow the use of SSLv3
2797 NO_TLSv1 Disallow the use of TLSv1.0
2798 NO_TLSv1_1 Disallow the use of TLSv1.1
2799 NO_TLSv1_2 Disallow the use of TLSv1.2
2801 Always create a new key when using
2802 temporary/ephemeral DH key exchanges
2803 ALL Enable various bug workarounds
2804 suggested as "harmless" by OpenSSL
2805 Be warned that this reduces SSL/TLS
2806 strength to some attacks.
2808 See the OpenSSL SSL_CTX_set_options documentation for a
2811 sslcafile=... A file containing additional CA certificates to use
2812 when verifying the peer certificate.
2814 sslcapath=... A directory containing additional CA certificates to
2815 use when verifying the peer certificate.
2817 sslcrlfile=... A certificate revocation list file to use when
2818 verifying the peer certificate.
2820 sslflags=... Specify various flags modifying the SSL implementation:
2823 Accept certificates even if they fail to
2826 Don't use the default CA list built in
2829 Don't verify the peer certificate
2830 matches the server name
2832 ssldomain= The peer name as advertised in it's certificate.
2833 Used for verifying the correctness of the received peer
2834 certificate. If not specified the peer hostname will be
2838 Enable the "Front-End-Https: On" header needed when
2839 using Squid as a SSL frontend in front of Microsoft OWA.
2840 See MS KB document Q307347 for details on this header.
2841 If set to auto the header will only be added if the
2842 request is forwarded as a https:// URL.
2845 ==== GENERAL OPTIONS ====
2848 A peer-specific connect timeout.
2849 Also see the peer_connect_timeout directive.
2851 connect-fail-limit=N
2852 How many times connecting to a peer must fail before
2853 it is marked as down. Default is 10.
2855 allow-miss Disable Squid's use of only-if-cached when forwarding
2856 requests to siblings. This is primarily useful when
2857 icp_hit_stale is used by the sibling. To extensive use
2858 of this option may result in forwarding loops, and you
2859 should avoid having two-way peerings with this option.
2860 For example to deny peer usage on requests from peer
2861 by denying cache_peer_access if the source is a peer.
2863 max-conn=N Limit the amount of connections Squid may open to this
2866 name=xxx Unique name for the peer.
2867 Required if you have multiple peers on the same host
2868 but different ports.
2869 This name can be used in cache_peer_access and similar
2870 directives to dentify the peer.
2871 Can be used by outgoing access controls through the
2874 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2875 requests to this peer. Use normal address selection instead.
2877 proxy-only objects fetched from the peer will not be stored locally.
2881 NAME: cache_peer_domain cache_host_domain
2886 Use to limit the domains for which a neighbor cache will be
2889 cache_peer_domain cache-host domain [domain ...]
2890 cache_peer_domain cache-host !domain
2892 For example, specifying
2894 cache_peer_domain parent.foo.net .edu
2896 has the effect such that UDP query packets are sent to
2897 'bigserver' only when the requested object exists on a
2898 server in the .edu domain. Prefixing the domainname
2899 with '!' means the cache will be queried for objects
2902 NOTE: * Any number of domains may be given for a cache-host,
2903 either on the same or separate lines.
2904 * When multiple domains are given for a particular
2905 cache-host, the first matched domain is applied.
2906 * Cache hosts with no domain restrictions are queried
2908 * There are no defaults.
2909 * There is also a 'cache_peer_access' tag in the ACL
2913 NAME: cache_peer_access
2918 Similar to 'cache_peer_domain' but provides more flexibility by
2921 cache_peer_access cache-host allow|deny [!]aclname ...
2923 The syntax is identical to 'http_access' and the other lists of
2924 ACL elements. See the comments for 'http_access' below, or
2925 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2928 NAME: neighbor_type_domain
2929 TYPE: hostdomaintype
2933 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2935 Modifying the neighbor type for specific domains is now
2936 possible. You can treat some domains differently than the
2937 default neighbor type specified on the 'cache_peer' line.
2938 Normally it should only be necessary to list domains which
2939 should be treated differently because the default neighbor type
2940 applies for hostnames which do not match domains listed here.
2943 cache_peer cache.foo.org parent 3128 3130
2944 neighbor_type_domain cache.foo.org sibling .com .net
2945 neighbor_type_domain cache.foo.org sibling .au .de
2948 NAME: dead_peer_timeout
2952 LOC: Config.Timeout.deadPeer
2954 This controls how long Squid waits to declare a peer cache
2955 as "dead." If there are no ICP replies received in this
2956 amount of time, Squid will declare the peer dead and not
2957 expect to receive any further ICP replies. However, it
2958 continues to send ICP queries, and will mark the peer as
2959 alive upon receipt of the first subsequent ICP reply.
2961 This timeout also affects when Squid expects to receive ICP
2962 replies from peers. If more than 'dead_peer' seconds have
2963 passed since the last ICP reply was received, Squid will not
2964 expect to receive an ICP reply on the next query. Thus, if
2965 your time between requests is greater than this timeout, you
2966 will see a lot of requests sent DIRECT to origin servers
2967 instead of to your parents.
2970 NAME: forward_max_tries
2973 LOC: Config.forward_max_tries
2975 Controls how many different forward paths Squid will try
2976 before giving up. See also forward_timeout.
2978 NOTE: connect_retries (default: none) can make each of these
2979 possible forwarding paths be tried multiple times.
2982 NAME: hierarchy_stoplist
2985 LOC: Config.hierarchy_stoplist
2987 A list of words which, if found in a URL, cause the object to
2988 be handled directly by this cache. In other words, use this
2989 to not query neighbor caches for certain objects. You may
2990 list this option multiple times.
2993 hierarchy_stoplist cgi-bin ?
2995 Note: never_direct overrides this option.
2999 MEMORY CACHE OPTIONS
3000 -----------------------------------------------------------------------------
3007 LOC: Config.memMaxSize
3009 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
3010 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
3011 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3012 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3014 'cache_mem' specifies the ideal amount of memory to be used
3016 * In-Transit objects
3018 * Negative-Cached objects
3020 Data for these objects are stored in 4 KB blocks. This
3021 parameter specifies the ideal upper limit on the total size of
3022 4 KB blocks allocated. In-Transit objects take the highest
3025 In-transit objects have priority over the others. When
3026 additional space is needed for incoming data, negative-cached
3027 and hot objects will be released. In other words, the
3028 negative-cached and hot objects will fill up any unused space
3029 not needed for in-transit objects.
3031 If circumstances require, this limit will be exceeded.
3032 Specifically, if your incoming request rate requires more than
3033 'cache_mem' of memory to hold in-transit objects, Squid will
3034 exceed this limit to satisfy the new requests. When the load
3035 decreases, blocks will be freed until the high-water mark is
3036 reached. Thereafter, blocks will be used to store hot
3039 If shared memory caching is enabled, Squid does not use the shared
3040 cache space for in-transit objects, but they still consume as much
3041 local memory as they need. For more details about the shared memory
3042 cache, see memory_cache_shared.
3045 NAME: maximum_object_size_in_memory
3049 LOC: Config.Store.maxInMemObjSize
3051 Objects greater than this size will not be attempted to kept in
3052 the memory cache. This should be set high enough to keep objects
3053 accessed frequently in memory to improve performance whilst low
3054 enough to keep larger objects from hoarding cache_mem.
3057 NAME: memory_cache_shared
3060 LOC: Config.memShared
3062 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3064 Controls whether the memory cache is shared among SMP workers.
3066 The shared memory cache is meant to occupy cache_mem bytes and replace
3067 the non-shared memory cache, although some entities may still be
3068 cached locally by workers for now (e.g., internal and in-transit
3069 objects may be served from a local memory cache even if shared memory
3070 caching is enabled).
3072 By default, the memory cache is shared if and only if all of the
3073 following conditions are satisfied: Squid runs in SMP mode with
3074 multiple workers, cache_mem is positive, and Squid environment
3075 supports required IPC primitives (e.g., POSIX shared memory segments
3076 and GCC-style atomic operations).
3078 To avoid blocking locks, shared memory uses opportunistic algorithms
3079 that do not guarantee that every cachable entity that could have been
3080 shared among SMP workers will actually be shared.
3082 Currently, entities exceeding 32KB in size cannot be shared.
3085 NAME: memory_cache_mode
3090 Controls which objects to keep in the memory cache (cache_mem)
3092 always Keep most recently fetched objects in memory (default)
3094 disk Only disk cache hits are kept in memory, which means
3095 an object must first be cached on disk and then hit
3096 a second time before cached in memory.
3098 network Only objects fetched from network is kept in memory
3101 NAME: memory_replacement_policy
3103 LOC: Config.memPolicy
3106 The memory replacement policy parameter determines which
3107 objects are purged from memory when memory space is needed.
3109 See cache_replacement_policy for details.
3114 -----------------------------------------------------------------------------
3117 NAME: cache_replacement_policy
3119 LOC: Config.replPolicy
3122 The cache replacement policy parameter determines which
3123 objects are evicted (replaced) when disk space is needed.
3125 lru : Squid's original list based LRU policy
3126 heap GDSF : Greedy-Dual Size Frequency
3127 heap LFUDA: Least Frequently Used with Dynamic Aging
3128 heap LRU : LRU policy implemented using a heap
3130 Applies to any cache_dir lines listed below this.
3132 The LRU policies keeps recently referenced objects.
3134 The heap GDSF policy optimizes object hit rate by keeping smaller
3135 popular objects in cache so it has a better chance of getting a
3136 hit. It achieves a lower byte hit rate than LFUDA though since
3137 it evicts larger (possibly popular) objects.
3139 The heap LFUDA policy keeps popular objects in cache regardless of
3140 their size and thus optimizes byte hit rate at the expense of
3141 hit rate since one large, popular object will prevent many
3142 smaller, slightly less popular objects from being cached.
3144 Both policies utilize a dynamic aging mechanism that prevents
3145 cache pollution that can otherwise occur with frequency-based
3146 replacement policies.
3148 NOTE: if using the LFUDA replacement policy you should increase
3149 the value of maximum_object_size above its default of 4096 KB to
3150 to maximize the potential byte hit rate improvement of LFUDA.
3152 For more information about the GDSF and LFUDA cache replacement
3153 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3154 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3160 LOC: Config.cacheSwap
3164 cache_dir Type Directory-Name Fs-specific-data [options]
3166 You can specify multiple cache_dir lines to spread the
3167 cache among different disk partitions.
3169 Type specifies the kind of storage system to use. Only "ufs"
3170 is built by default. To enable any of the other storage systems
3171 see the --enable-storeio configure option.
3173 'Directory' is a top-level directory where cache swap
3174 files will be stored. If you want to use an entire disk
3175 for caching, this can be the mount-point directory.
3176 The directory must exist and be writable by the Squid
3177 process. Squid will NOT create this directory for you.
3179 In SMP configurations, cache_dir must not precede the workers option
3180 and should use configuration macros or conditionals to give each
3181 worker interested in disk caching a dedicated cache directory.
3185 "ufs" is the old well-known Squid storage format that has always
3188 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3190 'Mbytes' is the amount of disk space (MB) to use under this
3191 directory. The default is 100 MB. Change this to suit your
3192 configuration. Do NOT put the size of your disk drive here.
3193 Instead, if you want Squid to use the entire disk drive,
3194 subtract 20% and use that value.
3196 'L1' is the number of first-level subdirectories which
3197 will be created under the 'Directory'. The default is 16.
3199 'L2' is the number of second-level subdirectories which
3200 will be created under each first-level directory. The default
3203 The aufs store type:
3205 "aufs" uses the same storage format as "ufs", utilizing
3206 POSIX-threads to avoid blocking the main Squid process on
3207 disk-I/O. This was formerly known in Squid as async-io.
3209 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3211 see argument descriptions under ufs above
3213 The diskd store type:
3215 "diskd" uses the same storage format as "ufs", utilizing a
3216 separate process to avoid blocking the main Squid process on
3219 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3221 see argument descriptions under ufs above
3223 Q1 specifies the number of unacknowledged I/O requests when Squid
3224 stops opening new files. If this many messages are in the queues,
3225 Squid won't open new files. Default is 64
3227 Q2 specifies the number of unacknowledged messages when Squid
3228 starts blocking. If this many messages are in the queues,
3229 Squid blocks until it receives some replies. Default is 72
3231 When Q1 < Q2 (the default), the cache directory is optimized
3232 for lower response time at the expense of a decrease in hit
3233 ratio. If Q1 > Q2, the cache directory is optimized for
3234 higher hit ratio at the expense of an increase in response
3237 The rock store type:
3239 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3241 The Rock Store type is a database-style storage. All cached
3242 entries are stored in a "database" file, using fixed-size slots,
3243 one entry per slot. The database size is specified in MB. The
3244 slot size is specified in bytes using the max-size option. See
3245 below for more info on the max-size option.
3247 swap-timeout=msec: Squid will not start writing a miss to or
3248 reading a hit from disk if it estimates that the swap operation
3249 will take more than the specified number of milliseconds. By
3250 default and when set to zero, disables the disk I/O time limit
3251 enforcement. Ignored when using blocking I/O module because
3252 blocking synchronous I/O does not allow Squid to estimate the
3253 expected swap wait time.
3255 max-swap-rate=swaps/sec: Artificially limits disk access using
3256 the specified I/O rate limit. Swap out requests that
3257 would cause the average I/O rate to exceed the limit are
3258 delayed. Individual swap in requests (i.e., hits or reads) are
3259 not delayed, but they do contribute to measured swap rate and
3260 since they are placed in the same FIFO queue as swap out
3261 requests, they may wait longer if max-swap-rate is smaller.
3262 This is necessary on file systems that buffer "too
3263 many" writes and then start blocking Squid and other processes
3264 while committing those writes to disk. Usually used together
3265 with swap-timeout to avoid excessive delays and queue overflows
3266 when disk demand exceeds available disk "bandwidth". By default
3267 and when set to zero, disables the disk I/O rate limit
3268 enforcement. Currently supported by IpcIo module only.
3271 The coss store type:
3273 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3274 production use and has thus been removed from this release.
3275 We hope that it can be made usable again soon.
3277 block-size=n defines the "block size" for COSS cache_dir's.
3278 Squid uses file numbers as block numbers. Since file numbers
3279 are limited to 24 bits, the block size determines the maximum
3280 size of the COSS partition. The default is 512 bytes, which
3281 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3282 you should not change the coss block size after Squid
3283 has written some objects to the cache_dir.
3285 The coss file store has changed from 2.5. Now it uses a file
3286 called 'stripe' in the directory names in the config - and
3287 this will be created by squid -z.
3291 no-store, no new objects should be stored to this cache_dir
3293 min-size=n, refers to the min object size in bytes this cache_dir
3294 will accept. It's used to restrict a cache_dir to only store
3295 large objects (e.g. aufs) while other storedirs are optimized
3296 for smaller objects (e.g. COSS). Defaults to 0.
3298 max-size=n, refers to the max object size in bytes this cache_dir
3299 supports. It is used to select the cache_dir to store the object.
3300 Note: To make optimal use of the max-size limits you should order
3301 the cache_dir lines with the smallest max-size value first and the
3302 ones with no max-size specification last.
3304 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3305 which can be changed with the --with-coss-membuf-size=N configure
3309 # Uncomment and adjust the following to add a disk cache directory.
3310 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3314 NAME: store_dir_select_algorithm
3316 LOC: Config.store_dir_select_algorithm
3319 Set this to 'round-robin' as an alternative.
3322 NAME: max_open_disk_fds
3324 LOC: Config.max_open_disk_fds
3327 To avoid having disk as the I/O bottleneck Squid can optionally
3328 bypass the on-disk cache if more than this amount of disk file
3329 descriptors are open.
3331 A value of 0 indicates no limit.
3334 NAME: minimum_object_size
3338 LOC: Config.Store.minObjectSize
3340 Objects smaller than this size will NOT be saved on disk. The
3341 value is specified in kilobytes, and the default is 0 KB, which
3342 means there is no minimum.
3345 NAME: maximum_object_size
3349 LOC: Config.Store.maxObjectSize
3351 Objects larger than this size will NOT be saved on disk. The
3352 value is specified in kilobytes, and the default is 4MB. If
3353 you wish to get a high BYTES hit ratio, you should probably
3354 increase this (one 32 MB object hit counts for 3200 10KB
3355 hits). If you wish to increase speed more than your want to
3356 save bandwidth you should leave this low.
3358 NOTE: if using the LFUDA replacement policy you should increase
3359 this value to maximize the byte hit rate improvement of LFUDA!
3360 See replacement_policy below for a discussion of this policy.
3363 NAME: cache_swap_low
3364 COMMENT: (percent, 0-100)
3367 LOC: Config.Swap.lowWaterMark
3370 NAME: cache_swap_high
3371 COMMENT: (percent, 0-100)
3374 LOC: Config.Swap.highWaterMark
3377 The low- and high-water marks for cache object replacement.
3378 Replacement begins when the swap (disk) usage is above the
3379 low-water mark and attempts to maintain utilization near the
3380 low-water mark. As swap utilization gets close to high-water
3381 mark object eviction becomes more aggressive. If utilization is
3382 close to the low-water mark less replacement is done each time.
3384 Defaults are 90% and 95%. If you have a large cache, 5% could be
3385 hundreds of MB. If this is the case you may wish to set these
3386 numbers closer together.
3391 -----------------------------------------------------------------------------
3401 logformat <name> <format specification>
3403 Defines an access log format.
3405 The <format specification> is a string with embedded % format codes
3407 % format codes all follow the same basic structure where all but
3408 the formatcode is optional. Output strings are automatically escaped
3409 as required according to their context and the output format
3410 modifiers are usually not needed, but can be specified if an explicit
3411 output format is desired.
3413 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3415 " output in quoted string format
3416 [ output in squid text log format as used by log_mime_hdrs
3417 # output in URL quoted format
3422 width minimum and/or maximum field width:
3423 [width_min][.width_max]
3424 When minimum starts with 0, the field is zero-padded.
3425 String values exceeding maximum width are truncated.
3427 {arg} argument such as header name etc
3431 % a literal % character
3432 sn Unique sequence number per log line entry
3433 err_code The ID of an error response served by Squid or
3434 a similar internal error identifier.
3435 err_detail Additional err_code-dependent error information.
3436 note The meta header specified by the argument. Also
3437 logs the adaptation meta headers set by the
3438 adaptation_meta configuration parameter.
3439 If no argument given all meta headers logged.
3441 Connection related format codes:
3443 >a Client source IP address
3445 >p Client source port
3446 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3447 >la Local IP address the client connected to
3448 >lp Local port number the client connected to
3450 la Local listening IP address the client connection was connected to.
3451 lp Local listening port number the client connection was connected to.
3453 <a Server IP address of the last server or peer connection
3454 <A Server FQDN or peer name
3455 <p Server port number of the last server or peer connection
3456 <la Local IP address of the last server or peer connection
3457 <lp Local port number of the last server or peer connection
3459 Time related format codes:
3461 ts Seconds since epoch
3462 tu subsecond time (milliseconds)
3463 tl Local time. Optional strftime format argument
3464 default %d/%b/%Y:%H:%M:%S %z
3465 tg GMT time. Optional strftime format argument
3466 default %d/%b/%Y:%H:%M:%S %z
3467 tr Response time (milliseconds)
3468 dt Total time spent making DNS lookups (milliseconds)
3470 Access Control related format codes:
3472 et Tag returned by external acl
3473 ea Log string returned by external acl
3474 un User name (any available)
3475 ul User name from authentication
3476 ue User name from external acl helper
3477 ui User name from ident
3478 us User name from SSL
3480 HTTP related format codes:
3482 [http::]>h Original request header. Optional header name argument
3483 on the format header[:[separator]element]
3484 [http::]>ha The HTTP request headers after adaptation and redirection.
3485 Optional header name argument as for >h
3486 [http::]<h Reply header. Optional header name argument
3488 [http::]>Hs HTTP status code sent to the client
3489 [http::]<Hs HTTP status code received from the next hop
3490 [http::]<bs Number of HTTP-equivalent message body bytes
3491 received from the next hop, excluding chunked
3492 transfer encoding and control messages.
3493 Generated FTP/Gopher listings are treated as
3495 [http::]mt MIME content type
3496 [http::]rm Request method (GET/POST etc)
3497 [http::]>rm Request method from client
3498 [http::]<rm Request method sent to server or peer
3499 [http::]ru Request URL from client (historic, filtered for logging)
3500 [http::]>ru Request URL from client
3501 [http::]<ru Request URL sent to server or peer
3502 [http::]rp Request URL-Path excluding hostname
3503 [http::]>rp Request URL-Path excluding hostname from client
3504 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3505 [http::]rv Request protocol version
3506 [http::]>rv Request protocol version from client
3507 [http::]<rv Request protocol version sent to server or peer
3508 [http::]<st Sent reply size including HTTP headers
3509 [http::]>st Received request size including HTTP headers. In the
3510 case of chunked requests the chunked encoding metadata
3512 [http::]>sh Received HTTP request headers size
3513 [http::]<sh Sent HTTP reply headers size
3514 [http::]st Request+Reply size including HTTP headers
3515 [http::]<sH Reply high offset sent
3516 [http::]<sS Upstream object size
3517 [http::]<pt Peer response time in milliseconds. The timer starts
3518 when the last request byte is sent to the next hop
3519 and stops when the last response byte is received.
3520 [http::]<tt Total server-side time in milliseconds. The timer
3521 starts with the first connect request (or write I/O)
3522 sent to the first selected peer. The timer stops
3523 with the last I/O with the last peer.
3525 Squid handling related format codes:
3527 Ss Squid request status (TCP_MISS etc)
3528 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3530 SSL-related format codes:
3532 ssl::bump_mode SslBump decision for the transaction:
3534 For CONNECT requests that initiated bumping of
3535 a connection and for any request received on
3536 an already bumped connection, Squid logs the
3537 corresponding SslBump mode ("server-first" or
3538 "client-first"). See the ssl_bump option for
3539 more information about these modes.
3541 A "none" token is logged for requests that
3542 triggered "ssl_bump" ACL evaluation matching
3543 either a "none" rule or no rules at all.
3545 In all other cases, a single dash ("-") is
3548 If ICAP is enabled, the following code becomes available (as
3549 well as ICAP log codes documented with the icap_log option):
3551 icap::tt Total ICAP processing time for the HTTP
3552 transaction. The timer ticks when ICAP
3553 ACLs are checked and when ICAP
3554 transaction is in progress.
3556 If adaptation is enabled the following three codes become available:
3558 adapt::<last_h The header of the last ICAP response or
3559 meta-information from the last eCAP
3560 transaction related to the HTTP transaction.
3561 Like <h, accepts an optional header name
3564 adapt::sum_trs Summed adaptation transaction response
3565 times recorded as a comma-separated list in
3566 the order of transaction start time. Each time
3567 value is recorded as an integer number,
3568 representing response time of one or more
3569 adaptation (ICAP or eCAP) transaction in
3570 milliseconds. When a failed transaction is
3571 being retried or repeated, its time is not
3572 logged individually but added to the
3573 replacement (next) transaction. See also:
3576 adapt::all_trs All adaptation transaction response times.
3577 Same as adaptation_strs but response times of
3578 individual transactions are never added
3579 together. Instead, all transaction response
3580 times are recorded individually.
3582 You can prefix adapt::*_trs format codes with adaptation
3583 service name in curly braces to record response time(s) specific
3584 to that service. For example: %{my_service}adapt::sum_trs
3586 If SSL is enabled, the following formating codes become available:
3588 %ssl::>cert_subject The Subject field of the received client
3589 SSL certificate or a dash ('-') if Squid has
3590 received an invalid/malformed certificate or
3591 no certificate at all. Consider encoding the
3592 logged value because Subject often has spaces.
3594 %ssl::>cert_issuer The Issuer field of the received client
3595 SSL certificate or a dash ('-') if Squid has
3596 received an invalid/malformed certificate or
3597 no certificate at all. Consider encoding the
3598 logged value because Issuer often has spaces.
3600 The default formats available (which do not need re-defining) are:
3602 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3603 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3604 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3605 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3606 logformat useragent %>a [%tl] "%{User-Agent}>h"
3608 NOTE: When the log_mime_hdrs directive is set to ON.
3609 The squid, common and combined formats have a safely encoded copy
3610 of the mime headers appended to each line within a pair of brackets.
3612 NOTE: The common and combined formats are not quite true to the Apache definition.
3613 The logs from Squid contain an extra status and hierarchy code appended.
3617 NAME: access_log cache_access_log
3619 LOC: Config.Log.accesslogs
3620 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3622 These files log client request activities. Has a line every HTTP or
3623 ICP request. The format is:
3624 access_log <module>:<place> [<logformat name> [acl acl ...]]
3625 access_log none [acl acl ...]]
3627 Will log to the specified module:place using the specified format (which
3628 must be defined in a logformat directive) those entries which match
3629 ALL the acl's specified (which must be defined in acl clauses).
3630 If no acl is specified, all requests will be logged to this destination.
3632 ===== Modules Currently available =====
3634 none Do not log any requests matching these ACL.
3635 Do not specify Place or logformat name.
3637 stdio Write each log line to disk immediately at the completion of
3639 Place: the filename and path to be written.
3641 daemon Very similar to stdio. But instead of writing to disk the log
3642 line is passed to a daemon helper for asychronous handling instead.
3643 Place: varies depending on the daemon.
3645 log_file_daemon Place: the file name and path to be written.
3647 syslog To log each request via syslog facility.
3648 Place: The syslog facility and priority level for these entries.
3649 Place Format: facility.priority
3651 where facility could be any of:
3652 authpriv, daemon, local0 ... local7 or user.
3654 And priority could be any of:
3655 err, warning, notice, info, debug.
3657 udp To send each log line as text data to a UDP receiver.
3658 Place: The destination host name or IP and port.
3659 Place Format: //host:port
3661 tcp To send each log line as text data to a TCP receiver.
3662 Place: The destination host name or IP and port.
3663 Place Format: //host:port
3666 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3672 LOC: Config.Log.icaplogs
3675 ICAP log files record ICAP transaction summaries, one line per
3678 The icap_log option format is:
3679 icap_log <filepath> [<logformat name> [acl acl ...]]
3680 icap_log none [acl acl ...]]
3682 Please see access_log option documentation for details. The two
3683 kinds of logs share the overall configuration approach and many
3686 ICAP processing of a single HTTP message or transaction may
3687 require multiple ICAP transactions. In such cases, multiple
3688 ICAP transaction log lines will correspond to a single access
3691 ICAP log uses logformat codes that make sense for an ICAP
3692 transaction. Header-related codes are applied to the HTTP header
3693 embedded in an ICAP server response, with the following caveats:
3694 For REQMOD, there is no HTTP response header unless the ICAP
3695 server performed request satisfaction. For RESPMOD, the HTTP
3696 request header is the header sent to the ICAP server. For
3697 OPTIONS, there are no HTTP headers.
3699 The following format codes are also available for ICAP logs:
3701 icap::<A ICAP server IP address. Similar to <A.
3703 icap::<service_name ICAP service name from the icap_service
3704 option in Squid configuration file.
3706 icap::ru ICAP Request-URI. Similar to ru.
3708 icap::rm ICAP request method (REQMOD, RESPMOD, or
3709 OPTIONS). Similar to existing rm.
3711 icap::>st Bytes sent to the ICAP server (TCP payload
3712 only; i.e., what Squid writes to the socket).
3714 icap::<st Bytes received from the ICAP server (TCP
3715 payload only; i.e., what Squid reads from
3718 icap::<bs Number of message body bytes received from the
3719 ICAP server. ICAP message body, if any, usually
3720 includes encapsulated HTTP message headers and
3721 possibly encapsulated HTTP message body. The
3722 HTTP body part is dechunked before its size is
3725 icap::tr Transaction response time (in
3726 milliseconds). The timer starts when
3727 the ICAP transaction is created and
3728 stops when the transaction is completed.
3731 icap::tio Transaction I/O time (in milliseconds). The
3732 timer starts when the first ICAP request
3733 byte is scheduled for sending. The timers
3734 stops when the last byte of the ICAP response
3737 icap::to Transaction outcome: ICAP_ERR* for all
3738 transaction errors, ICAP_OPT for OPTION
3739 transactions, ICAP_ECHO for 204
3740 responses, ICAP_MOD for message
3741 modification, and ICAP_SAT for request
3742 satisfaction. Similar to Ss.
3744 icap::Hs ICAP response status code. Similar to Hs.
3746 icap::>h ICAP request header(s). Similar to >h.
3748 icap::<h ICAP response header(s). Similar to <h.
3750 The default ICAP log format, which can be used without an explicit
3751 definition, is called icap_squid:
3753 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3755 See also: logformat, log_icap, and %adapt::<last_h
3758 NAME: logfile_daemon
3760 DEFAULT: @DEFAULT_LOGFILED@
3761 LOC: Log::TheConfig.logfile_daemon
3763 Specify the path to the logfile-writing daemon. This daemon is
3764 used to write the access and store logs, if configured.
3766 Squid sends a number of commands to the log daemon:
3767 L<data>\n - logfile data
3772 r<n>\n - set rotate count to <n>
3773 b<n>\n - 1 = buffer output, 0 = don't buffer output
3775 No responses is expected.
3780 LOC: Config.accessList.log
3782 COMMENT: allow|deny acl acl...
3784 This options allows you to control which requests gets logged
3785 to access.log (see access_log directive). Requests denied for
3786 logging will also not be accounted for in performance counters.
3788 This clause only supports fast acl types.
3789 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3795 LOC: Config.accessList.icap
3798 This options allows you to control which requests get logged
3799 to icap.log. See the icap_log directive for ICAP log details.
3802 NAME: cache_store_log
3805 LOC: Config.Log.store
3807 Logs the activities of the storage manager. Shows which
3808 objects are ejected from the cache, and which objects are
3809 saved and for how long.
3810 There are not really utilities to analyze this data, so you can safely
3811 disable it (the default).
3813 Store log uses modular logging outputs. See access_log for the list
3814 of modules supported.
3817 cache_store_log stdio:@DEFAULT_STORE_LOG@
3818 cache_store_log daemon:@DEFAULT_STORE_LOG@
3821 NAME: cache_swap_state cache_swap_log
3823 LOC: Config.Log.swap
3826 Location for the cache "swap.state" file. This index file holds
3827 the metadata of objects saved on disk. It is used to rebuild
3828 the cache during startup. Normally this file resides in each
3829 'cache_dir' directory, but you may specify an alternate
3830 pathname here. Note you must give a full filename, not just
3831 a directory. Since this is the index for the whole object
3832 list you CANNOT periodically rotate it!
3834 If %s can be used in the file name it will be replaced with a
3835 a representation of the cache_dir name where each / is replaced
3836 with '.'. This is needed to allow adding/removing cache_dir
3837 lines when cache_swap_log is being used.
3839 If have more than one 'cache_dir', and %s is not used in the name
3840 these swap logs will have names such as:
3846 The numbered extension (which is added automatically)
3847 corresponds to the order of the 'cache_dir' lines in this
3848 configuration file. If you change the order of the 'cache_dir'
3849 lines in this file, these index files will NOT correspond to
3850 the correct 'cache_dir' entry (unless you manually rename
3851 them). We recommend you do NOT use this option. It is
3852 better to keep these index files in each 'cache_dir' directory.
3855 NAME: logfile_rotate
3858 LOC: Config.Log.rotateNumber
3860 Specifies the number of logfile rotations to make when you
3861 type 'squid -k rotate'. The default is 10, which will rotate
3862 with extensions 0 through 9. Setting logfile_rotate to 0 will
3863 disable the file name rotation, but the logfiles are still closed
3864 and re-opened. This will enable you to rename the logfiles
3865 yourself just before sending the rotate signal.
3867 Note, the 'squid -k rotate' command normally sends a USR1
3868 signal to the running squid process. In certain situations
3869 (e.g. on Linux with Async I/O), USR1 is used for other
3870 purposes, so -k rotate uses another signal. It is best to get
3871 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3874 Note, from Squid-3.1 this option has no effect on the cache.log,
3875 that log can be rotated separately by using debug_options
3878 NAME: emulate_httpd_log
3881 Replace this with an access_log directive using the format 'common' or 'combined'.
3884 NAME: log_ip_on_direct
3887 Remove this option from your config. To log server or peer names use %<A in the log format.
3892 DEFAULT: @DEFAULT_MIME_TABLE@
3893 LOC: Config.mimeTablePathname
3895 Pathname to Squid's MIME table. You shouldn't need to change
3896 this, but the default file contains examples and formatting
3897 information if you do.
3903 LOC: Config.onoff.log_mime_hdrs
3906 The Cache can record both the request and the response MIME
3907 headers for each HTTP transaction. The headers are encoded
3908 safely and will appear as two bracketed fields at the end of
3909 the access log (for either the native or httpd-emulated log
3910 formats). To enable this logging set log_mime_hdrs to 'on'.
3916 Replace this with an access_log directive using the format 'useragent'.
3919 NAME: referer_log referrer_log
3922 Replace this with an access_log directive using the format 'referrer'.
3927 DEFAULT: @DEFAULT_PID_FILE@
3928 LOC: Config.pidFilename
3930 A filename to write the process-id to. To disable, enter "none".
3936 Remove this option from your config. To log FQDN use %>A in the log format.
3939 NAME: client_netmask
3941 LOC: Config.Addrs.client_netmask
3944 A netmask for client addresses in logfiles and cachemgr output.
3945 Change this to protect the privacy of your cache clients.
3946 A netmask of 255.255.255.0 will log all IP's in that range with
3947 the last digit set to '0'.
3953 Use a regular access.log with ACL limiting it to MISS events.
3956 NAME: strip_query_terms
3958 LOC: Config.onoff.strip_query_terms
3961 By default, Squid strips query terms from requested URLs before
3962 logging. This protects your user's privacy.
3969 LOC: Config.onoff.buffered_logs
3971 cache.log log file is written with stdio functions, and as such
3972 it can be buffered or unbuffered. By default it will be unbuffered.
3973 Buffering it can speed up the writing slightly (though you are
3974 unlikely to need to worry unless you run with tons of debugging
3975 enabled in which case performance will suffer badly anyway..).
3978 NAME: netdb_filename
3980 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3981 LOC: Config.netdbFilename
3984 A filename where Squid stores it's netdb state between restarts.
3985 To disable, enter "none".
3989 OPTIONS FOR TROUBLESHOOTING
3990 -----------------------------------------------------------------------------
3995 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3996 LOC: Debug::cache_log
3998 Cache logging file. This is where general information about
3999 your cache's behavior goes. You can increase the amount of data
4000 logged to this file and how often its rotated with "debug_options"
4006 LOC: Debug::debugOptions
4008 Logging options are set as section,level where each source file
4009 is assigned a unique section. Lower levels result in less
4010 output, Full debugging (level 9) can result in a very large
4011 log file, so be careful.
4013 The magic word "ALL" sets debugging levels for all sections.
4014 We recommend normally running with "ALL,1".
4016 The rotate=N option can be used to keep more or less of these logs
4017 than would otherwise be kept by logfile_rotate.
4018 For most uses a single log should be enough to monitor current
4019 events affecting Squid.
4024 LOC: Config.coredump_dir
4025 DEFAULT_IF_NONE: none
4027 By default Squid leaves core files in the directory from where
4028 it was started. If you set 'coredump_dir' to a directory
4029 that exists, Squid will chdir() to that directory at startup
4030 and coredump files will be left there.
4034 # Leave coredumps in the first cache dir
4035 coredump_dir @DEFAULT_SWAP_DIR@
4041 OPTIONS FOR FTP GATEWAYING
4042 -----------------------------------------------------------------------------
4048 LOC: Config.Ftp.anon_user
4050 If you want the anonymous login password to be more informative
4051 (and enable the use of picky ftp servers), set this to something
4052 reasonable for your domain, like wwwuser@somewhere.net
4054 The reason why this is domainless by default is the
4055 request can be made on the behalf of a user in any domain,
4056 depending on how the cache is used.
4057 Some ftp server also validate the email address is valid
4058 (for example perl.com).
4064 LOC: Config.Ftp.passive
4066 If your firewall does not allow Squid to use passive
4067 connections, turn off this option.
4069 Use of ftp_epsv_all option requires this to be ON.
4075 LOC: Config.Ftp.epsv_all
4077 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4079 NATs may be able to put the connection on a "fast path" through the
4080 translator, as the EPRT command will never be used and therefore,
4081 translation of the data portion of the segments will never be needed.
4083 When a client only expects to do two-way FTP transfers this may be
4085 If squid finds that it must do a three-way FTP transfer after issuing
4086 an EPSV ALL command, the FTP session will fail.
4088 If you have any doubts about this option do not use it.
4089 Squid will nicely attempt all other connection methods.
4091 Requires ftp_passive to be ON (default) for any effect.
4097 LOC: Config.Ftp.epsv
4099 FTP Protocol extensions permit the use of a special "EPSV" command.
4101 NATs may be able to put the connection on a "fast path" through the
4102 translator using EPSV, as the EPRT command will never be used
4103 and therefore, translation of the data portion of the segments
4104 will never be needed.
4106 Turning this OFF will prevent EPSV being attempted.
4107 WARNING: Doing so will convert Squid back to the old behavior with all
4108 the related problems with external NAT devices/layers.
4110 Requires ftp_passive to be ON (default) for any effect.
4116 LOC: Config.Ftp.eprt
4118 FTP Protocol extensions permit the use of a special "EPRT" command.
4120 This extension provides a protocol neutral alternative to the
4121 IPv4-only PORT command. When supported it enables active FTP data
4122 channels over IPv6 and efficient NAT handling.
4124 Turning this OFF will prevent EPRT being attempted and will skip
4125 straight to using PORT for IPv4 servers.
4127 Some devices are known to not handle this extension correctly and
4128 may result in crashes. Devices which suport EPRT enough to fail
4129 cleanly will result in Squid attempting PORT anyway. This directive
4130 should only be disabled when EPRT results in device failures.
4132 WARNING: Doing so will convert Squid back to the old behavior with all
4133 the related problems with external NAT devices/layers and IPv4-only FTP.
4136 NAME: ftp_sanitycheck
4139 LOC: Config.Ftp.sanitycheck
4141 For security and data integrity reasons Squid by default performs
4142 sanity checks of the addresses of FTP data connections ensure the
4143 data connection is to the requested server. If you need to allow
4144 FTP connections to servers using another IP address for the data
4145 connection turn this off.
4148 NAME: ftp_telnet_protocol
4151 LOC: Config.Ftp.telnet
4153 The FTP protocol is officially defined to use the telnet protocol
4154 as transport channel for the control connection. However, many
4155 implementations are broken and does not respect this aspect of
4158 If you have trouble accessing files with ASCII code 255 in the
4159 path or similar problems involving this ASCII code you can
4160 try setting this directive to off. If that helps, report to the
4161 operator of the FTP server in question that their FTP server
4162 is broken and does not follow the FTP standard.
4166 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4167 -----------------------------------------------------------------------------
4172 DEFAULT: @DEFAULT_DISKD@
4173 LOC: Config.Program.diskd
4175 Specify the location of the diskd executable.
4176 Note this is only useful if you have compiled in
4177 diskd as one of the store io modules.
4180 NAME: unlinkd_program
4183 DEFAULT: @DEFAULT_UNLINKD@
4184 LOC: Config.Program.unlinkd
4186 Specify the location of the executable for file deletion process.
4189 NAME: pinger_program
4191 DEFAULT: @DEFAULT_PINGER@
4192 LOC: Config.pinger.program
4195 Specify the location of the executable for the pinger process.
4201 LOC: Config.pinger.enable
4204 Control whether the pinger is active at run-time.
4205 Enables turning ICMP pinger on and off with a simple
4206 squid -k reconfigure.
4211 OPTIONS FOR URL REWRITING
4212 -----------------------------------------------------------------------------
4215 NAME: url_rewrite_program redirect_program
4217 LOC: Config.Program.redirect
4220 Specify the location of the executable URL rewriter to use.
4221 Since they can perform almost any function there isn't one included.
4223 For each requested URL, the rewriter will receive on line with the format
4225 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4228 After processing the request the helper must reply using the following format:
4230 [channel-ID <SP>] result [<SP> kv-pairs]
4232 The result code can be:
4234 OK status=30N url="..."
4235 Redirect the URL to the one supplied in 'url='.
4236 'status=' is optional and contains the status code to send
4237 the client in Squids HTTP response. It must be one of the
4238 HTTP redirect status codes: 301, 302, 303, 307, 308.
4239 When no status is given Squid will use 302.
4241 OK rewrite-url="..."
4242 Rewrite the URL to the one supplied in 'rewrite-url='.
4243 The new URL is fetched directly by Squid and returned to
4244 the client as the response to its request.
4247 Do not change the URL.
4250 An internal error occurred in the helper, preventing
4251 a result being identified.
4254 In the future, the interface protocol will be extended with
4255 key=value pairs ("kv-pairs" shown above). Helper programs
4256 should be prepared to receive and possibly ignore additional
4257 whitespace-separated tokens on each input line.
4259 When using the concurrency= option the protocol is changed by
4260 introducing a query channel tag in front of the request/response.
4261 The query channel tag is a number between 0 and concurrency-1.
4262 This value must be echoed back unchanged to Squid as the first part
4263 of the response relating to its request.
4265 WARNING: URL re-writing ability should be avoided whenever possible.
4266 Use the URL redirect form of response instead.
4268 Re-write creates a difference in the state held by the client
4269 and server. Possibly causing confusion when the server response
4270 contains snippets of its view state. Embeded URLs, response
4271 and content Location headers, etc. are not re-written by this
4274 By default, a URL rewriter is not used.
4277 NAME: url_rewrite_children redirect_children
4278 TYPE: HelperChildConfig
4279 DEFAULT: 20 startup=0 idle=1 concurrency=0
4280 LOC: Config.redirectChildren
4282 The maximum number of redirector processes to spawn. If you limit
4283 it too few Squid will have to wait for them to process a backlog of
4284 URLs, slowing it down. If you allow too many they will use RAM
4285 and other system resources noticably.
4287 The startup= and idle= options allow some measure of skew in your
4292 Sets a minimum of how many processes are to be spawned when Squid
4293 starts or reconfigures. When set to zero the first request will
4294 cause spawning of the first child process to handle it.
4296 Starting too few will cause an initial slowdown in traffic as Squid
4297 attempts to simultaneously spawn enough processes to cope.
4301 Sets a minimum of how many processes Squid is to try and keep available
4302 at all times. When traffic begins to rise above what the existing
4303 processes can handle this many more will be spawned up to the maximum
4304 configured. A minimum setting of 1 is required.
4308 The number of requests each redirector helper can handle in
4309 parallel. Defaults to 0 which indicates the redirector
4310 is a old-style single threaded redirector.
4312 When this directive is set to a value >= 1 then the protocol
4313 used to communicate with the helper is modified to include
4314 an ID in front of the request/response. The ID from the request
4315 must be echoed back with the response to that request.
4318 NAME: url_rewrite_host_header redirect_rewrites_host_header
4321 LOC: Config.onoff.redir_rewrites_host
4323 To preserve same-origin security policies in browsers and
4324 prevent Host: header forgery by redirectors Squid rewrites
4325 any Host: header in redirected requests.
4327 If you are running an accelerator this may not be a wanted
4328 effect of a redirector. This directive enables you disable
4329 Host: alteration in reverse-proxy traffic.
4331 WARNING: Entries are cached on the result of the URL rewriting
4332 process, so be careful if you have domain-virtual hosts.
4334 WARNING: Squid and other software verifies the URL and Host
4335 are matching, so be careful not to relay through other proxies
4336 or inspecting firewalls with this disabled.
4339 NAME: url_rewrite_access redirector_access
4342 LOC: Config.accessList.redirector
4344 If defined, this access list specifies which requests are
4345 sent to the redirector processes. By default all requests
4348 This clause supports both fast and slow acl types.
4349 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4352 NAME: url_rewrite_bypass redirector_bypass
4354 LOC: Config.onoff.redirector_bypass
4357 When this is 'on', a request will not go through the
4358 redirector if all redirectors are busy. If this is 'off'
4359 and the redirector queue grows too large, Squid will exit
4360 with a FATAL error and ask you to increase the number of
4361 redirectors. You should only enable this if the redirectors
4362 are not critical to your caching system. If you use
4363 redirectors for access control, and you enable this option,
4364 users may have access to pages they should not
4365 be allowed to request.
4369 OPTIONS FOR STORE ID
4370 -----------------------------------------------------------------------------
4373 NAME: store_id_program storeurl_rewrite_program
4375 LOC: Config.Program.store_id
4378 Specify the location of the executable StoreID helper to use.
4379 Since they can perform almost any function there isn't one included.
4381 For each requested URL, the helper will receive one line with the format
4383 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4386 After processing the request the helper must reply using the following format:
4388 [channel-ID <SP>] result [<SP> kv-pairs]
4390 The result code can be:
4393 Use the StoreID supplied in 'store-id='.
4396 The default is to use HTTP request URL as the store ID.
4399 An internal error occured in the helper, preventing
4400 a result being identified.
4403 Helper programs should be prepared to receive and possibly ignore additional
4404 kv-pairs with keys they do not support.
4406 When using the concurrency= option the protocol is changed by
4407 introducing a query channel tag in front of the request/response.
4408 The query channel tag is a number between 0 and concurrency-1.
4409 This value must be echoed back unchanged to Squid as the first part
4410 of the response relating to its request.
4412 NOTE: when using StoreID refresh_pattern will apply to the StoreID
4413 returned from the helper and not the URL.
4415 WARNING: Wrong StoreID value returned by a careless helper may result
4416 in the wrong cached response returned to the user.
4418 By default, a StoreID helper is not used.
4421 NAME: store_id_children storeurl_rewrite_children
4422 TYPE: HelperChildConfig
4423 DEFAULT: 20 startup=0 idle=1 concurrency=0
4424 LOC: Config.storeIdChildren
4426 The maximum number of StoreID helper processes to spawn. If you limit
4427 it too few Squid will have to wait for them to process a backlog of
4428 requests, slowing it down. If you allow too many they will use RAM
4429 and other system resources noticably.
4431 The startup= and idle= options allow some measure of skew in your
4436 Sets a minimum of how many processes are to be spawned when Squid
4437 starts or reconfigures. When set to zero the first request will
4438 cause spawning of the first child process to handle it.
4440 Starting too few will cause an initial slowdown in traffic as Squid
4441 attempts to simultaneously spawn enough processes to cope.
4445 Sets a minimum of how many processes Squid is to try and keep available
4446 at all times. When traffic begins to rise above what the existing
4447 processes can handle this many more will be spawned up to the maximum
4448 configured. A minimum setting of 1 is required.
4452 The number of requests each storeID helper can handle in
4453 parallel. Defaults to 0 which indicates the helper
4454 is a old-style single threaded program.
4456 When this directive is set to a value >= 1 then the protocol
4457 used to communicate with the helper is modified to include
4458 an ID in front of the request/response. The ID from the request
4459 must be echoed back with the response to that request.
4462 NAME: store_id_access storeurl_rewrite_access
4465 LOC: Config.accessList.store_id
4467 If defined, this access list specifies which requests are
4468 sent to the StoreID processes. By default all requests
4471 This clause supports both fast and slow acl types.
4472 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4475 NAME: store_id_bypass storeurl_rewrite_bypass
4477 LOC: Config.onoff.store_id_bypass
4480 When this is 'on', a request will not go through the
4481 helper if all helpers are busy. If this is 'off'
4482 and the helper queue grows too large, Squid will exit
4483 with a FATAL error and ask you to increase the number of
4484 helpers. You should only enable this if the helperss
4485 are not critical to your caching system. If you use
4486 helpers for critical caching components, and you enable this
4487 option, users may not get objects from cache.
4491 OPTIONS FOR TUNING THE CACHE
4492 -----------------------------------------------------------------------------
4495 NAME: cache no_cache
4498 LOC: Config.accessList.noCache
4500 A list of ACL elements which, if matched and denied, cause the request to
4501 not be satisfied from the cache and the reply to not be cached.
4502 In other words, use this to force certain objects to never be cached.
4504 You must use the words 'allow' or 'deny' to indicate whether items
4505 matching the ACL should be allowed or denied into the cache.
4507 Default is to allow all to be cached.
4509 This clause supports both fast and slow acl types.
4510 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4516 LOC: Config.maxStale
4519 This option puts an upper limit on how stale content Squid
4520 will serve from the cache if cache validation fails.
4521 Can be overriden by the refresh_pattern max-stale option.
4524 NAME: refresh_pattern
4525 TYPE: refreshpattern
4529 usage: refresh_pattern [-i] regex min percent max [options]
4531 By default, regular expressions are CASE-SENSITIVE. To make
4532 them case-insensitive, use the -i option.
4534 'Min' is the time (in minutes) an object without an explicit
4535 expiry time should be considered fresh. The recommended
4536 value is 0, any higher values may cause dynamic applications
4537 to be erroneously cached unless the application designer
4538 has taken the appropriate actions.
4540 'Percent' is a percentage of the objects age (time since last
4541 modification age) an object without explicit expiry time
4542 will be considered fresh.
4544 'Max' is an upper limit on how long objects without an explicit
4545 expiry time will be considered fresh.
4547 options: override-expire
4552 ignore-must-revalidate
4559 override-expire enforces min age even if the server
4560 sent an explicit expiry time (e.g., with the
4561 Expires: header or Cache-Control: max-age). Doing this
4562 VIOLATES the HTTP standard. Enabling this feature
4563 could make you liable for problems which it causes.
4565 Note: override-expire does not enforce staleness - it only extends
4566 freshness / min. If the server returns a Expires time which
4567 is longer than your max time, Squid will still consider
4568 the object fresh for that period of time.
4570 override-lastmod enforces min age even on objects
4571 that were modified recently.
4573 reload-into-ims changes client no-cache or ``reload''
4574 to If-Modified-Since requests. Doing this VIOLATES the
4575 HTTP standard. Enabling this feature could make you
4576 liable for problems which it causes.
4578 ignore-reload ignores a client no-cache or ``reload''
4579 header. Doing this VIOLATES the HTTP standard. Enabling
4580 this feature could make you liable for problems which
4583 ignore-no-store ignores any ``Cache-control: no-store''
4584 headers received from a server. Doing this VIOLATES
4585 the HTTP standard. Enabling this feature could make you
4586 liable for problems which it causes.
4588 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4589 headers received from a server. Doing this VIOLATES
4590 the HTTP standard. Enabling this feature could make you
4591 liable for problems which it causes.
4593 ignore-private ignores any ``Cache-control: private''
4594 headers received from a server. Doing this VIOLATES
4595 the HTTP standard. Enabling this feature could make you
4596 liable for problems which it causes.
4598 ignore-auth caches responses to requests with authorization,
4599 as if the originserver had sent ``Cache-control: public''
4600 in the response header. Doing this VIOLATES the HTTP standard.
4601 Enabling this feature could make you liable for problems which
4604 refresh-ims causes squid to contact the origin server
4605 when a client issues an If-Modified-Since request. This
4606 ensures that the client will receive an updated version
4607 if one is available.
4609 store-stale stores responses even if they don't have explicit
4610 freshness or a validator (i.e., Last-Modified or an ETag)
4611 present, or if they're already stale. By default, Squid will
4612 not cache such responses because they usually can't be
4613 reused. Note that such responses will be stale by default.
4615 max-stale=NN provide a maximum staleness factor. Squid won't
4616 serve objects more stale than this even if it failed to
4617 validate the object. Default: use the max_stale global limit.
4619 Basically a cached object is:
4621 FRESH if expires < now, else STALE
4623 FRESH if lm-factor < percent, else STALE
4627 The refresh_pattern lines are checked in the order listed here.
4628 The first entry which matches is used. If none of the entries
4629 match the default will be used.
4631 Note, you must uncomment all the default lines if you want
4632 to change one. The default setting is only active if none is
4637 # Add any of your own refresh_pattern entries above these.
4638 refresh_pattern ^ftp: 1440 20% 10080
4639 refresh_pattern ^gopher: 1440 0% 1440
4640 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4641 refresh_pattern . 0 20% 4320
4645 NAME: quick_abort_min
4649 LOC: Config.quickAbort.min
4652 NAME: quick_abort_max
4656 LOC: Config.quickAbort.max
4659 NAME: quick_abort_pct
4663 LOC: Config.quickAbort.pct
4665 The cache by default continues downloading aborted requests
4666 which are almost completed (less than 16 KB remaining). This
4667 may be undesirable on slow (e.g. SLIP) links and/or very busy
4668 caches. Impatient users may tie up file descriptors and
4669 bandwidth by repeatedly requesting and immediately aborting
4672 When the user aborts a request, Squid will check the
4673 quick_abort values to the amount of data transfered until
4676 If the transfer has less than 'quick_abort_min' KB remaining,
4677 it will finish the retrieval.
4679 If the transfer has more than 'quick_abort_max' KB remaining,
4680 it will abort the retrieval.
4682 If more than 'quick_abort_pct' of the transfer has completed,
4683 it will finish the retrieval.
4685 If you do not want any retrieval to continue after the client
4686 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4689 If you want retrievals to always continue if they are being
4690 cached set 'quick_abort_min' to '-1 KB'.
4693 NAME: read_ahead_gap
4694 COMMENT: buffer-size
4696 LOC: Config.readAheadGap
4699 The amount of data the cache will buffer ahead of what has been
4700 sent to the client when retrieving an object from another server.
4704 IFDEF: USE_HTTP_VIOLATIONS
4707 LOC: Config.negativeTtl
4710 Set the Default Time-to-Live (TTL) for failed requests.
4711 Certain types of failures (such as "connection refused" and
4712 "404 Not Found") are able to be negatively-cached for a short time.
4713 Modern web servers should provide Expires: header, however if they
4714 do not this can provide a minimum TTL.
4715 The default is not to cache errors with unknown expiry details.
4717 Note that this is different from negative caching of DNS lookups.
4719 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4720 this feature could make you liable for problems which it
4724 NAME: positive_dns_ttl
4727 LOC: Config.positiveDnsTtl
4730 Upper limit on how long Squid will cache positive DNS responses.
4731 Default is 6 hours (360 minutes). This directive must be set
4732 larger than negative_dns_ttl.
4735 NAME: negative_dns_ttl
4738 LOC: Config.negativeDnsTtl
4741 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4742 This also sets the lower cache limit on positive lookups.
4743 Minimum value is 1 second, and it is not recommendable to go
4744 much below 10 seconds.
4747 NAME: range_offset_limit
4748 COMMENT: size [acl acl...]
4750 LOC: Config.rangeOffsetLimit
4753 usage: (size) [units] [[!]aclname]
4755 Sets an upper limit on how far (number of bytes) into the file
4756 a Range request may be to cause Squid to prefetch the whole file.
4757 If beyond this limit, Squid forwards the Range request as it is and
4758 the result is NOT cached.
4760 This is to stop a far ahead range request (lets say start at 17MB)
4761 from making Squid fetch the whole object up to that point before
4762 sending anything to the client.
4764 Multiple range_offset_limit lines may be specified, and they will
4765 be searched from top to bottom on each request until a match is found.
4766 The first match found will be used. If no line matches a request, the
4767 default limit of 0 bytes will be used.
4769 'size' is the limit specified as a number of units.
4771 'units' specifies whether to use bytes, KB, MB, etc.
4772 If no units are specified bytes are assumed.
4774 A size of 0 causes Squid to never fetch more than the
4775 client requested. (default)
4777 A size of 'none' causes Squid to always fetch the object from the
4778 beginning so it may cache the result. (2.0 style)
4780 'aclname' is the name of a defined ACL.
4782 NP: Using 'none' as the byte value here will override any quick_abort settings
4783 that may otherwise apply to the range request. The range request will
4784 be fully fetched from start to finish regardless of the client
4785 actions. This affects bandwidth usage.
4788 NAME: minimum_expiry_time
4791 LOC: Config.minimum_expiry_time
4794 The minimum caching time according to (Expires - Date)
4795 Headers Squid honors if the object can't be revalidated
4796 defaults to 60 seconds. In reverse proxy environments it
4797 might be desirable to honor shorter object lifetimes. It
4798 is most likely better to make your server return a
4799 meaningful Last-Modified header however. In ESI environments
4800 where page fragments often have short lifetimes, this will
4801 often be best set to 0.
4804 NAME: store_avg_object_size
4808 LOC: Config.Store.avgObjectSize
4810 Average object size, used to estimate number of objects your
4811 cache can hold. The default is 13 KB.
4814 NAME: store_objects_per_bucket
4817 LOC: Config.Store.objectsPerBucket
4819 Target number of objects per bucket in the store hash table.
4820 Lowering this value increases the total number of buckets and
4821 also the storage maintenance rate. The default is 20.
4826 -----------------------------------------------------------------------------
4829 NAME: request_header_max_size
4833 LOC: Config.maxRequestHeaderSize
4835 This specifies the maximum size for HTTP headers in a request.
4836 Request headers are usually relatively small (about 512 bytes).
4837 Placing a limit on the request header size will catch certain
4838 bugs (for example with persistent connections) and possibly
4839 buffer-overflow or denial-of-service attacks.
4842 NAME: reply_header_max_size
4846 LOC: Config.maxReplyHeaderSize
4848 This specifies the maximum size for HTTP headers in a reply.
4849 Reply headers are usually relatively small (about 512 bytes).
4850 Placing a limit on the reply header size will catch certain
4851 bugs (for example with persistent connections) and possibly
4852 buffer-overflow or denial-of-service attacks.
4855 NAME: request_body_max_size
4859 LOC: Config.maxRequestBodySize
4861 This specifies the maximum size for an HTTP request body.
4862 In other words, the maximum size of a PUT/POST request.
4863 A user who attempts to send a request with a body larger
4864 than this limit receives an "Invalid Request" error message.
4865 If you set this parameter to a zero (the default), there will
4866 be no limit imposed.
4869 NAME: client_request_buffer_max_size
4873 LOC: Config.maxRequestBufferSize
4875 This specifies the maximum buffer size of a client request.
4876 It prevents squid eating too much memory when somebody uploads
4880 NAME: chunked_request_body_max_size
4884 LOC: Config.maxChunkedRequestBodySize
4886 A broken or confused HTTP/1.1 client may send a chunked HTTP
4887 request to Squid. Squid does not have full support for that
4888 feature yet. To cope with such requests, Squid buffers the
4889 entire request and then dechunks request body to create a
4890 plain HTTP/1.0 request with a known content length. The plain
4891 request is then used by the rest of Squid code as usual.
4893 The option value specifies the maximum size of the buffer used
4894 to hold the request before the conversion. If the chunked
4895 request size exceeds the specified limit, the conversion
4896 fails, and the client receives an "unsupported request" error,
4897 as if dechunking was disabled.
4899 Dechunking is enabled by default. To disable conversion of
4900 chunked requests, set the maximum to zero.
4902 Request dechunking feature and this option in particular are a
4903 temporary hack. When chunking requests and responses are fully
4904 supported, there will be no need to buffer a chunked request.
4908 IFDEF: USE_HTTP_VIOLATIONS
4911 LOC: Config.accessList.brokenPosts
4913 A list of ACL elements which, if matched, causes Squid to send
4914 an extra CRLF pair after the body of a PUT/POST request.
4916 Some HTTP servers has broken implementations of PUT/POST,
4917 and rely on an extra CRLF pair sent by some WWW clients.
4919 Quote from RFC2616 section 4.1 on this matter:
4921 Note: certain buggy HTTP/1.0 client implementations generate an
4922 extra CRLF's after a POST request. To restate what is explicitly
4923 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4924 a request with an extra CRLF.
4926 This clause only supports fast acl types.
4927 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4930 acl buggy_server url_regex ^http://....
4931 broken_posts allow buggy_server
4934 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4937 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4939 LOC: Adaptation::Config::use_indirect_client
4941 Controls whether the indirect client IP address (instead of the direct
4942 client IP address) is passed to adaptation services.
4944 See also: follow_x_forwarded_for adaptation_send_client_ip
4948 IFDEF: USE_HTTP_VIOLATIONS
4952 LOC: Config.onoff.via
4954 If set (default), Squid will include a Via header in requests and
4955 replies as required by RFC2616.
4961 LOC: Config.onoff.ie_refresh
4964 Microsoft Internet Explorer up until version 5.5 Service
4965 Pack 1 has an issue with transparent proxies, wherein it
4966 is impossible to force a refresh. Turning this on provides
4967 a partial fix to the problem, by causing all IMS-REFRESH
4968 requests from older IE versions to check the origin server
4969 for fresh content. This reduces hit ratio by some amount
4970 (~10% in my experience), but allows users to actually get
4971 fresh content when they want it. Note because Squid
4972 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4973 of 5.5 is unchanged from old versions of Squid (i.e. a
4974 forced refresh is impossible). Newer versions of IE will,
4975 hopefully, continue to have the new behavior and will be
4976 handled based on that assumption. This option defaults to
4977 the old Squid behavior, which is better for hit ratios but
4978 worse for clients using IE, if they need to be able to
4979 force fresh content.
4982 NAME: vary_ignore_expire
4985 LOC: Config.onoff.vary_ignore_expire
4988 Many HTTP servers supporting Vary gives such objects
4989 immediate expiry time with no cache-control header
4990 when requested by a HTTP/1.0 client. This option
4991 enables Squid to ignore such expiry times until
4992 HTTP/1.1 is fully implemented.
4994 WARNING: If turned on this may eventually cause some
4995 varying objects not intended for caching to get cached.
4998 NAME: request_entities
5000 LOC: Config.onoff.request_entities
5003 Squid defaults to deny GET and HEAD requests with request entities,
5004 as the meaning of such requests are undefined in the HTTP standard
5005 even if not explicitly forbidden.
5007 Set this directive to on if you have clients which insists
5008 on sending request entities in GET or HEAD requests. But be warned
5009 that there is server software (both proxies and web servers) which
5010 can fail to properly process this kind of request which may make you
5011 vulnerable to cache pollution attacks if enabled.
5014 NAME: request_header_access
5015 IFDEF: USE_HTTP_VIOLATIONS
5016 TYPE: http_header_access
5017 LOC: Config.request_header_access
5020 Usage: request_header_access header_name allow|deny [!]aclname ...
5022 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5023 this feature could make you liable for problems which it
5026 This option replaces the old 'anonymize_headers' and the
5027 older 'http_anonymizer' option with something that is much
5028 more configurable. A list of ACLs for each header name allows
5029 removal of specific header fields under specific conditions.
5031 This option only applies to outgoing HTTP request headers (i.e.,
5032 headers sent by Squid to the next HTTP hop such as a cache peer
5033 or an origin server). The option has no effect during cache hit
5034 detection. The equivalent adaptation vectoring point in ICAP
5035 terminology is post-cache REQMOD.
5037 The option is applied to individual outgoing request header
5038 fields. For each request header field F, Squid uses the first
5039 qualifying sets of request_header_access rules:
5041 1. Rules with header_name equal to F's name.
5042 2. Rules with header_name 'Other', provided F's name is not
5043 on the hard-coded list of commonly used HTTP header names.
5044 3. Rules with header_name 'All'.
5046 Within that qualifying rule set, rule ACLs are checked as usual.
5047 If ACLs of an "allow" rule match, the header field is allowed to
5048 go through as is. If ACLs of a "deny" rule match, the header is
5049 removed and request_header_replace is then checked to identify
5050 if the removed header has a replacement. If no rules within the
5051 set have matching ACLs, the header field is left as is.
5053 For example, to achieve the same behavior as the old
5054 'http_anonymizer standard' option, you should use:
5056 request_header_access From deny all
5057 request_header_access Referer deny all
5058 request_header_access Server deny all
5059 request_header_access User-Agent deny all
5060 request_header_access WWW-Authenticate deny all
5061 request_header_access Link deny all
5063 Or, to reproduce the old 'http_anonymizer paranoid' feature
5066 request_header_access Allow allow all
5067 request_header_access Authorization allow all
5068 request_header_access WWW-Authenticate allow all
5069 request_header_access Proxy-Authorization allow all
5070 request_header_access Proxy-Authenticate allow all
5071 request_header_access Cache-Control allow all
5072 request_header_access Content-Encoding allow all
5073 request_header_access Content-Length allow all
5074 request_header_access Content-Type allow all
5075 request_header_access Date allow all
5076 request_header_access Expires allow all
5077 request_header_access Host allow all
5078 request_header_access If-Modified-Since allow all
5079 request_header_access Last-Modified allow all
5080 request_header_access Location allow all
5081 request_header_access Pragma allow all
5082 request_header_access Accept allow all
5083 request_header_access Accept-Charset allow all
5084 request_header_access Accept-Encoding allow all
5085 request_header_access Accept-Language allow all
5086 request_header_access Content-Language allow all
5087 request_header_access Mime-Version allow all
5088 request_header_access Retry-After allow all
5089 request_header_access Title allow all
5090 request_header_access Connection allow all
5091 request_header_access All deny all
5093 although many of those are HTTP reply headers, and so should be
5094 controlled with the reply_header_access directive.
5096 By default, all headers are allowed (no anonymizing is
5100 NAME: reply_header_access
5101 IFDEF: USE_HTTP_VIOLATIONS
5102 TYPE: http_header_access
5103 LOC: Config.reply_header_access
5106 Usage: reply_header_access header_name allow|deny [!]aclname ...
5108 WARNING: Doing this VIOLATES the HTTP standard. Enabling
5109 this feature could make you liable for problems which it
5112 This option only applies to reply headers, i.e., from the
5113 server to the client.
5115 This is the same as request_header_access, but in the other
5116 direction. Please see request_header_access for detailed
5119 For example, to achieve the same behavior as the old
5120 'http_anonymizer standard' option, you should use:
5122 reply_header_access From deny all
5123 reply_header_access Referer deny all
5124 reply_header_access Server deny all
5125 reply_header_access User-Agent deny all
5126 reply_header_access WWW-Authenticate deny all
5127 reply_header_access Link deny all
5129 Or, to reproduce the old 'http_anonymizer paranoid' feature
5132 reply_header_access Allow allow all
5133 reply_header_access Authorization allow all
5134 reply_header_access WWW-Authenticate allow all
5135 reply_header_access Proxy-Authorization allow all
5136 reply_header_access Proxy-Authenticate allow all
5137 reply_header_access Cache-Control allow all
5138 reply_header_access Content-Encoding allow all
5139 reply_header_access Content-Length allow all
5140 reply_header_access Content-Type allow all
5141 reply_header_access Date allow all
5142 reply_header_access Expires allow all
5143 reply_header_access Host allow all
5144 reply_header_access If-Modified-Since allow all
5145 reply_header_access Last-Modified allow all
5146 reply_header_access Location allow all
5147 reply_header_access Pragma allow all
5148 reply_header_access Accept allow all
5149 reply_header_access Accept-Charset allow all
5150 reply_header_access Accept-Encoding allow all
5151 reply_header_access Accept-Language allow all
5152 reply_header_access Content-Language allow all
5153 reply_header_access Mime-Version allow all
5154 reply_header_access Retry-After allow all
5155 reply_header_access Title allow all
5156 reply_header_access Connection allow all
5157 reply_header_access All deny all
5159 although the HTTP request headers won't be usefully controlled
5160 by this directive -- see request_header_access for details.
5162 By default, all headers are allowed (no anonymizing is
5166 NAME: request_header_replace header_replace
5167 IFDEF: USE_HTTP_VIOLATIONS
5168 TYPE: http_header_replace
5169 LOC: Config.request_header_access
5172 Usage: request_header_replace header_name message
5173 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5175 This option allows you to change the contents of headers
5176 denied with request_header_access above, by replacing them
5177 with some fixed string. This replaces the old fake_user_agent
5180 This only applies to request headers, not reply headers.
5182 By default, headers are removed if denied.
5185 NAME: reply_header_replace
5186 IFDEF: USE_HTTP_VIOLATIONS
5187 TYPE: http_header_replace
5188 LOC: Config.reply_header_access
5191 Usage: reply_header_replace header_name message
5192 Example: reply_header_replace Server Foo/1.0
5194 This option allows you to change the contents of headers
5195 denied with reply_header_access above, by replacing them
5196 with some fixed string.
5198 This only applies to reply headers, not request headers.
5200 By default, headers are removed if denied.
5203 NAME: request_header_add
5204 TYPE: HeaderWithAclList
5205 LOC: Config.request_header_add
5208 Usage: request_header_add field-name field-value acl1 [acl2] ...
5209 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5211 This option adds header fields to outgoing HTTP requests (i.e.,
5212 request headers sent by Squid to the next HTTP hop such as a
5213 cache peer or an origin server). The option has no effect during
5214 cache hit detection. The equivalent adaptation vectoring point
5215 in ICAP terminology is post-cache REQMOD.
5217 Field-name is a token specifying an HTTP header name. If a
5218 standard HTTP header name is used, Squid does not check whether
5219 the new header conflicts with any existing headers or violates
5220 HTTP rules. If the request to be modified already contains a
5221 field with the same name, the old field is preserved but the
5222 header field values are not merged.
5224 Field-value is either a token or a quoted string. If quoted
5225 string format is used, then the surrounding quotes are removed
5226 while escape sequences and %macros are processed.
5228 In theory, all of the logformat codes can be used as %macros.
5229 However, unlike logging (which happens at the very end of
5230 transaction lifetime), the transaction may not yet have enough
5231 information to expand a macro when the new header value is needed.
5232 And some information may already be available to Squid but not yet
5233 committed where the macro expansion code can access it (report
5234 such instances!). The macro will be expanded into a single dash
5235 ('-') in such cases. Not all macros have been tested.
5237 One or more Squid ACLs may be specified to restrict header
5238 injection to matching requests. As always in squid.conf, all
5239 ACLs in an option ACL list must be satisfied for the insertion
5240 to happen. The request_header_add option supports fast ACLs
5249 This option used to log custom information about the master
5250 transaction. For example, an admin may configure Squid to log
5251 which "user group" the transaction belongs to, where "user group"
5252 will be determined based on a set of ACLs and not [just]
5253 authentication information.
5254 Values of key/value pairs can be logged using %{key}note macros:
5256 note key value acl ...
5257 logformat myFormat ... %{key}note ...
5260 NAME: relaxed_header_parser
5261 COMMENT: on|off|warn
5263 LOC: Config.onoff.relaxed_header_parser
5266 In the default "on" setting Squid accepts certain forms
5267 of non-compliant HTTP messages where it is unambiguous
5268 what the sending application intended even if the message
5269 is not correctly formatted. The messages is then normalized
5270 to the correct form when forwarded by Squid.
5272 If set to "warn" then a warning will be emitted in cache.log
5273 each time such HTTP error is encountered.
5275 If set to "off" then such HTTP errors will cause the request
5276 or response to be rejected.
5281 -----------------------------------------------------------------------------
5284 NAME: forward_timeout
5287 LOC: Config.Timeout.forward
5290 This parameter specifies how long Squid should at most attempt in
5291 finding a forwarding path for the request before giving up.
5294 NAME: connect_timeout
5297 LOC: Config.Timeout.connect
5300 This parameter specifies how long to wait for the TCP connect to
5301 the requested server or peer to complete before Squid should
5302 attempt to find another path where to forward the request.
5305 NAME: peer_connect_timeout
5308 LOC: Config.Timeout.peer_connect
5311 This parameter specifies how long to wait for a pending TCP
5312 connection to a peer cache. The default is 30 seconds. You
5313 may also set different timeout values for individual neighbors
5314 with the 'connect-timeout' option on a 'cache_peer' line.
5320 LOC: Config.Timeout.read
5323 The read_timeout is applied on server-side connections. After
5324 each successful read(), the timeout will be extended by this
5325 amount. If no data is read again after this amount of time,
5326 the request is aborted and logged with ERR_READ_TIMEOUT. The
5327 default is 15 minutes.
5333 LOC: Config.Timeout.write
5336 This timeout is tracked for all connections that have data
5337 available for writing and are waiting for the socket to become
5338 ready. After each successful write, the timeout is extended by
5339 the configured amount. If Squid has data to write but the
5340 connection is not ready for the configured duration, the
5341 transaction associated with the connection is terminated. The
5342 default is 15 minutes.
5345 NAME: request_timeout
5347 LOC: Config.Timeout.request
5350 How long to wait for complete HTTP request headers after initial
5351 connection establishment.
5354 NAME: client_idle_pconn_timeout persistent_request_timeout
5356 LOC: Config.Timeout.clientIdlePconn
5359 How long to wait for the next HTTP request on a persistent
5360 client connection after the previous request completes.
5363 NAME: client_lifetime
5366 LOC: Config.Timeout.lifetime
5369 The maximum amount of time a client (browser) is allowed to
5370 remain connected to the cache process. This protects the Cache
5371 from having a lot of sockets (and hence file descriptors) tied up
5372 in a CLOSE_WAIT state from remote clients that go away without
5373 properly shutting down (either because of a network failure or
5374 because of a poor client implementation). The default is one
5377 NOTE: The default value is intended to be much larger than any
5378 client would ever need to be connected to your cache. You
5379 should probably change client_lifetime only as a last resort.
5380 If you seem to have many client connections tying up
5381 filedescriptors, we recommend first tuning the read_timeout,
5382 request_timeout, persistent_request_timeout and quick_abort values.
5385 NAME: half_closed_clients
5387 LOC: Config.onoff.half_closed_clients
5390 Some clients may shutdown the sending side of their TCP
5391 connections, while leaving their receiving sides open. Sometimes,
5392 Squid can not tell the difference between a half-closed and a
5393 fully-closed TCP connection.
5395 By default, Squid will immediately close client connections when
5396 read(2) returns "no more data to read."
5398 Change this option to 'on' and Squid will keep open connections
5399 until a read(2) or write(2) on the socket returns an error.
5400 This may show some benefits for reverse proxies. But if not
5401 it is recommended to leave OFF.
5404 NAME: server_idle_pconn_timeout pconn_timeout
5406 LOC: Config.Timeout.serverIdlePconn
5409 Timeout for idle persistent connections to servers and other
5416 LOC: Ident::TheConfig.timeout
5419 Maximum time to wait for IDENT lookups to complete.
5421 If this is too high, and you enabled IDENT lookups from untrusted
5422 users, you might be susceptible to denial-of-service by having
5423 many ident requests going at once.
5426 NAME: shutdown_lifetime
5429 LOC: Config.shutdownLifetime
5432 When SIGTERM or SIGHUP is received, the cache is put into
5433 "shutdown pending" mode until all active sockets are closed.
5434 This value is the lifetime to set for all open descriptors
5435 during shutdown mode. Any active clients after this many
5436 seconds will receive a 'timeout' message.
5440 ADMINISTRATIVE PARAMETERS
5441 -----------------------------------------------------------------------------
5447 LOC: Config.adminEmail
5449 Email-address of local cache manager who will receive
5450 mail if the cache dies. The default is "webmaster."
5456 LOC: Config.EmailFrom
5458 From: email-address for mail sent when the cache dies.
5459 The default is to use 'appname@unique_hostname'.
5460 Default appname value is "squid", can be changed into
5461 src/globals.h before building squid.
5467 LOC: Config.EmailProgram
5469 Email program used to send mail if the cache dies.
5470 The default is "mail". The specified program must comply
5471 with the standard Unix mail syntax:
5472 mail-program recipient < mailfile
5474 Optional command line options can be specified.
5477 NAME: cache_effective_user
5479 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5480 LOC: Config.effectiveUser
5482 If you start Squid as root, it will change its effective/real
5483 UID/GID to the user specified below. The default is to change
5484 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5485 see also; cache_effective_group
5488 NAME: cache_effective_group
5491 LOC: Config.effectiveGroup
5493 Squid sets the GID to the effective user's default group ID
5494 (taken from the password file) and supplementary group list
5495 from the groups membership.
5497 If you want Squid to run with a specific GID regardless of
5498 the group memberships of the effective user then set this
5499 to the group (or GID) you want Squid to run as. When set
5500 all other group privileges of the effective user are ignored
5501 and only this GID is effective. If Squid is not started as
5502 root the user starting Squid MUST be member of the specified
5505 This option is not recommended by the Squid Team.
5506 Our preference is for administrators to configure a secure
5507 user account for squid with UID/GID matching system policies.
5510 NAME: httpd_suppress_version_string
5514 LOC: Config.onoff.httpd_suppress_version_string
5516 Suppress Squid version string info in HTTP headers and HTML error pages.
5519 NAME: visible_hostname
5521 LOC: Config.visibleHostname
5524 If you want to present a special hostname in error messages, etc,
5525 define this. Otherwise, the return value of gethostname()
5526 will be used. If you have multiple caches in a cluster and
5527 get errors about IP-forwarding you must set them to have individual
5528 names with this setting.
5531 NAME: unique_hostname
5533 LOC: Config.uniqueHostname
5536 If you want to have multiple machines with the same
5537 'visible_hostname' you must give each machine a different
5538 'unique_hostname' so forwarding loops can be detected.
5541 NAME: hostname_aliases
5543 LOC: Config.hostnameAliases
5546 A list of other DNS names your cache has.
5554 Minimum umask which should be enforced while the proxy
5555 is running, in addition to the umask set at startup.
5557 For a traditional octal representation of umasks, start
5562 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5563 -----------------------------------------------------------------------------
5565 This section contains parameters for the (optional) cache
5566 announcement service. This service is provided to help
5567 cache administrators locate one another in order to join or
5568 create cache hierarchies.
5570 An 'announcement' message is sent (via UDP) to the registration
5571 service by Squid. By default, the announcement message is NOT
5572 SENT unless you enable it with 'announce_period' below.
5574 The announcement message includes your hostname, plus the
5575 following information from this configuration file:
5581 All current information is processed regularly and made
5582 available on the Web at http://www.ircache.net/Cache/Tracker/.
5585 NAME: announce_period
5587 LOC: Config.Announce.period
5590 This is how frequently to send cache announcements. The
5591 default is `0' which disables sending the announcement
5594 To enable announcing your cache, just set an announce period.
5597 announce_period 1 day
5602 DEFAULT: tracker.ircache.net
5603 LOC: Config.Announce.host
5609 LOC: Config.Announce.file
5615 LOC: Config.Announce.port
5617 announce_host and announce_port set the hostname and port
5618 number where the registration message will be sent.
5620 Hostname will default to 'tracker.ircache.net' and port will
5621 default default to 3131. If the 'filename' argument is given,
5622 the contents of that file will be included in the announce
5627 HTTPD-ACCELERATOR OPTIONS
5628 -----------------------------------------------------------------------------
5631 NAME: httpd_accel_surrogate_id
5634 LOC: Config.Accel.surrogate_id
5636 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5637 need an identification token to allow control targeting. Because
5638 a farm of surrogates may all perform the same tasks, they may share
5639 an identification token.
5641 The default ID is the visible_hostname
5644 NAME: http_accel_surrogate_remote
5648 LOC: Config.onoff.surrogate_is_remote
5650 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
5651 Set this to on to have squid behave as a remote surrogate.
5655 IFDEF: USE_SQUID_ESI
5656 COMMENT: libxml2|expat|custom
5658 LOC: ESIParser::Type
5661 ESI markup is not strictly XML compatible. The custom ESI parser
5662 will give higher performance, but cannot handle non ASCII character
5667 DELAY POOL PARAMETERS
5668 -----------------------------------------------------------------------------
5672 TYPE: delay_pool_count
5674 IFDEF: USE_DELAY_POOLS
5677 This represents the number of delay pools to be used. For example,
5678 if you have one class 2 delay pool and one class 3 delays pool, you
5679 have a total of 2 delay pools.
5683 TYPE: delay_pool_class
5685 IFDEF: USE_DELAY_POOLS
5688 This defines the class of each delay pool. There must be exactly one
5689 delay_class line for each delay pool. For example, to define two
5690 delay pools, one of class 2 and one of class 3, the settings above
5694 delay_pools 4 # 4 delay pools
5695 delay_class 1 2 # pool 1 is a class 2 pool
5696 delay_class 2 3 # pool 2 is a class 3 pool
5697 delay_class 3 4 # pool 3 is a class 4 pool
5698 delay_class 4 5 # pool 4 is a class 5 pool
5700 The delay pool classes are:
5702 class 1 Everything is limited by a single aggregate
5705 class 2 Everything is limited by a single aggregate
5706 bucket as well as an "individual" bucket chosen
5707 from bits 25 through 32 of the IPv4 address.
5709 class 3 Everything is limited by a single aggregate
5710 bucket as well as a "network" bucket chosen
5711 from bits 17 through 24 of the IP address and a
5712 "individual" bucket chosen from bits 17 through
5713 32 of the IPv4 address.
5715 class 4 Everything in a class 3 delay pool, with an
5716 additional limit on a per user basis. This
5717 only takes effect if the username is established
5718 in advance - by forcing authentication in your
5721 class 5 Requests are grouped according their tag (see
5722 external_acl's tag= reply).
5725 Each pool also requires a delay_parameters directive to configure the pool size
5726 and speed limits used whenever the pool is applied to a request. Along with
5727 a set of delay_access directives to determine when it is used.
5729 NOTE: If an IP address is a.b.c.d
5730 -> bits 25 through 32 are "d"
5731 -> bits 17 through 24 are "c"
5732 -> bits 17 through 32 are "c * 256 + d"
5734 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5735 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5739 TYPE: delay_pool_access
5741 IFDEF: USE_DELAY_POOLS
5744 This is used to determine which delay pool a request falls into.
5746 delay_access is sorted per pool and the matching starts with pool 1,
5747 then pool 2, ..., and finally pool N. The first delay pool where the
5748 request is allowed is selected for the request. If it does not allow
5749 the request to any pool then the request is not delayed (default).
5751 For example, if you want some_big_clients in delay
5752 pool 1 and lotsa_little_clients in delay pool 2:
5755 delay_access 1 allow some_big_clients
5756 delay_access 1 deny all
5757 delay_access 2 allow lotsa_little_clients
5758 delay_access 2 deny all
5759 delay_access 3 allow authenticated_clients
5762 NAME: delay_parameters
5763 TYPE: delay_pool_rates
5765 IFDEF: USE_DELAY_POOLS
5768 This defines the parameters for a delay pool. Each delay pool has
5769 a number of "buckets" associated with it, as explained in the
5770 description of delay_class.
5772 For a class 1 delay pool, the syntax is:
5774 delay_parameters pool aggregate
5776 For a class 2 delay pool:
5778 delay_parameters pool aggregate individual
5780 For a class 3 delay pool:
5782 delay_parameters pool aggregate network individual
5784 For a class 4 delay pool:
5786 delay_parameters pool aggregate network individual user
5788 For a class 5 delay pool:
5790 delay_parameters pool tagrate
5792 The option variables are:
5794 pool a pool number - ie, a number between 1 and the
5795 number specified in delay_pools as used in
5798 aggregate the speed limit parameters for the aggregate bucket
5801 individual the speed limit parameters for the individual
5802 buckets (class 2, 3).
5804 network the speed limit parameters for the network buckets
5807 user the speed limit parameters for the user buckets
5810 tagrate the speed limit parameters for the tag buckets
5813 A pair of delay parameters is written restore/maximum, where restore is
5814 the number of bytes (not bits - modem and network speeds are usually
5815 quoted in bits) per second placed into the bucket, and maximum is the
5816 maximum number of bytes which can be in the bucket at any time.
5818 There must be one delay_parameters line for each delay pool.
5821 For example, if delay pool number 1 is a class 2 delay pool as in the
5822 above example, and is being used to strictly limit each host to 64Kbit/sec
5823 (plus overheads), with no overall limit, the line is:
5825 delay_parameters 1 -1/-1 8000/8000
5827 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5829 Note that the figure -1 is used to represent "unlimited".
5832 And, if delay pool number 2 is a class 3 delay pool as in the above
5833 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5834 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5835 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5836 to permit a decent web page to be downloaded at a decent speed
5837 (if the network is not being limited due to overuse) but slow down
5838 large downloads more significantly:
5840 delay_parameters 2 32000/32000 8000/8000 600/8000
5842 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5843 8 x 8000 KByte/sec -> 64Kbit/sec.
5844 8 x 600 Byte/sec -> 4800bit/sec.
5847 Finally, for a class 4 delay pool as in the example - each user will
5848 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5850 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5853 NAME: delay_initial_bucket_level
5854 COMMENT: (percent, 0-100)
5857 IFDEF: USE_DELAY_POOLS
5858 LOC: Config.Delay.initial
5860 The initial bucket percentage is used to determine how much is put
5861 in each bucket when squid starts, is reconfigured, or first notices
5862 a host accessing it (in class 2 and class 3, individual hosts and
5863 networks only have buckets associated with them once they have been
5868 CLIENT DELAY POOL PARAMETERS
5869 -----------------------------------------------------------------------------
5872 NAME: client_delay_pools
5873 TYPE: client_delay_pool_count
5875 IFDEF: USE_DELAY_POOLS
5876 LOC: Config.ClientDelay
5878 This option specifies the number of client delay pools used. It must
5879 preceed other client_delay_* options.
5882 client_delay_pools 2
5885 NAME: client_delay_initial_bucket_level
5886 COMMENT: (percent, 0-no_limit)
5889 IFDEF: USE_DELAY_POOLS
5890 LOC: Config.ClientDelay.initial
5892 This option determines the initial bucket size as a percentage of
5893 max_bucket_size from client_delay_parameters. Buckets are created
5894 at the time of the "first" connection from the matching IP. Idle
5895 buckets are periodically deleted up.
5897 You can specify more than 100 percent but note that such "oversized"
5898 buckets are not refilled until their size goes down to max_bucket_size
5899 from client_delay_parameters.
5902 client_delay_initial_bucket_level 50
5905 NAME: client_delay_parameters
5906 TYPE: client_delay_pool_rates
5908 IFDEF: USE_DELAY_POOLS
5909 LOC: Config.ClientDelay
5912 This option configures client-side bandwidth limits using the
5915 client_delay_parameters pool speed_limit max_bucket_size
5917 pool is an integer ID used for client_delay_access matching.
5919 speed_limit is bytes added to the bucket per second.
5921 max_bucket_size is the maximum size of a bucket, enforced after any
5922 speed_limit additions.
5924 Please see the delay_parameters option for more information and
5928 client_delay_parameters 1 1024 2048
5929 client_delay_parameters 2 51200 16384
5932 NAME: client_delay_access
5933 TYPE: client_delay_pool_access
5935 IFDEF: USE_DELAY_POOLS
5936 LOC: Config.ClientDelay
5939 This option determines the client-side delay pool for the
5942 client_delay_access pool_ID allow|deny acl_name
5944 All client_delay_access options are checked in their pool ID
5945 order, starting with pool 1. The first checked pool with allowed
5946 request is selected for the request. If no ACL matches or there
5947 are no client_delay_access options, the request bandwidth is not
5950 The ACL-selected pool is then used to find the
5951 client_delay_parameters for the request. Client-side pools are
5952 not used to aggregate clients. Clients are always aggregated
5953 based on their source IP addresses (one bucket per source IP).
5955 Please see delay_access for more examples.
5958 client_delay_access 1 allow low_rate_network
5959 client_delay_access 2 allow vips_network
5963 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5964 -----------------------------------------------------------------------------
5969 LOC: Config.Wccp.router
5973 Use this option to define your WCCP ``home'' router for
5976 wccp_router supports a single WCCP(v1) router
5978 wccp2_router supports multiple WCCPv2 routers
5980 only one of the two may be used at the same time and defines
5981 which version of WCCP to use.
5985 TYPE: IpAddress_list
5986 LOC: Config.Wccp2.router
5990 Use this option to define your WCCP ``home'' router for
5993 wccp_router supports a single WCCP(v1) router
5995 wccp2_router supports multiple WCCPv2 routers
5997 only one of the two may be used at the same time and defines
5998 which version of WCCP to use.
6003 LOC: Config.Wccp.version
6007 This directive is only relevant if you need to set up WCCP(v1)
6008 to some very old and end-of-life Cisco routers. In all other
6009 setups it must be left unset or at the default setting.
6010 It defines an internal version in the WCCP(v1) protocol,
6011 with version 4 being the officially documented protocol.
6013 According to some users, Cisco IOS 11.2 and earlier only
6014 support WCCP version 3. If you're using that or an earlier
6015 version of IOS, you may need to change this value to 3, otherwise
6016 do not specify this parameter.
6019 NAME: wccp2_rebuild_wait
6021 LOC: Config.Wccp2.rebuildwait
6025 If this is enabled Squid will wait for the cache dir rebuild to finish
6026 before sending the first wccp2 HereIAm packet
6029 NAME: wccp2_forwarding_method
6031 LOC: Config.Wccp2.forwarding_method
6035 WCCP2 allows the setting of forwarding methods between the
6036 router/switch and the cache. Valid values are as follows:
6038 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6039 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6041 Currently (as of IOS 12.4) cisco routers only support GRE.
6042 Cisco switches only support the L2 redirect assignment method.
6045 NAME: wccp2_return_method
6047 LOC: Config.Wccp2.return_method
6051 WCCP2 allows the setting of return methods between the
6052 router/switch and the cache for packets that the cache
6053 decides not to handle. Valid values are as follows:
6055 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
6056 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
6058 Currently (as of IOS 12.4) cisco routers only support GRE.
6059 Cisco switches only support the L2 redirect assignment.
6061 If the "ip wccp redirect exclude in" command has been
6062 enabled on the cache interface, then it is still safe for
6063 the proxy server to use a l2 redirect method even if this
6064 option is set to GRE.
6067 NAME: wccp2_assignment_method
6069 LOC: Config.Wccp2.assignment_method
6073 WCCP2 allows the setting of methods to assign the WCCP hash
6074 Valid values are as follows:
6076 hash - Hash assignment
6077 mask - Mask assignment
6079 As a general rule, cisco routers support the hash assignment method
6080 and cisco switches support the mask assignment method.
6085 LOC: Config.Wccp2.info
6086 DEFAULT_IF_NONE: standard 0
6089 WCCP2 allows for multiple traffic services. There are two
6090 types: "standard" and "dynamic". The standard type defines
6091 one service id - http (id 0). The dynamic service ids can be from
6092 51 to 255 inclusive. In order to use a dynamic service id
6093 one must define the type of traffic to be redirected; this is done
6094 using the wccp2_service_info option.
6096 The "standard" type does not require a wccp2_service_info option,
6097 just specifying the service id will suffice.
6099 MD5 service authentication can be enabled by adding
6100 "password=<password>" to the end of this service declaration.
6104 wccp2_service standard 0 # for the 'web-cache' standard service
6105 wccp2_service dynamic 80 # a dynamic service type which will be
6106 # fleshed out with subsequent options.
6107 wccp2_service standard 0 password=foo
6110 NAME: wccp2_service_info
6111 TYPE: wccp2_service_info
6112 LOC: Config.Wccp2.info
6116 Dynamic WCCPv2 services require further information to define the
6117 traffic you wish to have diverted.
6121 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
6122 priority=<priority> ports=<port>,<port>..
6124 The relevant WCCPv2 flags:
6125 + src_ip_hash, dst_ip_hash
6126 + source_port_hash, dst_port_hash
6127 + src_ip_alt_hash, dst_ip_alt_hash
6128 + src_port_alt_hash, dst_port_alt_hash
6131 The port list can be one to eight entries.
6135 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6136 priority=240 ports=80
6138 Note: the service id must have been defined by a previous
6139 'wccp2_service dynamic <id>' entry.
6144 LOC: Config.Wccp2.weight
6148 Each cache server gets assigned a set of the destination
6149 hash proportional to their weight.
6154 LOC: Config.Wccp.address
6161 LOC: Config.Wccp2.address
6165 Use this option if you require WCCP to use a specific
6168 The default behavior is to not bind to any specific address.
6172 PERSISTENT CONNECTION HANDLING
6173 -----------------------------------------------------------------------------
6175 Also see "pconn_timeout" in the TIMEOUTS section
6178 NAME: client_persistent_connections
6180 LOC: Config.onoff.client_pconns
6184 NAME: server_persistent_connections
6186 LOC: Config.onoff.server_pconns
6189 Persistent connection support for clients and servers. By
6190 default, Squid uses persistent connections (when allowed)
6191 with its clients and servers. You can use these options to
6192 disable persistent connections with clients and/or servers.
6195 NAME: persistent_connection_after_error
6197 LOC: Config.onoff.error_pconns
6200 With this directive the use of persistent connections after
6201 HTTP errors can be disabled. Useful if you have clients
6202 who fail to handle errors on persistent connections proper.
6205 NAME: detect_broken_pconn
6207 LOC: Config.onoff.detect_broken_server_pconns
6210 Some servers have been found to incorrectly signal the use
6211 of HTTP/1.0 persistent connections even on replies not
6212 compatible, causing significant delays. This server problem
6213 has mostly been seen on redirects.
6215 By enabling this directive Squid attempts to detect such
6216 broken replies and automatically assume the reply is finished
6217 after 10 seconds timeout.
6221 CACHE DIGEST OPTIONS
6222 -----------------------------------------------------------------------------
6225 NAME: digest_generation
6226 IFDEF: USE_CACHE_DIGESTS
6228 LOC: Config.onoff.digest_generation
6231 This controls whether the server will generate a Cache Digest
6232 of its contents. By default, Cache Digest generation is
6233 enabled if Squid is compiled with --enable-cache-digests defined.
6236 NAME: digest_bits_per_entry
6237 IFDEF: USE_CACHE_DIGESTS
6239 LOC: Config.digest.bits_per_entry
6242 This is the number of bits of the server's Cache Digest which
6243 will be associated with the Digest entry for a given HTTP
6244 Method and URL (public key) combination. The default is 5.
6247 NAME: digest_rebuild_period
6248 IFDEF: USE_CACHE_DIGESTS
6251 LOC: Config.digest.rebuild_period
6254 This is the wait time between Cache Digest rebuilds.
6257 NAME: digest_rewrite_period
6259 IFDEF: USE_CACHE_DIGESTS
6261 LOC: Config.digest.rewrite_period
6264 This is the wait time between Cache Digest writes to
6268 NAME: digest_swapout_chunk_size
6271 IFDEF: USE_CACHE_DIGESTS
6272 LOC: Config.digest.swapout_chunk_size
6275 This is the number of bytes of the Cache Digest to write to
6276 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6280 NAME: digest_rebuild_chunk_percentage
6281 COMMENT: (percent, 0-100)
6282 IFDEF: USE_CACHE_DIGESTS
6284 LOC: Config.digest.rebuild_chunk_percentage
6287 This is the percentage of the Cache Digest to be scanned at a
6288 time. By default it is set to 10% of the Cache Digest.
6293 -----------------------------------------------------------------------------
6298 LOC: Config.Port.snmp
6302 The port number where Squid listens for SNMP requests. To enable
6303 SNMP support set this to a suitable port number. Port number
6304 3401 is often used for the Squid SNMP agent. By default it's
6305 set to "0" (disabled)
6313 LOC: Config.accessList.snmp
6314 DEFAULT_IF_NONE: deny all
6317 Allowing or denying access to the SNMP port.
6319 All access to the agent is denied by default.
6322 snmp_access allow|deny [!]aclname ...
6324 This clause only supports fast acl types.
6325 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6327 snmp_access allow snmppublic localhost
6328 snmp_access deny all
6331 NAME: snmp_incoming_address
6333 LOC: Config.Addrs.snmp_incoming
6338 NAME: snmp_outgoing_address
6340 LOC: Config.Addrs.snmp_outgoing
6344 Just like 'udp_incoming_address', but for the SNMP port.
6346 snmp_incoming_address is used for the SNMP socket receiving
6347 messages from SNMP agents.
6348 snmp_outgoing_address is used for SNMP packets returned to SNMP
6351 The default snmp_incoming_address is to listen on all
6352 available network interfaces.
6354 If snmp_outgoing_address is not set it will use the same socket
6355 as snmp_incoming_address. Only change this if you want to have
6356 SNMP replies sent using another address than where this Squid
6357 listens for SNMP queries.
6359 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6360 the same value since they both use port 3401.
6365 -----------------------------------------------------------------------------
6368 NAME: icp_port udp_port
6371 LOC: Config.Port.icp
6373 The port number where Squid sends and receives ICP queries to
6374 and from neighbor caches. The standard UDP port for ICP is 3130.
6375 Default is disabled (0).
6378 icp_port @DEFAULT_ICP_PORT@
6385 LOC: Config.Port.htcp
6387 The port number where Squid sends and receives HTCP queries to
6388 and from neighbor caches. To turn it on you want to set it to
6389 4827. By default it is set to "0" (disabled).
6395 NAME: log_icp_queries
6399 LOC: Config.onoff.log_udp
6401 If set, ICP queries are logged to access.log. You may wish
6402 do disable this if your ICP load is VERY high to speed things
6403 up or to simplify log analysis.
6406 NAME: udp_incoming_address
6408 LOC:Config.Addrs.udp_incoming
6411 udp_incoming_address is used for UDP packets received from other
6414 The default behavior is to not bind to any specific address.
6416 Only change this if you want to have all UDP queries received on
6417 a specific interface/address.
6419 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6420 modules. Altering it will affect all of them in the same manner.
6422 see also; udp_outgoing_address
6424 NOTE, udp_incoming_address and udp_outgoing_address can not
6425 have the same value since they both use the same port.
6428 NAME: udp_outgoing_address
6430 LOC: Config.Addrs.udp_outgoing
6433 udp_outgoing_address is used for UDP packets sent out to other
6436 The default behavior is to not bind to any specific address.
6438 Instead it will use the same socket as udp_incoming_address.
6439 Only change this if you want to have UDP queries sent using another
6440 address than where this Squid listens for UDP queries from other
6443 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6444 modules. Altering it will affect all of them in the same manner.
6446 see also; udp_incoming_address
6448 NOTE, udp_incoming_address and udp_outgoing_address can not
6449 have the same value since they both use the same port.
6456 LOC: Config.onoff.icp_hit_stale
6458 If you want to return ICP_HIT for stale cache objects, set this
6459 option to 'on'. If you have sibling relationships with caches
6460 in other administrative domains, this should be 'off'. If you only
6461 have sibling relationships with caches under your control,
6462 it is probably okay to set this to 'on'.
6463 If set to 'on', your siblings should use the option "allow-miss"
6464 on their cache_peer lines for connecting to you.
6467 NAME: minimum_direct_hops
6470 LOC: Config.minDirectHops
6472 If using the ICMP pinging stuff, do direct fetches for sites
6473 which are no more than this many hops away.
6476 NAME: minimum_direct_rtt
6479 LOC: Config.minDirectRtt
6481 If using the ICMP pinging stuff, do direct fetches for sites
6482 which are no more than this many rtt milliseconds away.
6488 LOC: Config.Netdb.low
6494 LOC: Config.Netdb.high
6496 The low and high water marks for the ICMP measurement
6497 database. These are counts, not percents. The defaults are
6498 900 and 1000. When the high water mark is reached, database
6499 entries will be deleted until the low mark is reached.
6502 NAME: netdb_ping_period
6504 LOC: Config.Netdb.period
6507 The minimum period for measuring a site. There will be at
6508 least this much delay between successive pings to the same
6509 network. The default is five minutes.
6516 LOC: Config.onoff.query_icmp
6518 If you want to ask your peers to include ICMP data in their ICP
6519 replies, enable this option.
6521 If your peer has configured Squid (during compilation) with
6522 '--enable-icmp' that peer will send ICMP pings to origin server
6523 sites of the URLs it receives. If you enable this option the
6524 ICP replies from that peer will include the ICMP data (if available).
6525 Then, when choosing a parent cache, Squid will choose the parent with
6526 the minimal RTT to the origin server. When this happens, the
6527 hierarchy field of the access.log will be
6528 "CLOSEST_PARENT_MISS". This option is off by default.
6531 NAME: test_reachability
6535 LOC: Config.onoff.test_reachability
6537 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6538 instead of ICP_MISS if the target host is NOT in the ICMP
6539 database, or has a zero RTT.
6542 NAME: icp_query_timeout
6546 LOC: Config.Timeout.icp_query
6548 Normally Squid will automatically determine an optimal ICP
6549 query timeout value based on the round-trip-time of recent ICP
6550 queries. If you want to override the value determined by
6551 Squid, set this 'icp_query_timeout' to a non-zero value. This
6552 value is specified in MILLISECONDS, so, to use a 2-second
6553 timeout (the old default), you would write:
6555 icp_query_timeout 2000
6558 NAME: maximum_icp_query_timeout
6562 LOC: Config.Timeout.icp_query_max
6564 Normally the ICP query timeout is determined dynamically. But
6565 sometimes it can lead to very large values (say 5 seconds).
6566 Use this option to put an upper limit on the dynamic timeout
6567 value. Do NOT use this option to always use a fixed (instead
6568 of a dynamic) timeout value. To set a fixed timeout see the
6569 'icp_query_timeout' directive.
6572 NAME: minimum_icp_query_timeout
6576 LOC: Config.Timeout.icp_query_min
6578 Normally the ICP query timeout is determined dynamically. But
6579 sometimes it can lead to very small timeouts, even lower than
6580 the normal latency variance on your link due to traffic.
6581 Use this option to put an lower limit on the dynamic timeout
6582 value. Do NOT use this option to always use a fixed (instead
6583 of a dynamic) timeout value. To set a fixed timeout see the
6584 'icp_query_timeout' directive.
6587 NAME: background_ping_rate
6591 LOC: Config.backgroundPingRate
6593 Controls how often the ICP pings are sent to siblings that
6594 have background-ping set.
6598 MULTICAST ICP OPTIONS
6599 -----------------------------------------------------------------------------
6604 LOC: Config.mcast_group_list
6607 This tag specifies a list of multicast groups which your server
6608 should join to receive multicasted ICP queries.
6610 NOTE! Be very careful what you put here! Be sure you
6611 understand the difference between an ICP _query_ and an ICP
6612 _reply_. This option is to be set only if you want to RECEIVE
6613 multicast queries. Do NOT set this option to SEND multicast
6614 ICP (use cache_peer for that). ICP replies are always sent via
6615 unicast, so this option does not affect whether or not you will
6616 receive replies from multicast group members.
6618 You must be very careful to NOT use a multicast address which
6619 is already in use by another group of caches.
6621 If you are unsure about multicast, please read the Multicast
6622 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6624 Usage: mcast_groups 239.128.16.128 224.0.1.20
6626 By default, Squid doesn't listen on any multicast groups.
6629 NAME: mcast_miss_addr
6630 IFDEF: MULTICAST_MISS_STREAM
6632 LOC: Config.mcast_miss.addr
6635 If you enable this option, every "cache miss" URL will
6636 be sent out on the specified multicast address.
6638 Do not enable this option unless you are are absolutely
6639 certain you understand what you are doing.
6642 NAME: mcast_miss_ttl
6643 IFDEF: MULTICAST_MISS_STREAM
6645 LOC: Config.mcast_miss.ttl
6648 This is the time-to-live value for packets multicasted
6649 when multicasting off cache miss URLs is enabled. By
6650 default this is set to 'site scope', i.e. 16.
6653 NAME: mcast_miss_port
6654 IFDEF: MULTICAST_MISS_STREAM
6656 LOC: Config.mcast_miss.port
6659 This is the port number to be used in conjunction with
6663 NAME: mcast_miss_encode_key
6664 IFDEF: MULTICAST_MISS_STREAM
6666 LOC: Config.mcast_miss.encode_key
6667 DEFAULT: XXXXXXXXXXXXXXXX
6669 The URLs that are sent in the multicast miss stream are
6670 encrypted. This is the encryption key.
6673 NAME: mcast_icp_query_timeout
6677 LOC: Config.Timeout.mcast_icp_query
6679 For multicast peers, Squid regularly sends out ICP "probes" to
6680 count how many other peers are listening on the given multicast
6681 address. This value specifies how long Squid should wait to
6682 count all the replies. The default is 2000 msec, or 2
6687 INTERNAL ICON OPTIONS
6688 -----------------------------------------------------------------------------
6691 NAME: icon_directory
6693 LOC: Config.icons.directory
6694 DEFAULT: @DEFAULT_ICON_DIR@
6696 Where the icons are stored. These are normally kept in
6700 NAME: global_internal_static
6702 LOC: Config.onoff.global_internal_static
6705 This directive controls is Squid should intercept all requests for
6706 /squid-internal-static/ no matter which host the URL is requesting
6707 (default on setting), or if nothing special should be done for
6708 such URLs (off setting). The purpose of this directive is to make
6709 icons etc work better in complex cache hierarchies where it may
6710 not always be possible for all corners in the cache mesh to reach
6711 the server generating a directory listing.
6714 NAME: short_icon_urls
6716 LOC: Config.icons.use_short_names
6719 If this is enabled Squid will use short URLs for icons.
6720 If disabled it will revert to the old behavior of including
6721 it's own name and port in the URL.
6723 If you run a complex cache hierarchy with a mix of Squid and
6724 other proxies you may need to disable this directive.
6729 -----------------------------------------------------------------------------
6732 NAME: error_directory
6734 LOC: Config.errorDirectory
6737 If you wish to create your own versions of the default
6738 error files to customize them to suit your company copy
6739 the error/template files to another directory and point
6742 WARNING: This option will disable multi-language support
6743 on error pages if used.
6745 The squid developers are interested in making squid available in
6746 a wide variety of languages. If you are making translations for a
6747 language that Squid does not currently provide please consider
6748 contributing your translation back to the project.
6749 http://wiki.squid-cache.org/Translations
6751 The squid developers working on translations are happy to supply drop-in
6752 translated error files in exchange for any new language contributions.
6755 NAME: error_default_language
6756 IFDEF: USE_ERR_LOCALES
6758 LOC: Config.errorDefaultLanguage
6761 Set the default language which squid will send error pages in
6762 if no existing translation matches the clients language
6765 If unset (default) generic English will be used.
6767 The squid developers are interested in making squid available in
6768 a wide variety of languages. If you are interested in making
6769 translations for any language see the squid wiki for details.
6770 http://wiki.squid-cache.org/Translations
6773 NAME: error_log_languages
6774 IFDEF: USE_ERR_LOCALES
6776 LOC: Config.errorLogMissingLanguages
6779 Log to cache.log what languages users are attempting to
6780 auto-negotiate for translations.
6782 Successful negotiations are not logged. Only failures
6783 have meaning to indicate that Squid may need an upgrade
6784 of its error page translations.
6787 NAME: err_page_stylesheet
6789 LOC: Config.errorStylesheet
6790 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6792 CSS Stylesheet to pattern the display of Squid default error pages.
6794 For information on CSS see http://www.w3.org/Style/CSS/
6799 LOC: Config.errHtmlText
6802 HTML text to include in error messages. Make this a "mailto"
6803 URL to your admin address, or maybe just a link to your
6804 organizations Web page.
6806 To include this in your error messages, you must rewrite
6807 the error template files (found in the "errors" directory).
6808 Wherever you want the 'err_html_text' line to appear,
6809 insert a %L tag in the error template file.
6812 NAME: email_err_data
6815 LOC: Config.onoff.emailErrData
6818 If enabled, information about the occurred error will be
6819 included in the mailto links of the ERR pages (if %W is set)
6820 so that the email body contains the data.
6821 Syntax is <A HREF="mailto:%w%W">%w</A>
6826 LOC: Config.denyInfoList
6829 Usage: deny_info err_page_name acl
6830 or deny_info http://... acl
6831 or deny_info TCP_RESET acl
6833 This can be used to return a ERR_ page for requests which
6834 do not pass the 'http_access' rules. Squid remembers the last
6835 acl it evaluated in http_access, and if a 'deny_info' line exists
6836 for that ACL Squid returns a corresponding error page.
6838 The acl is typically the last acl on the http_access deny line which
6839 denied access. The exceptions to this rule are:
6840 - When Squid needs to request authentication credentials. It's then
6841 the first authentication related acl encountered
6842 - When none of the http_access lines matches. It's then the last
6843 acl processed on the last http_access line.
6844 - When the decision to deny access was made by an adaptation service,
6845 the acl name is the corresponding eCAP or ICAP service_name.
6847 NP: If providing your own custom error pages with error_directory
6848 you may also specify them by your custom file name:
6849 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6851 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6852 may be specified by prefixing the file name with the code and a colon.
6853 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6855 Alternatively you can tell Squid to reset the TCP connection
6856 by specifying TCP_RESET.
6858 Or you can specify an error URL or URL pattern. The browsers will
6859 get redirected to the specified URL after formatting tags have
6860 been replaced. Redirect will be done with 302 or 307 according to
6861 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6862 the URL. e.g. 303:http://example.com/
6865 %a - username (if available. Password NOT included)
6868 %E - Error description
6870 %H - Request domain name
6871 %i - Client IP Address
6873 %o - Message result from external ACL helper
6874 %p - Request Port number
6875 %P - Request Protocol name
6876 %R - Request URL path
6877 %T - Timestamp in RFC 1123 format
6878 %U - Full canonical URL from client
6879 (HTTPS URLs terminate with *)
6880 %u - Full canonical URL from client
6881 %w - Admin email from squid.conf
6883 %% - Literal percent (%) code
6888 OPTIONS INFLUENCING REQUEST FORWARDING
6889 -----------------------------------------------------------------------------
6892 NAME: nonhierarchical_direct
6894 LOC: Config.onoff.nonhierarchical_direct
6897 By default, Squid will send any non-hierarchical requests
6898 (matching hierarchy_stoplist or not cacheable request type) direct
6901 If you set this to off, Squid will prefer to send these
6902 requests to parents.
6904 Note that in most configurations, by turning this off you will only
6905 add latency to these request without any improvement in global hit
6908 If you are inside an firewall see never_direct instead of
6914 LOC: Config.onoff.prefer_direct
6917 Normally Squid tries to use parents for most requests. If you for some
6918 reason like it to first try going direct and only use a parent if
6919 going direct fails set this to on.
6921 By combining nonhierarchical_direct off and prefer_direct on you
6922 can set up Squid to use a parent as a backup path if going direct
6925 Note: If you want Squid to use parents for all requests see
6926 the never_direct directive. prefer_direct only modifies how Squid
6927 acts on cacheable requests.
6932 LOC: Config.accessList.AlwaysDirect
6935 Usage: always_direct allow|deny [!]aclname ...
6937 Here you can use ACL elements to specify requests which should
6938 ALWAYS be forwarded by Squid to the origin servers without using
6939 any peers. For example, to always directly forward requests for
6940 local servers ignoring any parents or siblings you may have use
6943 acl local-servers dstdomain my.domain.net
6944 always_direct allow local-servers
6946 To always forward FTP requests directly, use
6949 always_direct allow FTP
6951 NOTE: There is a similar, but opposite option named
6952 'never_direct'. You need to be aware that "always_direct deny
6953 foo" is NOT the same thing as "never_direct allow foo". You
6954 may need to use a deny rule to exclude a more-specific case of
6955 some other rule. Example:
6957 acl local-external dstdomain external.foo.net
6958 acl local-servers dstdomain .foo.net
6959 always_direct deny local-external
6960 always_direct allow local-servers
6962 NOTE: If your goal is to make the client forward the request
6963 directly to the origin server bypassing Squid then this needs
6964 to be done in the client configuration. Squid configuration
6965 can only tell Squid how Squid should fetch the object.
6967 NOTE: This directive is not related to caching. The replies
6968 is cached as usual even if you use always_direct. To not cache
6969 the replies see the 'cache' directive.
6971 This clause supports both fast and slow acl types.
6972 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6977 LOC: Config.accessList.NeverDirect
6980 Usage: never_direct allow|deny [!]aclname ...
6982 never_direct is the opposite of always_direct. Please read
6983 the description for always_direct if you have not already.
6985 With 'never_direct' you can use ACL elements to specify
6986 requests which should NEVER be forwarded directly to origin
6987 servers. For example, to force the use of a proxy for all
6988 requests, except those in your local domain use something like:
6990 acl local-servers dstdomain .foo.net
6991 never_direct deny local-servers
6992 never_direct allow all
6994 or if Squid is inside a firewall and there are local intranet
6995 servers inside the firewall use something like:
6997 acl local-intranet dstdomain .foo.net
6998 acl local-external dstdomain external.foo.net
6999 always_direct deny local-external
7000 always_direct allow local-intranet
7001 never_direct allow all
7003 This clause supports both fast and slow acl types.
7004 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
7008 ADVANCED NETWORKING OPTIONS
7009 -----------------------------------------------------------------------------
7012 NAME: incoming_udp_average incoming_icp_average
7015 LOC: Config.comm_incoming.udp.average
7017 Heavy voodoo here. I can't even believe you are reading this.
7018 Are you crazy? Don't even think about adjusting these unless
7019 you understand the algorithms in comm_select.c first!
7022 NAME: incoming_tcp_average incoming_http_average
7025 LOC: Config.comm_incoming.tcp.average
7027 Heavy voodoo here. I can't even believe you are reading this.
7028 Are you crazy? Don't even think about adjusting these unless
7029 you understand the algorithms in comm_select.c first!
7032 NAME: incoming_dns_average
7035 LOC: Config.comm_incoming.dns.average
7037 Heavy voodoo here. I can't even believe you are reading this.
7038 Are you crazy? Don't even think about adjusting these unless
7039 you understand the algorithms in comm_select.c first!
7042 NAME: min_udp_poll_cnt min_icp_poll_cnt
7045 LOC: Config.comm_incoming.udp.min_poll
7047 Heavy voodoo here. I can't even believe you are reading this.
7048 Are you crazy? Don't even think about adjusting these unless
7049 you understand the algorithms in comm_select.c first!
7052 NAME: min_dns_poll_cnt
7055 LOC: Config.comm_incoming.dns.min_poll
7057 Heavy voodoo here. I can't even believe you are reading this.
7058 Are you crazy? Don't even think about adjusting these unless
7059 you understand the algorithms in comm_select.c first!
7062 NAME: min_tcp_poll_cnt min_http_poll_cnt
7065 LOC: Config.comm_incoming.tcp.min_poll
7067 Heavy voodoo here. I can't even believe you are reading this.
7068 Are you crazy? Don't even think about adjusting these unless
7069 you understand the algorithms in comm_select.c first!
7075 LOC: Config.accept_filter
7079 The name of an accept(2) filter to install on Squid's
7080 listen socket(s). This feature is perhaps specific to
7081 FreeBSD and requires support in the kernel.
7083 The 'httpready' filter delays delivering new connections
7084 to Squid until a full HTTP request has been received.
7085 See the accf_http(9) man page for details.
7087 The 'dataready' filter delays delivering new connections
7088 to Squid until there is some data to process.
7089 See the accf_dataready(9) man page for details.
7093 The 'data' filter delays delivering of new connections
7094 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
7095 You may optionally specify a number of seconds to wait by
7096 'data=N' where N is the number of seconds. Defaults to 30
7097 if not specified. See the tcp(7) man page for details.
7100 accept_filter httpready
7105 NAME: client_ip_max_connections
7107 LOC: Config.client_ip_max_connections
7110 Set an absolute limit on the number of connections a single
7111 client IP can use. Any more than this and Squid will begin to drop
7112 new connections from the client until it closes some links.
7114 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
7115 connections from the client. For finer control use the ACL access controls.
7117 Requires client_db to be enabled (the default).
7119 WARNING: This may noticably slow down traffic received via external proxies
7120 or NAT devices and cause them to rebound error messages back to their clients.
7123 NAME: tcp_recv_bufsize
7127 LOC: Config.tcpRcvBufsz
7129 Size of receive buffer to set for TCP sockets. Probably just
7130 as easy to change your kernel's default. Set to zero to use
7131 the default buffer size.
7136 -----------------------------------------------------------------------------
7143 LOC: Adaptation::Icap::TheConfig.onoff
7146 If you want to enable the ICAP module support, set this to on.
7149 NAME: icap_connect_timeout
7152 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7155 This parameter specifies how long to wait for the TCP connect to
7156 the requested ICAP server to complete before giving up and either
7157 terminating the HTTP transaction or bypassing the failure.
7159 The default for optional services is peer_connect_timeout.
7160 The default for essential services is connect_timeout.
7161 If this option is explicitly set, its value applies to all services.
7164 NAME: icap_io_timeout
7168 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7171 This parameter specifies how long to wait for an I/O activity on
7172 an established, active ICAP connection before giving up and
7173 either terminating the HTTP transaction or bypassing the
7176 The default is read_timeout.
7179 NAME: icap_service_failure_limit
7180 COMMENT: limit [in memory-depth time-units]
7181 TYPE: icap_service_failure_limit
7183 LOC: Adaptation::Icap::TheConfig
7186 The limit specifies the number of failures that Squid tolerates
7187 when establishing a new TCP connection with an ICAP service. If
7188 the number of failures exceeds the limit, the ICAP service is
7189 not used for new ICAP requests until it is time to refresh its
7192 A negative value disables the limit. Without the limit, an ICAP
7193 service will not be considered down due to connectivity failures
7194 between ICAP OPTIONS requests.
7196 Squid forgets ICAP service failures older than the specified
7197 value of memory-depth. The memory fading algorithm
7198 is approximate because Squid does not remember individual
7199 errors but groups them instead, splitting the option
7200 value into ten time slots of equal length.
7202 When memory-depth is 0 and by default this option has no
7203 effect on service failure expiration.
7205 Squid always forgets failures when updating service settings
7206 using an ICAP OPTIONS transaction, regardless of this option
7210 # suspend service usage after 10 failures in 5 seconds:
7211 icap_service_failure_limit 10 in 5 seconds
7214 NAME: icap_service_revival_delay
7217 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7220 The delay specifies the number of seconds to wait after an ICAP
7221 OPTIONS request failure before requesting the options again. The
7222 failed ICAP service is considered "down" until fresh OPTIONS are
7225 The actual delay cannot be smaller than the hardcoded minimum
7226 delay of 30 seconds.
7229 NAME: icap_preview_enable
7233 LOC: Adaptation::Icap::TheConfig.preview_enable
7236 The ICAP Preview feature allows the ICAP server to handle the
7237 HTTP message by looking only at the beginning of the message body
7238 or even without receiving the body at all. In some environments,
7239 previews greatly speedup ICAP processing.
7241 During an ICAP OPTIONS transaction, the server may tell Squid what
7242 HTTP messages should be previewed and how big the preview should be.
7243 Squid will not use Preview if the server did not request one.
7245 To disable ICAP Preview for all ICAP services, regardless of
7246 individual ICAP server OPTIONS responses, set this option to "off".
7248 icap_preview_enable off
7251 NAME: icap_preview_size
7254 LOC: Adaptation::Icap::TheConfig.preview_size
7257 The default size of preview data to be sent to the ICAP server.
7258 -1 means no preview. This value might be overwritten on a per server
7259 basis by OPTIONS requests.
7262 NAME: icap_206_enable
7266 LOC: Adaptation::Icap::TheConfig.allow206_enable
7269 206 (Partial Content) responses is an ICAP extension that allows the
7270 ICAP agents to optionally combine adapted and original HTTP message
7271 content. The decision to combine is postponed until the end of the
7272 ICAP response. Squid supports Partial Content extension by default.
7274 Activation of the Partial Content extension is negotiated with each
7275 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7276 negotation correctly even if they do not support the extension, but
7277 some might fail. To disable Partial Content support for all ICAP
7278 services and to avoid any negotiation, set this option to "off".
7284 NAME: icap_default_options_ttl
7287 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7290 The default TTL value for ICAP OPTIONS responses that don't have
7291 an Options-TTL header.
7294 NAME: icap_persistent_connections
7298 LOC: Adaptation::Icap::TheConfig.reuse_connections
7301 Whether or not Squid should use persistent connections to
7305 NAME: adaptation_send_client_ip icap_send_client_ip
7307 IFDEF: USE_ADAPTATION
7309 LOC: Adaptation::Config::send_client_ip
7312 If enabled, Squid shares HTTP client IP information with adaptation
7313 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7314 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7316 See also: adaptation_uses_indirect_client
7319 NAME: adaptation_send_username icap_send_client_username
7321 IFDEF: USE_ADAPTATION
7323 LOC: Adaptation::Config::send_username
7326 This sends authenticated HTTP client username (if available) to
7327 the adaptation service.
7329 For ICAP, the username value is encoded based on the
7330 icap_client_username_encode option and is sent using the header
7331 specified by the icap_client_username_header option.
7334 NAME: icap_client_username_header
7337 LOC: Adaptation::Icap::TheConfig.client_username_header
7338 DEFAULT: X-Client-Username
7340 ICAP request header name to use for adaptation_send_username.
7343 NAME: icap_client_username_encode
7347 LOC: Adaptation::Icap::TheConfig.client_username_encode
7350 Whether to base64 encode the authenticated client username.
7354 TYPE: icap_service_type
7356 LOC: Adaptation::Icap::TheConfig
7359 Defines a single ICAP service using the following format:
7361 icap_service id vectoring_point uri [option ...]
7364 an opaque identifier or name which is used to direct traffic to
7365 this specific service. Must be unique among all adaptation
7366 services in squid.conf.
7368 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7369 This specifies at which point of transaction processing the
7370 ICAP service should be activated. *_postcache vectoring points
7371 are not yet supported.
7373 uri: icap://servername:port/servicepath
7374 ICAP server and service location.
7376 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7377 transactions. Squid does not enforce that requirement. You can specify
7378 services with the same service_url and different vectoring_points. You
7379 can even specify multiple identical services as long as their
7380 service_names differ.
7383 Service options are separated by white space. ICAP services support
7384 the following name=value options:
7387 If set to 'on' or '1', the ICAP service is treated as
7388 optional. If the service cannot be reached or malfunctions,
7389 Squid will try to ignore any errors and process the message as
7390 if the service was not enabled. No all ICAP errors can be
7391 bypassed. If set to 0, the ICAP service is treated as
7392 essential and all ICAP errors will result in an error page
7393 returned to the HTTP client.
7395 Bypass is off by default: services are treated as essential.
7398 If set to 'on' or '1', the ICAP service is allowed to
7399 dynamically change the current message adaptation plan by
7400 returning a chain of services to be used next. The services
7401 are specified using the X-Next-Services ICAP response header
7402 value, formatted as a comma-separated list of service names.
7403 Each named service should be configured in squid.conf. Other
7404 services are ignored. An empty X-Next-Services value results
7405 in an empty plan which ends the current adaptation.
7407 Dynamic adaptation plan may cross or cover multiple supported
7408 vectoring points in their natural processing order.
7410 Routing is not allowed by default: the ICAP X-Next-Services
7411 response header is ignored.
7414 Only has effect on split-stack systems. The default on those systems
7415 is to use IPv4-only connections. When set to 'on' this option will
7416 make Squid use IPv6-only connections to contact this ICAP service.
7418 on-overload=block|bypass|wait|force
7419 If the service Max-Connections limit has been reached, do
7420 one of the following for each new ICAP transaction:
7421 * block: send an HTTP error response to the client
7422 * bypass: ignore the "over-connected" ICAP service
7423 * wait: wait (in a FIFO queue) for an ICAP connection slot
7424 * force: proceed, ignoring the Max-Connections limit
7426 In SMP mode with N workers, each worker assumes the service
7427 connection limit is Max-Connections/N, even though not all
7428 workers may use a given service.
7430 The default value is "bypass" if service is bypassable,
7431 otherwise it is set to "wait".
7435 Use the given number as the Max-Connections limit, regardless
7436 of the Max-Connections value given by the service, if any.
7438 Older icap_service format without optional named parameters is
7439 deprecated but supported for backward compatibility.
7442 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7443 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7447 TYPE: icap_class_type
7452 This deprecated option was documented to define an ICAP service
7453 chain, even though it actually defined a set of similar, redundant
7454 services, and the chains were not supported.
7456 To define a set of redundant services, please use the
7457 adaptation_service_set directive. For service chains, use
7458 adaptation_service_chain.
7462 TYPE: icap_access_type
7467 This option is deprecated. Please use adaptation_access, which
7468 has the same ICAP functionality, but comes with better
7469 documentation, and eCAP support.
7474 -----------------------------------------------------------------------------
7481 LOC: Adaptation::Ecap::TheConfig.onoff
7484 Controls whether eCAP support is enabled.
7488 TYPE: ecap_service_type
7490 LOC: Adaptation::Ecap::TheConfig
7493 Defines a single eCAP service
7495 ecap_service id vectoring_point uri [option ...]
7498 an opaque identifier or name which is used to direct traffic to
7499 this specific service. Must be unique among all adaptation
7500 services in squid.conf.
7502 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7503 This specifies at which point of transaction processing the
7504 eCAP service should be activated. *_postcache vectoring points
7505 are not yet supported.
7507 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7508 Squid uses the eCAP service URI to match this configuration
7509 line with one of the dynamically loaded services. Each loaded
7510 eCAP service must have a unique URI. Obtain the right URI from
7511 the service provider.
7514 Service options are separated by white space. eCAP services support
7515 the following name=value options:
7518 If set to 'on' or '1', the eCAP service is treated as optional.
7519 If the service cannot be reached or malfunctions, Squid will try
7520 to ignore any errors and process the message as if the service
7521 was not enabled. No all eCAP errors can be bypassed.
7522 If set to 'off' or '0', the eCAP service is treated as essential
7523 and all eCAP errors will result in an error page returned to the
7526 Bypass is off by default: services are treated as essential.
7529 If set to 'on' or '1', the eCAP service is allowed to
7530 dynamically change the current message adaptation plan by
7531 returning a chain of services to be used next.
7533 Dynamic adaptation plan may cross or cover multiple supported
7534 vectoring points in their natural processing order.
7536 Routing is not allowed by default.
7538 Older ecap_service format without optional named parameters is
7539 deprecated but supported for backward compatibility.
7543 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7544 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7547 NAME: loadable_modules
7549 IFDEF: USE_LOADABLE_MODULES
7550 LOC: Config.loadable_module_names
7553 Instructs Squid to load the specified dynamic module(s) or activate
7554 preloaded module(s).
7556 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7560 MESSAGE ADAPTATION OPTIONS
7561 -----------------------------------------------------------------------------
7564 NAME: adaptation_service_set
7565 TYPE: adaptation_service_set_type
7566 IFDEF: USE_ADAPTATION
7571 Configures an ordered set of similar, redundant services. This is
7572 useful when hot standby or backup adaptation servers are available.
7574 adaptation_service_set set_name service_name1 service_name2 ...
7576 The named services are used in the set declaration order. The first
7577 applicable adaptation service from the set is used first. The next
7578 applicable service is tried if and only if the transaction with the
7579 previous service fails and the message waiting to be adapted is still
7582 When adaptation starts, broken services are ignored as if they were
7583 not a part of the set. A broken service is a down optional service.
7585 The services in a set must be attached to the same vectoring point
7586 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7588 If all services in a set are optional then adaptation failures are
7589 bypassable. If all services in the set are essential, then a
7590 transaction failure with one service may still be retried using
7591 another service from the set, but when all services fail, the master
7592 transaction fails as well.
7594 A set may contain a mix of optional and essential services, but that
7595 is likely to lead to surprising results because broken services become
7596 ignored (see above), making previously bypassable failures fatal.
7597 Technically, it is the bypassability of the last failed service that
7600 See also: adaptation_access adaptation_service_chain
7603 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7604 adaptation service_set svcLogger loggerLocal loggerRemote
7607 NAME: adaptation_service_chain
7608 TYPE: adaptation_service_chain_type
7609 IFDEF: USE_ADAPTATION
7614 Configures a list of complementary services that will be applied
7615 one-by-one, forming an adaptation chain or pipeline. This is useful
7616 when Squid must perform different adaptations on the same message.
7618 adaptation_service_chain chain_name service_name1 svc_name2 ...
7620 The named services are used in the chain declaration order. The first
7621 applicable adaptation service from the chain is used first. The next
7622 applicable service is applied to the successful adaptation results of
7623 the previous service in the chain.
7625 When adaptation starts, broken services are ignored as if they were
7626 not a part of the chain. A broken service is a down optional service.
7628 Request satisfaction terminates the adaptation chain because Squid
7629 does not currently allow declaration of RESPMOD services at the
7630 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7632 The services in a chain must be attached to the same vectoring point
7633 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7635 A chain may contain a mix of optional and essential services. If an
7636 essential adaptation fails (or the failure cannot be bypassed for
7637 other reasons), the master transaction fails. Otherwise, the failure
7638 is bypassed as if the failed adaptation service was not in the chain.
7640 See also: adaptation_access adaptation_service_set
7643 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
7646 NAME: adaptation_access
7647 TYPE: adaptation_access_type
7648 IFDEF: USE_ADAPTATION
7652 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
7654 adaptation_access service_name allow|deny [!]aclname...
7655 adaptation_access set_name allow|deny [!]aclname...
7657 At each supported vectoring point, the adaptation_access
7658 statements are processed in the order they appear in this
7659 configuration file. Statements pointing to the following services
7660 are ignored (i.e., skipped without checking their ACL):
7662 - services serving different vectoring points
7663 - "broken-but-bypassable" services
7664 - "up" services configured to ignore such transactions
7665 (e.g., based on the ICAP Transfer-Ignore header).
7667 When a set_name is used, all services in the set are checked
7668 using the same rules, to find the first applicable one. See
7669 adaptation_service_set for details.
7671 If an access list is checked and there is a match, the
7672 processing stops: For an "allow" rule, the corresponding
7673 adaptation service is used for the transaction. For a "deny"
7674 rule, no adaptation service is activated.
7676 It is currently not possible to apply more than one adaptation
7677 service at the same vectoring point to the same HTTP transaction.
7679 See also: icap_service and ecap_service
7682 adaptation_access service_1 allow all
7685 NAME: adaptation_service_iteration_limit
7687 IFDEF: USE_ADAPTATION
7688 LOC: Adaptation::Config::service_iteration_limit
7691 Limits the number of iterations allowed when applying adaptation
7692 services to a message. If your longest adaptation set or chain
7693 may have more than 16 services, increase the limit beyond its
7694 default value of 16. If detecting infinite iteration loops sooner
7695 is critical, make the iteration limit match the actual number
7696 of services in your longest adaptation set or chain.
7698 Infinite adaptation loops are most likely with routing services.
7700 See also: icap_service routing=1
7703 NAME: adaptation_masterx_shared_names
7705 IFDEF: USE_ADAPTATION
7706 LOC: Adaptation::Config::masterx_shared_name
7709 For each master transaction (i.e., the HTTP request and response
7710 sequence, including all related ICAP and eCAP exchanges), Squid
7711 maintains a table of metadata. The table entries are (name, value)
7712 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7713 with the master transaction.
7715 This option specifies the table entry names that Squid must accept
7716 from and forward to the adaptation transactions.
7718 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7719 shared table by returning an ICAP header field with a name
7720 specified in adaptation_masterx_shared_names.
7722 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7723 shared table by implementing the libecap::visitEachOption() API
7724 to provide an option with a name specified in
7725 adaptation_masterx_shared_names.
7727 Squid will store and forward the set entry to subsequent adaptation
7728 transactions within the same master transaction scope.
7730 Only one shared entry name is supported at this time.
7733 # share authentication information among ICAP services
7734 adaptation_masterx_shared_names X-Subscriber-ID
7737 NAME: adaptation_meta
7739 IFDEF: USE_ADAPTATION
7740 LOC: Adaptation::Config::metaHeaders
7743 This option allows Squid administrator to add custom ICAP request
7744 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7745 Use it to pass custom authentication tokens and other
7746 transaction-state related meta information to an ICAP/eCAP service.
7748 The addition of a meta header is ACL-driven:
7749 adaptation_meta name value [!]aclname ...
7751 Processing for a given header name stops after the first ACL list match.
7752 Thus, it is impossible to add two headers with the same name. If no ACL
7753 lists match for a given header name, no such header is added. For
7756 # do not debug transactions except for those that need debugging
7757 adaptation_meta X-Debug 1 needs_debugging
7759 # log all transactions except for those that must remain secret
7760 adaptation_meta X-Log 1 !keep_secret
7762 # mark transactions from users in the "G 1" group
7763 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7765 The "value" parameter may be a regular squid.conf token or a "double
7766 quoted string". Within the quoted string, use backslash (\) to escape
7767 any character, which is currently only useful for escaping backslashes
7768 and double quotes. For example,
7769 "this string has one backslash (\\) and two \"quotes\""
7771 Used adaptation_meta header values may be logged via %note
7772 logformat code. If multiple adaptation_meta headers with the same name
7773 are used during master transaction lifetime, the header values are
7774 logged in the order they were used and duplicate values are ignored
7775 (only the first repeated value will be logged).
7781 LOC: Adaptation::Icap::TheConfig.repeat
7782 DEFAULT_IF_NONE: deny all
7784 This ACL determines which retriable ICAP transactions are
7785 retried. Transactions that received a complete ICAP response
7786 and did not have to consume or produce HTTP bodies to receive
7787 that response are usually retriable.
7789 icap_retry allow|deny [!]aclname ...
7791 Squid automatically retries some ICAP I/O timeouts and errors
7792 due to persistent connection race conditions.
7794 See also: icap_retry_limit
7797 NAME: icap_retry_limit
7800 LOC: Adaptation::Icap::TheConfig.repeat_limit
7803 Limits the number of retries allowed. When set to zero (default),
7804 no retries are allowed.
7806 Communication errors due to persistent connection race
7807 conditions are unavoidable, automatically retried, and do not
7808 count against this limit.
7810 See also: icap_retry
7816 -----------------------------------------------------------------------------
7819 NAME: check_hostnames
7822 LOC: Config.onoff.check_hostnames
7824 For security and stability reasons Squid can check
7825 hostnames for Internet standard RFC compliance. If you want
7826 Squid to perform these checks turn this directive on.
7829 NAME: allow_underscore
7832 LOC: Config.onoff.allow_underscore
7834 Underscore characters is not strictly allowed in Internet hostnames
7835 but nevertheless used by many sites. Set this to off if you want
7836 Squid to be strict about the standard.
7837 This check is performed only when check_hostnames is set to on.
7840 NAME: cache_dns_program
7842 IFDEF: USE_DNSHELPER
7843 DEFAULT: @DEFAULT_DNSSERVER@
7844 LOC: Config.Program.dnsserver
7846 Specify the location of the executable for dnslookup process.
7850 TYPE: HelperChildConfig
7851 IFDEF: USE_DNSHELPER
7852 DEFAULT: 32 startup=1 idle=1
7853 LOC: Config.dnsChildren
7855 The maximum number of processes spawn to service DNS name lookups.
7856 If you limit it too few Squid will have to wait for them to process
7857 a backlog of requests, slowing it down. If you allow too many they
7858 will use RAM and other system resources noticably.
7859 The maximum this may be safely set to is 32.
7861 The startup= and idle= options allow some measure of skew in your
7866 Sets a minimum of how many processes are to be spawned when Squid
7867 starts or reconfigures. When set to zero the first request will
7868 cause spawning of the first child process to handle it.
7870 Starting too few will cause an initial slowdown in traffic as Squid
7871 attempts to simultaneously spawn enough processes to cope.
7875 Sets a minimum of how many processes Squid is to try and keep available
7876 at all times. When traffic begins to rise above what the existing
7877 processes can handle this many more will be spawned up to the maximum
7878 configured. A minimum setting of 1 is required.
7881 NAME: dns_retransmit_interval
7884 LOC: Config.Timeout.idns_retransmit
7885 IFDEF: !USE_DNSHELPER
7887 Initial retransmit interval for DNS queries. The interval is
7888 doubled each time all configured DNS servers have been tried.
7894 LOC: Config.Timeout.idns_query
7895 IFDEF: !USE_DNSHELPER
7897 DNS Query timeout. If no response is received to a DNS query
7898 within this time all DNS servers for the queried domain
7899 are assumed to be unavailable.
7902 NAME: dns_packet_max
7905 LOC: Config.dns.packet_max
7906 IFDEF: !USE_DNSHELPER
7908 Maximum number of bytes packet size to advertise via EDNS.
7909 Set to "none" to disable EDNS large packet support.
7911 For legacy reasons DNS UDP replies will default to 512 bytes which
7912 is too small for many responses. EDNS provides a means for Squid to
7913 negotiate receiving larger responses back immediately without having
7914 to failover with repeat requests. Responses larger than this limit
7915 will retain the old behaviour of failover to TCP DNS.
7917 Squid has no real fixed limit internally, but allowing packet sizes
7918 over 1500 bytes requires network jumbogram support and is usually not
7921 WARNING: The RFC also indicates that some older resolvers will reply
7922 with failure of the whole request if the extension is added. Some
7923 resolvers have already been identified which will reply with mangled
7924 EDNS response on occasion. Usually in response to many-KB jumbogram
7925 sizes being advertised by Squid.
7926 Squid will currently treat these both as an unable-to-resolve domain
7927 even if it would be resolvable without EDNS.
7934 LOC: Config.onoff.res_defnames
7936 Normally the RES_DEFNAMES resolver option is disabled
7937 (see res_init(3)). This prevents caches in a hierarchy
7938 from interpreting single-component hostnames locally. To allow
7939 Squid to handle single-component names, enable this option.
7942 NAME: dns_nameservers
7945 LOC: Config.dns_nameservers
7947 Use this if you want to specify a list of DNS name servers
7948 (IP addresses) to use instead of those given in your
7949 /etc/resolv.conf file.
7950 On Windows platforms, if no value is specified here or in
7951 the /etc/resolv.conf file, the list of DNS name servers are
7952 taken from the Windows registry, both static and dynamic DHCP
7953 configurations are supported.
7955 Example: dns_nameservers 10.0.0.1 192.172.0.4
7960 DEFAULT: @DEFAULT_HOSTS@
7961 LOC: Config.etcHostsPath
7963 Location of the host-local IP name-address associations
7964 database. Most Operating Systems have such a file on different
7966 - Un*X & Linux: /etc/hosts
7967 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7968 (%SystemRoot% value install default is c:\winnt)
7969 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7970 (%SystemRoot% value install default is c:\windows)
7971 - Windows 9x/Me: %windir%\hosts
7972 (%windir% value is usually c:\windows)
7973 - Cygwin: /etc/hosts
7975 The file contains newline-separated definitions, in the
7976 form ip_address_in_dotted_form name [name ...] names are
7977 whitespace-separated. Lines beginning with an hash (#)
7978 character are comments.
7980 The file is checked at startup and upon configuration.
7981 If set to 'none', it won't be checked.
7982 If append_domain is used, that domain will be added to
7983 domain-local (i.e. not containing any dot character) host
7989 LOC: Config.appendDomain
7992 Appends local domain name to hostnames without any dots in
7993 them. append_domain must begin with a period.
7995 Be warned there are now Internet names with no dots in
7996 them using only top-domain names, so setting this may
7997 cause some Internet sites to become unavailable.
8000 append_domain .yourdomain.com
8003 NAME: ignore_unknown_nameservers
8005 LOC: Config.onoff.ignore_unknown_nameservers
8007 IFDEF: !USE_DNSHELPER
8009 By default Squid checks that DNS responses are received
8010 from the same IP addresses they are sent to. If they
8011 don't match, Squid ignores the response and writes a warning
8012 message to cache.log. You can allow responses from unknown
8013 nameservers by setting this option to 'off'.
8019 LOC: Config.dns.v4_first
8020 IFDEF: !USE_DNSHELPER
8022 With the IPv6 Internet being as fast or faster than IPv4 Internet
8023 for most networks Squid prefers to contact websites over IPv6.
8025 This option reverses the order of preference to make Squid contact
8026 dual-stack websites over IPv4 first. Squid will still perform both
8027 IPv6 and IPv4 DNS lookups before connecting.
8030 This option will restrict the situations under which IPv6
8031 connectivity is used (and tested). Hiding network problems
8032 which would otherwise be detected and warned about.
8036 COMMENT: (number of entries)
8039 LOC: Config.ipcache.size
8046 LOC: Config.ipcache.low
8053 LOC: Config.ipcache.high
8055 The size, low-, and high-water marks for the IP cache.
8058 NAME: fqdncache_size
8059 COMMENT: (number of entries)
8062 LOC: Config.fqdncache.size
8064 Maximum number of FQDN cache entries.
8069 -----------------------------------------------------------------------------
8076 LOC: Config.onoff.mem_pools
8078 If set, Squid will keep pools of allocated (but unused) memory
8079 available for future use. If memory is a premium on your
8080 system and you believe your malloc library outperforms Squid
8081 routines, disable this.
8084 NAME: memory_pools_limit
8088 LOC: Config.MemPools.limit
8090 Used only with memory_pools on:
8091 memory_pools_limit 50 MB
8093 If set to a non-zero value, Squid will keep at most the specified
8094 limit of allocated (but unused) memory in memory pools. All free()
8095 requests that exceed this limit will be handled by your malloc
8096 library. Squid does not pre-allocate any memory, just safe-keeps
8097 objects that otherwise would be free()d. Thus, it is safe to set
8098 memory_pools_limit to a reasonably high value even if your
8099 configuration will use less memory.
8101 If set to none, Squid will keep all memory it can. That is, there
8102 will be no limit on the total amount of memory used for safe-keeping.
8104 To disable memory allocation optimization, do not set
8105 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
8107 An overhead for maintaining memory pools is not taken into account
8108 when the limit is checked. This overhead is close to four bytes per
8109 object kept. However, pools may actually _save_ memory because of
8110 reduced memory thrashing in your malloc library.
8114 COMMENT: on|off|transparent|truncate|delete
8117 LOC: opt_forwarded_for
8119 If set to "on", Squid will append your client's IP address
8120 in the HTTP requests it forwards. By default it looks like:
8122 X-Forwarded-For: 192.1.2.3
8124 If set to "off", it will appear as
8126 X-Forwarded-For: unknown
8128 If set to "transparent", Squid will not alter the
8129 X-Forwarded-For header in any way.
8131 If set to "delete", Squid will delete the entire
8132 X-Forwarded-For header.
8134 If set to "truncate", Squid will remove all existing
8135 X-Forwarded-For entries, and place the client IP as the sole entry.
8138 NAME: cachemgr_passwd
8139 TYPE: cachemgrpasswd
8141 LOC: Config.passwd_list
8143 Specify passwords for cachemgr operations.
8145 Usage: cachemgr_passwd password action action ...
8147 Some valid actions are (see cache manager menu for a full list):
8187 * Indicates actions which will not be performed without a
8188 valid password, others can be performed if not listed here.
8190 To disable an action, set the password to "disable".
8191 To allow performing an action without a password, set the
8194 Use the keyword "all" to set the same password for all actions.
8197 cachemgr_passwd secret shutdown
8198 cachemgr_passwd lesssssssecret info stats/objects
8199 cachemgr_passwd disable all
8206 LOC: Config.onoff.client_db
8208 If you want to disable collecting per-client statistics,
8209 turn off client_db here.
8212 NAME: refresh_all_ims
8216 LOC: Config.onoff.refresh_all_ims
8218 When you enable this option, squid will always check
8219 the origin server for an update when a client sends an
8220 If-Modified-Since request. Many browsers use IMS
8221 requests when the user requests a reload, and this
8222 ensures those clients receive the latest version.
8224 By default (off), squid may return a Not Modified response
8225 based on the age of the cached version.
8228 NAME: reload_into_ims
8229 IFDEF: USE_HTTP_VIOLATIONS
8233 LOC: Config.onoff.reload_into_ims
8235 When you enable this option, client no-cache or ``reload''
8236 requests will be changed to If-Modified-Since requests.
8237 Doing this VIOLATES the HTTP standard. Enabling this
8238 feature could make you liable for problems which it
8241 see also refresh_pattern for a more selective approach.
8244 NAME: connect_retries
8246 LOC: Config.connect_retries
8249 This sets the maximum number of connection attempts made for each
8250 TCP connection. The connect_retries attempts must all still
8251 complete within the connection timeout period.
8253 The default is not to re-try if the first connection attempt fails.
8254 The (not recommended) maximum is 10 tries.
8256 A warning message will be generated if it is set to a too-high
8257 value and the configured value will be over-ridden.
8259 Note: These re-tries are in addition to forward_max_tries
8260 which limit how many different addresses may be tried to find
8264 NAME: retry_on_error
8266 LOC: Config.retry.onerror
8269 If set to ON Squid will automatically retry requests when
8270 receiving an error response with status 403 (Forbidden),
8271 500 (Internal Error), 501 or 503 (Service not available).
8272 Status 502 and 504 (Gateway errors) are always retried.
8274 This is mainly useful if you are in a complex cache hierarchy to
8275 work around access control errors.
8277 NOTE: This retry will attempt to find another working destination.
8278 Which is different from the server which just failed.
8281 NAME: as_whois_server
8283 LOC: Config.as_whois_server
8284 DEFAULT: whois.ra.net
8286 WHOIS server to query for AS numbers. NOTE: AS numbers are
8287 queried only when Squid starts up, not for every request.
8292 LOC: Config.onoff.offline
8295 Enable this option and Squid will never try to validate cached
8299 NAME: uri_whitespace
8300 TYPE: uri_whitespace
8301 LOC: Config.uri_whitespace
8304 What to do with requests that have whitespace characters in the
8307 strip: The whitespace characters are stripped out of the URL.
8308 This is the behavior recommended by RFC2396.
8309 deny: The request is denied. The user receives an "Invalid
8311 allow: The request is allowed and the URI is not changed. The
8312 whitespace characters remain in the URI. Note the
8313 whitespace is passed to redirector processes if they
8315 encode: The request is allowed and the whitespace characters are
8316 encoded according to RFC1738. This could be considered
8317 a violation of the HTTP/1.1
8318 RFC because proxies are not allowed to rewrite URI's.
8319 chop: The request is allowed and the URI is chopped at the
8320 first whitespace. This might also be considered a
8326 LOC: Config.chroot_dir
8329 Specifies a directory where Squid should do a chroot() while
8330 initializing. This also causes Squid to fully drop root
8331 privileges after initializing. This means, for example, if you
8332 use a HTTP port less than 1024 and try to reconfigure, you may
8333 get an error saying that Squid can not open the port.
8336 NAME: balance_on_multiple_ip
8338 LOC: Config.onoff.balance_on_multiple_ip
8341 Modern IP resolvers in squid sort lookup results by preferred access.
8342 By default squid will use these IP in order and only rotates to
8343 the next listed when the most preffered fails.
8345 Some load balancing servers based on round robin DNS have been
8346 found not to preserve user session state across requests
8347 to different IP addresses.
8349 Enabling this directive Squid rotates IP's per request.
8352 NAME: pipeline_prefetch
8354 LOC: Config.onoff.pipeline_prefetch
8357 To boost the performance of pipelined requests to closer
8358 match that of a non-proxied environment Squid can try to fetch
8359 up to two requests in parallel from a pipeline.
8361 Defaults to off for bandwidth management and access logging
8364 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8367 NAME: high_response_time_warning
8370 LOC: Config.warnings.high_rptm
8373 If the one-minute median response time exceeds this value,
8374 Squid prints a WARNING with debug level 0 to get the
8375 administrators attention. The value is in milliseconds.
8378 NAME: high_page_fault_warning
8380 LOC: Config.warnings.high_pf
8383 If the one-minute average page fault rate exceeds this
8384 value, Squid prints a WARNING with debug level 0 to get
8385 the administrators attention. The value is in page faults
8389 NAME: high_memory_warning
8391 LOC: Config.warnings.high_memory
8394 If the memory usage (as determined by mallinfo) exceeds
8395 this amount, Squid prints a WARNING with debug level 0 to get
8396 the administrators attention.
8399 NAME: sleep_after_fork
8400 COMMENT: (microseconds)
8402 LOC: Config.sleep_after_fork
8405 When this is set to a non-zero value, the main Squid process
8406 sleeps the specified number of microseconds after a fork()
8407 system call. This sleep may help the situation where your
8408 system reports fork() failures due to lack of (virtual)
8409 memory. Note, however, if you have a lot of child
8410 processes, these sleep delays will add up and your
8411 Squid will not service requests for some amount of time
8412 until all the child processes have been started.
8413 On Windows value less then 1000 (1 milliseconds) are
8417 NAME: windows_ipaddrchangemonitor
8418 IFDEF: _SQUID_WINDOWS_
8422 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8424 On Windows Squid by default will monitor IP address changes and will
8425 reconfigure itself after any detected event. This is very useful for
8426 proxies connected to internet with dial-up interfaces.
8427 In some cases (a Proxy server acting as VPN gateway is one) it could be
8428 desiderable to disable this behaviour setting this to 'off'.
8429 Note: after changing this, Squid service must be restarted.
8434 IFDEF: USE_SQUID_EUI
8436 LOC: Eui::TheConfig.euiLookup
8438 Whether to lookup the EUI or MAC address of a connected client.
8441 NAME: max_filedescriptors max_filedesc
8444 LOC: Config.max_filedescriptors
8446 The maximum number of filedescriptors supported.
8448 The default "0" means Squid inherits the current ulimit setting.
8450 Note: Changing this requires a restart of Squid. Also
8451 not all comm loops supports large values.
8459 Number of main Squid processes or "workers" to fork and maintain.
8460 0: "no daemon" mode, like running "squid -N ..."
8461 1: "no SMP" mode, start one main Squid process daemon (default)
8462 N: start N main Squid process daemons (i.e., SMP mode)
8464 In SMP mode, each worker does nearly all what a single Squid daemon
8465 does (e.g., listen on http_port and forward HTTP requests).
8468 NAME: cpu_affinity_map
8469 TYPE: CpuAffinityMap
8470 LOC: Config.cpuAffinityMap
8473 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8475 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8477 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8479 affects processes 1 through 4 only and places them on the first
8480 four even cores, starting with core #1.
8482 CPU cores are numbered starting from 1. Requires support for
8483 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8485 Multiple cpu_affinity_map options are merged.