2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the documentation for the Squid configuration file.
34 This documentation can also be found online at:
35 http://www.squid-cache.org/Doc/config/
37 You may wish to look at the Squid home page and wiki for the
38 FAQ and other documentation:
39 http://www.squid-cache.org/
40 http://wiki.squid-cache.org/SquidFaq
41 http://wiki.squid-cache.org/ConfigExamples
43 This documentation shows what the defaults for various directives
44 happen to be. If you don't need to change the default, you should
45 leave the line out of your squid.conf in most cases.
47 In some cases "none" refers to no default setting at all,
48 while in other cases it refers to the value of the option
49 - the comments for that keyword indicate if this is the case.
54 Configuration options can be included using the "include" directive.
55 Include takes a list of files to include. Quoting and wildcards are
60 include /path/to/included/file/squid.acl.config
62 Includes can be nested up to a hard-coded depth of 16 levels.
63 This arbitrary restriction is to prevent recursive include references
64 from causing Squid entering an infinite loop whilst trying to load
68 Conditional configuration
70 If-statements can be used to make configuration directives
74 ... regular configuration directives ...
76 ... regular configuration directives ...]
79 The else part is optional. The keywords "if", "else", and "endif"
80 must be typed on their own lines, as if they were regular
81 configuration directives.
83 NOTE: An else-if condition is not supported.
85 These individual conditions types are supported:
88 Always evaluates to true.
90 Always evaluates to false.
92 Equality comparison of two integer numbers.
97 The following SMP-related preprocessor macros can be used.
99 ${process_name} expands to the current Squid process "name"
100 (e.g., squid1, squid2, or cache1).
102 ${process_number} expands to the current Squid process
103 identifier, which is an integer number (e.g., 1, 2, 3) unique
104 across all Squid processes.
107 # Options Removed in 3.2
108 NAME: ignore_expect_100
111 Remove this line. The HTTP/1.1 feature is now fully supported by default.
114 NAME: dns_v4_fallback
123 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
126 NAME: maximum_single_addr_tries
129 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
132 NAME: url_rewrite_concurrency
135 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
138 # Options Removed in 3.1
142 Remove this line. DNS is no longer tested on startup.
145 NAME: extension_methods
148 Remove this line. All valid methods for HTTP are accepted by default.
151 # 2.7 Options Removed/Replaced in 3.2
156 # 2.7 Options Removed/Replaced in 3.1
164 Remove this line. HTTP/1.1 is supported by default.
167 NAME: upgrade_http0.9
170 Remove this line. ICY/1.0 streaming protocol is supported by default.
173 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
176 Alter these entries. Use the qos_flows directive instead.
179 # Options Removed in 3.0
183 Since squid-3.0 replace with request_header_access or reply_header_access
184 depending on whether you wish to match client requests or server replies.
187 NAME: httpd_accel_no_pmtu_disc
190 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
194 OPTIONS FOR AUTHENTICATION
195 -----------------------------------------------------------------------------
204 This is used to define parameters for the various authentication
205 schemes supported by Squid.
207 format: auth_param scheme parameter [setting]
209 The order in which authentication schemes are presented to the client is
210 dependent on the order the scheme first appears in config file. IE
211 has a bug (it's not RFC 2617 compliant) in that it will use the basic
212 scheme if basic is the first entry presented, even if more secure
213 schemes are presented. For now use the order in the recommended
214 settings section below. If other browsers have difficulties (don't
215 recognize the schemes offered even if you are using basic) either
216 put basic first, or disable the other schemes (by commenting out their
219 Once an authentication scheme is fully configured, it can only be
220 shutdown by shutting squid down and restarting. Changes can be made on
221 the fly and activated with a reconfigure. I.E. You can change to a
222 different helper, but not unconfigure the helper completely.
224 Please note that while this directive defines how Squid processes
225 authentication it does not automatically activate authentication.
226 To use authentication you must in addition make use of ACLs based
227 on login name in http_access (proxy_auth, proxy_auth_regex or
228 external with %LOGIN used in the format tag). The browser will be
229 challenged for authentication on the first such acl encountered
230 in http_access processing and will also be re-challenged for new
231 login credentials if the request is being denied by a proxy_auth
234 WARNING: authentication can't be used in a transparently intercepting
235 proxy as the client then thinks it is talking to an origin server and
236 not the proxy. This is a limitation of bending the TCP/IP protocol to
237 transparently intercepting port 80, not a limitation in Squid.
238 Ports flagged 'transparent', 'intercept', or 'tproxy' have
239 authentication disabled.
241 === Parameters for the basic scheme follow. ===
244 Specify the command for the external authenticator. Such a program
245 reads a line containing "username password" and replies "OK" or
246 "ERR" in an endless loop. "ERR" responses may optionally be followed
247 by a error description available as %m in the returned error page.
248 If you use an authenticator, make sure you have 1 acl of type
251 By default, the basic authentication scheme is not used unless a
252 program is specified.
254 If you want to use the traditional NCSA proxy authentication, set
255 this line to something like
257 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
260 HTTP uses iso-latin-1 as characterset, while some authentication
261 backends such as LDAP expects UTF-8. If this is set to on Squid will
262 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
263 username & password to the helper.
265 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
266 The maximum number of authenticator processes to spawn. If you start too few
267 Squid will have to wait for them to process a backlog of credential
268 verifications, slowing it down. When password verifications are
269 done via a (slow) network you are likely to need lots of
270 authenticator processes.
272 The startup= and idle= options permit some skew in the exact amount
273 run. A minimum of startup=N will begin during startup and reconfigure.
274 Squid will start more in groups of up to idle=N in an attempt to meet
275 traffic needs and to keep idle=N free above those traffic needs up to
278 The concurrency= option sets the number of concurrent requests the
279 helper can process. The default of 0 is used for helpers who only
280 supports one request at a time. Setting this to a number greater than
281 0 changes the protocol used to include a channel number first on the
282 request/response line, allowing multiple requests to be sent to the
283 same helper in parallell without wating for the response.
284 Must not be set unless it's known the helper supports this.
286 auth_param basic children 20 startup=0 idle=1
289 Specifies the realm name which is to be reported to the
290 client for the basic proxy authentication scheme (part of
291 the text the user will see when prompted their username and
292 password). There is no default.
293 auth_param basic realm Squid proxy-caching web server
295 "credentialsttl" timetolive
296 Specifies how long squid assumes an externally validated
297 username:password pair is valid for - in other words how
298 often the helper program is called for that user. Set this
299 low to force revalidation with short lived passwords. Note
300 setting this high does not impact your susceptibility
301 to replay attacks unless you are using an one-time password
302 system (such as SecureID). If you are using such a system,
303 you will be vulnerable to replay attacks unless you also
304 use the max_user_ip ACL in an http_access rule.
306 "casesensitive" on|off
307 Specifies if usernames are case sensitive. Most user databases are
308 case insensitive allowing the same username to be spelled using both
309 lower and upper case letters, but some are case sensitive. This
310 makes a big difference for user_max_ip ACL processing and similar.
311 auth_param basic casesensitive off
313 === Parameters for the digest scheme follow ===
316 Specify the command for the external authenticator. Such
317 a program reads a line containing "username":"realm" and
318 replies with the appropriate H(A1) value hex encoded or
319 ERR if the user (or his H(A1) hash) does not exists.
320 See rfc 2616 for the definition of H(A1).
321 "ERR" responses may optionally be followed by a error description
322 available as %m in the returned error page.
324 By default, the digest authentication scheme is not used unless a
325 program is specified.
327 If you want to use a digest authenticator, set this line to
330 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
333 HTTP uses iso-latin-1 as characterset, while some authentication
334 backends such as LDAP expects UTF-8. If this is set to on Squid will
335 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
336 username & password to the helper.
338 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
339 The maximum number of authenticator processes to spawn (default 5).
340 If you start too few Squid will have to wait for them to
341 process a backlog of H(A1) calculations, slowing it down.
342 When the H(A1) calculations are done via a (slow) network
343 you are likely to need lots of authenticator processes.
345 The startup= and idle= options permit some skew in the exact amount
346 run. A minimum of startup=N will begin during startup and reconfigure.
347 Squid will start more in groups of up to idle=N in an attempt to meet
348 traffic needs and to keep idle=N free above those traffic needs up to
351 The concurrency= option sets the number of concurrent requests the
352 helper can process. The default of 0 is used for helpers who only
353 supports one request at a time. Setting this to a number greater than
354 0 changes the protocol used to include a channel number first on the
355 request/response line, allowing multiple requests to be sent to the
356 same helper in parallell without wating for the response.
357 Must not be set unless it's known the helper supports this.
359 auth_param digest children 20 startup=0 idle=1
362 Specifies the realm name which is to be reported to the
363 client for the digest proxy authentication scheme (part of
364 the text the user will see when prompted their username and
365 password). There is no default.
366 auth_param digest realm Squid proxy-caching web server
368 "nonce_garbage_interval" timeinterval
369 Specifies the interval that nonces that have been issued
370 to client_agent's are checked for validity.
372 "nonce_max_duration" timeinterval
373 Specifies the maximum length of time a given nonce will be
376 "nonce_max_count" number
377 Specifies the maximum number of times a given nonce can be
380 "nonce_strictness" on|off
381 Determines if squid requires strict increment-by-1 behavior
382 for nonce counts, or just incrementing (off - for use when
383 useragents generate nonce counts that occasionally miss 1
384 (ie, 1,2,4,6)). Default off.
386 "check_nonce_count" on|off
387 This directive if set to off can disable the nonce count check
388 completely to work around buggy digest qop implementations in
389 certain mainstream browser versions. Default on to check the
390 nonce count to protect from authentication replay attacks.
392 "post_workaround" on|off
393 This is a workaround to certain buggy browsers who sends
394 an incorrect request digest in POST requests when reusing
395 the same nonce as acquired earlier on a GET request.
397 === NTLM scheme options follow ===
400 Specify the command for the external NTLM authenticator.
401 Such a program reads exchanged NTLMSSP packets with
402 the browser via Squid until authentication is completed.
403 If you use an NTLM authenticator, make sure you have 1 acl
404 of type proxy_auth. By default, the NTLM authenticator_program
407 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
409 "children" numberofchildren [startup=N] [idle=N]
410 The maximum number of authenticator processes to spawn (default 5).
411 If you start too few Squid will have to wait for them to
412 process a backlog of credential verifications, slowing it
413 down. When credential verifications are done via a (slow)
414 network you are likely to need lots of authenticator
417 The startup= and idle= options permit some skew in the exact amount
418 run. A minimum of startup=N will begin during startup and reconfigure.
419 Squid will start more in groups of up to idle=N in an attempt to meet
420 traffic needs and to keep idle=N free above those traffic needs up to
423 auth_param ntlm children 20 startup=0 idle=1
426 If you experience problems with PUT/POST requests when using the
427 Negotiate authentication scheme then you can try setting this to
428 off. This will cause Squid to forcibly close the connection on
429 the initial requests where the browser asks which schemes are
430 supported by the proxy.
432 auth_param ntlm keep_alive on
434 === Options for configuring the NEGOTIATE auth-scheme follow ===
437 Specify the command for the external Negotiate authenticator.
438 This protocol is used in Microsoft Active-Directory enabled setups with
439 the Microsoft Internet Explorer or Mozilla Firefox browsers.
440 Its main purpose is to exchange credentials with the Squid proxy
441 using the Kerberos mechanisms.
442 If you use a Negotiate authenticator, make sure you have at least
443 one acl of type proxy_auth active. By default, the negotiate
444 authenticator_program is not used.
445 The only supported program for this role is the ntlm_auth
446 program distributed as part of Samba, version 4 or later.
448 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
450 "children" numberofchildren [startup=N] [idle=N]
451 The maximum number of authenticator processes to spawn (default 5).
452 If you start too few Squid will have to wait for them to
453 process a backlog of credential verifications, slowing it
454 down. When crendential verifications are done via a (slow)
455 network you are likely to need lots of authenticator
458 The startup= and idle= options permit some skew in the exact amount
459 run. A minimum of startup=N will begin during startup and reconfigure.
460 Squid will start more in groups of up to idle=N in an attempt to meet
461 traffic needs and to keep idle=N free above those traffic needs up to
464 auth_param negotiate children 20 startup=0 idle=1
467 If you experience problems with PUT/POST requests when using the
468 Negotiate authentication scheme then you can try setting this to
469 off. This will cause Squid to forcibly close the connection on
470 the initial requests where the browser asks which schemes are
471 supported by the proxy.
473 auth_param negotiate keep_alive on
478 #Recommended minimum configuration per scheme:
479 #auth_param negotiate program <uncomment and complete this line to activate>
480 #auth_param negotiate children 20 startup=0 idle=1
481 #auth_param negotiate keep_alive on
483 #auth_param ntlm program <uncomment and complete this line to activate>
484 #auth_param ntlm children 20 startup=0 idle=1
485 #auth_param ntlm keep_alive on
487 #auth_param digest program <uncomment and complete this line>
488 #auth_param digest children 20 startup=0 idle=1
489 #auth_param digest realm Squid proxy-caching web server
490 #auth_param digest nonce_garbage_interval 5 minutes
491 #auth_param digest nonce_max_duration 30 minutes
492 #auth_param digest nonce_max_count 50
494 #auth_param basic program <uncomment and complete this line>
495 #auth_param basic children 5 startup=5 idle=1
496 #auth_param basic realm Squid proxy-caching web server
497 #auth_param basic credentialsttl 2 hours
500 NAME: authenticate_cache_garbage_interval
503 LOC: Config.authenticateGCInterval
505 The time period between garbage collection across the username cache.
506 This is a tradeoff between memory utilization (long intervals - say
507 2 days) and CPU (short intervals - say 1 minute). Only change if you
511 NAME: authenticate_ttl
514 LOC: Config.authenticateTTL
516 The time a user & their credentials stay in the logged in
517 user cache since their last request. When the garbage
518 interval passes, all user credentials that have passed their
519 TTL are removed from memory.
522 NAME: authenticate_ip_ttl
524 LOC: Config.authenticateIpTTL
527 If you use proxy authentication and the 'max_user_ip' ACL,
528 this directive controls how long Squid remembers the IP
529 addresses associated with each user. Use a small value
530 (e.g., 60 seconds) if your users might change addresses
531 quickly, as is the case with dialups. You might be safe
532 using a larger value (e.g., 2 hours) in a corporate LAN
533 environment with relatively static address assignments.
538 -----------------------------------------------------------------------------
541 NAME: external_acl_type
542 TYPE: externalAclHelper
543 LOC: Config.externalAclHelperList
546 This option defines external acl classes using a helper program
547 to look up the status
549 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
553 ttl=n TTL in seconds for cached results (defaults to 3600
556 TTL for cached negative lookups (default same
559 Maximum number of acl helper processes spawned to service
560 external acl lookups of this type. (default 20)
562 Minimum number of acl helper processes to spawn during
563 startup and reconfigure to service external acl lookups
564 of this type. (default 0)
566 Number of acl helper processes to keep ahead of traffic
567 loads. Squid will spawn this many at once whenever load
568 rises above the capabilities of existing processes.
569 Up to the value of children-max. (default 1)
570 concurrency=n concurrency level per process. Only used with helpers
571 capable of processing more than one query at a time.
572 cache=n limit the result cache size, default is unbounded.
573 grace=n Percentage remaining of TTL where a refresh of a
574 cached entry should be initiated without needing to
575 wait for a new reply. (default is for no grace period)
576 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
577 ipv4 / ipv6 IP protocol used to communicate with this helper.
578 The default is to auto-detect IPv6 and use it when available.
580 FORMAT specifications
582 %LOGIN Authenticated user login name
583 %EXT_USER Username from previous external acl
584 %EXT_LOG Log details from previous external acl
585 %EXT_TAG Tag from previous external acl
586 %IDENT Ident user name
588 %SRCPORT Client source port
591 %PROTO Requested protocol
593 %PATH Requested URL path
594 %METHOD Request method
595 %MYADDR Squid interface address
596 %MYPORT Squid http_port number
597 %PATH Requested URL-path (including query-string if any)
598 %USER_CERT SSL User certificate in PEM format
599 %USER_CERTCHAIN SSL User certificate chain in PEM format
600 %USER_CERT_xx SSL User certificate subject attribute xx
601 %USER_CA_xx SSL User certificate issuer attribute xx
603 %>{Header} HTTP request header "Header"
605 HTTP request header "Hdr" list member "member"
607 HTTP request header list member using ; as
608 list separator. ; can be any non-alphanumeric
611 %<{Header} HTTP reply header "Header"
613 HTTP reply header "Hdr" list member "member"
615 HTTP reply header list member using ; as
616 list separator. ; can be any non-alphanumeric
619 %% The percent sign. Useful for helpers which need
620 an unchanging input format.
622 In addition to the above, any string specified in the referencing
623 acl will also be included in the helper request line, after the
624 specified formats (see the "acl external" directive)
626 The helper receives lines per the above format specification,
627 and returns lines starting with OK or ERR indicating the validity
628 of the request and optionally followed by additional keywords with
631 General result syntax:
633 OK/ERR keyword=value ...
637 user= The users name (login)
638 password= The users password (for login= cache_peer option)
639 message= Message describing the reason. Available as %o
641 tag= Apply a tag to a request (for both ERR and OK results)
642 Only sets a tag, does not alter existing tags.
643 log= String to be logged in access.log. Available as
644 %ea in logformat specifications
646 If protocol=3.0 (the default) then URL escaping is used to protect
647 each value in both requests and responses.
649 If using protocol=2.5 then all values need to be enclosed in quotes
650 if they may contain whitespace, or the whitespace escaped using \.
651 And quotes or \ characters within the keyword value must be \ escaped.
653 When using the concurrency= option the protocol is changed by
654 introducing a query channel tag infront of the request/response.
655 The query channel tag is a number between 0 and concurrency-1.
662 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
663 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
664 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
665 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
666 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
669 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
670 DEFAULT: localhost src 127.0.0.1/32 ::1
671 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
672 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
674 Defining an Access List
676 Every access list definition must begin with an aclname and acltype,
677 followed by either type-specific arguments or a quoted filename that
680 acl aclname acltype argument ...
681 acl aclname acltype "file" ...
683 When using "file", the file should contain one item per line.
685 By default, regular expressions are CASE-SENSITIVE.
686 To make them case-insensitive, use the -i option. To return case-sensitive
687 use the +i option between patterns, or make a new ACL line without -i.
689 Some acl types require suspending the current request in order
690 to access some external data source.
691 Those which do are marked with the tag [slow], those which
692 don't are marked as [fast].
693 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
694 for further information
696 ***** ACL TYPES AVAILABLE *****
698 acl aclname src ip-address/mask ... # clients IP address [fast]
699 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
700 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
701 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
703 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
704 # The arp ACL requires the special configure option --enable-arp-acl.
705 # Furthermore, the ARP ACL code is not portable to all operating systems.
706 # It works on Linux, Solaris, Windows, FreeBSD, and some
707 # other *BSD variants.
710 # NOTE: Squid can only determine the MAC address for clients that are on
711 # the same subnet. If the client is on a different subnet,
712 # then Squid cannot find out its MAC address.
714 acl aclname srcdomain .foo.com ...
715 # reverse lookup, from client IP [slow]
716 acl aclname dstdomain .foo.com ...
717 # Destination server from URL [fast]
718 acl aclname srcdom_regex [-i] \.foo\.com ...
719 # regex matching client name [slow]
720 acl aclname dstdom_regex [-i] \.foo\.com ...
721 # regex matching server [fast]
723 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
724 # based URL is used and no match is found. The name "none" is used
725 # if the reverse lookup fails.
727 acl aclname src_as number ...
728 acl aclname dst_as number ...
730 # Except for access control, AS numbers can be used for
731 # routing of requests to specific caches. Here's an
732 # example for routing all requests for AS#1241 and only
733 # those to mycache.mydomain.net:
734 # acl asexample dst_as 1241
735 # cache_peer_access mycache.mydomain.net allow asexample
736 # cache_peer_access mycache_mydomain.net deny all
738 acl aclname peername myPeer ...
740 # match against a named cache_peer entry
741 # set unique name= on cache_peer lines for reliable use.
743 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
753 # h1:m1 must be less than h2:m2
755 acl aclname url_regex [-i] ^http:// ...
756 # regex matching on whole URL [fast]
757 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
758 # regex matching on URL login field
759 acl aclname urlpath_regex [-i] \.gif$ ...
760 # regex matching on URL path [fast]
762 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
764 acl aclname localport 3128 ... # TCP port the client connected to [fast]
765 # NP: for interception mode this is usually '80'
767 acl aclname myportname 3128 ... # http(s)_port name [fast]
769 acl aclname proto HTTP FTP ... # request protocol [fast]
771 acl aclname method GET POST ... # HTTP request method [fast]
773 acl aclname http_status 200 301 500- 400-403 ...
774 # status code in reply [fast]
776 acl aclname browser [-i] regexp ...
777 # pattern match on User-Agent header (see also req_header below) [fast]
779 acl aclname referer_regex [-i] regexp ...
780 # pattern match on Referer header [fast]
781 # Referer is highly unreliable, so use with care
783 acl aclname ident username ...
784 acl aclname ident_regex [-i] pattern ...
785 # string match on ident output [slow]
786 # use REQUIRED to accept any non-null ident.
788 acl aclname proxy_auth [-i] username ...
789 acl aclname proxy_auth_regex [-i] pattern ...
790 # perform http authentication challenge to the client and match against
791 # supplied credentials [slow]
793 # takes a list of allowed usernames.
794 # use REQUIRED to accept any valid username.
796 # Will use proxy authentication in forward-proxy scenarios, and plain
797 # http authenticaiton in reverse-proxy scenarios
799 # NOTE: when a Proxy-Authentication header is sent but it is not
800 # needed during ACL checking the username is NOT logged
803 # NOTE: proxy_auth requires a EXTERNAL authentication program
804 # to check username/password combinations (see
805 # auth_param directive).
807 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
808 # as the browser needs to be configured for using a proxy in order
809 # to respond to proxy authentication.
811 acl aclname snmp_community string ...
812 # A community string to limit access to your SNMP Agent [fast]
815 # acl snmppublic snmp_community public
817 acl aclname maxconn number
818 # This will be matched when the client's IP address has
819 # more than <number> TCP connections established. [fast]
820 # NOTE: This only measures direct TCP links so X-Forwarded-For
821 # indirect clients are not counted.
823 acl aclname max_user_ip [-s] number
824 # This will be matched when the user attempts to log in from more
825 # than <number> different ip addresses. The authenticate_ip_ttl
826 # parameter controls the timeout on the ip entries. [fast]
827 # If -s is specified the limit is strict, denying browsing
828 # from any further IP addresses until the ttl has expired. Without
829 # -s Squid will just annoy the user by "randomly" denying requests.
830 # (the counter is reset each time the limit is reached and a
832 # NOTE: in acceleration mode or where there is mesh of child proxies,
833 # clients may appear to come from multiple addresses if they are
834 # going through proxy farms, so a limit of 1 may cause user problems.
836 acl aclname random probability
837 # Pseudo-randomly match requests. Based on the probability given.
838 # Probability may be written as a decimal (0.333), fraction (1/3)
839 # or ratio of matches:non-matches (3:5).
841 acl aclname req_mime_type [-i] mime-type ...
842 # regex match against the mime type of the request generated
843 # by the client. Can be used to detect file upload or some
844 # types HTTP tunneling requests [fast]
845 # NOTE: This does NOT match the reply. You cannot use this
846 # to match the returned file type.
848 acl aclname req_header header-name [-i] any\.regex\.here
849 # regex match against any of the known request headers. May be
850 # thought of as a superset of "browser", "referer" and "mime-type"
853 acl aclname rep_mime_type [-i] mime-type ...
854 # regex match against the mime type of the reply received by
855 # squid. Can be used to detect file download or some
856 # types HTTP tunneling requests. [fast]
857 # NOTE: This has no effect in http_access rules. It only has
858 # effect in rules that affect the reply data stream such as
861 acl aclname rep_header header-name [-i] any\.regex\.here
862 # regex match against any of the known reply headers. May be
863 # thought of as a superset of "browser", "referer" and "mime-type"
866 acl aclname external class_name [arguments...]
867 # external ACL lookup via a helper class defined by the
868 # external_acl_type directive [slow]
870 acl aclname user_cert attribute values...
871 # match against attributes in a user SSL certificate
872 # attribute is one of DN/C/O/CN/L/ST [fast]
874 acl aclname ca_cert attribute values...
875 # match against attributes a users issuing CA SSL certificate
876 # attribute is one of DN/C/O/CN/L/ST [fast]
878 acl aclname ext_user username ...
879 acl aclname ext_user_regex [-i] pattern ...
880 # string match on username returned by external acl helper [slow]
881 # use REQUIRED to accept any non-null user name.
883 acl aclname tag tagvalue ...
884 # string match on tag returned by external acl helper [slow]
886 acl aclname hier_code codename ...
887 # string match against squid hierarchy code(s); [fast]
888 # e.g., DIRECT, PARENT_HIT, NONE, etc.
890 # NOTE: This has no effect in http_access rules. It only has
891 # effect in rules that affect the reply data stream such as
895 acl aclname ssl_error errorname
896 # match against SSL certificate validation error [fast]
898 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
901 # The following can be used as shortcuts for certificate properties:
902 # [ssl::]certHasExpired: the "not after" field is in the past
903 # [ssl::]certNotYetValid: the "not before" field is in the future
904 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
905 # [ssl::]certSelfSigned: The certificate is self signed.
906 # [ssl::]certDomainMismatch: The certificate CN domain does not
907 # match the name the name of the host we are connecting to.
909 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
910 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
911 # predefined ACLs, just like the 'all' ACL.
913 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
914 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
918 acl macaddress arp 09:00:2b:23:45:67
919 acl myexample dst_as 1241
920 acl password proxy_auth REQUIRED
921 acl fileupload req_mime_type -i ^multipart/form-data$
922 acl javascript rep_mime_type -i ^application/x-javascript$
926 # Recommended minimum configuration:
929 # Example rule allowing access from your local networks.
930 # Adapt to list your (internal) IP networks from where browsing
932 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
933 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
934 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
935 acl localnet src fc00::/7 # RFC 4193 local private network range
936 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
938 acl SSL_ports port 443
939 acl Safe_ports port 80 # http
940 acl Safe_ports port 21 # ftp
941 acl Safe_ports port 443 # https
942 acl Safe_ports port 70 # gopher
943 acl Safe_ports port 210 # wais
944 acl Safe_ports port 1025-65535 # unregistered ports
945 acl Safe_ports port 280 # http-mgmt
946 acl Safe_ports port 488 # gss-http
947 acl Safe_ports port 591 # filemaker
948 acl Safe_ports port 777 # multiling http
949 acl CONNECT method CONNECT
953 NAME: follow_x_forwarded_for
955 IFDEF: FOLLOW_X_FORWARDED_FOR
956 LOC: Config.accessList.followXFF
957 DEFAULT_IF_NONE: deny all
959 Allowing or Denying the X-Forwarded-For header to be followed to
960 find the original source of a request.
962 Requests may pass through a chain of several other proxies
963 before reaching us. The X-Forwarded-For header will contain a
964 comma-separated list of the IP addresses in the chain, with the
965 rightmost address being the most recent.
967 If a request reaches us from a source that is allowed by this
968 configuration item, then we consult the X-Forwarded-For header
969 to see where that host received the request from. If the
970 X-Forwarded-For header contains multiple addresses, we continue
971 backtracking until we reach an address for which we are not allowed
972 to follow the X-Forwarded-For header, or until we reach the first
973 address in the list. For the purpose of ACL used in the
974 follow_x_forwarded_for directive the src ACL type always matches
975 the address we are testing and srcdomain matches its rDNS.
977 The end result of this process is an IP address that we will
978 refer to as the indirect client address. This address may
979 be treated as the client address for access control, ICAP, delay
980 pools and logging, depending on the acl_uses_indirect_client,
981 icap_uses_indirect_client, delay_pool_uses_indirect_client,
982 log_uses_indirect_client and tproxy_uses_indirect_client options.
984 This clause only supports fast acl types.
985 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
987 SECURITY CONSIDERATIONS:
989 Any host for which we follow the X-Forwarded-For header
990 can place incorrect information in the header, and Squid
991 will use the incorrect information as if it were the
992 source address of the request. This may enable remote
993 hosts to bypass any access control restrictions that are
994 based on the client's source addresses.
998 acl localhost src 127.0.0.1
999 acl my_other_proxy srcdomain .proxy.example.com
1000 follow_x_forwarded_for allow localhost
1001 follow_x_forwarded_for allow my_other_proxy
1004 NAME: acl_uses_indirect_client
1007 IFDEF: FOLLOW_X_FORWARDED_FOR
1009 LOC: Config.onoff.acl_uses_indirect_client
1011 Controls whether the indirect client address
1012 (see follow_x_forwarded_for) is used instead of the
1013 direct client address in acl matching.
1015 NOTE: maxconn ACL considers direct TCP links and indirect
1016 clients will always have zero. So no match.
1019 NAME: delay_pool_uses_indirect_client
1022 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1024 LOC: Config.onoff.delay_pool_uses_indirect_client
1026 Controls whether the indirect client address
1027 (see follow_x_forwarded_for) is used instead of the
1028 direct client address in delay pools.
1031 NAME: log_uses_indirect_client
1034 IFDEF: FOLLOW_X_FORWARDED_FOR
1036 LOC: Config.onoff.log_uses_indirect_client
1038 Controls whether the indirect client address
1039 (see follow_x_forwarded_for) is used instead of the
1040 direct client address in the access log.
1043 NAME: tproxy_uses_indirect_client
1046 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1048 LOC: Config.onoff.tproxy_uses_indirect_client
1050 Controls whether the indirect client address
1051 (see follow_x_forwarded_for) is used instead of the
1052 direct client address when spoofing the outgoing client.
1054 This has no effect on requests arriving in non-tproxy
1057 SECURITY WARNING: Usage of this option is dangerous
1058 and should not be used trivially. Correct configuration
1059 of follow_x_forewarded_for with a limited set of trusted
1060 sources is required to prevent abuse of your proxy.
1065 LOC: Config.accessList.http
1066 DEFAULT_IF_NONE: deny all
1068 Allowing or Denying access based on defined access lists
1070 Access to the HTTP port:
1071 http_access allow|deny [!]aclname ...
1073 NOTE on default values:
1075 If there are no "access" lines present, the default is to deny
1078 If none of the "access" lines cause a match, the default is the
1079 opposite of the last line in the list. If the last line was
1080 deny, the default is allow. Conversely, if the last line
1081 is allow, the default will be deny. For these reasons, it is a
1082 good idea to have an "deny all" entry at the end of your access
1083 lists to avoid potential confusion.
1085 This clause supports both fast and slow acl types.
1086 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1091 # Recommended minimum Access Permission configuration:
1093 # Only allow cachemgr access from localhost
1094 http_access allow localhost manager
1095 http_access deny manager
1097 # Deny requests to certain unsafe ports
1098 http_access deny !Safe_ports
1100 # Deny CONNECT to other than secure SSL ports
1101 http_access deny CONNECT !SSL_ports
1103 # We strongly recommend the following be uncommented to protect innocent
1104 # web applications running on the proxy server who think the only
1105 # one who can access services on "localhost" is a local user
1106 #http_access deny to_localhost
1109 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1112 # Example rule allowing access from your local networks.
1113 # Adapt localnet in the ACL section to list your (internal) IP networks
1114 # from where browsing should be allowed
1115 http_access allow localnet
1116 http_access allow localhost
1118 # And finally deny all other access to this proxy
1119 http_access deny all
1123 NAME: adapted_http_access http_access2
1125 LOC: Config.accessList.adapted_http
1128 Allowing or Denying access based on defined access lists
1130 Essentially identical to http_access, but runs after redirectors
1131 and ICAP/eCAP adaptation. Allowing access control based on their
1134 If not set then only http_access is used.
1137 NAME: http_reply_access
1139 LOC: Config.accessList.reply
1142 Allow replies to client requests. This is complementary to http_access.
1144 http_reply_access allow|deny [!] aclname ...
1146 NOTE: if there are no access lines present, the default is to allow
1149 If none of the access lines cause a match the opposite of the
1150 last line will apply. Thus it is good practice to end the rules
1151 with an "allow all" or "deny all" entry.
1153 This clause supports both fast and slow acl types.
1154 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1159 LOC: Config.accessList.icp
1160 DEFAULT_IF_NONE: deny all
1162 Allowing or Denying access to the ICP port based on defined
1165 icp_access allow|deny [!]aclname ...
1167 See http_access for details
1169 This clause only supports fast acl types.
1170 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1172 # Allow ICP queries from local networks only
1173 #icp_access allow localnet
1174 #icp_access deny all
1180 LOC: Config.accessList.htcp
1181 DEFAULT_IF_NONE: deny all
1183 Allowing or Denying access to the HTCP port based on defined
1186 htcp_access allow|deny [!]aclname ...
1188 See http_access for details
1190 NOTE: The default if no htcp_access lines are present is to
1191 deny all traffic. This default may cause problems with peers
1192 using the htcp option.
1194 This clause only supports fast acl types.
1195 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1197 # Allow HTCP queries from local networks only
1198 #htcp_access allow localnet
1199 #htcp_access deny all
1202 NAME: htcp_clr_access
1205 LOC: Config.accessList.htcp_clr
1206 DEFAULT_IF_NONE: deny all
1208 Allowing or Denying access to purge content using HTCP based
1209 on defined access lists
1211 htcp_clr_access allow|deny [!]aclname ...
1213 See http_access for details
1215 This clause only supports fast acl types.
1216 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1218 # Allow HTCP CLR requests from trusted peers
1219 acl htcp_clr_peer src 172.16.1.2
1220 htcp_clr_access allow htcp_clr_peer
1225 LOC: Config.accessList.miss
1228 Determins whether network access is permitted when satisfying a request.
1231 to force your neighbors to use you as a sibling instead of
1234 acl localclients src 172.16.0.0/16
1235 miss_access allow localclients
1236 miss_access deny !localclients
1238 This means only your local clients are allowed to fetch relayed/MISS
1239 replies from the network and all other clients can only fetch cached
1243 The default for this setting allows all clients who passed the
1244 http_access rules to relay via this proxy.
1246 This clause only supports fast acl types.
1247 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1250 NAME: ident_lookup_access
1253 DEFAULT_IF_NONE: deny all
1254 LOC: Ident::TheConfig.identLookup
1256 A list of ACL elements which, if matched, cause an ident
1257 (RFC 931) lookup to be performed for this request. For
1258 example, you might choose to always perform ident lookups
1259 for your main multi-user Unix boxes, but not for your Macs
1260 and PCs. By default, ident lookups are not performed for
1263 To enable ident lookups for specific client addresses, you
1264 can follow this example:
1266 acl ident_aware_hosts src 198.168.1.0/24
1267 ident_lookup_access allow ident_aware_hosts
1268 ident_lookup_access deny all
1270 Only src type ACL checks are fully supported. A srcdomain
1271 ACL might work at times, but it will not always provide
1274 This clause only supports fast acl types.
1275 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1278 NAME: reply_body_max_size
1279 COMMENT: size [acl acl...]
1282 LOC: Config.ReplyBodySize
1284 This option specifies the maximum size of a reply body. It can be
1285 used to prevent users from downloading very large files, such as
1286 MP3's and movies. When the reply headers are received, the
1287 reply_body_max_size lines are processed, and the first line where
1288 all (if any) listed ACLs are true is used as the maximum body size
1291 This size is checked twice. First when we get the reply headers,
1292 we check the content-length value. If the content length value exists
1293 and is larger than the allowed size, the request is denied and the
1294 user receives an error message that says "the request or reply
1295 is too large." If there is no content-length, and the reply
1296 size exceeds this limit, the client's connection is just closed
1297 and they will receive a partial reply.
1299 WARNING: downstream caches probably can not detect a partial reply
1300 if there is no content-length header, so they will cache
1301 partial responses and give them out as hits. You should NOT
1302 use this option if you have downstream caches.
1304 WARNING: A maximum size smaller than the size of squid's error messages
1305 will cause an infinite loop and crash squid. Ensure that the smallest
1306 non-zero value you use is greater that the maximum header size plus
1307 the size of your largest error page.
1309 If you set this parameter none (the default), there will be
1312 Configuration Format is:
1313 reply_body_max_size SIZE UNITS [acl ...]
1315 reply_body_max_size 10 MB
1321 -----------------------------------------------------------------------------
1324 NAME: http_port ascii_port
1327 LOC: Config.Sockaddr.http
1329 Usage: port [mode] [options]
1330 hostname:port [mode] [options]
1331 1.2.3.4:port [mode] [options]
1333 The socket addresses where Squid will listen for HTTP client
1334 requests. You may specify multiple socket addresses.
1335 There are three forms: port alone, hostname with port, and
1336 IP address with port. If you specify a hostname or IP
1337 address, Squid binds the socket to that specific
1338 address. Most likely, you do not need to bind to a specific
1339 address, so you can use the port number alone.
1341 If you are running Squid in accelerator mode, you
1342 probably want to listen on port 80 also, or instead.
1344 The -a command line option may be used to specify additional
1345 port(s) where Squid listens for proxy request. Such ports will
1346 be plain proxy ports with no options.
1348 You may specify multiple socket addresses on multiple lines.
1352 intercept Support for IP-Layer interception of
1353 outgoing requests without browser settings.
1354 NP: disables authentication and IPv6 on the port.
1356 tproxy Support Linux TPROXY for spoofing outgoing
1357 connections using the client IP address.
1358 NP: disables authentication and maybe IPv6 on the port.
1360 accel Accelerator / reverse proxy mode
1362 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1363 establish secure connection with the client and with
1364 the server, decrypt HTTPS messages as they pass through
1365 Squid, and treat them as unencrypted HTTP messages,
1366 becoming the man-in-the-middle.
1368 The ssl_bump option is required to fully enable
1369 bumping of CONNECT requests.
1371 Omitting the mode flag causes default forward proxy mode to be used.
1374 Accelerator Mode Options:
1376 defaultsite=domainname
1377 What to use for the Host: header if it is not present
1378 in a request. Determines what site (not origin server)
1379 accelerators should consider the default.
1381 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1383 protocol= Protocol to reconstruct accelerated requests with.
1384 Defaults to http for http_port and https for
1387 vport Virtual host port support. Using the http_port number
1388 instead of the port passed on Host: headers.
1390 vport=NN Virtual host port support. Using the specified port
1391 number instead of the port passed on Host: headers.
1394 Act as if this Squid is the origin server.
1395 This currently means generate new Date: and Expires:
1396 headers on HIT instead of adding Age:.
1398 ignore-cc Ignore request Cache-Control headers.
1400 WARNING: This option violates HTTP specifications if
1401 used in non-accelerator setups.
1403 allow-direct Allow direct forwarding in accelerator mode. Normally
1404 accelerated requests are denied direct forwarding as if
1405 never_direct was used.
1407 WARNING: this option opens accelerator mode to security
1408 vulnerabilities usually only affecting in interception
1409 mode. Make sure to protect forwarding with suitable
1410 http_access rules when using this.
1413 SSL Bump Mode Options:
1414 In addition to these options ssl-bump requires TLS/SSL options.
1416 generate-host-certificates[=<on|off>]
1417 Dynamically create SSL server certificates for the
1418 destination hosts of bumped CONNECT requests.When
1419 enabled, the cert and key options are used to sign
1420 generated certificates. Otherwise generated
1421 certificate will be selfsigned.
1422 If there is a CA certificate lifetime of the generated
1423 certificate equals lifetime of the CA certificate. If
1424 generated certificate is selfsigned lifetime is three
1426 This option is enabled by default when ssl-bump is used.
1427 See the ssl-bump option above for more information.
1429 dynamic_cert_mem_cache_size=SIZE
1430 Approximate total RAM size spent on cached generated
1431 certificates. If set to zero, caching is disabled. The
1432 default value is 4MB.
1436 cert= Path to SSL certificate (PEM format).
1438 key= Path to SSL private key file (PEM format)
1439 if not specified, the certificate file is
1440 assumed to be a combined certificate and
1443 version= The version of SSL/TLS supported
1444 1 automatic (default)
1451 cipher= Colon separated list of supported ciphers.
1452 NOTE: some ciphers such as EDH ciphers depend on
1453 additional settings. If those settings are
1454 omitted the ciphers may be silently ignored
1455 by the OpenSSL library.
1457 options= Various SSL implementation options. The most important
1459 NO_SSLv2 Disallow the use of SSLv2
1460 NO_SSLv3 Disallow the use of SSLv3
1461 NO_TLSv1 Disallow the use of TLSv1.0
1462 NO_TLSv1_1 Disallow the use of TLSv1.1
1463 NO_TLSv1_2 Disallow the use of TLSv1.2
1464 SINGLE_DH_USE Always create a new key when using
1465 temporary/ephemeral DH key exchanges
1466 ALL Enable various bug workarounds
1467 suggested as "harmless" by OpenSSL
1468 Be warned that this reduces SSL/TLS
1469 strength to some attacks.
1470 See OpenSSL SSL_CTX_set_options documentation for a
1471 complete list of options.
1473 clientca= File containing the list of CAs to use when
1474 requesting a client certificate.
1476 cafile= File containing additional CA certificates to
1477 use when verifying client certificates. If unset
1478 clientca will be used.
1480 capath= Directory containing additional CA certificates
1481 and CRL lists to use when verifying client certificates.
1483 crlfile= File of additional CRL lists to use when verifying
1484 the client certificate, in addition to CRLs stored in
1485 the capath. Implies VERIFY_CRL flag below.
1487 dhparams= File containing DH parameters for temporary/ephemeral
1488 DH key exchanges. See OpenSSL documentation for details
1489 on how to create this file.
1490 WARNING: EDH ciphers will be silently disabled if this
1493 sslflags= Various flags modifying the use of SSL:
1495 Don't request client certificates
1496 immediately, but wait until acl processing
1497 requires a certificate (not yet implemented).
1499 Don't use the default CA lists built in
1502 Don't allow for session reuse. Each connection
1503 will result in a new SSL session.
1505 Verify CRL lists when accepting client
1508 Verify CRL lists for all certificates in the
1509 client certificate chain.
1511 sslcontext= SSL session ID context identifier.
1515 connection-auth[=on|off]
1516 use connection-auth=off to tell Squid to prevent
1517 forwarding Microsoft connection oriented authentication
1518 (NTLM, Negotiate and Kerberos)
1520 disable-pmtu-discovery=
1521 Control Path-MTU discovery usage:
1522 off lets OS decide on what to do (default).
1523 transparent disable PMTU discovery when transparent
1525 always disable always PMTU discovery.
1527 In many setups of transparently intercepting proxies
1528 Path-MTU discovery can not work on traffic towards the
1529 clients. This is the case when the intercepting device
1530 does not fully track connections and fails to forward
1531 ICMP must fragment messages to the cache server. If you
1532 have such setup and experience that certain clients
1533 sporadically hang or never complete requests set
1534 disable-pmtu-discovery option to 'transparent'.
1536 name= Specifies a internal name for the port. Defaults to
1537 the port specification (port or addr:port)
1539 tcpkeepalive[=idle,interval,timeout]
1540 Enable TCP keepalive probes of idle connections.
1541 In seconds; idle is the initial time before TCP starts
1542 probing the connection, interval how often to probe, and
1543 timeout the time before giving up.
1545 If you run Squid on a dual-homed machine with an internal
1546 and an external interface we recommend you to specify the
1547 internal address:port in http_port. This way Squid will only be
1548 visible on the internal address.
1552 # Squid normally listens to port 3128
1553 http_port @DEFAULT_HTTP_PORT@
1561 LOC: Config.Sockaddr.https
1563 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1565 The socket address where Squid will listen for client requests made
1566 over TLS or SSL connections. Commonly referred to as HTTPS.
1568 This is most useful for situations where you are running squid in
1569 accelerator mode and you want to do the SSL work at the accelerator level.
1571 You may specify multiple socket addresses on multiple lines,
1572 each with their own SSL certificate and/or options.
1576 accel Accelerator / reverse proxy mode
1578 intercept Support for IP-Layer interception of
1579 outgoing requests without browser settings.
1580 NP: disables authentication and IPv6 on the port.
1582 tproxy Support Linux TPROXY for spoofing outgoing
1583 connections using the client IP address.
1584 NP: disables authentication and maybe IPv6 on the port.
1586 ssl-bump For each intercepted connection allowed by ssl_bump
1587 ACLs, establish a secure connection with the client and with
1588 the server, decrypt HTTPS messages as they pass through
1589 Squid, and treat them as unencrypted HTTP messages,
1590 becoming the man-in-the-middle.
1592 An "ssl_bump server-first" match is required to
1593 fully enable bumping of intercepted SSL connections.
1595 Requires tproxy or intercept.
1597 Omitting the mode flag causes default forward proxy mode to be used.
1600 See http_port for a list of generic options
1605 cert= Path to SSL certificate (PEM format).
1607 key= Path to SSL private key file (PEM format)
1608 if not specified, the certificate file is
1609 assumed to be a combined certificate and
1612 version= The version of SSL/TLS supported
1613 1 automatic (default)
1618 cipher= Colon separated list of supported ciphers.
1620 options= Various SSL engine options. The most important
1622 NO_SSLv2 Disallow the use of SSLv2
1623 NO_SSLv3 Disallow the use of SSLv3
1624 NO_TLSv1 Disallow the use of TLSv1
1625 SINGLE_DH_USE Always create a new key when using
1626 temporary/ephemeral DH key exchanges
1627 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1628 documentation for a complete list of options.
1630 clientca= File containing the list of CAs to use when
1631 requesting a client certificate.
1633 cafile= File containing additional CA certificates to
1634 use when verifying client certificates. If unset
1635 clientca will be used.
1637 capath= Directory containing additional CA certificates
1638 and CRL lists to use when verifying client certificates.
1640 crlfile= File of additional CRL lists to use when verifying
1641 the client certificate, in addition to CRLs stored in
1642 the capath. Implies VERIFY_CRL flag below.
1644 dhparams= File containing DH parameters for temporary/ephemeral
1647 sslflags= Various flags modifying the use of SSL:
1649 Don't request client certificates
1650 immediately, but wait until acl processing
1651 requires a certificate (not yet implemented).
1653 Don't use the default CA lists built in
1656 Don't allow for session reuse. Each connection
1657 will result in a new SSL session.
1659 Verify CRL lists when accepting client
1662 Verify CRL lists for all certificates in the
1663 client certificate chain.
1665 sslcontext= SSL session ID context identifier.
1667 generate-host-certificates[=<on|off>]
1668 Dynamically create SSL server certificates for the
1669 destination hosts of bumped SSL requests.When
1670 enabled, the cert and key options are used to sign
1671 generated certificates. Otherwise generated
1672 certificate will be selfsigned.
1673 If there is CA certificate life time of generated
1674 certificate equals lifetime of CA certificate. If
1675 generated certificate is selfsigned lifetime is three
1677 This option is enabled by default when SslBump is used.
1678 See the sslBump option above for more information.
1680 dynamic_cert_mem_cache_size=SIZE
1681 Approximate total RAM size spent on cached generated
1682 certificates. If set to zero, caching is disabled. The
1683 default value is 4MB.
1685 See http_port for a list of available options.
1688 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1691 LOC: Ip::Qos::TheConfig.tosToServer
1693 Allows you to select a TOS/Diffserv value for packets outgoing
1694 on the server side, based on an ACL.
1696 tcp_outgoing_tos ds-field [!]aclname ...
1698 Example where normal_service_net uses the TOS value 0x00
1699 and good_service_net uses 0x20
1701 acl normal_service_net src 10.0.0.0/24
1702 acl good_service_net src 10.0.1.0/24
1703 tcp_outgoing_tos 0x00 normal_service_net
1704 tcp_outgoing_tos 0x20 good_service_net
1706 TOS/DSCP values really only have local significance - so you should
1707 know what you're specifying. For more information, see RFC2474,
1708 RFC2475, and RFC3260.
1710 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1711 "default" to use whatever default your host has. Note that in
1712 practice often only multiples of 4 is usable as the two rightmost bits
1713 have been redefined for use by ECN (RFC 3168 section 23.1).
1715 Processing proceeds in the order specified, and stops at first fully
1719 NAME: clientside_tos
1722 LOC: Ip::Qos::TheConfig.tosToClient
1724 Allows you to select a TOS/Diffserv value for packets being transmitted
1725 on the client-side, based on an ACL.
1727 clientside_tos ds-field [!]aclname ...
1729 Example where normal_service_net uses the TOS value 0x00
1730 and good_service_net uses 0x20
1732 acl normal_service_net src 10.0.0.0/24
1733 acl good_service_net src 10.0.1.0/24
1734 clientside_tos 0x00 normal_service_net
1735 clientside_tos 0x20 good_service_net
1737 Note: This feature is incompatible with qos_flows. Any TOS values set here
1738 will be overwritten by TOS values in qos_flows.
1741 NAME: tcp_outgoing_mark
1743 IFDEF: SO_MARK&&USE_LIBCAP
1745 LOC: Ip::Qos::TheConfig.nfmarkToServer
1747 Allows you to apply a Netfilter mark value to outgoing packets
1748 on the server side, based on an ACL.
1750 tcp_outgoing_mark mark-value [!]aclname ...
1752 Example where normal_service_net uses the mark value 0x00
1753 and good_service_net uses 0x20
1755 acl normal_service_net src 10.0.0.0/24
1756 acl good_service_net src 10.0.1.0/24
1757 tcp_outgoing_mark 0x00 normal_service_net
1758 tcp_outgoing_mark 0x20 good_service_net
1761 NAME: clientside_mark
1763 IFDEF: SO_MARK&&USE_LIBCAP
1765 LOC: Ip::Qos::TheConfig.nfmarkToClient
1767 Allows you to apply a Netfilter mark value to packets being transmitted
1768 on the client-side, based on an ACL.
1770 clientside_mark mark-value [!]aclname ...
1772 Example where normal_service_net uses the mark value 0x00
1773 and good_service_net uses 0x20
1775 acl normal_service_net src 10.0.0.0/24
1776 acl good_service_net src 10.0.1.0/24
1777 clientside_mark 0x00 normal_service_net
1778 clientside_mark 0x20 good_service_net
1780 Note: This feature is incompatible with qos_flows. Any mark values set here
1781 will be overwritten by mark values in qos_flows.
1788 LOC: Ip::Qos::TheConfig
1790 Allows you to select a TOS/DSCP value to mark outgoing
1791 connections with, based on where the reply was sourced. For
1792 platforms using netfilter, allows you to set a netfilter mark
1793 value instead of, or in addition to, a TOS value.
1795 TOS values really only have local significance - so you should
1796 know what you're specifying. For more information, see RFC2474,
1797 RFC2475, and RFC3260.
1799 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1800 in practice often only multiples of 4 is usable as the two rightmost bits
1801 have been redefined for use by ECN (RFC 3168 section 23.1).
1803 Mark values can be any unsigned 32-bit integer value.
1805 This setting is configured by setting the following values:
1807 tos|mark Whether to set TOS or netfilter mark values
1809 local-hit=0xFF Value to mark local cache hits.
1811 sibling-hit=0xFF Value to mark hits from sibling peers.
1813 parent-hit=0xFF Value to mark hits from parent peers.
1815 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1816 over the preserve-miss feature (see below), unless
1817 mask is specified, in which case only the bits
1818 specified in the mask are written.
1820 The TOS variant of the following features are only possible on Linux
1821 and require your kernel to be patched with the TOS preserving ZPH
1822 patch, available from http://zph.bratcheda.org
1823 No patch is needed to preserve the netfilter mark, which will work
1824 with all variants of netfilter.
1826 disable-preserve-miss
1827 This option disables the preservation of the TOS or netfilter
1828 mark. By default, the existing TOS or netfilter mark value of
1829 the response coming from the remote server will be retained
1830 and masked with miss-mark.
1831 NOTE: in the case of a netfilter mark, the mark must be set on
1832 the connection (using the CONNMARK target) not on the packet
1836 Allows you to mask certain bits in the TOS or mark value
1837 received from the remote server, before copying the value to
1838 the TOS sent towards clients.
1839 Default for tos: 0xFF (TOS from server is not changed).
1840 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1842 All of these features require the --enable-zph-qos compilation flag
1843 (enabled by default). Netfilter marking also requires the
1844 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1845 libcap 2.09+ (--with-libcap).
1849 NAME: tcp_outgoing_address
1852 LOC: Config.accessList.outgoing_address
1854 Allows you to map requests to different outgoing IP addresses
1855 based on the username or source address of the user making
1858 tcp_outgoing_address ipaddr [[!]aclname] ...
1861 Forwarding clients with dedicated IPs for certain subnets.
1863 acl normal_service_net src 10.0.0.0/24
1864 acl good_service_net src 10.0.2.0/24
1866 tcp_outgoing_address 2001:db8::c001 good_service_net
1867 tcp_outgoing_address 10.1.0.2 good_service_net
1869 tcp_outgoing_address 2001:db8::beef normal_service_net
1870 tcp_outgoing_address 10.1.0.1 normal_service_net
1872 tcp_outgoing_address 2001:db8::1
1873 tcp_outgoing_address 10.1.0.3
1875 Processing proceeds in the order specified, and stops at first fully
1878 Squid will add an implicit IP version test to each line.
1879 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1880 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1883 NOTE: The use of this directive using client dependent ACLs is
1884 incompatible with the use of server side persistent connections. To
1885 ensure correct results it is best to set server_persistent_connections
1886 to off when using this directive in such configurations.
1888 NOTE: The use of this directive to set a local IP on outgoing TCP links
1889 is incompatible with using TPROXY to set client IP out outbound TCP links.
1890 When needing to contact peers use the no-tproxy cache_peer option and the
1891 client_dst_passthru directive re-enable normal forwarding such as this.
1895 NAME: host_verify_strict
1898 LOC: Config.onoff.hostStrictVerify
1900 Regardless of this option setting, when dealing with intercepted
1901 traffic, Squid always verifies that the destination IP address matches
1902 the Host header domain or IP (called 'authority form URL').
1904 This enforcement is performed to satisfy a MUST-level requirement in
1905 RFC 2616 section 14.23: "The Host field value MUST represent the naming
1906 authority of the origin server or gateway given by the original URL".
1909 Squid always responds with an HTTP 409 (Conflict) error
1910 page and logs a security warning if there is no match.
1912 Squid verifies that the destination IP address matches
1913 the Host header for forward-proxy and reverse-proxy traffic
1914 as well. For those traffic types, Squid also enables the
1915 following checks, comparing the corresponding Host header
1916 and Request-URI components:
1918 * The host names (domain or IP) must be identical,
1919 but valueless or missing Host header disables all checks.
1920 For the two host names to match, both must be either IP
1923 * Port numbers must be identical, but if a port is missing
1924 the scheme-default port is assumed.
1927 When set to OFF (the default):
1928 Squid allows suspicious requests to continue but logs a
1929 security warning and blocks caching of the response.
1931 * Forward-proxy traffic is not checked at all.
1933 * Reverse-proxy traffic is not checked at all.
1935 * Intercepted traffic which passes verification is handled
1936 according to client_dst_passthru.
1938 * Intercepted requests which fail verification are sent
1939 to the client original destination instead of DIRECT.
1940 This overrides 'client_dst_passthru off'.
1942 For now suspicious intercepted CONNECT requests are always
1943 responded to with an HTTP 409 (Conflict) error page.
1948 As described in CVE-2009-0801 when the Host: header alone is used
1949 to determine the destination of a request it becomes trivial for
1950 malicious scripts on remote websites to bypass browser same-origin
1951 security policy and sandboxing protections.
1953 The cause of this is that such applets are allowed to perform their
1954 own HTTP stack, in which case the same-origin policy of the browser
1955 sandbox only verifies that the applet tries to contact the same IP
1956 as from where it was loaded at the IP level. The Host: header may
1957 be different from the connected IP and approved origin.
1961 NAME: client_dst_passthru
1964 LOC: Config.onoff.client_dst_passthru
1966 With NAT or TPROXY intercepted traffic Squid may pass the request
1967 directly to the original client destination IP or seek a faster
1968 source using the HTTP Host header.
1970 Using Host to locate alternative servers can provide faster
1971 connectivity with a range of failure recovery options.
1972 But can also lead to connectivity trouble when the client and
1973 server are attempting stateful interactions unaware of the proxy.
1975 This option (on by default) prevents alternative DNS entries being
1976 located to send intercepted traffic DIRECT to an origin server.
1977 The clients original destination IP and port will be used instead.
1979 Regardless of this option setting, when dealing with intercepted
1980 traffic Squid will verify the Host: header and any traffic which
1981 fails Host verification will be treated as if this option were ON.
1983 see host_verify_strict for details on the verification process.
1988 -----------------------------------------------------------------------------
1991 NAME: ssl_unclean_shutdown
1995 LOC: Config.SSL.unclean_shutdown
1997 Some browsers (especially MSIE) bugs out on SSL shutdown
2004 LOC: Config.SSL.ssl_engine
2007 The OpenSSL engine to use. You will need to set this if you
2008 would like to use hardware SSL acceleration for example.
2011 NAME: sslproxy_client_certificate
2014 LOC: Config.ssl_client.cert
2017 Client SSL Certificate to use when proxying https:// URLs
2020 NAME: sslproxy_client_key
2023 LOC: Config.ssl_client.key
2026 Client SSL Key to use when proxying https:// URLs
2029 NAME: sslproxy_version
2032 LOC: Config.ssl_client.version
2035 SSL version level to use when proxying https:// URLs
2037 The versions of SSL/TLS supported:
2039 1 automatic (default)
2047 NAME: sslproxy_options
2050 LOC: Config.ssl_client.options
2053 SSL implementation options to use when proxying https:// URLs
2055 The most important being:
2057 NO_SSLv2 Disallow the use of SSLv2
2058 NO_SSLv3 Disallow the use of SSLv3
2059 NO_TLSv1 Disallow the use of TLSv1.0
2060 NO_TLSv1_1 Disallow the use of TLSv1.1
2061 NO_TLSv1_2 Disallow the use of TLSv1.2
2063 Always create a new key when using temporary/ephemeral
2066 Disable use of RFC5077 session tickets. Some servers
2067 may have problems understanding the TLS extension due
2068 to ambiguous specification in RFC4507.
2069 ALL Enable various bug workarounds suggested as "harmless"
2070 by OpenSSL. Be warned that this may reduce SSL/TLS
2071 strength to some attacks.
2073 See the OpenSSL SSL_CTX_set_options documentation for a
2074 complete list of possible options.
2077 NAME: sslproxy_cipher
2080 LOC: Config.ssl_client.cipher
2083 SSL cipher list to use when proxying https:// URLs
2085 Colon separated list of supported ciphers.
2088 NAME: sslproxy_cafile
2091 LOC: Config.ssl_client.cafile
2094 file containing CA certificates to use when verifying server
2095 certificates while proxying https:// URLs
2098 NAME: sslproxy_capath
2101 LOC: Config.ssl_client.capath
2104 directory containing CA certificates to use when verifying
2105 server certificates while proxying https:// URLs
2110 TYPE: sslproxy_ssl_bump
2111 LOC: Config.accessList.ssl_bump
2114 This option is consulted when a CONNECT request is received on
2115 an http_port (or a new connection is intercepted at an
2116 https_port), provided that port was configured with an ssl-bump
2117 flag. The subsequent data on the connection is either treated as
2118 HTTPS and decrypted OR tunneled at TCP level without decryption,
2119 depending on the first bumping "mode" which ACLs match.
2121 ssl_bump <mode> [!]acl ...
2123 The following bumping modes are supported:
2126 Allow bumping of the connection. Establish a secure connection
2127 with the client first, then connect to the server. This old mode
2128 does not allow Squid to mimic server SSL certificate and does
2129 not work with intercepted SSL connections.
2132 Allow bumping of the connection. Establish a secure connection
2133 with the server first, then establish a secure connection with
2134 the client, using a mimicked server certificate. Works with both
2135 CONNECT requests and intercepted SSL connections.
2138 Become a TCP tunnel without decoding the connection.
2139 Works with both CONNECT requests and intercepted SSL
2140 connections. This is the default behavior when no
2141 ssl_bump option is given or no ssl_bump ACLs match.
2143 By default, no connections are bumped.
2145 The first matching ssl_bump option wins. If no ACLs match, the
2146 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2147 does not have an implicit "negate the last given option" rule. You
2148 must make that rule explicit if you convert old ssl_bump allow/deny
2149 rules that rely on such an implicit rule.
2151 This clause supports both fast and slow acl types.
2152 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2154 See also: http_port ssl-bump, https_port ssl-bump
2157 # Example: Bump all requests except those originating from
2158 # localhost and those going to example.com.
2160 acl broken_sites dstdomain .example.com
2161 ssl_bump none localhost
2162 ssl_bump none broken_sites
2163 ssl_bump server-first all
2166 NAME: sslproxy_flags
2169 LOC: Config.ssl_client.flags
2172 Various flags modifying the use of SSL while proxying https:// URLs:
2173 DONT_VERIFY_PEER Accept certificates that fail verification.
2174 For refined control, see sslproxy_cert_error.
2175 NO_DEFAULT_CA Don't use the default CA list built in
2179 NAME: sslproxy_cert_error
2182 LOC: Config.ssl_client.cert_error
2185 Use this ACL to bypass server certificate validation errors.
2187 For example, the following lines will bypass all validation errors
2188 when talking to servers for example.com. All other
2189 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2191 acl BrokenButTrustedServers dstdomain example.com
2192 sslproxy_cert_error allow BrokenButTrustedServers
2193 sslproxy_cert_error deny all
2195 This clause only supports fast acl types.
2196 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2197 Using slow acl types may result in server crashes
2199 Without this option, all server certificate validation errors
2200 terminate the transaction. Bypassing validation errors is dangerous
2201 because an error usually implies that the server cannot be trusted and
2202 the connection may be insecure.
2204 See also: sslproxy_flags and DONT_VERIFY_PEER.
2206 Default setting: sslproxy_cert_error deny all
2209 NAME: sslproxy_cert_sign
2212 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2213 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2214 POSTSCRIPTUM: signTrusted all
2215 TYPE: sslproxy_cert_sign
2216 LOC: Config.ssl_client.cert_sign
2219 sslproxy_cert_sign <signing algorithm> acl ...
2221 The following certificate signing algorithms are supported:
2223 Sign using the configured CA certificate which is usually
2224 placed in and trusted by end-user browsers. This is the
2225 default for trusted origin server certificates.
2227 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2228 This is the default for untrusted origin server certificates
2229 that are not self-signed (see ssl::certUntrusted).
2231 Sign using a self-signed certificate with the right CN to
2232 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2233 browser. This is the default for self-signed origin server
2234 certificates (see ssl::certSelfSigned).
2236 This clause only supports fast acl types.
2238 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2239 signing algorithm to generate the certificate and ignores all
2240 subsequent sslproxy_cert_sign options (the first match wins). If no
2241 acl(s) match, the default signing algorithm is determined by errors
2242 detected when obtaining and validating the origin server certificate.
2244 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2245 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2246 CONNECT request that carries a domain name. In all other cases (CONNECT
2247 to an IP address or an intercepted SSL connection), Squid cannot detect
2248 the domain mismatch at certificate generation time when
2249 bump-server-first is used.
2252 NAME: sslproxy_cert_adapt
2255 TYPE: sslproxy_cert_adapt
2256 LOC: Config.ssl_client.cert_adapt
2259 sslproxy_cert_adapt <adaptation algorithm> acl ...
2261 The following certificate adaptation algorithms are supported:
2263 Sets the "Not After" property to the "Not After" property of
2264 the CA certificate used to sign generated certificates.
2266 Sets the "Not Before" property to the "Not Before" property of
2267 the CA certificate used to sign generated certificates.
2268 setCommonName or setCommonName{CN}
2269 Sets Subject.CN property to the host name specified as a
2270 CN parameter or, if no explicit CN parameter was specified,
2271 extracted from the CONNECT request. It is a misconfiguration
2272 to use setCommonName without an explicit parameter for
2273 intercepted or tproxied SSL connections.
2275 This clause only supports fast acl types.
2277 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2278 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2279 corresponding adaptation algorithm to generate the certificate and
2280 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2281 group (i.e., the first match wins within each algorithm group). If no
2282 acl(s) match, the default mimicking action takes place.
2284 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2285 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2286 CONNECT request that carries a domain name. In all other cases (CONNECT
2287 to an IP address or an intercepted SSL connection), Squid cannot detect
2288 the domain mismatch at certificate generation time when
2289 bump-server-first is used.
2292 NAME: sslpassword_program
2295 LOC: Config.Program.ssl_password
2298 Specify a program used for entering SSL key passphrases
2299 when using encrypted SSL certificate keys. If not specified
2300 keys must either be unencrypted, or Squid started with the -N
2301 option to allow it to query interactively for the passphrase.
2303 The key file name is given as argument to the program allowing
2304 selection of the right password if you have multiple encrypted
2309 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2310 -----------------------------------------------------------------------------
2313 NAME: sslcrtd_program
2316 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2317 LOC: Ssl::TheConfig.ssl_crtd
2319 Specify the location and options of the executable for ssl_crtd process.
2320 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2321 For more information use:
2322 @DEFAULT_SSL_CRTD@ -h
2325 NAME: sslcrtd_children
2326 TYPE: HelperChildConfig
2328 DEFAULT: 32 startup=5 idle=1
2329 LOC: Ssl::TheConfig.ssl_crtdChildren
2331 The maximum number of processes spawn to service ssl server.
2332 The maximum this may be safely set to is 32.
2334 The startup= and idle= options allow some measure of skew in your
2339 Sets the minimum number of processes to spawn when Squid
2340 starts or reconfigures. When set to zero the first request will
2341 cause spawning of the first child process to handle it.
2343 Starting too few children temporary slows Squid under load while it
2344 tries to spawn enough additional processes to cope with traffic.
2348 Sets a minimum of how many processes Squid is to try and keep available
2349 at all times. When traffic begins to rise above what the existing
2350 processes can handle this many more will be spawned up to the maximum
2351 configured. A minimum setting of 1 is required.
2353 You must have at least one ssl_crtd process.
2356 NAME: sslcrtvalidator_program
2360 LOC: Ssl::TheConfig.ssl_crt_validator
2362 Specify the location and options of the executable for ssl_crt_validator
2366 NAME: sslcrtvalidator_children
2367 TYPE: HelperChildConfig
2369 DEFAULT: 32 startup=5 idle=1 concurrency=1
2370 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2372 The maximum number of processes spawn to service ssl server.
2373 The maximum this may be safely set to is 32.
2375 The startup= and idle= options allow some measure of skew in your
2380 Sets the minimum number of processes to spawn when Squid
2381 starts or reconfigures. When set to zero the first request will
2382 cause spawning of the first child process to handle it.
2384 Starting too few children temporary slows Squid under load while it
2385 tries to spawn enough additional processes to cope with traffic.
2389 Sets a minimum of how many processes Squid is to try and keep available
2390 at all times. When traffic begins to rise above what the existing
2391 processes can handle this many more will be spawned up to the maximum
2392 configured. A minimum setting of 1 is required.
2396 The number of requests each certificate validator helper can handle in
2397 parallel. Defaults to 0 which indicates the certficate validator
2398 is a old-style single threaded redirector.
2400 When this directive is set to a value >= 1 then the protocol
2401 used to communicate with the helper is modified to include
2402 a request ID in front of the request/response. The request
2403 ID from the request must be echoed back with the response
2406 You must have at least one ssl_crt_validator process.
2410 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2411 -----------------------------------------------------------------------------
2419 To specify other caches in a hierarchy, use the format:
2421 cache_peer hostname type http-port icp-port [options]
2426 # hostname type port port options
2427 # -------------------- -------- ----- ----- -----------
2428 cache_peer parent.foo.net parent 3128 3130 default
2429 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2430 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2431 cache_peer example.com parent 80 0 default
2432 cache_peer cdn.example.com sibling 3128 0
2434 type: either 'parent', 'sibling', or 'multicast'.
2436 proxy-port: The port number where the peer accept HTTP requests.
2437 For other Squid proxies this is usually 3128
2438 For web servers this is usually 80
2440 icp-port: Used for querying neighbor caches about objects.
2441 Set to 0 if the peer does not support ICP or HTCP.
2442 See ICP and HTCP options below for additional details.
2445 ==== ICP OPTIONS ====
2447 You MUST also set icp_port and icp_access explicitly when using these options.
2448 The defaults will prevent peer traffic using ICP.
2451 no-query Disable ICP queries to this neighbor.
2454 Indicates the named peer is a member of a multicast group.
2455 ICP queries will not be sent directly to the peer, but ICP
2456 replies will be accepted from it.
2458 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2459 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2462 To only send ICP queries to this neighbor infrequently.
2463 This is used to keep the neighbor round trip time updated
2464 and is usually used in conjunction with weighted-round-robin.
2467 ==== HTCP OPTIONS ====
2469 You MUST also set htcp_port and htcp_access explicitly when using these options.
2470 The defaults will prevent peer traffic using HTCP.
2473 htcp Send HTCP, instead of ICP, queries to the neighbor.
2474 You probably also want to set the "icp-port" to 4827
2475 instead of 3130. This directive accepts a comma separated
2476 list of options described below.
2478 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2480 htcp=no-clr Send HTCP to the neighbor but without
2481 sending any CLR requests. This cannot be used with
2484 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2485 This cannot be used with no-clr.
2488 Send HTCP to the neighbor including CLRs but only when
2489 they do not result from PURGE requests.
2492 Forward any HTCP CLR requests this proxy receives to the peer.
2495 ==== PEER SELECTION METHODS ====
2497 The default peer selection method is ICP, with the first responding peer
2498 being used as source. These options can be used for better load balancing.
2501 default This is a parent cache which can be used as a "last-resort"
2502 if a peer cannot be located by any of the peer-selection methods.
2503 If specified more than once, only the first is used.
2505 round-robin Load-Balance parents which should be used in a round-robin
2506 fashion in the absence of any ICP queries.
2507 weight=N can be used to add bias.
2509 weighted-round-robin
2510 Load-Balance parents which should be used in a round-robin
2511 fashion with the frequency of each parent being based on the
2512 round trip time. Closer parents are used more often.
2513 Usually used for background-ping parents.
2514 weight=N can be used to add bias.
2516 carp Load-Balance parents which should be used as a CARP array.
2517 The requests will be distributed among the parents based on the
2518 CARP load balancing hash function based on their weight.
2520 userhash Load-balance parents based on the client proxy_auth or ident username.
2522 sourcehash Load-balance parents based on the client source IP.
2525 To be used only for cache peers of type "multicast".
2526 ALL members of this multicast group have "sibling"
2527 relationship with it, not "parent". This is to a multicast
2528 group when the requested object would be fetched only from
2529 a "parent" cache, anyway. It's useful, e.g., when
2530 configuring a pool of redundant Squid proxies, being
2531 members of the same multicast group.
2534 ==== PEER SELECTION OPTIONS ====
2536 weight=N use to affect the selection of a peer during any weighted
2537 peer-selection mechanisms.
2538 The weight must be an integer; default is 1,
2539 larger weights are favored more.
2540 This option does not affect parent selection if a peering
2541 protocol is not in use.
2543 basetime=N Specify a base amount to be subtracted from round trip
2545 It is subtracted before division by weight in calculating
2546 which parent to fectch from. If the rtt is less than the
2547 base time the rtt is set to a minimal value.
2549 ttl=N Specify a TTL to use when sending multicast ICP queries
2551 Only useful when sending to a multicast group.
2552 Because we don't accept ICP replies from random
2553 hosts, you must configure other group members as
2554 peers with the 'multicast-responder' option.
2556 no-delay To prevent access to this neighbor from influencing the
2559 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2560 enabled) for this host from the specified URL rather
2561 than the Squid default location.
2564 ==== CARP OPTIONS ====
2566 carp-key=key-specification
2567 use a different key than the full URL to hash against the peer.
2568 the key-specification is a comma-separated list of the keywords
2569 scheme, host, port, path, params
2570 Order is not important.
2572 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2574 originserver Causes this parent to be contacted as an origin server.
2575 Meant to be used in accelerator setups when the peer
2579 Set the Host header of requests forwarded to this peer.
2580 Useful in accelerator setups where the server (peer)
2581 expects a certain domain name but clients may request
2582 others. ie example.com or www.example.com
2584 no-digest Disable request of cache digests.
2587 Disables requesting ICMP RTT database (NetDB).
2590 ==== AUTHENTICATION OPTIONS ====
2593 If this is a personal/workgroup proxy and your parent
2594 requires proxy authentication.
2596 Note: The string can include URL escapes (i.e. %20 for
2597 spaces). This also means % must be written as %%.
2600 Send login details received from client to this peer.
2601 Both Proxy- and WWW-Authorization headers are passed
2602 without alteration to the peer.
2603 Authentication is not required by Squid for this to work.
2605 Note: This will pass any form of authentication but
2606 only Basic auth will work through a proxy unless the
2607 connection-auth options are also used.
2609 login=PASS Send login details received from client to this peer.
2610 Authentication is not required by this option.
2612 If there are no client-provided authentication headers
2613 to pass on, but username and password are available
2614 from an external ACL user= and password= result tags
2615 they may be sent instead.
2617 Note: To combine this with proxy_auth both proxies must
2618 share the same user database as HTTP only allows for
2619 a single login (one for proxy, one for origin server).
2620 Also be warned this will expose your users proxy
2621 password to the peer. USE WITH CAUTION
2624 Send the username to the upstream cache, but with a
2625 fixed password. This is meant to be used when the peer
2626 is in another administrative domain, but it is still
2627 needed to identify each user.
2628 The star can optionally be followed by some extra
2629 information which is added to the username. This can
2630 be used to identify this proxy to the peer, similar to
2631 the login=username:password option above.
2634 If this is a personal/workgroup proxy and your parent
2635 requires a secure proxy authentication.
2636 The first principal from the default keytab or defined by
2637 the environment variable KRB5_KTNAME will be used.
2639 WARNING: The connection may transmit requests from multiple
2640 clients. Negotiate often assumes end-to-end authentication
2641 and a single-client. Which is not strictly true here.
2643 login=NEGOTIATE:principal_name
2644 If this is a personal/workgroup proxy and your parent
2645 requires a secure proxy authentication.
2646 The principal principal_name from the default keytab or
2647 defined by the environment variable KRB5_KTNAME will be
2650 WARNING: The connection may transmit requests from multiple
2651 clients. Negotiate often assumes end-to-end authentication
2652 and a single-client. Which is not strictly true here.
2654 connection-auth=on|off
2655 Tell Squid that this peer does or not support Microsoft
2656 connection oriented authentication, and any such
2657 challenges received from there should be ignored.
2658 Default is auto to automatically determine the status
2662 ==== SSL / HTTPS / TLS OPTIONS ====
2664 ssl Encrypt connections to this peer with SSL/TLS.
2666 sslcert=/path/to/ssl/certificate
2667 A client SSL certificate to use when connecting to
2670 sslkey=/path/to/ssl/key
2671 The private SSL key corresponding to sslcert above.
2672 If 'sslkey' is not specified 'sslcert' is assumed to
2673 reference a combined file containing both the
2674 certificate and the key.
2676 sslversion=1|2|3|4|5|6
2677 The SSL version to use when connecting to this peer
2678 1 = automatic (default)
2685 sslcipher=... The list of valid SSL ciphers to use when connecting
2688 ssloptions=... Specify various SSL implementation options:
2690 NO_SSLv2 Disallow the use of SSLv2
2691 NO_SSLv3 Disallow the use of SSLv3
2692 NO_TLSv1 Disallow the use of TLSv1.0
2693 NO_TLSv1_1 Disallow the use of TLSv1.1
2694 NO_TLSv1_2 Disallow the use of TLSv1.2
2696 Always create a new key when using
2697 temporary/ephemeral DH key exchanges
2698 ALL Enable various bug workarounds
2699 suggested as "harmless" by OpenSSL
2700 Be warned that this reduces SSL/TLS
2701 strength to some attacks.
2703 See the OpenSSL SSL_CTX_set_options documentation for a
2706 sslcafile=... A file containing additional CA certificates to use
2707 when verifying the peer certificate.
2709 sslcapath=... A directory containing additional CA certificates to
2710 use when verifying the peer certificate.
2712 sslcrlfile=... A certificate revocation list file to use when
2713 verifying the peer certificate.
2715 sslflags=... Specify various flags modifying the SSL implementation:
2718 Accept certificates even if they fail to
2721 Don't use the default CA list built in
2724 Don't verify the peer certificate
2725 matches the server name
2727 ssldomain= The peer name as advertised in it's certificate.
2728 Used for verifying the correctness of the received peer
2729 certificate. If not specified the peer hostname will be
2733 Enable the "Front-End-Https: On" header needed when
2734 using Squid as a SSL frontend in front of Microsoft OWA.
2735 See MS KB document Q307347 for details on this header.
2736 If set to auto the header will only be added if the
2737 request is forwarded as a https:// URL.
2740 ==== GENERAL OPTIONS ====
2743 A peer-specific connect timeout.
2744 Also see the peer_connect_timeout directive.
2746 connect-fail-limit=N
2747 How many times connecting to a peer must fail before
2748 it is marked as down. Default is 10.
2750 allow-miss Disable Squid's use of only-if-cached when forwarding
2751 requests to siblings. This is primarily useful when
2752 icp_hit_stale is used by the sibling. To extensive use
2753 of this option may result in forwarding loops, and you
2754 should avoid having two-way peerings with this option.
2755 For example to deny peer usage on requests from peer
2756 by denying cache_peer_access if the source is a peer.
2758 max-conn=N Limit the amount of connections Squid may open to this
2761 name=xxx Unique name for the peer.
2762 Required if you have multiple peers on the same host
2763 but different ports.
2764 This name can be used in cache_peer_access and similar
2765 directives to dentify the peer.
2766 Can be used by outgoing access controls through the
2769 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2770 requests to this peer. Use normal address selection instead.
2772 proxy-only objects fetched from the peer will not be stored locally.
2776 NAME: cache_peer_domain cache_host_domain
2781 Use to limit the domains for which a neighbor cache will be
2784 cache_peer_domain cache-host domain [domain ...]
2785 cache_peer_domain cache-host !domain
2787 For example, specifying
2789 cache_peer_domain parent.foo.net .edu
2791 has the effect such that UDP query packets are sent to
2792 'bigserver' only when the requested object exists on a
2793 server in the .edu domain. Prefixing the domainname
2794 with '!' means the cache will be queried for objects
2797 NOTE: * Any number of domains may be given for a cache-host,
2798 either on the same or separate lines.
2799 * When multiple domains are given for a particular
2800 cache-host, the first matched domain is applied.
2801 * Cache hosts with no domain restrictions are queried
2803 * There are no defaults.
2804 * There is also a 'cache_peer_access' tag in the ACL
2808 NAME: cache_peer_access
2813 Similar to 'cache_peer_domain' but provides more flexibility by
2816 cache_peer_access cache-host allow|deny [!]aclname ...
2818 The syntax is identical to 'http_access' and the other lists of
2819 ACL elements. See the comments for 'http_access' below, or
2820 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2823 NAME: neighbor_type_domain
2824 TYPE: hostdomaintype
2828 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2830 Modifying the neighbor type for specific domains is now
2831 possible. You can treat some domains differently than the
2832 default neighbor type specified on the 'cache_peer' line.
2833 Normally it should only be necessary to list domains which
2834 should be treated differently because the default neighbor type
2835 applies for hostnames which do not match domains listed here.
2838 cache_peer cache.foo.org parent 3128 3130
2839 neighbor_type_domain cache.foo.org sibling .com .net
2840 neighbor_type_domain cache.foo.org sibling .au .de
2843 NAME: dead_peer_timeout
2847 LOC: Config.Timeout.deadPeer
2849 This controls how long Squid waits to declare a peer cache
2850 as "dead." If there are no ICP replies received in this
2851 amount of time, Squid will declare the peer dead and not
2852 expect to receive any further ICP replies. However, it
2853 continues to send ICP queries, and will mark the peer as
2854 alive upon receipt of the first subsequent ICP reply.
2856 This timeout also affects when Squid expects to receive ICP
2857 replies from peers. If more than 'dead_peer' seconds have
2858 passed since the last ICP reply was received, Squid will not
2859 expect to receive an ICP reply on the next query. Thus, if
2860 your time between requests is greater than this timeout, you
2861 will see a lot of requests sent DIRECT to origin servers
2862 instead of to your parents.
2865 NAME: forward_max_tries
2868 LOC: Config.forward_max_tries
2870 Controls how many different forward paths Squid will try
2871 before giving up. See also forward_timeout.
2873 NOTE: connect_retries (default: none) can make each of these
2874 possible forwarding paths be tried multiple times.
2877 NAME: hierarchy_stoplist
2880 LOC: Config.hierarchy_stoplist
2882 A list of words which, if found in a URL, cause the object to
2883 be handled directly by this cache. In other words, use this
2884 to not query neighbor caches for certain objects. You may
2885 list this option multiple times.
2888 hierarchy_stoplist cgi-bin ?
2890 Note: never_direct overrides this option.
2894 MEMORY CACHE OPTIONS
2895 -----------------------------------------------------------------------------
2902 LOC: Config.memMaxSize
2904 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2905 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2906 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2907 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2909 'cache_mem' specifies the ideal amount of memory to be used
2911 * In-Transit objects
2913 * Negative-Cached objects
2915 Data for these objects are stored in 4 KB blocks. This
2916 parameter specifies the ideal upper limit on the total size of
2917 4 KB blocks allocated. In-Transit objects take the highest
2920 In-transit objects have priority over the others. When
2921 additional space is needed for incoming data, negative-cached
2922 and hot objects will be released. In other words, the
2923 negative-cached and hot objects will fill up any unused space
2924 not needed for in-transit objects.
2926 If circumstances require, this limit will be exceeded.
2927 Specifically, if your incoming request rate requires more than
2928 'cache_mem' of memory to hold in-transit objects, Squid will
2929 exceed this limit to satisfy the new requests. When the load
2930 decreases, blocks will be freed until the high-water mark is
2931 reached. Thereafter, blocks will be used to store hot
2934 If shared memory caching is enabled, Squid does not use the shared
2935 cache space for in-transit objects, but they still consume as much
2936 local memory as they need. For more details about the shared memory
2937 cache, see memory_cache_shared.
2940 NAME: maximum_object_size_in_memory
2944 LOC: Config.Store.maxInMemObjSize
2946 Objects greater than this size will not be attempted to kept in
2947 the memory cache. This should be set high enough to keep objects
2948 accessed frequently in memory to improve performance whilst low
2949 enough to keep larger objects from hoarding cache_mem.
2952 NAME: memory_cache_shared
2955 LOC: Config.memShared
2957 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
2959 Controls whether the memory cache is shared among SMP workers.
2961 The shared memory cache is meant to occupy cache_mem bytes and replace
2962 the non-shared memory cache, although some entities may still be
2963 cached locally by workers for now (e.g., internal and in-transit
2964 objects may be served from a local memory cache even if shared memory
2965 caching is enabled).
2967 By default, the memory cache is shared if and only if all of the
2968 following conditions are satisfied: Squid runs in SMP mode with
2969 multiple workers, cache_mem is positive, and Squid environment
2970 supports required IPC primitives (e.g., POSIX shared memory segments
2971 and GCC-style atomic operations).
2973 To avoid blocking locks, shared memory uses opportunistic algorithms
2974 that do not guarantee that every cachable entity that could have been
2975 shared among SMP workers will actually be shared.
2977 Currently, entities exceeding 32KB in size cannot be shared.
2980 NAME: memory_cache_mode
2985 Controls which objects to keep in the memory cache (cache_mem)
2987 always Keep most recently fetched objects in memory (default)
2989 disk Only disk cache hits are kept in memory, which means
2990 an object must first be cached on disk and then hit
2991 a second time before cached in memory.
2993 network Only objects fetched from network is kept in memory
2996 NAME: memory_replacement_policy
2998 LOC: Config.memPolicy
3001 The memory replacement policy parameter determines which
3002 objects are purged from memory when memory space is needed.
3004 See cache_replacement_policy for details.
3009 -----------------------------------------------------------------------------
3012 NAME: cache_replacement_policy
3014 LOC: Config.replPolicy
3017 The cache replacement policy parameter determines which
3018 objects are evicted (replaced) when disk space is needed.
3020 lru : Squid's original list based LRU policy
3021 heap GDSF : Greedy-Dual Size Frequency
3022 heap LFUDA: Least Frequently Used with Dynamic Aging
3023 heap LRU : LRU policy implemented using a heap
3025 Applies to any cache_dir lines listed below this.
3027 The LRU policies keeps recently referenced objects.
3029 The heap GDSF policy optimizes object hit rate by keeping smaller
3030 popular objects in cache so it has a better chance of getting a
3031 hit. It achieves a lower byte hit rate than LFUDA though since
3032 it evicts larger (possibly popular) objects.
3034 The heap LFUDA policy keeps popular objects in cache regardless of
3035 their size and thus optimizes byte hit rate at the expense of
3036 hit rate since one large, popular object will prevent many
3037 smaller, slightly less popular objects from being cached.
3039 Both policies utilize a dynamic aging mechanism that prevents
3040 cache pollution that can otherwise occur with frequency-based
3041 replacement policies.
3043 NOTE: if using the LFUDA replacement policy you should increase
3044 the value of maximum_object_size above its default of 4096 KB to
3045 to maximize the potential byte hit rate improvement of LFUDA.
3047 For more information about the GDSF and LFUDA cache replacement
3048 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3049 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3055 LOC: Config.cacheSwap
3059 cache_dir Type Directory-Name Fs-specific-data [options]
3061 You can specify multiple cache_dir lines to spread the
3062 cache among different disk partitions.
3064 Type specifies the kind of storage system to use. Only "ufs"
3065 is built by default. To enable any of the other storage systems
3066 see the --enable-storeio configure option.
3068 'Directory' is a top-level directory where cache swap
3069 files will be stored. If you want to use an entire disk
3070 for caching, this can be the mount-point directory.
3071 The directory must exist and be writable by the Squid
3072 process. Squid will NOT create this directory for you.
3074 In SMP configurations, cache_dir must not precede the workers option
3075 and should use configuration macros or conditionals to give each
3076 worker interested in disk caching a dedicated cache directory.
3080 "ufs" is the old well-known Squid storage format that has always
3083 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3085 'Mbytes' is the amount of disk space (MB) to use under this
3086 directory. The default is 100 MB. Change this to suit your
3087 configuration. Do NOT put the size of your disk drive here.
3088 Instead, if you want Squid to use the entire disk drive,
3089 subtract 20% and use that value.
3091 'L1' is the number of first-level subdirectories which
3092 will be created under the 'Directory'. The default is 16.
3094 'L2' is the number of second-level subdirectories which
3095 will be created under each first-level directory. The default
3098 The aufs store type:
3100 "aufs" uses the same storage format as "ufs", utilizing
3101 POSIX-threads to avoid blocking the main Squid process on
3102 disk-I/O. This was formerly known in Squid as async-io.
3104 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3106 see argument descriptions under ufs above
3108 The diskd store type:
3110 "diskd" uses the same storage format as "ufs", utilizing a
3111 separate process to avoid blocking the main Squid process on
3114 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3116 see argument descriptions under ufs above
3118 Q1 specifies the number of unacknowledged I/O requests when Squid
3119 stops opening new files. If this many messages are in the queues,
3120 Squid won't open new files. Default is 64
3122 Q2 specifies the number of unacknowledged messages when Squid
3123 starts blocking. If this many messages are in the queues,
3124 Squid blocks until it receives some replies. Default is 72
3126 When Q1 < Q2 (the default), the cache directory is optimized
3127 for lower response time at the expense of a decrease in hit
3128 ratio. If Q1 > Q2, the cache directory is optimized for
3129 higher hit ratio at the expense of an increase in response
3132 The rock store type:
3134 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3136 The Rock Store type is a database-style storage. All cached
3137 entries are stored in a "database" file, using fixed-size slots,
3138 one entry per slot. The database size is specified in MB. The
3139 slot size is specified in bytes using the max-size option. See
3140 below for more info on the max-size option.
3142 swap-timeout=msec: Squid will not start writing a miss to or
3143 reading a hit from disk if it estimates that the swap operation
3144 will take more than the specified number of milliseconds. By
3145 default and when set to zero, disables the disk I/O time limit
3146 enforcement. Ignored when using blocking I/O module because
3147 blocking synchronous I/O does not allow Squid to estimate the
3148 expected swap wait time.
3150 max-swap-rate=swaps/sec: Artificially limits disk access using
3151 the specified I/O rate limit. Swap out requests that
3152 would cause the average I/O rate to exceed the limit are
3153 delayed. Individual swap in requests (i.e., hits or reads) are
3154 not delayed, but they do contribute to measured swap rate and
3155 since they are placed in the same FIFO queue as swap out
3156 requests, they may wait longer if max-swap-rate is smaller.
3157 This is necessary on file systems that buffer "too
3158 many" writes and then start blocking Squid and other processes
3159 while committing those writes to disk. Usually used together
3160 with swap-timeout to avoid excessive delays and queue overflows
3161 when disk demand exceeds available disk "bandwidth". By default
3162 and when set to zero, disables the disk I/O rate limit
3163 enforcement. Currently supported by IpcIo module only.
3166 The coss store type:
3168 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3169 production use and has thus been removed from this release.
3170 We hope that it can be made usable again soon.
3172 block-size=n defines the "block size" for COSS cache_dir's.
3173 Squid uses file numbers as block numbers. Since file numbers
3174 are limited to 24 bits, the block size determines the maximum
3175 size of the COSS partition. The default is 512 bytes, which
3176 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3177 you should not change the coss block size after Squid
3178 has written some objects to the cache_dir.
3180 The coss file store has changed from 2.5. Now it uses a file
3181 called 'stripe' in the directory names in the config - and
3182 this will be created by squid -z.
3186 no-store, no new objects should be stored to this cache_dir
3188 min-size=n, refers to the min object size in bytes this cache_dir
3189 will accept. It's used to restrict a cache_dir to only store
3190 large objects (e.g. aufs) while other storedirs are optimized
3191 for smaller objects (e.g. COSS). Defaults to 0.
3193 max-size=n, refers to the max object size in bytes this cache_dir
3194 supports. It is used to select the cache_dir to store the object.
3195 Note: To make optimal use of the max-size limits you should order
3196 the cache_dir lines with the smallest max-size value first and the
3197 ones with no max-size specification last.
3199 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3200 which can be changed with the --with-coss-membuf-size=N configure
3204 # Uncomment and adjust the following to add a disk cache directory.
3205 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3209 NAME: store_dir_select_algorithm
3211 LOC: Config.store_dir_select_algorithm
3214 Set this to 'round-robin' as an alternative.
3217 NAME: max_open_disk_fds
3219 LOC: Config.max_open_disk_fds
3222 To avoid having disk as the I/O bottleneck Squid can optionally
3223 bypass the on-disk cache if more than this amount of disk file
3224 descriptors are open.
3226 A value of 0 indicates no limit.
3229 NAME: minimum_object_size
3233 LOC: Config.Store.minObjectSize
3235 Objects smaller than this size will NOT be saved on disk. The
3236 value is specified in kilobytes, and the default is 0 KB, which
3237 means there is no minimum.
3240 NAME: maximum_object_size
3244 LOC: Config.Store.maxObjectSize
3246 Objects larger than this size will NOT be saved on disk. The
3247 value is specified in kilobytes, and the default is 4MB. If
3248 you wish to get a high BYTES hit ratio, you should probably
3249 increase this (one 32 MB object hit counts for 3200 10KB
3250 hits). If you wish to increase speed more than your want to
3251 save bandwidth you should leave this low.
3253 NOTE: if using the LFUDA replacement policy you should increase
3254 this value to maximize the byte hit rate improvement of LFUDA!
3255 See replacement_policy below for a discussion of this policy.
3258 NAME: cache_swap_low
3259 COMMENT: (percent, 0-100)
3262 LOC: Config.Swap.lowWaterMark
3265 NAME: cache_swap_high
3266 COMMENT: (percent, 0-100)
3269 LOC: Config.Swap.highWaterMark
3272 The low- and high-water marks for cache object replacement.
3273 Replacement begins when the swap (disk) usage is above the
3274 low-water mark and attempts to maintain utilization near the
3275 low-water mark. As swap utilization gets close to high-water
3276 mark object eviction becomes more aggressive. If utilization is
3277 close to the low-water mark less replacement is done each time.
3279 Defaults are 90% and 95%. If you have a large cache, 5% could be
3280 hundreds of MB. If this is the case you may wish to set these
3281 numbers closer together.
3286 -----------------------------------------------------------------------------
3296 logformat <name> <format specification>
3298 Defines an access log format.
3300 The <format specification> is a string with embedded % format codes
3302 % format codes all follow the same basic structure where all but
3303 the formatcode is optional. Output strings are automatically escaped
3304 as required according to their context and the output format
3305 modifiers are usually not needed, but can be specified if an explicit
3306 output format is desired.
3308 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3310 " output in quoted string format
3311 [ output in squid text log format as used by log_mime_hdrs
3312 # output in URL quoted format
3317 width minimum and/or maximum field width:
3318 [width_min][.width_max]
3319 When minimum starts with 0, the field is zero-padded.
3320 String values exceeding maximum width are truncated.
3322 {arg} argument such as header name etc
3326 % a literal % character
3327 sn Unique sequence number per log line entry
3328 err_code The ID of an error response served by Squid or
3329 a similar internal error identifier.
3330 err_detail Additional err_code-dependent error information.
3331 note The meta header specified by the argument. Also
3332 logs the adaptation meta headers set by the
3333 adaptation_meta configuration parameter.
3334 If no argument given all meta headers logged.
3336 Connection related format codes:
3338 >a Client source IP address
3340 >p Client source port
3341 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3342 >la Local IP address the client connected to
3343 >lp Local port number the client connected to
3345 la Local listening IP address the client connection was connected to.
3346 lp Local listening port number the client connection was connected to.
3348 <a Server IP address of the last server or peer connection
3349 <A Server FQDN or peer name
3350 <p Server port number of the last server or peer connection
3351 <la Local IP address of the last server or peer connection
3352 <lp Local port number of the last server or peer connection
3354 Time related format codes:
3356 ts Seconds since epoch
3357 tu subsecond time (milliseconds)
3358 tl Local time. Optional strftime format argument
3359 default %d/%b/%Y:%H:%M:%S %z
3360 tg GMT time. Optional strftime format argument
3361 default %d/%b/%Y:%H:%M:%S %z
3362 tr Response time (milliseconds)
3363 dt Total time spent making DNS lookups (milliseconds)
3365 Access Control related format codes:
3367 et Tag returned by external acl
3368 ea Log string returned by external acl
3369 un User name (any available)
3370 ul User name from authentication
3371 ue User name from external acl helper
3372 ui User name from ident
3373 us User name from SSL
3375 HTTP related format codes:
3377 [http::]>h Original request header. Optional header name argument
3378 on the format header[:[separator]element]
3379 [http::]>ha The HTTP request headers after adaptation and redirection.
3380 Optional header name argument as for >h
3381 [http::]<h Reply header. Optional header name argument
3383 [http::]>Hs HTTP status code sent to the client
3384 [http::]<Hs HTTP status code received from the next hop
3385 [http::]<bs Number of HTTP-equivalent message body bytes
3386 received from the next hop, excluding chunked
3387 transfer encoding and control messages.
3388 Generated FTP/Gopher listings are treated as
3390 [http::]mt MIME content type
3391 [http::]rm Request method (GET/POST etc)
3392 [http::]>rm Request method from client
3393 [http::]<rm Request method sent to server or peer
3394 [http::]ru Request URL from client (historic, filtered for logging)
3395 [http::]>ru Request URL from client
3396 [http::]<ru Request URL sent to server or peer
3397 [http::]rp Request URL-Path excluding hostname
3398 [http::]>rp Request URL-Path excluding hostname from client
3399 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3400 [http::]rv Request protocol version
3401 [http::]>rv Request protocol version from client
3402 [http::]<rv Request protocol version sent to server or peer
3403 [http::]<st Sent reply size including HTTP headers
3404 [http::]>st Received request size including HTTP headers. In the
3405 case of chunked requests the chunked encoding metadata
3407 [http::]>sh Received HTTP request headers size
3408 [http::]<sh Sent HTTP reply headers size
3409 [http::]st Request+Reply size including HTTP headers
3410 [http::]<sH Reply high offset sent
3411 [http::]<sS Upstream object size
3412 [http::]<pt Peer response time in milliseconds. The timer starts
3413 when the last request byte is sent to the next hop
3414 and stops when the last response byte is received.
3415 [http::]<tt Total server-side time in milliseconds. The timer
3416 starts with the first connect request (or write I/O)
3417 sent to the first selected peer. The timer stops
3418 with the last I/O with the last peer.
3420 Squid handling related format codes:
3422 Ss Squid request status (TCP_MISS etc)
3423 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3425 SSL-related format codes:
3427 ssl::bump_mode SslBump decision for the transaction:
3429 For CONNECT requests that initiated bumping of
3430 a connection and for any request received on
3431 an already bumped connection, Squid logs the
3432 corresponding SslBump mode ("server-first" or
3433 "client-first"). See the ssl_bump option for
3434 more information about these modes.
3436 A "none" token is logged for requests that
3437 triggered "ssl_bump" ACL evaluation matching
3438 either a "none" rule or no rules at all.
3440 In all other cases, a single dash ("-") is
3443 If ICAP is enabled, the following code becomes available (as
3444 well as ICAP log codes documented with the icap_log option):
3446 icap::tt Total ICAP processing time for the HTTP
3447 transaction. The timer ticks when ICAP
3448 ACLs are checked and when ICAP
3449 transaction is in progress.
3451 If adaptation is enabled the following three codes become available:
3453 adapt::<last_h The header of the last ICAP response or
3454 meta-information from the last eCAP
3455 transaction related to the HTTP transaction.
3456 Like <h, accepts an optional header name
3459 adapt::sum_trs Summed adaptation transaction response
3460 times recorded as a comma-separated list in
3461 the order of transaction start time. Each time
3462 value is recorded as an integer number,
3463 representing response time of one or more
3464 adaptation (ICAP or eCAP) transaction in
3465 milliseconds. When a failed transaction is
3466 being retried or repeated, its time is not
3467 logged individually but added to the
3468 replacement (next) transaction. See also:
3471 adapt::all_trs All adaptation transaction response times.
3472 Same as adaptation_strs but response times of
3473 individual transactions are never added
3474 together. Instead, all transaction response
3475 times are recorded individually.
3477 You can prefix adapt::*_trs format codes with adaptation
3478 service name in curly braces to record response time(s) specific
3479 to that service. For example: %{my_service}adapt::sum_trs
3481 If SSL is enabled, the following formating codes become available:
3483 %ssl::>cert_subject The Subject field of the received client
3484 SSL certificate or a dash ('-') if Squid has
3485 received an invalid/malformed certificate or
3486 no certificate at all. Consider encoding the
3487 logged value because Subject often has spaces.
3489 %ssl::>cert_issuer The Issuer field of the received client
3490 SSL certificate or a dash ('-') if Squid has
3491 received an invalid/malformed certificate or
3492 no certificate at all. Consider encoding the
3493 logged value because Issuer often has spaces.
3495 The default formats available (which do not need re-defining) are:
3497 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3498 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3499 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3500 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3501 logformat useragent %>a [%tl] "%{User-Agent}>h"
3503 NOTE: When the log_mime_hdrs directive is set to ON.
3504 The squid, common and combined formats have a safely encoded copy
3505 of the mime headers appended to each line within a pair of brackets.
3507 NOTE: The common and combined formats are not quite true to the Apache definition.
3508 The logs from Squid contain an extra status and hierarchy code appended.
3512 NAME: access_log cache_access_log
3514 LOC: Config.Log.accesslogs
3515 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3517 These files log client request activities. Has a line every HTTP or
3518 ICP request. The format is:
3519 access_log <module>:<place> [<logformat name> [acl acl ...]]
3520 access_log none [acl acl ...]]
3522 Will log to the specified module:place using the specified format (which
3523 must be defined in a logformat directive) those entries which match
3524 ALL the acl's specified (which must be defined in acl clauses).
3525 If no acl is specified, all requests will be logged to this destination.
3527 ===== Modules Currently available =====
3529 none Do not log any requests matching these ACL.
3530 Do not specify Place or logformat name.
3532 stdio Write each log line to disk immediately at the completion of
3534 Place: the filename and path to be written.
3536 daemon Very similar to stdio. But instead of writing to disk the log
3537 line is passed to a daemon helper for asychronous handling instead.
3538 Place: varies depending on the daemon.
3540 log_file_daemon Place: the file name and path to be written.
3542 syslog To log each request via syslog facility.
3543 Place: The syslog facility and priority level for these entries.
3544 Place Format: facility.priority
3546 where facility could be any of:
3547 authpriv, daemon, local0 ... local7 or user.
3549 And priority could be any of:
3550 err, warning, notice, info, debug.
3552 udp To send each log line as text data to a UDP receiver.
3553 Place: The destination host name or IP and port.
3554 Place Format: //host:port
3556 tcp To send each log line as text data to a TCP receiver.
3557 Place: The destination host name or IP and port.
3558 Place Format: //host:port
3561 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3567 LOC: Config.Log.icaplogs
3570 ICAP log files record ICAP transaction summaries, one line per
3573 The icap_log option format is:
3574 icap_log <filepath> [<logformat name> [acl acl ...]]
3575 icap_log none [acl acl ...]]
3577 Please see access_log option documentation for details. The two
3578 kinds of logs share the overall configuration approach and many
3581 ICAP processing of a single HTTP message or transaction may
3582 require multiple ICAP transactions. In such cases, multiple
3583 ICAP transaction log lines will correspond to a single access
3586 ICAP log uses logformat codes that make sense for an ICAP
3587 transaction. Header-related codes are applied to the HTTP header
3588 embedded in an ICAP server response, with the following caveats:
3589 For REQMOD, there is no HTTP response header unless the ICAP
3590 server performed request satisfaction. For RESPMOD, the HTTP
3591 request header is the header sent to the ICAP server. For
3592 OPTIONS, there are no HTTP headers.
3594 The following format codes are also available for ICAP logs:
3596 icap::<A ICAP server IP address. Similar to <A.
3598 icap::<service_name ICAP service name from the icap_service
3599 option in Squid configuration file.
3601 icap::ru ICAP Request-URI. Similar to ru.
3603 icap::rm ICAP request method (REQMOD, RESPMOD, or
3604 OPTIONS). Similar to existing rm.
3606 icap::>st Bytes sent to the ICAP server (TCP payload
3607 only; i.e., what Squid writes to the socket).
3609 icap::<st Bytes received from the ICAP server (TCP
3610 payload only; i.e., what Squid reads from
3613 icap::<bs Number of message body bytes received from the
3614 ICAP server. ICAP message body, if any, usually
3615 includes encapsulated HTTP message headers and
3616 possibly encapsulated HTTP message body. The
3617 HTTP body part is dechunked before its size is
3620 icap::tr Transaction response time (in
3621 milliseconds). The timer starts when
3622 the ICAP transaction is created and
3623 stops when the transaction is completed.
3626 icap::tio Transaction I/O time (in milliseconds). The
3627 timer starts when the first ICAP request
3628 byte is scheduled for sending. The timers
3629 stops when the last byte of the ICAP response
3632 icap::to Transaction outcome: ICAP_ERR* for all
3633 transaction errors, ICAP_OPT for OPTION
3634 transactions, ICAP_ECHO for 204
3635 responses, ICAP_MOD for message
3636 modification, and ICAP_SAT for request
3637 satisfaction. Similar to Ss.
3639 icap::Hs ICAP response status code. Similar to Hs.
3641 icap::>h ICAP request header(s). Similar to >h.
3643 icap::<h ICAP response header(s). Similar to <h.
3645 The default ICAP log format, which can be used without an explicit
3646 definition, is called icap_squid:
3648 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3650 See also: logformat, log_icap, and %adapt::<last_h
3653 NAME: logfile_daemon
3655 DEFAULT: @DEFAULT_LOGFILED@
3656 LOC: Log::TheConfig.logfile_daemon
3658 Specify the path to the logfile-writing daemon. This daemon is
3659 used to write the access and store logs, if configured.
3661 Squid sends a number of commands to the log daemon:
3662 L<data>\n - logfile data
3667 r<n>\n - set rotate count to <n>
3668 b<n>\n - 1 = buffer output, 0 = don't buffer output
3670 No responses is expected.
3675 LOC: Config.accessList.log
3677 COMMENT: allow|deny acl acl...
3679 This options allows you to control which requests gets logged
3680 to access.log (see access_log directive). Requests denied for
3681 logging will also not be accounted for in performance counters.
3683 This clause only supports fast acl types.
3684 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3690 LOC: Config.accessList.icap
3693 This options allows you to control which requests get logged
3694 to icap.log. See the icap_log directive for ICAP log details.
3697 NAME: cache_store_log
3700 LOC: Config.Log.store
3702 Logs the activities of the storage manager. Shows which
3703 objects are ejected from the cache, and which objects are
3704 saved and for how long.
3705 There are not really utilities to analyze this data, so you can safely
3706 disable it (the default).
3708 Store log uses modular logging outputs. See access_log for the list
3709 of modules supported.
3712 cache_store_log stdio:@DEFAULT_STORE_LOG@
3713 cache_store_log daemon:@DEFAULT_STORE_LOG@
3716 NAME: cache_swap_state cache_swap_log
3718 LOC: Config.Log.swap
3721 Location for the cache "swap.state" file. This index file holds
3722 the metadata of objects saved on disk. It is used to rebuild
3723 the cache during startup. Normally this file resides in each
3724 'cache_dir' directory, but you may specify an alternate
3725 pathname here. Note you must give a full filename, not just
3726 a directory. Since this is the index for the whole object
3727 list you CANNOT periodically rotate it!
3729 If %s can be used in the file name it will be replaced with a
3730 a representation of the cache_dir name where each / is replaced
3731 with '.'. This is needed to allow adding/removing cache_dir
3732 lines when cache_swap_log is being used.
3734 If have more than one 'cache_dir', and %s is not used in the name
3735 these swap logs will have names such as:
3741 The numbered extension (which is added automatically)
3742 corresponds to the order of the 'cache_dir' lines in this
3743 configuration file. If you change the order of the 'cache_dir'
3744 lines in this file, these index files will NOT correspond to
3745 the correct 'cache_dir' entry (unless you manually rename
3746 them). We recommend you do NOT use this option. It is
3747 better to keep these index files in each 'cache_dir' directory.
3750 NAME: logfile_rotate
3753 LOC: Config.Log.rotateNumber
3755 Specifies the number of logfile rotations to make when you
3756 type 'squid -k rotate'. The default is 10, which will rotate
3757 with extensions 0 through 9. Setting logfile_rotate to 0 will
3758 disable the file name rotation, but the logfiles are still closed
3759 and re-opened. This will enable you to rename the logfiles
3760 yourself just before sending the rotate signal.
3762 Note, the 'squid -k rotate' command normally sends a USR1
3763 signal to the running squid process. In certain situations
3764 (e.g. on Linux with Async I/O), USR1 is used for other
3765 purposes, so -k rotate uses another signal. It is best to get
3766 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3769 Note, from Squid-3.1 this option has no effect on the cache.log,
3770 that log can be rotated separately by using debug_options
3773 NAME: emulate_httpd_log
3776 Replace this with an access_log directive using the format 'common' or 'combined'.
3779 NAME: log_ip_on_direct
3782 Remove this option from your config. To log server or peer names use %<A in the log format.
3787 DEFAULT: @DEFAULT_MIME_TABLE@
3788 LOC: Config.mimeTablePathname
3790 Pathname to Squid's MIME table. You shouldn't need to change
3791 this, but the default file contains examples and formatting
3792 information if you do.
3798 LOC: Config.onoff.log_mime_hdrs
3801 The Cache can record both the request and the response MIME
3802 headers for each HTTP transaction. The headers are encoded
3803 safely and will appear as two bracketed fields at the end of
3804 the access log (for either the native or httpd-emulated log
3805 formats). To enable this logging set log_mime_hdrs to 'on'.
3811 Replace this with an access_log directive using the format 'useragent'.
3814 NAME: referer_log referrer_log
3817 Replace this with an access_log directive using the format 'referrer'.
3822 DEFAULT: @DEFAULT_PID_FILE@
3823 LOC: Config.pidFilename
3825 A filename to write the process-id to. To disable, enter "none".
3831 Remove this option from your config. To log FQDN use %>A in the log format.
3834 NAME: client_netmask
3836 LOC: Config.Addrs.client_netmask
3839 A netmask for client addresses in logfiles and cachemgr output.
3840 Change this to protect the privacy of your cache clients.
3841 A netmask of 255.255.255.0 will log all IP's in that range with
3842 the last digit set to '0'.
3848 Use a regular access.log with ACL limiting it to MISS events.
3851 NAME: strip_query_terms
3853 LOC: Config.onoff.strip_query_terms
3856 By default, Squid strips query terms from requested URLs before
3857 logging. This protects your user's privacy.
3864 LOC: Config.onoff.buffered_logs
3866 cache.log log file is written with stdio functions, and as such
3867 it can be buffered or unbuffered. By default it will be unbuffered.
3868 Buffering it can speed up the writing slightly (though you are
3869 unlikely to need to worry unless you run with tons of debugging
3870 enabled in which case performance will suffer badly anyway..).
3873 NAME: netdb_filename
3875 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3876 LOC: Config.netdbFilename
3879 A filename where Squid stores it's netdb state between restarts.
3880 To disable, enter "none".
3884 OPTIONS FOR TROUBLESHOOTING
3885 -----------------------------------------------------------------------------
3890 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3891 LOC: Debug::cache_log
3893 Cache logging file. This is where general information about
3894 your cache's behavior goes. You can increase the amount of data
3895 logged to this file and how often its rotated with "debug_options"
3901 LOC: Debug::debugOptions
3903 Logging options are set as section,level where each source file
3904 is assigned a unique section. Lower levels result in less
3905 output, Full debugging (level 9) can result in a very large
3906 log file, so be careful.
3908 The magic word "ALL" sets debugging levels for all sections.
3909 We recommend normally running with "ALL,1".
3911 The rotate=N option can be used to keep more or less of these logs
3912 than would otherwise be kept by logfile_rotate.
3913 For most uses a single log should be enough to monitor current
3914 events affecting Squid.
3919 LOC: Config.coredump_dir
3920 DEFAULT_IF_NONE: none
3922 By default Squid leaves core files in the directory from where
3923 it was started. If you set 'coredump_dir' to a directory
3924 that exists, Squid will chdir() to that directory at startup
3925 and coredump files will be left there.
3929 # Leave coredumps in the first cache dir
3930 coredump_dir @DEFAULT_SWAP_DIR@
3936 OPTIONS FOR FTP GATEWAYING
3937 -----------------------------------------------------------------------------
3943 LOC: Config.Ftp.anon_user
3945 If you want the anonymous login password to be more informative
3946 (and enable the use of picky ftp servers), set this to something
3947 reasonable for your domain, like wwwuser@somewhere.net
3949 The reason why this is domainless by default is the
3950 request can be made on the behalf of a user in any domain,
3951 depending on how the cache is used.
3952 Some ftp server also validate the email address is valid
3953 (for example perl.com).
3959 LOC: Config.Ftp.passive
3961 If your firewall does not allow Squid to use passive
3962 connections, turn off this option.
3964 Use of ftp_epsv_all option requires this to be ON.
3970 LOC: Config.Ftp.epsv_all
3972 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
3974 NATs may be able to put the connection on a "fast path" through the
3975 translator, as the EPRT command will never be used and therefore,
3976 translation of the data portion of the segments will never be needed.
3978 When a client only expects to do two-way FTP transfers this may be
3980 If squid finds that it must do a three-way FTP transfer after issuing
3981 an EPSV ALL command, the FTP session will fail.
3983 If you have any doubts about this option do not use it.
3984 Squid will nicely attempt all other connection methods.
3986 Requires ftp_passive to be ON (default) for any effect.
3992 LOC: Config.Ftp.epsv
3994 FTP Protocol extensions permit the use of a special "EPSV" command.
3996 NATs may be able to put the connection on a "fast path" through the
3997 translator using EPSV, as the EPRT command will never be used
3998 and therefore, translation of the data portion of the segments
3999 will never be needed.
4001 Turning this OFF will prevent EPSV being attempted.
4002 WARNING: Doing so will convert Squid back to the old behavior with all
4003 the related problems with external NAT devices/layers.
4005 Requires ftp_passive to be ON (default) for any effect.
4011 LOC: Config.Ftp.eprt
4013 FTP Protocol extensions permit the use of a special "EPRT" command.
4015 This extension provides a protocol neutral alternative to the
4016 IPv4-only PORT command. When supported it enables active FTP data
4017 channels over IPv6 and efficient NAT handling.
4019 Turning this OFF will prevent EPRT being attempted and will skip
4020 straight to using PORT for IPv4 servers.
4022 Some devices are known to not handle this extension correctly and
4023 may result in crashes. Devices which suport EPRT enough to fail
4024 cleanly will result in Squid attempting PORT anyway. This directive
4025 should only be disabled when EPRT results in device failures.
4027 WARNING: Doing so will convert Squid back to the old behavior with all
4028 the related problems with external NAT devices/layers and IPv4-only FTP.
4031 NAME: ftp_sanitycheck
4034 LOC: Config.Ftp.sanitycheck
4036 For security and data integrity reasons Squid by default performs
4037 sanity checks of the addresses of FTP data connections ensure the
4038 data connection is to the requested server. If you need to allow
4039 FTP connections to servers using another IP address for the data
4040 connection turn this off.
4043 NAME: ftp_telnet_protocol
4046 LOC: Config.Ftp.telnet
4048 The FTP protocol is officially defined to use the telnet protocol
4049 as transport channel for the control connection. However, many
4050 implementations are broken and does not respect this aspect of
4053 If you have trouble accessing files with ASCII code 255 in the
4054 path or similar problems involving this ASCII code you can
4055 try setting this directive to off. If that helps, report to the
4056 operator of the FTP server in question that their FTP server
4057 is broken and does not follow the FTP standard.
4061 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4062 -----------------------------------------------------------------------------
4067 DEFAULT: @DEFAULT_DISKD@
4068 LOC: Config.Program.diskd
4070 Specify the location of the diskd executable.
4071 Note this is only useful if you have compiled in
4072 diskd as one of the store io modules.
4075 NAME: unlinkd_program
4078 DEFAULT: @DEFAULT_UNLINKD@
4079 LOC: Config.Program.unlinkd
4081 Specify the location of the executable for file deletion process.
4084 NAME: pinger_program
4086 DEFAULT: @DEFAULT_PINGER@
4087 LOC: Config.pinger.program
4090 Specify the location of the executable for the pinger process.
4096 LOC: Config.pinger.enable
4099 Control whether the pinger is active at run-time.
4100 Enables turning ICMP pinger on and off with a simple
4101 squid -k reconfigure.
4106 OPTIONS FOR URL REWRITING
4107 -----------------------------------------------------------------------------
4110 NAME: url_rewrite_program redirect_program
4112 LOC: Config.Program.redirect
4115 Specify the location of the executable URL rewriter to use.
4116 Since they can perform almost any function there isn't one included.
4118 For each requested URL, the rewriter will receive on line with the format
4120 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
4122 In the future, the rewriter interface will be extended with
4123 key=value pairs ("kvpairs" shown above). Rewriter programs
4124 should be prepared to receive and possibly ignore additional
4125 whitespace-separated tokens on each input line.
4127 And the rewriter may return a rewritten URL. The other components of
4128 the request line does not need to be returned (ignored if they are).
4130 The rewriter can also indicate that a client-side redirect should
4131 be performed to the new URL. This is done by prefixing the returned
4132 URL with "301:" (moved permanently) or 302: (moved temporarily), etc.
4134 By default, a URL rewriter is not used.
4137 NAME: url_rewrite_children redirect_children
4138 TYPE: HelperChildConfig
4139 DEFAULT: 20 startup=0 idle=1 concurrency=0
4140 LOC: Config.redirectChildren
4142 The maximum number of redirector processes to spawn. If you limit
4143 it too few Squid will have to wait for them to process a backlog of
4144 URLs, slowing it down. If you allow too many they will use RAM
4145 and other system resources noticably.
4147 The startup= and idle= options allow some measure of skew in your
4152 Sets a minimum of how many processes are to be spawned when Squid
4153 starts or reconfigures. When set to zero the first request will
4154 cause spawning of the first child process to handle it.
4156 Starting too few will cause an initial slowdown in traffic as Squid
4157 attempts to simultaneously spawn enough processes to cope.
4161 Sets a minimum of how many processes Squid is to try and keep available
4162 at all times. When traffic begins to rise above what the existing
4163 processes can handle this many more will be spawned up to the maximum
4164 configured. A minimum setting of 1 is required.
4168 The number of requests each redirector helper can handle in
4169 parallel. Defaults to 0 which indicates the redirector
4170 is a old-style single threaded redirector.
4172 When this directive is set to a value >= 1 then the protocol
4173 used to communicate with the helper is modified to include
4174 a request ID in front of the request/response. The request
4175 ID from the request must be echoed back with the response
4179 NAME: url_rewrite_host_header redirect_rewrites_host_header
4182 LOC: Config.onoff.redir_rewrites_host
4184 To preserve same-origin security policies in browsers and
4185 prevent Host: header forgery by redirectors Squid rewrites
4186 any Host: header in redirected requests.
4188 If you are running an accelerator this may not be a wanted
4189 effect of a redirector. This directive enables you disable
4190 Host: alteration in reverse-proxy traffic.
4192 WARNING: Entries are cached on the result of the URL rewriting
4193 process, so be careful if you have domain-virtual hosts.
4195 WARNING: Squid and other software verifies the URL and Host
4196 are matching, so be careful not to relay through other proxies
4197 or inspecting firewalls with this disabled.
4200 NAME: url_rewrite_access redirector_access
4203 LOC: Config.accessList.redirector
4205 If defined, this access list specifies which requests are
4206 sent to the redirector processes. By default all requests
4209 This clause supports both fast and slow acl types.
4210 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4213 NAME: url_rewrite_bypass redirector_bypass
4215 LOC: Config.onoff.redirector_bypass
4218 When this is 'on', a request will not go through the
4219 redirector if all redirectors are busy. If this is 'off'
4220 and the redirector queue grows too large, Squid will exit
4221 with a FATAL error and ask you to increase the number of
4222 redirectors. You should only enable this if the redirectors
4223 are not critical to your caching system. If you use
4224 redirectors for access control, and you enable this option,
4225 users may have access to pages they should not
4226 be allowed to request.
4230 OPTIONS FOR TUNING THE CACHE
4231 -----------------------------------------------------------------------------
4234 NAME: cache no_cache
4237 LOC: Config.accessList.noCache
4239 A list of ACL elements which, if matched and denied, cause the request to
4240 not be satisfied from the cache and the reply to not be cached.
4241 In other words, use this to force certain objects to never be cached.
4243 You must use the words 'allow' or 'deny' to indicate whether items
4244 matching the ACL should be allowed or denied into the cache.
4246 Default is to allow all to be cached.
4248 This clause supports both fast and slow acl types.
4249 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4255 LOC: Config.maxStale
4258 This option puts an upper limit on how stale content Squid
4259 will serve from the cache if cache validation fails.
4260 Can be overriden by the refresh_pattern max-stale option.
4263 NAME: refresh_pattern
4264 TYPE: refreshpattern
4268 usage: refresh_pattern [-i] regex min percent max [options]
4270 By default, regular expressions are CASE-SENSITIVE. To make
4271 them case-insensitive, use the -i option.
4273 'Min' is the time (in minutes) an object without an explicit
4274 expiry time should be considered fresh. The recommended
4275 value is 0, any higher values may cause dynamic applications
4276 to be erroneously cached unless the application designer
4277 has taken the appropriate actions.
4279 'Percent' is a percentage of the objects age (time since last
4280 modification age) an object without explicit expiry time
4281 will be considered fresh.
4283 'Max' is an upper limit on how long objects without an explicit
4284 expiry time will be considered fresh.
4286 options: override-expire
4291 ignore-must-revalidate
4298 override-expire enforces min age even if the server
4299 sent an explicit expiry time (e.g., with the
4300 Expires: header or Cache-Control: max-age). Doing this
4301 VIOLATES the HTTP standard. Enabling this feature
4302 could make you liable for problems which it causes.
4304 Note: override-expire does not enforce staleness - it only extends
4305 freshness / min. If the server returns a Expires time which
4306 is longer than your max time, Squid will still consider
4307 the object fresh for that period of time.
4309 override-lastmod enforces min age even on objects
4310 that were modified recently.
4312 reload-into-ims changes client no-cache or ``reload''
4313 to If-Modified-Since requests. Doing this VIOLATES the
4314 HTTP standard. Enabling this feature could make you
4315 liable for problems which it causes.
4317 ignore-reload ignores a client no-cache or ``reload''
4318 header. Doing this VIOLATES the HTTP standard. Enabling
4319 this feature could make you liable for problems which
4322 ignore-no-store ignores any ``Cache-control: no-store''
4323 headers received from a server. Doing this VIOLATES
4324 the HTTP standard. Enabling this feature could make you
4325 liable for problems which it causes.
4327 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4328 headers received from a server. Doing this VIOLATES
4329 the HTTP standard. Enabling this feature could make you
4330 liable for problems which it causes.
4332 ignore-private ignores any ``Cache-control: private''
4333 headers received from a server. Doing this VIOLATES
4334 the HTTP standard. Enabling this feature could make you
4335 liable for problems which it causes.
4337 ignore-auth caches responses to requests with authorization,
4338 as if the originserver had sent ``Cache-control: public''
4339 in the response header. Doing this VIOLATES the HTTP standard.
4340 Enabling this feature could make you liable for problems which
4343 refresh-ims causes squid to contact the origin server
4344 when a client issues an If-Modified-Since request. This
4345 ensures that the client will receive an updated version
4346 if one is available.
4348 store-stale stores responses even if they don't have explicit
4349 freshness or a validator (i.e., Last-Modified or an ETag)
4350 present, or if they're already stale. By default, Squid will
4351 not cache such responses because they usually can't be
4352 reused. Note that such responses will be stale by default.
4354 max-stale=NN provide a maximum staleness factor. Squid won't
4355 serve objects more stale than this even if it failed to
4356 validate the object. Default: use the max_stale global limit.
4358 Basically a cached object is:
4360 FRESH if expires < now, else STALE
4362 FRESH if lm-factor < percent, else STALE
4366 The refresh_pattern lines are checked in the order listed here.
4367 The first entry which matches is used. If none of the entries
4368 match the default will be used.
4370 Note, you must uncomment all the default lines if you want
4371 to change one. The default setting is only active if none is
4376 # Add any of your own refresh_pattern entries above these.
4377 refresh_pattern ^ftp: 1440 20% 10080
4378 refresh_pattern ^gopher: 1440 0% 1440
4379 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4380 refresh_pattern . 0 20% 4320
4384 NAME: quick_abort_min
4388 LOC: Config.quickAbort.min
4391 NAME: quick_abort_max
4395 LOC: Config.quickAbort.max
4398 NAME: quick_abort_pct
4402 LOC: Config.quickAbort.pct
4404 The cache by default continues downloading aborted requests
4405 which are almost completed (less than 16 KB remaining). This
4406 may be undesirable on slow (e.g. SLIP) links and/or very busy
4407 caches. Impatient users may tie up file descriptors and
4408 bandwidth by repeatedly requesting and immediately aborting
4411 When the user aborts a request, Squid will check the
4412 quick_abort values to the amount of data transfered until
4415 If the transfer has less than 'quick_abort_min' KB remaining,
4416 it will finish the retrieval.
4418 If the transfer has more than 'quick_abort_max' KB remaining,
4419 it will abort the retrieval.
4421 If more than 'quick_abort_pct' of the transfer has completed,
4422 it will finish the retrieval.
4424 If you do not want any retrieval to continue after the client
4425 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4428 If you want retrievals to always continue if they are being
4429 cached set 'quick_abort_min' to '-1 KB'.
4432 NAME: read_ahead_gap
4433 COMMENT: buffer-size
4435 LOC: Config.readAheadGap
4438 The amount of data the cache will buffer ahead of what has been
4439 sent to the client when retrieving an object from another server.
4443 IFDEF: USE_HTTP_VIOLATIONS
4446 LOC: Config.negativeTtl
4449 Set the Default Time-to-Live (TTL) for failed requests.
4450 Certain types of failures (such as "connection refused" and
4451 "404 Not Found") are able to be negatively-cached for a short time.
4452 Modern web servers should provide Expires: header, however if they
4453 do not this can provide a minimum TTL.
4454 The default is not to cache errors with unknown expiry details.
4456 Note that this is different from negative caching of DNS lookups.
4458 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4459 this feature could make you liable for problems which it
4463 NAME: positive_dns_ttl
4466 LOC: Config.positiveDnsTtl
4469 Upper limit on how long Squid will cache positive DNS responses.
4470 Default is 6 hours (360 minutes). This directive must be set
4471 larger than negative_dns_ttl.
4474 NAME: negative_dns_ttl
4477 LOC: Config.negativeDnsTtl
4480 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4481 This also sets the lower cache limit on positive lookups.
4482 Minimum value is 1 second, and it is not recommendable to go
4483 much below 10 seconds.
4486 NAME: range_offset_limit
4487 COMMENT: size [acl acl...]
4489 LOC: Config.rangeOffsetLimit
4492 usage: (size) [units] [[!]aclname]
4494 Sets an upper limit on how far (number of bytes) into the file
4495 a Range request may be to cause Squid to prefetch the whole file.
4496 If beyond this limit, Squid forwards the Range request as it is and
4497 the result is NOT cached.
4499 This is to stop a far ahead range request (lets say start at 17MB)
4500 from making Squid fetch the whole object up to that point before
4501 sending anything to the client.
4503 Multiple range_offset_limit lines may be specified, and they will
4504 be searched from top to bottom on each request until a match is found.
4505 The first match found will be used. If no line matches a request, the
4506 default limit of 0 bytes will be used.
4508 'size' is the limit specified as a number of units.
4510 'units' specifies whether to use bytes, KB, MB, etc.
4511 If no units are specified bytes are assumed.
4513 A size of 0 causes Squid to never fetch more than the
4514 client requested. (default)
4516 A size of 'none' causes Squid to always fetch the object from the
4517 beginning so it may cache the result. (2.0 style)
4519 'aclname' is the name of a defined ACL.
4521 NP: Using 'none' as the byte value here will override any quick_abort settings
4522 that may otherwise apply to the range request. The range request will
4523 be fully fetched from start to finish regardless of the client
4524 actions. This affects bandwidth usage.
4527 NAME: minimum_expiry_time
4530 LOC: Config.minimum_expiry_time
4533 The minimum caching time according to (Expires - Date)
4534 Headers Squid honors if the object can't be revalidated
4535 defaults to 60 seconds. In reverse proxy environments it
4536 might be desirable to honor shorter object lifetimes. It
4537 is most likely better to make your server return a
4538 meaningful Last-Modified header however. In ESI environments
4539 where page fragments often have short lifetimes, this will
4540 often be best set to 0.
4543 NAME: store_avg_object_size
4547 LOC: Config.Store.avgObjectSize
4549 Average object size, used to estimate number of objects your
4550 cache can hold. The default is 13 KB.
4553 NAME: store_objects_per_bucket
4556 LOC: Config.Store.objectsPerBucket
4558 Target number of objects per bucket in the store hash table.
4559 Lowering this value increases the total number of buckets and
4560 also the storage maintenance rate. The default is 20.
4565 -----------------------------------------------------------------------------
4568 NAME: request_header_max_size
4572 LOC: Config.maxRequestHeaderSize
4574 This specifies the maximum size for HTTP headers in a request.
4575 Request headers are usually relatively small (about 512 bytes).
4576 Placing a limit on the request header size will catch certain
4577 bugs (for example with persistent connections) and possibly
4578 buffer-overflow or denial-of-service attacks.
4581 NAME: reply_header_max_size
4585 LOC: Config.maxReplyHeaderSize
4587 This specifies the maximum size for HTTP headers in a reply.
4588 Reply headers are usually relatively small (about 512 bytes).
4589 Placing a limit on the reply header size will catch certain
4590 bugs (for example with persistent connections) and possibly
4591 buffer-overflow or denial-of-service attacks.
4594 NAME: request_body_max_size
4598 LOC: Config.maxRequestBodySize
4600 This specifies the maximum size for an HTTP request body.
4601 In other words, the maximum size of a PUT/POST request.
4602 A user who attempts to send a request with a body larger
4603 than this limit receives an "Invalid Request" error message.
4604 If you set this parameter to a zero (the default), there will
4605 be no limit imposed.
4608 NAME: client_request_buffer_max_size
4612 LOC: Config.maxRequestBufferSize
4614 This specifies the maximum buffer size of a client request.
4615 It prevents squid eating too much memory when somebody uploads
4619 NAME: chunked_request_body_max_size
4623 LOC: Config.maxChunkedRequestBodySize
4625 A broken or confused HTTP/1.1 client may send a chunked HTTP
4626 request to Squid. Squid does not have full support for that
4627 feature yet. To cope with such requests, Squid buffers the
4628 entire request and then dechunks request body to create a
4629 plain HTTP/1.0 request with a known content length. The plain
4630 request is then used by the rest of Squid code as usual.
4632 The option value specifies the maximum size of the buffer used
4633 to hold the request before the conversion. If the chunked
4634 request size exceeds the specified limit, the conversion
4635 fails, and the client receives an "unsupported request" error,
4636 as if dechunking was disabled.
4638 Dechunking is enabled by default. To disable conversion of
4639 chunked requests, set the maximum to zero.
4641 Request dechunking feature and this option in particular are a
4642 temporary hack. When chunking requests and responses are fully
4643 supported, there will be no need to buffer a chunked request.
4647 IFDEF: USE_HTTP_VIOLATIONS
4650 LOC: Config.accessList.brokenPosts
4652 A list of ACL elements which, if matched, causes Squid to send
4653 an extra CRLF pair after the body of a PUT/POST request.
4655 Some HTTP servers has broken implementations of PUT/POST,
4656 and rely on an extra CRLF pair sent by some WWW clients.
4658 Quote from RFC2616 section 4.1 on this matter:
4660 Note: certain buggy HTTP/1.0 client implementations generate an
4661 extra CRLF's after a POST request. To restate what is explicitly
4662 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4663 a request with an extra CRLF.
4665 This clause only supports fast acl types.
4666 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4669 acl buggy_server url_regex ^http://....
4670 broken_posts allow buggy_server
4673 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4676 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4678 LOC: Adaptation::Config::use_indirect_client
4680 Controls whether the indirect client IP address (instead of the direct
4681 client IP address) is passed to adaptation services.
4683 See also: follow_x_forwarded_for adaptation_send_client_ip
4687 IFDEF: USE_HTTP_VIOLATIONS
4691 LOC: Config.onoff.via
4693 If set (default), Squid will include a Via header in requests and
4694 replies as required by RFC2616.
4700 LOC: Config.onoff.ie_refresh
4703 Microsoft Internet Explorer up until version 5.5 Service
4704 Pack 1 has an issue with transparent proxies, wherein it
4705 is impossible to force a refresh. Turning this on provides
4706 a partial fix to the problem, by causing all IMS-REFRESH
4707 requests from older IE versions to check the origin server
4708 for fresh content. This reduces hit ratio by some amount
4709 (~10% in my experience), but allows users to actually get
4710 fresh content when they want it. Note because Squid
4711 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4712 of 5.5 is unchanged from old versions of Squid (i.e. a
4713 forced refresh is impossible). Newer versions of IE will,
4714 hopefully, continue to have the new behavior and will be
4715 handled based on that assumption. This option defaults to
4716 the old Squid behavior, which is better for hit ratios but
4717 worse for clients using IE, if they need to be able to
4718 force fresh content.
4721 NAME: vary_ignore_expire
4724 LOC: Config.onoff.vary_ignore_expire
4727 Many HTTP servers supporting Vary gives such objects
4728 immediate expiry time with no cache-control header
4729 when requested by a HTTP/1.0 client. This option
4730 enables Squid to ignore such expiry times until
4731 HTTP/1.1 is fully implemented.
4733 WARNING: If turned on this may eventually cause some
4734 varying objects not intended for caching to get cached.
4737 NAME: request_entities
4739 LOC: Config.onoff.request_entities
4742 Squid defaults to deny GET and HEAD requests with request entities,
4743 as the meaning of such requests are undefined in the HTTP standard
4744 even if not explicitly forbidden.
4746 Set this directive to on if you have clients which insists
4747 on sending request entities in GET or HEAD requests. But be warned
4748 that there is server software (both proxies and web servers) which
4749 can fail to properly process this kind of request which may make you
4750 vulnerable to cache pollution attacks if enabled.
4753 NAME: request_header_access
4754 IFDEF: USE_HTTP_VIOLATIONS
4755 TYPE: http_header_access
4756 LOC: Config.request_header_access
4759 Usage: request_header_access header_name allow|deny [!]aclname ...
4761 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4762 this feature could make you liable for problems which it
4765 This option replaces the old 'anonymize_headers' and the
4766 older 'http_anonymizer' option with something that is much
4767 more configurable. A list of ACLs for each header name allows
4768 removal of specific header fields under specific conditions.
4770 This option only applies to outgoing HTTP request headers (i.e.,
4771 headers sent by Squid to the next HTTP hop such as a cache peer
4772 or an origin server). The option has no effect during cache hit
4773 detection. The equivalent adaptation vectoring point in ICAP
4774 terminology is post-cache REQMOD.
4776 The option is applied to individual outgoing request header
4777 fields. For each request header field F, Squid uses the first
4778 qualifying sets of request_header_access rules:
4780 1. Rules with header_name equal to F's name.
4781 2. Rules with header_name 'Other', provided F's name is not
4782 on the hard-coded list of commonly used HTTP header names.
4783 3. Rules with header_name 'All'.
4785 Within that qualifying rule set, rule ACLs are checked as usual.
4786 If ACLs of an "allow" rule match, the header field is allowed to
4787 go through as is. If ACLs of a "deny" rule match, the header is
4788 removed and request_header_replace is then checked to identify
4789 if the removed header has a replacement. If no rules within the
4790 set have matching ACLs, the header field is left as is.
4792 For example, to achieve the same behavior as the old
4793 'http_anonymizer standard' option, you should use:
4795 request_header_access From deny all
4796 request_header_access Referer deny all
4797 request_header_access Server deny all
4798 request_header_access User-Agent deny all
4799 request_header_access WWW-Authenticate deny all
4800 request_header_access Link deny all
4802 Or, to reproduce the old 'http_anonymizer paranoid' feature
4805 request_header_access Allow allow all
4806 request_header_access Authorization allow all
4807 request_header_access WWW-Authenticate allow all
4808 request_header_access Proxy-Authorization allow all
4809 request_header_access Proxy-Authenticate allow all
4810 request_header_access Cache-Control allow all
4811 request_header_access Content-Encoding allow all
4812 request_header_access Content-Length allow all
4813 request_header_access Content-Type allow all
4814 request_header_access Date allow all
4815 request_header_access Expires allow all
4816 request_header_access Host allow all
4817 request_header_access If-Modified-Since allow all
4818 request_header_access Last-Modified allow all
4819 request_header_access Location allow all
4820 request_header_access Pragma allow all
4821 request_header_access Accept allow all
4822 request_header_access Accept-Charset allow all
4823 request_header_access Accept-Encoding allow all
4824 request_header_access Accept-Language allow all
4825 request_header_access Content-Language allow all
4826 request_header_access Mime-Version allow all
4827 request_header_access Retry-After allow all
4828 request_header_access Title allow all
4829 request_header_access Connection allow all
4830 request_header_access All deny all
4832 although many of those are HTTP reply headers, and so should be
4833 controlled with the reply_header_access directive.
4835 By default, all headers are allowed (no anonymizing is
4839 NAME: reply_header_access
4840 IFDEF: USE_HTTP_VIOLATIONS
4841 TYPE: http_header_access
4842 LOC: Config.reply_header_access
4845 Usage: reply_header_access header_name allow|deny [!]aclname ...
4847 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4848 this feature could make you liable for problems which it
4851 This option only applies to reply headers, i.e., from the
4852 server to the client.
4854 This is the same as request_header_access, but in the other
4855 direction. Please see request_header_access for detailed
4858 For example, to achieve the same behavior as the old
4859 'http_anonymizer standard' option, you should use:
4861 reply_header_access From deny all
4862 reply_header_access Referer deny all
4863 reply_header_access Server deny all
4864 reply_header_access User-Agent deny all
4865 reply_header_access WWW-Authenticate deny all
4866 reply_header_access Link deny all
4868 Or, to reproduce the old 'http_anonymizer paranoid' feature
4871 reply_header_access Allow allow all
4872 reply_header_access Authorization allow all
4873 reply_header_access WWW-Authenticate allow all
4874 reply_header_access Proxy-Authorization allow all
4875 reply_header_access Proxy-Authenticate allow all
4876 reply_header_access Cache-Control allow all
4877 reply_header_access Content-Encoding allow all
4878 reply_header_access Content-Length allow all
4879 reply_header_access Content-Type allow all
4880 reply_header_access Date allow all
4881 reply_header_access Expires allow all
4882 reply_header_access Host allow all
4883 reply_header_access If-Modified-Since allow all
4884 reply_header_access Last-Modified allow all
4885 reply_header_access Location allow all
4886 reply_header_access Pragma allow all
4887 reply_header_access Accept allow all
4888 reply_header_access Accept-Charset allow all
4889 reply_header_access Accept-Encoding allow all
4890 reply_header_access Accept-Language allow all
4891 reply_header_access Content-Language allow all
4892 reply_header_access Mime-Version allow all
4893 reply_header_access Retry-After allow all
4894 reply_header_access Title allow all
4895 reply_header_access Connection allow all
4896 reply_header_access All deny all
4898 although the HTTP request headers won't be usefully controlled
4899 by this directive -- see request_header_access for details.
4901 By default, all headers are allowed (no anonymizing is
4905 NAME: request_header_replace header_replace
4906 IFDEF: USE_HTTP_VIOLATIONS
4907 TYPE: http_header_replace
4908 LOC: Config.request_header_access
4911 Usage: request_header_replace header_name message
4912 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
4914 This option allows you to change the contents of headers
4915 denied with request_header_access above, by replacing them
4916 with some fixed string. This replaces the old fake_user_agent
4919 This only applies to request headers, not reply headers.
4921 By default, headers are removed if denied.
4924 NAME: reply_header_replace
4925 IFDEF: USE_HTTP_VIOLATIONS
4926 TYPE: http_header_replace
4927 LOC: Config.reply_header_access
4930 Usage: reply_header_replace header_name message
4931 Example: reply_header_replace Server Foo/1.0
4933 This option allows you to change the contents of headers
4934 denied with reply_header_access above, by replacing them
4935 with some fixed string.
4937 This only applies to reply headers, not request headers.
4939 By default, headers are removed if denied.
4942 NAME: request_header_add
4943 TYPE: HeaderWithAclList
4944 LOC: Config.request_header_add
4947 Usage: request_header_add field-name field-value acl1 [acl2] ...
4948 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
4950 This option adds header fields to outgoing HTTP requests (i.e.,
4951 request headers sent by Squid to the next HTTP hop such as a
4952 cache peer or an origin server). The option has no effect during
4953 cache hit detection. The equivalent adaptation vectoring point
4954 in ICAP terminology is post-cache REQMOD.
4956 Field-name is a token specifying an HTTP header name. If a
4957 standard HTTP header name is used, Squid does not check whether
4958 the new header conflicts with any existing headers or violates
4959 HTTP rules. If the request to be modified already contains a
4960 field with the same name, the old field is preserved but the
4961 header field values are not merged.
4963 Field-value is either a token or a quoted string. If quoted
4964 string format is used, then the surrounding quotes are removed
4965 while escape sequences and %macros are processed.
4967 In theory, all of the logformat codes can be used as %macros.
4968 However, unlike logging (which happens at the very end of
4969 transaction lifetime), the transaction may not yet have enough
4970 information to expand a macro when the new header value is needed.
4971 And some information may already be available to Squid but not yet
4972 committed where the macro expansion code can access it (report
4973 such instances!). The macro will be expanded into a single dash
4974 ('-') in such cases. Not all macros have been tested.
4976 One or more Squid ACLs may be specified to restrict header
4977 injection to matching requests. As always in squid.conf, all
4978 ACLs in an option ACL list must be satisfied for the insertion
4979 to happen. The request_header_add option supports fast ACLs
4988 This option used to log custom information about the master
4989 transaction. For example, an admin may configure Squid to log
4990 which "user group" the transaction belongs to, where "user group"
4991 will be determined based on a set of ACLs and not [just]
4992 authentication information.
4993 Values of key/value pairs can be logged using %{key}note macros:
4995 note key value acl ...
4996 logformat myFormat ... %{key}note ...
4999 NAME: relaxed_header_parser
5000 COMMENT: on|off|warn
5002 LOC: Config.onoff.relaxed_header_parser
5005 In the default "on" setting Squid accepts certain forms
5006 of non-compliant HTTP messages where it is unambiguous
5007 what the sending application intended even if the message
5008 is not correctly formatted. The messages is then normalized
5009 to the correct form when forwarded by Squid.
5011 If set to "warn" then a warning will be emitted in cache.log
5012 each time such HTTP error is encountered.
5014 If set to "off" then such HTTP errors will cause the request
5015 or response to be rejected.
5020 -----------------------------------------------------------------------------
5023 NAME: forward_timeout
5026 LOC: Config.Timeout.forward
5029 This parameter specifies how long Squid should at most attempt in
5030 finding a forwarding path for the request before giving up.
5033 NAME: connect_timeout
5036 LOC: Config.Timeout.connect
5039 This parameter specifies how long to wait for the TCP connect to
5040 the requested server or peer to complete before Squid should
5041 attempt to find another path where to forward the request.
5044 NAME: peer_connect_timeout
5047 LOC: Config.Timeout.peer_connect
5050 This parameter specifies how long to wait for a pending TCP
5051 connection to a peer cache. The default is 30 seconds. You
5052 may also set different timeout values for individual neighbors
5053 with the 'connect-timeout' option on a 'cache_peer' line.
5059 LOC: Config.Timeout.read
5062 The read_timeout is applied on server-side connections. After
5063 each successful read(), the timeout will be extended by this
5064 amount. If no data is read again after this amount of time,
5065 the request is aborted and logged with ERR_READ_TIMEOUT. The
5066 default is 15 minutes.
5072 LOC: Config.Timeout.write
5075 This timeout is tracked for all connections that have data
5076 available for writing and are waiting for the socket to become
5077 ready. After each successful write, the timeout is extended by
5078 the configured amount. If Squid has data to write but the
5079 connection is not ready for the configured duration, the
5080 transaction associated with the connection is terminated. The
5081 default is 15 minutes.
5084 NAME: request_timeout
5086 LOC: Config.Timeout.request
5089 How long to wait for complete HTTP request headers after initial
5090 connection establishment.
5093 NAME: client_idle_pconn_timeout persistent_request_timeout
5095 LOC: Config.Timeout.clientIdlePconn
5098 How long to wait for the next HTTP request on a persistent
5099 client connection after the previous request completes.
5102 NAME: client_lifetime
5105 LOC: Config.Timeout.lifetime
5108 The maximum amount of time a client (browser) is allowed to
5109 remain connected to the cache process. This protects the Cache
5110 from having a lot of sockets (and hence file descriptors) tied up
5111 in a CLOSE_WAIT state from remote clients that go away without
5112 properly shutting down (either because of a network failure or
5113 because of a poor client implementation). The default is one
5116 NOTE: The default value is intended to be much larger than any
5117 client would ever need to be connected to your cache. You
5118 should probably change client_lifetime only as a last resort.
5119 If you seem to have many client connections tying up
5120 filedescriptors, we recommend first tuning the read_timeout,
5121 request_timeout, persistent_request_timeout and quick_abort values.
5124 NAME: half_closed_clients
5126 LOC: Config.onoff.half_closed_clients
5129 Some clients may shutdown the sending side of their TCP
5130 connections, while leaving their receiving sides open. Sometimes,
5131 Squid can not tell the difference between a half-closed and a
5132 fully-closed TCP connection.
5134 By default, Squid will immediately close client connections when
5135 read(2) returns "no more data to read."
5137 Change this option to 'on' and Squid will keep open connections
5138 until a read(2) or write(2) on the socket returns an error.
5139 This may show some benefits for reverse proxies. But if not
5140 it is recommended to leave OFF.
5143 NAME: server_idle_pconn_timeout pconn_timeout
5145 LOC: Config.Timeout.serverIdlePconn
5148 Timeout for idle persistent connections to servers and other
5155 LOC: Ident::TheConfig.timeout
5158 Maximum time to wait for IDENT lookups to complete.
5160 If this is too high, and you enabled IDENT lookups from untrusted
5161 users, you might be susceptible to denial-of-service by having
5162 many ident requests going at once.
5165 NAME: shutdown_lifetime
5168 LOC: Config.shutdownLifetime
5171 When SIGTERM or SIGHUP is received, the cache is put into
5172 "shutdown pending" mode until all active sockets are closed.
5173 This value is the lifetime to set for all open descriptors
5174 during shutdown mode. Any active clients after this many
5175 seconds will receive a 'timeout' message.
5179 ADMINISTRATIVE PARAMETERS
5180 -----------------------------------------------------------------------------
5186 LOC: Config.adminEmail
5188 Email-address of local cache manager who will receive
5189 mail if the cache dies. The default is "webmaster."
5195 LOC: Config.EmailFrom
5197 From: email-address for mail sent when the cache dies.
5198 The default is to use 'appname@unique_hostname'.
5199 Default appname value is "squid", can be changed into
5200 src/globals.h before building squid.
5206 LOC: Config.EmailProgram
5208 Email program used to send mail if the cache dies.
5209 The default is "mail". The specified program must comply
5210 with the standard Unix mail syntax:
5211 mail-program recipient < mailfile
5213 Optional command line options can be specified.
5216 NAME: cache_effective_user
5218 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5219 LOC: Config.effectiveUser
5221 If you start Squid as root, it will change its effective/real
5222 UID/GID to the user specified below. The default is to change
5223 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5224 see also; cache_effective_group
5227 NAME: cache_effective_group
5230 LOC: Config.effectiveGroup
5232 Squid sets the GID to the effective user's default group ID
5233 (taken from the password file) and supplementary group list
5234 from the groups membership.
5236 If you want Squid to run with a specific GID regardless of
5237 the group memberships of the effective user then set this
5238 to the group (or GID) you want Squid to run as. When set
5239 all other group privileges of the effective user are ignored
5240 and only this GID is effective. If Squid is not started as
5241 root the user starting Squid MUST be member of the specified
5244 This option is not recommended by the Squid Team.
5245 Our preference is for administrators to configure a secure
5246 user account for squid with UID/GID matching system policies.
5249 NAME: httpd_suppress_version_string
5253 LOC: Config.onoff.httpd_suppress_version_string
5255 Suppress Squid version string info in HTTP headers and HTML error pages.
5258 NAME: visible_hostname
5260 LOC: Config.visibleHostname
5263 If you want to present a special hostname in error messages, etc,
5264 define this. Otherwise, the return value of gethostname()
5265 will be used. If you have multiple caches in a cluster and
5266 get errors about IP-forwarding you must set them to have individual
5267 names with this setting.
5270 NAME: unique_hostname
5272 LOC: Config.uniqueHostname
5275 If you want to have multiple machines with the same
5276 'visible_hostname' you must give each machine a different
5277 'unique_hostname' so forwarding loops can be detected.
5280 NAME: hostname_aliases
5282 LOC: Config.hostnameAliases
5285 A list of other DNS names your cache has.
5293 Minimum umask which should be enforced while the proxy
5294 is running, in addition to the umask set at startup.
5296 For a traditional octal representation of umasks, start
5301 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5302 -----------------------------------------------------------------------------
5304 This section contains parameters for the (optional) cache
5305 announcement service. This service is provided to help
5306 cache administrators locate one another in order to join or
5307 create cache hierarchies.
5309 An 'announcement' message is sent (via UDP) to the registration
5310 service by Squid. By default, the announcement message is NOT
5311 SENT unless you enable it with 'announce_period' below.
5313 The announcement message includes your hostname, plus the
5314 following information from this configuration file:
5320 All current information is processed regularly and made
5321 available on the Web at http://www.ircache.net/Cache/Tracker/.
5324 NAME: announce_period
5326 LOC: Config.Announce.period
5329 This is how frequently to send cache announcements. The
5330 default is `0' which disables sending the announcement
5333 To enable announcing your cache, just set an announce period.
5336 announce_period 1 day
5341 DEFAULT: tracker.ircache.net
5342 LOC: Config.Announce.host
5348 LOC: Config.Announce.file
5354 LOC: Config.Announce.port
5356 announce_host and announce_port set the hostname and port
5357 number where the registration message will be sent.
5359 Hostname will default to 'tracker.ircache.net' and port will
5360 default default to 3131. If the 'filename' argument is given,
5361 the contents of that file will be included in the announce
5366 HTTPD-ACCELERATOR OPTIONS
5367 -----------------------------------------------------------------------------
5370 NAME: httpd_accel_surrogate_id
5373 LOC: Config.Accel.surrogate_id
5375 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5376 need an identification token to allow control targeting. Because
5377 a farm of surrogates may all perform the same tasks, they may share
5378 an identification token.
5380 The default ID is the visible_hostname
5383 NAME: http_accel_surrogate_remote
5387 LOC: Config.onoff.surrogate_is_remote
5389 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
5390 Set this to on to have squid behave as a remote surrogate.
5394 IFDEF: USE_SQUID_ESI
5395 COMMENT: libxml2|expat|custom
5397 LOC: ESIParser::Type
5400 ESI markup is not strictly XML compatible. The custom ESI parser
5401 will give higher performance, but cannot handle non ASCII character
5406 DELAY POOL PARAMETERS
5407 -----------------------------------------------------------------------------
5411 TYPE: delay_pool_count
5413 IFDEF: USE_DELAY_POOLS
5416 This represents the number of delay pools to be used. For example,
5417 if you have one class 2 delay pool and one class 3 delays pool, you
5418 have a total of 2 delay pools.
5422 TYPE: delay_pool_class
5424 IFDEF: USE_DELAY_POOLS
5427 This defines the class of each delay pool. There must be exactly one
5428 delay_class line for each delay pool. For example, to define two
5429 delay pools, one of class 2 and one of class 3, the settings above
5433 delay_pools 4 # 4 delay pools
5434 delay_class 1 2 # pool 1 is a class 2 pool
5435 delay_class 2 3 # pool 2 is a class 3 pool
5436 delay_class 3 4 # pool 3 is a class 4 pool
5437 delay_class 4 5 # pool 4 is a class 5 pool
5439 The delay pool classes are:
5441 class 1 Everything is limited by a single aggregate
5444 class 2 Everything is limited by a single aggregate
5445 bucket as well as an "individual" bucket chosen
5446 from bits 25 through 32 of the IPv4 address.
5448 class 3 Everything is limited by a single aggregate
5449 bucket as well as a "network" bucket chosen
5450 from bits 17 through 24 of the IP address and a
5451 "individual" bucket chosen from bits 17 through
5452 32 of the IPv4 address.
5454 class 4 Everything in a class 3 delay pool, with an
5455 additional limit on a per user basis. This
5456 only takes effect if the username is established
5457 in advance - by forcing authentication in your
5460 class 5 Requests are grouped according their tag (see
5461 external_acl's tag= reply).
5464 Each pool also requires a delay_parameters directive to configure the pool size
5465 and speed limits used whenever the pool is applied to a request. Along with
5466 a set of delay_access directives to determine when it is used.
5468 NOTE: If an IP address is a.b.c.d
5469 -> bits 25 through 32 are "d"
5470 -> bits 17 through 24 are "c"
5471 -> bits 17 through 32 are "c * 256 + d"
5473 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5474 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5478 TYPE: delay_pool_access
5480 IFDEF: USE_DELAY_POOLS
5483 This is used to determine which delay pool a request falls into.
5485 delay_access is sorted per pool and the matching starts with pool 1,
5486 then pool 2, ..., and finally pool N. The first delay pool where the
5487 request is allowed is selected for the request. If it does not allow
5488 the request to any pool then the request is not delayed (default).
5490 For example, if you want some_big_clients in delay
5491 pool 1 and lotsa_little_clients in delay pool 2:
5494 delay_access 1 allow some_big_clients
5495 delay_access 1 deny all
5496 delay_access 2 allow lotsa_little_clients
5497 delay_access 2 deny all
5498 delay_access 3 allow authenticated_clients
5501 NAME: delay_parameters
5502 TYPE: delay_pool_rates
5504 IFDEF: USE_DELAY_POOLS
5507 This defines the parameters for a delay pool. Each delay pool has
5508 a number of "buckets" associated with it, as explained in the
5509 description of delay_class.
5511 For a class 1 delay pool, the syntax is:
5513 delay_parameters pool aggregate
5515 For a class 2 delay pool:
5517 delay_parameters pool aggregate individual
5519 For a class 3 delay pool:
5521 delay_parameters pool aggregate network individual
5523 For a class 4 delay pool:
5525 delay_parameters pool aggregate network individual user
5527 For a class 5 delay pool:
5529 delay_parameters pool tagrate
5531 The option variables are:
5533 pool a pool number - ie, a number between 1 and the
5534 number specified in delay_pools as used in
5537 aggregate the speed limit parameters for the aggregate bucket
5540 individual the speed limit parameters for the individual
5541 buckets (class 2, 3).
5543 network the speed limit parameters for the network buckets
5546 user the speed limit parameters for the user buckets
5549 tagrate the speed limit parameters for the tag buckets
5552 A pair of delay parameters is written restore/maximum, where restore is
5553 the number of bytes (not bits - modem and network speeds are usually
5554 quoted in bits) per second placed into the bucket, and maximum is the
5555 maximum number of bytes which can be in the bucket at any time.
5557 There must be one delay_parameters line for each delay pool.
5560 For example, if delay pool number 1 is a class 2 delay pool as in the
5561 above example, and is being used to strictly limit each host to 64Kbit/sec
5562 (plus overheads), with no overall limit, the line is:
5564 delay_parameters 1 -1/-1 8000/8000
5566 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5568 Note that the figure -1 is used to represent "unlimited".
5571 And, if delay pool number 2 is a class 3 delay pool as in the above
5572 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5573 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5574 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5575 to permit a decent web page to be downloaded at a decent speed
5576 (if the network is not being limited due to overuse) but slow down
5577 large downloads more significantly:
5579 delay_parameters 2 32000/32000 8000/8000 600/8000
5581 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5582 8 x 8000 KByte/sec -> 64Kbit/sec.
5583 8 x 600 Byte/sec -> 4800bit/sec.
5586 Finally, for a class 4 delay pool as in the example - each user will
5587 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5589 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5592 NAME: delay_initial_bucket_level
5593 COMMENT: (percent, 0-100)
5596 IFDEF: USE_DELAY_POOLS
5597 LOC: Config.Delay.initial
5599 The initial bucket percentage is used to determine how much is put
5600 in each bucket when squid starts, is reconfigured, or first notices
5601 a host accessing it (in class 2 and class 3, individual hosts and
5602 networks only have buckets associated with them once they have been
5607 CLIENT DELAY POOL PARAMETERS
5608 -----------------------------------------------------------------------------
5611 NAME: client_delay_pools
5612 TYPE: client_delay_pool_count
5614 IFDEF: USE_DELAY_POOLS
5615 LOC: Config.ClientDelay
5617 This option specifies the number of client delay pools used. It must
5618 preceed other client_delay_* options.
5621 client_delay_pools 2
5624 NAME: client_delay_initial_bucket_level
5625 COMMENT: (percent, 0-no_limit)
5628 IFDEF: USE_DELAY_POOLS
5629 LOC: Config.ClientDelay.initial
5631 This option determines the initial bucket size as a percentage of
5632 max_bucket_size from client_delay_parameters. Buckets are created
5633 at the time of the "first" connection from the matching IP. Idle
5634 buckets are periodically deleted up.
5636 You can specify more than 100 percent but note that such "oversized"
5637 buckets are not refilled until their size goes down to max_bucket_size
5638 from client_delay_parameters.
5641 client_delay_initial_bucket_level 50
5644 NAME: client_delay_parameters
5645 TYPE: client_delay_pool_rates
5647 IFDEF: USE_DELAY_POOLS
5648 LOC: Config.ClientDelay
5651 This option configures client-side bandwidth limits using the
5654 client_delay_parameters pool speed_limit max_bucket_size
5656 pool is an integer ID used for client_delay_access matching.
5658 speed_limit is bytes added to the bucket per second.
5660 max_bucket_size is the maximum size of a bucket, enforced after any
5661 speed_limit additions.
5663 Please see the delay_parameters option for more information and
5667 client_delay_parameters 1 1024 2048
5668 client_delay_parameters 2 51200 16384
5671 NAME: client_delay_access
5672 TYPE: client_delay_pool_access
5674 IFDEF: USE_DELAY_POOLS
5675 LOC: Config.ClientDelay
5678 This option determines the client-side delay pool for the
5681 client_delay_access pool_ID allow|deny acl_name
5683 All client_delay_access options are checked in their pool ID
5684 order, starting with pool 1. The first checked pool with allowed
5685 request is selected for the request. If no ACL matches or there
5686 are no client_delay_access options, the request bandwidth is not
5689 The ACL-selected pool is then used to find the
5690 client_delay_parameters for the request. Client-side pools are
5691 not used to aggregate clients. Clients are always aggregated
5692 based on their source IP addresses (one bucket per source IP).
5694 Please see delay_access for more examples.
5697 client_delay_access 1 allow low_rate_network
5698 client_delay_access 2 allow vips_network
5702 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5703 -----------------------------------------------------------------------------
5708 LOC: Config.Wccp.router
5712 Use this option to define your WCCP ``home'' router for
5715 wccp_router supports a single WCCP(v1) router
5717 wccp2_router supports multiple WCCPv2 routers
5719 only one of the two may be used at the same time and defines
5720 which version of WCCP to use.
5724 TYPE: IpAddress_list
5725 LOC: Config.Wccp2.router
5729 Use this option to define your WCCP ``home'' router for
5732 wccp_router supports a single WCCP(v1) router
5734 wccp2_router supports multiple WCCPv2 routers
5736 only one of the two may be used at the same time and defines
5737 which version of WCCP to use.
5742 LOC: Config.Wccp.version
5746 This directive is only relevant if you need to set up WCCP(v1)
5747 to some very old and end-of-life Cisco routers. In all other
5748 setups it must be left unset or at the default setting.
5749 It defines an internal version in the WCCP(v1) protocol,
5750 with version 4 being the officially documented protocol.
5752 According to some users, Cisco IOS 11.2 and earlier only
5753 support WCCP version 3. If you're using that or an earlier
5754 version of IOS, you may need to change this value to 3, otherwise
5755 do not specify this parameter.
5758 NAME: wccp2_rebuild_wait
5760 LOC: Config.Wccp2.rebuildwait
5764 If this is enabled Squid will wait for the cache dir rebuild to finish
5765 before sending the first wccp2 HereIAm packet
5768 NAME: wccp2_forwarding_method
5770 LOC: Config.Wccp2.forwarding_method
5774 WCCP2 allows the setting of forwarding methods between the
5775 router/switch and the cache. Valid values are as follows:
5777 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5778 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5780 Currently (as of IOS 12.4) cisco routers only support GRE.
5781 Cisco switches only support the L2 redirect assignment method.
5784 NAME: wccp2_return_method
5786 LOC: Config.Wccp2.return_method
5790 WCCP2 allows the setting of return methods between the
5791 router/switch and the cache for packets that the cache
5792 decides not to handle. Valid values are as follows:
5794 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5795 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5797 Currently (as of IOS 12.4) cisco routers only support GRE.
5798 Cisco switches only support the L2 redirect assignment.
5800 If the "ip wccp redirect exclude in" command has been
5801 enabled on the cache interface, then it is still safe for
5802 the proxy server to use a l2 redirect method even if this
5803 option is set to GRE.
5806 NAME: wccp2_assignment_method
5808 LOC: Config.Wccp2.assignment_method
5812 WCCP2 allows the setting of methods to assign the WCCP hash
5813 Valid values are as follows:
5815 hash - Hash assignment
5816 mask - Mask assignment
5818 As a general rule, cisco routers support the hash assignment method
5819 and cisco switches support the mask assignment method.
5824 LOC: Config.Wccp2.info
5825 DEFAULT_IF_NONE: standard 0
5828 WCCP2 allows for multiple traffic services. There are two
5829 types: "standard" and "dynamic". The standard type defines
5830 one service id - http (id 0). The dynamic service ids can be from
5831 51 to 255 inclusive. In order to use a dynamic service id
5832 one must define the type of traffic to be redirected; this is done
5833 using the wccp2_service_info option.
5835 The "standard" type does not require a wccp2_service_info option,
5836 just specifying the service id will suffice.
5838 MD5 service authentication can be enabled by adding
5839 "password=<password>" to the end of this service declaration.
5843 wccp2_service standard 0 # for the 'web-cache' standard service
5844 wccp2_service dynamic 80 # a dynamic service type which will be
5845 # fleshed out with subsequent options.
5846 wccp2_service standard 0 password=foo
5849 NAME: wccp2_service_info
5850 TYPE: wccp2_service_info
5851 LOC: Config.Wccp2.info
5855 Dynamic WCCPv2 services require further information to define the
5856 traffic you wish to have diverted.
5860 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
5861 priority=<priority> ports=<port>,<port>..
5863 The relevant WCCPv2 flags:
5864 + src_ip_hash, dst_ip_hash
5865 + source_port_hash, dst_port_hash
5866 + src_ip_alt_hash, dst_ip_alt_hash
5867 + src_port_alt_hash, dst_port_alt_hash
5870 The port list can be one to eight entries.
5874 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
5875 priority=240 ports=80
5877 Note: the service id must have been defined by a previous
5878 'wccp2_service dynamic <id>' entry.
5883 LOC: Config.Wccp2.weight
5887 Each cache server gets assigned a set of the destination
5888 hash proportional to their weight.
5893 LOC: Config.Wccp.address
5900 LOC: Config.Wccp2.address
5904 Use this option if you require WCCP to use a specific
5907 The default behavior is to not bind to any specific address.
5911 PERSISTENT CONNECTION HANDLING
5912 -----------------------------------------------------------------------------
5914 Also see "pconn_timeout" in the TIMEOUTS section
5917 NAME: client_persistent_connections
5919 LOC: Config.onoff.client_pconns
5923 NAME: server_persistent_connections
5925 LOC: Config.onoff.server_pconns
5928 Persistent connection support for clients and servers. By
5929 default, Squid uses persistent connections (when allowed)
5930 with its clients and servers. You can use these options to
5931 disable persistent connections with clients and/or servers.
5934 NAME: persistent_connection_after_error
5936 LOC: Config.onoff.error_pconns
5939 With this directive the use of persistent connections after
5940 HTTP errors can be disabled. Useful if you have clients
5941 who fail to handle errors on persistent connections proper.
5944 NAME: detect_broken_pconn
5946 LOC: Config.onoff.detect_broken_server_pconns
5949 Some servers have been found to incorrectly signal the use
5950 of HTTP/1.0 persistent connections even on replies not
5951 compatible, causing significant delays. This server problem
5952 has mostly been seen on redirects.
5954 By enabling this directive Squid attempts to detect such
5955 broken replies and automatically assume the reply is finished
5956 after 10 seconds timeout.
5960 CACHE DIGEST OPTIONS
5961 -----------------------------------------------------------------------------
5964 NAME: digest_generation
5965 IFDEF: USE_CACHE_DIGESTS
5967 LOC: Config.onoff.digest_generation
5970 This controls whether the server will generate a Cache Digest
5971 of its contents. By default, Cache Digest generation is
5972 enabled if Squid is compiled with --enable-cache-digests defined.
5975 NAME: digest_bits_per_entry
5976 IFDEF: USE_CACHE_DIGESTS
5978 LOC: Config.digest.bits_per_entry
5981 This is the number of bits of the server's Cache Digest which
5982 will be associated with the Digest entry for a given HTTP
5983 Method and URL (public key) combination. The default is 5.
5986 NAME: digest_rebuild_period
5987 IFDEF: USE_CACHE_DIGESTS
5990 LOC: Config.digest.rebuild_period
5993 This is the wait time between Cache Digest rebuilds.
5996 NAME: digest_rewrite_period
5998 IFDEF: USE_CACHE_DIGESTS
6000 LOC: Config.digest.rewrite_period
6003 This is the wait time between Cache Digest writes to
6007 NAME: digest_swapout_chunk_size
6010 IFDEF: USE_CACHE_DIGESTS
6011 LOC: Config.digest.swapout_chunk_size
6014 This is the number of bytes of the Cache Digest to write to
6015 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6019 NAME: digest_rebuild_chunk_percentage
6020 COMMENT: (percent, 0-100)
6021 IFDEF: USE_CACHE_DIGESTS
6023 LOC: Config.digest.rebuild_chunk_percentage
6026 This is the percentage of the Cache Digest to be scanned at a
6027 time. By default it is set to 10% of the Cache Digest.
6032 -----------------------------------------------------------------------------
6037 LOC: Config.Port.snmp
6041 The port number where Squid listens for SNMP requests. To enable
6042 SNMP support set this to a suitable port number. Port number
6043 3401 is often used for the Squid SNMP agent. By default it's
6044 set to "0" (disabled)
6052 LOC: Config.accessList.snmp
6053 DEFAULT_IF_NONE: deny all
6056 Allowing or denying access to the SNMP port.
6058 All access to the agent is denied by default.
6061 snmp_access allow|deny [!]aclname ...
6063 This clause only supports fast acl types.
6064 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6066 snmp_access allow snmppublic localhost
6067 snmp_access deny all
6070 NAME: snmp_incoming_address
6072 LOC: Config.Addrs.snmp_incoming
6077 NAME: snmp_outgoing_address
6079 LOC: Config.Addrs.snmp_outgoing
6083 Just like 'udp_incoming_address', but for the SNMP port.
6085 snmp_incoming_address is used for the SNMP socket receiving
6086 messages from SNMP agents.
6087 snmp_outgoing_address is used for SNMP packets returned to SNMP
6090 The default snmp_incoming_address is to listen on all
6091 available network interfaces.
6093 If snmp_outgoing_address is not set it will use the same socket
6094 as snmp_incoming_address. Only change this if you want to have
6095 SNMP replies sent using another address than where this Squid
6096 listens for SNMP queries.
6098 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6099 the same value since they both use port 3401.
6104 -----------------------------------------------------------------------------
6107 NAME: icp_port udp_port
6110 LOC: Config.Port.icp
6112 The port number where Squid sends and receives ICP queries to
6113 and from neighbor caches. The standard UDP port for ICP is 3130.
6114 Default is disabled (0).
6117 icp_port @DEFAULT_ICP_PORT@
6124 LOC: Config.Port.htcp
6126 The port number where Squid sends and receives HTCP queries to
6127 and from neighbor caches. To turn it on you want to set it to
6128 4827. By default it is set to "0" (disabled).
6134 NAME: log_icp_queries
6138 LOC: Config.onoff.log_udp
6140 If set, ICP queries are logged to access.log. You may wish
6141 do disable this if your ICP load is VERY high to speed things
6142 up or to simplify log analysis.
6145 NAME: udp_incoming_address
6147 LOC:Config.Addrs.udp_incoming
6150 udp_incoming_address is used for UDP packets received from other
6153 The default behavior is to not bind to any specific address.
6155 Only change this if you want to have all UDP queries received on
6156 a specific interface/address.
6158 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6159 modules. Altering it will affect all of them in the same manner.
6161 see also; udp_outgoing_address
6163 NOTE, udp_incoming_address and udp_outgoing_address can not
6164 have the same value since they both use the same port.
6167 NAME: udp_outgoing_address
6169 LOC: Config.Addrs.udp_outgoing
6172 udp_outgoing_address is used for UDP packets sent out to other
6175 The default behavior is to not bind to any specific address.
6177 Instead it will use the same socket as udp_incoming_address.
6178 Only change this if you want to have UDP queries sent using another
6179 address than where this Squid listens for UDP queries from other
6182 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6183 modules. Altering it will affect all of them in the same manner.
6185 see also; udp_incoming_address
6187 NOTE, udp_incoming_address and udp_outgoing_address can not
6188 have the same value since they both use the same port.
6195 LOC: Config.onoff.icp_hit_stale
6197 If you want to return ICP_HIT for stale cache objects, set this
6198 option to 'on'. If you have sibling relationships with caches
6199 in other administrative domains, this should be 'off'. If you only
6200 have sibling relationships with caches under your control,
6201 it is probably okay to set this to 'on'.
6202 If set to 'on', your siblings should use the option "allow-miss"
6203 on their cache_peer lines for connecting to you.
6206 NAME: minimum_direct_hops
6209 LOC: Config.minDirectHops
6211 If using the ICMP pinging stuff, do direct fetches for sites
6212 which are no more than this many hops away.
6215 NAME: minimum_direct_rtt
6218 LOC: Config.minDirectRtt
6220 If using the ICMP pinging stuff, do direct fetches for sites
6221 which are no more than this many rtt milliseconds away.
6227 LOC: Config.Netdb.low
6233 LOC: Config.Netdb.high
6235 The low and high water marks for the ICMP measurement
6236 database. These are counts, not percents. The defaults are
6237 900 and 1000. When the high water mark is reached, database
6238 entries will be deleted until the low mark is reached.
6241 NAME: netdb_ping_period
6243 LOC: Config.Netdb.period
6246 The minimum period for measuring a site. There will be at
6247 least this much delay between successive pings to the same
6248 network. The default is five minutes.
6255 LOC: Config.onoff.query_icmp
6257 If you want to ask your peers to include ICMP data in their ICP
6258 replies, enable this option.
6260 If your peer has configured Squid (during compilation) with
6261 '--enable-icmp' that peer will send ICMP pings to origin server
6262 sites of the URLs it receives. If you enable this option the
6263 ICP replies from that peer will include the ICMP data (if available).
6264 Then, when choosing a parent cache, Squid will choose the parent with
6265 the minimal RTT to the origin server. When this happens, the
6266 hierarchy field of the access.log will be
6267 "CLOSEST_PARENT_MISS". This option is off by default.
6270 NAME: test_reachability
6274 LOC: Config.onoff.test_reachability
6276 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6277 instead of ICP_MISS if the target host is NOT in the ICMP
6278 database, or has a zero RTT.
6281 NAME: icp_query_timeout
6285 LOC: Config.Timeout.icp_query
6287 Normally Squid will automatically determine an optimal ICP
6288 query timeout value based on the round-trip-time of recent ICP
6289 queries. If you want to override the value determined by
6290 Squid, set this 'icp_query_timeout' to a non-zero value. This
6291 value is specified in MILLISECONDS, so, to use a 2-second
6292 timeout (the old default), you would write:
6294 icp_query_timeout 2000
6297 NAME: maximum_icp_query_timeout
6301 LOC: Config.Timeout.icp_query_max
6303 Normally the ICP query timeout is determined dynamically. But
6304 sometimes it can lead to very large values (say 5 seconds).
6305 Use this option to put an upper limit on the dynamic timeout
6306 value. Do NOT use this option to always use a fixed (instead
6307 of a dynamic) timeout value. To set a fixed timeout see the
6308 'icp_query_timeout' directive.
6311 NAME: minimum_icp_query_timeout
6315 LOC: Config.Timeout.icp_query_min
6317 Normally the ICP query timeout is determined dynamically. But
6318 sometimes it can lead to very small timeouts, even lower than
6319 the normal latency variance on your link due to traffic.
6320 Use this option to put an lower limit on the dynamic timeout
6321 value. Do NOT use this option to always use a fixed (instead
6322 of a dynamic) timeout value. To set a fixed timeout see the
6323 'icp_query_timeout' directive.
6326 NAME: background_ping_rate
6330 LOC: Config.backgroundPingRate
6332 Controls how often the ICP pings are sent to siblings that
6333 have background-ping set.
6337 MULTICAST ICP OPTIONS
6338 -----------------------------------------------------------------------------
6343 LOC: Config.mcast_group_list
6346 This tag specifies a list of multicast groups which your server
6347 should join to receive multicasted ICP queries.
6349 NOTE! Be very careful what you put here! Be sure you
6350 understand the difference between an ICP _query_ and an ICP
6351 _reply_. This option is to be set only if you want to RECEIVE
6352 multicast queries. Do NOT set this option to SEND multicast
6353 ICP (use cache_peer for that). ICP replies are always sent via
6354 unicast, so this option does not affect whether or not you will
6355 receive replies from multicast group members.
6357 You must be very careful to NOT use a multicast address which
6358 is already in use by another group of caches.
6360 If you are unsure about multicast, please read the Multicast
6361 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6363 Usage: mcast_groups 239.128.16.128 224.0.1.20
6365 By default, Squid doesn't listen on any multicast groups.
6368 NAME: mcast_miss_addr
6369 IFDEF: MULTICAST_MISS_STREAM
6371 LOC: Config.mcast_miss.addr
6374 If you enable this option, every "cache miss" URL will
6375 be sent out on the specified multicast address.
6377 Do not enable this option unless you are are absolutely
6378 certain you understand what you are doing.
6381 NAME: mcast_miss_ttl
6382 IFDEF: MULTICAST_MISS_STREAM
6384 LOC: Config.mcast_miss.ttl
6387 This is the time-to-live value for packets multicasted
6388 when multicasting off cache miss URLs is enabled. By
6389 default this is set to 'site scope', i.e. 16.
6392 NAME: mcast_miss_port
6393 IFDEF: MULTICAST_MISS_STREAM
6395 LOC: Config.mcast_miss.port
6398 This is the port number to be used in conjunction with
6402 NAME: mcast_miss_encode_key
6403 IFDEF: MULTICAST_MISS_STREAM
6405 LOC: Config.mcast_miss.encode_key
6406 DEFAULT: XXXXXXXXXXXXXXXX
6408 The URLs that are sent in the multicast miss stream are
6409 encrypted. This is the encryption key.
6412 NAME: mcast_icp_query_timeout
6416 LOC: Config.Timeout.mcast_icp_query
6418 For multicast peers, Squid regularly sends out ICP "probes" to
6419 count how many other peers are listening on the given multicast
6420 address. This value specifies how long Squid should wait to
6421 count all the replies. The default is 2000 msec, or 2
6426 INTERNAL ICON OPTIONS
6427 -----------------------------------------------------------------------------
6430 NAME: icon_directory
6432 LOC: Config.icons.directory
6433 DEFAULT: @DEFAULT_ICON_DIR@
6435 Where the icons are stored. These are normally kept in
6439 NAME: global_internal_static
6441 LOC: Config.onoff.global_internal_static
6444 This directive controls is Squid should intercept all requests for
6445 /squid-internal-static/ no matter which host the URL is requesting
6446 (default on setting), or if nothing special should be done for
6447 such URLs (off setting). The purpose of this directive is to make
6448 icons etc work better in complex cache hierarchies where it may
6449 not always be possible for all corners in the cache mesh to reach
6450 the server generating a directory listing.
6453 NAME: short_icon_urls
6455 LOC: Config.icons.use_short_names
6458 If this is enabled Squid will use short URLs for icons.
6459 If disabled it will revert to the old behavior of including
6460 it's own name and port in the URL.
6462 If you run a complex cache hierarchy with a mix of Squid and
6463 other proxies you may need to disable this directive.
6468 -----------------------------------------------------------------------------
6471 NAME: error_directory
6473 LOC: Config.errorDirectory
6476 If you wish to create your own versions of the default
6477 error files to customize them to suit your company copy
6478 the error/template files to another directory and point
6481 WARNING: This option will disable multi-language support
6482 on error pages if used.
6484 The squid developers are interested in making squid available in
6485 a wide variety of languages. If you are making translations for a
6486 language that Squid does not currently provide please consider
6487 contributing your translation back to the project.
6488 http://wiki.squid-cache.org/Translations
6490 The squid developers working on translations are happy to supply drop-in
6491 translated error files in exchange for any new language contributions.
6494 NAME: error_default_language
6495 IFDEF: USE_ERR_LOCALES
6497 LOC: Config.errorDefaultLanguage
6500 Set the default language which squid will send error pages in
6501 if no existing translation matches the clients language
6504 If unset (default) generic English will be used.
6506 The squid developers are interested in making squid available in
6507 a wide variety of languages. If you are interested in making
6508 translations for any language see the squid wiki for details.
6509 http://wiki.squid-cache.org/Translations
6512 NAME: error_log_languages
6513 IFDEF: USE_ERR_LOCALES
6515 LOC: Config.errorLogMissingLanguages
6518 Log to cache.log what languages users are attempting to
6519 auto-negotiate for translations.
6521 Successful negotiations are not logged. Only failures
6522 have meaning to indicate that Squid may need an upgrade
6523 of its error page translations.
6526 NAME: err_page_stylesheet
6528 LOC: Config.errorStylesheet
6529 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6531 CSS Stylesheet to pattern the display of Squid default error pages.
6533 For information on CSS see http://www.w3.org/Style/CSS/
6538 LOC: Config.errHtmlText
6541 HTML text to include in error messages. Make this a "mailto"
6542 URL to your admin address, or maybe just a link to your
6543 organizations Web page.
6545 To include this in your error messages, you must rewrite
6546 the error template files (found in the "errors" directory).
6547 Wherever you want the 'err_html_text' line to appear,
6548 insert a %L tag in the error template file.
6551 NAME: email_err_data
6554 LOC: Config.onoff.emailErrData
6557 If enabled, information about the occurred error will be
6558 included in the mailto links of the ERR pages (if %W is set)
6559 so that the email body contains the data.
6560 Syntax is <A HREF="mailto:%w%W">%w</A>
6565 LOC: Config.denyInfoList
6568 Usage: deny_info err_page_name acl
6569 or deny_info http://... acl
6570 or deny_info TCP_RESET acl
6572 This can be used to return a ERR_ page for requests which
6573 do not pass the 'http_access' rules. Squid remembers the last
6574 acl it evaluated in http_access, and if a 'deny_info' line exists
6575 for that ACL Squid returns a corresponding error page.
6577 The acl is typically the last acl on the http_access deny line which
6578 denied access. The exceptions to this rule are:
6579 - When Squid needs to request authentication credentials. It's then
6580 the first authentication related acl encountered
6581 - When none of the http_access lines matches. It's then the last
6582 acl processed on the last http_access line.
6583 - When the decision to deny access was made by an adaptation service,
6584 the acl name is the corresponding eCAP or ICAP service_name.
6586 NP: If providing your own custom error pages with error_directory
6587 you may also specify them by your custom file name:
6588 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6590 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6591 may be specified by prefixing the file name with the code and a colon.
6592 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6594 Alternatively you can tell Squid to reset the TCP connection
6595 by specifying TCP_RESET.
6597 Or you can specify an error URL or URL pattern. The browsers will
6598 get redirected to the specified URL after formatting tags have
6599 been replaced. Redirect will be done with 302 or 307 according to
6600 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6601 the URL. e.g. 303:http://example.com/
6604 %a - username (if available. Password NOT included)
6607 %E - Error description
6609 %H - Request domain name
6610 %i - Client IP Address
6612 %o - Message result from external ACL helper
6613 %p - Request Port number
6614 %P - Request Protocol name
6615 %R - Request URL path
6616 %T - Timestamp in RFC 1123 format
6617 %U - Full canonical URL from client
6618 (HTTPS URLs terminate with *)
6619 %u - Full canonical URL from client
6620 %w - Admin email from squid.conf
6622 %% - Literal percent (%) code
6627 OPTIONS INFLUENCING REQUEST FORWARDING
6628 -----------------------------------------------------------------------------
6631 NAME: nonhierarchical_direct
6633 LOC: Config.onoff.nonhierarchical_direct
6636 By default, Squid will send any non-hierarchical requests
6637 (matching hierarchy_stoplist or not cacheable request type) direct
6640 If you set this to off, Squid will prefer to send these
6641 requests to parents.
6643 Note that in most configurations, by turning this off you will only
6644 add latency to these request without any improvement in global hit
6647 If you are inside an firewall see never_direct instead of
6653 LOC: Config.onoff.prefer_direct
6656 Normally Squid tries to use parents for most requests. If you for some
6657 reason like it to first try going direct and only use a parent if
6658 going direct fails set this to on.
6660 By combining nonhierarchical_direct off and prefer_direct on you
6661 can set up Squid to use a parent as a backup path if going direct
6664 Note: If you want Squid to use parents for all requests see
6665 the never_direct directive. prefer_direct only modifies how Squid
6666 acts on cacheable requests.
6671 LOC: Config.accessList.AlwaysDirect
6674 Usage: always_direct allow|deny [!]aclname ...
6676 Here you can use ACL elements to specify requests which should
6677 ALWAYS be forwarded by Squid to the origin servers without using
6678 any peers. For example, to always directly forward requests for
6679 local servers ignoring any parents or siblings you may have use
6682 acl local-servers dstdomain my.domain.net
6683 always_direct allow local-servers
6685 To always forward FTP requests directly, use
6688 always_direct allow FTP
6690 NOTE: There is a similar, but opposite option named
6691 'never_direct'. You need to be aware that "always_direct deny
6692 foo" is NOT the same thing as "never_direct allow foo". You
6693 may need to use a deny rule to exclude a more-specific case of
6694 some other rule. Example:
6696 acl local-external dstdomain external.foo.net
6697 acl local-servers dstdomain .foo.net
6698 always_direct deny local-external
6699 always_direct allow local-servers
6701 NOTE: If your goal is to make the client forward the request
6702 directly to the origin server bypassing Squid then this needs
6703 to be done in the client configuration. Squid configuration
6704 can only tell Squid how Squid should fetch the object.
6706 NOTE: This directive is not related to caching. The replies
6707 is cached as usual even if you use always_direct. To not cache
6708 the replies see the 'cache' directive.
6710 This clause supports both fast and slow acl types.
6711 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6716 LOC: Config.accessList.NeverDirect
6719 Usage: never_direct allow|deny [!]aclname ...
6721 never_direct is the opposite of always_direct. Please read
6722 the description for always_direct if you have not already.
6724 With 'never_direct' you can use ACL elements to specify
6725 requests which should NEVER be forwarded directly to origin
6726 servers. For example, to force the use of a proxy for all
6727 requests, except those in your local domain use something like:
6729 acl local-servers dstdomain .foo.net
6730 never_direct deny local-servers
6731 never_direct allow all
6733 or if Squid is inside a firewall and there are local intranet
6734 servers inside the firewall use something like:
6736 acl local-intranet dstdomain .foo.net
6737 acl local-external dstdomain external.foo.net
6738 always_direct deny local-external
6739 always_direct allow local-intranet
6740 never_direct allow all
6742 This clause supports both fast and slow acl types.
6743 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6747 ADVANCED NETWORKING OPTIONS
6748 -----------------------------------------------------------------------------
6751 NAME: incoming_udp_average incoming_icp_average
6754 LOC: Config.comm_incoming.udp.average
6756 Heavy voodoo here. I can't even believe you are reading this.
6757 Are you crazy? Don't even think about adjusting these unless
6758 you understand the algorithms in comm_select.c first!
6761 NAME: incoming_tcp_average incoming_http_average
6764 LOC: Config.comm_incoming.tcp.average
6766 Heavy voodoo here. I can't even believe you are reading this.
6767 Are you crazy? Don't even think about adjusting these unless
6768 you understand the algorithms in comm_select.c first!
6771 NAME: incoming_dns_average
6774 LOC: Config.comm_incoming.dns.average
6776 Heavy voodoo here. I can't even believe you are reading this.
6777 Are you crazy? Don't even think about adjusting these unless
6778 you understand the algorithms in comm_select.c first!
6781 NAME: min_udp_poll_cnt min_icp_poll_cnt
6784 LOC: Config.comm_incoming.udp.min_poll
6786 Heavy voodoo here. I can't even believe you are reading this.
6787 Are you crazy? Don't even think about adjusting these unless
6788 you understand the algorithms in comm_select.c first!
6791 NAME: min_dns_poll_cnt
6794 LOC: Config.comm_incoming.dns.min_poll
6796 Heavy voodoo here. I can't even believe you are reading this.
6797 Are you crazy? Don't even think about adjusting these unless
6798 you understand the algorithms in comm_select.c first!
6801 NAME: min_tcp_poll_cnt min_http_poll_cnt
6804 LOC: Config.comm_incoming.tcp.min_poll
6806 Heavy voodoo here. I can't even believe you are reading this.
6807 Are you crazy? Don't even think about adjusting these unless
6808 you understand the algorithms in comm_select.c first!
6814 LOC: Config.accept_filter
6818 The name of an accept(2) filter to install on Squid's
6819 listen socket(s). This feature is perhaps specific to
6820 FreeBSD and requires support in the kernel.
6822 The 'httpready' filter delays delivering new connections
6823 to Squid until a full HTTP request has been received.
6824 See the accf_http(9) man page for details.
6826 The 'dataready' filter delays delivering new connections
6827 to Squid until there is some data to process.
6828 See the accf_dataready(9) man page for details.
6832 The 'data' filter delays delivering of new connections
6833 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
6834 You may optionally specify a number of seconds to wait by
6835 'data=N' where N is the number of seconds. Defaults to 30
6836 if not specified. See the tcp(7) man page for details.
6839 accept_filter httpready
6844 NAME: client_ip_max_connections
6846 LOC: Config.client_ip_max_connections
6849 Set an absolute limit on the number of connections a single
6850 client IP can use. Any more than this and Squid will begin to drop
6851 new connections from the client until it closes some links.
6853 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
6854 connections from the client. For finer control use the ACL access controls.
6856 Requires client_db to be enabled (the default).
6858 WARNING: This may noticably slow down traffic received via external proxies
6859 or NAT devices and cause them to rebound error messages back to their clients.
6862 NAME: tcp_recv_bufsize
6866 LOC: Config.tcpRcvBufsz
6868 Size of receive buffer to set for TCP sockets. Probably just
6869 as easy to change your kernel's default. Set to zero to use
6870 the default buffer size.
6875 -----------------------------------------------------------------------------
6882 LOC: Adaptation::Icap::TheConfig.onoff
6885 If you want to enable the ICAP module support, set this to on.
6888 NAME: icap_connect_timeout
6891 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
6894 This parameter specifies how long to wait for the TCP connect to
6895 the requested ICAP server to complete before giving up and either
6896 terminating the HTTP transaction or bypassing the failure.
6898 The default for optional services is peer_connect_timeout.
6899 The default for essential services is connect_timeout.
6900 If this option is explicitly set, its value applies to all services.
6903 NAME: icap_io_timeout
6907 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
6910 This parameter specifies how long to wait for an I/O activity on
6911 an established, active ICAP connection before giving up and
6912 either terminating the HTTP transaction or bypassing the
6915 The default is read_timeout.
6918 NAME: icap_service_failure_limit
6919 COMMENT: limit [in memory-depth time-units]
6920 TYPE: icap_service_failure_limit
6922 LOC: Adaptation::Icap::TheConfig
6925 The limit specifies the number of failures that Squid tolerates
6926 when establishing a new TCP connection with an ICAP service. If
6927 the number of failures exceeds the limit, the ICAP service is
6928 not used for new ICAP requests until it is time to refresh its
6931 A negative value disables the limit. Without the limit, an ICAP
6932 service will not be considered down due to connectivity failures
6933 between ICAP OPTIONS requests.
6935 Squid forgets ICAP service failures older than the specified
6936 value of memory-depth. The memory fading algorithm
6937 is approximate because Squid does not remember individual
6938 errors but groups them instead, splitting the option
6939 value into ten time slots of equal length.
6941 When memory-depth is 0 and by default this option has no
6942 effect on service failure expiration.
6944 Squid always forgets failures when updating service settings
6945 using an ICAP OPTIONS transaction, regardless of this option
6949 # suspend service usage after 10 failures in 5 seconds:
6950 icap_service_failure_limit 10 in 5 seconds
6953 NAME: icap_service_revival_delay
6956 LOC: Adaptation::Icap::TheConfig.service_revival_delay
6959 The delay specifies the number of seconds to wait after an ICAP
6960 OPTIONS request failure before requesting the options again. The
6961 failed ICAP service is considered "down" until fresh OPTIONS are
6964 The actual delay cannot be smaller than the hardcoded minimum
6965 delay of 30 seconds.
6968 NAME: icap_preview_enable
6972 LOC: Adaptation::Icap::TheConfig.preview_enable
6975 The ICAP Preview feature allows the ICAP server to handle the
6976 HTTP message by looking only at the beginning of the message body
6977 or even without receiving the body at all. In some environments,
6978 previews greatly speedup ICAP processing.
6980 During an ICAP OPTIONS transaction, the server may tell Squid what
6981 HTTP messages should be previewed and how big the preview should be.
6982 Squid will not use Preview if the server did not request one.
6984 To disable ICAP Preview for all ICAP services, regardless of
6985 individual ICAP server OPTIONS responses, set this option to "off".
6987 icap_preview_enable off
6990 NAME: icap_preview_size
6993 LOC: Adaptation::Icap::TheConfig.preview_size
6996 The default size of preview data to be sent to the ICAP server.
6997 -1 means no preview. This value might be overwritten on a per server
6998 basis by OPTIONS requests.
7001 NAME: icap_206_enable
7005 LOC: Adaptation::Icap::TheConfig.allow206_enable
7008 206 (Partial Content) responses is an ICAP extension that allows the
7009 ICAP agents to optionally combine adapted and original HTTP message
7010 content. The decision to combine is postponed until the end of the
7011 ICAP response. Squid supports Partial Content extension by default.
7013 Activation of the Partial Content extension is negotiated with each
7014 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7015 negotation correctly even if they do not support the extension, but
7016 some might fail. To disable Partial Content support for all ICAP
7017 services and to avoid any negotiation, set this option to "off".
7023 NAME: icap_default_options_ttl
7026 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7029 The default TTL value for ICAP OPTIONS responses that don't have
7030 an Options-TTL header.
7033 NAME: icap_persistent_connections
7037 LOC: Adaptation::Icap::TheConfig.reuse_connections
7040 Whether or not Squid should use persistent connections to
7044 NAME: adaptation_send_client_ip icap_send_client_ip
7046 IFDEF: USE_ADAPTATION
7048 LOC: Adaptation::Config::send_client_ip
7051 If enabled, Squid shares HTTP client IP information with adaptation
7052 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7053 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7055 See also: adaptation_uses_indirect_client
7058 NAME: adaptation_send_username icap_send_client_username
7060 IFDEF: USE_ADAPTATION
7062 LOC: Adaptation::Config::send_username
7065 This sends authenticated HTTP client username (if available) to
7066 the adaptation service.
7068 For ICAP, the username value is encoded based on the
7069 icap_client_username_encode option and is sent using the header
7070 specified by the icap_client_username_header option.
7073 NAME: icap_client_username_header
7076 LOC: Adaptation::Icap::TheConfig.client_username_header
7077 DEFAULT: X-Client-Username
7079 ICAP request header name to use for adaptation_send_username.
7082 NAME: icap_client_username_encode
7086 LOC: Adaptation::Icap::TheConfig.client_username_encode
7089 Whether to base64 encode the authenticated client username.
7093 TYPE: icap_service_type
7095 LOC: Adaptation::Icap::TheConfig
7098 Defines a single ICAP service using the following format:
7100 icap_service id vectoring_point uri [option ...]
7103 an opaque identifier or name which is used to direct traffic to
7104 this specific service. Must be unique among all adaptation
7105 services in squid.conf.
7107 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7108 This specifies at which point of transaction processing the
7109 ICAP service should be activated. *_postcache vectoring points
7110 are not yet supported.
7112 uri: icap://servername:port/servicepath
7113 ICAP server and service location.
7115 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7116 transactions. Squid does not enforce that requirement. You can specify
7117 services with the same service_url and different vectoring_points. You
7118 can even specify multiple identical services as long as their
7119 service_names differ.
7122 Service options are separated by white space. ICAP services support
7123 the following name=value options:
7126 If set to 'on' or '1', the ICAP service is treated as
7127 optional. If the service cannot be reached or malfunctions,
7128 Squid will try to ignore any errors and process the message as
7129 if the service was not enabled. No all ICAP errors can be
7130 bypassed. If set to 0, the ICAP service is treated as
7131 essential and all ICAP errors will result in an error page
7132 returned to the HTTP client.
7134 Bypass is off by default: services are treated as essential.
7137 If set to 'on' or '1', the ICAP service is allowed to
7138 dynamically change the current message adaptation plan by
7139 returning a chain of services to be used next. The services
7140 are specified using the X-Next-Services ICAP response header
7141 value, formatted as a comma-separated list of service names.
7142 Each named service should be configured in squid.conf. Other
7143 services are ignored. An empty X-Next-Services value results
7144 in an empty plan which ends the current adaptation.
7146 Dynamic adaptation plan may cross or cover multiple supported
7147 vectoring points in their natural processing order.
7149 Routing is not allowed by default: the ICAP X-Next-Services
7150 response header is ignored.
7153 Only has effect on split-stack systems. The default on those systems
7154 is to use IPv4-only connections. When set to 'on' this option will
7155 make Squid use IPv6-only connections to contact this ICAP service.
7157 on-overload=block|bypass|wait|force
7158 If the service Max-Connections limit has been reached, do
7159 one of the following for each new ICAP transaction:
7160 * block: send an HTTP error response to the client
7161 * bypass: ignore the "over-connected" ICAP service
7162 * wait: wait (in a FIFO queue) for an ICAP connection slot
7163 * force: proceed, ignoring the Max-Connections limit
7165 In SMP mode with N workers, each worker assumes the service
7166 connection limit is Max-Connections/N, even though not all
7167 workers may use a given service.
7169 The default value is "bypass" if service is bypassable,
7170 otherwise it is set to "wait".
7174 Use the given number as the Max-Connections limit, regardless
7175 of the Max-Connections value given by the service, if any.
7177 Older icap_service format without optional named parameters is
7178 deprecated but supported for backward compatibility.
7181 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7182 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7186 TYPE: icap_class_type
7191 This deprecated option was documented to define an ICAP service
7192 chain, even though it actually defined a set of similar, redundant
7193 services, and the chains were not supported.
7195 To define a set of redundant services, please use the
7196 adaptation_service_set directive. For service chains, use
7197 adaptation_service_chain.
7201 TYPE: icap_access_type
7206 This option is deprecated. Please use adaptation_access, which
7207 has the same ICAP functionality, but comes with better
7208 documentation, and eCAP support.
7213 -----------------------------------------------------------------------------
7220 LOC: Adaptation::Ecap::TheConfig.onoff
7223 Controls whether eCAP support is enabled.
7227 TYPE: ecap_service_type
7229 LOC: Adaptation::Ecap::TheConfig
7232 Defines a single eCAP service
7234 ecap_service id vectoring_point uri [option ...]
7237 an opaque identifier or name which is used to direct traffic to
7238 this specific service. Must be unique among all adaptation
7239 services in squid.conf.
7241 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7242 This specifies at which point of transaction processing the
7243 eCAP service should be activated. *_postcache vectoring points
7244 are not yet supported.
7246 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7247 Squid uses the eCAP service URI to match this configuration
7248 line with one of the dynamically loaded services. Each loaded
7249 eCAP service must have a unique URI. Obtain the right URI from
7250 the service provider.
7253 Service options are separated by white space. eCAP services support
7254 the following name=value options:
7257 If set to 'on' or '1', the eCAP service is treated as optional.
7258 If the service cannot be reached or malfunctions, Squid will try
7259 to ignore any errors and process the message as if the service
7260 was not enabled. No all eCAP errors can be bypassed.
7261 If set to 'off' or '0', the eCAP service is treated as essential
7262 and all eCAP errors will result in an error page returned to the
7265 Bypass is off by default: services are treated as essential.
7268 If set to 'on' or '1', the eCAP service is allowed to
7269 dynamically change the current message adaptation plan by
7270 returning a chain of services to be used next.
7272 Dynamic adaptation plan may cross or cover multiple supported
7273 vectoring points in their natural processing order.
7275 Routing is not allowed by default.
7277 Older ecap_service format without optional named parameters is
7278 deprecated but supported for backward compatibility.
7282 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7283 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7286 NAME: loadable_modules
7288 IFDEF: USE_LOADABLE_MODULES
7289 LOC: Config.loadable_module_names
7292 Instructs Squid to load the specified dynamic module(s) or activate
7293 preloaded module(s).
7295 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7299 MESSAGE ADAPTATION OPTIONS
7300 -----------------------------------------------------------------------------
7303 NAME: adaptation_service_set
7304 TYPE: adaptation_service_set_type
7305 IFDEF: USE_ADAPTATION
7310 Configures an ordered set of similar, redundant services. This is
7311 useful when hot standby or backup adaptation servers are available.
7313 adaptation_service_set set_name service_name1 service_name2 ...
7315 The named services are used in the set declaration order. The first
7316 applicable adaptation service from the set is used first. The next
7317 applicable service is tried if and only if the transaction with the
7318 previous service fails and the message waiting to be adapted is still
7321 When adaptation starts, broken services are ignored as if they were
7322 not a part of the set. A broken service is a down optional service.
7324 The services in a set must be attached to the same vectoring point
7325 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7327 If all services in a set are optional then adaptation failures are
7328 bypassable. If all services in the set are essential, then a
7329 transaction failure with one service may still be retried using
7330 another service from the set, but when all services fail, the master
7331 transaction fails as well.
7333 A set may contain a mix of optional and essential services, but that
7334 is likely to lead to surprising results because broken services become
7335 ignored (see above), making previously bypassable failures fatal.
7336 Technically, it is the bypassability of the last failed service that
7339 See also: adaptation_access adaptation_service_chain
7342 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7343 adaptation service_set svcLogger loggerLocal loggerRemote
7346 NAME: adaptation_service_chain
7347 TYPE: adaptation_service_chain_type
7348 IFDEF: USE_ADAPTATION
7353 Configures a list of complementary services that will be applied
7354 one-by-one, forming an adaptation chain or pipeline. This is useful
7355 when Squid must perform different adaptations on the same message.
7357 adaptation_service_chain chain_name service_name1 svc_name2 ...
7359 The named services are used in the chain declaration order. The first
7360 applicable adaptation service from the chain is used first. The next
7361 applicable service is applied to the successful adaptation results of
7362 the previous service in the chain.
7364 When adaptation starts, broken services are ignored as if they were
7365 not a part of the chain. A broken service is a down optional service.
7367 Request satisfaction terminates the adaptation chain because Squid
7368 does not currently allow declaration of RESPMOD services at the
7369 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7371 The services in a chain must be attached to the same vectoring point
7372 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7374 A chain may contain a mix of optional and essential services. If an
7375 essential adaptation fails (or the failure cannot be bypassed for
7376 other reasons), the master transaction fails. Otherwise, the failure
7377 is bypassed as if the failed adaptation service was not in the chain.
7379 See also: adaptation_access adaptation_service_set
7382 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
7385 NAME: adaptation_access
7386 TYPE: adaptation_access_type
7387 IFDEF: USE_ADAPTATION
7391 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
7393 adaptation_access service_name allow|deny [!]aclname...
7394 adaptation_access set_name allow|deny [!]aclname...
7396 At each supported vectoring point, the adaptation_access
7397 statements are processed in the order they appear in this
7398 configuration file. Statements pointing to the following services
7399 are ignored (i.e., skipped without checking their ACL):
7401 - services serving different vectoring points
7402 - "broken-but-bypassable" services
7403 - "up" services configured to ignore such transactions
7404 (e.g., based on the ICAP Transfer-Ignore header).
7406 When a set_name is used, all services in the set are checked
7407 using the same rules, to find the first applicable one. See
7408 adaptation_service_set for details.
7410 If an access list is checked and there is a match, the
7411 processing stops: For an "allow" rule, the corresponding
7412 adaptation service is used for the transaction. For a "deny"
7413 rule, no adaptation service is activated.
7415 It is currently not possible to apply more than one adaptation
7416 service at the same vectoring point to the same HTTP transaction.
7418 See also: icap_service and ecap_service
7421 adaptation_access service_1 allow all
7424 NAME: adaptation_service_iteration_limit
7426 IFDEF: USE_ADAPTATION
7427 LOC: Adaptation::Config::service_iteration_limit
7430 Limits the number of iterations allowed when applying adaptation
7431 services to a message. If your longest adaptation set or chain
7432 may have more than 16 services, increase the limit beyond its
7433 default value of 16. If detecting infinite iteration loops sooner
7434 is critical, make the iteration limit match the actual number
7435 of services in your longest adaptation set or chain.
7437 Infinite adaptation loops are most likely with routing services.
7439 See also: icap_service routing=1
7442 NAME: adaptation_masterx_shared_names
7444 IFDEF: USE_ADAPTATION
7445 LOC: Adaptation::Config::masterx_shared_name
7448 For each master transaction (i.e., the HTTP request and response
7449 sequence, including all related ICAP and eCAP exchanges), Squid
7450 maintains a table of metadata. The table entries are (name, value)
7451 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7452 with the master transaction.
7454 This option specifies the table entry names that Squid must accept
7455 from and forward to the adaptation transactions.
7457 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7458 shared table by returning an ICAP header field with a name
7459 specified in adaptation_masterx_shared_names.
7461 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7462 shared table by implementing the libecap::visitEachOption() API
7463 to provide an option with a name specified in
7464 adaptation_masterx_shared_names.
7466 Squid will store and forward the set entry to subsequent adaptation
7467 transactions within the same master transaction scope.
7469 Only one shared entry name is supported at this time.
7472 # share authentication information among ICAP services
7473 adaptation_masterx_shared_names X-Subscriber-ID
7476 NAME: adaptation_meta
7478 IFDEF: USE_ADAPTATION
7479 LOC: Adaptation::Config::metaHeaders
7482 This option allows Squid administrator to add custom ICAP request
7483 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7484 Use it to pass custom authentication tokens and other
7485 transaction-state related meta information to an ICAP/eCAP service.
7487 The addition of a meta header is ACL-driven:
7488 adaptation_meta name value [!]aclname ...
7490 Processing for a given header name stops after the first ACL list match.
7491 Thus, it is impossible to add two headers with the same name. If no ACL
7492 lists match for a given header name, no such header is added. For
7495 # do not debug transactions except for those that need debugging
7496 adaptation_meta X-Debug 1 needs_debugging
7498 # log all transactions except for those that must remain secret
7499 adaptation_meta X-Log 1 !keep_secret
7501 # mark transactions from users in the "G 1" group
7502 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7504 The "value" parameter may be a regular squid.conf token or a "double
7505 quoted string". Within the quoted string, use backslash (\) to escape
7506 any character, which is currently only useful for escaping backslashes
7507 and double quotes. For example,
7508 "this string has one backslash (\\) and two \"quotes\""
7510 Used adaptation_meta header values may be logged via %note
7511 logformat code. If multiple adaptation_meta headers with the same name
7512 are used during master transaction lifetime, the header values are
7513 logged in the order they were used and duplicate values are ignored
7514 (only the first repeated value will be logged).
7520 LOC: Adaptation::Icap::TheConfig.repeat
7521 DEFAULT_IF_NONE: deny all
7523 This ACL determines which retriable ICAP transactions are
7524 retried. Transactions that received a complete ICAP response
7525 and did not have to consume or produce HTTP bodies to receive
7526 that response are usually retriable.
7528 icap_retry allow|deny [!]aclname ...
7530 Squid automatically retries some ICAP I/O timeouts and errors
7531 due to persistent connection race conditions.
7533 See also: icap_retry_limit
7536 NAME: icap_retry_limit
7539 LOC: Adaptation::Icap::TheConfig.repeat_limit
7542 Limits the number of retries allowed. When set to zero (default),
7543 no retries are allowed.
7545 Communication errors due to persistent connection race
7546 conditions are unavoidable, automatically retried, and do not
7547 count against this limit.
7549 See also: icap_retry
7555 -----------------------------------------------------------------------------
7558 NAME: check_hostnames
7561 LOC: Config.onoff.check_hostnames
7563 For security and stability reasons Squid can check
7564 hostnames for Internet standard RFC compliance. If you want
7565 Squid to perform these checks turn this directive on.
7568 NAME: allow_underscore
7571 LOC: Config.onoff.allow_underscore
7573 Underscore characters is not strictly allowed in Internet hostnames
7574 but nevertheless used by many sites. Set this to off if you want
7575 Squid to be strict about the standard.
7576 This check is performed only when check_hostnames is set to on.
7579 NAME: cache_dns_program
7581 IFDEF: USE_DNSHELPER
7582 DEFAULT: @DEFAULT_DNSSERVER@
7583 LOC: Config.Program.dnsserver
7585 Specify the location of the executable for dnslookup process.
7589 TYPE: HelperChildConfig
7590 IFDEF: USE_DNSHELPER
7591 DEFAULT: 32 startup=1 idle=1
7592 LOC: Config.dnsChildren
7594 The maximum number of processes spawn to service DNS name lookups.
7595 If you limit it too few Squid will have to wait for them to process
7596 a backlog of requests, slowing it down. If you allow too many they
7597 will use RAM and other system resources noticably.
7598 The maximum this may be safely set to is 32.
7600 The startup= and idle= options allow some measure of skew in your
7605 Sets a minimum of how many processes are to be spawned when Squid
7606 starts or reconfigures. When set to zero the first request will
7607 cause spawning of the first child process to handle it.
7609 Starting too few will cause an initial slowdown in traffic as Squid
7610 attempts to simultaneously spawn enough processes to cope.
7614 Sets a minimum of how many processes Squid is to try and keep available
7615 at all times. When traffic begins to rise above what the existing
7616 processes can handle this many more will be spawned up to the maximum
7617 configured. A minimum setting of 1 is required.
7620 NAME: dns_retransmit_interval
7623 LOC: Config.Timeout.idns_retransmit
7624 IFDEF: !USE_DNSHELPER
7626 Initial retransmit interval for DNS queries. The interval is
7627 doubled each time all configured DNS servers have been tried.
7633 LOC: Config.Timeout.idns_query
7634 IFDEF: !USE_DNSHELPER
7636 DNS Query timeout. If no response is received to a DNS query
7637 within this time all DNS servers for the queried domain
7638 are assumed to be unavailable.
7641 NAME: dns_packet_max
7644 LOC: Config.dns.packet_max
7645 IFDEF: !USE_DNSHELPER
7647 Maximum number of bytes packet size to advertise via EDNS.
7648 Set to "none" to disable EDNS large packet support.
7650 For legacy reasons DNS UDP replies will default to 512 bytes which
7651 is too small for many responses. EDNS provides a means for Squid to
7652 negotiate receiving larger responses back immediately without having
7653 to failover with repeat requests. Responses larger than this limit
7654 will retain the old behaviour of failover to TCP DNS.
7656 Squid has no real fixed limit internally, but allowing packet sizes
7657 over 1500 bytes requires network jumbogram support and is usually not
7660 WARNING: The RFC also indicates that some older resolvers will reply
7661 with failure of the whole request if the extension is added. Some
7662 resolvers have already been identified which will reply with mangled
7663 EDNS response on occasion. Usually in response to many-KB jumbogram
7664 sizes being advertised by Squid.
7665 Squid will currently treat these both as an unable-to-resolve domain
7666 even if it would be resolvable without EDNS.
7673 LOC: Config.onoff.res_defnames
7675 Normally the RES_DEFNAMES resolver option is disabled
7676 (see res_init(3)). This prevents caches in a hierarchy
7677 from interpreting single-component hostnames locally. To allow
7678 Squid to handle single-component names, enable this option.
7681 NAME: dns_nameservers
7684 LOC: Config.dns_nameservers
7686 Use this if you want to specify a list of DNS name servers
7687 (IP addresses) to use instead of those given in your
7688 /etc/resolv.conf file.
7689 On Windows platforms, if no value is specified here or in
7690 the /etc/resolv.conf file, the list of DNS name servers are
7691 taken from the Windows registry, both static and dynamic DHCP
7692 configurations are supported.
7694 Example: dns_nameservers 10.0.0.1 192.172.0.4
7699 DEFAULT: @DEFAULT_HOSTS@
7700 LOC: Config.etcHostsPath
7702 Location of the host-local IP name-address associations
7703 database. Most Operating Systems have such a file on different
7705 - Un*X & Linux: /etc/hosts
7706 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7707 (%SystemRoot% value install default is c:\winnt)
7708 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7709 (%SystemRoot% value install default is c:\windows)
7710 - Windows 9x/Me: %windir%\hosts
7711 (%windir% value is usually c:\windows)
7712 - Cygwin: /etc/hosts
7714 The file contains newline-separated definitions, in the
7715 form ip_address_in_dotted_form name [name ...] names are
7716 whitespace-separated. Lines beginning with an hash (#)
7717 character are comments.
7719 The file is checked at startup and upon configuration.
7720 If set to 'none', it won't be checked.
7721 If append_domain is used, that domain will be added to
7722 domain-local (i.e. not containing any dot character) host
7728 LOC: Config.appendDomain
7731 Appends local domain name to hostnames without any dots in
7732 them. append_domain must begin with a period.
7734 Be warned there are now Internet names with no dots in
7735 them using only top-domain names, so setting this may
7736 cause some Internet sites to become unavailable.
7739 append_domain .yourdomain.com
7742 NAME: ignore_unknown_nameservers
7744 LOC: Config.onoff.ignore_unknown_nameservers
7746 IFDEF: !USE_DNSHELPER
7748 By default Squid checks that DNS responses are received
7749 from the same IP addresses they are sent to. If they
7750 don't match, Squid ignores the response and writes a warning
7751 message to cache.log. You can allow responses from unknown
7752 nameservers by setting this option to 'off'.
7758 LOC: Config.dns.v4_first
7759 IFDEF: !USE_DNSHELPER
7761 With the IPv6 Internet being as fast or faster than IPv4 Internet
7762 for most networks Squid prefers to contact websites over IPv6.
7764 This option reverses the order of preference to make Squid contact
7765 dual-stack websites over IPv4 first. Squid will still perform both
7766 IPv6 and IPv4 DNS lookups before connecting.
7769 This option will restrict the situations under which IPv6
7770 connectivity is used (and tested). Hiding network problems
7771 which would otherwise be detected and warned about.
7775 COMMENT: (number of entries)
7778 LOC: Config.ipcache.size
7785 LOC: Config.ipcache.low
7792 LOC: Config.ipcache.high
7794 The size, low-, and high-water marks for the IP cache.
7797 NAME: fqdncache_size
7798 COMMENT: (number of entries)
7801 LOC: Config.fqdncache.size
7803 Maximum number of FQDN cache entries.
7808 -----------------------------------------------------------------------------
7815 LOC: Config.onoff.mem_pools
7817 If set, Squid will keep pools of allocated (but unused) memory
7818 available for future use. If memory is a premium on your
7819 system and you believe your malloc library outperforms Squid
7820 routines, disable this.
7823 NAME: memory_pools_limit
7827 LOC: Config.MemPools.limit
7829 Used only with memory_pools on:
7830 memory_pools_limit 50 MB
7832 If set to a non-zero value, Squid will keep at most the specified
7833 limit of allocated (but unused) memory in memory pools. All free()
7834 requests that exceed this limit will be handled by your malloc
7835 library. Squid does not pre-allocate any memory, just safe-keeps
7836 objects that otherwise would be free()d. Thus, it is safe to set
7837 memory_pools_limit to a reasonably high value even if your
7838 configuration will use less memory.
7840 If set to none, Squid will keep all memory it can. That is, there
7841 will be no limit on the total amount of memory used for safe-keeping.
7843 To disable memory allocation optimization, do not set
7844 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
7846 An overhead for maintaining memory pools is not taken into account
7847 when the limit is checked. This overhead is close to four bytes per
7848 object kept. However, pools may actually _save_ memory because of
7849 reduced memory thrashing in your malloc library.
7853 COMMENT: on|off|transparent|truncate|delete
7856 LOC: opt_forwarded_for
7858 If set to "on", Squid will append your client's IP address
7859 in the HTTP requests it forwards. By default it looks like:
7861 X-Forwarded-For: 192.1.2.3
7863 If set to "off", it will appear as
7865 X-Forwarded-For: unknown
7867 If set to "transparent", Squid will not alter the
7868 X-Forwarded-For header in any way.
7870 If set to "delete", Squid will delete the entire
7871 X-Forwarded-For header.
7873 If set to "truncate", Squid will remove all existing
7874 X-Forwarded-For entries, and place the client IP as the sole entry.
7877 NAME: cachemgr_passwd
7878 TYPE: cachemgrpasswd
7880 LOC: Config.passwd_list
7882 Specify passwords for cachemgr operations.
7884 Usage: cachemgr_passwd password action action ...
7886 Some valid actions are (see cache manager menu for a full list):
7926 * Indicates actions which will not be performed without a
7927 valid password, others can be performed if not listed here.
7929 To disable an action, set the password to "disable".
7930 To allow performing an action without a password, set the
7933 Use the keyword "all" to set the same password for all actions.
7936 cachemgr_passwd secret shutdown
7937 cachemgr_passwd lesssssssecret info stats/objects
7938 cachemgr_passwd disable all
7945 LOC: Config.onoff.client_db
7947 If you want to disable collecting per-client statistics,
7948 turn off client_db here.
7951 NAME: refresh_all_ims
7955 LOC: Config.onoff.refresh_all_ims
7957 When you enable this option, squid will always check
7958 the origin server for an update when a client sends an
7959 If-Modified-Since request. Many browsers use IMS
7960 requests when the user requests a reload, and this
7961 ensures those clients receive the latest version.
7963 By default (off), squid may return a Not Modified response
7964 based on the age of the cached version.
7967 NAME: reload_into_ims
7968 IFDEF: USE_HTTP_VIOLATIONS
7972 LOC: Config.onoff.reload_into_ims
7974 When you enable this option, client no-cache or ``reload''
7975 requests will be changed to If-Modified-Since requests.
7976 Doing this VIOLATES the HTTP standard. Enabling this
7977 feature could make you liable for problems which it
7980 see also refresh_pattern for a more selective approach.
7983 NAME: connect_retries
7985 LOC: Config.connect_retries
7988 This sets the maximum number of connection attempts made for each
7989 TCP connection. The connect_retries attempts must all still
7990 complete within the connection timeout period.
7992 The default is not to re-try if the first connection attempt fails.
7993 The (not recommended) maximum is 10 tries.
7995 A warning message will be generated if it is set to a too-high
7996 value and the configured value will be over-ridden.
7998 Note: These re-tries are in addition to forward_max_tries
7999 which limit how many different addresses may be tried to find
8003 NAME: retry_on_error
8005 LOC: Config.retry.onerror
8008 If set to ON Squid will automatically retry requests when
8009 receiving an error response with status 403 (Forbidden),
8010 500 (Internal Error), 501 or 503 (Service not available).
8011 Status 502 and 504 (Gateway errors) are always retried.
8013 This is mainly useful if you are in a complex cache hierarchy to
8014 work around access control errors.
8016 NOTE: This retry will attempt to find another working destination.
8017 Which is different from the server which just failed.
8020 NAME: as_whois_server
8022 LOC: Config.as_whois_server
8023 DEFAULT: whois.ra.net
8025 WHOIS server to query for AS numbers. NOTE: AS numbers are
8026 queried only when Squid starts up, not for every request.
8031 LOC: Config.onoff.offline
8034 Enable this option and Squid will never try to validate cached
8038 NAME: uri_whitespace
8039 TYPE: uri_whitespace
8040 LOC: Config.uri_whitespace
8043 What to do with requests that have whitespace characters in the
8046 strip: The whitespace characters are stripped out of the URL.
8047 This is the behavior recommended by RFC2396.
8048 deny: The request is denied. The user receives an "Invalid
8050 allow: The request is allowed and the URI is not changed. The
8051 whitespace characters remain in the URI. Note the
8052 whitespace is passed to redirector processes if they
8054 encode: The request is allowed and the whitespace characters are
8055 encoded according to RFC1738. This could be considered
8056 a violation of the HTTP/1.1
8057 RFC because proxies are not allowed to rewrite URI's.
8058 chop: The request is allowed and the URI is chopped at the
8059 first whitespace. This might also be considered a
8065 LOC: Config.chroot_dir
8068 Specifies a directory where Squid should do a chroot() while
8069 initializing. This also causes Squid to fully drop root
8070 privileges after initializing. This means, for example, if you
8071 use a HTTP port less than 1024 and try to reconfigure, you may
8072 get an error saying that Squid can not open the port.
8075 NAME: balance_on_multiple_ip
8077 LOC: Config.onoff.balance_on_multiple_ip
8080 Modern IP resolvers in squid sort lookup results by preferred access.
8081 By default squid will use these IP in order and only rotates to
8082 the next listed when the most preffered fails.
8084 Some load balancing servers based on round robin DNS have been
8085 found not to preserve user session state across requests
8086 to different IP addresses.
8088 Enabling this directive Squid rotates IP's per request.
8091 NAME: pipeline_prefetch
8093 LOC: Config.onoff.pipeline_prefetch
8096 To boost the performance of pipelined requests to closer
8097 match that of a non-proxied environment Squid can try to fetch
8098 up to two requests in parallel from a pipeline.
8100 Defaults to off for bandwidth management and access logging
8103 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8106 NAME: high_response_time_warning
8109 LOC: Config.warnings.high_rptm
8112 If the one-minute median response time exceeds this value,
8113 Squid prints a WARNING with debug level 0 to get the
8114 administrators attention. The value is in milliseconds.
8117 NAME: high_page_fault_warning
8119 LOC: Config.warnings.high_pf
8122 If the one-minute average page fault rate exceeds this
8123 value, Squid prints a WARNING with debug level 0 to get
8124 the administrators attention. The value is in page faults
8128 NAME: high_memory_warning
8130 LOC: Config.warnings.high_memory
8133 If the memory usage (as determined by mallinfo) exceeds
8134 this amount, Squid prints a WARNING with debug level 0 to get
8135 the administrators attention.
8138 NAME: sleep_after_fork
8139 COMMENT: (microseconds)
8141 LOC: Config.sleep_after_fork
8144 When this is set to a non-zero value, the main Squid process
8145 sleeps the specified number of microseconds after a fork()
8146 system call. This sleep may help the situation where your
8147 system reports fork() failures due to lack of (virtual)
8148 memory. Note, however, if you have a lot of child
8149 processes, these sleep delays will add up and your
8150 Squid will not service requests for some amount of time
8151 until all the child processes have been started.
8152 On Windows value less then 1000 (1 milliseconds) are
8156 NAME: windows_ipaddrchangemonitor
8157 IFDEF: _SQUID_WINDOWS_
8161 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8163 On Windows Squid by default will monitor IP address changes and will
8164 reconfigure itself after any detected event. This is very useful for
8165 proxies connected to internet with dial-up interfaces.
8166 In some cases (a Proxy server acting as VPN gateway is one) it could be
8167 desiderable to disable this behaviour setting this to 'off'.
8168 Note: after changing this, Squid service must be restarted.
8173 IFDEF: USE_SQUID_EUI
8175 LOC: Eui::TheConfig.euiLookup
8177 Whether to lookup the EUI or MAC address of a connected client.
8180 NAME: max_filedescriptors max_filedesc
8183 LOC: Config.max_filedescriptors
8185 The maximum number of filedescriptors supported.
8187 The default "0" means Squid inherits the current ulimit setting.
8189 Note: Changing this requires a restart of Squid. Also
8190 not all comm loops supports large values.
8198 Number of main Squid processes or "workers" to fork and maintain.
8199 0: "no daemon" mode, like running "squid -N ..."
8200 1: "no SMP" mode, start one main Squid process daemon (default)
8201 N: start N main Squid process daemons (i.e., SMP mode)
8203 In SMP mode, each worker does nearly all what a single Squid daemon
8204 does (e.g., listen on http_port and forward HTTP requests).
8207 NAME: cpu_affinity_map
8208 TYPE: CpuAffinityMap
8209 LOC: Config.cpuAffinityMap
8212 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8214 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8216 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8218 affects processes 1 through 4 only and places them on the first
8219 four even cores, starting with core #1.
8221 CPU cores are numbered starting from 1. Requires support for
8222 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8224 Multiple cpu_affinity_map options are merged.