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: url_rewrite_concurrency
129 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
132 # Options Removed in 3.1
136 Remove this line. DNS is no longer tested on startup.
139 NAME: extension_methods
142 Remove this line. All valid methods for HTTP are accepted by default.
145 # 2.7 Options Removed/Replaced in 3.2
150 # 2.7 Options Removed/Replaced in 3.1
158 Remove this line. HTTP/1.1 is supported by default.
161 NAME: upgrade_http0.9
164 Remove this line. ICY/1.0 streaming protocol is supported by default.
167 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
170 Alter these entries. Use the qos_flows directive instead.
173 # Options Removed in 3.0
177 Since squid-3.0 replace with request_header_access or reply_header_access
178 depending on whether you wish to match client requests or server replies.
181 NAME: httpd_accel_no_pmtu_disc
184 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
188 OPTIONS FOR AUTHENTICATION
189 -----------------------------------------------------------------------------
198 This is used to define parameters for the various authentication
199 schemes supported by Squid.
201 format: auth_param scheme parameter [setting]
203 The order in which authentication schemes are presented to the client is
204 dependent on the order the scheme first appears in config file. IE
205 has a bug (it's not RFC 2617 compliant) in that it will use the basic
206 scheme if basic is the first entry presented, even if more secure
207 schemes are presented. For now use the order in the recommended
208 settings section below. If other browsers have difficulties (don't
209 recognize the schemes offered even if you are using basic) either
210 put basic first, or disable the other schemes (by commenting out their
213 Once an authentication scheme is fully configured, it can only be
214 shutdown by shutting squid down and restarting. Changes can be made on
215 the fly and activated with a reconfigure. I.E. You can change to a
216 different helper, but not unconfigure the helper completely.
218 Please note that while this directive defines how Squid processes
219 authentication it does not automatically activate authentication.
220 To use authentication you must in addition make use of ACLs based
221 on login name in http_access (proxy_auth, proxy_auth_regex or
222 external with %LOGIN used in the format tag). The browser will be
223 challenged for authentication on the first such acl encountered
224 in http_access processing and will also be re-challenged for new
225 login credentials if the request is being denied by a proxy_auth
228 WARNING: authentication can't be used in a transparently intercepting
229 proxy as the client then thinks it is talking to an origin server and
230 not the proxy. This is a limitation of bending the TCP/IP protocol to
231 transparently intercepting port 80, not a limitation in Squid.
232 Ports flagged 'transparent', 'intercept', or 'tproxy' have
233 authentication disabled.
235 === Parameters for the basic scheme follow. ===
238 Specify the command for the external authenticator. Such a program
239 reads a line containing "username password" and replies "OK" or
240 "ERR" in an endless loop. "ERR" responses may optionally be followed
241 by a error description available as %m in the returned error page.
242 If you use an authenticator, make sure you have 1 acl of type
245 By default, the basic authentication scheme is not used unless a
246 program is specified.
248 If you want to use the traditional NCSA proxy authentication, set
249 this line to something like
251 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
254 HTTP uses iso-latin-1 as characterset, while some authentication
255 backends such as LDAP expects UTF-8. If this is set to on Squid will
256 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
257 username & password to the helper.
259 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
260 The maximum number of authenticator processes to spawn. If you start too few
261 Squid will have to wait for them to process a backlog of credential
262 verifications, slowing it down. When password verifications are
263 done via a (slow) network you are likely to need lots of
264 authenticator processes.
266 The startup= and idle= options permit some skew in the exact amount
267 run. A minimum of startup=N will begin during startup and reconfigure.
268 Squid will start more in groups of up to idle=N in an attempt to meet
269 traffic needs and to keep idle=N free above those traffic needs up to
272 The concurrency= option sets the number of concurrent requests the
273 helper can process. The default of 0 is used for helpers who only
274 supports one request at a time. Setting this to a number greater than
275 0 changes the protocol used to include a channel number first on the
276 request/response line, allowing multiple requests to be sent to the
277 same helper in parallell without wating for the response.
278 Must not be set unless it's known the helper supports this.
280 auth_param basic children 20 startup=0 idle=1
283 Specifies the realm name which is to be reported to the
284 client for the basic proxy authentication scheme (part of
285 the text the user will see when prompted their username and
286 password). There is no default.
287 auth_param basic realm Squid proxy-caching web server
289 "credentialsttl" timetolive
290 Specifies how long squid assumes an externally validated
291 username:password pair is valid for - in other words how
292 often the helper program is called for that user. Set this
293 low to force revalidation with short lived passwords. Note
294 setting this high does not impact your susceptibility
295 to replay attacks unless you are using an one-time password
296 system (such as SecureID). If you are using such a system,
297 you will be vulnerable to replay attacks unless you also
298 use the max_user_ip ACL in an http_access rule.
300 "casesensitive" on|off
301 Specifies if usernames are case sensitive. Most user databases are
302 case insensitive allowing the same username to be spelled using both
303 lower and upper case letters, but some are case sensitive. This
304 makes a big difference for user_max_ip ACL processing and similar.
305 auth_param basic casesensitive off
307 === Parameters for the digest scheme follow ===
310 Specify the command for the external authenticator. Such
311 a program reads a line containing "username":"realm" and
312 replies with the appropriate H(A1) value hex encoded or
313 ERR if the user (or his H(A1) hash) does not exists.
314 See rfc 2616 for the definition of H(A1).
315 "ERR" responses may optionally be followed by a error description
316 available as %m in the returned error page.
318 By default, the digest authentication scheme is not used unless a
319 program is specified.
321 If you want to use a digest authenticator, set this line to
324 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
327 HTTP uses iso-latin-1 as characterset, while some authentication
328 backends such as LDAP expects UTF-8. If this is set to on Squid will
329 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
330 username & password to the helper.
332 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
333 The maximum number of authenticator processes to spawn (default 5).
334 If you start too few Squid will have to wait for them to
335 process a backlog of H(A1) calculations, slowing it down.
336 When the H(A1) calculations are done via a (slow) network
337 you are likely to need lots of authenticator processes.
339 The startup= and idle= options permit some skew in the exact amount
340 run. A minimum of startup=N will begin during startup and reconfigure.
341 Squid will start more in groups of up to idle=N in an attempt to meet
342 traffic needs and to keep idle=N free above those traffic needs up to
345 The concurrency= option sets the number of concurrent requests the
346 helper can process. The default of 0 is used for helpers who only
347 supports one request at a time. Setting this to a number greater than
348 0 changes the protocol used to include a channel number first on the
349 request/response line, allowing multiple requests to be sent to the
350 same helper in parallell without wating for the response.
351 Must not be set unless it's known the helper supports this.
353 auth_param digest children 20 startup=0 idle=1
356 Specifies the realm name which is to be reported to the
357 client for the digest proxy authentication scheme (part of
358 the text the user will see when prompted their username and
359 password). There is no default.
360 auth_param digest realm Squid proxy-caching web server
362 "nonce_garbage_interval" timeinterval
363 Specifies the interval that nonces that have been issued
364 to client_agent's are checked for validity.
366 "nonce_max_duration" timeinterval
367 Specifies the maximum length of time a given nonce will be
370 "nonce_max_count" number
371 Specifies the maximum number of times a given nonce can be
374 "nonce_strictness" on|off
375 Determines if squid requires strict increment-by-1 behavior
376 for nonce counts, or just incrementing (off - for use when
377 useragents generate nonce counts that occasionally miss 1
378 (ie, 1,2,4,6)). Default off.
380 "check_nonce_count" on|off
381 This directive if set to off can disable the nonce count check
382 completely to work around buggy digest qop implementations in
383 certain mainstream browser versions. Default on to check the
384 nonce count to protect from authentication replay attacks.
386 "post_workaround" on|off
387 This is a workaround to certain buggy browsers who sends
388 an incorrect request digest in POST requests when reusing
389 the same nonce as acquired earlier on a GET request.
391 === NTLM scheme options follow ===
394 Specify the command for the external NTLM authenticator.
395 Such a program reads exchanged NTLMSSP packets with
396 the browser via Squid until authentication is completed.
397 If you use an NTLM authenticator, make sure you have 1 acl
398 of type proxy_auth. By default, the NTLM authenticator_program
401 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
403 "children" numberofchildren [startup=N] [idle=N]
404 The maximum number of authenticator processes to spawn (default 5).
405 If you start too few Squid will have to wait for them to
406 process a backlog of credential verifications, slowing it
407 down. When credential verifications are done via a (slow)
408 network you are likely to need lots of authenticator
411 The startup= and idle= options permit some skew in the exact amount
412 run. A minimum of startup=N will begin during startup and reconfigure.
413 Squid will start more in groups of up to idle=N in an attempt to meet
414 traffic needs and to keep idle=N free above those traffic needs up to
417 auth_param ntlm children 20 startup=0 idle=1
420 If you experience problems with PUT/POST requests when using the
421 Negotiate authentication scheme then you can try setting this to
422 off. This will cause Squid to forcibly close the connection on
423 the initial requests where the browser asks which schemes are
424 supported by the proxy.
426 auth_param ntlm keep_alive on
428 === Options for configuring the NEGOTIATE auth-scheme follow ===
431 Specify the command for the external Negotiate authenticator.
432 This protocol is used in Microsoft Active-Directory enabled setups with
433 the Microsoft Internet Explorer or Mozilla Firefox browsers.
434 Its main purpose is to exchange credentials with the Squid proxy
435 using the Kerberos mechanisms.
436 If you use a Negotiate authenticator, make sure you have at least
437 one acl of type proxy_auth active. By default, the negotiate
438 authenticator_program is not used.
439 The only supported program for this role is the ntlm_auth
440 program distributed as part of Samba, version 4 or later.
442 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
444 "children" numberofchildren [startup=N] [idle=N]
445 The maximum number of authenticator processes to spawn (default 5).
446 If you start too few Squid will have to wait for them to
447 process a backlog of credential verifications, slowing it
448 down. When crendential verifications are done via a (slow)
449 network you are likely to need lots of authenticator
452 The startup= and idle= options permit some skew in the exact amount
453 run. A minimum of startup=N will begin during startup and reconfigure.
454 Squid will start more in groups of up to idle=N in an attempt to meet
455 traffic needs and to keep idle=N free above those traffic needs up to
458 auth_param negotiate children 20 startup=0 idle=1
461 If you experience problems with PUT/POST requests when using the
462 Negotiate authentication scheme then you can try setting this to
463 off. This will cause Squid to forcibly close the connection on
464 the initial requests where the browser asks which schemes are
465 supported by the proxy.
467 auth_param negotiate keep_alive on
472 #Recommended minimum configuration per scheme:
473 #auth_param negotiate program <uncomment and complete this line to activate>
474 #auth_param negotiate children 20 startup=0 idle=1
475 #auth_param negotiate keep_alive on
477 #auth_param ntlm program <uncomment and complete this line to activate>
478 #auth_param ntlm children 20 startup=0 idle=1
479 #auth_param ntlm keep_alive on
481 #auth_param digest program <uncomment and complete this line>
482 #auth_param digest children 20 startup=0 idle=1
483 #auth_param digest realm Squid proxy-caching web server
484 #auth_param digest nonce_garbage_interval 5 minutes
485 #auth_param digest nonce_max_duration 30 minutes
486 #auth_param digest nonce_max_count 50
488 #auth_param basic program <uncomment and complete this line>
489 #auth_param basic children 5 startup=5 idle=1
490 #auth_param basic realm Squid proxy-caching web server
491 #auth_param basic credentialsttl 2 hours
494 NAME: authenticate_cache_garbage_interval
497 LOC: Config.authenticateGCInterval
499 The time period between garbage collection across the username cache.
500 This is a tradeoff between memory utilization (long intervals - say
501 2 days) and CPU (short intervals - say 1 minute). Only change if you
505 NAME: authenticate_ttl
508 LOC: Config.authenticateTTL
510 The time a user & their credentials stay in the logged in
511 user cache since their last request. When the garbage
512 interval passes, all user credentials that have passed their
513 TTL are removed from memory.
516 NAME: authenticate_ip_ttl
518 LOC: Config.authenticateIpTTL
521 If you use proxy authentication and the 'max_user_ip' ACL,
522 this directive controls how long Squid remembers the IP
523 addresses associated with each user. Use a small value
524 (e.g., 60 seconds) if your users might change addresses
525 quickly, as is the case with dialups. You might be safe
526 using a larger value (e.g., 2 hours) in a corporate LAN
527 environment with relatively static address assignments.
532 -----------------------------------------------------------------------------
535 NAME: external_acl_type
536 TYPE: externalAclHelper
537 LOC: Config.externalAclHelperList
540 This option defines external acl classes using a helper program
541 to look up the status
543 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
547 ttl=n TTL in seconds for cached results (defaults to 3600
550 TTL for cached negative lookups (default same
553 Maximum number of acl helper processes spawned to service
554 external acl lookups of this type. (default 20)
556 Minimum number of acl helper processes to spawn during
557 startup and reconfigure to service external acl lookups
558 of this type. (default 0)
560 Number of acl helper processes to keep ahead of traffic
561 loads. Squid will spawn this many at once whenever load
562 rises above the capabilities of existing processes.
563 Up to the value of children-max. (default 1)
564 concurrency=n concurrency level per process. Only used with helpers
565 capable of processing more than one query at a time.
566 cache=n limit the result cache size, default is unbounded.
567 grace=n Percentage remaining of TTL where a refresh of a
568 cached entry should be initiated without needing to
569 wait for a new reply. (default is for no grace period)
570 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
571 ipv4 / ipv6 IP protocol used to communicate with this helper.
572 The default is to auto-detect IPv6 and use it when available.
574 FORMAT specifications
576 %LOGIN Authenticated user login name
577 %EXT_USER Username from previous external acl
578 %EXT_LOG Log details from previous external acl
579 %EXT_TAG Tag from previous external acl
580 %IDENT Ident user name
582 %SRCPORT Client source port
585 %PROTO Requested protocol
587 %PATH Requested URL path
588 %METHOD Request method
589 %MYADDR Squid interface address
590 %MYPORT Squid http_port number
591 %PATH Requested URL-path (including query-string if any)
592 %USER_CERT SSL User certificate in PEM format
593 %USER_CERTCHAIN SSL User certificate chain in PEM format
594 %USER_CERT_xx SSL User certificate subject attribute xx
595 %USER_CA_xx SSL User certificate issuer attribute xx
597 %>{Header} HTTP request header "Header"
599 HTTP request header "Hdr" list member "member"
601 HTTP request header list member using ; as
602 list separator. ; can be any non-alphanumeric
605 %<{Header} HTTP reply header "Header"
607 HTTP reply header "Hdr" list member "member"
609 HTTP reply header list member using ; as
610 list separator. ; can be any non-alphanumeric
613 %% The percent sign. Useful for helpers which need
614 an unchanging input format.
616 In addition to the above, any string specified in the referencing
617 acl will also be included in the helper request line, after the
618 specified formats (see the "acl external" directive)
620 The helper receives lines per the above format specification,
621 and returns lines starting with OK or ERR indicating the validity
622 of the request and optionally followed by additional keywords with
625 General result syntax:
627 OK/ERR keyword=value ...
631 user= The users name (login)
632 password= The users password (for login= cache_peer option)
633 message= Message describing the reason. Available as %o
635 tag= Apply a tag to a request (for both ERR and OK results)
636 Only sets a tag, does not alter existing tags.
637 log= String to be logged in access.log. Available as
638 %ea in logformat specifications
640 If protocol=3.0 (the default) then URL escaping is used to protect
641 each value in both requests and responses.
643 If using protocol=2.5 then all values need to be enclosed in quotes
644 if they may contain whitespace, or the whitespace escaped using \.
645 And quotes or \ characters within the keyword value must be \ escaped.
647 When using the concurrency= option the protocol is changed by
648 introducing a query channel tag infront of the request/response.
649 The query channel tag is a number between 0 and concurrency-1.
656 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
657 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
658 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
659 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
660 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
663 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
664 DEFAULT: localhost src 127.0.0.1/32 ::1
665 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
666 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
668 Defining an Access List
670 Every access list definition must begin with an aclname and acltype,
671 followed by either type-specific arguments or a quoted filename that
674 acl aclname acltype argument ...
675 acl aclname acltype "file" ...
677 When using "file", the file should contain one item per line.
679 By default, regular expressions are CASE-SENSITIVE.
680 To make them case-insensitive, use the -i option. To return case-sensitive
681 use the +i option between patterns, or make a new ACL line without -i.
683 Some acl types require suspending the current request in order
684 to access some external data source.
685 Those which do are marked with the tag [slow], those which
686 don't are marked as [fast].
687 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
688 for further information
690 ***** ACL TYPES AVAILABLE *****
692 acl aclname src ip-address/mask ... # clients IP address [fast]
693 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
694 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
695 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
697 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
698 # The arp ACL requires the special configure option --enable-arp-acl.
699 # Furthermore, the ARP ACL code is not portable to all operating systems.
700 # It works on Linux, Solaris, Windows, FreeBSD, and some
701 # other *BSD variants.
704 # NOTE: Squid can only determine the MAC address for clients that are on
705 # the same subnet. If the client is on a different subnet,
706 # then Squid cannot find out its MAC address.
708 acl aclname srcdomain .foo.com ...
709 # reverse lookup, from client IP [slow]
710 acl aclname dstdomain .foo.com ...
711 # Destination server from URL [fast]
712 acl aclname srcdom_regex [-i] \.foo\.com ...
713 # regex matching client name [slow]
714 acl aclname dstdom_regex [-i] \.foo\.com ...
715 # regex matching server [fast]
717 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
718 # based URL is used and no match is found. The name "none" is used
719 # if the reverse lookup fails.
721 acl aclname src_as number ...
722 acl aclname dst_as number ...
724 # Except for access control, AS numbers can be used for
725 # routing of requests to specific caches. Here's an
726 # example for routing all requests for AS#1241 and only
727 # those to mycache.mydomain.net:
728 # acl asexample dst_as 1241
729 # cache_peer_access mycache.mydomain.net allow asexample
730 # cache_peer_access mycache_mydomain.net deny all
732 acl aclname peername myPeer ...
734 # match against a named cache_peer entry
735 # set unique name= on cache_peer lines for reliable use.
737 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
747 # h1:m1 must be less than h2:m2
749 acl aclname url_regex [-i] ^http:// ...
750 # regex matching on whole URL [fast]
751 acl aclname urlpath_regex [-i] \.gif$ ...
752 # regex matching on URL path [fast]
754 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
756 acl aclname localport 3128 ... # TCP port the client connected to [fast]
757 # NP: for interception mode this is usually '80'
759 acl aclname myportname 3128 ... # http(s)_port name [fast]
761 acl aclname proto HTTP FTP ... # request protocol [fast]
763 acl aclname method GET POST ... # HTTP request method [fast]
765 acl aclname http_status 200 301 500- 400-403 ...
766 # status code in reply [fast]
768 acl aclname browser [-i] regexp ...
769 # pattern match on User-Agent header (see also req_header below) [fast]
771 acl aclname referer_regex [-i] regexp ...
772 # pattern match on Referer header [fast]
773 # Referer is highly unreliable, so use with care
775 acl aclname ident username ...
776 acl aclname ident_regex [-i] pattern ...
777 # string match on ident output [slow]
778 # use REQUIRED to accept any non-null ident.
780 acl aclname proxy_auth [-i] username ...
781 acl aclname proxy_auth_regex [-i] pattern ...
782 # perform http authentication challenge to the client and match against
783 # supplied credentials [slow]
785 # takes a list of allowed usernames.
786 # use REQUIRED to accept any valid username.
788 # Will use proxy authentication in forward-proxy scenarios, and plain
789 # http authenticaiton in reverse-proxy scenarios
791 # NOTE: when a Proxy-Authentication header is sent but it is not
792 # needed during ACL checking the username is NOT logged
795 # NOTE: proxy_auth requires a EXTERNAL authentication program
796 # to check username/password combinations (see
797 # auth_param directive).
799 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
800 # as the browser needs to be configured for using a proxy in order
801 # to respond to proxy authentication.
803 acl aclname snmp_community string ...
804 # A community string to limit access to your SNMP Agent [fast]
807 # acl snmppublic snmp_community public
809 acl aclname maxconn number
810 # This will be matched when the client's IP address has
811 # more than <number> TCP connections established. [fast]
812 # NOTE: This only measures direct TCP links so X-Forwarded-For
813 # indirect clients are not counted.
815 acl aclname max_user_ip [-s] number
816 # This will be matched when the user attempts to log in from more
817 # than <number> different ip addresses. The authenticate_ip_ttl
818 # parameter controls the timeout on the ip entries. [fast]
819 # If -s is specified the limit is strict, denying browsing
820 # from any further IP addresses until the ttl has expired. Without
821 # -s Squid will just annoy the user by "randomly" denying requests.
822 # (the counter is reset each time the limit is reached and a
824 # NOTE: in acceleration mode or where there is mesh of child proxies,
825 # clients may appear to come from multiple addresses if they are
826 # going through proxy farms, so a limit of 1 may cause user problems.
828 acl aclname random probability
829 # Pseudo-randomly match requests. Based on the probability given.
830 # Probability may be written as a decimal (0.333), fraction (1/3)
831 # or ratio of matches:non-matches (3:5).
833 acl aclname req_mime_type [-i] mime-type ...
834 # regex match against the mime type of the request generated
835 # by the client. Can be used to detect file upload or some
836 # types HTTP tunneling requests [fast]
837 # NOTE: This does NOT match the reply. You cannot use this
838 # to match the returned file type.
840 acl aclname req_header header-name [-i] any\.regex\.here
841 # regex match against any of the known request headers. May be
842 # thought of as a superset of "browser", "referer" and "mime-type"
845 acl aclname rep_mime_type [-i] mime-type ...
846 # regex match against the mime type of the reply received by
847 # squid. Can be used to detect file download or some
848 # types HTTP tunneling requests. [fast]
849 # NOTE: This has no effect in http_access rules. It only has
850 # effect in rules that affect the reply data stream such as
853 acl aclname rep_header header-name [-i] any\.regex\.here
854 # regex match against any of the known reply headers. May be
855 # thought of as a superset of "browser", "referer" and "mime-type"
858 acl aclname external class_name [arguments...]
859 # external ACL lookup via a helper class defined by the
860 # external_acl_type directive [slow]
862 acl aclname user_cert attribute values...
863 # match against attributes in a user SSL certificate
864 # attribute is one of DN/C/O/CN/L/ST [fast]
866 acl aclname ca_cert attribute values...
867 # match against attributes a users issuing CA SSL certificate
868 # attribute is one of DN/C/O/CN/L/ST [fast]
870 acl aclname ext_user username ...
871 acl aclname ext_user_regex [-i] pattern ...
872 # string match on username returned by external acl helper [slow]
873 # use REQUIRED to accept any non-null user name.
875 acl aclname tag tagvalue ...
876 # string match on tag returned by external acl helper [slow]
878 acl aclname hier_code codename ...
879 # string match against squid hierarchy code(s); [fast]
880 # e.g., DIRECT, PARENT_HIT, NONE, etc.
882 # NOTE: This has no effect in http_access rules. It only has
883 # effect in rules that affect the reply data stream such as
887 acl aclname ssl_error errorname
888 # match against SSL certificate validation error [fast]
890 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
893 # The following can be used as shortcuts for certificate properties:
894 # [ssl::]certHasExpired: the "not after" field is in the past
895 # [ssl::]certNotYetValid: the "not before" field is in the future
896 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
897 # [ssl::]certSelfSigned: The certificate is self signed.
898 # [ssl::]certDomainMismatch: The certificate CN domain does not
899 # match the name the name of the host we are connecting to.
901 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
902 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
903 # predefined ACLs, just like the 'all' ACL.
905 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
906 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
910 acl macaddress arp 09:00:2b:23:45:67
911 acl myexample dst_as 1241
912 acl password proxy_auth REQUIRED
913 acl fileupload req_mime_type -i ^multipart/form-data$
914 acl javascript rep_mime_type -i ^application/x-javascript$
918 # Recommended minimum configuration:
921 # Example rule allowing access from your local networks.
922 # Adapt to list your (internal) IP networks from where browsing
924 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
925 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
926 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
927 acl localnet src fc00::/7 # RFC 4193 local private network range
928 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
930 acl SSL_ports port 443
931 acl Safe_ports port 80 # http
932 acl Safe_ports port 21 # ftp
933 acl Safe_ports port 443 # https
934 acl Safe_ports port 70 # gopher
935 acl Safe_ports port 210 # wais
936 acl Safe_ports port 1025-65535 # unregistered ports
937 acl Safe_ports port 280 # http-mgmt
938 acl Safe_ports port 488 # gss-http
939 acl Safe_ports port 591 # filemaker
940 acl Safe_ports port 777 # multiling http
941 acl CONNECT method CONNECT
945 NAME: follow_x_forwarded_for
947 IFDEF: FOLLOW_X_FORWARDED_FOR
948 LOC: Config.accessList.followXFF
949 DEFAULT_IF_NONE: deny all
951 Allowing or Denying the X-Forwarded-For header to be followed to
952 find the original source of a request.
954 Requests may pass through a chain of several other proxies
955 before reaching us. The X-Forwarded-For header will contain a
956 comma-separated list of the IP addresses in the chain, with the
957 rightmost address being the most recent.
959 If a request reaches us from a source that is allowed by this
960 configuration item, then we consult the X-Forwarded-For header
961 to see where that host received the request from. If the
962 X-Forwarded-For header contains multiple addresses, we continue
963 backtracking until we reach an address for which we are not allowed
964 to follow the X-Forwarded-For header, or until we reach the first
965 address in the list. For the purpose of ACL used in the
966 follow_x_forwarded_for directive the src ACL type always matches
967 the address we are testing and srcdomain matches its rDNS.
969 The end result of this process is an IP address that we will
970 refer to as the indirect client address. This address may
971 be treated as the client address for access control, ICAP, delay
972 pools and logging, depending on the acl_uses_indirect_client,
973 icap_uses_indirect_client, delay_pool_uses_indirect_client,
974 log_uses_indirect_client and tproxy_uses_indirect_client options.
976 This clause only supports fast acl types.
977 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
979 SECURITY CONSIDERATIONS:
981 Any host for which we follow the X-Forwarded-For header
982 can place incorrect information in the header, and Squid
983 will use the incorrect information as if it were the
984 source address of the request. This may enable remote
985 hosts to bypass any access control restrictions that are
986 based on the client's source addresses.
990 acl localhost src 127.0.0.1
991 acl my_other_proxy srcdomain .proxy.example.com
992 follow_x_forwarded_for allow localhost
993 follow_x_forwarded_for allow my_other_proxy
996 NAME: acl_uses_indirect_client
999 IFDEF: FOLLOW_X_FORWARDED_FOR
1001 LOC: Config.onoff.acl_uses_indirect_client
1003 Controls whether the indirect client address
1004 (see follow_x_forwarded_for) is used instead of the
1005 direct client address in acl matching.
1007 NOTE: maxconn ACL considers direct TCP links and indirect
1008 clients will always have zero. So no match.
1011 NAME: delay_pool_uses_indirect_client
1014 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1016 LOC: Config.onoff.delay_pool_uses_indirect_client
1018 Controls whether the indirect client address
1019 (see follow_x_forwarded_for) is used instead of the
1020 direct client address in delay pools.
1023 NAME: log_uses_indirect_client
1026 IFDEF: FOLLOW_X_FORWARDED_FOR
1028 LOC: Config.onoff.log_uses_indirect_client
1030 Controls whether the indirect client address
1031 (see follow_x_forwarded_for) is used instead of the
1032 direct client address in the access log.
1035 NAME: tproxy_uses_indirect_client
1038 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1040 LOC: Config.onoff.tproxy_uses_indirect_client
1042 Controls whether the indirect client address
1043 (see follow_x_forwarded_for) is used instead of the
1044 direct client address when spoofing the outgoing client.
1046 This has no effect on requests arriving in non-tproxy
1049 SECURITY WARNING: Usage of this option is dangerous
1050 and should not be used trivially. Correct configuration
1051 of follow_x_forewarded_for with a limited set of trusted
1052 sources is required to prevent abuse of your proxy.
1057 LOC: Config.accessList.http
1058 DEFAULT_IF_NONE: deny all
1060 Allowing or Denying access based on defined access lists
1062 Access to the HTTP port:
1063 http_access allow|deny [!]aclname ...
1065 NOTE on default values:
1067 If there are no "access" lines present, the default is to deny
1070 If none of the "access" lines cause a match, the default is the
1071 opposite of the last line in the list. If the last line was
1072 deny, the default is allow. Conversely, if the last line
1073 is allow, the default will be deny. For these reasons, it is a
1074 good idea to have an "deny all" entry at the end of your access
1075 lists to avoid potential confusion.
1077 This clause supports both fast and slow acl types.
1078 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1083 # Recommended minimum Access Permission configuration:
1085 # Only allow cachemgr access from localhost
1086 http_access allow localhost manager
1087 http_access deny manager
1089 # Deny requests to certain unsafe ports
1090 http_access deny !Safe_ports
1092 # Deny CONNECT to other than secure SSL ports
1093 http_access deny CONNECT !SSL_ports
1095 # We strongly recommend the following be uncommented to protect innocent
1096 # web applications running on the proxy server who think the only
1097 # one who can access services on "localhost" is a local user
1098 #http_access deny to_localhost
1101 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1104 # Example rule allowing access from your local networks.
1105 # Adapt localnet in the ACL section to list your (internal) IP networks
1106 # from where browsing should be allowed
1107 http_access allow localnet
1108 http_access allow localhost
1110 # And finally deny all other access to this proxy
1111 http_access deny all
1115 NAME: adapted_http_access http_access2
1117 LOC: Config.accessList.adapted_http
1120 Allowing or Denying access based on defined access lists
1122 Essentially identical to http_access, but runs after redirectors
1123 and ICAP/eCAP adaptation. Allowing access control based on their
1126 If not set then only http_access is used.
1129 NAME: http_reply_access
1131 LOC: Config.accessList.reply
1134 Allow replies to client requests. This is complementary to http_access.
1136 http_reply_access allow|deny [!] aclname ...
1138 NOTE: if there are no access lines present, the default is to allow
1141 If none of the access lines cause a match the opposite of the
1142 last line will apply. Thus it is good practice to end the rules
1143 with an "allow all" or "deny all" entry.
1145 This clause supports both fast and slow acl types.
1146 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1151 LOC: Config.accessList.icp
1152 DEFAULT_IF_NONE: deny all
1154 Allowing or Denying access to the ICP port based on defined
1157 icp_access allow|deny [!]aclname ...
1159 See http_access for details
1161 This clause only supports fast acl types.
1162 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1164 # Allow ICP queries from local networks only
1165 #icp_access allow localnet
1166 #icp_access deny all
1172 LOC: Config.accessList.htcp
1173 DEFAULT_IF_NONE: deny all
1175 Allowing or Denying access to the HTCP port based on defined
1178 htcp_access allow|deny [!]aclname ...
1180 See http_access for details
1182 NOTE: The default if no htcp_access lines are present is to
1183 deny all traffic. This default may cause problems with peers
1184 using the htcp option.
1186 This clause only supports fast acl types.
1187 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1189 # Allow HTCP queries from local networks only
1190 #htcp_access allow localnet
1191 #htcp_access deny all
1194 NAME: htcp_clr_access
1197 LOC: Config.accessList.htcp_clr
1198 DEFAULT_IF_NONE: deny all
1200 Allowing or Denying access to purge content using HTCP based
1201 on defined access lists
1203 htcp_clr_access allow|deny [!]aclname ...
1205 See http_access for details
1207 This clause only supports fast acl types.
1208 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1210 # Allow HTCP CLR requests from trusted peers
1211 acl htcp_clr_peer src 172.16.1.2
1212 htcp_clr_access allow htcp_clr_peer
1217 LOC: Config.accessList.miss
1220 Determins whether network access is permitted when satisfying a request.
1223 to force your neighbors to use you as a sibling instead of
1226 acl localclients src 172.16.0.0/16
1227 miss_access allow localclients
1228 miss_access deny !localclients
1230 This means only your local clients are allowed to fetch relayed/MISS
1231 replies from the network and all other clients can only fetch cached
1235 The default for this setting allows all clients who passed the
1236 http_access rules to relay via this proxy.
1238 This clause only supports fast acl types.
1239 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1242 NAME: ident_lookup_access
1245 DEFAULT_IF_NONE: deny all
1246 LOC: Ident::TheConfig.identLookup
1248 A list of ACL elements which, if matched, cause an ident
1249 (RFC 931) lookup to be performed for this request. For
1250 example, you might choose to always perform ident lookups
1251 for your main multi-user Unix boxes, but not for your Macs
1252 and PCs. By default, ident lookups are not performed for
1255 To enable ident lookups for specific client addresses, you
1256 can follow this example:
1258 acl ident_aware_hosts src 198.168.1.0/24
1259 ident_lookup_access allow ident_aware_hosts
1260 ident_lookup_access deny all
1262 Only src type ACL checks are fully supported. A srcdomain
1263 ACL might work at times, but it will not always provide
1266 This clause only supports fast acl types.
1267 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1270 NAME: reply_body_max_size
1271 COMMENT: size [acl acl...]
1274 LOC: Config.ReplyBodySize
1276 This option specifies the maximum size of a reply body. It can be
1277 used to prevent users from downloading very large files, such as
1278 MP3's and movies. When the reply headers are received, the
1279 reply_body_max_size lines are processed, and the first line where
1280 all (if any) listed ACLs are true is used as the maximum body size
1283 This size is checked twice. First when we get the reply headers,
1284 we check the content-length value. If the content length value exists
1285 and is larger than the allowed size, the request is denied and the
1286 user receives an error message that says "the request or reply
1287 is too large." If there is no content-length, and the reply
1288 size exceeds this limit, the client's connection is just closed
1289 and they will receive a partial reply.
1291 WARNING: downstream caches probably can not detect a partial reply
1292 if there is no content-length header, so they will cache
1293 partial responses and give them out as hits. You should NOT
1294 use this option if you have downstream caches.
1296 WARNING: A maximum size smaller than the size of squid's error messages
1297 will cause an infinite loop and crash squid. Ensure that the smallest
1298 non-zero value you use is greater that the maximum header size plus
1299 the size of your largest error page.
1301 If you set this parameter none (the default), there will be
1304 Configuration Format is:
1305 reply_body_max_size SIZE UNITS [acl ...]
1307 reply_body_max_size 10 MB
1313 -----------------------------------------------------------------------------
1316 NAME: http_port ascii_port
1319 LOC: Config.Sockaddr.http
1321 Usage: port [mode] [options]
1322 hostname:port [mode] [options]
1323 1.2.3.4:port [mode] [options]
1325 The socket addresses where Squid will listen for HTTP client
1326 requests. You may specify multiple socket addresses.
1327 There are three forms: port alone, hostname with port, and
1328 IP address with port. If you specify a hostname or IP
1329 address, Squid binds the socket to that specific
1330 address. Most likely, you do not need to bind to a specific
1331 address, so you can use the port number alone.
1333 If you are running Squid in accelerator mode, you
1334 probably want to listen on port 80 also, or instead.
1336 The -a command line option may be used to specify additional
1337 port(s) where Squid listens for proxy request. Such ports will
1338 be plain proxy ports with no options.
1340 You may specify multiple socket addresses on multiple lines.
1344 intercept Support for IP-Layer interception of
1345 outgoing requests without browser settings.
1346 NP: disables authentication and IPv6 on the port.
1348 tproxy Support Linux TPROXY for spoofing outgoing
1349 connections using the client IP address.
1350 NP: disables authentication and maybe IPv6 on the port.
1352 accel Accelerator / reverse proxy mode
1354 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1355 establish secure connection with the client and with
1356 the server, decrypt HTTPS messages as they pass through
1357 Squid, and treat them as unencrypted HTTP messages,
1358 becoming the man-in-the-middle.
1360 The ssl_bump option is required to fully enable
1361 bumping of CONNECT requests.
1363 Omitting the mode flag causes default forward proxy mode to be used.
1366 Accelerator Mode Options:
1368 defaultsite=domainname
1369 What to use for the Host: header if it is not present
1370 in a request. Determines what site (not origin server)
1371 accelerators should consider the default.
1373 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1375 protocol= Protocol to reconstruct accelerated requests with.
1376 Defaults to http for http_port and https for
1379 vport Virtual host port support. Using the http_port number
1380 instead of the port passed on Host: headers.
1382 vport=NN Virtual host port support. Using the specified port
1383 number instead of the port passed on Host: headers.
1386 Act as if this Squid is the origin server.
1387 This currently means generate new Date: and Expires:
1388 headers on HIT instead of adding Age:.
1390 ignore-cc Ignore request Cache-Control headers.
1392 WARNING: This option violates HTTP specifications if
1393 used in non-accelerator setups.
1395 allow-direct Allow direct forwarding in accelerator mode. Normally
1396 accelerated requests are denied direct forwarding as if
1397 never_direct was used.
1399 WARNING: this option opens accelerator mode to security
1400 vulnerabilities usually only affecting in interception
1401 mode. Make sure to protect forwarding with suitable
1402 http_access rules when using this.
1405 SSL Bump Mode Options:
1406 In addition to these options ssl-bump requires TLS/SSL options.
1408 generate-host-certificates[=<on|off>]
1409 Dynamically create SSL server certificates for the
1410 destination hosts of bumped CONNECT requests.When
1411 enabled, the cert and key options are used to sign
1412 generated certificates. Otherwise generated
1413 certificate will be selfsigned.
1414 If there is a CA certificate lifetime of the generated
1415 certificate equals lifetime of the CA certificate. If
1416 generated certificate is selfsigned lifetime is three
1418 This option is enabled by default when ssl-bump is used.
1419 See the ssl-bump option above for more information.
1421 dynamic_cert_mem_cache_size=SIZE
1422 Approximate total RAM size spent on cached generated
1423 certificates. If set to zero, caching is disabled. The
1424 default value is 4MB.
1428 cert= Path to SSL certificate (PEM format).
1430 key= Path to SSL private key file (PEM format)
1431 if not specified, the certificate file is
1432 assumed to be a combined certificate and
1435 version= The version of SSL/TLS supported
1436 1 automatic (default)
1443 cipher= Colon separated list of supported ciphers.
1444 NOTE: some ciphers such as EDH ciphers depend on
1445 additional settings. If those settings are
1446 omitted the ciphers may be silently ignored
1447 by the OpenSSL library.
1449 options= Various SSL implementation options. The most important
1451 NO_SSLv2 Disallow the use of SSLv2
1452 NO_SSLv3 Disallow the use of SSLv3
1453 NO_TLSv1 Disallow the use of TLSv1.0
1454 NO_TLSv1_1 Disallow the use of TLSv1.1
1455 NO_TLSv1_2 Disallow the use of TLSv1.2
1456 SINGLE_DH_USE Always create a new key when using
1457 temporary/ephemeral DH key exchanges
1458 ALL Enable various bug workarounds
1459 suggested as "harmless" by OpenSSL
1460 Be warned that this reduces SSL/TLS
1461 strength to some attacks.
1462 See OpenSSL SSL_CTX_set_options documentation for a
1463 complete list of options.
1465 clientca= File containing the list of CAs to use when
1466 requesting a client certificate.
1468 cafile= File containing additional CA certificates to
1469 use when verifying client certificates. If unset
1470 clientca will be used.
1472 capath= Directory containing additional CA certificates
1473 and CRL lists to use when verifying client certificates.
1475 crlfile= File of additional CRL lists to use when verifying
1476 the client certificate, in addition to CRLs stored in
1477 the capath. Implies VERIFY_CRL flag below.
1479 dhparams= File containing DH parameters for temporary/ephemeral
1480 DH key exchanges. See OpenSSL documentation for details
1481 on how to create this file.
1482 WARNING: EDH ciphers will be silently disabled if this
1485 sslflags= Various flags modifying the use of SSL:
1487 Don't request client certificates
1488 immediately, but wait until acl processing
1489 requires a certificate (not yet implemented).
1491 Don't use the default CA lists built in
1494 Don't allow for session reuse. Each connection
1495 will result in a new SSL session.
1497 Verify CRL lists when accepting client
1500 Verify CRL lists for all certificates in the
1501 client certificate chain.
1503 sslcontext= SSL session ID context identifier.
1507 connection-auth[=on|off]
1508 use connection-auth=off to tell Squid to prevent
1509 forwarding Microsoft connection oriented authentication
1510 (NTLM, Negotiate and Kerberos)
1512 disable-pmtu-discovery=
1513 Control Path-MTU discovery usage:
1514 off lets OS decide on what to do (default).
1515 transparent disable PMTU discovery when transparent
1517 always disable always PMTU discovery.
1519 In many setups of transparently intercepting proxies
1520 Path-MTU discovery can not work on traffic towards the
1521 clients. This is the case when the intercepting device
1522 does not fully track connections and fails to forward
1523 ICMP must fragment messages to the cache server. If you
1524 have such setup and experience that certain clients
1525 sporadically hang or never complete requests set
1526 disable-pmtu-discovery option to 'transparent'.
1528 name= Specifies a internal name for the port. Defaults to
1529 the port specification (port or addr:port)
1531 tcpkeepalive[=idle,interval,timeout]
1532 Enable TCP keepalive probes of idle connections.
1533 In seconds; idle is the initial time before TCP starts
1534 probing the connection, interval how often to probe, and
1535 timeout the time before giving up.
1537 If you run Squid on a dual-homed machine with an internal
1538 and an external interface we recommend you to specify the
1539 internal address:port in http_port. This way Squid will only be
1540 visible on the internal address.
1544 # Squid normally listens to port 3128
1545 http_port @DEFAULT_HTTP_PORT@
1553 LOC: Config.Sockaddr.https
1555 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1557 The socket address where Squid will listen for client requests made
1558 over TLS or SSL connections. Commonly referred to as HTTPS.
1560 This is most useful for situations where you are running squid in
1561 accelerator mode and you want to do the SSL work at the accelerator level.
1563 You may specify multiple socket addresses on multiple lines,
1564 each with their own SSL certificate and/or options.
1568 accel Accelerator / reverse proxy mode
1570 intercept Support for IP-Layer interception of
1571 outgoing requests without browser settings.
1572 NP: disables authentication and IPv6 on the port.
1574 tproxy Support Linux TPROXY for spoofing outgoing
1575 connections using the client IP address.
1576 NP: disables authentication and maybe IPv6 on the port.
1578 ssl-bump For each intercepted connection allowed by ssl_bump
1579 ACLs, establish a secure connection with the client and with
1580 the server, decrypt HTTPS messages as they pass through
1581 Squid, and treat them as unencrypted HTTP messages,
1582 becoming the man-in-the-middle.
1584 An "ssl_bump server-first" match is required to
1585 fully enable bumping of intercepted SSL connections.
1587 Requires tproxy or intercept.
1589 Omitting the mode flag causes default forward proxy mode to be used.
1592 See http_port for a list of generic options
1597 cert= Path to SSL certificate (PEM format).
1599 key= Path to SSL private key file (PEM format)
1600 if not specified, the certificate file is
1601 assumed to be a combined certificate and
1604 version= The version of SSL/TLS supported
1605 1 automatic (default)
1610 cipher= Colon separated list of supported ciphers.
1612 options= Various SSL engine options. The most important
1614 NO_SSLv2 Disallow the use of SSLv2
1615 NO_SSLv3 Disallow the use of SSLv3
1616 NO_TLSv1 Disallow the use of TLSv1
1617 SINGLE_DH_USE Always create a new key when using
1618 temporary/ephemeral DH key exchanges
1619 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1620 documentation for a complete list of options.
1622 clientca= File containing the list of CAs to use when
1623 requesting a client certificate.
1625 cafile= File containing additional CA certificates to
1626 use when verifying client certificates. If unset
1627 clientca will be used.
1629 capath= Directory containing additional CA certificates
1630 and CRL lists to use when verifying client certificates.
1632 crlfile= File of additional CRL lists to use when verifying
1633 the client certificate, in addition to CRLs stored in
1634 the capath. Implies VERIFY_CRL flag below.
1636 dhparams= File containing DH parameters for temporary/ephemeral
1639 sslflags= Various flags modifying the use of SSL:
1641 Don't request client certificates
1642 immediately, but wait until acl processing
1643 requires a certificate (not yet implemented).
1645 Don't use the default CA lists built in
1648 Don't allow for session reuse. Each connection
1649 will result in a new SSL session.
1651 Verify CRL lists when accepting client
1654 Verify CRL lists for all certificates in the
1655 client certificate chain.
1657 sslcontext= SSL session ID context identifier.
1659 generate-host-certificates[=<on|off>]
1660 Dynamically create SSL server certificates for the
1661 destination hosts of bumped SSL requests.When
1662 enabled, the cert and key options are used to sign
1663 generated certificates. Otherwise generated
1664 certificate will be selfsigned.
1665 If there is CA certificate life time of generated
1666 certificate equals lifetime of CA certificate. If
1667 generated certificate is selfsigned lifetime is three
1669 This option is enabled by default when SslBump is used.
1670 See the sslBump option above for more information.
1672 dynamic_cert_mem_cache_size=SIZE
1673 Approximate total RAM size spent on cached generated
1674 certificates. If set to zero, caching is disabled. The
1675 default value is 4MB.
1677 See http_port for a list of available options.
1680 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1683 LOC: Ip::Qos::TheConfig.tosToServer
1685 Allows you to select a TOS/Diffserv value for packets outgoing
1686 on the server side, based on an ACL.
1688 tcp_outgoing_tos ds-field [!]aclname ...
1690 Example where normal_service_net uses the TOS value 0x00
1691 and good_service_net uses 0x20
1693 acl normal_service_net src 10.0.0.0/24
1694 acl good_service_net src 10.0.1.0/24
1695 tcp_outgoing_tos 0x00 normal_service_net
1696 tcp_outgoing_tos 0x20 good_service_net
1698 TOS/DSCP values really only have local significance - so you should
1699 know what you're specifying. For more information, see RFC2474,
1700 RFC2475, and RFC3260.
1702 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1703 "default" to use whatever default your host has. Note that in
1704 practice often only multiples of 4 is usable as the two rightmost bits
1705 have been redefined for use by ECN (RFC 3168 section 23.1).
1707 Processing proceeds in the order specified, and stops at first fully
1711 NAME: clientside_tos
1714 LOC: Ip::Qos::TheConfig.tosToClient
1716 Allows you to select a TOS/Diffserv value for packets being transmitted
1717 on the client-side, based on an ACL.
1719 clientside_tos ds-field [!]aclname ...
1721 Example where normal_service_net uses the TOS value 0x00
1722 and good_service_net uses 0x20
1724 acl normal_service_net src 10.0.0.0/24
1725 acl good_service_net src 10.0.1.0/24
1726 clientside_tos 0x00 normal_service_net
1727 clientside_tos 0x20 good_service_net
1729 Note: This feature is incompatible with qos_flows. Any TOS values set here
1730 will be overwritten by TOS values in qos_flows.
1733 NAME: tcp_outgoing_mark
1735 IFDEF: SO_MARK&&USE_LIBCAP
1737 LOC: Ip::Qos::TheConfig.nfmarkToServer
1739 Allows you to apply a Netfilter mark value to outgoing packets
1740 on the server side, based on an ACL.
1742 tcp_outgoing_mark mark-value [!]aclname ...
1744 Example where normal_service_net uses the mark value 0x00
1745 and good_service_net uses 0x20
1747 acl normal_service_net src 10.0.0.0/24
1748 acl good_service_net src 10.0.1.0/24
1749 tcp_outgoing_mark 0x00 normal_service_net
1750 tcp_outgoing_mark 0x20 good_service_net
1753 NAME: clientside_mark
1755 IFDEF: SO_MARK&&USE_LIBCAP
1757 LOC: Ip::Qos::TheConfig.nfmarkToClient
1759 Allows you to apply a Netfilter mark value to packets being transmitted
1760 on the client-side, based on an ACL.
1762 clientside_mark mark-value [!]aclname ...
1764 Example where normal_service_net uses the mark value 0x00
1765 and good_service_net uses 0x20
1767 acl normal_service_net src 10.0.0.0/24
1768 acl good_service_net src 10.0.1.0/24
1769 clientside_mark 0x00 normal_service_net
1770 clientside_mark 0x20 good_service_net
1772 Note: This feature is incompatible with qos_flows. Any mark values set here
1773 will be overwritten by mark values in qos_flows.
1780 LOC: Ip::Qos::TheConfig
1782 Allows you to select a TOS/DSCP value to mark outgoing
1783 connections with, based on where the reply was sourced. For
1784 platforms using netfilter, allows you to set a netfilter mark
1785 value instead of, or in addition to, a TOS value.
1787 TOS values really only have local significance - so you should
1788 know what you're specifying. For more information, see RFC2474,
1789 RFC2475, and RFC3260.
1791 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1792 in practice often only multiples of 4 is usable as the two rightmost bits
1793 have been redefined for use by ECN (RFC 3168 section 23.1).
1795 Mark values can be any unsigned 32-bit integer value.
1797 This setting is configured by setting the following values:
1799 tos|mark Whether to set TOS or netfilter mark values
1801 local-hit=0xFF Value to mark local cache hits.
1803 sibling-hit=0xFF Value to mark hits from sibling peers.
1805 parent-hit=0xFF Value to mark hits from parent peers.
1807 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1808 over the preserve-miss feature (see below), unless
1809 mask is specified, in which case only the bits
1810 specified in the mask are written.
1812 The TOS variant of the following features are only possible on Linux
1813 and require your kernel to be patched with the TOS preserving ZPH
1814 patch, available from http://zph.bratcheda.org
1815 No patch is needed to preserve the netfilter mark, which will work
1816 with all variants of netfilter.
1818 disable-preserve-miss
1819 This option disables the preservation of the TOS or netfilter
1820 mark. By default, the existing TOS or netfilter mark value of
1821 the response coming from the remote server will be retained
1822 and masked with miss-mark.
1823 NOTE: in the case of a netfilter mark, the mark must be set on
1824 the connection (using the CONNMARK target) not on the packet
1828 Allows you to mask certain bits in the TOS or mark value
1829 received from the remote server, before copying the value to
1830 the TOS sent towards clients.
1831 Default for tos: 0xFF (TOS from server is not changed).
1832 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1834 All of these features require the --enable-zph-qos compilation flag
1835 (enabled by default). Netfilter marking also requires the
1836 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1837 libcap 2.09+ (--with-libcap).
1841 NAME: tcp_outgoing_address
1844 LOC: Config.accessList.outgoing_address
1846 Allows you to map requests to different outgoing IP addresses
1847 based on the username or source address of the user making
1850 tcp_outgoing_address ipaddr [[!]aclname] ...
1853 Forwarding clients with dedicated IPs for certain subnets.
1855 acl normal_service_net src 10.0.0.0/24
1856 acl good_service_net src 10.0.2.0/24
1858 tcp_outgoing_address 2001:db8::c001 good_service_net
1859 tcp_outgoing_address 10.1.0.2 good_service_net
1861 tcp_outgoing_address 2001:db8::beef normal_service_net
1862 tcp_outgoing_address 10.1.0.1 normal_service_net
1864 tcp_outgoing_address 2001:db8::1
1865 tcp_outgoing_address 10.1.0.3
1867 Processing proceeds in the order specified, and stops at first fully
1870 Squid will add an implicit IP version test to each line.
1871 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1872 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1875 NOTE: The use of this directive using client dependent ACLs is
1876 incompatible with the use of server side persistent connections. To
1877 ensure correct results it is best to set server_persistent_connections
1878 to off when using this directive in such configurations.
1880 NOTE: The use of this directive to set a local IP on outgoing TCP links
1881 is incompatible with using TPROXY to set client IP out outbound TCP links.
1882 When needing to contact peers use the no-tproxy cache_peer option and the
1883 client_dst_passthru directive re-enable normal forwarding such as this.
1887 NAME: host_verify_strict
1890 LOC: Config.onoff.hostStrictVerify
1892 Regardless of this option setting, when dealing with intercepted
1893 traffic, Squid always verifies that the destination IP address matches
1894 the Host header domain or IP (called 'authority form URL').
1896 This enforcement is performed to satisfy a MUST-level requirement in
1897 RFC 2616 section 14.23: "The Host field value MUST represent the naming
1898 authority of the origin server or gateway given by the original URL".
1901 Squid always responds with an HTTP 409 (Conflict) error
1902 page and logs a security warning if there is no match.
1904 Squid verifies that the destination IP address matches
1905 the Host header for forward-proxy and reverse-proxy traffic
1906 as well. For those traffic types, Squid also enables the
1907 following checks, comparing the corresponding Host header
1908 and Request-URI components:
1910 * The host names (domain or IP) must be identical,
1911 but valueless or missing Host header disables all checks.
1912 For the two host names to match, both must be either IP
1915 * Port numbers must be identical, but if a port is missing
1916 the scheme-default port is assumed.
1919 When set to OFF (the default):
1920 Squid allows suspicious requests to continue but logs a
1921 security warning and blocks caching of the response.
1923 * Forward-proxy traffic is not checked at all.
1925 * Reverse-proxy traffic is not checked at all.
1927 * Intercepted traffic which passes verification is handled
1928 according to client_dst_passthru.
1930 * Intercepted requests which fail verification are sent
1931 to the client original destination instead of DIRECT.
1932 This overrides 'client_dst_passthru off'.
1934 For now suspicious intercepted CONNECT requests are always
1935 responded to with an HTTP 409 (Conflict) error page.
1940 As described in CVE-2009-0801 when the Host: header alone is used
1941 to determine the destination of a request it becomes trivial for
1942 malicious scripts on remote websites to bypass browser same-origin
1943 security policy and sandboxing protections.
1945 The cause of this is that such applets are allowed to perform their
1946 own HTTP stack, in which case the same-origin policy of the browser
1947 sandbox only verifies that the applet tries to contact the same IP
1948 as from where it was loaded at the IP level. The Host: header may
1949 be different from the connected IP and approved origin.
1953 NAME: client_dst_passthru
1956 LOC: Config.onoff.client_dst_passthru
1958 With NAT or TPROXY intercepted traffic Squid may pass the request
1959 directly to the original client destination IP or seek a faster
1960 source using the HTTP Host header.
1962 Using Host to locate alternative servers can provide faster
1963 connectivity with a range of failure recovery options.
1964 But can also lead to connectivity trouble when the client and
1965 server are attempting stateful interactions unaware of the proxy.
1967 This option (on by default) prevents alternative DNS entries being
1968 located to send intercepted traffic DIRECT to an origin server.
1969 The clients original destination IP and port will be used instead.
1971 Regardless of this option setting, when dealing with intercepted
1972 traffic Squid will verify the Host: header and any traffic which
1973 fails Host verification will be treated as if this option were ON.
1975 see host_verify_strict for details on the verification process.
1980 -----------------------------------------------------------------------------
1983 NAME: ssl_unclean_shutdown
1987 LOC: Config.SSL.unclean_shutdown
1989 Some browsers (especially MSIE) bugs out on SSL shutdown
1996 LOC: Config.SSL.ssl_engine
1999 The OpenSSL engine to use. You will need to set this if you
2000 would like to use hardware SSL acceleration for example.
2003 NAME: sslproxy_client_certificate
2006 LOC: Config.ssl_client.cert
2009 Client SSL Certificate to use when proxying https:// URLs
2012 NAME: sslproxy_client_key
2015 LOC: Config.ssl_client.key
2018 Client SSL Key to use when proxying https:// URLs
2021 NAME: sslproxy_version
2024 LOC: Config.ssl_client.version
2027 SSL version level to use when proxying https:// URLs
2029 The versions of SSL/TLS supported:
2031 1 automatic (default)
2039 NAME: sslproxy_options
2042 LOC: Config.ssl_client.options
2045 SSL implementation options to use when proxying https:// URLs
2047 The most important being:
2049 NO_SSLv2 Disallow the use of SSLv2
2050 NO_SSLv3 Disallow the use of SSLv3
2051 NO_TLSv1 Disallow the use of TLSv1.0
2052 NO_TLSv1_1 Disallow the use of TLSv1.1
2053 NO_TLSv1_2 Disallow the use of TLSv1.2
2055 Always create a new key when using temporary/ephemeral
2058 Disable use of RFC5077 session tickets. Some servers
2059 may have problems understanding the TLS extension due
2060 to ambiguous specification in RFC4507.
2061 ALL Enable various bug workarounds suggested as "harmless"
2062 by OpenSSL. Be warned that this may reduce SSL/TLS
2063 strength to some attacks.
2065 See the OpenSSL SSL_CTX_set_options documentation for a
2066 complete list of possible options.
2069 NAME: sslproxy_cipher
2072 LOC: Config.ssl_client.cipher
2075 SSL cipher list to use when proxying https:// URLs
2077 Colon separated list of supported ciphers.
2080 NAME: sslproxy_cafile
2083 LOC: Config.ssl_client.cafile
2086 file containing CA certificates to use when verifying server
2087 certificates while proxying https:// URLs
2090 NAME: sslproxy_capath
2093 LOC: Config.ssl_client.capath
2096 directory containing CA certificates to use when verifying
2097 server certificates while proxying https:// URLs
2102 TYPE: sslproxy_ssl_bump
2103 LOC: Config.accessList.ssl_bump
2106 This option is consulted when a CONNECT request is received on
2107 an http_port (or a new connection is intercepted at an
2108 https_port), provided that port was configured with an ssl-bump
2109 flag. The subsequent data on the connection is either treated as
2110 HTTPS and decrypted OR tunneled at TCP level without decryption,
2111 depending on the first bumping "mode" which ACLs match.
2113 ssl_bump <mode> [!]acl ...
2115 The following bumping modes are supported:
2118 Allow bumping of the connection. Establish a secure connection
2119 with the client first, then connect to the server. This old mode
2120 does not allow Squid to mimic server SSL certificate and does
2121 not work with intercepted SSL connections.
2124 Allow bumping of the connection. Establish a secure connection
2125 with the server first, then establish a secure connection with
2126 the client, using a mimicked server certificate. Works with both
2127 CONNECT requests and intercepted SSL connections.
2130 Become a TCP tunnel without decoding the connection.
2131 Works with both CONNECT requests and intercepted SSL
2132 connections. This is the default behavior when no
2133 ssl_bump option is given or no ssl_bump ACLs match.
2135 By default, no connections are bumped.
2137 The first matching ssl_bump option wins. If no ACLs match, the
2138 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2139 does not have an implicit "negate the last given option" rule. You
2140 must make that rule explicit if you convert old ssl_bump allow/deny
2141 rules that rely on such an implicit rule.
2143 This clause supports both fast and slow acl types.
2144 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2146 See also: http_port ssl-bump, https_port ssl-bump
2149 # Example: Bump all requests except those originating from
2150 # localhost and those going to example.com.
2152 acl broken_sites dstdomain .example.com
2153 ssl_bump none localhost
2154 ssl_bump none broken_sites
2155 ssl_bump server-first all
2158 NAME: sslproxy_flags
2161 LOC: Config.ssl_client.flags
2164 Various flags modifying the use of SSL while proxying https:// URLs:
2165 DONT_VERIFY_PEER Accept certificates that fail verification.
2166 For refined control, see sslproxy_cert_error.
2167 NO_DEFAULT_CA Don't use the default CA list built in
2171 NAME: sslproxy_cert_error
2174 LOC: Config.ssl_client.cert_error
2177 Use this ACL to bypass server certificate validation errors.
2179 For example, the following lines will bypass all validation errors
2180 when talking to servers for example.com. All other
2181 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2183 acl BrokenButTrustedServers dstdomain example.com
2184 sslproxy_cert_error allow BrokenButTrustedServers
2185 sslproxy_cert_error deny all
2187 This clause only supports fast acl types.
2188 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2189 Using slow acl types may result in server crashes
2191 Without this option, all server certificate validation errors
2192 terminate the transaction. Bypassing validation errors is dangerous
2193 because an error usually implies that the server cannot be trusted and
2194 the connection may be insecure.
2196 See also: sslproxy_flags and DONT_VERIFY_PEER.
2198 Default setting: sslproxy_cert_error deny all
2201 NAME: sslproxy_cert_sign
2204 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2205 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2206 POSTSCRIPTUM: signTrusted all
2207 TYPE: sslproxy_cert_sign
2208 LOC: Config.ssl_client.cert_sign
2211 sslproxy_cert_sign <signing algorithm> acl ...
2213 The following certificate signing algorithms are supported:
2215 Sign using the configured CA certificate which is usually
2216 placed in and trusted by end-user browsers. This is the
2217 default for trusted origin server certificates.
2219 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2220 This is the default for untrusted origin server certificates
2221 that are not self-signed (see ssl::certUntrusted).
2223 Sign using a self-signed certificate with the right CN to
2224 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2225 browser. This is the default for self-signed origin server
2226 certificates (see ssl::certSelfSigned).
2228 This clause only supports fast acl types.
2230 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2231 signing algorithm to generate the certificate and ignores all
2232 subsequent sslproxy_cert_sign options (the first match wins). If no
2233 acl(s) match, the default signing algorithm is determined by errors
2234 detected when obtaining and validating the origin server certificate.
2236 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2237 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2238 CONNECT request that carries a domain name. In all other cases (CONNECT
2239 to an IP address or an intercepted SSL connection), Squid cannot detect
2240 the domain mismatch at certificate generation time when
2241 bump-server-first is used.
2244 NAME: sslproxy_cert_adapt
2247 TYPE: sslproxy_cert_adapt
2248 LOC: Config.ssl_client.cert_adapt
2251 sslproxy_cert_adapt <adaptation algorithm> acl ...
2253 The following certificate adaptation algorithms are supported:
2255 Sets the "Not After" property to the "Not After" property of
2256 the CA certificate used to sign generated certificates.
2258 Sets the "Not Before" property to the "Not Before" property of
2259 the CA certificate used to sign generated certificates.
2260 setCommonName or setCommonName{CN}
2261 Sets Subject.CN property to the host name specified as a
2262 CN parameter or, if no explicit CN parameter was specified,
2263 extracted from the CONNECT request. It is a misconfiguration
2264 to use setCommonName without an explicit parameter for
2265 intercepted or tproxied SSL connections.
2267 This clause only supports fast acl types.
2269 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2270 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2271 corresponding adaptation algorithm to generate the certificate and
2272 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2273 group (i.e., the first match wins within each algorithm group). If no
2274 acl(s) match, the default mimicking action takes place.
2276 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2277 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2278 CONNECT request that carries a domain name. In all other cases (CONNECT
2279 to an IP address or an intercepted SSL connection), Squid cannot detect
2280 the domain mismatch at certificate generation time when
2281 bump-server-first is used.
2284 NAME: sslpassword_program
2287 LOC: Config.Program.ssl_password
2290 Specify a program used for entering SSL key passphrases
2291 when using encrypted SSL certificate keys. If not specified
2292 keys must either be unencrypted, or Squid started with the -N
2293 option to allow it to query interactively for the passphrase.
2295 The key file name is given as argument to the program allowing
2296 selection of the right password if you have multiple encrypted
2301 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2302 -----------------------------------------------------------------------------
2305 NAME: sslcrtd_program
2308 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2309 LOC: Ssl::TheConfig.ssl_crtd
2311 Specify the location and options of the executable for ssl_crtd process.
2312 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2313 For more information use:
2314 @DEFAULT_SSL_CRTD@ -h
2317 NAME: sslcrtd_children
2318 TYPE: HelperChildConfig
2320 DEFAULT: 32 startup=5 idle=1
2321 LOC: Ssl::TheConfig.ssl_crtdChildren
2323 The maximum number of processes spawn to service ssl server.
2324 The maximum this may be safely set to is 32.
2326 The startup= and idle= options allow some measure of skew in your
2331 Sets the minimum number of processes to spawn when Squid
2332 starts or reconfigures. When set to zero the first request will
2333 cause spawning of the first child process to handle it.
2335 Starting too few children temporary slows Squid under load while it
2336 tries to spawn enough additional processes to cope with traffic.
2340 Sets a minimum of how many processes Squid is to try and keep available
2341 at all times. When traffic begins to rise above what the existing
2342 processes can handle this many more will be spawned up to the maximum
2343 configured. A minimum setting of 1 is required.
2345 You must have at least one ssl_crtd process.
2349 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2350 -----------------------------------------------------------------------------
2358 To specify other caches in a hierarchy, use the format:
2360 cache_peer hostname type http-port icp-port [options]
2365 # hostname type port port options
2366 # -------------------- -------- ----- ----- -----------
2367 cache_peer parent.foo.net parent 3128 3130 default
2368 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2369 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2370 cache_peer example.com parent 80 0 default
2371 cache_peer cdn.example.com sibling 3128 0
2373 type: either 'parent', 'sibling', or 'multicast'.
2375 proxy-port: The port number where the peer accept HTTP requests.
2376 For other Squid proxies this is usually 3128
2377 For web servers this is usually 80
2379 icp-port: Used for querying neighbor caches about objects.
2380 Set to 0 if the peer does not support ICP or HTCP.
2381 See ICP and HTCP options below for additional details.
2384 ==== ICP OPTIONS ====
2386 You MUST also set icp_port and icp_access explicitly when using these options.
2387 The defaults will prevent peer traffic using ICP.
2390 no-query Disable ICP queries to this neighbor.
2393 Indicates the named peer is a member of a multicast group.
2394 ICP queries will not be sent directly to the peer, but ICP
2395 replies will be accepted from it.
2397 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2398 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2401 To only send ICP queries to this neighbor infrequently.
2402 This is used to keep the neighbor round trip time updated
2403 and is usually used in conjunction with weighted-round-robin.
2406 ==== HTCP OPTIONS ====
2408 You MUST also set htcp_port and htcp_access explicitly when using these options.
2409 The defaults will prevent peer traffic using HTCP.
2412 htcp Send HTCP, instead of ICP, queries to the neighbor.
2413 You probably also want to set the "icp-port" to 4827
2414 instead of 3130. This directive accepts a comma separated
2415 list of options described below.
2417 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2419 htcp=no-clr Send HTCP to the neighbor but without
2420 sending any CLR requests. This cannot be used with
2423 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2424 This cannot be used with no-clr.
2427 Send HTCP to the neighbor including CLRs but only when
2428 they do not result from PURGE requests.
2431 Forward any HTCP CLR requests this proxy receives to the peer.
2434 ==== PEER SELECTION METHODS ====
2436 The default peer selection method is ICP, with the first responding peer
2437 being used as source. These options can be used for better load balancing.
2440 default This is a parent cache which can be used as a "last-resort"
2441 if a peer cannot be located by any of the peer-selection methods.
2442 If specified more than once, only the first is used.
2444 round-robin Load-Balance parents which should be used in a round-robin
2445 fashion in the absence of any ICP queries.
2446 weight=N can be used to add bias.
2448 weighted-round-robin
2449 Load-Balance parents which should be used in a round-robin
2450 fashion with the frequency of each parent being based on the
2451 round trip time. Closer parents are used more often.
2452 Usually used for background-ping parents.
2453 weight=N can be used to add bias.
2455 carp Load-Balance parents which should be used as a CARP array.
2456 The requests will be distributed among the parents based on the
2457 CARP load balancing hash function based on their weight.
2459 userhash Load-balance parents based on the client proxy_auth or ident username.
2461 sourcehash Load-balance parents based on the client source IP.
2464 To be used only for cache peers of type "multicast".
2465 ALL members of this multicast group have "sibling"
2466 relationship with it, not "parent". This is to a multicast
2467 group when the requested object would be fetched only from
2468 a "parent" cache, anyway. It's useful, e.g., when
2469 configuring a pool of redundant Squid proxies, being
2470 members of the same multicast group.
2473 ==== PEER SELECTION OPTIONS ====
2475 weight=N use to affect the selection of a peer during any weighted
2476 peer-selection mechanisms.
2477 The weight must be an integer; default is 1,
2478 larger weights are favored more.
2479 This option does not affect parent selection if a peering
2480 protocol is not in use.
2482 basetime=N Specify a base amount to be subtracted from round trip
2484 It is subtracted before division by weight in calculating
2485 which parent to fectch from. If the rtt is less than the
2486 base time the rtt is set to a minimal value.
2488 ttl=N Specify a TTL to use when sending multicast ICP queries
2490 Only useful when sending to a multicast group.
2491 Because we don't accept ICP replies from random
2492 hosts, you must configure other group members as
2493 peers with the 'multicast-responder' option.
2495 no-delay To prevent access to this neighbor from influencing the
2498 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2499 enabled) for this host from the specified URL rather
2500 than the Squid default location.
2503 ==== CARP OPTIONS ====
2505 carp-key=key-specification
2506 use a different key than the full URL to hash against the peer.
2507 the key-specification is a comma-separated list of the keywords
2508 scheme, host, port, path, params
2509 Order is not important.
2511 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2513 originserver Causes this parent to be contacted as an origin server.
2514 Meant to be used in accelerator setups when the peer
2518 Set the Host header of requests forwarded to this peer.
2519 Useful in accelerator setups where the server (peer)
2520 expects a certain domain name but clients may request
2521 others. ie example.com or www.example.com
2523 no-digest Disable request of cache digests.
2526 Disables requesting ICMP RTT database (NetDB).
2529 ==== AUTHENTICATION OPTIONS ====
2532 If this is a personal/workgroup proxy and your parent
2533 requires proxy authentication.
2535 Note: The string can include URL escapes (i.e. %20 for
2536 spaces). This also means % must be written as %%.
2539 Send login details received from client to this peer.
2540 Both Proxy- and WWW-Authorization headers are passed
2541 without alteration to the peer.
2542 Authentication is not required by Squid for this to work.
2544 Note: This will pass any form of authentication but
2545 only Basic auth will work through a proxy unless the
2546 connection-auth options are also used.
2548 login=PASS Send login details received from client to this peer.
2549 Authentication is not required by this option.
2551 If there are no client-provided authentication headers
2552 to pass on, but username and password are available
2553 from an external ACL user= and password= result tags
2554 they may be sent instead.
2556 Note: To combine this with proxy_auth both proxies must
2557 share the same user database as HTTP only allows for
2558 a single login (one for proxy, one for origin server).
2559 Also be warned this will expose your users proxy
2560 password to the peer. USE WITH CAUTION
2563 Send the username to the upstream cache, but with a
2564 fixed password. This is meant to be used when the peer
2565 is in another administrative domain, but it is still
2566 needed to identify each user.
2567 The star can optionally be followed by some extra
2568 information which is added to the username. This can
2569 be used to identify this proxy to the peer, similar to
2570 the login=username:password option above.
2573 If this is a personal/workgroup proxy and your parent
2574 requires a secure proxy authentication.
2575 The first principal from the default keytab or defined by
2576 the environment variable KRB5_KTNAME will be used.
2578 WARNING: The connection may transmit requests from multiple
2579 clients. Negotiate often assumes end-to-end authentication
2580 and a single-client. Which is not strictly true here.
2582 login=NEGOTIATE:principal_name
2583 If this is a personal/workgroup proxy and your parent
2584 requires a secure proxy authentication.
2585 The principal principal_name from the default keytab or
2586 defined by the environment variable KRB5_KTNAME will be
2589 WARNING: The connection may transmit requests from multiple
2590 clients. Negotiate often assumes end-to-end authentication
2591 and a single-client. Which is not strictly true here.
2593 connection-auth=on|off
2594 Tell Squid that this peer does or not support Microsoft
2595 connection oriented authentication, and any such
2596 challenges received from there should be ignored.
2597 Default is auto to automatically determine the status
2601 ==== SSL / HTTPS / TLS OPTIONS ====
2603 ssl Encrypt connections to this peer with SSL/TLS.
2605 sslcert=/path/to/ssl/certificate
2606 A client SSL certificate to use when connecting to
2609 sslkey=/path/to/ssl/key
2610 The private SSL key corresponding to sslcert above.
2611 If 'sslkey' is not specified 'sslcert' is assumed to
2612 reference a combined file containing both the
2613 certificate and the key.
2615 sslversion=1|2|3|4|5|6
2616 The SSL version to use when connecting to this peer
2617 1 = automatic (default)
2624 sslcipher=... The list of valid SSL ciphers to use when connecting
2627 ssloptions=... Specify various SSL implementation options:
2629 NO_SSLv2 Disallow the use of SSLv2
2630 NO_SSLv3 Disallow the use of SSLv3
2631 NO_TLSv1 Disallow the use of TLSv1.0
2632 NO_TLSv1_1 Disallow the use of TLSv1.1
2633 NO_TLSv1_2 Disallow the use of TLSv1.2
2635 Always create a new key when using
2636 temporary/ephemeral DH key exchanges
2637 ALL Enable various bug workarounds
2638 suggested as "harmless" by OpenSSL
2639 Be warned that this reduces SSL/TLS
2640 strength to some attacks.
2642 See the OpenSSL SSL_CTX_set_options documentation for a
2645 sslcafile=... A file containing additional CA certificates to use
2646 when verifying the peer certificate.
2648 sslcapath=... A directory containing additional CA certificates to
2649 use when verifying the peer certificate.
2651 sslcrlfile=... A certificate revocation list file to use when
2652 verifying the peer certificate.
2654 sslflags=... Specify various flags modifying the SSL implementation:
2657 Accept certificates even if they fail to
2660 Don't use the default CA list built in
2663 Don't verify the peer certificate
2664 matches the server name
2666 ssldomain= The peer name as advertised in it's certificate.
2667 Used for verifying the correctness of the received peer
2668 certificate. If not specified the peer hostname will be
2672 Enable the "Front-End-Https: On" header needed when
2673 using Squid as a SSL frontend in front of Microsoft OWA.
2674 See MS KB document Q307347 for details on this header.
2675 If set to auto the header will only be added if the
2676 request is forwarded as a https:// URL.
2679 ==== GENERAL OPTIONS ====
2682 A peer-specific connect timeout.
2683 Also see the peer_connect_timeout directive.
2685 connect-fail-limit=N
2686 How many times connecting to a peer must fail before
2687 it is marked as down. Default is 10.
2689 allow-miss Disable Squid's use of only-if-cached when forwarding
2690 requests to siblings. This is primarily useful when
2691 icp_hit_stale is used by the sibling. To extensive use
2692 of this option may result in forwarding loops, and you
2693 should avoid having two-way peerings with this option.
2694 For example to deny peer usage on requests from peer
2695 by denying cache_peer_access if the source is a peer.
2697 max-conn=N Limit the amount of connections Squid may open to this
2700 name=xxx Unique name for the peer.
2701 Required if you have multiple peers on the same host
2702 but different ports.
2703 This name can be used in cache_peer_access and similar
2704 directives to dentify the peer.
2705 Can be used by outgoing access controls through the
2708 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2709 requests to this peer. Use normal address selection instead.
2711 proxy-only objects fetched from the peer will not be stored locally.
2715 NAME: cache_peer_domain cache_host_domain
2720 Use to limit the domains for which a neighbor cache will be
2723 cache_peer_domain cache-host domain [domain ...]
2724 cache_peer_domain cache-host !domain
2726 For example, specifying
2728 cache_peer_domain parent.foo.net .edu
2730 has the effect such that UDP query packets are sent to
2731 'bigserver' only when the requested object exists on a
2732 server in the .edu domain. Prefixing the domainname
2733 with '!' means the cache will be queried for objects
2736 NOTE: * Any number of domains may be given for a cache-host,
2737 either on the same or separate lines.
2738 * When multiple domains are given for a particular
2739 cache-host, the first matched domain is applied.
2740 * Cache hosts with no domain restrictions are queried
2742 * There are no defaults.
2743 * There is also a 'cache_peer_access' tag in the ACL
2747 NAME: cache_peer_access
2752 Similar to 'cache_peer_domain' but provides more flexibility by
2755 cache_peer_access cache-host allow|deny [!]aclname ...
2757 The syntax is identical to 'http_access' and the other lists of
2758 ACL elements. See the comments for 'http_access' below, or
2759 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2762 NAME: neighbor_type_domain
2763 TYPE: hostdomaintype
2767 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2769 Modifying the neighbor type for specific domains is now
2770 possible. You can treat some domains differently than the
2771 default neighbor type specified on the 'cache_peer' line.
2772 Normally it should only be necessary to list domains which
2773 should be treated differently because the default neighbor type
2774 applies for hostnames which do not match domains listed here.
2777 cache_peer cache.foo.org parent 3128 3130
2778 neighbor_type_domain cache.foo.org sibling .com .net
2779 neighbor_type_domain cache.foo.org sibling .au .de
2782 NAME: dead_peer_timeout
2786 LOC: Config.Timeout.deadPeer
2788 This controls how long Squid waits to declare a peer cache
2789 as "dead." If there are no ICP replies received in this
2790 amount of time, Squid will declare the peer dead and not
2791 expect to receive any further ICP replies. However, it
2792 continues to send ICP queries, and will mark the peer as
2793 alive upon receipt of the first subsequent ICP reply.
2795 This timeout also affects when Squid expects to receive ICP
2796 replies from peers. If more than 'dead_peer' seconds have
2797 passed since the last ICP reply was received, Squid will not
2798 expect to receive an ICP reply on the next query. Thus, if
2799 your time between requests is greater than this timeout, you
2800 will see a lot of requests sent DIRECT to origin servers
2801 instead of to your parents.
2804 NAME: forward_max_tries
2807 LOC: Config.forward_max_tries
2809 Controls how many different forward paths Squid will try
2810 before giving up. See also forward_timeout.
2812 NOTE: connect_retries (default: none) can make each of these
2813 possible forwarding paths be tried multiple times.
2816 NAME: hierarchy_stoplist
2819 LOC: Config.hierarchy_stoplist
2821 A list of words which, if found in a URL, cause the object to
2822 be handled directly by this cache. In other words, use this
2823 to not query neighbor caches for certain objects. You may
2824 list this option multiple times.
2827 hierarchy_stoplist cgi-bin ?
2829 Note: never_direct overrides this option.
2833 MEMORY CACHE OPTIONS
2834 -----------------------------------------------------------------------------
2841 LOC: Config.memMaxSize
2843 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2844 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2845 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2846 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2848 'cache_mem' specifies the ideal amount of memory to be used
2850 * In-Transit objects
2852 * Negative-Cached objects
2854 Data for these objects are stored in 4 KB blocks. This
2855 parameter specifies the ideal upper limit on the total size of
2856 4 KB blocks allocated. In-Transit objects take the highest
2859 In-transit objects have priority over the others. When
2860 additional space is needed for incoming data, negative-cached
2861 and hot objects will be released. In other words, the
2862 negative-cached and hot objects will fill up any unused space
2863 not needed for in-transit objects.
2865 If circumstances require, this limit will be exceeded.
2866 Specifically, if your incoming request rate requires more than
2867 'cache_mem' of memory to hold in-transit objects, Squid will
2868 exceed this limit to satisfy the new requests. When the load
2869 decreases, blocks will be freed until the high-water mark is
2870 reached. Thereafter, blocks will be used to store hot
2873 If shared memory caching is enabled, Squid does not use the shared
2874 cache space for in-transit objects, but they still consume as much
2875 local memory as they need. For more details about the shared memory
2876 cache, see memory_cache_shared.
2879 NAME: maximum_object_size_in_memory
2883 LOC: Config.Store.maxInMemObjSize
2885 Objects greater than this size will not be attempted to kept in
2886 the memory cache. This should be set high enough to keep objects
2887 accessed frequently in memory to improve performance whilst low
2888 enough to keep larger objects from hoarding cache_mem.
2891 NAME: memory_cache_shared
2894 LOC: Config.memShared
2896 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
2898 Controls whether the memory cache is shared among SMP workers.
2900 The shared memory cache is meant to occupy cache_mem bytes and replace
2901 the non-shared memory cache, although some entities may still be
2902 cached locally by workers for now (e.g., internal and in-transit
2903 objects may be served from a local memory cache even if shared memory
2904 caching is enabled).
2906 By default, the memory cache is shared if and only if all of the
2907 following conditions are satisfied: Squid runs in SMP mode with
2908 multiple workers, cache_mem is positive, and Squid environment
2909 supports required IPC primitives (e.g., POSIX shared memory segments
2910 and GCC-style atomic operations).
2912 To avoid blocking locks, shared memory uses opportunistic algorithms
2913 that do not guarantee that every cachable entity that could have been
2914 shared among SMP workers will actually be shared.
2916 Currently, entities exceeding 32KB in size cannot be shared.
2919 NAME: memory_cache_mode
2924 Controls which objects to keep in the memory cache (cache_mem)
2926 always Keep most recently fetched objects in memory (default)
2928 disk Only disk cache hits are kept in memory, which means
2929 an object must first be cached on disk and then hit
2930 a second time before cached in memory.
2932 network Only objects fetched from network is kept in memory
2935 NAME: memory_replacement_policy
2937 LOC: Config.memPolicy
2940 The memory replacement policy parameter determines which
2941 objects are purged from memory when memory space is needed.
2943 See cache_replacement_policy for details.
2948 -----------------------------------------------------------------------------
2951 NAME: cache_replacement_policy
2953 LOC: Config.replPolicy
2956 The cache replacement policy parameter determines which
2957 objects are evicted (replaced) when disk space is needed.
2959 lru : Squid's original list based LRU policy
2960 heap GDSF : Greedy-Dual Size Frequency
2961 heap LFUDA: Least Frequently Used with Dynamic Aging
2962 heap LRU : LRU policy implemented using a heap
2964 Applies to any cache_dir lines listed below this.
2966 The LRU policies keeps recently referenced objects.
2968 The heap GDSF policy optimizes object hit rate by keeping smaller
2969 popular objects in cache so it has a better chance of getting a
2970 hit. It achieves a lower byte hit rate than LFUDA though since
2971 it evicts larger (possibly popular) objects.
2973 The heap LFUDA policy keeps popular objects in cache regardless of
2974 their size and thus optimizes byte hit rate at the expense of
2975 hit rate since one large, popular object will prevent many
2976 smaller, slightly less popular objects from being cached.
2978 Both policies utilize a dynamic aging mechanism that prevents
2979 cache pollution that can otherwise occur with frequency-based
2980 replacement policies.
2982 NOTE: if using the LFUDA replacement policy you should increase
2983 the value of maximum_object_size above its default of 4096 KB to
2984 to maximize the potential byte hit rate improvement of LFUDA.
2986 For more information about the GDSF and LFUDA cache replacement
2987 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
2988 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
2994 LOC: Config.cacheSwap
2998 cache_dir Type Directory-Name Fs-specific-data [options]
3000 You can specify multiple cache_dir lines to spread the
3001 cache among different disk partitions.
3003 Type specifies the kind of storage system to use. Only "ufs"
3004 is built by default. To enable any of the other storage systems
3005 see the --enable-storeio configure option.
3007 'Directory' is a top-level directory where cache swap
3008 files will be stored. If you want to use an entire disk
3009 for caching, this can be the mount-point directory.
3010 The directory must exist and be writable by the Squid
3011 process. Squid will NOT create this directory for you.
3013 In SMP configurations, cache_dir must not precede the workers option
3014 and should use configuration macros or conditionals to give each
3015 worker interested in disk caching a dedicated cache directory.
3019 "ufs" is the old well-known Squid storage format that has always
3022 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3024 'Mbytes' is the amount of disk space (MB) to use under this
3025 directory. The default is 100 MB. Change this to suit your
3026 configuration. Do NOT put the size of your disk drive here.
3027 Instead, if you want Squid to use the entire disk drive,
3028 subtract 20% and use that value.
3030 'L1' is the number of first-level subdirectories which
3031 will be created under the 'Directory'. The default is 16.
3033 'L2' is the number of second-level subdirectories which
3034 will be created under each first-level directory. The default
3037 The aufs store type:
3039 "aufs" uses the same storage format as "ufs", utilizing
3040 POSIX-threads to avoid blocking the main Squid process on
3041 disk-I/O. This was formerly known in Squid as async-io.
3043 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3045 see argument descriptions under ufs above
3047 The diskd store type:
3049 "diskd" uses the same storage format as "ufs", utilizing a
3050 separate process to avoid blocking the main Squid process on
3053 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3055 see argument descriptions under ufs above
3057 Q1 specifies the number of unacknowledged I/O requests when Squid
3058 stops opening new files. If this many messages are in the queues,
3059 Squid won't open new files. Default is 64
3061 Q2 specifies the number of unacknowledged messages when Squid
3062 starts blocking. If this many messages are in the queues,
3063 Squid blocks until it receives some replies. Default is 72
3065 When Q1 < Q2 (the default), the cache directory is optimized
3066 for lower response time at the expense of a decrease in hit
3067 ratio. If Q1 > Q2, the cache directory is optimized for
3068 higher hit ratio at the expense of an increase in response
3071 The rock store type:
3073 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3075 The Rock Store type is a database-style storage. All cached
3076 entries are stored in a "database" file, using fixed-size slots,
3077 one entry per slot. The database size is specified in MB. The
3078 slot size is specified in bytes using the max-size option. See
3079 below for more info on the max-size option.
3081 swap-timeout=msec: Squid will not start writing a miss to or
3082 reading a hit from disk if it estimates that the swap operation
3083 will take more than the specified number of milliseconds. By
3084 default and when set to zero, disables the disk I/O time limit
3085 enforcement. Ignored when using blocking I/O module because
3086 blocking synchronous I/O does not allow Squid to estimate the
3087 expected swap wait time.
3089 max-swap-rate=swaps/sec: Artificially limits disk access using
3090 the specified I/O rate limit. Swap out requests that
3091 would cause the average I/O rate to exceed the limit are
3092 delayed. Individual swap in requests (i.e., hits or reads) are
3093 not delayed, but they do contribute to measured swap rate and
3094 since they are placed in the same FIFO queue as swap out
3095 requests, they may wait longer if max-swap-rate is smaller.
3096 This is necessary on file systems that buffer "too
3097 many" writes and then start blocking Squid and other processes
3098 while committing those writes to disk. Usually used together
3099 with swap-timeout to avoid excessive delays and queue overflows
3100 when disk demand exceeds available disk "bandwidth". By default
3101 and when set to zero, disables the disk I/O rate limit
3102 enforcement. Currently supported by IpcIo module only.
3105 The coss store type:
3107 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3108 production use and has thus been removed from this release.
3109 We hope that it can be made usable again soon.
3111 block-size=n defines the "block size" for COSS cache_dir's.
3112 Squid uses file numbers as block numbers. Since file numbers
3113 are limited to 24 bits, the block size determines the maximum
3114 size of the COSS partition. The default is 512 bytes, which
3115 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3116 you should not change the coss block size after Squid
3117 has written some objects to the cache_dir.
3119 The coss file store has changed from 2.5. Now it uses a file
3120 called 'stripe' in the directory names in the config - and
3121 this will be created by squid -z.
3125 no-store, no new objects should be stored to this cache_dir
3127 min-size=n, refers to the min object size in bytes this cache_dir
3128 will accept. It's used to restrict a cache_dir to only store
3129 large objects (e.g. aufs) while other storedirs are optimized
3130 for smaller objects (e.g. COSS). Defaults to 0.
3132 max-size=n, refers to the max object size in bytes this cache_dir
3133 supports. It is used to select the cache_dir to store the object.
3134 Note: To make optimal use of the max-size limits you should order
3135 the cache_dir lines with the smallest max-size value first and the
3136 ones with no max-size specification last.
3138 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3139 which can be changed with the --with-coss-membuf-size=N configure
3143 # Uncomment and adjust the following to add a disk cache directory.
3144 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3148 NAME: store_dir_select_algorithm
3150 LOC: Config.store_dir_select_algorithm
3153 Set this to 'round-robin' as an alternative.
3156 NAME: max_open_disk_fds
3158 LOC: Config.max_open_disk_fds
3161 To avoid having disk as the I/O bottleneck Squid can optionally
3162 bypass the on-disk cache if more than this amount of disk file
3163 descriptors are open.
3165 A value of 0 indicates no limit.
3168 NAME: minimum_object_size
3172 LOC: Config.Store.minObjectSize
3174 Objects smaller than this size will NOT be saved on disk. The
3175 value is specified in kilobytes, and the default is 0 KB, which
3176 means there is no minimum.
3179 NAME: maximum_object_size
3183 LOC: Config.Store.maxObjectSize
3185 Objects larger than this size will NOT be saved on disk. The
3186 value is specified in kilobytes, and the default is 4MB. If
3187 you wish to get a high BYTES hit ratio, you should probably
3188 increase this (one 32 MB object hit counts for 3200 10KB
3189 hits). If you wish to increase speed more than your want to
3190 save bandwidth you should leave this low.
3192 NOTE: if using the LFUDA replacement policy you should increase
3193 this value to maximize the byte hit rate improvement of LFUDA!
3194 See replacement_policy below for a discussion of this policy.
3197 NAME: cache_swap_low
3198 COMMENT: (percent, 0-100)
3201 LOC: Config.Swap.lowWaterMark
3204 NAME: cache_swap_high
3205 COMMENT: (percent, 0-100)
3208 LOC: Config.Swap.highWaterMark
3211 The low- and high-water marks for cache object replacement.
3212 Replacement begins when the swap (disk) usage is above the
3213 low-water mark and attempts to maintain utilization near the
3214 low-water mark. As swap utilization gets close to high-water
3215 mark object eviction becomes more aggressive. If utilization is
3216 close to the low-water mark less replacement is done each time.
3218 Defaults are 90% and 95%. If you have a large cache, 5% could be
3219 hundreds of MB. If this is the case you may wish to set these
3220 numbers closer together.
3225 -----------------------------------------------------------------------------
3235 logformat <name> <format specification>
3237 Defines an access log format.
3239 The <format specification> is a string with embedded % format codes
3241 % format codes all follow the same basic structure where all but
3242 the formatcode is optional. Output strings are automatically escaped
3243 as required according to their context and the output format
3244 modifiers are usually not needed, but can be specified if an explicit
3245 output format is desired.
3247 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3249 " output in quoted string format
3250 [ output in squid text log format as used by log_mime_hdrs
3251 # output in URL quoted format
3256 width minimum and/or maximum field width:
3257 [width_min][.width_max]
3258 When minimum starts with 0, the field is zero-padded.
3259 String values exceeding maximum width are truncated.
3261 {arg} argument such as header name etc
3265 % a literal % character
3266 sn Unique sequence number per log line entry
3267 err_code The ID of an error response served by Squid or
3268 a similar internal error identifier.
3269 err_detail Additional err_code-dependent error information.
3271 Connection related format codes:
3273 >a Client source IP address
3275 >p Client source port
3276 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3277 >la Local IP address the client connected to
3278 >lp Local port number the client connected to
3280 la Local listening IP address the client connection was connected to.
3281 lp Local listening port number the client connection was connected to.
3283 <a Server IP address of the last server or peer connection
3284 <A Server FQDN or peer name
3285 <p Server port number of the last server or peer connection
3286 <la Local IP address of the last server or peer connection
3287 <lp Local port number of the last server or peer connection
3289 Time related format codes:
3291 ts Seconds since epoch
3292 tu subsecond time (milliseconds)
3293 tl Local time. Optional strftime format argument
3294 default %d/%b/%Y:%H:%M:%S %z
3295 tg GMT time. Optional strftime format argument
3296 default %d/%b/%Y:%H:%M:%S %z
3297 tr Response time (milliseconds)
3298 dt Total time spent making DNS lookups (milliseconds)
3300 Access Control related format codes:
3302 et Tag returned by external acl
3303 ea Log string returned by external acl
3304 un User name (any available)
3305 ul User name from authentication
3306 ue User name from external acl helper
3307 ui User name from ident
3308 us User name from SSL
3310 HTTP related format codes:
3312 [http::]>h Original request header. Optional header name argument
3313 on the format header[:[separator]element]
3314 [http::]>ha The HTTP request headers after adaptation and redirection.
3315 Optional header name argument as for >h
3316 [http::]<h Reply header. Optional header name argument
3318 [http::]>Hs HTTP status code sent to the client
3319 [http::]<Hs HTTP status code received from the next hop
3320 [http::]<bs Number of HTTP-equivalent message body bytes
3321 received from the next hop, excluding chunked
3322 transfer encoding and control messages.
3323 Generated FTP/Gopher listings are treated as
3325 [http::]mt MIME content type
3326 [http::]rm Request method (GET/POST etc)
3327 [http::]>rm Request method from client
3328 [http::]<rm Request method sent to server or peer
3329 [http::]ru Request URL from client (historic, filtered for logging)
3330 [http::]>ru Request URL from client
3331 [http::]<ru Request URL sent to server or peer
3332 [http::]rp Request URL-Path excluding hostname
3333 [http::]>rp Request URL-Path excluding hostname from client
3334 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3335 [http::]rv Request protocol version
3336 [http::]>rv Request protocol version from client
3337 [http::]<rv Request protocol version sent to server or peer
3338 [http::]<st Sent reply size including HTTP headers
3339 [http::]>st Received request size including HTTP headers. In the
3340 case of chunked requests the chunked encoding metadata
3342 [http::]>sh Received HTTP request headers size
3343 [http::]<sh Sent HTTP reply headers size
3344 [http::]st Request+Reply size including HTTP headers
3345 [http::]<sH Reply high offset sent
3346 [http::]<sS Upstream object size
3347 [http::]<pt Peer response time in milliseconds. The timer starts
3348 when the last request byte is sent to the next hop
3349 and stops when the last response byte is received.
3350 [http::]<tt Total server-side time in milliseconds. The timer
3351 starts with the first connect request (or write I/O)
3352 sent to the first selected peer. The timer stops
3353 with the last I/O with the last peer.
3355 Squid handling related format codes:
3357 Ss Squid request status (TCP_MISS etc)
3358 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3360 SSL-related format codes:
3362 ssl::bump_mode SslBump decision for the transaction:
3364 For CONNECT requests that initiated bumping of
3365 a connection and for any request received on
3366 an already bumped connection, Squid logs the
3367 corresponding SslBump mode ("server-first" or
3368 "client-first"). See the ssl_bump option for
3369 more information about these modes.
3371 A "none" token is logged for requests that
3372 triggered "ssl_bump" ACL evaluation matching
3373 either a "none" rule or no rules at all.
3375 In all other cases, a single dash ("-") is
3378 If ICAP is enabled, the following code becomes available (as
3379 well as ICAP log codes documented with the icap_log option):
3381 icap::tt Total ICAP processing time for the HTTP
3382 transaction. The timer ticks when ICAP
3383 ACLs are checked and when ICAP
3384 transaction is in progress.
3386 If adaptation is enabled the following three codes become available:
3388 adapt::<last_h The header of the last ICAP response or
3389 meta-information from the last eCAP
3390 transaction related to the HTTP transaction.
3391 Like <h, accepts an optional header name
3394 adapt::sum_trs Summed adaptation transaction response
3395 times recorded as a comma-separated list in
3396 the order of transaction start time. Each time
3397 value is recorded as an integer number,
3398 representing response time of one or more
3399 adaptation (ICAP or eCAP) transaction in
3400 milliseconds. When a failed transaction is
3401 being retried or repeated, its time is not
3402 logged individually but added to the
3403 replacement (next) transaction. See also:
3406 adapt::all_trs All adaptation transaction response times.
3407 Same as adaptation_strs but response times of
3408 individual transactions are never added
3409 together. Instead, all transaction response
3410 times are recorded individually.
3412 You can prefix adapt::*_trs format codes with adaptation
3413 service name in curly braces to record response time(s) specific
3414 to that service. For example: %{my_service}adapt::sum_trs
3416 If SSL is enabled, the following formating codes become available:
3418 %ssl::>cert_subject The Subject field of the received client
3419 SSL certificate or a dash ('-') if Squid has
3420 received an invalid/malformed certificate or
3421 no certificate at all. Consider encoding the
3422 logged value because Subject often has spaces.
3424 %ssl::>cert_issuer The Issuer field of the received client
3425 SSL certificate or a dash ('-') if Squid has
3426 received an invalid/malformed certificate or
3427 no certificate at all. Consider encoding the
3428 logged value because Issuer often has spaces.
3430 The default formats available (which do not need re-defining) are:
3432 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3433 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3434 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3435 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3436 logformat useragent %>a [%tl] "%{User-Agent}>h"
3438 NOTE: When the log_mime_hdrs directive is set to ON.
3439 The squid, common and combined formats have a safely encoded copy
3440 of the mime headers appended to each line within a pair of brackets.
3442 NOTE: The common and combined formats are not quite true to the Apache definition.
3443 The logs from Squid contain an extra status and hierarchy code appended.
3447 NAME: access_log cache_access_log
3449 LOC: Config.Log.accesslogs
3450 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3452 These files log client request activities. Has a line every HTTP or
3453 ICP request. The format is:
3454 access_log <module>:<place> [<logformat name> [acl acl ...]]
3455 access_log none [acl acl ...]]
3457 Will log to the specified module:place using the specified format (which
3458 must be defined in a logformat directive) those entries which match
3459 ALL the acl's specified (which must be defined in acl clauses).
3460 If no acl is specified, all requests will be logged to this destination.
3462 ===== Modules Currently available =====
3464 none Do not log any requests matching these ACL.
3465 Do not specify Place or logformat name.
3467 stdio Write each log line to disk immediately at the completion of
3469 Place: the filename and path to be written.
3471 daemon Very similar to stdio. But instead of writing to disk the log
3472 line is passed to a daemon helper for asychronous handling instead.
3473 Place: varies depending on the daemon.
3475 log_file_daemon Place: the file name and path to be written.
3477 syslog To log each request via syslog facility.
3478 Place: The syslog facility and priority level for these entries.
3479 Place Format: facility.priority
3481 where facility could be any of:
3482 authpriv, daemon, local0 ... local7 or user.
3484 And priority could be any of:
3485 err, warning, notice, info, debug.
3487 udp To send each log line as text data to a UDP receiver.
3488 Place: The destination host name or IP and port.
3489 Place Format: //host:port
3491 tcp To send each log line as text data to a TCP receiver.
3492 Place: The destination host name or IP and port.
3493 Place Format: //host:port
3496 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3502 LOC: Config.Log.icaplogs
3505 ICAP log files record ICAP transaction summaries, one line per
3508 The icap_log option format is:
3509 icap_log <filepath> [<logformat name> [acl acl ...]]
3510 icap_log none [acl acl ...]]
3512 Please see access_log option documentation for details. The two
3513 kinds of logs share the overall configuration approach and many
3516 ICAP processing of a single HTTP message or transaction may
3517 require multiple ICAP transactions. In such cases, multiple
3518 ICAP transaction log lines will correspond to a single access
3521 ICAP log uses logformat codes that make sense for an ICAP
3522 transaction. Header-related codes are applied to the HTTP header
3523 embedded in an ICAP server response, with the following caveats:
3524 For REQMOD, there is no HTTP response header unless the ICAP
3525 server performed request satisfaction. For RESPMOD, the HTTP
3526 request header is the header sent to the ICAP server. For
3527 OPTIONS, there are no HTTP headers.
3529 The following format codes are also available for ICAP logs:
3531 icap::<A ICAP server IP address. Similar to <A.
3533 icap::<service_name ICAP service name from the icap_service
3534 option in Squid configuration file.
3536 icap::ru ICAP Request-URI. Similar to ru.
3538 icap::rm ICAP request method (REQMOD, RESPMOD, or
3539 OPTIONS). Similar to existing rm.
3541 icap::>st Bytes sent to the ICAP server (TCP payload
3542 only; i.e., what Squid writes to the socket).
3544 icap::<st Bytes received from the ICAP server (TCP
3545 payload only; i.e., what Squid reads from
3548 icap::<bs Number of message body bytes received from the
3549 ICAP server. ICAP message body, if any, usually
3550 includes encapsulated HTTP message headers and
3551 possibly encapsulated HTTP message body. The
3552 HTTP body part is dechunked before its size is
3555 icap::tr Transaction response time (in
3556 milliseconds). The timer starts when
3557 the ICAP transaction is created and
3558 stops when the transaction is completed.
3561 icap::tio Transaction I/O time (in milliseconds). The
3562 timer starts when the first ICAP request
3563 byte is scheduled for sending. The timers
3564 stops when the last byte of the ICAP response
3567 icap::to Transaction outcome: ICAP_ERR* for all
3568 transaction errors, ICAP_OPT for OPTION
3569 transactions, ICAP_ECHO for 204
3570 responses, ICAP_MOD for message
3571 modification, and ICAP_SAT for request
3572 satisfaction. Similar to Ss.
3574 icap::Hs ICAP response status code. Similar to Hs.
3576 icap::>h ICAP request header(s). Similar to >h.
3578 icap::<h ICAP response header(s). Similar to <h.
3580 The default ICAP log format, which can be used without an explicit
3581 definition, is called icap_squid:
3583 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3585 See also: logformat, log_icap, and %adapt::<last_h
3588 NAME: logfile_daemon
3590 DEFAULT: @DEFAULT_LOGFILED@
3591 LOC: Log::TheConfig.logfile_daemon
3593 Specify the path to the logfile-writing daemon. This daemon is
3594 used to write the access and store logs, if configured.
3596 Squid sends a number of commands to the log daemon:
3597 L<data>\n - logfile data
3602 r<n>\n - set rotate count to <n>
3603 b<n>\n - 1 = buffer output, 0 = don't buffer output
3605 No responses is expected.
3610 LOC: Config.accessList.log
3612 COMMENT: allow|deny acl acl...
3614 This options allows you to control which requests gets logged
3615 to access.log (see access_log directive). Requests denied for
3616 logging will also not be accounted for in performance counters.
3618 This clause only supports fast acl types.
3619 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3625 LOC: Config.accessList.icap
3628 This options allows you to control which requests get logged
3629 to icap.log. See the icap_log directive for ICAP log details.
3632 NAME: cache_store_log
3635 LOC: Config.Log.store
3637 Logs the activities of the storage manager. Shows which
3638 objects are ejected from the cache, and which objects are
3639 saved and for how long. To disable, enter "none" or remove the line.
3640 There are not really utilities to analyze this data, so you can safely
3644 cache_store_log @DEFAULT_STORE_LOG@
3647 NAME: cache_swap_state cache_swap_log
3649 LOC: Config.Log.swap
3652 Location for the cache "swap.state" file. This index file holds
3653 the metadata of objects saved on disk. It is used to rebuild
3654 the cache during startup. Normally this file resides in each
3655 'cache_dir' directory, but you may specify an alternate
3656 pathname here. Note you must give a full filename, not just
3657 a directory. Since this is the index for the whole object
3658 list you CANNOT periodically rotate it!
3660 If %s can be used in the file name it will be replaced with a
3661 a representation of the cache_dir name where each / is replaced
3662 with '.'. This is needed to allow adding/removing cache_dir
3663 lines when cache_swap_log is being used.
3665 If have more than one 'cache_dir', and %s is not used in the name
3666 these swap logs will have names such as:
3672 The numbered extension (which is added automatically)
3673 corresponds to the order of the 'cache_dir' lines in this
3674 configuration file. If you change the order of the 'cache_dir'
3675 lines in this file, these index files will NOT correspond to
3676 the correct 'cache_dir' entry (unless you manually rename
3677 them). We recommend you do NOT use this option. It is
3678 better to keep these index files in each 'cache_dir' directory.
3681 NAME: logfile_rotate
3684 LOC: Config.Log.rotateNumber
3686 Specifies the number of logfile rotations to make when you
3687 type 'squid -k rotate'. The default is 10, which will rotate
3688 with extensions 0 through 9. Setting logfile_rotate to 0 will
3689 disable the file name rotation, but the logfiles are still closed
3690 and re-opened. This will enable you to rename the logfiles
3691 yourself just before sending the rotate signal.
3693 Note, the 'squid -k rotate' command normally sends a USR1
3694 signal to the running squid process. In certain situations
3695 (e.g. on Linux with Async I/O), USR1 is used for other
3696 purposes, so -k rotate uses another signal. It is best to get
3697 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3700 Note, from Squid-3.1 this option has no effect on the cache.log,
3701 that log can be rotated separately by using debug_options
3704 NAME: emulate_httpd_log
3707 Replace this with an access_log directive using the format 'common' or 'combined'.
3710 NAME: log_ip_on_direct
3713 Remove this option from your config. To log server or peer names use %<A in the log format.
3718 DEFAULT: @DEFAULT_MIME_TABLE@
3719 LOC: Config.mimeTablePathname
3721 Pathname to Squid's MIME table. You shouldn't need to change
3722 this, but the default file contains examples and formatting
3723 information if you do.
3729 LOC: Config.onoff.log_mime_hdrs
3732 The Cache can record both the request and the response MIME
3733 headers for each HTTP transaction. The headers are encoded
3734 safely and will appear as two bracketed fields at the end of
3735 the access log (for either the native or httpd-emulated log
3736 formats). To enable this logging set log_mime_hdrs to 'on'.
3742 Replace this with an access_log directive using the format 'useragent'.
3745 NAME: referer_log referrer_log
3748 Replace this with an access_log directive using the format 'referrer'.
3753 DEFAULT: @DEFAULT_PID_FILE@
3754 LOC: Config.pidFilename
3756 A filename to write the process-id to. To disable, enter "none".
3762 Remove this option from your config. To log FQDN use %>A in the log format.
3765 NAME: client_netmask
3767 LOC: Config.Addrs.client_netmask
3770 A netmask for client addresses in logfiles and cachemgr output.
3771 Change this to protect the privacy of your cache clients.
3772 A netmask of 255.255.255.0 will log all IP's in that range with
3773 the last digit set to '0'.
3779 Use a regular access.log with ACL limiting it to MISS events.
3782 NAME: strip_query_terms
3784 LOC: Config.onoff.strip_query_terms
3787 By default, Squid strips query terms from requested URLs before
3788 logging. This protects your user's privacy.
3795 LOC: Config.onoff.buffered_logs
3797 cache.log log file is written with stdio functions, and as such
3798 it can be buffered or unbuffered. By default it will be unbuffered.
3799 Buffering it can speed up the writing slightly (though you are
3800 unlikely to need to worry unless you run with tons of debugging
3801 enabled in which case performance will suffer badly anyway..).
3804 NAME: netdb_filename
3806 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3807 LOC: Config.netdbFilename
3810 A filename where Squid stores it's netdb state between restarts.
3811 To disable, enter "none".
3815 OPTIONS FOR TROUBLESHOOTING
3816 -----------------------------------------------------------------------------
3821 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3822 LOC: Debug::cache_log
3824 Cache logging file. This is where general information about
3825 your cache's behavior goes. You can increase the amount of data
3826 logged to this file and how often its rotated with "debug_options"
3832 LOC: Debug::debugOptions
3834 Logging options are set as section,level where each source file
3835 is assigned a unique section. Lower levels result in less
3836 output, Full debugging (level 9) can result in a very large
3837 log file, so be careful.
3839 The magic word "ALL" sets debugging levels for all sections.
3840 We recommend normally running with "ALL,1".
3842 The rotate=N option can be used to keep more or less of these logs
3843 than would otherwise be kept by logfile_rotate.
3844 For most uses a single log should be enough to monitor current
3845 events affecting Squid.
3850 LOC: Config.coredump_dir
3851 DEFAULT_IF_NONE: none
3853 By default Squid leaves core files in the directory from where
3854 it was started. If you set 'coredump_dir' to a directory
3855 that exists, Squid will chdir() to that directory at startup
3856 and coredump files will be left there.
3860 # Leave coredumps in the first cache dir
3861 coredump_dir @DEFAULT_SWAP_DIR@
3867 OPTIONS FOR FTP GATEWAYING
3868 -----------------------------------------------------------------------------
3874 LOC: Config.Ftp.anon_user
3876 If you want the anonymous login password to be more informative
3877 (and enable the use of picky ftp servers), set this to something
3878 reasonable for your domain, like wwwuser@somewhere.net
3880 The reason why this is domainless by default is the
3881 request can be made on the behalf of a user in any domain,
3882 depending on how the cache is used.
3883 Some ftp server also validate the email address is valid
3884 (for example perl.com).
3890 LOC: Config.Ftp.passive
3892 If your firewall does not allow Squid to use passive
3893 connections, turn off this option.
3895 Use of ftp_epsv_all option requires this to be ON.
3901 LOC: Config.Ftp.epsv_all
3903 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
3905 NATs may be able to put the connection on a "fast path" through the
3906 translator, as the EPRT command will never be used and therefore,
3907 translation of the data portion of the segments will never be needed.
3909 When a client only expects to do two-way FTP transfers this may be
3911 If squid finds that it must do a three-way FTP transfer after issuing
3912 an EPSV ALL command, the FTP session will fail.
3914 If you have any doubts about this option do not use it.
3915 Squid will nicely attempt all other connection methods.
3917 Requires ftp_passive to be ON (default) for any effect.
3923 LOC: Config.Ftp.epsv
3925 FTP Protocol extensions permit the use of a special "EPSV" command.
3927 NATs may be able to put the connection on a "fast path" through the
3928 translator using EPSV, as the EPRT command will never be used
3929 and therefore, translation of the data portion of the segments
3930 will never be needed.
3932 Turning this OFF will prevent EPSV being attempted.
3933 WARNING: Doing so will convert Squid back to the old behavior with all
3934 the related problems with external NAT devices/layers.
3936 Requires ftp_passive to be ON (default) for any effect.
3942 LOC: Config.Ftp.eprt
3944 FTP Protocol extensions permit the use of a special "EPRT" command.
3946 This extension provides a protocol neutral alternative to the
3947 IPv4-only PORT command. When supported it enables active FTP data
3948 channels over IPv6 and efficient NAT handling.
3950 Turning this OFF will prevent EPRT being attempted and will skip
3951 straight to using PORT for IPv4 servers.
3953 Some devices are known to not handle this extension correctly and
3954 may result in crashes. Devices which suport EPRT enough to fail
3955 cleanly will result in Squid attempting PORT anyway. This directive
3956 should only be disabled when EPRT results in device failures.
3958 WARNING: Doing so will convert Squid back to the old behavior with all
3959 the related problems with external NAT devices/layers and IPv4-only FTP.
3962 NAME: ftp_sanitycheck
3965 LOC: Config.Ftp.sanitycheck
3967 For security and data integrity reasons Squid by default performs
3968 sanity checks of the addresses of FTP data connections ensure the
3969 data connection is to the requested server. If you need to allow
3970 FTP connections to servers using another IP address for the data
3971 connection turn this off.
3974 NAME: ftp_telnet_protocol
3977 LOC: Config.Ftp.telnet
3979 The FTP protocol is officially defined to use the telnet protocol
3980 as transport channel for the control connection. However, many
3981 implementations are broken and does not respect this aspect of
3984 If you have trouble accessing files with ASCII code 255 in the
3985 path or similar problems involving this ASCII code you can
3986 try setting this directive to off. If that helps, report to the
3987 operator of the FTP server in question that their FTP server
3988 is broken and does not follow the FTP standard.
3992 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
3993 -----------------------------------------------------------------------------
3998 DEFAULT: @DEFAULT_DISKD@
3999 LOC: Config.Program.diskd
4001 Specify the location of the diskd executable.
4002 Note this is only useful if you have compiled in
4003 diskd as one of the store io modules.
4006 NAME: unlinkd_program
4009 DEFAULT: @DEFAULT_UNLINKD@
4010 LOC: Config.Program.unlinkd
4012 Specify the location of the executable for file deletion process.
4015 NAME: pinger_program
4017 DEFAULT: @DEFAULT_PINGER@
4018 LOC: Config.pinger.program
4021 Specify the location of the executable for the pinger process.
4027 LOC: Config.pinger.enable
4030 Control whether the pinger is active at run-time.
4031 Enables turning ICMP pinger on and off with a simple
4032 squid -k reconfigure.
4037 OPTIONS FOR URL REWRITING
4038 -----------------------------------------------------------------------------
4041 NAME: url_rewrite_program redirect_program
4043 LOC: Config.Program.redirect
4046 Specify the location of the executable URL rewriter to use.
4047 Since they can perform almost any function there isn't one included.
4049 For each requested URL, the rewriter will receive on line with the format
4051 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
4053 In the future, the rewriter interface will be extended with
4054 key=value pairs ("kvpairs" shown above). Rewriter programs
4055 should be prepared to receive and possibly ignore additional
4056 whitespace-separated tokens on each input line.
4058 And the rewriter may return a rewritten URL. The other components of
4059 the request line does not need to be returned (ignored if they are).
4061 The rewriter can also indicate that a client-side redirect should
4062 be performed to the new URL. This is done by prefixing the returned
4063 URL with "301:" (moved permanently) or 302: (moved temporarily), etc.
4065 By default, a URL rewriter is not used.
4068 NAME: url_rewrite_children redirect_children
4069 TYPE: HelperChildConfig
4070 DEFAULT: 20 startup=0 idle=1 concurrency=0
4071 LOC: Config.redirectChildren
4073 The maximum number of redirector processes to spawn. If you limit
4074 it too few Squid will have to wait for them to process a backlog of
4075 URLs, slowing it down. If you allow too many they will use RAM
4076 and other system resources noticably.
4078 The startup= and idle= options allow some measure of skew in your
4083 Sets a minimum of how many processes are to be spawned when Squid
4084 starts or reconfigures. When set to zero the first request will
4085 cause spawning of the first child process to handle it.
4087 Starting too few will cause an initial slowdown in traffic as Squid
4088 attempts to simultaneously spawn enough processes to cope.
4092 Sets a minimum of how many processes Squid is to try and keep available
4093 at all times. When traffic begins to rise above what the existing
4094 processes can handle this many more will be spawned up to the maximum
4095 configured. A minimum setting of 1 is required.
4099 The number of requests each redirector helper can handle in
4100 parallel. Defaults to 0 which indicates the redirector
4101 is a old-style single threaded redirector.
4103 When this directive is set to a value >= 1 then the protocol
4104 used to communicate with the helper is modified to include
4105 a request ID in front of the request/response. The request
4106 ID from the request must be echoed back with the response
4110 NAME: url_rewrite_host_header redirect_rewrites_host_header
4113 LOC: Config.onoff.redir_rewrites_host
4115 To preserve same-origin security policies in browsers and
4116 prevent Host: header forgery by redirectors Squid rewrites
4117 any Host: header in redirected requests.
4119 If you are running an accelerator this may not be a wanted
4120 effect of a redirector. This directive enables you disable
4121 Host: alteration in reverse-proxy traffic.
4123 WARNING: Entries are cached on the result of the URL rewriting
4124 process, so be careful if you have domain-virtual hosts.
4126 WARNING: Squid and other software verifies the URL and Host
4127 are matching, so be careful not to relay through other proxies
4128 or inspecting firewalls with this disabled.
4131 NAME: url_rewrite_access redirector_access
4134 LOC: Config.accessList.redirector
4136 If defined, this access list specifies which requests are
4137 sent to the redirector processes. By default all requests
4140 This clause supports both fast and slow acl types.
4141 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4144 NAME: url_rewrite_bypass redirector_bypass
4146 LOC: Config.onoff.redirector_bypass
4149 When this is 'on', a request will not go through the
4150 redirector if all redirectors are busy. If this is 'off'
4151 and the redirector queue grows too large, Squid will exit
4152 with a FATAL error and ask you to increase the number of
4153 redirectors. You should only enable this if the redirectors
4154 are not critical to your caching system. If you use
4155 redirectors for access control, and you enable this option,
4156 users may have access to pages they should not
4157 be allowed to request.
4161 OPTIONS FOR TUNING THE CACHE
4162 -----------------------------------------------------------------------------
4165 NAME: cache no_cache
4168 LOC: Config.accessList.noCache
4170 A list of ACL elements which, if matched and denied, cause the request to
4171 not be satisfied from the cache and the reply to not be cached.
4172 In other words, use this to force certain objects to never be cached.
4174 You must use the words 'allow' or 'deny' to indicate whether items
4175 matching the ACL should be allowed or denied into the cache.
4177 Default is to allow all to be cached.
4179 This clause supports both fast and slow acl types.
4180 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4186 LOC: Config.maxStale
4189 This option puts an upper limit on how stale content Squid
4190 will serve from the cache if cache validation fails.
4191 Can be overriden by the refresh_pattern max-stale option.
4194 NAME: refresh_pattern
4195 TYPE: refreshpattern
4199 usage: refresh_pattern [-i] regex min percent max [options]
4201 By default, regular expressions are CASE-SENSITIVE. To make
4202 them case-insensitive, use the -i option.
4204 'Min' is the time (in minutes) an object without an explicit
4205 expiry time should be considered fresh. The recommended
4206 value is 0, any higher values may cause dynamic applications
4207 to be erroneously cached unless the application designer
4208 has taken the appropriate actions.
4210 'Percent' is a percentage of the objects age (time since last
4211 modification age) an object without explicit expiry time
4212 will be considered fresh.
4214 'Max' is an upper limit on how long objects without an explicit
4215 expiry time will be considered fresh.
4217 options: override-expire
4223 ignore-must-revalidate
4230 override-expire enforces min age even if the server
4231 sent an explicit expiry time (e.g., with the
4232 Expires: header or Cache-Control: max-age). Doing this
4233 VIOLATES the HTTP standard. Enabling this feature
4234 could make you liable for problems which it causes.
4236 Note: override-expire does not enforce staleness - it only extends
4237 freshness / min. If the server returns a Expires time which
4238 is longer than your max time, Squid will still consider
4239 the object fresh for that period of time.
4241 override-lastmod enforces min age even on objects
4242 that were modified recently.
4244 reload-into-ims changes client no-cache or ``reload''
4245 to If-Modified-Since requests. Doing this VIOLATES the
4246 HTTP standard. Enabling this feature could make you
4247 liable for problems which it causes.
4249 ignore-reload ignores a client no-cache or ``reload''
4250 header. Doing this VIOLATES the HTTP standard. Enabling
4251 this feature could make you liable for problems which
4254 ignore-no-cache ignores any ``Pragma: no-cache'' and
4255 ``Cache-control: no-cache'' headers received from a server.
4256 The HTTP RFC never allows the use of this (Pragma) header
4257 from a server, only a client, though plenty of servers
4260 ignore-no-store ignores any ``Cache-control: no-store''
4261 headers received from a server. Doing this VIOLATES
4262 the HTTP standard. Enabling this feature could make you
4263 liable for problems which it causes.
4265 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4266 headers received from a server. Doing this VIOLATES
4267 the HTTP standard. Enabling this feature could make you
4268 liable for problems which it causes.
4270 ignore-private ignores any ``Cache-control: private''
4271 headers received from a server. Doing this VIOLATES
4272 the HTTP standard. Enabling this feature could make you
4273 liable for problems which it causes.
4275 ignore-auth caches responses to requests with authorization,
4276 as if the originserver had sent ``Cache-control: public''
4277 in the response header. Doing this VIOLATES the HTTP standard.
4278 Enabling this feature could make you liable for problems which
4281 refresh-ims causes squid to contact the origin server
4282 when a client issues an If-Modified-Since request. This
4283 ensures that the client will receive an updated version
4284 if one is available.
4286 store-stale stores responses even if they don't have explicit
4287 freshness or a validator (i.e., Last-Modified or an ETag)
4288 present, or if they're already stale. By default, Squid will
4289 not cache such responses because they usually can't be
4290 reused. Note that such responses will be stale by default.
4292 max-stale=NN provide a maximum staleness factor. Squid won't
4293 serve objects more stale than this even if it failed to
4294 validate the object. Default: use the max_stale global limit.
4296 Basically a cached object is:
4298 FRESH if expires < now, else STALE
4300 FRESH if lm-factor < percent, else STALE
4304 The refresh_pattern lines are checked in the order listed here.
4305 The first entry which matches is used. If none of the entries
4306 match the default will be used.
4308 Note, you must uncomment all the default lines if you want
4309 to change one. The default setting is only active if none is
4314 # Add any of your own refresh_pattern entries above these.
4315 refresh_pattern ^ftp: 1440 20% 10080
4316 refresh_pattern ^gopher: 1440 0% 1440
4317 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4318 refresh_pattern . 0 20% 4320
4322 NAME: quick_abort_min
4326 LOC: Config.quickAbort.min
4329 NAME: quick_abort_max
4333 LOC: Config.quickAbort.max
4336 NAME: quick_abort_pct
4340 LOC: Config.quickAbort.pct
4342 The cache by default continues downloading aborted requests
4343 which are almost completed (less than 16 KB remaining). This
4344 may be undesirable on slow (e.g. SLIP) links and/or very busy
4345 caches. Impatient users may tie up file descriptors and
4346 bandwidth by repeatedly requesting and immediately aborting
4349 When the user aborts a request, Squid will check the
4350 quick_abort values to the amount of data transfered until
4353 If the transfer has less than 'quick_abort_min' KB remaining,
4354 it will finish the retrieval.
4356 If the transfer has more than 'quick_abort_max' KB remaining,
4357 it will abort the retrieval.
4359 If more than 'quick_abort_pct' of the transfer has completed,
4360 it will finish the retrieval.
4362 If you do not want any retrieval to continue after the client
4363 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4366 If you want retrievals to always continue if they are being
4367 cached set 'quick_abort_min' to '-1 KB'.
4370 NAME: read_ahead_gap
4371 COMMENT: buffer-size
4373 LOC: Config.readAheadGap
4376 The amount of data the cache will buffer ahead of what has been
4377 sent to the client when retrieving an object from another server.
4381 IFDEF: USE_HTTP_VIOLATIONS
4384 LOC: Config.negativeTtl
4387 Set the Default Time-to-Live (TTL) for failed requests.
4388 Certain types of failures (such as "connection refused" and
4389 "404 Not Found") are able to be negatively-cached for a short time.
4390 Modern web servers should provide Expires: header, however if they
4391 do not this can provide a minimum TTL.
4392 The default is not to cache errors with unknown expiry details.
4394 Note that this is different from negative caching of DNS lookups.
4396 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4397 this feature could make you liable for problems which it
4401 NAME: positive_dns_ttl
4404 LOC: Config.positiveDnsTtl
4407 Upper limit on how long Squid will cache positive DNS responses.
4408 Default is 6 hours (360 minutes). This directive must be set
4409 larger than negative_dns_ttl.
4412 NAME: negative_dns_ttl
4415 LOC: Config.negativeDnsTtl
4418 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4419 This also sets the lower cache limit on positive lookups.
4420 Minimum value is 1 second, and it is not recommendable to go
4421 much below 10 seconds.
4424 NAME: range_offset_limit
4425 COMMENT: size [acl acl...]
4427 LOC: Config.rangeOffsetLimit
4430 usage: (size) [units] [[!]aclname]
4432 Sets an upper limit on how far (number of bytes) into the file
4433 a Range request may be to cause Squid to prefetch the whole file.
4434 If beyond this limit, Squid forwards the Range request as it is and
4435 the result is NOT cached.
4437 This is to stop a far ahead range request (lets say start at 17MB)
4438 from making Squid fetch the whole object up to that point before
4439 sending anything to the client.
4441 Multiple range_offset_limit lines may be specified, and they will
4442 be searched from top to bottom on each request until a match is found.
4443 The first match found will be used. If no line matches a request, the
4444 default limit of 0 bytes will be used.
4446 'size' is the limit specified as a number of units.
4448 'units' specifies whether to use bytes, KB, MB, etc.
4449 If no units are specified bytes are assumed.
4451 A size of 0 causes Squid to never fetch more than the
4452 client requested. (default)
4454 A size of 'none' causes Squid to always fetch the object from the
4455 beginning so it may cache the result. (2.0 style)
4457 'aclname' is the name of a defined ACL.
4459 NP: Using 'none' as the byte value here will override any quick_abort settings
4460 that may otherwise apply to the range request. The range request will
4461 be fully fetched from start to finish regardless of the client
4462 actions. This affects bandwidth usage.
4465 NAME: minimum_expiry_time
4468 LOC: Config.minimum_expiry_time
4471 The minimum caching time according to (Expires - Date)
4472 Headers Squid honors if the object can't be revalidated
4473 defaults to 60 seconds. In reverse proxy environments it
4474 might be desirable to honor shorter object lifetimes. It
4475 is most likely better to make your server return a
4476 meaningful Last-Modified header however. In ESI environments
4477 where page fragments often have short lifetimes, this will
4478 often be best set to 0.
4481 NAME: store_avg_object_size
4485 LOC: Config.Store.avgObjectSize
4487 Average object size, used to estimate number of objects your
4488 cache can hold. The default is 13 KB.
4491 NAME: store_objects_per_bucket
4494 LOC: Config.Store.objectsPerBucket
4496 Target number of objects per bucket in the store hash table.
4497 Lowering this value increases the total number of buckets and
4498 also the storage maintenance rate. The default is 20.
4503 -----------------------------------------------------------------------------
4506 NAME: request_header_max_size
4510 LOC: Config.maxRequestHeaderSize
4512 This specifies the maximum size for HTTP headers in a request.
4513 Request headers are usually relatively small (about 512 bytes).
4514 Placing a limit on the request header size will catch certain
4515 bugs (for example with persistent connections) and possibly
4516 buffer-overflow or denial-of-service attacks.
4519 NAME: reply_header_max_size
4523 LOC: Config.maxReplyHeaderSize
4525 This specifies the maximum size for HTTP headers in a reply.
4526 Reply headers are usually relatively small (about 512 bytes).
4527 Placing a limit on the reply header size will catch certain
4528 bugs (for example with persistent connections) and possibly
4529 buffer-overflow or denial-of-service attacks.
4532 NAME: request_body_max_size
4536 LOC: Config.maxRequestBodySize
4538 This specifies the maximum size for an HTTP request body.
4539 In other words, the maximum size of a PUT/POST request.
4540 A user who attempts to send a request with a body larger
4541 than this limit receives an "Invalid Request" error message.
4542 If you set this parameter to a zero (the default), there will
4543 be no limit imposed.
4546 NAME: client_request_buffer_max_size
4550 LOC: Config.maxRequestBufferSize
4552 This specifies the maximum buffer size of a client request.
4553 It prevents squid eating too much memory when somebody uploads
4557 NAME: chunked_request_body_max_size
4561 LOC: Config.maxChunkedRequestBodySize
4563 A broken or confused HTTP/1.1 client may send a chunked HTTP
4564 request to Squid. Squid does not have full support for that
4565 feature yet. To cope with such requests, Squid buffers the
4566 entire request and then dechunks request body to create a
4567 plain HTTP/1.0 request with a known content length. The plain
4568 request is then used by the rest of Squid code as usual.
4570 The option value specifies the maximum size of the buffer used
4571 to hold the request before the conversion. If the chunked
4572 request size exceeds the specified limit, the conversion
4573 fails, and the client receives an "unsupported request" error,
4574 as if dechunking was disabled.
4576 Dechunking is enabled by default. To disable conversion of
4577 chunked requests, set the maximum to zero.
4579 Request dechunking feature and this option in particular are a
4580 temporary hack. When chunking requests and responses are fully
4581 supported, there will be no need to buffer a chunked request.
4585 IFDEF: USE_HTTP_VIOLATIONS
4588 LOC: Config.accessList.brokenPosts
4590 A list of ACL elements which, if matched, causes Squid to send
4591 an extra CRLF pair after the body of a PUT/POST request.
4593 Some HTTP servers has broken implementations of PUT/POST,
4594 and rely on an extra CRLF pair sent by some WWW clients.
4596 Quote from RFC2616 section 4.1 on this matter:
4598 Note: certain buggy HTTP/1.0 client implementations generate an
4599 extra CRLF's after a POST request. To restate what is explicitly
4600 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4601 a request with an extra CRLF.
4603 This clause only supports fast acl types.
4604 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4607 acl buggy_server url_regex ^http://....
4608 broken_posts allow buggy_server
4611 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4614 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4616 LOC: Adaptation::Config::use_indirect_client
4618 Controls whether the indirect client IP address (instead of the direct
4619 client IP address) is passed to adaptation services.
4621 See also: follow_x_forwarded_for adaptation_send_client_ip
4625 IFDEF: USE_HTTP_VIOLATIONS
4629 LOC: Config.onoff.via
4631 If set (default), Squid will include a Via header in requests and
4632 replies as required by RFC2616.
4638 LOC: Config.onoff.ie_refresh
4641 Microsoft Internet Explorer up until version 5.5 Service
4642 Pack 1 has an issue with transparent proxies, wherein it
4643 is impossible to force a refresh. Turning this on provides
4644 a partial fix to the problem, by causing all IMS-REFRESH
4645 requests from older IE versions to check the origin server
4646 for fresh content. This reduces hit ratio by some amount
4647 (~10% in my experience), but allows users to actually get
4648 fresh content when they want it. Note because Squid
4649 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4650 of 5.5 is unchanged from old versions of Squid (i.e. a
4651 forced refresh is impossible). Newer versions of IE will,
4652 hopefully, continue to have the new behavior and will be
4653 handled based on that assumption. This option defaults to
4654 the old Squid behavior, which is better for hit ratios but
4655 worse for clients using IE, if they need to be able to
4656 force fresh content.
4659 NAME: vary_ignore_expire
4662 LOC: Config.onoff.vary_ignore_expire
4665 Many HTTP servers supporting Vary gives such objects
4666 immediate expiry time with no cache-control header
4667 when requested by a HTTP/1.0 client. This option
4668 enables Squid to ignore such expiry times until
4669 HTTP/1.1 is fully implemented.
4671 WARNING: If turned on this may eventually cause some
4672 varying objects not intended for caching to get cached.
4675 NAME: request_entities
4677 LOC: Config.onoff.request_entities
4680 Squid defaults to deny GET and HEAD requests with request entities,
4681 as the meaning of such requests are undefined in the HTTP standard
4682 even if not explicitly forbidden.
4684 Set this directive to on if you have clients which insists
4685 on sending request entities in GET or HEAD requests. But be warned
4686 that there is server software (both proxies and web servers) which
4687 can fail to properly process this kind of request which may make you
4688 vulnerable to cache pollution attacks if enabled.
4691 NAME: request_header_access
4692 IFDEF: USE_HTTP_VIOLATIONS
4693 TYPE: http_header_access
4694 LOC: Config.request_header_access
4697 Usage: request_header_access header_name allow|deny [!]aclname ...
4699 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4700 this feature could make you liable for problems which it
4703 This option replaces the old 'anonymize_headers' and the
4704 older 'http_anonymizer' option with something that is much
4705 more configurable. A list of ACLs for each header name allows
4706 removal of specific header fields under specific conditions.
4708 This option only applies to outgoing HTTP request headers (i.e.,
4709 headers sent by Squid to the next HTTP hop such as a cache peer
4710 or an origin server). The option has no effect during cache hit
4711 detection. The equivalent adaptation vectoring point in ICAP
4712 terminology is post-cache REQMOD.
4714 The option is applied to individual outgoing request header
4715 fields. For each request header field F, Squid uses the first
4716 qualifying sets of request_header_access rules:
4718 1. Rules with header_name equal to F's name.
4719 2. Rules with header_name 'Other', provided F's name is not
4720 on the hard-coded list of commonly used HTTP header names.
4721 3. Rules with header_name 'All'.
4723 Within that qualifying rule set, rule ACLs are checked as usual.
4724 If ACLs of an "allow" rule match, the header field is allowed to
4725 go through as is. If ACLs of a "deny" rule match, the header is
4726 removed and request_header_replace is then checked to identify
4727 if the removed header has a replacement. If no rules within the
4728 set have matching ACLs, the header field is left as is.
4730 For example, to achieve the same behavior as the old
4731 'http_anonymizer standard' option, you should use:
4733 request_header_access From deny all
4734 request_header_access Referer deny all
4735 request_header_access Server deny all
4736 request_header_access User-Agent deny all
4737 request_header_access WWW-Authenticate deny all
4738 request_header_access Link deny all
4740 Or, to reproduce the old 'http_anonymizer paranoid' feature
4743 request_header_access Allow allow all
4744 request_header_access Authorization allow all
4745 request_header_access WWW-Authenticate allow all
4746 request_header_access Proxy-Authorization allow all
4747 request_header_access Proxy-Authenticate allow all
4748 request_header_access Cache-Control allow all
4749 request_header_access Content-Encoding allow all
4750 request_header_access Content-Length allow all
4751 request_header_access Content-Type allow all
4752 request_header_access Date allow all
4753 request_header_access Expires allow all
4754 request_header_access Host allow all
4755 request_header_access If-Modified-Since allow all
4756 request_header_access Last-Modified allow all
4757 request_header_access Location allow all
4758 request_header_access Pragma allow all
4759 request_header_access Accept allow all
4760 request_header_access Accept-Charset allow all
4761 request_header_access Accept-Encoding allow all
4762 request_header_access Accept-Language allow all
4763 request_header_access Content-Language allow all
4764 request_header_access Mime-Version allow all
4765 request_header_access Retry-After allow all
4766 request_header_access Title allow all
4767 request_header_access Connection allow all
4768 request_header_access All deny all
4770 although many of those are HTTP reply headers, and so should be
4771 controlled with the reply_header_access directive.
4773 By default, all headers are allowed (no anonymizing is
4777 NAME: reply_header_access
4778 IFDEF: USE_HTTP_VIOLATIONS
4779 TYPE: http_header_access
4780 LOC: Config.reply_header_access
4783 Usage: reply_header_access header_name allow|deny [!]aclname ...
4785 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4786 this feature could make you liable for problems which it
4789 This option only applies to reply headers, i.e., from the
4790 server to the client.
4792 This is the same as request_header_access, but in the other
4793 direction. Please see request_header_access for detailed
4796 For example, to achieve the same behavior as the old
4797 'http_anonymizer standard' option, you should use:
4799 reply_header_access From deny all
4800 reply_header_access Referer deny all
4801 reply_header_access Server deny all
4802 reply_header_access User-Agent deny all
4803 reply_header_access WWW-Authenticate deny all
4804 reply_header_access Link deny all
4806 Or, to reproduce the old 'http_anonymizer paranoid' feature
4809 reply_header_access Allow allow all
4810 reply_header_access Authorization allow all
4811 reply_header_access WWW-Authenticate allow all
4812 reply_header_access Proxy-Authorization allow all
4813 reply_header_access Proxy-Authenticate allow all
4814 reply_header_access Cache-Control allow all
4815 reply_header_access Content-Encoding allow all
4816 reply_header_access Content-Length allow all
4817 reply_header_access Content-Type allow all
4818 reply_header_access Date allow all
4819 reply_header_access Expires allow all
4820 reply_header_access Host allow all
4821 reply_header_access If-Modified-Since allow all
4822 reply_header_access Last-Modified allow all
4823 reply_header_access Location allow all
4824 reply_header_access Pragma allow all
4825 reply_header_access Accept allow all
4826 reply_header_access Accept-Charset allow all
4827 reply_header_access Accept-Encoding allow all
4828 reply_header_access Accept-Language allow all
4829 reply_header_access Content-Language allow all
4830 reply_header_access Mime-Version allow all
4831 reply_header_access Retry-After allow all
4832 reply_header_access Title allow all
4833 reply_header_access Connection allow all
4834 reply_header_access All deny all
4836 although the HTTP request headers won't be usefully controlled
4837 by this directive -- see request_header_access for details.
4839 By default, all headers are allowed (no anonymizing is
4843 NAME: request_header_replace header_replace
4844 IFDEF: USE_HTTP_VIOLATIONS
4845 TYPE: http_header_replace
4846 LOC: Config.request_header_access
4849 Usage: request_header_replace header_name message
4850 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
4852 This option allows you to change the contents of headers
4853 denied with request_header_access above, by replacing them
4854 with some fixed string. This replaces the old fake_user_agent
4857 This only applies to request headers, not reply headers.
4859 By default, headers are removed if denied.
4862 NAME: reply_header_replace
4863 IFDEF: USE_HTTP_VIOLATIONS
4864 TYPE: http_header_replace
4865 LOC: Config.reply_header_access
4868 Usage: reply_header_replace header_name message
4869 Example: reply_header_replace Server Foo/1.0
4871 This option allows you to change the contents of headers
4872 denied with reply_header_access above, by replacing them
4873 with some fixed string.
4875 This only applies to reply headers, not request headers.
4877 By default, headers are removed if denied.
4880 NAME: request_header_add
4881 TYPE: HeaderWithAclList
4882 LOC: Config.request_header_add
4885 Usage: request_header_add field-name field-value acl1 [acl2] ...
4886 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
4888 This option adds header fields to outgoing HTTP requests (i.e.,
4889 request headers sent by Squid to the next HTTP hop such as a
4890 cache peer or an origin server). The option has no effect during
4891 cache hit detection. The equivalent adaptation vectoring point
4892 in ICAP terminology is post-cache REQMOD.
4894 Field-name is a token specifying an HTTP header name. If a
4895 standard HTTP header name is used, Squid does not check whether
4896 the new header conflicts with any existing headers or violates
4897 HTTP rules. If the request to be modified already contains a
4898 field with the same name, the old field is preserved but the
4899 header field values are not merged.
4901 Field-value is either a token or a quoted string. If quoted
4902 string format is used, then the surrounding quotes are removed
4903 while escape sequences and %macros are processed.
4905 In theory, all of the logformat codes can be used as %macros.
4906 However, unlike logging (which happens at the very end of
4907 transaction lifetime), the transaction may not yet have enough
4908 information to expand a macro when the new header value is needed.
4909 And some information may already be available to Squid but not yet
4910 committed where the macro expansion code can access it (report
4911 such instances!). The macro will be expanded into a single dash
4912 ('-') in such cases. Not all macros have been tested.
4914 One or more Squid ACLs may be specified to restrict header
4915 injection to matching requests. As always in squid.conf, all
4916 ACLs in an option ACL list must be satisfied for the insertion
4917 to happen. The request_header_add option supports fast ACLs
4921 NAME: relaxed_header_parser
4922 COMMENT: on|off|warn
4924 LOC: Config.onoff.relaxed_header_parser
4927 In the default "on" setting Squid accepts certain forms
4928 of non-compliant HTTP messages where it is unambiguous
4929 what the sending application intended even if the message
4930 is not correctly formatted. The messages is then normalized
4931 to the correct form when forwarded by Squid.
4933 If set to "warn" then a warning will be emitted in cache.log
4934 each time such HTTP error is encountered.
4936 If set to "off" then such HTTP errors will cause the request
4937 or response to be rejected.
4942 -----------------------------------------------------------------------------
4945 NAME: forward_timeout
4948 LOC: Config.Timeout.forward
4951 This parameter specifies how long Squid should at most attempt in
4952 finding a forwarding path for the request before giving up.
4955 NAME: connect_timeout
4958 LOC: Config.Timeout.connect
4961 This parameter specifies how long to wait for the TCP connect to
4962 the requested server or peer to complete before Squid should
4963 attempt to find another path where to forward the request.
4966 NAME: peer_connect_timeout
4969 LOC: Config.Timeout.peer_connect
4972 This parameter specifies how long to wait for a pending TCP
4973 connection to a peer cache. The default is 30 seconds. You
4974 may also set different timeout values for individual neighbors
4975 with the 'connect-timeout' option on a 'cache_peer' line.
4981 LOC: Config.Timeout.read
4984 The read_timeout is applied on server-side connections. After
4985 each successful read(), the timeout will be extended by this
4986 amount. If no data is read again after this amount of time,
4987 the request is aborted and logged with ERR_READ_TIMEOUT. The
4988 default is 15 minutes.
4994 LOC: Config.Timeout.write
4997 This timeout is tracked for all connections that have data
4998 available for writing and are waiting for the socket to become
4999 ready. After each successful write, the timeout is extended by
5000 the configured amount. If Squid has data to write but the
5001 connection is not ready for the configured duration, the
5002 transaction associated with the connection is terminated. The
5003 default is 15 minutes.
5006 NAME: request_timeout
5008 LOC: Config.Timeout.request
5011 How long to wait for complete HTTP request headers after initial
5012 connection establishment.
5015 NAME: client_idle_pconn_timeout persistent_request_timeout
5017 LOC: Config.Timeout.clientIdlePconn
5020 How long to wait for the next HTTP request on a persistent
5021 client connection after the previous request completes.
5024 NAME: client_lifetime
5027 LOC: Config.Timeout.lifetime
5030 The maximum amount of time a client (browser) is allowed to
5031 remain connected to the cache process. This protects the Cache
5032 from having a lot of sockets (and hence file descriptors) tied up
5033 in a CLOSE_WAIT state from remote clients that go away without
5034 properly shutting down (either because of a network failure or
5035 because of a poor client implementation). The default is one
5038 NOTE: The default value is intended to be much larger than any
5039 client would ever need to be connected to your cache. You
5040 should probably change client_lifetime only as a last resort.
5041 If you seem to have many client connections tying up
5042 filedescriptors, we recommend first tuning the read_timeout,
5043 request_timeout, persistent_request_timeout and quick_abort values.
5046 NAME: half_closed_clients
5048 LOC: Config.onoff.half_closed_clients
5051 Some clients may shutdown the sending side of their TCP
5052 connections, while leaving their receiving sides open. Sometimes,
5053 Squid can not tell the difference between a half-closed and a
5054 fully-closed TCP connection.
5056 By default, Squid will immediately close client connections when
5057 read(2) returns "no more data to read."
5059 Change this option to 'on' and Squid will keep open connections
5060 until a read(2) or write(2) on the socket returns an error.
5061 This may show some benefits for reverse proxies. But if not
5062 it is recommended to leave OFF.
5065 NAME: server_idle_pconn_timeout pconn_timeout
5067 LOC: Config.Timeout.serverIdlePconn
5070 Timeout for idle persistent connections to servers and other
5077 LOC: Ident::TheConfig.timeout
5080 Maximum time to wait for IDENT lookups to complete.
5082 If this is too high, and you enabled IDENT lookups from untrusted
5083 users, you might be susceptible to denial-of-service by having
5084 many ident requests going at once.
5087 NAME: shutdown_lifetime
5090 LOC: Config.shutdownLifetime
5093 When SIGTERM or SIGHUP is received, the cache is put into
5094 "shutdown pending" mode until all active sockets are closed.
5095 This value is the lifetime to set for all open descriptors
5096 during shutdown mode. Any active clients after this many
5097 seconds will receive a 'timeout' message.
5101 ADMINISTRATIVE PARAMETERS
5102 -----------------------------------------------------------------------------
5108 LOC: Config.adminEmail
5110 Email-address of local cache manager who will receive
5111 mail if the cache dies. The default is "webmaster."
5117 LOC: Config.EmailFrom
5119 From: email-address for mail sent when the cache dies.
5120 The default is to use 'appname@unique_hostname'.
5121 Default appname value is "squid", can be changed into
5122 src/globals.h before building squid.
5128 LOC: Config.EmailProgram
5130 Email program used to send mail if the cache dies.
5131 The default is "mail". The specified program must comply
5132 with the standard Unix mail syntax:
5133 mail-program recipient < mailfile
5135 Optional command line options can be specified.
5138 NAME: cache_effective_user
5140 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5141 LOC: Config.effectiveUser
5143 If you start Squid as root, it will change its effective/real
5144 UID/GID to the user specified below. The default is to change
5145 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5146 see also; cache_effective_group
5149 NAME: cache_effective_group
5152 LOC: Config.effectiveGroup
5154 Squid sets the GID to the effective user's default group ID
5155 (taken from the password file) and supplementary group list
5156 from the groups membership.
5158 If you want Squid to run with a specific GID regardless of
5159 the group memberships of the effective user then set this
5160 to the group (or GID) you want Squid to run as. When set
5161 all other group privileges of the effective user are ignored
5162 and only this GID is effective. If Squid is not started as
5163 root the user starting Squid MUST be member of the specified
5166 This option is not recommended by the Squid Team.
5167 Our preference is for administrators to configure a secure
5168 user account for squid with UID/GID matching system policies.
5171 NAME: httpd_suppress_version_string
5175 LOC: Config.onoff.httpd_suppress_version_string
5177 Suppress Squid version string info in HTTP headers and HTML error pages.
5180 NAME: visible_hostname
5182 LOC: Config.visibleHostname
5185 If you want to present a special hostname in error messages, etc,
5186 define this. Otherwise, the return value of gethostname()
5187 will be used. If you have multiple caches in a cluster and
5188 get errors about IP-forwarding you must set them to have individual
5189 names with this setting.
5192 NAME: unique_hostname
5194 LOC: Config.uniqueHostname
5197 If you want to have multiple machines with the same
5198 'visible_hostname' you must give each machine a different
5199 'unique_hostname' so forwarding loops can be detected.
5202 NAME: hostname_aliases
5204 LOC: Config.hostnameAliases
5207 A list of other DNS names your cache has.
5215 Minimum umask which should be enforced while the proxy
5216 is running, in addition to the umask set at startup.
5218 For a traditional octal representation of umasks, start
5223 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5224 -----------------------------------------------------------------------------
5226 This section contains parameters for the (optional) cache
5227 announcement service. This service is provided to help
5228 cache administrators locate one another in order to join or
5229 create cache hierarchies.
5231 An 'announcement' message is sent (via UDP) to the registration
5232 service by Squid. By default, the announcement message is NOT
5233 SENT unless you enable it with 'announce_period' below.
5235 The announcement message includes your hostname, plus the
5236 following information from this configuration file:
5242 All current information is processed regularly and made
5243 available on the Web at http://www.ircache.net/Cache/Tracker/.
5246 NAME: announce_period
5248 LOC: Config.Announce.period
5251 This is how frequently to send cache announcements. The
5252 default is `0' which disables sending the announcement
5255 To enable announcing your cache, just set an announce period.
5258 announce_period 1 day
5263 DEFAULT: tracker.ircache.net
5264 LOC: Config.Announce.host
5270 LOC: Config.Announce.file
5276 LOC: Config.Announce.port
5278 announce_host and announce_port set the hostname and port
5279 number where the registration message will be sent.
5281 Hostname will default to 'tracker.ircache.net' and port will
5282 default default to 3131. If the 'filename' argument is given,
5283 the contents of that file will be included in the announce
5288 HTTPD-ACCELERATOR OPTIONS
5289 -----------------------------------------------------------------------------
5292 NAME: httpd_accel_surrogate_id
5295 LOC: Config.Accel.surrogate_id
5297 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5298 need an identification token to allow control targeting. Because
5299 a farm of surrogates may all perform the same tasks, they may share
5300 an identification token.
5302 The default ID is the visible_hostname
5305 NAME: http_accel_surrogate_remote
5309 LOC: Config.onoff.surrogate_is_remote
5311 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
5312 Set this to on to have squid behave as a remote surrogate.
5316 IFDEF: USE_SQUID_ESI
5317 COMMENT: libxml2|expat|custom
5319 LOC: ESIParser::Type
5322 ESI markup is not strictly XML compatible. The custom ESI parser
5323 will give higher performance, but cannot handle non ASCII character
5328 DELAY POOL PARAMETERS
5329 -----------------------------------------------------------------------------
5333 TYPE: delay_pool_count
5335 IFDEF: USE_DELAY_POOLS
5338 This represents the number of delay pools to be used. For example,
5339 if you have one class 2 delay pool and one class 3 delays pool, you
5340 have a total of 2 delay pools.
5344 TYPE: delay_pool_class
5346 IFDEF: USE_DELAY_POOLS
5349 This defines the class of each delay pool. There must be exactly one
5350 delay_class line for each delay pool. For example, to define two
5351 delay pools, one of class 2 and one of class 3, the settings above
5355 delay_pools 4 # 4 delay pools
5356 delay_class 1 2 # pool 1 is a class 2 pool
5357 delay_class 2 3 # pool 2 is a class 3 pool
5358 delay_class 3 4 # pool 3 is a class 4 pool
5359 delay_class 4 5 # pool 4 is a class 5 pool
5361 The delay pool classes are:
5363 class 1 Everything is limited by a single aggregate
5366 class 2 Everything is limited by a single aggregate
5367 bucket as well as an "individual" bucket chosen
5368 from bits 25 through 32 of the IPv4 address.
5370 class 3 Everything is limited by a single aggregate
5371 bucket as well as a "network" bucket chosen
5372 from bits 17 through 24 of the IP address and a
5373 "individual" bucket chosen from bits 17 through
5374 32 of the IPv4 address.
5376 class 4 Everything in a class 3 delay pool, with an
5377 additional limit on a per user basis. This
5378 only takes effect if the username is established
5379 in advance - by forcing authentication in your
5382 class 5 Requests are grouped according their tag (see
5383 external_acl's tag= reply).
5386 Each pool also requires a delay_parameters directive to configure the pool size
5387 and speed limits used whenever the pool is applied to a request. Along with
5388 a set of delay_access directives to determine when it is used.
5390 NOTE: If an IP address is a.b.c.d
5391 -> bits 25 through 32 are "d"
5392 -> bits 17 through 24 are "c"
5393 -> bits 17 through 32 are "c * 256 + d"
5395 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5396 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5400 TYPE: delay_pool_access
5402 IFDEF: USE_DELAY_POOLS
5405 This is used to determine which delay pool a request falls into.
5407 delay_access is sorted per pool and the matching starts with pool 1,
5408 then pool 2, ..., and finally pool N. The first delay pool where the
5409 request is allowed is selected for the request. If it does not allow
5410 the request to any pool then the request is not delayed (default).
5412 For example, if you want some_big_clients in delay
5413 pool 1 and lotsa_little_clients in delay pool 2:
5416 delay_access 1 allow some_big_clients
5417 delay_access 1 deny all
5418 delay_access 2 allow lotsa_little_clients
5419 delay_access 2 deny all
5420 delay_access 3 allow authenticated_clients
5423 NAME: delay_parameters
5424 TYPE: delay_pool_rates
5426 IFDEF: USE_DELAY_POOLS
5429 This defines the parameters for a delay pool. Each delay pool has
5430 a number of "buckets" associated with it, as explained in the
5431 description of delay_class.
5433 For a class 1 delay pool, the syntax is:
5435 delay_parameters pool aggregate
5437 For a class 2 delay pool:
5439 delay_parameters pool aggregate individual
5441 For a class 3 delay pool:
5443 delay_parameters pool aggregate network individual
5445 For a class 4 delay pool:
5447 delay_parameters pool aggregate network individual user
5449 For a class 5 delay pool:
5451 delay_parameters pool tagrate
5453 The option variables are:
5455 pool a pool number - ie, a number between 1 and the
5456 number specified in delay_pools as used in
5459 aggregate the speed limit parameters for the aggregate bucket
5462 individual the speed limit parameters for the individual
5463 buckets (class 2, 3).
5465 network the speed limit parameters for the network buckets
5468 user the speed limit parameters for the user buckets
5471 tagrate the speed limit parameters for the tag buckets
5474 A pair of delay parameters is written restore/maximum, where restore is
5475 the number of bytes (not bits - modem and network speeds are usually
5476 quoted in bits) per second placed into the bucket, and maximum is the
5477 maximum number of bytes which can be in the bucket at any time.
5479 There must be one delay_parameters line for each delay pool.
5482 For example, if delay pool number 1 is a class 2 delay pool as in the
5483 above example, and is being used to strictly limit each host to 64Kbit/sec
5484 (plus overheads), with no overall limit, the line is:
5486 delay_parameters 1 -1/-1 8000/8000
5488 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5490 Note that the figure -1 is used to represent "unlimited".
5493 And, if delay pool number 2 is a class 3 delay pool as in the above
5494 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5495 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5496 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5497 to permit a decent web page to be downloaded at a decent speed
5498 (if the network is not being limited due to overuse) but slow down
5499 large downloads more significantly:
5501 delay_parameters 2 32000/32000 8000/8000 600/8000
5503 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5504 8 x 8000 KByte/sec -> 64Kbit/sec.
5505 8 x 600 Byte/sec -> 4800bit/sec.
5508 Finally, for a class 4 delay pool as in the example - each user will
5509 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5511 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5514 NAME: delay_initial_bucket_level
5515 COMMENT: (percent, 0-100)
5518 IFDEF: USE_DELAY_POOLS
5519 LOC: Config.Delay.initial
5521 The initial bucket percentage is used to determine how much is put
5522 in each bucket when squid starts, is reconfigured, or first notices
5523 a host accessing it (in class 2 and class 3, individual hosts and
5524 networks only have buckets associated with them once they have been
5529 CLIENT DELAY POOL PARAMETERS
5530 -----------------------------------------------------------------------------
5533 NAME: client_delay_pools
5534 TYPE: client_delay_pool_count
5536 IFDEF: USE_DELAY_POOLS
5537 LOC: Config.ClientDelay
5539 This option specifies the number of client delay pools used. It must
5540 preceed other client_delay_* options.
5543 client_delay_pools 2
5546 NAME: client_delay_initial_bucket_level
5547 COMMENT: (percent, 0-no_limit)
5550 IFDEF: USE_DELAY_POOLS
5551 LOC: Config.ClientDelay.initial
5553 This option determines the initial bucket size as a percentage of
5554 max_bucket_size from client_delay_parameters. Buckets are created
5555 at the time of the "first" connection from the matching IP. Idle
5556 buckets are periodically deleted up.
5558 You can specify more than 100 percent but note that such "oversized"
5559 buckets are not refilled until their size goes down to max_bucket_size
5560 from client_delay_parameters.
5563 client_delay_initial_bucket_level 50
5566 NAME: client_delay_parameters
5567 TYPE: client_delay_pool_rates
5569 IFDEF: USE_DELAY_POOLS
5570 LOC: Config.ClientDelay
5573 This option configures client-side bandwidth limits using the
5576 client_delay_parameters pool speed_limit max_bucket_size
5578 pool is an integer ID used for client_delay_access matching.
5580 speed_limit is bytes added to the bucket per second.
5582 max_bucket_size is the maximum size of a bucket, enforced after any
5583 speed_limit additions.
5585 Please see the delay_parameters option for more information and
5589 client_delay_parameters 1 1024 2048
5590 client_delay_parameters 2 51200 16384
5593 NAME: client_delay_access
5594 TYPE: client_delay_pool_access
5596 IFDEF: USE_DELAY_POOLS
5597 LOC: Config.ClientDelay
5600 This option determines the client-side delay pool for the
5603 client_delay_access pool_ID allow|deny acl_name
5605 All client_delay_access options are checked in their pool ID
5606 order, starting with pool 1. The first checked pool with allowed
5607 request is selected for the request. If no ACL matches or there
5608 are no client_delay_access options, the request bandwidth is not
5611 The ACL-selected pool is then used to find the
5612 client_delay_parameters for the request. Client-side pools are
5613 not used to aggregate clients. Clients are always aggregated
5614 based on their source IP addresses (one bucket per source IP).
5616 Please see delay_access for more examples.
5619 client_delay_access 1 allow low_rate_network
5620 client_delay_access 2 allow vips_network
5624 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5625 -----------------------------------------------------------------------------
5630 LOC: Config.Wccp.router
5634 Use this option to define your WCCP ``home'' router for
5637 wccp_router supports a single WCCP(v1) router
5639 wccp2_router supports multiple WCCPv2 routers
5641 only one of the two may be used at the same time and defines
5642 which version of WCCP to use.
5646 TYPE: IpAddress_list
5647 LOC: Config.Wccp2.router
5651 Use this option to define your WCCP ``home'' router for
5654 wccp_router supports a single WCCP(v1) router
5656 wccp2_router supports multiple WCCPv2 routers
5658 only one of the two may be used at the same time and defines
5659 which version of WCCP to use.
5664 LOC: Config.Wccp.version
5668 This directive is only relevant if you need to set up WCCP(v1)
5669 to some very old and end-of-life Cisco routers. In all other
5670 setups it must be left unset or at the default setting.
5671 It defines an internal version in the WCCP(v1) protocol,
5672 with version 4 being the officially documented protocol.
5674 According to some users, Cisco IOS 11.2 and earlier only
5675 support WCCP version 3. If you're using that or an earlier
5676 version of IOS, you may need to change this value to 3, otherwise
5677 do not specify this parameter.
5680 NAME: wccp2_rebuild_wait
5682 LOC: Config.Wccp2.rebuildwait
5686 If this is enabled Squid will wait for the cache dir rebuild to finish
5687 before sending the first wccp2 HereIAm packet
5690 NAME: wccp2_forwarding_method
5692 LOC: Config.Wccp2.forwarding_method
5696 WCCP2 allows the setting of forwarding methods between the
5697 router/switch and the cache. Valid values are as follows:
5699 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5700 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5702 Currently (as of IOS 12.4) cisco routers only support GRE.
5703 Cisco switches only support the L2 redirect assignment method.
5706 NAME: wccp2_return_method
5708 LOC: Config.Wccp2.return_method
5712 WCCP2 allows the setting of return methods between the
5713 router/switch and the cache for packets that the cache
5714 decides not to handle. Valid values are as follows:
5716 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5717 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5719 Currently (as of IOS 12.4) cisco routers only support GRE.
5720 Cisco switches only support the L2 redirect assignment.
5722 If the "ip wccp redirect exclude in" command has been
5723 enabled on the cache interface, then it is still safe for
5724 the proxy server to use a l2 redirect method even if this
5725 option is set to GRE.
5728 NAME: wccp2_assignment_method
5730 LOC: Config.Wccp2.assignment_method
5734 WCCP2 allows the setting of methods to assign the WCCP hash
5735 Valid values are as follows:
5737 hash - Hash assignment
5738 mask - Mask assignment
5740 As a general rule, cisco routers support the hash assignment method
5741 and cisco switches support the mask assignment method.
5746 LOC: Config.Wccp2.info
5747 DEFAULT_IF_NONE: standard 0
5750 WCCP2 allows for multiple traffic services. There are two
5751 types: "standard" and "dynamic". The standard type defines
5752 one service id - http (id 0). The dynamic service ids can be from
5753 51 to 255 inclusive. In order to use a dynamic service id
5754 one must define the type of traffic to be redirected; this is done
5755 using the wccp2_service_info option.
5757 The "standard" type does not require a wccp2_service_info option,
5758 just specifying the service id will suffice.
5760 MD5 service authentication can be enabled by adding
5761 "password=<password>" to the end of this service declaration.
5765 wccp2_service standard 0 # for the 'web-cache' standard service
5766 wccp2_service dynamic 80 # a dynamic service type which will be
5767 # fleshed out with subsequent options.
5768 wccp2_service standard 0 password=foo
5771 NAME: wccp2_service_info
5772 TYPE: wccp2_service_info
5773 LOC: Config.Wccp2.info
5777 Dynamic WCCPv2 services require further information to define the
5778 traffic you wish to have diverted.
5782 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
5783 priority=<priority> ports=<port>,<port>..
5785 The relevant WCCPv2 flags:
5786 + src_ip_hash, dst_ip_hash
5787 + source_port_hash, dst_port_hash
5788 + src_ip_alt_hash, dst_ip_alt_hash
5789 + src_port_alt_hash, dst_port_alt_hash
5792 The port list can be one to eight entries.
5796 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
5797 priority=240 ports=80
5799 Note: the service id must have been defined by a previous
5800 'wccp2_service dynamic <id>' entry.
5805 LOC: Config.Wccp2.weight
5809 Each cache server gets assigned a set of the destination
5810 hash proportional to their weight.
5815 LOC: Config.Wccp.address
5822 LOC: Config.Wccp2.address
5826 Use this option if you require WCCP to use a specific
5829 The default behavior is to not bind to any specific address.
5833 PERSISTENT CONNECTION HANDLING
5834 -----------------------------------------------------------------------------
5836 Also see "pconn_timeout" in the TIMEOUTS section
5839 NAME: client_persistent_connections
5841 LOC: Config.onoff.client_pconns
5845 NAME: server_persistent_connections
5847 LOC: Config.onoff.server_pconns
5850 Persistent connection support for clients and servers. By
5851 default, Squid uses persistent connections (when allowed)
5852 with its clients and servers. You can use these options to
5853 disable persistent connections with clients and/or servers.
5856 NAME: persistent_connection_after_error
5858 LOC: Config.onoff.error_pconns
5861 With this directive the use of persistent connections after
5862 HTTP errors can be disabled. Useful if you have clients
5863 who fail to handle errors on persistent connections proper.
5866 NAME: detect_broken_pconn
5868 LOC: Config.onoff.detect_broken_server_pconns
5871 Some servers have been found to incorrectly signal the use
5872 of HTTP/1.0 persistent connections even on replies not
5873 compatible, causing significant delays. This server problem
5874 has mostly been seen on redirects.
5876 By enabling this directive Squid attempts to detect such
5877 broken replies and automatically assume the reply is finished
5878 after 10 seconds timeout.
5882 CACHE DIGEST OPTIONS
5883 -----------------------------------------------------------------------------
5886 NAME: digest_generation
5887 IFDEF: USE_CACHE_DIGESTS
5889 LOC: Config.onoff.digest_generation
5892 This controls whether the server will generate a Cache Digest
5893 of its contents. By default, Cache Digest generation is
5894 enabled if Squid is compiled with --enable-cache-digests defined.
5897 NAME: digest_bits_per_entry
5898 IFDEF: USE_CACHE_DIGESTS
5900 LOC: Config.digest.bits_per_entry
5903 This is the number of bits of the server's Cache Digest which
5904 will be associated with the Digest entry for a given HTTP
5905 Method and URL (public key) combination. The default is 5.
5908 NAME: digest_rebuild_period
5909 IFDEF: USE_CACHE_DIGESTS
5912 LOC: Config.digest.rebuild_period
5915 This is the wait time between Cache Digest rebuilds.
5918 NAME: digest_rewrite_period
5920 IFDEF: USE_CACHE_DIGESTS
5922 LOC: Config.digest.rewrite_period
5925 This is the wait time between Cache Digest writes to
5929 NAME: digest_swapout_chunk_size
5932 IFDEF: USE_CACHE_DIGESTS
5933 LOC: Config.digest.swapout_chunk_size
5936 This is the number of bytes of the Cache Digest to write to
5937 disk at a time. It defaults to 4096 bytes (4KB), the Squid
5941 NAME: digest_rebuild_chunk_percentage
5942 COMMENT: (percent, 0-100)
5943 IFDEF: USE_CACHE_DIGESTS
5945 LOC: Config.digest.rebuild_chunk_percentage
5948 This is the percentage of the Cache Digest to be scanned at a
5949 time. By default it is set to 10% of the Cache Digest.
5954 -----------------------------------------------------------------------------
5959 LOC: Config.Port.snmp
5963 The port number where Squid listens for SNMP requests. To enable
5964 SNMP support set this to a suitable port number. Port number
5965 3401 is often used for the Squid SNMP agent. By default it's
5966 set to "0" (disabled)
5974 LOC: Config.accessList.snmp
5975 DEFAULT_IF_NONE: deny all
5978 Allowing or denying access to the SNMP port.
5980 All access to the agent is denied by default.
5983 snmp_access allow|deny [!]aclname ...
5985 This clause only supports fast acl types.
5986 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5988 snmp_access allow snmppublic localhost
5989 snmp_access deny all
5992 NAME: snmp_incoming_address
5994 LOC: Config.Addrs.snmp_incoming
5999 NAME: snmp_outgoing_address
6001 LOC: Config.Addrs.snmp_outgoing
6005 Just like 'udp_incoming_address', but for the SNMP port.
6007 snmp_incoming_address is used for the SNMP socket receiving
6008 messages from SNMP agents.
6009 snmp_outgoing_address is used for SNMP packets returned to SNMP
6012 The default snmp_incoming_address is to listen on all
6013 available network interfaces.
6015 If snmp_outgoing_address is not set it will use the same socket
6016 as snmp_incoming_address. Only change this if you want to have
6017 SNMP replies sent using another address than where this Squid
6018 listens for SNMP queries.
6020 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6021 the same value since they both use port 3401.
6026 -----------------------------------------------------------------------------
6029 NAME: icp_port udp_port
6032 LOC: Config.Port.icp
6034 The port number where Squid sends and receives ICP queries to
6035 and from neighbor caches. The standard UDP port for ICP is 3130.
6036 Default is disabled (0).
6039 icp_port @DEFAULT_ICP_PORT@
6046 LOC: Config.Port.htcp
6048 The port number where Squid sends and receives HTCP queries to
6049 and from neighbor caches. To turn it on you want to set it to
6050 4827. By default it is set to "0" (disabled).
6056 NAME: log_icp_queries
6060 LOC: Config.onoff.log_udp
6062 If set, ICP queries are logged to access.log. You may wish
6063 do disable this if your ICP load is VERY high to speed things
6064 up or to simplify log analysis.
6067 NAME: udp_incoming_address
6069 LOC:Config.Addrs.udp_incoming
6072 udp_incoming_address is used for UDP packets received from other
6075 The default behavior is to not bind to any specific address.
6077 Only change this if you want to have all UDP queries received on
6078 a specific interface/address.
6080 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6081 modules. Altering it will affect all of them in the same manner.
6083 see also; udp_outgoing_address
6085 NOTE, udp_incoming_address and udp_outgoing_address can not
6086 have the same value since they both use the same port.
6089 NAME: udp_outgoing_address
6091 LOC: Config.Addrs.udp_outgoing
6094 udp_outgoing_address is used for UDP packets sent out to other
6097 The default behavior is to not bind to any specific address.
6099 Instead it will use the same socket as udp_incoming_address.
6100 Only change this if you want to have UDP queries sent using another
6101 address than where this Squid listens for UDP queries from other
6104 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6105 modules. Altering it will affect all of them in the same manner.
6107 see also; udp_incoming_address
6109 NOTE, udp_incoming_address and udp_outgoing_address can not
6110 have the same value since they both use the same port.
6117 LOC: Config.onoff.icp_hit_stale
6119 If you want to return ICP_HIT for stale cache objects, set this
6120 option to 'on'. If you have sibling relationships with caches
6121 in other administrative domains, this should be 'off'. If you only
6122 have sibling relationships with caches under your control,
6123 it is probably okay to set this to 'on'.
6124 If set to 'on', your siblings should use the option "allow-miss"
6125 on their cache_peer lines for connecting to you.
6128 NAME: minimum_direct_hops
6131 LOC: Config.minDirectHops
6133 If using the ICMP pinging stuff, do direct fetches for sites
6134 which are no more than this many hops away.
6137 NAME: minimum_direct_rtt
6140 LOC: Config.minDirectRtt
6142 If using the ICMP pinging stuff, do direct fetches for sites
6143 which are no more than this many rtt milliseconds away.
6149 LOC: Config.Netdb.low
6155 LOC: Config.Netdb.high
6157 The low and high water marks for the ICMP measurement
6158 database. These are counts, not percents. The defaults are
6159 900 and 1000. When the high water mark is reached, database
6160 entries will be deleted until the low mark is reached.
6163 NAME: netdb_ping_period
6165 LOC: Config.Netdb.period
6168 The minimum period for measuring a site. There will be at
6169 least this much delay between successive pings to the same
6170 network. The default is five minutes.
6177 LOC: Config.onoff.query_icmp
6179 If you want to ask your peers to include ICMP data in their ICP
6180 replies, enable this option.
6182 If your peer has configured Squid (during compilation) with
6183 '--enable-icmp' that peer will send ICMP pings to origin server
6184 sites of the URLs it receives. If you enable this option the
6185 ICP replies from that peer will include the ICMP data (if available).
6186 Then, when choosing a parent cache, Squid will choose the parent with
6187 the minimal RTT to the origin server. When this happens, the
6188 hierarchy field of the access.log will be
6189 "CLOSEST_PARENT_MISS". This option is off by default.
6192 NAME: test_reachability
6196 LOC: Config.onoff.test_reachability
6198 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6199 instead of ICP_MISS if the target host is NOT in the ICMP
6200 database, or has a zero RTT.
6203 NAME: icp_query_timeout
6207 LOC: Config.Timeout.icp_query
6209 Normally Squid will automatically determine an optimal ICP
6210 query timeout value based on the round-trip-time of recent ICP
6211 queries. If you want to override the value determined by
6212 Squid, set this 'icp_query_timeout' to a non-zero value. This
6213 value is specified in MILLISECONDS, so, to use a 2-second
6214 timeout (the old default), you would write:
6216 icp_query_timeout 2000
6219 NAME: maximum_icp_query_timeout
6223 LOC: Config.Timeout.icp_query_max
6225 Normally the ICP query timeout is determined dynamically. But
6226 sometimes it can lead to very large values (say 5 seconds).
6227 Use this option to put an upper limit on the dynamic timeout
6228 value. Do NOT use this option to always use a fixed (instead
6229 of a dynamic) timeout value. To set a fixed timeout see the
6230 'icp_query_timeout' directive.
6233 NAME: minimum_icp_query_timeout
6237 LOC: Config.Timeout.icp_query_min
6239 Normally the ICP query timeout is determined dynamically. But
6240 sometimes it can lead to very small timeouts, even lower than
6241 the normal latency variance on your link due to traffic.
6242 Use this option to put an lower limit on the dynamic timeout
6243 value. Do NOT use this option to always use a fixed (instead
6244 of a dynamic) timeout value. To set a fixed timeout see the
6245 'icp_query_timeout' directive.
6248 NAME: background_ping_rate
6252 LOC: Config.backgroundPingRate
6254 Controls how often the ICP pings are sent to siblings that
6255 have background-ping set.
6259 MULTICAST ICP OPTIONS
6260 -----------------------------------------------------------------------------
6265 LOC: Config.mcast_group_list
6268 This tag specifies a list of multicast groups which your server
6269 should join to receive multicasted ICP queries.
6271 NOTE! Be very careful what you put here! Be sure you
6272 understand the difference between an ICP _query_ and an ICP
6273 _reply_. This option is to be set only if you want to RECEIVE
6274 multicast queries. Do NOT set this option to SEND multicast
6275 ICP (use cache_peer for that). ICP replies are always sent via
6276 unicast, so this option does not affect whether or not you will
6277 receive replies from multicast group members.
6279 You must be very careful to NOT use a multicast address which
6280 is already in use by another group of caches.
6282 If you are unsure about multicast, please read the Multicast
6283 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6285 Usage: mcast_groups 239.128.16.128 224.0.1.20
6287 By default, Squid doesn't listen on any multicast groups.
6290 NAME: mcast_miss_addr
6291 IFDEF: MULTICAST_MISS_STREAM
6293 LOC: Config.mcast_miss.addr
6296 If you enable this option, every "cache miss" URL will
6297 be sent out on the specified multicast address.
6299 Do not enable this option unless you are are absolutely
6300 certain you understand what you are doing.
6303 NAME: mcast_miss_ttl
6304 IFDEF: MULTICAST_MISS_STREAM
6306 LOC: Config.mcast_miss.ttl
6309 This is the time-to-live value for packets multicasted
6310 when multicasting off cache miss URLs is enabled. By
6311 default this is set to 'site scope', i.e. 16.
6314 NAME: mcast_miss_port
6315 IFDEF: MULTICAST_MISS_STREAM
6317 LOC: Config.mcast_miss.port
6320 This is the port number to be used in conjunction with
6324 NAME: mcast_miss_encode_key
6325 IFDEF: MULTICAST_MISS_STREAM
6327 LOC: Config.mcast_miss.encode_key
6328 DEFAULT: XXXXXXXXXXXXXXXX
6330 The URLs that are sent in the multicast miss stream are
6331 encrypted. This is the encryption key.
6334 NAME: mcast_icp_query_timeout
6338 LOC: Config.Timeout.mcast_icp_query
6340 For multicast peers, Squid regularly sends out ICP "probes" to
6341 count how many other peers are listening on the given multicast
6342 address. This value specifies how long Squid should wait to
6343 count all the replies. The default is 2000 msec, or 2
6348 INTERNAL ICON OPTIONS
6349 -----------------------------------------------------------------------------
6352 NAME: icon_directory
6354 LOC: Config.icons.directory
6355 DEFAULT: @DEFAULT_ICON_DIR@
6357 Where the icons are stored. These are normally kept in
6361 NAME: global_internal_static
6363 LOC: Config.onoff.global_internal_static
6366 This directive controls is Squid should intercept all requests for
6367 /squid-internal-static/ no matter which host the URL is requesting
6368 (default on setting), or if nothing special should be done for
6369 such URLs (off setting). The purpose of this directive is to make
6370 icons etc work better in complex cache hierarchies where it may
6371 not always be possible for all corners in the cache mesh to reach
6372 the server generating a directory listing.
6375 NAME: short_icon_urls
6377 LOC: Config.icons.use_short_names
6380 If this is enabled Squid will use short URLs for icons.
6381 If disabled it will revert to the old behavior of including
6382 it's own name and port in the URL.
6384 If you run a complex cache hierarchy with a mix of Squid and
6385 other proxies you may need to disable this directive.
6390 -----------------------------------------------------------------------------
6393 NAME: error_directory
6395 LOC: Config.errorDirectory
6398 If you wish to create your own versions of the default
6399 error files to customize them to suit your company copy
6400 the error/template files to another directory and point
6403 WARNING: This option will disable multi-language support
6404 on error pages if used.
6406 The squid developers are interested in making squid available in
6407 a wide variety of languages. If you are making translations for a
6408 language that Squid does not currently provide please consider
6409 contributing your translation back to the project.
6410 http://wiki.squid-cache.org/Translations
6412 The squid developers working on translations are happy to supply drop-in
6413 translated error files in exchange for any new language contributions.
6416 NAME: error_default_language
6417 IFDEF: USE_ERR_LOCALES
6419 LOC: Config.errorDefaultLanguage
6422 Set the default language which squid will send error pages in
6423 if no existing translation matches the clients language
6426 If unset (default) generic English will be used.
6428 The squid developers are interested in making squid available in
6429 a wide variety of languages. If you are interested in making
6430 translations for any language see the squid wiki for details.
6431 http://wiki.squid-cache.org/Translations
6434 NAME: error_log_languages
6435 IFDEF: USE_ERR_LOCALES
6437 LOC: Config.errorLogMissingLanguages
6440 Log to cache.log what languages users are attempting to
6441 auto-negotiate for translations.
6443 Successful negotiations are not logged. Only failures
6444 have meaning to indicate that Squid may need an upgrade
6445 of its error page translations.
6448 NAME: err_page_stylesheet
6450 LOC: Config.errorStylesheet
6451 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6453 CSS Stylesheet to pattern the display of Squid default error pages.
6455 For information on CSS see http://www.w3.org/Style/CSS/
6460 LOC: Config.errHtmlText
6463 HTML text to include in error messages. Make this a "mailto"
6464 URL to your admin address, or maybe just a link to your
6465 organizations Web page.
6467 To include this in your error messages, you must rewrite
6468 the error template files (found in the "errors" directory).
6469 Wherever you want the 'err_html_text' line to appear,
6470 insert a %L tag in the error template file.
6473 NAME: email_err_data
6476 LOC: Config.onoff.emailErrData
6479 If enabled, information about the occurred error will be
6480 included in the mailto links of the ERR pages (if %W is set)
6481 so that the email body contains the data.
6482 Syntax is <A HREF="mailto:%w%W">%w</A>
6487 LOC: Config.denyInfoList
6490 Usage: deny_info err_page_name acl
6491 or deny_info http://... acl
6492 or deny_info TCP_RESET acl
6494 This can be used to return a ERR_ page for requests which
6495 do not pass the 'http_access' rules. Squid remembers the last
6496 acl it evaluated in http_access, and if a 'deny_info' line exists
6497 for that ACL Squid returns a corresponding error page.
6499 The acl is typically the last acl on the http_access deny line which
6500 denied access. The exceptions to this rule are:
6501 - When Squid needs to request authentication credentials. It's then
6502 the first authentication related acl encountered
6503 - When none of the http_access lines matches. It's then the last
6504 acl processed on the last http_access line.
6505 - When the decision to deny access was made by an adaptation service,
6506 the acl name is the corresponding eCAP or ICAP service_name.
6508 NP: If providing your own custom error pages with error_directory
6509 you may also specify them by your custom file name:
6510 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6512 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6513 may be specified by prefixing the file name with the code and a colon.
6514 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6516 Alternatively you can tell Squid to reset the TCP connection
6517 by specifying TCP_RESET.
6519 Or you can specify an error URL or URL pattern. The browsers will
6520 get redirected to the specified URL after formatting tags have
6521 been replaced. Redirect will be done with 302 or 307 according to
6522 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6523 the URL. e.g. 303:http://example.com/
6526 %a - username (if available. Password NOT included)
6529 %E - Error description
6531 %H - Request domain name
6532 %i - Client IP Address
6534 %o - Message result from external ACL helper
6535 %p - Request Port number
6536 %P - Request Protocol name
6537 %R - Request URL path
6538 %T - Timestamp in RFC 1123 format
6539 %U - Full canonical URL from client
6540 (HTTPS URLs terminate with *)
6541 %u - Full canonical URL from client
6542 %w - Admin email from squid.conf
6544 %% - Literal percent (%) code
6549 OPTIONS INFLUENCING REQUEST FORWARDING
6550 -----------------------------------------------------------------------------
6553 NAME: nonhierarchical_direct
6555 LOC: Config.onoff.nonhierarchical_direct
6558 By default, Squid will send any non-hierarchical requests
6559 (matching hierarchy_stoplist or not cacheable request type) direct
6562 If you set this to off, Squid will prefer to send these
6563 requests to parents.
6565 Note that in most configurations, by turning this off you will only
6566 add latency to these request without any improvement in global hit
6569 If you are inside an firewall see never_direct instead of
6575 LOC: Config.onoff.prefer_direct
6578 Normally Squid tries to use parents for most requests. If you for some
6579 reason like it to first try going direct and only use a parent if
6580 going direct fails set this to on.
6582 By combining nonhierarchical_direct off and prefer_direct on you
6583 can set up Squid to use a parent as a backup path if going direct
6586 Note: If you want Squid to use parents for all requests see
6587 the never_direct directive. prefer_direct only modifies how Squid
6588 acts on cacheable requests.
6593 LOC: Config.accessList.AlwaysDirect
6596 Usage: always_direct allow|deny [!]aclname ...
6598 Here you can use ACL elements to specify requests which should
6599 ALWAYS be forwarded by Squid to the origin servers without using
6600 any peers. For example, to always directly forward requests for
6601 local servers ignoring any parents or siblings you may have use
6604 acl local-servers dstdomain my.domain.net
6605 always_direct allow local-servers
6607 To always forward FTP requests directly, use
6610 always_direct allow FTP
6612 NOTE: There is a similar, but opposite option named
6613 'never_direct'. You need to be aware that "always_direct deny
6614 foo" is NOT the same thing as "never_direct allow foo". You
6615 may need to use a deny rule to exclude a more-specific case of
6616 some other rule. Example:
6618 acl local-external dstdomain external.foo.net
6619 acl local-servers dstdomain .foo.net
6620 always_direct deny local-external
6621 always_direct allow local-servers
6623 NOTE: If your goal is to make the client forward the request
6624 directly to the origin server bypassing Squid then this needs
6625 to be done in the client configuration. Squid configuration
6626 can only tell Squid how Squid should fetch the object.
6628 NOTE: This directive is not related to caching. The replies
6629 is cached as usual even if you use always_direct. To not cache
6630 the replies see the 'cache' directive.
6632 This clause supports both fast and slow acl types.
6633 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6638 LOC: Config.accessList.NeverDirect
6641 Usage: never_direct allow|deny [!]aclname ...
6643 never_direct is the opposite of always_direct. Please read
6644 the description for always_direct if you have not already.
6646 With 'never_direct' you can use ACL elements to specify
6647 requests which should NEVER be forwarded directly to origin
6648 servers. For example, to force the use of a proxy for all
6649 requests, except those in your local domain use something like:
6651 acl local-servers dstdomain .foo.net
6652 never_direct deny local-servers
6653 never_direct allow all
6655 or if Squid is inside a firewall and there are local intranet
6656 servers inside the firewall use something like:
6658 acl local-intranet dstdomain .foo.net
6659 acl local-external dstdomain external.foo.net
6660 always_direct deny local-external
6661 always_direct allow local-intranet
6662 never_direct allow all
6664 This clause supports both fast and slow acl types.
6665 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6669 ADVANCED NETWORKING OPTIONS
6670 -----------------------------------------------------------------------------
6673 NAME: incoming_udp_average incoming_icp_average
6676 LOC: Config.comm_incoming.udp.average
6678 Heavy voodoo here. I can't even believe you are reading this.
6679 Are you crazy? Don't even think about adjusting these unless
6680 you understand the algorithms in comm_select.c first!
6683 NAME: incoming_tcp_average incoming_http_average
6686 LOC: Config.comm_incoming.tcp.average
6688 Heavy voodoo here. I can't even believe you are reading this.
6689 Are you crazy? Don't even think about adjusting these unless
6690 you understand the algorithms in comm_select.c first!
6693 NAME: incoming_dns_average
6696 LOC: Config.comm_incoming.dns.average
6698 Heavy voodoo here. I can't even believe you are reading this.
6699 Are you crazy? Don't even think about adjusting these unless
6700 you understand the algorithms in comm_select.c first!
6703 NAME: min_udp_poll_cnt min_icp_poll_cnt
6706 LOC: Config.comm_incoming.udp.min_poll
6708 Heavy voodoo here. I can't even believe you are reading this.
6709 Are you crazy? Don't even think about adjusting these unless
6710 you understand the algorithms in comm_select.c first!
6713 NAME: min_dns_poll_cnt
6716 LOC: Config.comm_incoming.dns.min_poll
6718 Heavy voodoo here. I can't even believe you are reading this.
6719 Are you crazy? Don't even think about adjusting these unless
6720 you understand the algorithms in comm_select.c first!
6723 NAME: min_tcp_poll_cnt min_http_poll_cnt
6726 LOC: Config.comm_incoming.tcp.min_poll
6728 Heavy voodoo here. I can't even believe you are reading this.
6729 Are you crazy? Don't even think about adjusting these unless
6730 you understand the algorithms in comm_select.c first!
6736 LOC: Config.accept_filter
6740 The name of an accept(2) filter to install on Squid's
6741 listen socket(s). This feature is perhaps specific to
6742 FreeBSD and requires support in the kernel.
6744 The 'httpready' filter delays delivering new connections
6745 to Squid until a full HTTP request has been received.
6746 See the accf_http(9) man page for details.
6748 The 'dataready' filter delays delivering new connections
6749 to Squid until there is some data to process.
6750 See the accf_dataready(9) man page for details.
6754 The 'data' filter delays delivering of new connections
6755 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
6756 You may optionally specify a number of seconds to wait by
6757 'data=N' where N is the number of seconds. Defaults to 30
6758 if not specified. See the tcp(7) man page for details.
6761 accept_filter httpready
6766 NAME: client_ip_max_connections
6768 LOC: Config.client_ip_max_connections
6771 Set an absolute limit on the number of connections a single
6772 client IP can use. Any more than this and Squid will begin to drop
6773 new connections from the client until it closes some links.
6775 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
6776 connections from the client. For finer control use the ACL access controls.
6778 Requires client_db to be enabled (the default).
6780 WARNING: This may noticably slow down traffic received via external proxies
6781 or NAT devices and cause them to rebound error messages back to their clients.
6784 NAME: tcp_recv_bufsize
6788 LOC: Config.tcpRcvBufsz
6790 Size of receive buffer to set for TCP sockets. Probably just
6791 as easy to change your kernel's default. Set to zero to use
6792 the default buffer size.
6797 -----------------------------------------------------------------------------
6804 LOC: Adaptation::Icap::TheConfig.onoff
6807 If you want to enable the ICAP module support, set this to on.
6810 NAME: icap_connect_timeout
6813 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
6816 This parameter specifies how long to wait for the TCP connect to
6817 the requested ICAP server to complete before giving up and either
6818 terminating the HTTP transaction or bypassing the failure.
6820 The default for optional services is peer_connect_timeout.
6821 The default for essential services is connect_timeout.
6822 If this option is explicitly set, its value applies to all services.
6825 NAME: icap_io_timeout
6829 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
6832 This parameter specifies how long to wait for an I/O activity on
6833 an established, active ICAP connection before giving up and
6834 either terminating the HTTP transaction or bypassing the
6837 The default is read_timeout.
6840 NAME: icap_service_failure_limit
6841 COMMENT: limit [in memory-depth time-units]
6842 TYPE: icap_service_failure_limit
6844 LOC: Adaptation::Icap::TheConfig
6847 The limit specifies the number of failures that Squid tolerates
6848 when establishing a new TCP connection with an ICAP service. If
6849 the number of failures exceeds the limit, the ICAP service is
6850 not used for new ICAP requests until it is time to refresh its
6853 A negative value disables the limit. Without the limit, an ICAP
6854 service will not be considered down due to connectivity failures
6855 between ICAP OPTIONS requests.
6857 Squid forgets ICAP service failures older than the specified
6858 value of memory-depth. The memory fading algorithm
6859 is approximate because Squid does not remember individual
6860 errors but groups them instead, splitting the option
6861 value into ten time slots of equal length.
6863 When memory-depth is 0 and by default this option has no
6864 effect on service failure expiration.
6866 Squid always forgets failures when updating service settings
6867 using an ICAP OPTIONS transaction, regardless of this option
6871 # suspend service usage after 10 failures in 5 seconds:
6872 icap_service_failure_limit 10 in 5 seconds
6875 NAME: icap_service_revival_delay
6878 LOC: Adaptation::Icap::TheConfig.service_revival_delay
6881 The delay specifies the number of seconds to wait after an ICAP
6882 OPTIONS request failure before requesting the options again. The
6883 failed ICAP service is considered "down" until fresh OPTIONS are
6886 The actual delay cannot be smaller than the hardcoded minimum
6887 delay of 30 seconds.
6890 NAME: icap_preview_enable
6894 LOC: Adaptation::Icap::TheConfig.preview_enable
6897 The ICAP Preview feature allows the ICAP server to handle the
6898 HTTP message by looking only at the beginning of the message body
6899 or even without receiving the body at all. In some environments,
6900 previews greatly speedup ICAP processing.
6902 During an ICAP OPTIONS transaction, the server may tell Squid what
6903 HTTP messages should be previewed and how big the preview should be.
6904 Squid will not use Preview if the server did not request one.
6906 To disable ICAP Preview for all ICAP services, regardless of
6907 individual ICAP server OPTIONS responses, set this option to "off".
6909 icap_preview_enable off
6912 NAME: icap_preview_size
6915 LOC: Adaptation::Icap::TheConfig.preview_size
6918 The default size of preview data to be sent to the ICAP server.
6919 -1 means no preview. This value might be overwritten on a per server
6920 basis by OPTIONS requests.
6923 NAME: icap_206_enable
6927 LOC: Adaptation::Icap::TheConfig.allow206_enable
6930 206 (Partial Content) responses is an ICAP extension that allows the
6931 ICAP agents to optionally combine adapted and original HTTP message
6932 content. The decision to combine is postponed until the end of the
6933 ICAP response. Squid supports Partial Content extension by default.
6935 Activation of the Partial Content extension is negotiated with each
6936 ICAP service during OPTIONS exchange. Most ICAP servers should handle
6937 negotation correctly even if they do not support the extension, but
6938 some might fail. To disable Partial Content support for all ICAP
6939 services and to avoid any negotiation, set this option to "off".
6945 NAME: icap_default_options_ttl
6948 LOC: Adaptation::Icap::TheConfig.default_options_ttl
6951 The default TTL value for ICAP OPTIONS responses that don't have
6952 an Options-TTL header.
6955 NAME: icap_persistent_connections
6959 LOC: Adaptation::Icap::TheConfig.reuse_connections
6962 Whether or not Squid should use persistent connections to
6966 NAME: adaptation_send_client_ip icap_send_client_ip
6968 IFDEF: USE_ADAPTATION
6970 LOC: Adaptation::Config::send_client_ip
6973 If enabled, Squid shares HTTP client IP information with adaptation
6974 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
6975 For eCAP, Squid sets the libecap::metaClientIp transaction option.
6977 See also: adaptation_uses_indirect_client
6980 NAME: adaptation_send_username icap_send_client_username
6982 IFDEF: USE_ADAPTATION
6984 LOC: Adaptation::Config::send_username
6987 This sends authenticated HTTP client username (if available) to
6988 the adaptation service.
6990 For ICAP, the username value is encoded based on the
6991 icap_client_username_encode option and is sent using the header
6992 specified by the icap_client_username_header option.
6995 NAME: icap_client_username_header
6998 LOC: Adaptation::Icap::TheConfig.client_username_header
6999 DEFAULT: X-Client-Username
7001 ICAP request header name to use for adaptation_send_username.
7004 NAME: icap_client_username_encode
7008 LOC: Adaptation::Icap::TheConfig.client_username_encode
7011 Whether to base64 encode the authenticated client username.
7015 TYPE: icap_service_type
7017 LOC: Adaptation::Icap::TheConfig
7020 Defines a single ICAP service using the following format:
7022 icap_service id vectoring_point uri [option ...]
7025 an opaque identifier or name which is used to direct traffic to
7026 this specific service. Must be unique among all adaptation
7027 services in squid.conf.
7029 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7030 This specifies at which point of transaction processing the
7031 ICAP service should be activated. *_postcache vectoring points
7032 are not yet supported.
7034 uri: icap://servername:port/servicepath
7035 ICAP server and service location.
7037 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7038 transactions. Squid does not enforce that requirement. You can specify
7039 services with the same service_url and different vectoring_points. You
7040 can even specify multiple identical services as long as their
7041 service_names differ.
7044 Service options are separated by white space. ICAP services support
7045 the following name=value options:
7048 If set to 'on' or '1', the ICAP service is treated as
7049 optional. If the service cannot be reached or malfunctions,
7050 Squid will try to ignore any errors and process the message as
7051 if the service was not enabled. No all ICAP errors can be
7052 bypassed. If set to 0, the ICAP service is treated as
7053 essential and all ICAP errors will result in an error page
7054 returned to the HTTP client.
7056 Bypass is off by default: services are treated as essential.
7059 If set to 'on' or '1', the ICAP service is allowed to
7060 dynamically change the current message adaptation plan by
7061 returning a chain of services to be used next. The services
7062 are specified using the X-Next-Services ICAP response header
7063 value, formatted as a comma-separated list of service names.
7064 Each named service should be configured in squid.conf. Other
7065 services are ignored. An empty X-Next-Services value results
7066 in an empty plan which ends the current adaptation.
7068 Dynamic adaptation plan may cross or cover multiple supported
7069 vectoring points in their natural processing order.
7071 Routing is not allowed by default: the ICAP X-Next-Services
7072 response header is ignored.
7075 Only has effect on split-stack systems. The default on those systems
7076 is to use IPv4-only connections. When set to 'on' this option will
7077 make Squid use IPv6-only connections to contact this ICAP service.
7079 on-overload=block|bypass|wait|force
7080 If the service Max-Connections limit has been reached, do
7081 one of the following for each new ICAP transaction:
7082 * block: send an HTTP error response to the client
7083 * bypass: ignore the "over-connected" ICAP service
7084 * wait: wait (in a FIFO queue) for an ICAP connection slot
7085 * force: proceed, ignoring the Max-Connections limit
7087 In SMP mode with N workers, each worker assumes the service
7088 connection limit is Max-Connections/N, even though not all
7089 workers may use a given service.
7091 The default value is "bypass" if service is bypassable,
7092 otherwise it is set to "wait".
7096 Use the given number as the Max-Connections limit, regardless
7097 of the Max-Connections value given by the service, if any.
7099 Older icap_service format without optional named parameters is
7100 deprecated but supported for backward compatibility.
7103 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7104 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7108 TYPE: icap_class_type
7113 This deprecated option was documented to define an ICAP service
7114 chain, even though it actually defined a set of similar, redundant
7115 services, and the chains were not supported.
7117 To define a set of redundant services, please use the
7118 adaptation_service_set directive. For service chains, use
7119 adaptation_service_chain.
7123 TYPE: icap_access_type
7128 This option is deprecated. Please use adaptation_access, which
7129 has the same ICAP functionality, but comes with better
7130 documentation, and eCAP support.
7135 -----------------------------------------------------------------------------
7142 LOC: Adaptation::Ecap::TheConfig.onoff
7145 Controls whether eCAP support is enabled.
7149 TYPE: ecap_service_type
7151 LOC: Adaptation::Ecap::TheConfig
7154 Defines a single eCAP service
7156 ecap_service id vectoring_point uri [option ...]
7159 an opaque identifier or name which is used to direct traffic to
7160 this specific service. Must be unique among all adaptation
7161 services in squid.conf.
7163 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7164 This specifies at which point of transaction processing the
7165 eCAP service should be activated. *_postcache vectoring points
7166 are not yet supported.
7168 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7169 Squid uses the eCAP service URI to match this configuration
7170 line with one of the dynamically loaded services. Each loaded
7171 eCAP service must have a unique URI. Obtain the right URI from
7172 the service provider.
7175 Service options are separated by white space. eCAP services support
7176 the following name=value options:
7179 If set to 'on' or '1', the eCAP service is treated as optional.
7180 If the service cannot be reached or malfunctions, Squid will try
7181 to ignore any errors and process the message as if the service
7182 was not enabled. No all eCAP errors can be bypassed.
7183 If set to 'off' or '0', the eCAP service is treated as essential
7184 and all eCAP errors will result in an error page returned to the
7187 Bypass is off by default: services are treated as essential.
7190 If set to 'on' or '1', the eCAP service is allowed to
7191 dynamically change the current message adaptation plan by
7192 returning a chain of services to be used next.
7194 Dynamic adaptation plan may cross or cover multiple supported
7195 vectoring points in their natural processing order.
7197 Routing is not allowed by default.
7199 Older ecap_service format without optional named parameters is
7200 deprecated but supported for backward compatibility.
7204 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7205 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7208 NAME: loadable_modules
7210 IFDEF: USE_LOADABLE_MODULES
7211 LOC: Config.loadable_module_names
7214 Instructs Squid to load the specified dynamic module(s) or activate
7215 preloaded module(s).
7217 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7221 MESSAGE ADAPTATION OPTIONS
7222 -----------------------------------------------------------------------------
7225 NAME: adaptation_service_set
7226 TYPE: adaptation_service_set_type
7227 IFDEF: USE_ADAPTATION
7232 Configures an ordered set of similar, redundant services. This is
7233 useful when hot standby or backup adaptation servers are available.
7235 adaptation_service_set set_name service_name1 service_name2 ...
7237 The named services are used in the set declaration order. The first
7238 applicable adaptation service from the set is used first. The next
7239 applicable service is tried if and only if the transaction with the
7240 previous service fails and the message waiting to be adapted is still
7243 When adaptation starts, broken services are ignored as if they were
7244 not a part of the set. A broken service is a down optional service.
7246 The services in a set must be attached to the same vectoring point
7247 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7249 If all services in a set are optional then adaptation failures are
7250 bypassable. If all services in the set are essential, then a
7251 transaction failure with one service may still be retried using
7252 another service from the set, but when all services fail, the master
7253 transaction fails as well.
7255 A set may contain a mix of optional and essential services, but that
7256 is likely to lead to surprising results because broken services become
7257 ignored (see above), making previously bypassable failures fatal.
7258 Technically, it is the bypassability of the last failed service that
7261 See also: adaptation_access adaptation_service_chain
7264 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7265 adaptation service_set svcLogger loggerLocal loggerRemote
7268 NAME: adaptation_service_chain
7269 TYPE: adaptation_service_chain_type
7270 IFDEF: USE_ADAPTATION
7275 Configures a list of complementary services that will be applied
7276 one-by-one, forming an adaptation chain or pipeline. This is useful
7277 when Squid must perform different adaptations on the same message.
7279 adaptation_service_chain chain_name service_name1 svc_name2 ...
7281 The named services are used in the chain declaration order. The first
7282 applicable adaptation service from the chain is used first. The next
7283 applicable service is applied to the successful adaptation results of
7284 the previous service in the chain.
7286 When adaptation starts, broken services are ignored as if they were
7287 not a part of the chain. A broken service is a down optional service.
7289 Request satisfaction terminates the adaptation chain because Squid
7290 does not currently allow declaration of RESPMOD services at the
7291 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7293 The services in a chain must be attached to the same vectoring point
7294 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7296 A chain may contain a mix of optional and essential services. If an
7297 essential adaptation fails (or the failure cannot be bypassed for
7298 other reasons), the master transaction fails. Otherwise, the failure
7299 is bypassed as if the failed adaptation service was not in the chain.
7301 See also: adaptation_access adaptation_service_set
7304 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
7307 NAME: adaptation_access
7308 TYPE: adaptation_access_type
7309 IFDEF: USE_ADAPTATION
7313 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
7315 adaptation_access service_name allow|deny [!]aclname...
7316 adaptation_access set_name allow|deny [!]aclname...
7318 At each supported vectoring point, the adaptation_access
7319 statements are processed in the order they appear in this
7320 configuration file. Statements pointing to the following services
7321 are ignored (i.e., skipped without checking their ACL):
7323 - services serving different vectoring points
7324 - "broken-but-bypassable" services
7325 - "up" services configured to ignore such transactions
7326 (e.g., based on the ICAP Transfer-Ignore header).
7328 When a set_name is used, all services in the set are checked
7329 using the same rules, to find the first applicable one. See
7330 adaptation_service_set for details.
7332 If an access list is checked and there is a match, the
7333 processing stops: For an "allow" rule, the corresponding
7334 adaptation service is used for the transaction. For a "deny"
7335 rule, no adaptation service is activated.
7337 It is currently not possible to apply more than one adaptation
7338 service at the same vectoring point to the same HTTP transaction.
7340 See also: icap_service and ecap_service
7343 adaptation_access service_1 allow all
7346 NAME: adaptation_service_iteration_limit
7348 IFDEF: USE_ADAPTATION
7349 LOC: Adaptation::Config::service_iteration_limit
7352 Limits the number of iterations allowed when applying adaptation
7353 services to a message. If your longest adaptation set or chain
7354 may have more than 16 services, increase the limit beyond its
7355 default value of 16. If detecting infinite iteration loops sooner
7356 is critical, make the iteration limit match the actual number
7357 of services in your longest adaptation set or chain.
7359 Infinite adaptation loops are most likely with routing services.
7361 See also: icap_service routing=1
7364 NAME: adaptation_masterx_shared_names
7366 IFDEF: USE_ADAPTATION
7367 LOC: Adaptation::Config::masterx_shared_name
7370 For each master transaction (i.e., the HTTP request and response
7371 sequence, including all related ICAP and eCAP exchanges), Squid
7372 maintains a table of metadata. The table entries are (name, value)
7373 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7374 with the master transaction.
7376 This option specifies the table entry names that Squid must accept
7377 from and forward to the adaptation transactions.
7379 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7380 shared table by returning an ICAP header field with a name
7381 specified in adaptation_masterx_shared_names.
7383 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7384 shared table by implementing the libecap::visitEachOption() API
7385 to provide an option with a name specified in
7386 adaptation_masterx_shared_names.
7388 Squid will store and forward the set entry to subsequent adaptation
7389 transactions within the same master transaction scope.
7391 Only one shared entry name is supported at this time.
7394 # share authentication information among ICAP services
7395 adaptation_masterx_shared_names X-Subscriber-ID
7398 NAME: adaptation_meta
7399 TYPE: adaptation_meta_type
7400 IFDEF: USE_ADAPTATION
7401 LOC: Adaptation::Config::metaHeaders
7404 This option allows Squid administrator to add custom ICAP request
7405 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7406 Use it to pass custom authentication tokens and other
7407 transaction-state related meta information to an ICAP/eCAP service.
7409 The addition of a meta header is ACL-driven:
7410 adaptation_meta name value [!]aclname ...
7412 Processing for a given header name stops after the first ACL list match.
7413 Thus, it is impossible to add two headers with the same name. If no ACL
7414 lists match for a given header name, no such header is added. For
7417 # do not debug transactions except for those that need debugging
7418 adaptation_meta X-Debug 1 needs_debugging
7420 # log all transactions except for those that must remain secret
7421 adaptation_meta X-Log 1 !keep_secret
7423 # mark transactions from users in the "G 1" group
7424 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7426 The "value" parameter may be a regular squid.conf token or a "double
7427 quoted string". Within the quoted string, use backslash (\) to escape
7428 any character, which is currently only useful for escaping backslashes
7429 and double quotes. For example,
7430 "this string has one backslash (\\) and two \"quotes\""
7436 LOC: Adaptation::Icap::TheConfig.repeat
7437 DEFAULT_IF_NONE: deny all
7439 This ACL determines which retriable ICAP transactions are
7440 retried. Transactions that received a complete ICAP response
7441 and did not have to consume or produce HTTP bodies to receive
7442 that response are usually retriable.
7444 icap_retry allow|deny [!]aclname ...
7446 Squid automatically retries some ICAP I/O timeouts and errors
7447 due to persistent connection race conditions.
7449 See also: icap_retry_limit
7452 NAME: icap_retry_limit
7455 LOC: Adaptation::Icap::TheConfig.repeat_limit
7458 Limits the number of retries allowed. When set to zero (default),
7459 no retries are allowed.
7461 Communication errors due to persistent connection race
7462 conditions are unavoidable, automatically retried, and do not
7463 count against this limit.
7465 See also: icap_retry
7471 -----------------------------------------------------------------------------
7474 NAME: check_hostnames
7477 LOC: Config.onoff.check_hostnames
7479 For security and stability reasons Squid can check
7480 hostnames for Internet standard RFC compliance. If you want
7481 Squid to perform these checks turn this directive on.
7484 NAME: allow_underscore
7487 LOC: Config.onoff.allow_underscore
7489 Underscore characters is not strictly allowed in Internet hostnames
7490 but nevertheless used by many sites. Set this to off if you want
7491 Squid to be strict about the standard.
7492 This check is performed only when check_hostnames is set to on.
7495 NAME: cache_dns_program
7497 IFDEF: USE_DNSHELPER
7498 DEFAULT: @DEFAULT_DNSSERVER@
7499 LOC: Config.Program.dnsserver
7501 Specify the location of the executable for dnslookup process.
7505 TYPE: HelperChildConfig
7506 IFDEF: USE_DNSHELPER
7507 DEFAULT: 32 startup=1 idle=1
7508 LOC: Config.dnsChildren
7510 The maximum number of processes spawn to service DNS name lookups.
7511 If you limit it too few Squid will have to wait for them to process
7512 a backlog of requests, slowing it down. If you allow too many they
7513 will use RAM and other system resources noticably.
7514 The maximum this may be safely set to is 32.
7516 The startup= and idle= options allow some measure of skew in your
7521 Sets a minimum of how many processes are to be spawned when Squid
7522 starts or reconfigures. When set to zero the first request will
7523 cause spawning of the first child process to handle it.
7525 Starting too few will cause an initial slowdown in traffic as Squid
7526 attempts to simultaneously spawn enough processes to cope.
7530 Sets a minimum of how many processes Squid is to try and keep available
7531 at all times. When traffic begins to rise above what the existing
7532 processes can handle this many more will be spawned up to the maximum
7533 configured. A minimum setting of 1 is required.
7536 NAME: dns_retransmit_interval
7539 LOC: Config.Timeout.idns_retransmit
7540 IFDEF: !USE_DNSHELPER
7542 Initial retransmit interval for DNS queries. The interval is
7543 doubled each time all configured DNS servers have been tried.
7549 LOC: Config.Timeout.idns_query
7550 IFDEF: !USE_DNSHELPER
7552 DNS Query timeout. If no response is received to a DNS query
7553 within this time all DNS servers for the queried domain
7554 are assumed to be unavailable.
7557 NAME: dns_packet_max
7560 LOC: Config.dns.packet_max
7561 IFDEF: !USE_DNSHELPER
7563 Maximum number of bytes packet size to advertise via EDNS.
7564 Set to "none" to disable EDNS large packet support.
7566 For legacy reasons DNS UDP replies will default to 512 bytes which
7567 is too small for many responses. EDNS provides a means for Squid to
7568 negotiate receiving larger responses back immediately without having
7569 to failover with repeat requests. Responses larger than this limit
7570 will retain the old behaviour of failover to TCP DNS.
7572 Squid has no real fixed limit internally, but allowing packet sizes
7573 over 1500 bytes requires network jumbogram support and is usually not
7576 WARNING: The RFC also indicates that some older resolvers will reply
7577 with failure of the whole request if the extension is added. Some
7578 resolvers have already been identified which will reply with mangled
7579 EDNS response on occasion. Usually in response to many-KB jumbogram
7580 sizes being advertised by Squid.
7581 Squid will currently treat these both as an unable-to-resolve domain
7582 even if it would be resolvable without EDNS.
7589 LOC: Config.onoff.res_defnames
7591 Normally the RES_DEFNAMES resolver option is disabled
7592 (see res_init(3)). This prevents caches in a hierarchy
7593 from interpreting single-component hostnames locally. To allow
7594 Squid to handle single-component names, enable this option.
7597 NAME: dns_nameservers
7600 LOC: Config.dns_nameservers
7602 Use this if you want to specify a list of DNS name servers
7603 (IP addresses) to use instead of those given in your
7604 /etc/resolv.conf file.
7605 On Windows platforms, if no value is specified here or in
7606 the /etc/resolv.conf file, the list of DNS name servers are
7607 taken from the Windows registry, both static and dynamic DHCP
7608 configurations are supported.
7610 Example: dns_nameservers 10.0.0.1 192.172.0.4
7615 DEFAULT: @DEFAULT_HOSTS@
7616 LOC: Config.etcHostsPath
7618 Location of the host-local IP name-address associations
7619 database. Most Operating Systems have such a file on different
7621 - Un*X & Linux: /etc/hosts
7622 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7623 (%SystemRoot% value install default is c:\winnt)
7624 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7625 (%SystemRoot% value install default is c:\windows)
7626 - Windows 9x/Me: %windir%\hosts
7627 (%windir% value is usually c:\windows)
7628 - Cygwin: /etc/hosts
7630 The file contains newline-separated definitions, in the
7631 form ip_address_in_dotted_form name [name ...] names are
7632 whitespace-separated. Lines beginning with an hash (#)
7633 character are comments.
7635 The file is checked at startup and upon configuration.
7636 If set to 'none', it won't be checked.
7637 If append_domain is used, that domain will be added to
7638 domain-local (i.e. not containing any dot character) host
7644 LOC: Config.appendDomain
7647 Appends local domain name to hostnames without any dots in
7648 them. append_domain must begin with a period.
7650 Be warned there are now Internet names with no dots in
7651 them using only top-domain names, so setting this may
7652 cause some Internet sites to become unavailable.
7655 append_domain .yourdomain.com
7658 NAME: ignore_unknown_nameservers
7660 LOC: Config.onoff.ignore_unknown_nameservers
7662 IFDEF: !USE_DNSHELPER
7664 By default Squid checks that DNS responses are received
7665 from the same IP addresses they are sent to. If they
7666 don't match, Squid ignores the response and writes a warning
7667 message to cache.log. You can allow responses from unknown
7668 nameservers by setting this option to 'off'.
7674 LOC: Config.dns.v4_first
7675 IFDEF: !USE_DNSHELPER
7677 With the IPv6 Internet being as fast or faster than IPv4 Internet
7678 for most networks Squid prefers to contact websites over IPv6.
7680 This option reverses the order of preference to make Squid contact
7681 dual-stack websites over IPv4 first. Squid will still perform both
7682 IPv6 and IPv4 DNS lookups before connecting.
7685 This option will restrict the situations under which IPv6
7686 connectivity is used (and tested). Hiding network problems
7687 which would otherwise be detected and warned about.
7691 COMMENT: (number of entries)
7694 LOC: Config.ipcache.size
7701 LOC: Config.ipcache.low
7708 LOC: Config.ipcache.high
7710 The size, low-, and high-water marks for the IP cache.
7713 NAME: fqdncache_size
7714 COMMENT: (number of entries)
7717 LOC: Config.fqdncache.size
7719 Maximum number of FQDN cache entries.
7724 -----------------------------------------------------------------------------
7731 LOC: Config.onoff.mem_pools
7733 If set, Squid will keep pools of allocated (but unused) memory
7734 available for future use. If memory is a premium on your
7735 system and you believe your malloc library outperforms Squid
7736 routines, disable this.
7739 NAME: memory_pools_limit
7743 LOC: Config.MemPools.limit
7745 Used only with memory_pools on:
7746 memory_pools_limit 50 MB
7748 If set to a non-zero value, Squid will keep at most the specified
7749 limit of allocated (but unused) memory in memory pools. All free()
7750 requests that exceed this limit will be handled by your malloc
7751 library. Squid does not pre-allocate any memory, just safe-keeps
7752 objects that otherwise would be free()d. Thus, it is safe to set
7753 memory_pools_limit to a reasonably high value even if your
7754 configuration will use less memory.
7756 If set to none, Squid will keep all memory it can. That is, there
7757 will be no limit on the total amount of memory used for safe-keeping.
7759 To disable memory allocation optimization, do not set
7760 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
7762 An overhead for maintaining memory pools is not taken into account
7763 when the limit is checked. This overhead is close to four bytes per
7764 object kept. However, pools may actually _save_ memory because of
7765 reduced memory thrashing in your malloc library.
7769 COMMENT: on|off|transparent|truncate|delete
7772 LOC: opt_forwarded_for
7774 If set to "on", Squid will append your client's IP address
7775 in the HTTP requests it forwards. By default it looks like:
7777 X-Forwarded-For: 192.1.2.3
7779 If set to "off", it will appear as
7781 X-Forwarded-For: unknown
7783 If set to "transparent", Squid will not alter the
7784 X-Forwarded-For header in any way.
7786 If set to "delete", Squid will delete the entire
7787 X-Forwarded-For header.
7789 If set to "truncate", Squid will remove all existing
7790 X-Forwarded-For entries, and place the client IP as the sole entry.
7793 NAME: cachemgr_passwd
7794 TYPE: cachemgrpasswd
7796 LOC: Config.passwd_list
7798 Specify passwords for cachemgr operations.
7800 Usage: cachemgr_passwd password action action ...
7802 Some valid actions are (see cache manager menu for a full list):
7842 * Indicates actions which will not be performed without a
7843 valid password, others can be performed if not listed here.
7845 To disable an action, set the password to "disable".
7846 To allow performing an action without a password, set the
7849 Use the keyword "all" to set the same password for all actions.
7852 cachemgr_passwd secret shutdown
7853 cachemgr_passwd lesssssssecret info stats/objects
7854 cachemgr_passwd disable all
7861 LOC: Config.onoff.client_db
7863 If you want to disable collecting per-client statistics,
7864 turn off client_db here.
7867 NAME: refresh_all_ims
7871 LOC: Config.onoff.refresh_all_ims
7873 When you enable this option, squid will always check
7874 the origin server for an update when a client sends an
7875 If-Modified-Since request. Many browsers use IMS
7876 requests when the user requests a reload, and this
7877 ensures those clients receive the latest version.
7879 By default (off), squid may return a Not Modified response
7880 based on the age of the cached version.
7883 NAME: reload_into_ims
7884 IFDEF: USE_HTTP_VIOLATIONS
7888 LOC: Config.onoff.reload_into_ims
7890 When you enable this option, client no-cache or ``reload''
7891 requests will be changed to If-Modified-Since requests.
7892 Doing this VIOLATES the HTTP standard. Enabling this
7893 feature could make you liable for problems which it
7896 see also refresh_pattern for a more selective approach.
7899 NAME: connect_retries
7901 LOC: Config.connect_retries
7904 This sets the maximum number of connection attempts made for each
7905 TCP connection. The connect_retries attempts must all still
7906 complete within the connection timeout period.
7908 The default is not to re-try if the first connection attempt fails.
7909 The (not recommended) maximum is 10 tries.
7911 A warning message will be generated if it is set to a too-high
7912 value and the configured value will be over-ridden.
7914 Note: These re-tries are in addition to forward_max_tries
7915 which limit how many different addresses may be tried to find
7919 NAME: retry_on_error
7921 LOC: Config.retry.onerror
7924 If set to ON Squid will automatically retry requests when
7925 receiving an error response with status 403 (Forbidden),
7926 500 (Internal Error), 501 or 503 (Service not available).
7927 Status 502 and 504 (Gateway errors) are always retried.
7929 This is mainly useful if you are in a complex cache hierarchy to
7930 work around access control errors.
7932 NOTE: This retry will attempt to find another working destination.
7933 Which is different from the server which just failed.
7936 NAME: as_whois_server
7938 LOC: Config.as_whois_server
7939 DEFAULT: whois.ra.net
7941 WHOIS server to query for AS numbers. NOTE: AS numbers are
7942 queried only when Squid starts up, not for every request.
7947 LOC: Config.onoff.offline
7950 Enable this option and Squid will never try to validate cached
7954 NAME: uri_whitespace
7955 TYPE: uri_whitespace
7956 LOC: Config.uri_whitespace
7959 What to do with requests that have whitespace characters in the
7962 strip: The whitespace characters are stripped out of the URL.
7963 This is the behavior recommended by RFC2396.
7964 deny: The request is denied. The user receives an "Invalid
7966 allow: The request is allowed and the URI is not changed. The
7967 whitespace characters remain in the URI. Note the
7968 whitespace is passed to redirector processes if they
7970 encode: The request is allowed and the whitespace characters are
7971 encoded according to RFC1738. This could be considered
7972 a violation of the HTTP/1.1
7973 RFC because proxies are not allowed to rewrite URI's.
7974 chop: The request is allowed and the URI is chopped at the
7975 first whitespace. This might also be considered a
7981 LOC: Config.chroot_dir
7984 Specifies a directory where Squid should do a chroot() while
7985 initializing. This also causes Squid to fully drop root
7986 privileges after initializing. This means, for example, if you
7987 use a HTTP port less than 1024 and try to reconfigure, you may
7988 get an error saying that Squid can not open the port.
7991 NAME: balance_on_multiple_ip
7993 LOC: Config.onoff.balance_on_multiple_ip
7996 Modern IP resolvers in squid sort lookup results by preferred access.
7997 By default squid will use these IP in order and only rotates to
7998 the next listed when the most preffered fails.
8000 Some load balancing servers based on round robin DNS have been
8001 found not to preserve user session state across requests
8002 to different IP addresses.
8004 Enabling this directive Squid rotates IP's per request.
8007 NAME: pipeline_prefetch
8009 LOC: Config.onoff.pipeline_prefetch
8012 To boost the performance of pipelined requests to closer
8013 match that of a non-proxied environment Squid can try to fetch
8014 up to two requests in parallel from a pipeline.
8016 Defaults to off for bandwidth management and access logging
8019 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8022 NAME: high_response_time_warning
8025 LOC: Config.warnings.high_rptm
8028 If the one-minute median response time exceeds this value,
8029 Squid prints a WARNING with debug level 0 to get the
8030 administrators attention. The value is in milliseconds.
8033 NAME: high_page_fault_warning
8035 LOC: Config.warnings.high_pf
8038 If the one-minute average page fault rate exceeds this
8039 value, Squid prints a WARNING with debug level 0 to get
8040 the administrators attention. The value is in page faults
8044 NAME: high_memory_warning
8046 LOC: Config.warnings.high_memory
8049 If the memory usage (as determined by mallinfo) exceeds
8050 this amount, Squid prints a WARNING with debug level 0 to get
8051 the administrators attention.
8054 NAME: sleep_after_fork
8055 COMMENT: (microseconds)
8057 LOC: Config.sleep_after_fork
8060 When this is set to a non-zero value, the main Squid process
8061 sleeps the specified number of microseconds after a fork()
8062 system call. This sleep may help the situation where your
8063 system reports fork() failures due to lack of (virtual)
8064 memory. Note, however, if you have a lot of child
8065 processes, these sleep delays will add up and your
8066 Squid will not service requests for some amount of time
8067 until all the child processes have been started.
8068 On Windows value less then 1000 (1 milliseconds) are
8072 NAME: windows_ipaddrchangemonitor
8073 IFDEF: _SQUID_MSWIN_
8077 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8079 On Windows Squid by default will monitor IP address changes and will
8080 reconfigure itself after any detected event. This is very useful for
8081 proxies connected to internet with dial-up interfaces.
8082 In some cases (a Proxy server acting as VPN gateway is one) it could be
8083 desiderable to disable this behaviour setting this to 'off'.
8084 Note: after changing this, Squid service must be restarted.
8089 IFDEF: USE_SQUID_EUI
8091 LOC: Eui::TheConfig.euiLookup
8093 Whether to lookup the EUI or MAC address of a connected client.
8096 NAME: max_filedescriptors max_filedesc
8099 LOC: Config.max_filedescriptors
8101 The maximum number of filedescriptors supported.
8103 The default "0" means Squid inherits the current ulimit setting.
8105 Note: Changing this requires a restart of Squid. Also
8106 not all comm loops supports large values.
8114 Number of main Squid processes or "workers" to fork and maintain.
8115 0: "no daemon" mode, like running "squid -N ..."
8116 1: "no SMP" mode, start one main Squid process daemon (default)
8117 N: start N main Squid process daemons (i.e., SMP mode)
8119 In SMP mode, each worker does nearly all what a single Squid daemon
8120 does (e.g., listen on http_port and forward HTTP requests).
8123 NAME: cpu_affinity_map
8124 TYPE: CpuAffinityMap
8125 LOC: Config.cpuAffinityMap
8128 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8130 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8132 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8134 affects processes 1 through 4 only and places them on the first
8135 four even cores, starting with core #1.
8137 CPU cores are numbered starting from 1. Requires support for
8138 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8140 Multiple cpu_affinity_map options are merged.