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 default Squid configuration file. You may wish
34 to look at the Squid home page (http://www.squid-cache.org/)
35 for the FAQ and other documentation.
37 The default Squid config file shows what the defaults for
38 various options happen to be. If you don't need to change the
39 default, you shouldn't uncomment the line. Doing so may cause
40 run-time problems. In some cases "none" refers to no default
41 setting at all, while in other cases it refers to a valid
42 option - the comments for that keyword indicate if this is the
48 Configuration options can be included using the "include" directive.
49 Include takes a list of files to include. Quoting and wildcards is
54 include /path/to/included/file/squid.acl.config
56 Includes can be nested up to a hard-coded depth of 16 levels.
57 This arbitrary restriction is to prevent recursive include references
58 from causing Squid entering an infinite loop whilst trying to load
62 Conditional configuration
64 If-statements can be used to make configuration directives
68 ... regular configuration directives ...
70 ... regular configuration directives ...]
73 The else part is optional. The keywords "if", "else", and "endif"
74 must be typed on their own lines, as if they were regular
75 configuration directives.
77 These individual conditions types are supported:
80 Always evaluates to true.
82 Always evaluates to false.
84 Equality comparison of two integer numbers.
89 The following SMP-related preprocessor macros can be used.
91 ${process_name} expands to the current Squid process "name"
92 (e.g., squid1, squid2, or cache1).
94 ${process_number} expands to the current Squid process
95 identifier, which is an integer number (e.g., 1, 2, 3) unique
96 across all Squid processes.
99 # Options Removed in 3.2
100 NAME: ignore_expect_100
103 Remove this line. The HTTP/1.1 feature is now fully supported by default.
109 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
112 NAME: url_rewrite_concurrency
115 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
118 # Options Removed in 3.1
122 Remove this line. DNS is no longer tested on startup.
125 NAME: extension_methods
128 Remove this line. All valid methods for HTTP are accepted by default.
131 # 2.7 Options Removed/Replaced in 3.1
139 Remove this line. HTTP/1.1 is supported by default.
142 NAME: upgrade_http0.9
145 Remove this line. ICY/1.0 streaming protocol is supported by default.
148 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
151 Alter these entries. Use the qos_flows directive instead.
154 # Options Removed in 3.0
158 Since squid-3.0 replace with request_header_access or reply_header_access
159 depending on whether you wish to match client requests or server replies.
162 NAME: httpd_accel_no_pmtu_disc
165 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
169 OPTIONS FOR AUTHENTICATION
170 -----------------------------------------------------------------------------
179 This is used to define parameters for the various authentication
180 schemes supported by Squid.
182 format: auth_param scheme parameter [setting]
184 The order in which authentication schemes are presented to the client is
185 dependent on the order the scheme first appears in config file. IE
186 has a bug (it's not RFC 2617 compliant) in that it will use the basic
187 scheme if basic is the first entry presented, even if more secure
188 schemes are presented. For now use the order in the recommended
189 settings section below. If other browsers have difficulties (don't
190 recognize the schemes offered even if you are using basic) either
191 put basic first, or disable the other schemes (by commenting out their
194 Once an authentication scheme is fully configured, it can only be
195 shutdown by shutting squid down and restarting. Changes can be made on
196 the fly and activated with a reconfigure. I.E. You can change to a
197 different helper, but not unconfigure the helper completely.
199 Please note that while this directive defines how Squid processes
200 authentication it does not automatically activate authentication.
201 To use authentication you must in addition make use of ACLs based
202 on login name in http_access (proxy_auth, proxy_auth_regex or
203 external with %LOGIN used in the format tag). The browser will be
204 challenged for authentication on the first such acl encountered
205 in http_access processing and will also be re-challenged for new
206 login credentials if the request is being denied by a proxy_auth
209 WARNING: authentication can't be used in a transparently intercepting
210 proxy as the client then thinks it is talking to an origin server and
211 not the proxy. This is a limitation of bending the TCP/IP protocol to
212 transparently intercepting port 80, not a limitation in Squid.
213 Ports flagged 'transparent', 'intercept', or 'tproxy' have
214 authentication disabled.
216 === Parameters for the basic scheme follow. ===
219 Specify the command for the external authenticator. Such a program
220 reads a line containing "username password" and replies "OK" or
221 "ERR" in an endless loop. "ERR" responses may optionally be followed
222 by a error description available as %m in the returned error page.
223 If you use an authenticator, make sure you have 1 acl of type
226 By default, the basic authentication scheme is not used unless a
227 program is specified.
229 If you want to use the traditional NCSA proxy authentication, set
230 this line to something like
232 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
235 HTTP uses iso-latin-1 as characterset, while some authentication
236 backends such as LDAP expects UTF-8. If this is set to on Squid will
237 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
238 username & password to the helper.
240 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
241 The maximum number of authenticator processes to spawn. If you start too few
242 Squid will have to wait for them to process a backlog of credential
243 verifications, slowing it down. When password verifications are
244 done via a (slow) network you are likely to need lots of
245 authenticator processes.
247 The startup= and idle= options permit some skew in the exact amount
248 run. A minimum of startup=N will begin during startup and reconfigure.
249 Squid will start more in groups of up to idle=N in an attempt to meet
250 traffic needs and to keep idle=N free above those traffic needs up to
253 The concurrency= option sets the number of concurrent requests the
254 helper can process. The default of 0 is used for helpers who only
255 supports one request at a time. Setting this to a number greater than
256 0 changes the protocol used to include a channel number first on the
257 request/response line, allowing multiple requests to be sent to the
258 same helper in parallell without wating for the response.
259 Must not be set unless it's known the helper supports this.
261 auth_param basic children 20 startup=0 idle=1
264 Specifies the realm name which is to be reported to the
265 client for the basic proxy authentication scheme (part of
266 the text the user will see when prompted their username and
267 password). There is no default.
268 auth_param basic realm Squid proxy-caching web server
270 "credentialsttl" timetolive
271 Specifies how long squid assumes an externally validated
272 username:password pair is valid for - in other words how
273 often the helper program is called for that user. Set this
274 low to force revalidation with short lived passwords. Note
275 setting this high does not impact your susceptibility
276 to replay attacks unless you are using an one-time password
277 system (such as SecureID). If you are using such a system,
278 you will be vulnerable to replay attacks unless you also
279 use the max_user_ip ACL in an http_access rule.
281 "casesensitive" on|off
282 Specifies if usernames are case sensitive. Most user databases are
283 case insensitive allowing the same username to be spelled using both
284 lower and upper case letters, but some are case sensitive. This
285 makes a big difference for user_max_ip ACL processing and similar.
286 auth_param basic casesensitive off
288 === Parameters for the digest scheme follow ===
291 Specify the command for the external authenticator. Such
292 a program reads a line containing "username":"realm" and
293 replies with the appropriate H(A1) value hex encoded or
294 ERR if the user (or his H(A1) hash) does not exists.
295 See rfc 2616 for the definition of H(A1).
296 "ERR" responses may optionally be followed by a error description
297 available as %m in the returned error page.
299 By default, the digest authentication scheme is not used unless a
300 program is specified.
302 If you want to use a digest authenticator, set this line to
305 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
308 HTTP uses iso-latin-1 as characterset, while some authentication
309 backends such as LDAP expects UTF-8. If this is set to on Squid will
310 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
311 username & password to the helper.
313 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
314 The maximum number of authenticator processes to spawn (default 5).
315 If you start too few Squid will have to wait for them to
316 process a backlog of H(A1) calculations, slowing it down.
317 When the H(A1) calculations are done via a (slow) network
318 you are likely to need lots of authenticator processes.
320 The startup= and idle= options permit some skew in the exact amount
321 run. A minimum of startup=N will begin during startup and reconfigure.
322 Squid will start more in groups of up to idle=N in an attempt to meet
323 traffic needs and to keep idle=N free above those traffic needs up to
326 The concurrency= option sets the number of concurrent requests the
327 helper can process. The default of 0 is used for helpers who only
328 supports one request at a time. Setting this to a number greater than
329 0 changes the protocol used to include a channel number first on the
330 request/response line, allowing multiple requests to be sent to the
331 same helper in parallell without wating for the response.
332 Must not be set unless it's known the helper supports this.
334 auth_param digest children 20 startup=0 idle=1
337 Specifies the realm name which is to be reported to the
338 client for the digest proxy authentication scheme (part of
339 the text the user will see when prompted their username and
340 password). There is no default.
341 auth_param digest realm Squid proxy-caching web server
343 "nonce_garbage_interval" timeinterval
344 Specifies the interval that nonces that have been issued
345 to client_agent's are checked for validity.
347 "nonce_max_duration" timeinterval
348 Specifies the maximum length of time a given nonce will be
351 "nonce_max_count" number
352 Specifies the maximum number of times a given nonce can be
355 "nonce_strictness" on|off
356 Determines if squid requires strict increment-by-1 behavior
357 for nonce counts, or just incrementing (off - for use when
358 useragents generate nonce counts that occasionally miss 1
359 (ie, 1,2,4,6)). Default off.
361 "check_nonce_count" on|off
362 This directive if set to off can disable the nonce count check
363 completely to work around buggy digest qop implementations in
364 certain mainstream browser versions. Default on to check the
365 nonce count to protect from authentication replay attacks.
367 "post_workaround" on|off
368 This is a workaround to certain buggy browsers who sends
369 an incorrect request digest in POST requests when reusing
370 the same nonce as acquired earlier on a GET request.
372 === NTLM scheme options follow ===
375 Specify the command for the external NTLM authenticator.
376 Such a program reads exchanged NTLMSSP packets with
377 the browser via Squid until authentication is completed.
378 If you use an NTLM authenticator, make sure you have 1 acl
379 of type proxy_auth. By default, the NTLM authenticator_program
382 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
384 "children" numberofchildren [startup=N] [idle=N]
385 The maximum number of authenticator processes to spawn (default 5).
386 If you start too few Squid will have to wait for them to
387 process a backlog of credential verifications, slowing it
388 down. When credential verifications are done via a (slow)
389 network you are likely to need lots of authenticator
392 The startup= and idle= options permit some skew in the exact amount
393 run. A minimum of startup=N will begin during startup and reconfigure.
394 Squid will start more in groups of up to idle=N in an attempt to meet
395 traffic needs and to keep idle=N free above those traffic needs up to
398 auth_param ntlm children 20 startup=0 idle=1
401 If you experience problems with PUT/POST requests when using the
402 Negotiate authentication scheme then you can try setting this to
403 off. This will cause Squid to forcibly close the connection on
404 the initial requests where the browser asks which schemes are
405 supported by the proxy.
407 auth_param ntlm keep_alive on
409 === Options for configuring the NEGOTIATE auth-scheme follow ===
412 Specify the command for the external Negotiate authenticator.
413 This protocol is used in Microsoft Active-Directory enabled setups with
414 the Microsoft Internet Explorer or Mozilla Firefox browsers.
415 Its main purpose is to exchange credentials with the Squid proxy
416 using the Kerberos mechanisms.
417 If you use a Negotiate authenticator, make sure you have at least
418 one acl of type proxy_auth active. By default, the negotiate
419 authenticator_program is not used.
420 The only supported program for this role is the ntlm_auth
421 program distributed as part of Samba, version 4 or later.
423 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
425 "children" numberofchildren [startup=N] [idle=N]
426 The maximum number of authenticator processes to spawn (default 5).
427 If you start too few Squid will have to wait for them to
428 process a backlog of credential verifications, slowing it
429 down. When crendential verifications are done via a (slow)
430 network you are likely to need lots of authenticator
433 The startup= and idle= options permit some skew in the exact amount
434 run. A minimum of startup=N will begin during startup and reconfigure.
435 Squid will start more in groups of up to idle=N in an attempt to meet
436 traffic needs and to keep idle=N free above those traffic needs up to
439 auth_param negotiate children 20 startup=0 idle=1
442 If you experience problems with PUT/POST requests when using the
443 Negotiate authentication scheme then you can try setting this to
444 off. This will cause Squid to forcibly close the connection on
445 the initial requests where the browser asks which schemes are
446 supported by the proxy.
448 auth_param negotiate keep_alive on
453 #Recommended minimum configuration per scheme:
454 #auth_param negotiate program <uncomment and complete this line to activate>
455 #auth_param negotiate children 20 startup=0 idle=1
456 #auth_param negotiate keep_alive on
458 #auth_param ntlm program <uncomment and complete this line to activate>
459 #auth_param ntlm children 20 startup=0 idle=1
460 #auth_param ntlm keep_alive on
462 #auth_param digest program <uncomment and complete this line>
463 #auth_param digest children 20 startup=0 idle=1
464 #auth_param digest realm Squid proxy-caching web server
465 #auth_param digest nonce_garbage_interval 5 minutes
466 #auth_param digest nonce_max_duration 30 minutes
467 #auth_param digest nonce_max_count 50
469 #auth_param basic program <uncomment and complete this line>
470 #auth_param basic children 5 startup=5 idle=1
471 #auth_param basic realm Squid proxy-caching web server
472 #auth_param basic credentialsttl 2 hours
475 NAME: authenticate_cache_garbage_interval
478 LOC: Config.authenticateGCInterval
480 The time period between garbage collection across the username cache.
481 This is a tradeoff between memory utilization (long intervals - say
482 2 days) and CPU (short intervals - say 1 minute). Only change if you
486 NAME: authenticate_ttl
489 LOC: Config.authenticateTTL
491 The time a user & their credentials stay in the logged in
492 user cache since their last request. When the garbage
493 interval passes, all user credentials that have passed their
494 TTL are removed from memory.
497 NAME: authenticate_ip_ttl
499 LOC: Config.authenticateIpTTL
502 If you use proxy authentication and the 'max_user_ip' ACL,
503 this directive controls how long Squid remembers the IP
504 addresses associated with each user. Use a small value
505 (e.g., 60 seconds) if your users might change addresses
506 quickly, as is the case with dialups. You might be safe
507 using a larger value (e.g., 2 hours) in a corporate LAN
508 environment with relatively static address assignments.
513 -----------------------------------------------------------------------------
516 NAME: external_acl_type
517 TYPE: externalAclHelper
518 LOC: Config.externalAclHelperList
521 This option defines external acl classes using a helper program
522 to look up the status
524 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
528 ttl=n TTL in seconds for cached results (defaults to 3600
531 TTL for cached negative lookups (default same
534 Maximum number of acl helper processes spawned to service
535 external acl lookups of this type. (default 20)
537 Minimum number of acl helper processes to spawn during
538 startup and reconfigure to service external acl lookups
539 of this type. (default 0)
541 Number of acl helper processes to keep ahead of traffic
542 loads. Squid will spawn this many at once whenever load
543 rises above the capabilities of existing processes.
544 Up to the value of children-max. (default 1)
545 concurrency=n concurrency level per process. Only used with helpers
546 capable of processing more than one query at a time.
547 cache=n limit the result cache size, default is unbounded.
548 grace=n Percentage remaining of TTL where a refresh of a
549 cached entry should be initiated without needing to
550 wait for a new reply. (default is for no grace period)
551 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
552 ipv4 / ipv6 IP-mode used to communicate to this helper.
553 For compatability with older configurations and helpers
554 the default is currently 'ipv4'.
556 FORMAT specifications
558 %LOGIN Authenticated user login name
559 %EXT_USER Username from previous external acl
560 %EXT_LOG Log details from previous external acl
561 %EXT_TAG Tag from previous external acl
562 %IDENT Ident user name
564 %SRCPORT Client source port
567 %PROTO Requested protocol
569 %PATH Requested URL path
570 %METHOD Request method
571 %MYADDR Squid interface address
572 %MYPORT Squid http_port number
573 %PATH Requested URL-path (including query-string if any)
574 %USER_CERT SSL User certificate in PEM format
575 %USER_CERTCHAIN SSL User certificate chain in PEM format
576 %USER_CERT_xx SSL User certificate subject attribute xx
577 %USER_CA_xx SSL User certificate issuer attribute xx
579 %>{Header} HTTP request header "Header"
581 HTTP request header "Hdr" list member "member"
583 HTTP request header list member using ; as
584 list separator. ; can be any non-alphanumeric
587 %<{Header} HTTP reply header "Header"
589 HTTP reply header "Hdr" list member "member"
591 HTTP reply header list member using ; as
592 list separator. ; can be any non-alphanumeric
595 %% The percent sign. Useful for helpers which need
596 an unchanging input format.
598 In addition to the above, any string specified in the referencing
599 acl will also be included in the helper request line, after the
600 specified formats (see the "acl external" directive)
602 The helper receives lines per the above format specification,
603 and returns lines starting with OK or ERR indicating the validity
604 of the request and optionally followed by additional keywords with
607 General result syntax:
609 OK/ERR keyword=value ...
613 user= The users name (login)
614 password= The users password (for login= cache_peer option)
615 message= Message describing the reason. Available as %o
617 tag= Apply a tag to a request (for both ERR and OK results)
618 Only sets a tag, does not alter existing tags.
619 log= String to be logged in access.log. Available as
620 %ea in logformat specifications
622 If protocol=3.0 (the default) then URL escaping is used to protect
623 each value in both requests and responses.
625 If using protocol=2.5 then all values need to be enclosed in quotes
626 if they may contain whitespace, or the whitespace escaped using \.
627 And quotes or \ characters within the keyword value must be \ escaped.
629 When using the concurrency= option the protocol is changed by
630 introducing a query channel tag infront of the request/response.
631 The query channel tag is a number between 0 and concurrency-1.
638 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
639 DEFAULT: localhost src 127.0.0.1/32 ::1
640 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
641 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
643 Defining an Access List
645 Every access list definition must begin with an aclname and acltype,
646 followed by either type-specific arguments or a quoted filename that
649 acl aclname acltype argument ...
650 acl aclname acltype "file" ...
652 When using "file", the file should contain one item per line.
654 By default, regular expressions are CASE-SENSITIVE.
655 To make them case-insensitive, use the -i option. To return case-sensitive
656 use the +i option between patterns, or make a new ACL line without -i.
658 Some acl types require suspending the current request in order
659 to access some external data source.
660 Those which do are marked with the tag [slow], those which
661 don't are marked as [fast].
662 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
663 for further information
665 ***** ACL TYPES AVAILABLE *****
667 acl aclname src ip-address/mask ... # clients IP address [fast]
668 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
669 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
670 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
672 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
673 # The arp ACL requires the special configure option --enable-arp-acl.
674 # Furthermore, the ARP ACL code is not portable to all operating systems.
675 # It works on Linux, Solaris, Windows, FreeBSD, and some
676 # other *BSD variants.
679 # NOTE: Squid can only determine the MAC address for clients that are on
680 # the same subnet. If the client is on a different subnet,
681 # then Squid cannot find out its MAC address.
683 acl aclname srcdomain .foo.com ...
684 # reverse lookup, from client IP [slow]
685 acl aclname dstdomain .foo.com ...
686 # Destination server from URL [fast]
687 acl aclname srcdom_regex [-i] \.foo\.com ...
688 # regex matching client name [slow]
689 acl aclname dstdom_regex [-i] \.foo\.com ...
690 # regex matching server [fast]
692 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
693 # based URL is used and no match is found. The name "none" is used
694 # if the reverse lookup fails.
696 acl aclname src_as number ...
697 acl aclname dst_as number ...
699 # Except for access control, AS numbers can be used for
700 # routing of requests to specific caches. Here's an
701 # example for routing all requests for AS#1241 and only
702 # those to mycache.mydomain.net:
703 # acl asexample dst_as 1241
704 # cache_peer_access mycache.mydomain.net allow asexample
705 # cache_peer_access mycache_mydomain.net deny all
707 acl aclname peername myPeer ...
709 # match against a named cache_peer entry
710 # set unique name= on cache_peer lines for reliable use.
712 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
722 # h1:m1 must be less than h2:m2
724 acl aclname url_regex [-i] ^http:// ...
725 # regex matching on whole URL [fast]
726 acl aclname urlpath_regex [-i] \.gif$ ...
727 # regex matching on URL path [fast]
729 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
731 acl aclname localport 3128 ... # TCP port the client connected to [fast]
732 # NP: for interception mode this is usually '80'
734 acl aclname myportname 3128 ... # http(s)_port name [fast]
736 acl aclname proto HTTP FTP ... # request protocol [fast]
738 acl aclname method GET POST ... # HTTP request method [fast]
740 acl aclname http_status 200 301 500- 400-403 ...
741 # status code in reply [fast]
743 acl aclname browser [-i] regexp ...
744 # pattern match on User-Agent header (see also req_header below) [fast]
746 acl aclname referer_regex [-i] regexp ...
747 # pattern match on Referer header [fast]
748 # Referer is highly unreliable, so use with care
750 acl aclname ident username ...
751 acl aclname ident_regex [-i] pattern ...
752 # string match on ident output [slow]
753 # use REQUIRED to accept any non-null ident.
755 acl aclname proxy_auth [-i] username ...
756 acl aclname proxy_auth_regex [-i] pattern ...
757 # perform http authentication challenge to the client and match against
758 # supplied credentials [slow]
760 # takes a list of allowed usernames.
761 # use REQUIRED to accept any valid username.
763 # Will use proxy authentication in forward-proxy scenarios, and plain
764 # http authenticaiton in reverse-proxy scenarios
766 # NOTE: when a Proxy-Authentication header is sent but it is not
767 # needed during ACL checking the username is NOT logged
770 # NOTE: proxy_auth requires a EXTERNAL authentication program
771 # to check username/password combinations (see
772 # auth_param directive).
774 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
775 # as the browser needs to be configured for using a proxy in order
776 # to respond to proxy authentication.
778 acl aclname snmp_community string ...
779 # A community string to limit access to your SNMP Agent [fast]
782 # acl snmppublic snmp_community public
784 acl aclname maxconn number
785 # This will be matched when the client's IP address has
786 # more than <number> TCP connections established. [fast]
787 # NOTE: This only measures direct TCP links so X-Forwarded-For
788 # indirect clients are not counted.
790 acl aclname max_user_ip [-s] number
791 # This will be matched when the user attempts to log in from more
792 # than <number> different ip addresses. The authenticate_ip_ttl
793 # parameter controls the timeout on the ip entries. [fast]
794 # If -s is specified the limit is strict, denying browsing
795 # from any further IP addresses until the ttl has expired. Without
796 # -s Squid will just annoy the user by "randomly" denying requests.
797 # (the counter is reset each time the limit is reached and a
799 # NOTE: in acceleration mode or where there is mesh of child proxies,
800 # clients may appear to come from multiple addresses if they are
801 # going through proxy farms, so a limit of 1 may cause user problems.
803 acl aclname random probability
804 # Pseudo-randomly match requests. Based on the probability given.
805 # Probability may be written as a decimal (0.333), fraction (1/3)
806 # or ratio of matches:non-matches (3:5).
808 acl aclname req_mime_type [-i] mime-type ...
809 # regex match against the mime type of the request generated
810 # by the client. Can be used to detect file upload or some
811 # types HTTP tunneling requests [fast]
812 # NOTE: This does NOT match the reply. You cannot use this
813 # to match the returned file type.
815 acl aclname req_header header-name [-i] any\.regex\.here
816 # regex match against any of the known request headers. May be
817 # thought of as a superset of "browser", "referer" and "mime-type"
820 acl aclname rep_mime_type [-i] mime-type ...
821 # regex match against the mime type of the reply received by
822 # squid. Can be used to detect file download or some
823 # types HTTP tunneling requests. [fast]
824 # NOTE: This has no effect in http_access rules. It only has
825 # effect in rules that affect the reply data stream such as
828 acl aclname rep_header header-name [-i] any\.regex\.here
829 # regex match against any of the known reply headers. May be
830 # thought of as a superset of "browser", "referer" and "mime-type"
833 acl aclname external class_name [arguments...]
834 # external ACL lookup via a helper class defined by the
835 # external_acl_type directive [slow]
837 acl aclname user_cert attribute values...
838 # match against attributes in a user SSL certificate
839 # attribute is one of DN/C/O/CN/L/ST [fast]
841 acl aclname ca_cert attribute values...
842 # match against attributes a users issuing CA SSL certificate
843 # attribute is one of DN/C/O/CN/L/ST [fast]
845 acl aclname ext_user username ...
846 acl aclname ext_user_regex [-i] pattern ...
847 # string match on username returned by external acl helper [slow]
848 # use REQUIRED to accept any non-null user name.
850 acl aclname tag tagvalue ...
851 # string match on tag returned by external acl helper [slow]
853 acl aclname hier_code codename ...
854 # string match against squid hierarchy code(s); [fast]
855 # e.g., DIRECT, PARENT_HIT, NONE, etc.
857 # NOTE: This has no effect in http_access rules. It only has
858 # effect in rules that affect the reply data stream such as
862 acl macaddress arp 09:00:2b:23:45:67
863 acl myexample dst_as 1241
864 acl password proxy_auth REQUIRED
865 acl fileupload req_mime_type -i ^multipart/form-data$
866 acl javascript rep_mime_type -i ^application/x-javascript$
870 # Recommended minimum configuration:
873 # Example rule allowing access from your local networks.
874 # Adapt to list your (internal) IP networks from where browsing
876 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
877 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
878 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
879 acl localnet src fc00::/7 # RFC 4193 local private network range
880 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
882 acl SSL_ports port 443
883 acl Safe_ports port 80 # http
884 acl Safe_ports port 21 # ftp
885 acl Safe_ports port 443 # https
886 acl Safe_ports port 70 # gopher
887 acl Safe_ports port 210 # wais
888 acl Safe_ports port 1025-65535 # unregistered ports
889 acl Safe_ports port 280 # http-mgmt
890 acl Safe_ports port 488 # gss-http
891 acl Safe_ports port 591 # filemaker
892 acl Safe_ports port 777 # multiling http
893 acl CONNECT method CONNECT
897 NAME: follow_x_forwarded_for
899 IFDEF: FOLLOW_X_FORWARDED_FOR
900 LOC: Config.accessList.followXFF
901 DEFAULT_IF_NONE: deny all
903 Allowing or Denying the X-Forwarded-For header to be followed to
904 find the original source of a request.
906 Requests may pass through a chain of several other proxies
907 before reaching us. The X-Forwarded-For header will contain a
908 comma-separated list of the IP addresses in the chain, with the
909 rightmost address being the most recent.
911 If a request reaches us from a source that is allowed by this
912 configuration item, then we consult the X-Forwarded-For header
913 to see where that host received the request from. If the
914 X-Forwarded-For header contains multiple addresses, we continue
915 backtracking until we reach an address for which we are not allowed
916 to follow the X-Forwarded-For header, or until we reach the first
917 address in the list. For the purpose of ACL used in the
918 follow_x_forwarded_for directive the src ACL type always matches
919 the address we are testing and srcdomain matches its rDNS.
921 The end result of this process is an IP address that we will
922 refer to as the indirect client address. This address may
923 be treated as the client address for access control, ICAP, delay
924 pools and logging, depending on the acl_uses_indirect_client,
925 icap_uses_indirect_client, delay_pool_uses_indirect_client,
926 log_uses_indirect_client and tproxy_uses_indirect_client options.
928 This clause only supports fast acl types.
929 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
931 SECURITY CONSIDERATIONS:
933 Any host for which we follow the X-Forwarded-For header
934 can place incorrect information in the header, and Squid
935 will use the incorrect information as if it were the
936 source address of the request. This may enable remote
937 hosts to bypass any access control restrictions that are
938 based on the client's source addresses.
942 acl localhost src 127.0.0.1
943 acl my_other_proxy srcdomain .proxy.example.com
944 follow_x_forwarded_for allow localhost
945 follow_x_forwarded_for allow my_other_proxy
948 NAME: acl_uses_indirect_client
951 IFDEF: FOLLOW_X_FORWARDED_FOR
953 LOC: Config.onoff.acl_uses_indirect_client
955 Controls whether the indirect client address
956 (see follow_x_forwarded_for) is used instead of the
957 direct client address in acl matching.
959 NOTE: maxconn ACL considers direct TCP links and indirect
960 clients will always have zero. So no match.
963 NAME: delay_pool_uses_indirect_client
966 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
968 LOC: Config.onoff.delay_pool_uses_indirect_client
970 Controls whether the indirect client address
971 (see follow_x_forwarded_for) is used instead of the
972 direct client address in delay pools.
975 NAME: log_uses_indirect_client
978 IFDEF: FOLLOW_X_FORWARDED_FOR
980 LOC: Config.onoff.log_uses_indirect_client
982 Controls whether the indirect client address
983 (see follow_x_forwarded_for) is used instead of the
984 direct client address in the access log.
987 NAME: tproxy_uses_indirect_client
990 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
992 LOC: Config.onoff.tproxy_uses_indirect_client
994 Controls whether the indirect client address
995 (see follow_x_forwarded_for) is used instead of the
996 direct client address when spoofing the outgoing client.
998 This has no effect on requests arriving in non-tproxy
1001 SECURITY WARNING: Usage of this option is dangerous
1002 and should not be used trivially. Correct configuration
1003 of follow_x_forewarded_for with a limited set of trusted
1004 sources is required to prevent abuse of your proxy.
1009 LOC: Config.accessList.http
1010 DEFAULT_IF_NONE: deny all
1012 Allowing or Denying access based on defined access lists
1014 Access to the HTTP port:
1015 http_access allow|deny [!]aclname ...
1017 NOTE on default values:
1019 If there are no "access" lines present, the default is to deny
1022 If none of the "access" lines cause a match, the default is the
1023 opposite of the last line in the list. If the last line was
1024 deny, the default is allow. Conversely, if the last line
1025 is allow, the default will be deny. For these reasons, it is a
1026 good idea to have an "deny all" entry at the end of your access
1027 lists to avoid potential confusion.
1029 This clause supports both fast and slow acl types.
1030 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1035 # Recommended minimum Access Permission configuration:
1037 # Only allow cachemgr access from localhost
1038 http_access allow localhost manager
1039 http_access deny manager
1041 # Deny requests to certain unsafe ports
1042 http_access deny !Safe_ports
1044 # Deny CONNECT to other than secure SSL ports
1045 http_access deny CONNECT !SSL_ports
1047 # We strongly recommend the following be uncommented to protect innocent
1048 # web applications running on the proxy server who think the only
1049 # one who can access services on "localhost" is a local user
1050 #http_access deny to_localhost
1053 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1056 # Example rule allowing access from your local networks.
1057 # Adapt localnet in the ACL section to list your (internal) IP networks
1058 # from where browsing should be allowed
1059 http_access allow localnet
1060 http_access allow localhost
1062 # And finally deny all other access to this proxy
1063 http_access deny all
1067 NAME: adapted_http_access http_access2
1069 LOC: Config.accessList.adapted_http
1072 Allowing or Denying access based on defined access lists
1074 Essentially identical to http_access, but runs after redirectors
1075 and ICAP/eCAP adaptation. Allowing access control based on their
1078 If not set then only http_access is used.
1081 NAME: http_reply_access
1083 LOC: Config.accessList.reply
1086 Allow replies to client requests. This is complementary to http_access.
1088 http_reply_access allow|deny [!] aclname ...
1090 NOTE: if there are no access lines present, the default is to allow
1093 If none of the access lines cause a match the opposite of the
1094 last line will apply. Thus it is good practice to end the rules
1095 with an "allow all" or "deny all" entry.
1097 This clause supports both fast and slow acl types.
1098 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1103 LOC: Config.accessList.icp
1104 DEFAULT_IF_NONE: deny all
1106 Allowing or Denying access to the ICP port based on defined
1109 icp_access allow|deny [!]aclname ...
1111 See http_access for details
1113 This clause only supports fast acl types.
1114 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1116 # Allow ICP queries from local networks only
1117 #icp_access allow localnet
1118 #icp_access deny all
1124 LOC: Config.accessList.htcp
1125 DEFAULT_IF_NONE: deny all
1127 Allowing or Denying access to the HTCP port based on defined
1130 htcp_access allow|deny [!]aclname ...
1132 See http_access for details
1134 NOTE: The default if no htcp_access lines are present is to
1135 deny all traffic. This default may cause problems with peers
1136 using the htcp option.
1138 This clause only supports fast acl types.
1139 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1141 # Allow HTCP queries from local networks only
1142 #htcp_access allow localnet
1143 #htcp_access deny all
1146 NAME: htcp_clr_access
1149 LOC: Config.accessList.htcp_clr
1150 DEFAULT_IF_NONE: deny all
1152 Allowing or Denying access to purge content using HTCP based
1153 on defined access lists
1155 htcp_clr_access allow|deny [!]aclname ...
1157 See http_access for details
1159 This clause only supports fast acl types.
1160 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1162 # Allow HTCP CLR requests from trusted peers
1163 acl htcp_clr_peer src 172.16.1.2
1164 htcp_clr_access allow htcp_clr_peer
1169 LOC: Config.accessList.miss
1172 Determins whether network access is permitted when satisfying a request.
1175 to force your neighbors to use you as a sibling instead of
1178 acl localclients src 172.16.0.0/16
1179 miss_access allow localclients
1180 miss_access deny !localclients
1182 This means only your local clients are allowed to fetch relayed/MISS
1183 replies from the network and all other clients can only fetch cached
1187 The default for this setting allows all clients who passed the
1188 http_access rules to relay via this proxy.
1190 This clause only supports fast acl types.
1191 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1194 NAME: ident_lookup_access
1197 DEFAULT_IF_NONE: deny all
1198 LOC: Ident::TheConfig.identLookup
1200 A list of ACL elements which, if matched, cause an ident
1201 (RFC 931) lookup to be performed for this request. For
1202 example, you might choose to always perform ident lookups
1203 for your main multi-user Unix boxes, but not for your Macs
1204 and PCs. By default, ident lookups are not performed for
1207 To enable ident lookups for specific client addresses, you
1208 can follow this example:
1210 acl ident_aware_hosts src 198.168.1.0/24
1211 ident_lookup_access allow ident_aware_hosts
1212 ident_lookup_access deny all
1214 Only src type ACL checks are fully supported. A srcdomain
1215 ACL might work at times, but it will not always provide
1218 This clause only supports fast acl types.
1219 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1222 NAME: reply_body_max_size
1223 COMMENT: size [acl acl...]
1226 LOC: Config.ReplyBodySize
1228 This option specifies the maximum size of a reply body. It can be
1229 used to prevent users from downloading very large files, such as
1230 MP3's and movies. When the reply headers are received, the
1231 reply_body_max_size lines are processed, and the first line where
1232 all (if any) listed ACLs are true is used as the maximum body size
1235 This size is checked twice. First when we get the reply headers,
1236 we check the content-length value. If the content length value exists
1237 and is larger than the allowed size, the request is denied and the
1238 user receives an error message that says "the request or reply
1239 is too large." If there is no content-length, and the reply
1240 size exceeds this limit, the client's connection is just closed
1241 and they will receive a partial reply.
1243 WARNING: downstream caches probably can not detect a partial reply
1244 if there is no content-length header, so they will cache
1245 partial responses and give them out as hits. You should NOT
1246 use this option if you have downstream caches.
1248 WARNING: A maximum size smaller than the size of squid's error messages
1249 will cause an infinite loop and crash squid. Ensure that the smallest
1250 non-zero value you use is greater that the maximum header size plus
1251 the size of your largest error page.
1253 If you set this parameter none (the default), there will be
1256 Configuration Format is:
1257 reply_body_max_size SIZE UNITS [acl ...]
1259 reply_body_max_size 10 MB
1265 -----------------------------------------------------------------------------
1268 NAME: http_port ascii_port
1269 TYPE: http_port_list
1271 LOC: Config.Sockaddr.http
1273 Usage: port [mode] [options]
1274 hostname:port [mode] [options]
1275 1.2.3.4:port [mode] [options]
1277 The socket addresses where Squid will listen for HTTP client
1278 requests. You may specify multiple socket addresses.
1279 There are three forms: port alone, hostname with port, and
1280 IP address with port. If you specify a hostname or IP
1281 address, Squid binds the socket to that specific
1282 address. Most likely, you do not need to bind to a specific
1283 address, so you can use the port number alone.
1285 If you are running Squid in accelerator mode, you
1286 probably want to listen on port 80 also, or instead.
1288 The -a command line option may be used to specify additional
1289 port(s) where Squid listens for proxy request. Such ports will
1290 be plain proxy ports with no options.
1292 You may specify multiple socket addresses on multiple lines.
1296 intercept Support for IP-Layer interception of
1297 outgoing requests without browser settings.
1298 NP: disables authentication and IPv6 on the port.
1300 tproxy Support Linux TPROXY for spoofing outgoing
1301 connections using the client IP address.
1302 NP: disables authentication and maybe IPv6 on the port.
1304 accel Accelerator / reverse proxy mode
1306 ssl-bump Intercept each CONNECT request matching ssl_bump ACL,
1307 establish secure connection with the client and with
1308 the server, decrypt HTTP messages as they pass through
1309 Squid, and treat them as unencrypted HTTP messages,
1310 becoming the man-in-the-middle.
1312 The ssl_bump option is required to fully enable
1313 the SslBump feature.
1315 Omitting the mode flag causes default forward proxy mode to be used.
1318 Accelerator Mode Options:
1320 defaultsite=domainname
1321 What to use for the Host: header if it is not present
1322 in a request. Determines what site (not origin server)
1323 accelerators should consider the default.
1325 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1327 protocol= Protocol to reconstruct accelerated requests with.
1328 Defaults to http for http_port and https for
1331 vport Virtual host port support. Using the http_port number
1332 instead of the port passed on Host: headers.
1334 vport=NN Virtual host port support. Using the specified port
1335 number instead of the port passed on Host: headers.
1338 Act as if this Squid is the origin server.
1339 This currently means generate new Date: and Expires:
1340 headers on HIT instead of adding Age:.
1342 ignore-cc Ignore request Cache-Control headers.
1344 WARNING: This option violates HTTP specifications if
1345 used in non-accelerator setups.
1347 allow-direct Allow direct forwarding in accelerator mode. Normally
1348 accelerated requests are denied direct forwarding as if
1349 never_direct was used.
1351 WARNING: this option opens accelerator mode to security
1352 vulnerabilities usually only affecting in interception
1353 mode. Make sure to protect forwarding with suitable
1354 http_access rules when using this.
1357 SSL Bump Mode Options:
1359 cert= Path to SSL certificate (PEM format).
1361 key= Path to SSL private key file (PEM format)
1362 if not specified, the certificate file is
1363 assumed to be a combined certificate and
1366 version= The version of SSL/TLS supported
1367 1 automatic (default)
1372 cipher= Colon separated list of supported ciphers.
1374 options= Various SSL engine options. The most important
1376 NO_SSLv2 Disallow the use of SSLv2
1377 NO_SSLv3 Disallow the use of SSLv3
1378 NO_TLSv1 Disallow the use of TLSv1
1379 SINGLE_DH_USE Always create a new key when using
1380 temporary/ephemeral DH key exchanges
1381 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1382 documentation for a complete list of options.
1384 clientca= File containing the list of CAs to use when
1385 requesting a client certificate.
1387 cafile= File containing additional CA certificates to
1388 use when verifying client certificates. If unset
1389 clientca will be used.
1391 capath= Directory containing additional CA certificates
1392 and CRL lists to use when verifying client certificates.
1394 crlfile= File of additional CRL lists to use when verifying
1395 the client certificate, in addition to CRLs stored in
1396 the capath. Implies VERIFY_CRL flag below.
1398 dhparams= File containing DH parameters for temporary/ephemeral
1401 sslflags= Various flags modifying the use of SSL:
1403 Don't request client certificates
1404 immediately, but wait until acl processing
1405 requires a certificate (not yet implemented).
1407 Don't use the default CA lists built in
1410 Don't allow for session reuse. Each connection
1411 will result in a new SSL session.
1413 Verify CRL lists when accepting client
1416 Verify CRL lists for all certificates in the
1417 client certificate chain.
1419 sslcontext= SSL session ID context identifier.
1421 generate-host-certificates[=<on|off>]
1422 Dynamically create SSL server certificates for the
1423 destination hosts of bumped CONNECT requests.When
1424 enabled, the cert and key options are used to sign
1425 generated certificates. Otherwise generated
1426 certificate will be selfsigned.
1427 If there is CA certificate life time of generated
1428 certificate equals lifetime of CA certificate. If
1429 generated certificate is selfsigned lifetime is three
1431 This option is enabled by default when SslBump is used.
1432 See the sslBump option above for more information.
1434 dynamic_cert_mem_cache_size=SIZE
1435 Approximate total RAM size spent on cached generated
1436 certificates. If set to zero, caching is disabled. The
1437 default value is 4MB. An average XXX-bit certificate
1438 consumes about XXX bytes of RAM.
1442 connection-auth[=on|off]
1443 use connection-auth=off to tell Squid to prevent
1444 forwarding Microsoft connection oriented authentication
1445 (NTLM, Negotiate and Kerberos)
1447 disable-pmtu-discovery=
1448 Control Path-MTU discovery usage:
1449 off lets OS decide on what to do (default).
1450 transparent disable PMTU discovery when transparent
1452 always disable always PMTU discovery.
1454 In many setups of transparently intercepting proxies
1455 Path-MTU discovery can not work on traffic towards the
1456 clients. This is the case when the intercepting device
1457 does not fully track connections and fails to forward
1458 ICMP must fragment messages to the cache server. If you
1459 have such setup and experience that certain clients
1460 sporadically hang or never complete requests set
1461 disable-pmtu-discovery option to 'transparent'.
1463 name= Specifies a internal name for the port. Defaults to
1464 the port specification (port or addr:port)
1466 tcpkeepalive[=idle,interval,timeout]
1467 Enable TCP keepalive probes of idle connections.
1468 In seconds; idle is the initial time before TCP starts
1469 probing the connection, interval how often to probe, and
1470 timeout the time before giving up.
1472 If you run Squid on a dual-homed machine with an internal
1473 and an external interface we recommend you to specify the
1474 internal address:port in http_port. This way Squid will only be
1475 visible on the internal address.
1479 # Squid normally listens to port 3128
1480 http_port @DEFAULT_HTTP_PORT@
1486 TYPE: https_port_list
1488 LOC: Config.Sockaddr.https
1490 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1492 The socket address where Squid will listen for HTTPS client
1495 This is really only useful for situations where you are running
1496 squid in accelerator mode and you want to do the SSL work at the
1499 You may specify multiple socket addresses on multiple lines,
1500 each with their own SSL certificate and/or options.
1504 accel Accelerator / reverse proxy mode
1506 Omitting the mode flag causes default forward proxy mode to be used.
1509 See http_port for a list of generic options
1514 cert= Path to SSL certificate (PEM format).
1516 key= Path to SSL private key file (PEM format)
1517 if not specified, the certificate file is
1518 assumed to be a combined certificate and
1521 version= The version of SSL/TLS supported
1522 1 automatic (default)
1527 cipher= Colon separated list of supported ciphers.
1529 options= Various SSL engine options. The most important
1531 NO_SSLv2 Disallow the use of SSLv2
1532 NO_SSLv3 Disallow the use of SSLv3
1533 NO_TLSv1 Disallow the use of TLSv1
1534 SINGLE_DH_USE Always create a new key when using
1535 temporary/ephemeral DH key exchanges
1536 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1537 documentation for a complete list of options.
1539 clientca= File containing the list of CAs to use when
1540 requesting a client certificate.
1542 cafile= File containing additional CA certificates to
1543 use when verifying client certificates. If unset
1544 clientca will be used.
1546 capath= Directory containing additional CA certificates
1547 and CRL lists to use when verifying client certificates.
1549 crlfile= File of additional CRL lists to use when verifying
1550 the client certificate, in addition to CRLs stored in
1551 the capath. Implies VERIFY_CRL flag below.
1553 dhparams= File containing DH parameters for temporary/ephemeral
1556 sslflags= Various flags modifying the use of SSL:
1558 Don't request client certificates
1559 immediately, but wait until acl processing
1560 requires a certificate (not yet implemented).
1562 Don't use the default CA lists built in
1565 Don't allow for session reuse. Each connection
1566 will result in a new SSL session.
1568 Verify CRL lists when accepting client
1571 Verify CRL lists for all certificates in the
1572 client certificate chain.
1574 sslcontext= SSL session ID context identifier.
1578 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1581 LOC: Ip::Qos::TheConfig.tosToServer
1583 Allows you to select a TOS/Diffserv value for packets outgoing
1584 on the server side, based on an ACL.
1586 tcp_outgoing_tos ds-field [!]aclname ...
1588 Example where normal_service_net uses the TOS value 0x00
1589 and good_service_net uses 0x20
1591 acl normal_service_net src 10.0.0.0/24
1592 acl good_service_net src 10.0.1.0/24
1593 tcp_outgoing_tos 0x00 normal_service_net
1594 tcp_outgoing_tos 0x20 good_service_net
1596 TOS/DSCP values really only have local significance - so you should
1597 know what you're specifying. For more information, see RFC2474,
1598 RFC2475, and RFC3260.
1600 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1601 "default" to use whatever default your host has. Note that in
1602 practice often only multiples of 4 is usable as the two rightmost bits
1603 have been redefined for use by ECN (RFC 3168 section 23.1).
1605 Processing proceeds in the order specified, and stops at first fully
1609 NAME: clientside_tos
1612 LOC: Ip::Qos::TheConfig.tosToClient
1614 Allows you to select a TOS/Diffserv value for packets being transmitted
1615 on the client-side, based on an ACL.
1617 clientside_tos ds-field [!]aclname ...
1619 Example where normal_service_net uses the TOS value 0x00
1620 and good_service_net uses 0x20
1622 acl normal_service_net src 10.0.0.0/24
1623 acl good_service_net src 10.0.1.0/24
1624 clientside_tos 0x00 normal_service_net
1625 clientside_tos 0x20 good_service_net
1627 Note: This feature is incompatible with qos_flows. Any TOS values set here
1628 will be overwritten by TOS values in qos_flows.
1631 NAME: tcp_outgoing_mark
1633 IFDEF: SO_MARK&&USE_LIBCAP
1635 LOC: Ip::Qos::TheConfig.nfmarkToServer
1637 Allows you to apply a Netfilter mark value to outgoing packets
1638 on the server side, based on an ACL.
1640 tcp_outgoing_mark mark-value [!]aclname ...
1642 Example where normal_service_net uses the mark value 0x00
1643 and good_service_net uses 0x20
1645 acl normal_service_net src 10.0.0.0/24
1646 acl good_service_net src 10.0.1.0/24
1647 tcp_outgoing_mark 0x00 normal_service_net
1648 tcp_outgoing_mark 0x20 good_service_net
1651 NAME: clientside_mark
1653 IFDEF: SO_MARK&&USE_LIBCAP
1655 LOC: Ip::Qos::TheConfig.nfmarkToClient
1657 Allows you to apply a Netfilter mark value to packets being transmitted
1658 on the client-side, based on an ACL.
1660 clientside_mark mark-value [!]aclname ...
1662 Example where normal_service_net uses the mark value 0x00
1663 and good_service_net uses 0x20
1665 acl normal_service_net src 10.0.0.0/24
1666 acl good_service_net src 10.0.1.0/24
1667 clientside_mark 0x00 normal_service_net
1668 clientside_mark 0x20 good_service_net
1670 Note: This feature is incompatible with qos_flows. Any mark values set here
1671 will be overwritten by mark values in qos_flows.
1678 LOC: Ip::Qos::TheConfig
1680 Allows you to select a TOS/DSCP value to mark outgoing
1681 connections with, based on where the reply was sourced. For
1682 platforms using netfilter, allows you to set a netfilter mark
1683 value instead of, or in addition to, a TOS value.
1685 TOS values really only have local significance - so you should
1686 know what you're specifying. For more information, see RFC2474,
1687 RFC2475, and RFC3260.
1689 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1690 in practice often only multiples of 4 is usable as the two rightmost bits
1691 have been redefined for use by ECN (RFC 3168 section 23.1).
1693 Mark values can be any unsigned 32-bit integer value.
1695 This setting is configured by setting the following values:
1697 tos|mark Whether to set TOS or netfilter mark values
1699 local-hit=0xFF Value to mark local cache hits.
1701 sibling-hit=0xFF Value to mark hits from sibling peers.
1703 parent-hit=0xFF Value to mark hits from parent peers.
1705 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1706 over the preserve-miss feature (see below), unless
1707 mask is specified, in which case only the bits
1708 specified in the mask are written.
1710 The TOS variant of the following features are only possible on Linux
1711 and require your kernel to be patched with the TOS preserving ZPH
1712 patch, available from http://zph.bratcheda.org
1713 No patch is needed to preserve the netfilter mark, which will work
1714 with all variants of netfilter.
1716 disable-preserve-miss
1717 This option disables the preservation of the TOS or netfilter
1718 mark. By default, the existing TOS or netfilter mark value of
1719 the response coming from the remote server will be retained
1720 and masked with miss-mark.
1721 NOTE: in the case of a netfilter mark, the mark must be set on
1722 the connection (using the CONNMARK target) not on the packet
1726 Allows you to mask certain bits in the TOS or mark value
1727 received from the remote server, before copying the value to
1728 the TOS sent towards clients.
1729 Default for tos: 0xFF (TOS from server is not changed).
1730 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1732 All of these features require the --enable-zph-qos compilation flag
1733 (enabled by default). Netfilter marking also requires the
1734 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1735 libcap 2.09+ (--with-libcap).
1739 NAME: tcp_outgoing_address
1742 LOC: Config.accessList.outgoing_address
1744 Allows you to map requests to different outgoing IP addresses
1745 based on the username or source address of the user making
1748 tcp_outgoing_address ipaddr [[!]aclname] ...
1751 Forwarding clients with dedicated IPs for certain subnets.
1753 acl normal_service_net src 10.0.0.0/24
1754 acl good_service_net src 10.0.2.0/24
1756 tcp_outgoing_address 2001:db8::c001 good_service_net
1757 tcp_outgoing_address 10.1.0.2 good_service_net
1759 tcp_outgoing_address 2001:db8::beef normal_service_net
1760 tcp_outgoing_address 10.1.0.1 normal_service_net
1762 tcp_outgoing_address 2001:db8::1
1763 tcp_outgoing_address 10.1.0.3
1765 Processing proceeds in the order specified, and stops at first fully
1768 Squid will add an implicit IP version test to each line.
1769 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1770 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1773 NOTE: The use of this directive using client dependent ACLs is
1774 incompatible with the use of server side persistent connections. To
1775 ensure correct results it is best to set server_persistent_connections
1776 to off when using this directive in such configurations.
1778 NOTE: The use of this directive to set a local IP on outgoing TCP links
1779 is incompatible with using TPROXY to set client IP out outbound TCP links.
1780 When needing to contact peers use the no-tproxy cache_peer option and the
1781 client_dst_passthru directive re-enable normal forwarding such as this.
1785 NAME: host_verify_strict
1788 LOC: Config.onoff.hostStrictVerify
1790 Regardless of this option setting, when dealing with intercepted
1791 traffic, Squid always verifies that the destination IP address matches
1792 the Host header domain or IP (called 'authority form URL'). Squid
1793 responds with an HTTP 409 (Conflict) error page and logs a security
1794 warning if there is no match.
1796 When set to ON, Squid verifies that the destination IP address matches
1797 the Host header for forward-proxy and reverse-proxy traffic as well. For
1798 those traffic types, Squid also enables the following checks, comparing
1799 the corresponding Host header and Request-URI components:
1801 * The host names (domain or IP) must be identical,
1802 but valueless or missing Host header disables all checks.
1803 For the two host names to match, both must be either IP or FQDN.
1805 * Port numbers must be identical,
1806 but if a port is missing, the scheme-default port is assumed.
1808 This enforcement is performed to satisfy a MUST-level requirement in
1809 RFC 2616 section 14.23: "The Host field value MUST represent the naming
1810 authority of the origin server or gateway given by the original URL".
1813 NAME: client_dst_passthru
1816 LOC: Config.onoff.client_dst_passthru
1818 With NAT or TPROXY intercepted traffic Squid may pass the request
1819 directly to the original client destination IP or seek a faster
1822 This option (on by default) prevents cache_peer and alternative DNS
1823 entries being used on intercepted traffic. Both of which lead to
1824 the security vulnerability outlined below.
1828 This directive should only be disabled if cache_peer are required.
1830 As described in CVE-2009-0801 when the Host: header alone is used
1831 to determine the destination of a request it becomes trivial for
1832 malicious scripts on remote websites to bypass browser same-origin
1833 security policy and sandboxing protections.
1835 The cause of this is that such applets are allowed to perform their
1836 own HTTP stack, in which case the same-origin policy of the browser
1837 sandbox only verifies that the applet tries to contact the same IP
1838 as from where it was loaded at the IP level. The Host: header may
1839 be different from the connected IP and approved origin.
1845 -----------------------------------------------------------------------------
1848 NAME: ssl_unclean_shutdown
1852 LOC: Config.SSL.unclean_shutdown
1854 Some browsers (especially MSIE) bugs out on SSL shutdown
1861 LOC: Config.SSL.ssl_engine
1864 The OpenSSL engine to use. You will need to set this if you
1865 would like to use hardware SSL acceleration for example.
1868 NAME: sslproxy_client_certificate
1871 LOC: Config.ssl_client.cert
1874 Client SSL Certificate to use when proxying https:// URLs
1877 NAME: sslproxy_client_key
1880 LOC: Config.ssl_client.key
1883 Client SSL Key to use when proxying https:// URLs
1886 NAME: sslproxy_version
1889 LOC: Config.ssl_client.version
1892 SSL version level to use when proxying https:// URLs
1895 NAME: sslproxy_options
1898 LOC: Config.ssl_client.options
1901 SSL engine options to use when proxying https:// URLs
1903 The most important being:
1905 NO_SSLv2 Disallow the use of SSLv2
1906 NO_SSLv3 Disallow the use of SSLv3
1907 NO_TLSv1 Disallow the use of TLSv1
1909 Always create a new key when using
1910 temporary/ephemeral DH key exchanges
1912 These options vary depending on your SSL engine.
1913 See the OpenSSL SSL_CTX_set_options documentation for a
1914 complete list of possible options.
1917 NAME: sslproxy_cipher
1920 LOC: Config.ssl_client.cipher
1923 SSL cipher list to use when proxying https:// URLs
1925 Colon separated list of supported ciphers.
1928 NAME: sslproxy_cafile
1931 LOC: Config.ssl_client.cafile
1934 file containing CA certificates to use when verifying server
1935 certificates while proxying https:// URLs
1938 NAME: sslproxy_capath
1941 LOC: Config.ssl_client.capath
1944 directory containing CA certificates to use when verifying
1945 server certificates while proxying https:// URLs
1951 LOC: Config.accessList.ssl_bump
1954 This ACL controls which CONNECT requests to an http_port
1955 marked with an sslBump flag are actually "bumped". Please
1956 see the sslBump flag of an http_port option for more details
1957 about decoding proxied SSL connections.
1959 By default, no requests are bumped.
1961 See also: http_port ssl-bump
1963 This clause supports both fast and slow acl types.
1964 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1967 # Example: Bump all requests except those originating from localhost and
1968 # those going to webax.com or example.com sites.
1970 acl localhost src 127.0.0.1/32
1971 acl broken_sites dstdomain .webax.com
1972 acl broken_sites dstdomain .example.com
1973 ssl_bump deny localhost
1974 ssl_bump deny broken_sites
1978 NAME: sslproxy_flags
1981 LOC: Config.ssl_client.flags
1984 Various flags modifying the use of SSL while proxying https:// URLs:
1985 DONT_VERIFY_PEER Accept certificates that fail verification.
1986 For refined control, see sslproxy_cert_error.
1987 NO_DEFAULT_CA Don't use the default CA list built in
1991 NAME: sslproxy_cert_error
1994 LOC: Config.ssl_client.cert_error
1997 Use this ACL to bypass server certificate validation errors.
1999 For example, the following lines will bypass all validation errors
2000 when talking to servers located at 172.16.0.0/16. All other
2001 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2003 acl BrokenServersAtTrustedIP dst 172.16.0.0/16
2004 sslproxy_cert_error allow BrokenServersAtTrustedIP
2005 sslproxy_cert_error deny all
2007 This clause only supports fast acl types.
2008 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2009 Using slow acl types may result in server crashes
2011 Without this option, all server certificate validation errors
2012 terminate the transaction. Bypassing validation errors is dangerous
2013 because an error usually implies that the server cannot be trusted and
2014 the connection may be insecure.
2016 See also: sslproxy_flags and DONT_VERIFY_PEER.
2018 Default setting: sslproxy_cert_error deny all
2021 NAME: sslpassword_program
2024 LOC: Config.Program.ssl_password
2027 Specify a program used for entering SSL key passphrases
2028 when using encrypted SSL certificate keys. If not specified
2029 keys must either be unencrypted, or Squid started with the -N
2030 option to allow it to query interactively for the passphrase.
2032 The key file name is given as argument to the program allowing
2033 selection of the right password if you have multiple encrypted
2038 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2039 -----------------------------------------------------------------------------
2042 NAME: sslcrtd_program
2045 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2046 LOC: Ssl::TheConfig.ssl_crtd
2048 Specify the location and options of the executable for ssl_crtd process.
2049 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2050 For more information use:
2051 @DEFAULT_SSL_CRTD@ -h
2054 NAME: sslcrtd_children
2055 TYPE: HelperChildConfig
2057 DEFAULT: 32 startup=5 idle=1
2058 LOC: Ssl::TheConfig.ssl_crtdChildren
2060 The maximum number of processes spawn to service ssl server.
2061 The maximum this may be safely set to is 32.
2063 The startup= and idle= options allow some measure of skew in your
2068 Sets the minimum number of processes to spawn when Squid
2069 starts or reconfigures. When set to zero the first request will
2070 cause spawning of the first child process to handle it.
2072 Starting too few children temporary slows Squid under load while it
2073 tries to spawn enough additional processes to cope with traffic.
2077 Sets a minimum of how many processes Squid is to try and keep available
2078 at all times. When traffic begins to rise above what the existing
2079 processes can handle this many more will be spawned up to the maximum
2080 configured. A minimum setting of 1 is required.
2082 You must have at least one ssl_crtd process.
2086 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2087 -----------------------------------------------------------------------------
2095 To specify other caches in a hierarchy, use the format:
2097 cache_peer hostname type http-port icp-port [options]
2102 # hostname type port port options
2103 # -------------------- -------- ----- ----- -----------
2104 cache_peer parent.foo.net parent 3128 3130 default
2105 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2106 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2107 cache_peer example.com parent 80 0 default
2108 cache_peer cdn.example.com sibling 3128 0
2110 type: either 'parent', 'sibling', or 'multicast'.
2112 proxy-port: The port number where the peer accept HTTP requests.
2113 For other Squid proxies this is usually 3128
2114 For web servers this is usually 80
2116 icp-port: Used for querying neighbor caches about objects.
2117 Set to 0 if the peer does not support ICP or HTCP.
2118 See ICP and HTCP options below for additional details.
2121 ==== ICP OPTIONS ====
2123 You MUST also set icp_port and icp_access explicitly when using these options.
2124 The defaults will prevent peer traffic using ICP.
2127 no-query Disable ICP queries to this neighbor.
2130 Indicates the named peer is a member of a multicast group.
2131 ICP queries will not be sent directly to the peer, but ICP
2132 replies will be accepted from it.
2134 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2135 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2138 To only send ICP queries to this neighbor infrequently.
2139 This is used to keep the neighbor round trip time updated
2140 and is usually used in conjunction with weighted-round-robin.
2143 ==== HTCP OPTIONS ====
2145 You MUST also set htcp_port and htcp_access explicitly when using these options.
2146 The defaults will prevent peer traffic using HTCP.
2149 htcp Send HTCP, instead of ICP, queries to the neighbor.
2150 You probably also want to set the "icp-port" to 4827
2151 instead of 3130. This directive accepts a comma separated
2152 list of options described below.
2154 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2156 htcp=no-clr Send HTCP to the neighbor but without
2157 sending any CLR requests. This cannot be used with
2160 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2161 This cannot be used with no-clr.
2164 Send HTCP to the neighbor including CLRs but only when
2165 they do not result from PURGE requests.
2168 Forward any HTCP CLR requests this proxy receives to the peer.
2171 ==== PEER SELECTION METHODS ====
2173 The default peer selection method is ICP, with the first responding peer
2174 being used as source. These options can be used for better load balancing.
2177 default This is a parent cache which can be used as a "last-resort"
2178 if a peer cannot be located by any of the peer-selection methods.
2179 If specified more than once, only the first is used.
2181 round-robin Load-Balance parents which should be used in a round-robin
2182 fashion in the absence of any ICP queries.
2183 weight=N can be used to add bias.
2185 weighted-round-robin
2186 Load-Balance parents which should be used in a round-robin
2187 fashion with the frequency of each parent being based on the
2188 round trip time. Closer parents are used more often.
2189 Usually used for background-ping parents.
2190 weight=N can be used to add bias.
2192 carp Load-Balance parents which should be used as a CARP array.
2193 The requests will be distributed among the parents based on the
2194 CARP load balancing hash function based on their weight.
2196 userhash Load-balance parents based on the client proxy_auth or ident username.
2198 sourcehash Load-balance parents based on the client source IP.
2201 To be used only for cache peers of type "multicast".
2202 ALL members of this multicast group have "sibling"
2203 relationship with it, not "parent". This is to a multicast
2204 group when the requested object would be fetched only from
2205 a "parent" cache, anyway. It's useful, e.g., when
2206 configuring a pool of redundant Squid proxies, being
2207 members of the same multicast group.
2210 ==== PEER SELECTION OPTIONS ====
2212 weight=N use to affect the selection of a peer during any weighted
2213 peer-selection mechanisms.
2214 The weight must be an integer; default is 1,
2215 larger weights are favored more.
2216 This option does not affect parent selection if a peering
2217 protocol is not in use.
2219 basetime=N Specify a base amount to be subtracted from round trip
2221 It is subtracted before division by weight in calculating
2222 which parent to fectch from. If the rtt is less than the
2223 base time the rtt is set to a minimal value.
2225 ttl=N Specify a TTL to use when sending multicast ICP queries
2227 Only useful when sending to a multicast group.
2228 Because we don't accept ICP replies from random
2229 hosts, you must configure other group members as
2230 peers with the 'multicast-responder' option.
2232 no-delay To prevent access to this neighbor from influencing the
2235 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2236 enabled) for this host from the specified URL rather
2237 than the Squid default location.
2240 ==== CARP OPTIONS ====
2242 carp-key=key-specification
2243 use a different key than the full URL to hash against the peer.
2244 the key-specification is a comma-separated list of the keywords
2245 scheme, host, port, path, params
2246 Order is not important.
2248 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2250 originserver Causes this parent to be contacted as an origin server.
2251 Meant to be used in accelerator setups when the peer
2255 Set the Host header of requests forwarded to this peer.
2256 Useful in accelerator setups where the server (peer)
2257 expects a certain domain name but clients may request
2258 others. ie example.com or www.example.com
2260 no-digest Disable request of cache digests.
2263 Disables requesting ICMP RTT database (NetDB).
2266 ==== AUTHENTICATION OPTIONS ====
2269 If this is a personal/workgroup proxy and your parent
2270 requires proxy authentication.
2272 Note: The string can include URL escapes (i.e. %20 for
2273 spaces). This also means % must be written as %%.
2276 Send login details received from client to this peer.
2277 Both Proxy- and WWW-Authorization headers are passed
2278 without alteration to the peer.
2279 Authentication is not required by Squid for this to work.
2281 Note: This will pass any form of authentication but
2282 only Basic auth will work through a proxy unless the
2283 connection-auth options are also used.
2285 login=PASS Send login details received from client to this peer.
2286 Authentication is not required by this option.
2288 If there are no client-provided authentication headers
2289 to pass on, but username and password are available
2290 from an external ACL user= and password= result tags
2291 they may be sent instead.
2293 Note: To combine this with proxy_auth both proxies must
2294 share the same user database as HTTP only allows for
2295 a single login (one for proxy, one for origin server).
2296 Also be warned this will expose your users proxy
2297 password to the peer. USE WITH CAUTION
2300 Send the username to the upstream cache, but with a
2301 fixed password. This is meant to be used when the peer
2302 is in another administrative domain, but it is still
2303 needed to identify each user.
2304 The star can optionally be followed by some extra
2305 information which is added to the username. This can
2306 be used to identify this proxy to the peer, similar to
2307 the login=username:password option above.
2310 If this is a personal/workgroup proxy and your parent
2311 requires a secure proxy authentication.
2312 The first principal from the default keytab or defined by
2313 the environment variable KRB5_KTNAME will be used.
2315 WARNING: The connection may transmit requests from multiple
2316 clients. Negotiate often assumes end-to-end authentication
2317 and a single-client. Which is not strictly true here.
2319 login=NEGOTIATE:principal_name
2320 If this is a personal/workgroup proxy and your parent
2321 requires a secure proxy authentication.
2322 The principal principal_name from the default keytab or
2323 defined by the environment variable KRB5_KTNAME will be
2326 WARNING: The connection may transmit requests from multiple
2327 clients. Negotiate often assumes end-to-end authentication
2328 and a single-client. Which is not strictly true here.
2330 connection-auth=on|off
2331 Tell Squid that this peer does or not support Microsoft
2332 connection oriented authentication, and any such
2333 challenges received from there should be ignored.
2334 Default is auto to automatically determine the status
2338 ==== SSL / HTTPS / TLS OPTIONS ====
2340 ssl Encrypt connections to this peer with SSL/TLS.
2342 sslcert=/path/to/ssl/certificate
2343 A client SSL certificate to use when connecting to
2346 sslkey=/path/to/ssl/key
2347 The private SSL key corresponding to sslcert above.
2348 If 'sslkey' is not specified 'sslcert' is assumed to
2349 reference a combined file containing both the
2350 certificate and the key.
2353 The SSL version to use when connecting to this peer
2354 1 = automatic (default)
2359 sslcipher=... The list of valid SSL ciphers to use when connecting
2362 ssloptions=... Specify various SSL engine options:
2363 NO_SSLv2 Disallow the use of SSLv2
2364 NO_SSLv3 Disallow the use of SSLv3
2365 NO_TLSv1 Disallow the use of TLSv1
2366 See src/ssl_support.c or the OpenSSL documentation for
2367 a more complete list.
2369 sslcafile=... A file containing additional CA certificates to use
2370 when verifying the peer certificate.
2372 sslcapath=... A directory containing additional CA certificates to
2373 use when verifying the peer certificate.
2375 sslcrlfile=... A certificate revocation list file to use when
2376 verifying the peer certificate.
2378 sslflags=... Specify various flags modifying the SSL implementation:
2381 Accept certificates even if they fail to
2384 Don't use the default CA list built in
2387 Don't verify the peer certificate
2388 matches the server name
2390 ssldomain= The peer name as advertised in it's certificate.
2391 Used for verifying the correctness of the received peer
2392 certificate. If not specified the peer hostname will be
2396 Enable the "Front-End-Https: On" header needed when
2397 using Squid as a SSL frontend in front of Microsoft OWA.
2398 See MS KB document Q307347 for details on this header.
2399 If set to auto the header will only be added if the
2400 request is forwarded as a https:// URL.
2403 ==== GENERAL OPTIONS ====
2406 A peer-specific connect timeout.
2407 Also see the peer_connect_timeout directive.
2409 connect-fail-limit=N
2410 How many times connecting to a peer must fail before
2411 it is marked as down. Default is 10.
2413 allow-miss Disable Squid's use of only-if-cached when forwarding
2414 requests to siblings. This is primarily useful when
2415 icp_hit_stale is used by the sibling. To extensive use
2416 of this option may result in forwarding loops, and you
2417 should avoid having two-way peerings with this option.
2418 For example to deny peer usage on requests from peer
2419 by denying cache_peer_access if the source is a peer.
2421 max-conn=N Limit the amount of connections Squid may open to this
2424 name=xxx Unique name for the peer.
2425 Required if you have multiple peers on the same host
2426 but different ports.
2427 This name can be used in cache_peer_access and similar
2428 directives to dentify the peer.
2429 Can be used by outgoing access controls through the
2432 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2433 requests to this peer. Use normal address selection instead.
2435 proxy-only objects fetched from the peer will not be stored locally.
2439 NAME: cache_peer_domain cache_host_domain
2444 Use to limit the domains for which a neighbor cache will be
2447 cache_peer_domain cache-host domain [domain ...]
2448 cache_peer_domain cache-host !domain
2450 For example, specifying
2452 cache_peer_domain parent.foo.net .edu
2454 has the effect such that UDP query packets are sent to
2455 'bigserver' only when the requested object exists on a
2456 server in the .edu domain. Prefixing the domainname
2457 with '!' means the cache will be queried for objects
2460 NOTE: * Any number of domains may be given for a cache-host,
2461 either on the same or separate lines.
2462 * When multiple domains are given for a particular
2463 cache-host, the first matched domain is applied.
2464 * Cache hosts with no domain restrictions are queried
2466 * There are no defaults.
2467 * There is also a 'cache_peer_access' tag in the ACL
2471 NAME: cache_peer_access
2476 Similar to 'cache_peer_domain' but provides more flexibility by
2479 cache_peer_access cache-host allow|deny [!]aclname ...
2481 The syntax is identical to 'http_access' and the other lists of
2482 ACL elements. See the comments for 'http_access' below, or
2483 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2486 NAME: neighbor_type_domain
2487 TYPE: hostdomaintype
2491 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2493 Modifying the neighbor type for specific domains is now
2494 possible. You can treat some domains differently than the
2495 default neighbor type specified on the 'cache_peer' line.
2496 Normally it should only be necessary to list domains which
2497 should be treated differently because the default neighbor type
2498 applies for hostnames which do not match domains listed here.
2501 cache_peer cache.foo.org parent 3128 3130
2502 neighbor_type_domain cache.foo.org sibling .com .net
2503 neighbor_type_domain cache.foo.org sibling .au .de
2506 NAME: dead_peer_timeout
2510 LOC: Config.Timeout.deadPeer
2512 This controls how long Squid waits to declare a peer cache
2513 as "dead." If there are no ICP replies received in this
2514 amount of time, Squid will declare the peer dead and not
2515 expect to receive any further ICP replies. However, it
2516 continues to send ICP queries, and will mark the peer as
2517 alive upon receipt of the first subsequent ICP reply.
2519 This timeout also affects when Squid expects to receive ICP
2520 replies from peers. If more than 'dead_peer' seconds have
2521 passed since the last ICP reply was received, Squid will not
2522 expect to receive an ICP reply on the next query. Thus, if
2523 your time between requests is greater than this timeout, you
2524 will see a lot of requests sent DIRECT to origin servers
2525 instead of to your parents.
2528 NAME: forward_max_tries
2531 LOC: Config.forward_max_tries
2533 Controls how many different forward paths Squid will try
2534 before giving up. See also forward_timeout.
2536 NOTE: connect_retries (default: none) can make each of these
2537 possible forwarding paths be tried multiple times.
2540 NAME: hierarchy_stoplist
2543 LOC: Config.hierarchy_stoplist
2545 A list of words which, if found in a URL, cause the object to
2546 be handled directly by this cache. In other words, use this
2547 to not query neighbor caches for certain objects. You may
2548 list this option multiple times.
2551 hierarchy_stoplist cgi-bin ?
2553 Note: never_direct overrides this option.
2557 MEMORY CACHE OPTIONS
2558 -----------------------------------------------------------------------------
2565 LOC: Config.memMaxSize
2567 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2568 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2569 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2570 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2572 'cache_mem' specifies the ideal amount of memory to be used
2574 * In-Transit objects
2576 * Negative-Cached objects
2578 Data for these objects are stored in 4 KB blocks. This
2579 parameter specifies the ideal upper limit on the total size of
2580 4 KB blocks allocated. In-Transit objects take the highest
2583 In-transit objects have priority over the others. When
2584 additional space is needed for incoming data, negative-cached
2585 and hot objects will be released. In other words, the
2586 negative-cached and hot objects will fill up any unused space
2587 not needed for in-transit objects.
2589 If circumstances require, this limit will be exceeded.
2590 Specifically, if your incoming request rate requires more than
2591 'cache_mem' of memory to hold in-transit objects, Squid will
2592 exceed this limit to satisfy the new requests. When the load
2593 decreases, blocks will be freed until the high-water mark is
2594 reached. Thereafter, blocks will be used to store hot
2597 If shared memory caching is enabled, Squid does not use the shared
2598 cache space for in-transit objects, but they still consume as much
2599 local memory as they need. For more details about the shared memory
2600 cache, see memory_cache_shared.
2603 NAME: maximum_object_size_in_memory
2607 LOC: Config.Store.maxInMemObjSize
2609 Objects greater than this size will not be attempted to kept in
2610 the memory cache. This should be set high enough to keep objects
2611 accessed frequently in memory to improve performance whilst low
2612 enough to keep larger objects from hoarding cache_mem.
2615 NAME: memory_cache_shared
2618 LOC: Config.memShared
2620 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
2622 Controls whether the memory cache is shared among SMP workers.
2624 The shared memory cache is meant to occupy cache_mem bytes and replace
2625 the non-shared memory cache, although some entities may still be
2626 cached locally by workers for now (e.g., internal and in-transit
2627 objects may be served from a local memory cache even if shared memory
2628 caching is enabled).
2630 By default, the memory cache is shared if and only if all of the
2631 following conditions are satisfied: Squid runs in SMP mode with
2632 multiple workers, cache_mem is positive, and Squid environment
2633 supports required IPC primitives (e.g., POSIX shared memory segments
2634 and GCC-style atomic operations).
2636 To avoid blocking locks, shared memory uses opportunistic algorithms
2637 that do not guarantee that every cachable entity that could have been
2638 shared among SMP workers will actually be shared.
2640 Currently, entities exceeding 32KB in size cannot be shared.
2643 NAME: memory_cache_mode
2648 Controls which objects to keep in the memory cache (cache_mem)
2650 always Keep most recently fetched objects in memory (default)
2652 disk Only disk cache hits are kept in memory, which means
2653 an object must first be cached on disk and then hit
2654 a second time before cached in memory.
2656 network Only objects fetched from network is kept in memory
2659 NAME: memory_replacement_policy
2661 LOC: Config.memPolicy
2664 The memory replacement policy parameter determines which
2665 objects are purged from memory when memory space is needed.
2667 See cache_replacement_policy for details.
2672 -----------------------------------------------------------------------------
2675 NAME: cache_replacement_policy
2677 LOC: Config.replPolicy
2680 The cache replacement policy parameter determines which
2681 objects are evicted (replaced) when disk space is needed.
2683 lru : Squid's original list based LRU policy
2684 heap GDSF : Greedy-Dual Size Frequency
2685 heap LFUDA: Least Frequently Used with Dynamic Aging
2686 heap LRU : LRU policy implemented using a heap
2688 Applies to any cache_dir lines listed below this.
2690 The LRU policies keeps recently referenced objects.
2692 The heap GDSF policy optimizes object hit rate by keeping smaller
2693 popular objects in cache so it has a better chance of getting a
2694 hit. It achieves a lower byte hit rate than LFUDA though since
2695 it evicts larger (possibly popular) objects.
2697 The heap LFUDA policy keeps popular objects in cache regardless of
2698 their size and thus optimizes byte hit rate at the expense of
2699 hit rate since one large, popular object will prevent many
2700 smaller, slightly less popular objects from being cached.
2702 Both policies utilize a dynamic aging mechanism that prevents
2703 cache pollution that can otherwise occur with frequency-based
2704 replacement policies.
2706 NOTE: if using the LFUDA replacement policy you should increase
2707 the value of maximum_object_size above its default of 4096 KB to
2708 to maximize the potential byte hit rate improvement of LFUDA.
2710 For more information about the GDSF and LFUDA cache replacement
2711 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
2712 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
2718 LOC: Config.cacheSwap
2722 cache_dir Type Directory-Name Fs-specific-data [options]
2724 You can specify multiple cache_dir lines to spread the
2725 cache among different disk partitions.
2727 Type specifies the kind of storage system to use. Only "ufs"
2728 is built by default. To enable any of the other storage systems
2729 see the --enable-storeio configure option.
2731 'Directory' is a top-level directory where cache swap
2732 files will be stored. If you want to use an entire disk
2733 for caching, this can be the mount-point directory.
2734 The directory must exist and be writable by the Squid
2735 process. Squid will NOT create this directory for you.
2737 In SMP configurations, cache_dir must not precede the workers option
2738 and should use configuration macros or conditionals to give each
2739 worker interested in disk caching a dedicated cache directory.
2743 "ufs" is the old well-known Squid storage format that has always
2746 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
2748 'Mbytes' is the amount of disk space (MB) to use under this
2749 directory. The default is 100 MB. Change this to suit your
2750 configuration. Do NOT put the size of your disk drive here.
2751 Instead, if you want Squid to use the entire disk drive,
2752 subtract 20% and use that value.
2754 'L1' is the number of first-level subdirectories which
2755 will be created under the 'Directory'. The default is 16.
2757 'L2' is the number of second-level subdirectories which
2758 will be created under each first-level directory. The default
2761 The aufs store type:
2763 "aufs" uses the same storage format as "ufs", utilizing
2764 POSIX-threads to avoid blocking the main Squid process on
2765 disk-I/O. This was formerly known in Squid as async-io.
2767 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
2769 see argument descriptions under ufs above
2771 The diskd store type:
2773 "diskd" uses the same storage format as "ufs", utilizing a
2774 separate process to avoid blocking the main Squid process on
2777 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
2779 see argument descriptions under ufs above
2781 Q1 specifies the number of unacknowledged I/O requests when Squid
2782 stops opening new files. If this many messages are in the queues,
2783 Squid won't open new files. Default is 64
2785 Q2 specifies the number of unacknowledged messages when Squid
2786 starts blocking. If this many messages are in the queues,
2787 Squid blocks until it receives some replies. Default is 72
2789 When Q1 < Q2 (the default), the cache directory is optimized
2790 for lower response time at the expense of a decrease in hit
2791 ratio. If Q1 > Q2, the cache directory is optimized for
2792 higher hit ratio at the expense of an increase in response
2795 The rock store type:
2797 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
2799 The Rock Store type is a database-style storage. All cached
2800 entries are stored in a "database" file, using fixed-size slots,
2801 one entry per slot. The database size is specified in MB. The
2802 slot size is specified in bytes using the max-size option. See
2803 below for more info on the max-size option.
2805 swap-timeout=msec: Squid will not start writing a miss to or
2806 reading a hit from disk if it estimates that the swap operation
2807 will take more than the specified number of milliseconds. By
2808 default and when set to zero, disables the disk I/O time limit
2809 enforcement. Ignored when using blocking I/O module because
2810 blocking synchronous I/O does not allow Squid to estimate the
2811 expected swap wait time.
2813 max-swap-rate=swaps/sec: Artificially limits disk access using
2814 the specified I/O rate limit. Swap out requests that
2815 would cause the average I/O rate to exceed the limit are
2816 delayed. Individual swap in requests (i.e., hits or reads) are
2817 not delayed, but they do contribute to measured swap rate and
2818 since they are placed in the same FIFO queue as swap out
2819 requests, they may wait longer if max-swap-rate is smaller.
2820 This is necessary on file systems that buffer "too
2821 many" writes and then start blocking Squid and other processes
2822 while committing those writes to disk. Usually used together
2823 with swap-timeout to avoid excessive delays and queue overflows
2824 when disk demand exceeds available disk "bandwidth". By default
2825 and when set to zero, disables the disk I/O rate limit
2826 enforcement. Currently supported by IpcIo module only.
2829 The coss store type:
2831 NP: COSS filesystem in Squid-3 has been deemed too unstable for
2832 production use and has thus been removed from this release.
2833 We hope that it can be made usable again soon.
2835 block-size=n defines the "block size" for COSS cache_dir's.
2836 Squid uses file numbers as block numbers. Since file numbers
2837 are limited to 24 bits, the block size determines the maximum
2838 size of the COSS partition. The default is 512 bytes, which
2839 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
2840 you should not change the coss block size after Squid
2841 has written some objects to the cache_dir.
2843 The coss file store has changed from 2.5. Now it uses a file
2844 called 'stripe' in the directory names in the config - and
2845 this will be created by squid -z.
2849 no-store, no new objects should be stored to this cache_dir
2851 min-size=n, refers to the min object size in bytes this cache_dir
2852 will accept. It's used to restrict a cache_dir to only store
2853 large objects (e.g. aufs) while other storedirs are optimized
2854 for smaller objects (e.g. COSS). Defaults to 0.
2856 max-size=n, refers to the max object size in bytes this cache_dir
2857 supports. It is used to select the cache_dir to store the object.
2858 Note: To make optimal use of the max-size limits you should order
2859 the cache_dir lines with the smallest max-size value first and the
2860 ones with no max-size specification last.
2862 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
2863 which can be changed with the --with-coss-membuf-size=N configure
2867 # Uncomment and adjust the following to add a disk cache directory.
2868 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
2872 NAME: store_dir_select_algorithm
2874 LOC: Config.store_dir_select_algorithm
2877 Set this to 'round-robin' as an alternative.
2880 NAME: max_open_disk_fds
2882 LOC: Config.max_open_disk_fds
2885 To avoid having disk as the I/O bottleneck Squid can optionally
2886 bypass the on-disk cache if more than this amount of disk file
2887 descriptors are open.
2889 A value of 0 indicates no limit.
2892 NAME: minimum_object_size
2896 LOC: Config.Store.minObjectSize
2898 Objects smaller than this size will NOT be saved on disk. The
2899 value is specified in kilobytes, and the default is 0 KB, which
2900 means there is no minimum.
2903 NAME: maximum_object_size
2907 LOC: Config.Store.maxObjectSize
2909 Objects larger than this size will NOT be saved on disk. The
2910 value is specified in kilobytes, and the default is 4MB. If
2911 you wish to get a high BYTES hit ratio, you should probably
2912 increase this (one 32 MB object hit counts for 3200 10KB
2913 hits). If you wish to increase speed more than your want to
2914 save bandwidth you should leave this low.
2916 NOTE: if using the LFUDA replacement policy you should increase
2917 this value to maximize the byte hit rate improvement of LFUDA!
2918 See replacement_policy below for a discussion of this policy.
2921 NAME: cache_swap_low
2922 COMMENT: (percent, 0-100)
2925 LOC: Config.Swap.lowWaterMark
2928 NAME: cache_swap_high
2929 COMMENT: (percent, 0-100)
2932 LOC: Config.Swap.highWaterMark
2935 The low- and high-water marks for cache object replacement.
2936 Replacement begins when the swap (disk) usage is above the
2937 low-water mark and attempts to maintain utilization near the
2938 low-water mark. As swap utilization gets close to high-water
2939 mark object eviction becomes more aggressive. If utilization is
2940 close to the low-water mark less replacement is done each time.
2942 Defaults are 90% and 95%. If you have a large cache, 5% could be
2943 hundreds of MB. If this is the case you may wish to set these
2944 numbers closer together.
2949 -----------------------------------------------------------------------------
2959 logformat <name> <format specification>
2961 Defines an access log format.
2963 The <format specification> is a string with embedded % format codes
2965 % format codes all follow the same basic structure where all but
2966 the formatcode is optional. Output strings are automatically escaped
2967 as required according to their context and the output format
2968 modifiers are usually not needed, but can be specified if an explicit
2969 output format is desired.
2971 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
2973 " output in quoted string format
2974 [ output in squid text log format as used by log_mime_hdrs
2975 # output in URL quoted format
2980 width minimum and/or maximum field width:
2981 [width_min][.width_max]
2982 When minimum starts with 0, the field is zero-padded.
2983 String values exceeding maximum width are truncated.
2985 {arg} argument such as header name etc
2989 % a literal % character
2990 sn Unique sequence number per log line entry
2991 err_code The ID of an error response served by Squid or
2992 a similar internal error identifier.
2993 err_detail Additional err_code-dependent error information.
2995 Connection related format codes:
2997 >a Client source IP address
2999 >p Client source port
3000 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3001 >la Local IP address the client connected to
3002 >lp Local port number the client connected to
3004 la Local listening IP address the client connection was connected to.
3005 lp Local listening port number the client connection was connected to.
3007 <a Server IP address of the last server or peer connection
3008 <A Server FQDN or peer name
3009 <p Server port number of the last server or peer connection
3010 <la Local IP address of the last server or peer connection
3011 <lp Local port number of the last server or peer connection
3013 Time related format codes:
3015 ts Seconds since epoch
3016 tu subsecond time (milliseconds)
3017 tl Local time. Optional strftime format argument
3018 default %d/%b/%Y:%H:%M:%S %z
3019 tg GMT time. Optional strftime format argument
3020 default %d/%b/%Y:%H:%M:%S %z
3021 tr Response time (milliseconds)
3022 dt Total time spent making DNS lookups (milliseconds)
3024 Access Control related format codes:
3026 et Tag returned by external acl
3027 ea Log string returned by external acl
3028 un User name (any available)
3029 ul User name from authentication
3030 ue User name from external acl helper
3031 ui User name from ident
3032 us User name from SSL
3034 HTTP related format codes:
3036 [http::]>h Original request header. Optional header name argument
3037 on the format header[:[separator]element]
3038 [http::]>ha The HTTP request headers after adaptation and redirection.
3039 Optional header name argument as for >h
3040 [http::]<h Reply header. Optional header name argument
3042 [http::]>Hs HTTP status code sent to the client
3043 [http::]<Hs HTTP status code received from the next hop
3044 [http::]<bs Number of HTTP-equivalent message body bytes
3045 received from the next hop, excluding chunked
3046 transfer encoding and control messages.
3047 Generated FTP/Gopher listings are treated as
3049 [http::]mt MIME content type
3050 [http::]rm Request method (GET/POST etc)
3051 [http::]>rm Request method from client
3052 [http::]<rm Request method sent to server or peer
3053 [http::]ru Request URL from client (historic, filtered for logging)
3054 [http::]>ru Request URL from client
3055 [http::]<ru Request URL sent to server or peer
3056 [http::]rp Request URL-Path excluding hostname
3057 [http::]>rp Request URL-Path excluding hostname from client
3058 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3059 [http::]rv Request protocol version
3060 [http::]>rv Request protocol version from client
3061 [http::]<rv Request protocol version sent to server or peer
3062 [http::]<st Sent reply size including HTTP headers
3063 [http::]>st Received request size including HTTP headers. In the
3064 case of chunked requests the chunked encoding metadata
3066 [http::]>sh Received HTTP request headers size
3067 [http::]<sh Sent HTTP reply headers size
3068 [http::]st Request+Reply size including HTTP headers
3069 [http::]<sH Reply high offset sent
3070 [http::]<sS Upstream object size
3071 [http::]<pt Peer response time in milliseconds. The timer starts
3072 when the last request byte is sent to the next hop
3073 and stops when the last response byte is received.
3074 [http::]<tt Total server-side time in milliseconds. The timer
3075 starts with the first connect request (or write I/O)
3076 sent to the first selected peer. The timer stops
3077 with the last I/O with the last peer.
3079 Squid handling related format codes:
3081 Ss Squid request status (TCP_MISS etc)
3082 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3084 If ICAP is enabled, the following code becomes available (as
3085 well as ICAP log codes documented with the icap_log option):
3087 icap::tt Total ICAP processing time for the HTTP
3088 transaction. The timer ticks when ICAP
3089 ACLs are checked and when ICAP
3090 transaction is in progress.
3092 If adaptation is enabled the following three codes become available:
3094 adapt::<last_h The header of the last ICAP response or
3095 meta-information from the last eCAP
3096 transaction related to the HTTP transaction.
3097 Like <h, accepts an optional header name
3100 adapt::sum_trs Summed adaptation transaction response
3101 times recorded as a comma-separated list in
3102 the order of transaction start time. Each time
3103 value is recorded as an integer number,
3104 representing response time of one or more
3105 adaptation (ICAP or eCAP) transaction in
3106 milliseconds. When a failed transaction is
3107 being retried or repeated, its time is not
3108 logged individually but added to the
3109 replacement (next) transaction. See also:
3112 adapt::all_trs All adaptation transaction response times.
3113 Same as adaptation_strs but response times of
3114 individual transactions are never added
3115 together. Instead, all transaction response
3116 times are recorded individually.
3118 You can prefix adapt::*_trs format codes with adaptation
3119 service name in curly braces to record response time(s) specific
3120 to that service. For example: %{my_service}adapt::sum_trs
3122 The default formats available (which do not need re-defining) are:
3124 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3125 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3126 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3127 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3128 logformat useragent %>a [%tl] "%{User-Agent}>h"
3130 NOTE: When the log_mime_hdrs directive is set to ON.
3131 The squid, common and combined formats have a safely encoded copy
3132 of the mime headers appended to each line within a pair of brackets.
3134 NOTE: The common and combined formats are not quite true to the Apache definition.
3135 The logs from Squid contain an extra status and hierarchy code appended.
3139 NAME: access_log cache_access_log
3141 LOC: Config.Log.accesslogs
3142 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3144 These files log client request activities. Has a line every HTTP or
3145 ICP request. The format is:
3146 access_log <module>:<place> [<logformat name> [acl acl ...]]
3147 access_log none [acl acl ...]]
3149 Will log to the specified module:place using the specified format (which
3150 must be defined in a logformat directive) those entries which match
3151 ALL the acl's specified (which must be defined in acl clauses).
3152 If no acl is specified, all requests will be logged to this destination.
3154 ===== Modules Currently available =====
3156 none Do not log any requests matching these ACL.
3157 Do not specify Place or logformat name.
3159 stdio Write each log line to disk immediately at the completion of
3161 Place: the filename and path to be written.
3163 daemon Very similar to stdio. But instead of writing to disk the log
3164 line is passed to a daemon helper for asychronous handling instead.
3165 Place: varies depending on the daemon.
3167 log_file_daemon Place: the file name and path to be written.
3169 syslog To log each request via syslog facility.
3170 Place: The syslog facility and priority level for these entries.
3171 Place Format: facility.priority
3173 where facility could be any of:
3174 authpriv, daemon, local0 ... local7 or user.
3176 And priority could be any of:
3177 err, warning, notice, info, debug.
3179 udp To send each log line as text data to a UDP receiver.
3180 Place: The destination host name or IP and port.
3181 Place Format: //host:port
3183 tcp To send each log line as text data to a TCP receiver.
3184 Place: The destination host name or IP and port.
3185 Place Format: //host:port
3188 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3194 LOC: Config.Log.icaplogs
3197 ICAP log files record ICAP transaction summaries, one line per
3200 The icap_log option format is:
3201 icap_log <filepath> [<logformat name> [acl acl ...]]
3202 icap_log none [acl acl ...]]
3204 Please see access_log option documentation for details. The two
3205 kinds of logs share the overall configuration approach and many
3208 ICAP processing of a single HTTP message or transaction may
3209 require multiple ICAP transactions. In such cases, multiple
3210 ICAP transaction log lines will correspond to a single access
3213 ICAP log uses logformat codes that make sense for an ICAP
3214 transaction. Header-related codes are applied to the HTTP header
3215 embedded in an ICAP server response, with the following caveats:
3216 For REQMOD, there is no HTTP response header unless the ICAP
3217 server performed request satisfaction. For RESPMOD, the HTTP
3218 request header is the header sent to the ICAP server. For
3219 OPTIONS, there are no HTTP headers.
3221 The following format codes are also available for ICAP logs:
3223 icap::<A ICAP server IP address. Similar to <A.
3225 icap::<service_name ICAP service name from the icap_service
3226 option in Squid configuration file.
3228 icap::ru ICAP Request-URI. Similar to ru.
3230 icap::rm ICAP request method (REQMOD, RESPMOD, or
3231 OPTIONS). Similar to existing rm.
3233 icap::>st Bytes sent to the ICAP server (TCP payload
3234 only; i.e., what Squid writes to the socket).
3236 icap::<st Bytes received from the ICAP server (TCP
3237 payload only; i.e., what Squid reads from
3240 icap::<bs Number of message body bytes received from the
3241 ICAP server. ICAP message body, if any, usually
3242 includes encapsulated HTTP message headers and
3243 possibly encapsulated HTTP message body. The
3244 HTTP body part is dechunked before its size is
3247 icap::tr Transaction response time (in
3248 milliseconds). The timer starts when
3249 the ICAP transaction is created and
3250 stops when the transaction is completed.
3253 icap::tio Transaction I/O time (in milliseconds). The
3254 timer starts when the first ICAP request
3255 byte is scheduled for sending. The timers
3256 stops when the last byte of the ICAP response
3259 icap::to Transaction outcome: ICAP_ERR* for all
3260 transaction errors, ICAP_OPT for OPTION
3261 transactions, ICAP_ECHO for 204
3262 responses, ICAP_MOD for message
3263 modification, and ICAP_SAT for request
3264 satisfaction. Similar to Ss.
3266 icap::Hs ICAP response status code. Similar to Hs.
3268 icap::>h ICAP request header(s). Similar to >h.
3270 icap::<h ICAP response header(s). Similar to <h.
3272 The default ICAP log format, which can be used without an explicit
3273 definition, is called icap_squid:
3275 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3277 See also: logformat, log_icap, and %adapt::<last_h
3280 NAME: logfile_daemon
3282 DEFAULT: @DEFAULT_LOGFILED@
3283 LOC: Log::TheConfig.logfile_daemon
3285 Specify the path to the logfile-writing daemon. This daemon is
3286 used to write the access and store logs, if configured.
3288 Squid sends a number of commands to the log daemon:
3289 L<data>\n - logfile data
3294 r<n>\n - set rotate count to <n>
3295 b<n>\n - 1 = buffer output, 0 = don't buffer output
3297 No responses is expected.
3302 LOC: Config.accessList.log
3304 COMMENT: allow|deny acl acl...
3306 This options allows you to control which requests gets logged
3307 to access.log (see access_log directive). Requests denied for
3308 logging will also not be accounted for in performance counters.
3310 This clause only supports fast acl types.
3311 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3317 LOC: Config.accessList.icap
3320 This options allows you to control which requests get logged
3321 to icap.log. See the icap_log directive for ICAP log details.
3324 NAME: cache_store_log
3327 LOC: Config.Log.store
3329 Logs the activities of the storage manager. Shows which
3330 objects are ejected from the cache, and which objects are
3331 saved and for how long. To disable, enter "none" or remove the line.
3332 There are not really utilities to analyze this data, so you can safely
3336 cache_store_log @DEFAULT_STORE_LOG@
3339 NAME: cache_swap_state cache_swap_log
3341 LOC: Config.Log.swap
3344 Location for the cache "swap.state" file. This index file holds
3345 the metadata of objects saved on disk. It is used to rebuild
3346 the cache during startup. Normally this file resides in each
3347 'cache_dir' directory, but you may specify an alternate
3348 pathname here. Note you must give a full filename, not just
3349 a directory. Since this is the index for the whole object
3350 list you CANNOT periodically rotate it!
3352 If %s can be used in the file name it will be replaced with a
3353 a representation of the cache_dir name where each / is replaced
3354 with '.'. This is needed to allow adding/removing cache_dir
3355 lines when cache_swap_log is being used.
3357 If have more than one 'cache_dir', and %s is not used in the name
3358 these swap logs will have names such as:
3364 The numbered extension (which is added automatically)
3365 corresponds to the order of the 'cache_dir' lines in this
3366 configuration file. If you change the order of the 'cache_dir'
3367 lines in this file, these index files will NOT correspond to
3368 the correct 'cache_dir' entry (unless you manually rename
3369 them). We recommend you do NOT use this option. It is
3370 better to keep these index files in each 'cache_dir' directory.
3373 NAME: logfile_rotate
3376 LOC: Config.Log.rotateNumber
3378 Specifies the number of logfile rotations to make when you
3379 type 'squid -k rotate'. The default is 10, which will rotate
3380 with extensions 0 through 9. Setting logfile_rotate to 0 will
3381 disable the file name rotation, but the logfiles are still closed
3382 and re-opened. This will enable you to rename the logfiles
3383 yourself just before sending the rotate signal.
3385 Note, the 'squid -k rotate' command normally sends a USR1
3386 signal to the running squid process. In certain situations
3387 (e.g. on Linux with Async I/O), USR1 is used for other
3388 purposes, so -k rotate uses another signal. It is best to get
3389 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3392 Note, from Squid-3.1 this option has no effect on the cache.log,
3393 that log can be rotated separately by using debug_options
3396 NAME: emulate_httpd_log
3399 Replace this with an access_log directive using the format 'common' or 'combined'.
3402 NAME: log_ip_on_direct
3405 Remove this option from your config. To log server or peer names use %<A in the log format.
3410 DEFAULT: @DEFAULT_MIME_TABLE@
3411 LOC: Config.mimeTablePathname
3413 Pathname to Squid's MIME table. You shouldn't need to change
3414 this, but the default file contains examples and formatting
3415 information if you do.
3421 LOC: Config.onoff.log_mime_hdrs
3424 The Cache can record both the request and the response MIME
3425 headers for each HTTP transaction. The headers are encoded
3426 safely and will appear as two bracketed fields at the end of
3427 the access log (for either the native or httpd-emulated log
3428 formats). To enable this logging set log_mime_hdrs to 'on'.
3434 Replace this with an access_log directive using the format 'useragent'.
3437 NAME: referer_log referrer_log
3440 Replace this with an access_log directive using the format 'referrer'.
3445 DEFAULT: @DEFAULT_PID_FILE@
3446 LOC: Config.pidFilename
3448 A filename to write the process-id to. To disable, enter "none".
3454 Remove this option from your config. To log FQDN use %>A in the log format.
3457 NAME: client_netmask
3459 LOC: Config.Addrs.client_netmask
3462 A netmask for client addresses in logfiles and cachemgr output.
3463 Change this to protect the privacy of your cache clients.
3464 A netmask of 255.255.255.0 will log all IP's in that range with
3465 the last digit set to '0'.
3471 Use a regular access.log with ACL limiting it to MISS events.
3474 NAME: strip_query_terms
3476 LOC: Config.onoff.strip_query_terms
3479 By default, Squid strips query terms from requested URLs before
3480 logging. This protects your user's privacy.
3487 LOC: Config.onoff.buffered_logs
3489 cache.log log file is written with stdio functions, and as such
3490 it can be buffered or unbuffered. By default it will be unbuffered.
3491 Buffering it can speed up the writing slightly (though you are
3492 unlikely to need to worry unless you run with tons of debugging
3493 enabled in which case performance will suffer badly anyway..).
3496 NAME: netdb_filename
3498 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3499 LOC: Config.netdbFilename
3502 A filename where Squid stores it's netdb state between restarts.
3503 To disable, enter "none".
3507 OPTIONS FOR TROUBLESHOOTING
3508 -----------------------------------------------------------------------------
3513 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3514 LOC: Debug::cache_log
3516 Cache logging file. This is where general information about
3517 your cache's behavior goes. You can increase the amount of data
3518 logged to this file and how often its rotated with "debug_options"
3524 LOC: Debug::debugOptions
3526 Logging options are set as section,level where each source file
3527 is assigned a unique section. Lower levels result in less
3528 output, Full debugging (level 9) can result in a very large
3529 log file, so be careful.
3531 The magic word "ALL" sets debugging levels for all sections.
3532 We recommend normally running with "ALL,1".
3534 The rotate=N option can be used to keep more or less of these logs
3535 than would otherwise be kept by logfile_rotate.
3536 For most uses a single log should be enough to monitor current
3537 events affecting Squid.
3542 LOC: Config.coredump_dir
3543 DEFAULT_IF_NONE: none
3545 By default Squid leaves core files in the directory from where
3546 it was started. If you set 'coredump_dir' to a directory
3547 that exists, Squid will chdir() to that directory at startup
3548 and coredump files will be left there.
3552 # Leave coredumps in the first cache dir
3553 coredump_dir @DEFAULT_SWAP_DIR@
3559 OPTIONS FOR FTP GATEWAYING
3560 -----------------------------------------------------------------------------
3566 LOC: Config.Ftp.anon_user
3568 If you want the anonymous login password to be more informative
3569 (and enable the use of picky ftp servers), set this to something
3570 reasonable for your domain, like wwwuser@somewhere.net
3572 The reason why this is domainless by default is the
3573 request can be made on the behalf of a user in any domain,
3574 depending on how the cache is used.
3575 Some ftp server also validate the email address is valid
3576 (for example perl.com).
3582 LOC: Config.Ftp.passive
3584 If your firewall does not allow Squid to use passive
3585 connections, turn off this option.
3587 Use of ftp_epsv_all option requires this to be ON.
3593 LOC: Config.Ftp.epsv_all
3595 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
3597 NATs may be able to put the connection on a "fast path" through the
3598 translator, as the EPRT command will never be used and therefore,
3599 translation of the data portion of the segments will never be needed.
3601 When a client only expects to do two-way FTP transfers this may be
3603 If squid finds that it must do a three-way FTP transfer after issuing
3604 an EPSV ALL command, the FTP session will fail.
3606 If you have any doubts about this option do not use it.
3607 Squid will nicely attempt all other connection methods.
3609 Requires ftp_passive to be ON (default) for any effect.
3615 LOC: Config.Ftp.epsv
3617 FTP Protocol extensions permit the use of a special "EPSV" command.
3619 NATs may be able to put the connection on a "fast path" through the
3620 translator using EPSV, as the EPRT command will never be used
3621 and therefore, translation of the data portion of the segments
3622 will never be needed.
3624 Turning this OFF will prevent EPSV being attempted.
3625 WARNING: Doing so will convert Squid back to the old behavior with all
3626 the related problems with external NAT devices/layers.
3628 Requires ftp_passive to be ON (default) for any effect.
3634 LOC: Config.Ftp.eprt
3636 FTP Protocol extensions permit the use of a special "EPRT" command.
3638 This extension provides a protocol neutral alternative to the
3639 IPv4-only PORT command. When supported it enables active FTP data
3640 channels over IPv6 and efficient NAT handling.
3642 Turning this OFF will prevent EPRT being attempted and will skip
3643 straight to using PORT for IPv4 servers.
3645 Some devices are known to not handle this extension correctly and
3646 may result in crashes. Devices which suport EPRT enough to fail
3647 cleanly will result in Squid attempting PORT anyway. This directive
3648 should only be disabled when EPRT results in device failures.
3650 WARNING: Doing so will convert Squid back to the old behavior with all
3651 the related problems with external NAT devices/layers and IPv4-only FTP.
3654 NAME: ftp_sanitycheck
3657 LOC: Config.Ftp.sanitycheck
3659 For security and data integrity reasons Squid by default performs
3660 sanity checks of the addresses of FTP data connections ensure the
3661 data connection is to the requested server. If you need to allow
3662 FTP connections to servers using another IP address for the data
3663 connection turn this off.
3666 NAME: ftp_telnet_protocol
3669 LOC: Config.Ftp.telnet
3671 The FTP protocol is officially defined to use the telnet protocol
3672 as transport channel for the control connection. However, many
3673 implementations are broken and does not respect this aspect of
3676 If you have trouble accessing files with ASCII code 255 in the
3677 path or similar problems involving this ASCII code you can
3678 try setting this directive to off. If that helps, report to the
3679 operator of the FTP server in question that their FTP server
3680 is broken and does not follow the FTP standard.
3684 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
3685 -----------------------------------------------------------------------------
3690 DEFAULT: @DEFAULT_DISKD@
3691 LOC: Config.Program.diskd
3693 Specify the location of the diskd executable.
3694 Note this is only useful if you have compiled in
3695 diskd as one of the store io modules.
3698 NAME: unlinkd_program
3701 DEFAULT: @DEFAULT_UNLINKD@
3702 LOC: Config.Program.unlinkd
3704 Specify the location of the executable for file deletion process.
3707 NAME: pinger_program
3709 DEFAULT: @DEFAULT_PINGER@
3710 LOC: Config.pinger.program
3713 Specify the location of the executable for the pinger process.
3719 LOC: Config.pinger.enable
3722 Control whether the pinger is active at run-time.
3723 Enables turning ICMP pinger on and off with a simple
3724 squid -k reconfigure.
3729 OPTIONS FOR URL REWRITING
3730 -----------------------------------------------------------------------------
3733 NAME: url_rewrite_program redirect_program
3735 LOC: Config.Program.redirect
3738 Specify the location of the executable URL rewriter to use.
3739 Since they can perform almost any function there isn't one included.
3741 For each requested URL, the rewriter will receive on line with the format
3743 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
3745 In the future, the rewriter interface will be extended with
3746 key=value pairs ("kvpairs" shown above). Rewriter programs
3747 should be prepared to receive and possibly ignore additional
3748 whitespace-separated tokens on each input line.
3750 And the rewriter may return a rewritten URL. The other components of
3751 the request line does not need to be returned (ignored if they are).
3753 The rewriter can also indicate that a client-side redirect should
3754 be performed to the new URL. This is done by prefixing the returned
3755 URL with "301:" (moved permanently) or 302: (moved temporarily), etc.
3757 By default, a URL rewriter is not used.
3760 NAME: url_rewrite_children redirect_children
3761 TYPE: HelperChildConfig
3762 DEFAULT: 20 startup=0 idle=1 concurrency=0
3763 LOC: Config.redirectChildren
3765 The maximum number of redirector processes to spawn. If you limit
3766 it too few Squid will have to wait for them to process a backlog of
3767 URLs, slowing it down. If you allow too many they will use RAM
3768 and other system resources noticably.
3770 The startup= and idle= options allow some measure of skew in your
3775 Sets a minimum of how many processes are to be spawned when Squid
3776 starts or reconfigures. When set to zero the first request will
3777 cause spawning of the first child process to handle it.
3779 Starting too few will cause an initial slowdown in traffic as Squid
3780 attempts to simultaneously spawn enough processes to cope.
3784 Sets a minimum of how many processes Squid is to try and keep available
3785 at all times. When traffic begins to rise above what the existing
3786 processes can handle this many more will be spawned up to the maximum
3787 configured. A minimum setting of 1 is required.
3791 The number of requests each redirector helper can handle in
3792 parallel. Defaults to 0 which indicates the redirector
3793 is a old-style single threaded redirector.
3795 When this directive is set to a value >= 1 then the protocol
3796 used to communicate with the helper is modified to include
3797 a request ID in front of the request/response. The request
3798 ID from the request must be echoed back with the response
3802 NAME: url_rewrite_host_header redirect_rewrites_host_header
3805 LOC: Config.onoff.redir_rewrites_host
3807 To preserve same-origin security policies in browsers and
3808 prevent Host: header forgery by redirectors Squid rewrites
3809 any Host: header in redirected requests.
3811 If you are running an accelerator this may not be a wanted
3812 effect of a redirector. This directive enables you disable
3813 Host: alteration in reverse-proxy traffic.
3815 WARNING: Entries are cached on the result of the URL rewriting
3816 process, so be careful if you have domain-virtual hosts.
3818 WARNING: Squid and other software verifies the URL and Host
3819 are matching, so be careful not to relay through other proxies
3820 or inspecting firewalls with this disabled.
3823 NAME: url_rewrite_access redirector_access
3826 LOC: Config.accessList.redirector
3828 If defined, this access list specifies which requests are
3829 sent to the redirector processes. By default all requests
3832 This clause supports both fast and slow acl types.
3833 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3836 NAME: url_rewrite_bypass redirector_bypass
3838 LOC: Config.onoff.redirector_bypass
3841 When this is 'on', a request will not go through the
3842 redirector if all redirectors are busy. If this is 'off'
3843 and the redirector queue grows too large, Squid will exit
3844 with a FATAL error and ask you to increase the number of
3845 redirectors. You should only enable this if the redirectors
3846 are not critical to your caching system. If you use
3847 redirectors for access control, and you enable this option,
3848 users may have access to pages they should not
3849 be allowed to request.
3853 OPTIONS FOR TUNING THE CACHE
3854 -----------------------------------------------------------------------------
3857 NAME: cache no_cache
3860 LOC: Config.accessList.noCache
3862 A list of ACL elements which, if matched and denied, cause the request to
3863 not be satisfied from the cache and the reply to not be cached.
3864 In other words, use this to force certain objects to never be cached.
3866 You must use the words 'allow' or 'deny' to indicate whether items
3867 matching the ACL should be allowed or denied into the cache.
3869 Default is to allow all to be cached.
3871 This clause supports both fast and slow acl types.
3872 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3878 LOC: Config.maxStale
3881 This option puts an upper limit on how stale content Squid
3882 will serve from the cache if cache validation fails.
3883 Can be overriden by the refresh_pattern max-stale option.
3886 NAME: refresh_pattern
3887 TYPE: refreshpattern
3891 usage: refresh_pattern [-i] regex min percent max [options]
3893 By default, regular expressions are CASE-SENSITIVE. To make
3894 them case-insensitive, use the -i option.
3896 'Min' is the time (in minutes) an object without an explicit
3897 expiry time should be considered fresh. The recommended
3898 value is 0, any higher values may cause dynamic applications
3899 to be erroneously cached unless the application designer
3900 has taken the appropriate actions.
3902 'Percent' is a percentage of the objects age (time since last
3903 modification age) an object without explicit expiry time
3904 will be considered fresh.
3906 'Max' is an upper limit on how long objects without an explicit
3907 expiry time will be considered fresh.
3909 options: override-expire
3915 ignore-must-revalidate
3922 override-expire enforces min age even if the server
3923 sent an explicit expiry time (e.g., with the
3924 Expires: header or Cache-Control: max-age). Doing this
3925 VIOLATES the HTTP standard. Enabling this feature
3926 could make you liable for problems which it causes.
3928 Note: override-expire does not enforce staleness - it only extends
3929 freshness / min. If the server returns a Expires time which
3930 is longer than your max time, Squid will still consider
3931 the object fresh for that period of time.
3933 override-lastmod enforces min age even on objects
3934 that were modified recently.
3936 reload-into-ims changes client no-cache or ``reload''
3937 to If-Modified-Since requests. Doing this VIOLATES the
3938 HTTP standard. Enabling this feature could make you
3939 liable for problems which it causes.
3941 ignore-reload ignores a client no-cache or ``reload''
3942 header. Doing this VIOLATES the HTTP standard. Enabling
3943 this feature could make you liable for problems which
3946 ignore-no-cache ignores any ``Pragma: no-cache'' and
3947 ``Cache-control: no-cache'' headers received from a server.
3948 The HTTP RFC never allows the use of this (Pragma) header
3949 from a server, only a client, though plenty of servers
3952 ignore-no-store ignores any ``Cache-control: no-store''
3953 headers received from a server. Doing this VIOLATES
3954 the HTTP standard. Enabling this feature could make you
3955 liable for problems which it causes.
3957 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
3958 headers received from a server. Doing this VIOLATES
3959 the HTTP standard. Enabling this feature could make you
3960 liable for problems which it causes.
3962 ignore-private ignores any ``Cache-control: private''
3963 headers received from a server. Doing this VIOLATES
3964 the HTTP standard. Enabling this feature could make you
3965 liable for problems which it causes.
3967 ignore-auth caches responses to requests with authorization,
3968 as if the originserver had sent ``Cache-control: public''
3969 in the response header. Doing this VIOLATES the HTTP standard.
3970 Enabling this feature could make you liable for problems which
3973 refresh-ims causes squid to contact the origin server
3974 when a client issues an If-Modified-Since request. This
3975 ensures that the client will receive an updated version
3976 if one is available.
3978 store-stale stores responses even if they don't have explicit
3979 freshness or a validator (i.e., Last-Modified or an ETag)
3980 present, or if they're already stale. By default, Squid will
3981 not cache such responses because they usually can't be
3982 reused. Note that such responses will be stale by default.
3984 max-stale=NN provide a maximum staleness factor. Squid won't
3985 serve objects more stale than this even if it failed to
3986 validate the object. Default: use the max_stale global limit.
3988 Basically a cached object is:
3990 FRESH if expires < now, else STALE
3992 FRESH if lm-factor < percent, else STALE
3996 The refresh_pattern lines are checked in the order listed here.
3997 The first entry which matches is used. If none of the entries
3998 match the default will be used.
4000 Note, you must uncomment all the default lines if you want
4001 to change one. The default setting is only active if none is
4006 # Add any of your own refresh_pattern entries above these.
4007 refresh_pattern ^ftp: 1440 20% 10080
4008 refresh_pattern ^gopher: 1440 0% 1440
4009 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4010 refresh_pattern . 0 20% 4320
4014 NAME: quick_abort_min
4018 LOC: Config.quickAbort.min
4021 NAME: quick_abort_max
4025 LOC: Config.quickAbort.max
4028 NAME: quick_abort_pct
4032 LOC: Config.quickAbort.pct
4034 The cache by default continues downloading aborted requests
4035 which are almost completed (less than 16 KB remaining). This
4036 may be undesirable on slow (e.g. SLIP) links and/or very busy
4037 caches. Impatient users may tie up file descriptors and
4038 bandwidth by repeatedly requesting and immediately aborting
4041 When the user aborts a request, Squid will check the
4042 quick_abort values to the amount of data transfered until
4045 If the transfer has less than 'quick_abort_min' KB remaining,
4046 it will finish the retrieval.
4048 If the transfer has more than 'quick_abort_max' KB remaining,
4049 it will abort the retrieval.
4051 If more than 'quick_abort_pct' of the transfer has completed,
4052 it will finish the retrieval.
4054 If you do not want any retrieval to continue after the client
4055 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4058 If you want retrievals to always continue if they are being
4059 cached set 'quick_abort_min' to '-1 KB'.
4062 NAME: read_ahead_gap
4063 COMMENT: buffer-size
4065 LOC: Config.readAheadGap
4068 The amount of data the cache will buffer ahead of what has been
4069 sent to the client when retrieving an object from another server.
4073 IFDEF: USE_HTTP_VIOLATIONS
4076 LOC: Config.negativeTtl
4079 Set the Default Time-to-Live (TTL) for failed requests.
4080 Certain types of failures (such as "connection refused" and
4081 "404 Not Found") are able to be negatively-cached for a short time.
4082 Modern web servers should provide Expires: header, however if they
4083 do not this can provide a minimum TTL.
4084 The default is not to cache errors with unknown expiry details.
4086 Note that this is different from negative caching of DNS lookups.
4088 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4089 this feature could make you liable for problems which it
4093 NAME: positive_dns_ttl
4096 LOC: Config.positiveDnsTtl
4099 Upper limit on how long Squid will cache positive DNS responses.
4100 Default is 6 hours (360 minutes). This directive must be set
4101 larger than negative_dns_ttl.
4104 NAME: negative_dns_ttl
4107 LOC: Config.negativeDnsTtl
4110 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4111 This also sets the lower cache limit on positive lookups.
4112 Minimum value is 1 second, and it is not recommendable to go
4113 much below 10 seconds.
4116 NAME: range_offset_limit
4117 COMMENT: size [acl acl...]
4119 LOC: Config.rangeOffsetLimit
4122 usage: (size) [units] [[!]aclname]
4124 Sets an upper limit on how far (number of bytes) into the file
4125 a Range request may be to cause Squid to prefetch the whole file.
4126 If beyond this limit, Squid forwards the Range request as it is and
4127 the result is NOT cached.
4129 This is to stop a far ahead range request (lets say start at 17MB)
4130 from making Squid fetch the whole object up to that point before
4131 sending anything to the client.
4133 Multiple range_offset_limit lines may be specified, and they will
4134 be searched from top to bottom on each request until a match is found.
4135 The first match found will be used. If no line matches a request, the
4136 default limit of 0 bytes will be used.
4138 'size' is the limit specified as a number of units.
4140 'units' specifies whether to use bytes, KB, MB, etc.
4141 If no units are specified bytes are assumed.
4143 A size of 0 causes Squid to never fetch more than the
4144 client requested. (default)
4146 A size of 'none' causes Squid to always fetch the object from the
4147 beginning so it may cache the result. (2.0 style)
4149 'aclname' is the name of a defined ACL.
4151 NP: Using 'none' as the byte value here will override any quick_abort settings
4152 that may otherwise apply to the range request. The range request will
4153 be fully fetched from start to finish regardless of the client
4154 actions. This affects bandwidth usage.
4157 NAME: minimum_expiry_time
4160 LOC: Config.minimum_expiry_time
4163 The minimum caching time according to (Expires - Date)
4164 Headers Squid honors if the object can't be revalidated
4165 defaults to 60 seconds. In reverse proxy environments it
4166 might be desirable to honor shorter object lifetimes. It
4167 is most likely better to make your server return a
4168 meaningful Last-Modified header however. In ESI environments
4169 where page fragments often have short lifetimes, this will
4170 often be best set to 0.
4173 NAME: store_avg_object_size
4177 LOC: Config.Store.avgObjectSize
4179 Average object size, used to estimate number of objects your
4180 cache can hold. The default is 13 KB.
4183 NAME: store_objects_per_bucket
4186 LOC: Config.Store.objectsPerBucket
4188 Target number of objects per bucket in the store hash table.
4189 Lowering this value increases the total number of buckets and
4190 also the storage maintenance rate. The default is 20.
4195 -----------------------------------------------------------------------------
4198 NAME: request_header_max_size
4202 LOC: Config.maxRequestHeaderSize
4204 This specifies the maximum size for HTTP headers in a request.
4205 Request headers are usually relatively small (about 512 bytes).
4206 Placing a limit on the request header size will catch certain
4207 bugs (for example with persistent connections) and possibly
4208 buffer-overflow or denial-of-service attacks.
4211 NAME: reply_header_max_size
4215 LOC: Config.maxReplyHeaderSize
4217 This specifies the maximum size for HTTP headers in a reply.
4218 Reply headers are usually relatively small (about 512 bytes).
4219 Placing a limit on the reply header size will catch certain
4220 bugs (for example with persistent connections) and possibly
4221 buffer-overflow or denial-of-service attacks.
4224 NAME: request_body_max_size
4228 LOC: Config.maxRequestBodySize
4230 This specifies the maximum size for an HTTP request body.
4231 In other words, the maximum size of a PUT/POST request.
4232 A user who attempts to send a request with a body larger
4233 than this limit receives an "Invalid Request" error message.
4234 If you set this parameter to a zero (the default), there will
4235 be no limit imposed.
4238 NAME: client_request_buffer_max_size
4242 LOC: Config.maxRequestBufferSize
4244 This specifies the maximum buffer size of a client request.
4245 It prevents squid eating too much memory when somebody uploads
4249 NAME: chunked_request_body_max_size
4253 LOC: Config.maxChunkedRequestBodySize
4255 A broken or confused HTTP/1.1 client may send a chunked HTTP
4256 request to Squid. Squid does not have full support for that
4257 feature yet. To cope with such requests, Squid buffers the
4258 entire request and then dechunks request body to create a
4259 plain HTTP/1.0 request with a known content length. The plain
4260 request is then used by the rest of Squid code as usual.
4262 The option value specifies the maximum size of the buffer used
4263 to hold the request before the conversion. If the chunked
4264 request size exceeds the specified limit, the conversion
4265 fails, and the client receives an "unsupported request" error,
4266 as if dechunking was disabled.
4268 Dechunking is enabled by default. To disable conversion of
4269 chunked requests, set the maximum to zero.
4271 Request dechunking feature and this option in particular are a
4272 temporary hack. When chunking requests and responses are fully
4273 supported, there will be no need to buffer a chunked request.
4277 IFDEF: USE_HTTP_VIOLATIONS
4280 LOC: Config.accessList.brokenPosts
4282 A list of ACL elements which, if matched, causes Squid to send
4283 an extra CRLF pair after the body of a PUT/POST request.
4285 Some HTTP servers has broken implementations of PUT/POST,
4286 and rely on an extra CRLF pair sent by some WWW clients.
4288 Quote from RFC2616 section 4.1 on this matter:
4290 Note: certain buggy HTTP/1.0 client implementations generate an
4291 extra CRLF's after a POST request. To restate what is explicitly
4292 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4293 a request with an extra CRLF.
4295 This clause only supports fast acl types.
4296 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4299 acl buggy_server url_regex ^http://....
4300 broken_posts allow buggy_server
4303 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4306 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4308 LOC: Adaptation::Config::use_indirect_client
4310 Controls whether the indirect client IP address (instead of the direct
4311 client IP address) is passed to adaptation services.
4313 See also: follow_x_forwarded_for adaptation_send_client_ip
4317 IFDEF: USE_HTTP_VIOLATIONS
4321 LOC: Config.onoff.via
4323 If set (default), Squid will include a Via header in requests and
4324 replies as required by RFC2616.
4330 LOC: Config.onoff.ie_refresh
4333 Microsoft Internet Explorer up until version 5.5 Service
4334 Pack 1 has an issue with transparent proxies, wherein it
4335 is impossible to force a refresh. Turning this on provides
4336 a partial fix to the problem, by causing all IMS-REFRESH
4337 requests from older IE versions to check the origin server
4338 for fresh content. This reduces hit ratio by some amount
4339 (~10% in my experience), but allows users to actually get
4340 fresh content when they want it. Note because Squid
4341 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4342 of 5.5 is unchanged from old versions of Squid (i.e. a
4343 forced refresh is impossible). Newer versions of IE will,
4344 hopefully, continue to have the new behavior and will be
4345 handled based on that assumption. This option defaults to
4346 the old Squid behavior, which is better for hit ratios but
4347 worse for clients using IE, if they need to be able to
4348 force fresh content.
4351 NAME: vary_ignore_expire
4354 LOC: Config.onoff.vary_ignore_expire
4357 Many HTTP servers supporting Vary gives such objects
4358 immediate expiry time with no cache-control header
4359 when requested by a HTTP/1.0 client. This option
4360 enables Squid to ignore such expiry times until
4361 HTTP/1.1 is fully implemented.
4363 WARNING: If turned on this may eventually cause some
4364 varying objects not intended for caching to get cached.
4367 NAME: request_entities
4369 LOC: Config.onoff.request_entities
4372 Squid defaults to deny GET and HEAD requests with request entities,
4373 as the meaning of such requests are undefined in the HTTP standard
4374 even if not explicitly forbidden.
4376 Set this directive to on if you have clients which insists
4377 on sending request entities in GET or HEAD requests. But be warned
4378 that there is server software (both proxies and web servers) which
4379 can fail to properly process this kind of request which may make you
4380 vulnerable to cache pollution attacks if enabled.
4383 NAME: request_header_access
4384 IFDEF: USE_HTTP_VIOLATIONS
4385 TYPE: http_header_access[]
4386 LOC: Config.request_header_access
4389 Usage: request_header_access header_name allow|deny [!]aclname ...
4391 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4392 this feature could make you liable for problems which it
4395 This option replaces the old 'anonymize_headers' and the
4396 older 'http_anonymizer' option with something that is much
4397 more configurable. This new method creates a list of ACLs
4398 for each header, allowing you very fine-tuned header
4401 This option only applies to request headers, i.e., from the
4402 client to the server.
4404 You can only specify known headers for the header name.
4405 Other headers are reclassified as 'Other'. You can also
4406 refer to all the headers with 'All'.
4408 For example, to achieve the same behavior as the old
4409 'http_anonymizer standard' option, you should use:
4411 request_header_access From deny all
4412 request_header_access Referer deny all
4413 request_header_access Server deny all
4414 request_header_access User-Agent deny all
4415 request_header_access WWW-Authenticate deny all
4416 request_header_access Link deny all
4418 Or, to reproduce the old 'http_anonymizer paranoid' feature
4421 request_header_access Allow allow all
4422 request_header_access Authorization allow all
4423 request_header_access WWW-Authenticate allow all
4424 request_header_access Proxy-Authorization allow all
4425 request_header_access Proxy-Authenticate allow all
4426 request_header_access Cache-Control allow all
4427 request_header_access Content-Encoding allow all
4428 request_header_access Content-Length allow all
4429 request_header_access Content-Type allow all
4430 request_header_access Date allow all
4431 request_header_access Expires allow all
4432 request_header_access Host allow all
4433 request_header_access If-Modified-Since allow all
4434 request_header_access Last-Modified allow all
4435 request_header_access Location allow all
4436 request_header_access Pragma allow all
4437 request_header_access Accept allow all
4438 request_header_access Accept-Charset allow all
4439 request_header_access Accept-Encoding allow all
4440 request_header_access Accept-Language allow all
4441 request_header_access Content-Language allow all
4442 request_header_access Mime-Version allow all
4443 request_header_access Retry-After allow all
4444 request_header_access Title allow all
4445 request_header_access Connection allow all
4446 request_header_access All deny all
4448 although many of those are HTTP reply headers, and so should be
4449 controlled with the reply_header_access directive.
4451 By default, all headers are allowed (no anonymizing is
4455 NAME: reply_header_access
4456 IFDEF: USE_HTTP_VIOLATIONS
4457 TYPE: http_header_access[]
4458 LOC: Config.reply_header_access
4461 Usage: reply_header_access header_name allow|deny [!]aclname ...
4463 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4464 this feature could make you liable for problems which it
4467 This option only applies to reply headers, i.e., from the
4468 server to the client.
4470 This is the same as request_header_access, but in the other
4473 This option replaces the old 'anonymize_headers' and the
4474 older 'http_anonymizer' option with something that is much
4475 more configurable. This new method creates a list of ACLs
4476 for each header, allowing you very fine-tuned header
4479 You can only specify known headers for the header name.
4480 Other headers are reclassified as 'Other'. You can also
4481 refer to all the headers with 'All'.
4483 For example, to achieve the same behavior as the old
4484 'http_anonymizer standard' option, you should use:
4486 reply_header_access From deny all
4487 reply_header_access Referer deny all
4488 reply_header_access Server deny all
4489 reply_header_access User-Agent deny all
4490 reply_header_access WWW-Authenticate deny all
4491 reply_header_access Link deny all
4493 Or, to reproduce the old 'http_anonymizer paranoid' feature
4496 reply_header_access Allow allow all
4497 reply_header_access Authorization allow all
4498 reply_header_access WWW-Authenticate allow all
4499 reply_header_access Proxy-Authorization allow all
4500 reply_header_access Proxy-Authenticate allow all
4501 reply_header_access Cache-Control allow all
4502 reply_header_access Content-Encoding allow all
4503 reply_header_access Content-Length allow all
4504 reply_header_access Content-Type allow all
4505 reply_header_access Date allow all
4506 reply_header_access Expires allow all
4507 reply_header_access Host allow all
4508 reply_header_access If-Modified-Since allow all
4509 reply_header_access Last-Modified allow all
4510 reply_header_access Location allow all
4511 reply_header_access Pragma allow all
4512 reply_header_access Accept allow all
4513 reply_header_access Accept-Charset allow all
4514 reply_header_access Accept-Encoding allow all
4515 reply_header_access Accept-Language allow all
4516 reply_header_access Content-Language allow all
4517 reply_header_access Mime-Version allow all
4518 reply_header_access Retry-After allow all
4519 reply_header_access Title allow all
4520 reply_header_access Connection allow all
4521 reply_header_access All deny all
4523 although the HTTP request headers won't be usefully controlled
4524 by this directive -- see request_header_access for details.
4526 By default, all headers are allowed (no anonymizing is
4530 NAME: request_header_replace header_replace
4531 IFDEF: USE_HTTP_VIOLATIONS
4532 TYPE: http_header_replace[]
4533 LOC: Config.request_header_access
4536 Usage: request_header_replace header_name message
4537 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
4539 This option allows you to change the contents of headers
4540 denied with request_header_access above, by replacing them
4541 with some fixed string. This replaces the old fake_user_agent
4544 This only applies to request headers, not reply headers.
4546 By default, headers are removed if denied.
4549 NAME: reply_header_replace
4550 IFDEF: USE_HTTP_VIOLATIONS
4551 TYPE: http_header_replace[]
4552 LOC: Config.reply_header_access
4555 Usage: reply_header_replace header_name message
4556 Example: reply_header_replace Server Foo/1.0
4558 This option allows you to change the contents of headers
4559 denied with reply_header_access above, by replacing them
4560 with some fixed string.
4562 This only applies to reply headers, not request headers.
4564 By default, headers are removed if denied.
4567 NAME: relaxed_header_parser
4568 COMMENT: on|off|warn
4570 LOC: Config.onoff.relaxed_header_parser
4573 In the default "on" setting Squid accepts certain forms
4574 of non-compliant HTTP messages where it is unambiguous
4575 what the sending application intended even if the message
4576 is not correctly formatted. The messages is then normalized
4577 to the correct form when forwarded by Squid.
4579 If set to "warn" then a warning will be emitted in cache.log
4580 each time such HTTP error is encountered.
4582 If set to "off" then such HTTP errors will cause the request
4583 or response to be rejected.
4588 -----------------------------------------------------------------------------
4591 NAME: forward_timeout
4594 LOC: Config.Timeout.forward
4597 This parameter specifies how long Squid should at most attempt in
4598 finding a forwarding path for the request before giving up.
4601 NAME: connect_timeout
4604 LOC: Config.Timeout.connect
4607 This parameter specifies how long to wait for the TCP connect to
4608 the requested server or peer to complete before Squid should
4609 attempt to find another path where to forward the request.
4612 NAME: peer_connect_timeout
4615 LOC: Config.Timeout.peer_connect
4618 This parameter specifies how long to wait for a pending TCP
4619 connection to a peer cache. The default is 30 seconds. You
4620 may also set different timeout values for individual neighbors
4621 with the 'connect-timeout' option on a 'cache_peer' line.
4627 LOC: Config.Timeout.read
4630 The read_timeout is applied on server-side connections. After
4631 each successful read(), the timeout will be extended by this
4632 amount. If no data is read again after this amount of time,
4633 the request is aborted and logged with ERR_READ_TIMEOUT. The
4634 default is 15 minutes.
4640 LOC: Config.Timeout.write
4643 This timeout is tracked for all connections that have data
4644 available for writing and are waiting for the socket to become
4645 ready. After each successful write, the timeout is extended by
4646 the configured amount. If Squid has data to write but the
4647 connection is not ready for the configured duration, the
4648 transaction associated with the connection is terminated. The
4649 default is 15 minutes.
4652 NAME: request_timeout
4654 LOC: Config.Timeout.request
4657 How long to wait for an HTTP request after initial
4658 connection establishment.
4661 NAME: client_idle_pconn_timeout persistent_request_timeout
4663 LOC: Config.Timeout.clientIdlePconn
4666 How long to wait for the next HTTP request on a persistent
4667 client connection after the previous request completes.
4670 NAME: client_lifetime
4673 LOC: Config.Timeout.lifetime
4676 The maximum amount of time a client (browser) is allowed to
4677 remain connected to the cache process. This protects the Cache
4678 from having a lot of sockets (and hence file descriptors) tied up
4679 in a CLOSE_WAIT state from remote clients that go away without
4680 properly shutting down (either because of a network failure or
4681 because of a poor client implementation). The default is one
4684 NOTE: The default value is intended to be much larger than any
4685 client would ever need to be connected to your cache. You
4686 should probably change client_lifetime only as a last resort.
4687 If you seem to have many client connections tying up
4688 filedescriptors, we recommend first tuning the read_timeout,
4689 request_timeout, persistent_request_timeout and quick_abort values.
4692 NAME: half_closed_clients
4694 LOC: Config.onoff.half_closed_clients
4697 Some clients may shutdown the sending side of their TCP
4698 connections, while leaving their receiving sides open. Sometimes,
4699 Squid can not tell the difference between a half-closed and a
4700 fully-closed TCP connection.
4702 By default, Squid will immediately close client connections when
4703 read(2) returns "no more data to read."
4705 Change this option to 'on' and Squid will keep open connections
4706 until a read(2) or write(2) on the socket returns an error.
4707 This may show some benefits for reverse proxies. But if not
4708 it is recommended to leave OFF.
4711 NAME: server_idle_pconn_timeout pconn_timeout
4713 LOC: Config.Timeout.serverIdlePconn
4716 Timeout for idle persistent connections to servers and other
4723 LOC: Ident::TheConfig.timeout
4726 Maximum time to wait for IDENT lookups to complete.
4728 If this is too high, and you enabled IDENT lookups from untrusted
4729 users, you might be susceptible to denial-of-service by having
4730 many ident requests going at once.
4733 NAME: shutdown_lifetime
4736 LOC: Config.shutdownLifetime
4739 When SIGTERM or SIGHUP is received, the cache is put into
4740 "shutdown pending" mode until all active sockets are closed.
4741 This value is the lifetime to set for all open descriptors
4742 during shutdown mode. Any active clients after this many
4743 seconds will receive a 'timeout' message.
4747 ADMINISTRATIVE PARAMETERS
4748 -----------------------------------------------------------------------------
4754 LOC: Config.adminEmail
4756 Email-address of local cache manager who will receive
4757 mail if the cache dies. The default is "webmaster."
4763 LOC: Config.EmailFrom
4765 From: email-address for mail sent when the cache dies.
4766 The default is to use 'appname@unique_hostname'.
4767 Default appname value is "squid", can be changed into
4768 src/globals.h before building squid.
4774 LOC: Config.EmailProgram
4776 Email program used to send mail if the cache dies.
4777 The default is "mail". The specified program must comply
4778 with the standard Unix mail syntax:
4779 mail-program recipient < mailfile
4781 Optional command line options can be specified.
4784 NAME: cache_effective_user
4786 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
4787 LOC: Config.effectiveUser
4789 If you start Squid as root, it will change its effective/real
4790 UID/GID to the user specified below. The default is to change
4791 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
4792 see also; cache_effective_group
4795 NAME: cache_effective_group
4798 LOC: Config.effectiveGroup
4800 Squid sets the GID to the effective user's default group ID
4801 (taken from the password file) and supplementary group list
4802 from the groups membership.
4804 If you want Squid to run with a specific GID regardless of
4805 the group memberships of the effective user then set this
4806 to the group (or GID) you want Squid to run as. When set
4807 all other group privileges of the effective user are ignored
4808 and only this GID is effective. If Squid is not started as
4809 root the user starting Squid MUST be member of the specified
4812 This option is not recommended by the Squid Team.
4813 Our preference is for administrators to configure a secure
4814 user account for squid with UID/GID matching system policies.
4817 NAME: httpd_suppress_version_string
4821 LOC: Config.onoff.httpd_suppress_version_string
4823 Suppress Squid version string info in HTTP headers and HTML error pages.
4826 NAME: visible_hostname
4828 LOC: Config.visibleHostname
4831 If you want to present a special hostname in error messages, etc,
4832 define this. Otherwise, the return value of gethostname()
4833 will be used. If you have multiple caches in a cluster and
4834 get errors about IP-forwarding you must set them to have individual
4835 names with this setting.
4838 NAME: unique_hostname
4840 LOC: Config.uniqueHostname
4843 If you want to have multiple machines with the same
4844 'visible_hostname' you must give each machine a different
4845 'unique_hostname' so forwarding loops can be detected.
4848 NAME: hostname_aliases
4850 LOC: Config.hostnameAliases
4853 A list of other DNS names your cache has.
4861 Minimum umask which should be enforced while the proxy
4862 is running, in addition to the umask set at startup.
4864 For a traditional octal representation of umasks, start
4869 OPTIONS FOR THE CACHE REGISTRATION SERVICE
4870 -----------------------------------------------------------------------------
4872 This section contains parameters for the (optional) cache
4873 announcement service. This service is provided to help
4874 cache administrators locate one another in order to join or
4875 create cache hierarchies.
4877 An 'announcement' message is sent (via UDP) to the registration
4878 service by Squid. By default, the announcement message is NOT
4879 SENT unless you enable it with 'announce_period' below.
4881 The announcement message includes your hostname, plus the
4882 following information from this configuration file:
4888 All current information is processed regularly and made
4889 available on the Web at http://www.ircache.net/Cache/Tracker/.
4892 NAME: announce_period
4894 LOC: Config.Announce.period
4897 This is how frequently to send cache announcements. The
4898 default is `0' which disables sending the announcement
4901 To enable announcing your cache, just set an announce period.
4904 announce_period 1 day
4909 DEFAULT: tracker.ircache.net
4910 LOC: Config.Announce.host
4916 LOC: Config.Announce.file
4922 LOC: Config.Announce.port
4924 announce_host and announce_port set the hostname and port
4925 number where the registration message will be sent.
4927 Hostname will default to 'tracker.ircache.net' and port will
4928 default default to 3131. If the 'filename' argument is given,
4929 the contents of that file will be included in the announce
4934 HTTPD-ACCELERATOR OPTIONS
4935 -----------------------------------------------------------------------------
4938 NAME: httpd_accel_surrogate_id
4941 LOC: Config.Accel.surrogate_id
4943 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
4944 need an identification token to allow control targeting. Because
4945 a farm of surrogates may all perform the same tasks, they may share
4946 an identification token.
4948 The default ID is the visible_hostname
4951 NAME: http_accel_surrogate_remote
4955 LOC: Config.onoff.surrogate_is_remote
4957 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
4958 Set this to on to have squid behave as a remote surrogate.
4962 IFDEF: USE_SQUID_ESI
4963 COMMENT: libxml2|expat|custom
4965 LOC: ESIParser::Type
4968 ESI markup is not strictly XML compatible. The custom ESI parser
4969 will give higher performance, but cannot handle non ASCII character
4974 DELAY POOL PARAMETERS
4975 -----------------------------------------------------------------------------
4979 TYPE: delay_pool_count
4981 IFDEF: USE_DELAY_POOLS
4984 This represents the number of delay pools to be used. For example,
4985 if you have one class 2 delay pool and one class 3 delays pool, you
4986 have a total of 2 delay pools.
4990 TYPE: delay_pool_class
4992 IFDEF: USE_DELAY_POOLS
4995 This defines the class of each delay pool. There must be exactly one
4996 delay_class line for each delay pool. For example, to define two
4997 delay pools, one of class 2 and one of class 3, the settings above
5001 delay_pools 4 # 4 delay pools
5002 delay_class 1 2 # pool 1 is a class 2 pool
5003 delay_class 2 3 # pool 2 is a class 3 pool
5004 delay_class 3 4 # pool 3 is a class 4 pool
5005 delay_class 4 5 # pool 4 is a class 5 pool
5007 The delay pool classes are:
5009 class 1 Everything is limited by a single aggregate
5012 class 2 Everything is limited by a single aggregate
5013 bucket as well as an "individual" bucket chosen
5014 from bits 25 through 32 of the IPv4 address.
5016 class 3 Everything is limited by a single aggregate
5017 bucket as well as a "network" bucket chosen
5018 from bits 17 through 24 of the IP address and a
5019 "individual" bucket chosen from bits 17 through
5020 32 of the IPv4 address.
5022 class 4 Everything in a class 3 delay pool, with an
5023 additional limit on a per user basis. This
5024 only takes effect if the username is established
5025 in advance - by forcing authentication in your
5028 class 5 Requests are grouped according their tag (see
5029 external_acl's tag= reply).
5032 Each pool also requires a delay_parameters directive to configure the pool size
5033 and speed limits used whenever the pool is applied to a request. Along with
5034 a set of delay_access directives to determine when it is used.
5036 NOTE: If an IP address is a.b.c.d
5037 -> bits 25 through 32 are "d"
5038 -> bits 17 through 24 are "c"
5039 -> bits 17 through 32 are "c * 256 + d"
5041 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5042 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5046 TYPE: delay_pool_access
5048 IFDEF: USE_DELAY_POOLS
5051 This is used to determine which delay pool a request falls into.
5053 delay_access is sorted per pool and the matching starts with pool 1,
5054 then pool 2, ..., and finally pool N. The first delay pool where the
5055 request is allowed is selected for the request. If it does not allow
5056 the request to any pool then the request is not delayed (default).
5058 For example, if you want some_big_clients in delay
5059 pool 1 and lotsa_little_clients in delay pool 2:
5062 delay_access 1 allow some_big_clients
5063 delay_access 1 deny all
5064 delay_access 2 allow lotsa_little_clients
5065 delay_access 2 deny all
5066 delay_access 3 allow authenticated_clients
5069 NAME: delay_parameters
5070 TYPE: delay_pool_rates
5072 IFDEF: USE_DELAY_POOLS
5075 This defines the parameters for a delay pool. Each delay pool has
5076 a number of "buckets" associated with it, as explained in the
5077 description of delay_class.
5079 For a class 1 delay pool, the syntax is:
5081 delay_parameters pool aggregate
5083 For a class 2 delay pool:
5085 delay_parameters pool aggregate individual
5087 For a class 3 delay pool:
5089 delay_parameters pool aggregate network individual
5091 For a class 4 delay pool:
5093 delay_parameters pool aggregate network individual user
5095 For a class 5 delay pool:
5097 delay_parameters pool tagrate
5099 The option variables are:
5101 pool a pool number - ie, a number between 1 and the
5102 number specified in delay_pools as used in
5105 aggregate the speed limit parameters for the aggregate bucket
5108 individual the speed limit parameters for the individual
5109 buckets (class 2, 3).
5111 network the speed limit parameters for the network buckets
5114 user the speed limit parameters for the user buckets
5117 tagrate the speed limit parameters for the tag buckets
5120 A pair of delay parameters is written restore/maximum, where restore is
5121 the number of bytes (not bits - modem and network speeds are usually
5122 quoted in bits) per second placed into the bucket, and maximum is the
5123 maximum number of bytes which can be in the bucket at any time.
5125 There must be one delay_parameters line for each delay pool.
5128 For example, if delay pool number 1 is a class 2 delay pool as in the
5129 above example, and is being used to strictly limit each host to 64Kbit/sec
5130 (plus overheads), with no overall limit, the line is:
5132 delay_parameters 1 -1/-1 8000/8000
5134 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5136 Note that the figure -1 is used to represent "unlimited".
5139 And, if delay pool number 2 is a class 3 delay pool as in the above
5140 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5141 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5142 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5143 to permit a decent web page to be downloaded at a decent speed
5144 (if the network is not being limited due to overuse) but slow down
5145 large downloads more significantly:
5147 delay_parameters 2 32000/32000 8000/8000 600/8000
5149 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5150 8 x 8000 KByte/sec -> 64Kbit/sec.
5151 8 x 600 Byte/sec -> 4800bit/sec.
5154 Finally, for a class 4 delay pool as in the example - each user will
5155 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5157 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5160 NAME: delay_initial_bucket_level
5161 COMMENT: (percent, 0-100)
5164 IFDEF: USE_DELAY_POOLS
5165 LOC: Config.Delay.initial
5167 The initial bucket percentage is used to determine how much is put
5168 in each bucket when squid starts, is reconfigured, or first notices
5169 a host accessing it (in class 2 and class 3, individual hosts and
5170 networks only have buckets associated with them once they have been
5175 CLIENT DELAY POOL PARAMETERS
5176 -----------------------------------------------------------------------------
5179 NAME: client_delay_pools
5180 TYPE: client_delay_pool_count
5182 IFDEF: USE_DELAY_POOLS
5183 LOC: Config.ClientDelay
5185 This option specifies the number of client delay pools used. It must
5186 preceed other client_delay_* options.
5189 client_delay_pools 2
5192 NAME: client_delay_initial_bucket_level
5193 COMMENT: (percent, 0-no_limit)
5196 IFDEF: USE_DELAY_POOLS
5197 LOC: Config.ClientDelay.initial
5199 This option determines the initial bucket size as a percentage of
5200 max_bucket_size from client_delay_parameters. Buckets are created
5201 at the time of the "first" connection from the matching IP. Idle
5202 buckets are periodically deleted up.
5204 You can specify more than 100 percent but note that such "oversized"
5205 buckets are not refilled until their size goes down to max_bucket_size
5206 from client_delay_parameters.
5209 client_delay_initial_bucket_level 50
5212 NAME: client_delay_parameters
5213 TYPE: client_delay_pool_rates
5215 IFDEF: USE_DELAY_POOLS
5216 LOC: Config.ClientDelay
5219 This option configures client-side bandwidth limits using the
5222 client_delay_parameters pool speed_limit max_bucket_size
5224 pool is an integer ID used for client_delay_access matching.
5226 speed_limit is bytes added to the bucket per second.
5228 max_bucket_size is the maximum size of a bucket, enforced after any
5229 speed_limit additions.
5231 Please see the delay_parameters option for more information and
5235 client_delay_parameters 1 1024 2048
5236 client_delay_parameters 2 51200 16384
5239 NAME: client_delay_access
5240 TYPE: client_delay_pool_access
5242 IFDEF: USE_DELAY_POOLS
5243 LOC: Config.ClientDelay
5246 This option determines the client-side delay pool for the
5249 client_delay_access pool_ID allow|deny acl_name
5251 All client_delay_access options are checked in their pool ID
5252 order, starting with pool 1. The first checked pool with allowed
5253 request is selected for the request. If no ACL matches or there
5254 are no client_delay_access options, the request bandwidth is not
5257 The ACL-selected pool is then used to find the
5258 client_delay_parameters for the request. Client-side pools are
5259 not used to aggregate clients. Clients are always aggregated
5260 based on their source IP addresses (one bucket per source IP).
5262 Please see delay_access for more examples.
5265 client_delay_access 1 allow low_rate_network
5266 client_delay_access 2 allow vips_network
5270 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5271 -----------------------------------------------------------------------------
5276 LOC: Config.Wccp.router
5280 Use this option to define your WCCP ``home'' router for
5283 wccp_router supports a single WCCP(v1) router
5285 wccp2_router supports multiple WCCPv2 routers
5287 only one of the two may be used at the same time and defines
5288 which version of WCCP to use.
5292 TYPE: IpAddress_list
5293 LOC: Config.Wccp2.router
5297 Use this option to define your WCCP ``home'' router for
5300 wccp_router supports a single WCCP(v1) router
5302 wccp2_router supports multiple WCCPv2 routers
5304 only one of the two may be used at the same time and defines
5305 which version of WCCP to use.
5310 LOC: Config.Wccp.version
5314 This directive is only relevant if you need to set up WCCP(v1)
5315 to some very old and end-of-life Cisco routers. In all other
5316 setups it must be left unset or at the default setting.
5317 It defines an internal version in the WCCP(v1) protocol,
5318 with version 4 being the officially documented protocol.
5320 According to some users, Cisco IOS 11.2 and earlier only
5321 support WCCP version 3. If you're using that or an earlier
5322 version of IOS, you may need to change this value to 3, otherwise
5323 do not specify this parameter.
5326 NAME: wccp2_rebuild_wait
5328 LOC: Config.Wccp2.rebuildwait
5332 If this is enabled Squid will wait for the cache dir rebuild to finish
5333 before sending the first wccp2 HereIAm packet
5336 NAME: wccp2_forwarding_method
5338 LOC: Config.Wccp2.forwarding_method
5342 WCCP2 allows the setting of forwarding methods between the
5343 router/switch and the cache. Valid values are as follows:
5345 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5346 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5348 Currently (as of IOS 12.4) cisco routers only support GRE.
5349 Cisco switches only support the L2 redirect assignment method.
5352 NAME: wccp2_return_method
5354 LOC: Config.Wccp2.return_method
5358 WCCP2 allows the setting of return methods between the
5359 router/switch and the cache for packets that the cache
5360 decides not to handle. Valid values are as follows:
5362 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5363 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5365 Currently (as of IOS 12.4) cisco routers only support GRE.
5366 Cisco switches only support the L2 redirect assignment.
5368 If the "ip wccp redirect exclude in" command has been
5369 enabled on the cache interface, then it is still safe for
5370 the proxy server to use a l2 redirect method even if this
5371 option is set to GRE.
5374 NAME: wccp2_assignment_method
5376 LOC: Config.Wccp2.assignment_method
5380 WCCP2 allows the setting of methods to assign the WCCP hash
5381 Valid values are as follows:
5383 hash - Hash assignment
5384 mask - Mask assignment
5386 As a general rule, cisco routers support the hash assignment method
5387 and cisco switches support the mask assignment method.
5392 LOC: Config.Wccp2.info
5393 DEFAULT_IF_NONE: standard 0
5396 WCCP2 allows for multiple traffic services. There are two
5397 types: "standard" and "dynamic". The standard type defines
5398 one service id - http (id 0). The dynamic service ids can be from
5399 51 to 255 inclusive. In order to use a dynamic service id
5400 one must define the type of traffic to be redirected; this is done
5401 using the wccp2_service_info option.
5403 The "standard" type does not require a wccp2_service_info option,
5404 just specifying the service id will suffice.
5406 MD5 service authentication can be enabled by adding
5407 "password=<password>" to the end of this service declaration.
5411 wccp2_service standard 0 # for the 'web-cache' standard service
5412 wccp2_service dynamic 80 # a dynamic service type which will be
5413 # fleshed out with subsequent options.
5414 wccp2_service standard 0 password=foo
5417 NAME: wccp2_service_info
5418 TYPE: wccp2_service_info
5419 LOC: Config.Wccp2.info
5423 Dynamic WCCPv2 services require further information to define the
5424 traffic you wish to have diverted.
5428 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
5429 priority=<priority> ports=<port>,<port>..
5431 The relevant WCCPv2 flags:
5432 + src_ip_hash, dst_ip_hash
5433 + source_port_hash, dst_port_hash
5434 + src_ip_alt_hash, dst_ip_alt_hash
5435 + src_port_alt_hash, dst_port_alt_hash
5438 The port list can be one to eight entries.
5442 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
5443 priority=240 ports=80
5445 Note: the service id must have been defined by a previous
5446 'wccp2_service dynamic <id>' entry.
5451 LOC: Config.Wccp2.weight
5455 Each cache server gets assigned a set of the destination
5456 hash proportional to their weight.
5461 LOC: Config.Wccp.address
5468 LOC: Config.Wccp2.address
5472 Use this option if you require WCCP to use a specific
5475 The default behavior is to not bind to any specific address.
5479 PERSISTENT CONNECTION HANDLING
5480 -----------------------------------------------------------------------------
5482 Also see "pconn_timeout" in the TIMEOUTS section
5485 NAME: client_persistent_connections
5487 LOC: Config.onoff.client_pconns
5491 NAME: server_persistent_connections
5493 LOC: Config.onoff.server_pconns
5496 Persistent connection support for clients and servers. By
5497 default, Squid uses persistent connections (when allowed)
5498 with its clients and servers. You can use these options to
5499 disable persistent connections with clients and/or servers.
5502 NAME: persistent_connection_after_error
5504 LOC: Config.onoff.error_pconns
5507 With this directive the use of persistent connections after
5508 HTTP errors can be disabled. Useful if you have clients
5509 who fail to handle errors on persistent connections proper.
5512 NAME: detect_broken_pconn
5514 LOC: Config.onoff.detect_broken_server_pconns
5517 Some servers have been found to incorrectly signal the use
5518 of HTTP/1.0 persistent connections even on replies not
5519 compatible, causing significant delays. This server problem
5520 has mostly been seen on redirects.
5522 By enabling this directive Squid attempts to detect such
5523 broken replies and automatically assume the reply is finished
5524 after 10 seconds timeout.
5528 CACHE DIGEST OPTIONS
5529 -----------------------------------------------------------------------------
5532 NAME: digest_generation
5533 IFDEF: USE_CACHE_DIGESTS
5535 LOC: Config.onoff.digest_generation
5538 This controls whether the server will generate a Cache Digest
5539 of its contents. By default, Cache Digest generation is
5540 enabled if Squid is compiled with --enable-cache-digests defined.
5543 NAME: digest_bits_per_entry
5544 IFDEF: USE_CACHE_DIGESTS
5546 LOC: Config.digest.bits_per_entry
5549 This is the number of bits of the server's Cache Digest which
5550 will be associated with the Digest entry for a given HTTP
5551 Method and URL (public key) combination. The default is 5.
5554 NAME: digest_rebuild_period
5555 IFDEF: USE_CACHE_DIGESTS
5558 LOC: Config.digest.rebuild_period
5561 This is the wait time between Cache Digest rebuilds.
5564 NAME: digest_rewrite_period
5566 IFDEF: USE_CACHE_DIGESTS
5568 LOC: Config.digest.rewrite_period
5571 This is the wait time between Cache Digest writes to
5575 NAME: digest_swapout_chunk_size
5578 IFDEF: USE_CACHE_DIGESTS
5579 LOC: Config.digest.swapout_chunk_size
5582 This is the number of bytes of the Cache Digest to write to
5583 disk at a time. It defaults to 4096 bytes (4KB), the Squid
5587 NAME: digest_rebuild_chunk_percentage
5588 COMMENT: (percent, 0-100)
5589 IFDEF: USE_CACHE_DIGESTS
5591 LOC: Config.digest.rebuild_chunk_percentage
5594 This is the percentage of the Cache Digest to be scanned at a
5595 time. By default it is set to 10% of the Cache Digest.
5600 -----------------------------------------------------------------------------
5605 LOC: Config.Port.snmp
5609 The port number where Squid listens for SNMP requests. To enable
5610 SNMP support set this to a suitable port number. Port number
5611 3401 is often used for the Squid SNMP agent. By default it's
5612 set to "0" (disabled)
5620 LOC: Config.accessList.snmp
5621 DEFAULT_IF_NONE: deny all
5624 Allowing or denying access to the SNMP port.
5626 All access to the agent is denied by default.
5629 snmp_access allow|deny [!]aclname ...
5631 This clause only supports fast acl types.
5632 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5634 snmp_access allow snmppublic localhost
5635 snmp_access deny all
5638 NAME: snmp_incoming_address
5640 LOC: Config.Addrs.snmp_incoming
5645 NAME: snmp_outgoing_address
5647 LOC: Config.Addrs.snmp_outgoing
5651 Just like 'udp_incoming_address', but for the SNMP port.
5653 snmp_incoming_address is used for the SNMP socket receiving
5654 messages from SNMP agents.
5655 snmp_outgoing_address is used for SNMP packets returned to SNMP
5658 The default snmp_incoming_address is to listen on all
5659 available network interfaces.
5661 If snmp_outgoing_address is not set it will use the same socket
5662 as snmp_incoming_address. Only change this if you want to have
5663 SNMP replies sent using another address than where this Squid
5664 listens for SNMP queries.
5666 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
5667 the same value since they both use port 3401.
5672 -----------------------------------------------------------------------------
5675 NAME: icp_port udp_port
5678 LOC: Config.Port.icp
5680 The port number where Squid sends and receives ICP queries to
5681 and from neighbor caches. The standard UDP port for ICP is 3130.
5682 Default is disabled (0).
5685 icp_port @DEFAULT_ICP_PORT@
5692 LOC: Config.Port.htcp
5694 The port number where Squid sends and receives HTCP queries to
5695 and from neighbor caches. To turn it on you want to set it to
5696 4827. By default it is set to "0" (disabled).
5702 NAME: log_icp_queries
5706 LOC: Config.onoff.log_udp
5708 If set, ICP queries are logged to access.log. You may wish
5709 do disable this if your ICP load is VERY high to speed things
5710 up or to simplify log analysis.
5713 NAME: udp_incoming_address
5715 LOC:Config.Addrs.udp_incoming
5718 udp_incoming_address is used for UDP packets received from other
5721 The default behavior is to not bind to any specific address.
5723 Only change this if you want to have all UDP queries received on
5724 a specific interface/address.
5726 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
5727 modules. Altering it will affect all of them in the same manner.
5729 see also; udp_outgoing_address
5731 NOTE, udp_incoming_address and udp_outgoing_address can not
5732 have the same value since they both use the same port.
5735 NAME: udp_outgoing_address
5737 LOC: Config.Addrs.udp_outgoing
5740 udp_outgoing_address is used for UDP packets sent out to other
5743 The default behavior is to not bind to any specific address.
5745 Instead it will use the same socket as udp_incoming_address.
5746 Only change this if you want to have UDP queries sent using another
5747 address than where this Squid listens for UDP queries from other
5750 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
5751 modules. Altering it will affect all of them in the same manner.
5753 see also; udp_incoming_address
5755 NOTE, udp_incoming_address and udp_outgoing_address can not
5756 have the same value since they both use the same port.
5763 LOC: Config.onoff.icp_hit_stale
5765 If you want to return ICP_HIT for stale cache objects, set this
5766 option to 'on'. If you have sibling relationships with caches
5767 in other administrative domains, this should be 'off'. If you only
5768 have sibling relationships with caches under your control,
5769 it is probably okay to set this to 'on'.
5770 If set to 'on', your siblings should use the option "allow-miss"
5771 on their cache_peer lines for connecting to you.
5774 NAME: minimum_direct_hops
5777 LOC: Config.minDirectHops
5779 If using the ICMP pinging stuff, do direct fetches for sites
5780 which are no more than this many hops away.
5783 NAME: minimum_direct_rtt
5786 LOC: Config.minDirectRtt
5788 If using the ICMP pinging stuff, do direct fetches for sites
5789 which are no more than this many rtt milliseconds away.
5795 LOC: Config.Netdb.low
5801 LOC: Config.Netdb.high
5803 The low and high water marks for the ICMP measurement
5804 database. These are counts, not percents. The defaults are
5805 900 and 1000. When the high water mark is reached, database
5806 entries will be deleted until the low mark is reached.
5809 NAME: netdb_ping_period
5811 LOC: Config.Netdb.period
5814 The minimum period for measuring a site. There will be at
5815 least this much delay between successive pings to the same
5816 network. The default is five minutes.
5823 LOC: Config.onoff.query_icmp
5825 If you want to ask your peers to include ICMP data in their ICP
5826 replies, enable this option.
5828 If your peer has configured Squid (during compilation) with
5829 '--enable-icmp' that peer will send ICMP pings to origin server
5830 sites of the URLs it receives. If you enable this option the
5831 ICP replies from that peer will include the ICMP data (if available).
5832 Then, when choosing a parent cache, Squid will choose the parent with
5833 the minimal RTT to the origin server. When this happens, the
5834 hierarchy field of the access.log will be
5835 "CLOSEST_PARENT_MISS". This option is off by default.
5838 NAME: test_reachability
5842 LOC: Config.onoff.test_reachability
5844 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
5845 instead of ICP_MISS if the target host is NOT in the ICMP
5846 database, or has a zero RTT.
5849 NAME: icp_query_timeout
5853 LOC: Config.Timeout.icp_query
5855 Normally Squid will automatically determine an optimal ICP
5856 query timeout value based on the round-trip-time of recent ICP
5857 queries. If you want to override the value determined by
5858 Squid, set this 'icp_query_timeout' to a non-zero value. This
5859 value is specified in MILLISECONDS, so, to use a 2-second
5860 timeout (the old default), you would write:
5862 icp_query_timeout 2000
5865 NAME: maximum_icp_query_timeout
5869 LOC: Config.Timeout.icp_query_max
5871 Normally the ICP query timeout is determined dynamically. But
5872 sometimes it can lead to very large values (say 5 seconds).
5873 Use this option to put an upper limit on the dynamic timeout
5874 value. Do NOT use this option to always use a fixed (instead
5875 of a dynamic) timeout value. To set a fixed timeout see the
5876 'icp_query_timeout' directive.
5879 NAME: minimum_icp_query_timeout
5883 LOC: Config.Timeout.icp_query_min
5885 Normally the ICP query timeout is determined dynamically. But
5886 sometimes it can lead to very small timeouts, even lower than
5887 the normal latency variance on your link due to traffic.
5888 Use this option to put an lower limit on the dynamic timeout
5889 value. Do NOT use this option to always use a fixed (instead
5890 of a dynamic) timeout value. To set a fixed timeout see the
5891 'icp_query_timeout' directive.
5894 NAME: background_ping_rate
5898 LOC: Config.backgroundPingRate
5900 Controls how often the ICP pings are sent to siblings that
5901 have background-ping set.
5905 MULTICAST ICP OPTIONS
5906 -----------------------------------------------------------------------------
5911 LOC: Config.mcast_group_list
5914 This tag specifies a list of multicast groups which your server
5915 should join to receive multicasted ICP queries.
5917 NOTE! Be very careful what you put here! Be sure you
5918 understand the difference between an ICP _query_ and an ICP
5919 _reply_. This option is to be set only if you want to RECEIVE
5920 multicast queries. Do NOT set this option to SEND multicast
5921 ICP (use cache_peer for that). ICP replies are always sent via
5922 unicast, so this option does not affect whether or not you will
5923 receive replies from multicast group members.
5925 You must be very careful to NOT use a multicast address which
5926 is already in use by another group of caches.
5928 If you are unsure about multicast, please read the Multicast
5929 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
5931 Usage: mcast_groups 239.128.16.128 224.0.1.20
5933 By default, Squid doesn't listen on any multicast groups.
5936 NAME: mcast_miss_addr
5937 IFDEF: MULTICAST_MISS_STREAM
5939 LOC: Config.mcast_miss.addr
5942 If you enable this option, every "cache miss" URL will
5943 be sent out on the specified multicast address.
5945 Do not enable this option unless you are are absolutely
5946 certain you understand what you are doing.
5949 NAME: mcast_miss_ttl
5950 IFDEF: MULTICAST_MISS_STREAM
5952 LOC: Config.mcast_miss.ttl
5955 This is the time-to-live value for packets multicasted
5956 when multicasting off cache miss URLs is enabled. By
5957 default this is set to 'site scope', i.e. 16.
5960 NAME: mcast_miss_port
5961 IFDEF: MULTICAST_MISS_STREAM
5963 LOC: Config.mcast_miss.port
5966 This is the port number to be used in conjunction with
5970 NAME: mcast_miss_encode_key
5971 IFDEF: MULTICAST_MISS_STREAM
5973 LOC: Config.mcast_miss.encode_key
5974 DEFAULT: XXXXXXXXXXXXXXXX
5976 The URLs that are sent in the multicast miss stream are
5977 encrypted. This is the encryption key.
5980 NAME: mcast_icp_query_timeout
5984 LOC: Config.Timeout.mcast_icp_query
5986 For multicast peers, Squid regularly sends out ICP "probes" to
5987 count how many other peers are listening on the given multicast
5988 address. This value specifies how long Squid should wait to
5989 count all the replies. The default is 2000 msec, or 2
5994 INTERNAL ICON OPTIONS
5995 -----------------------------------------------------------------------------
5998 NAME: icon_directory
6000 LOC: Config.icons.directory
6001 DEFAULT: @DEFAULT_ICON_DIR@
6003 Where the icons are stored. These are normally kept in
6007 NAME: global_internal_static
6009 LOC: Config.onoff.global_internal_static
6012 This directive controls is Squid should intercept all requests for
6013 /squid-internal-static/ no matter which host the URL is requesting
6014 (default on setting), or if nothing special should be done for
6015 such URLs (off setting). The purpose of this directive is to make
6016 icons etc work better in complex cache hierarchies where it may
6017 not always be possible for all corners in the cache mesh to reach
6018 the server generating a directory listing.
6021 NAME: short_icon_urls
6023 LOC: Config.icons.use_short_names
6026 If this is enabled Squid will use short URLs for icons.
6027 If disabled it will revert to the old behavior of including
6028 it's own name and port in the URL.
6030 If you run a complex cache hierarchy with a mix of Squid and
6031 other proxies you may need to disable this directive.
6036 -----------------------------------------------------------------------------
6039 NAME: error_directory
6041 LOC: Config.errorDirectory
6044 If you wish to create your own versions of the default
6045 error files to customize them to suit your company copy
6046 the error/template files to another directory and point
6049 WARNING: This option will disable multi-language support
6050 on error pages if used.
6052 The squid developers are interested in making squid available in
6053 a wide variety of languages. If you are making translations for a
6054 language that Squid does not currently provide please consider
6055 contributing your translation back to the project.
6056 http://wiki.squid-cache.org/Translations
6058 The squid developers working on translations are happy to supply drop-in
6059 translated error files in exchange for any new language contributions.
6062 NAME: error_default_language
6063 IFDEF: USE_ERR_LOCALES
6065 LOC: Config.errorDefaultLanguage
6068 Set the default language which squid will send error pages in
6069 if no existing translation matches the clients language
6072 If unset (default) generic English will be used.
6074 The squid developers are interested in making squid available in
6075 a wide variety of languages. If you are interested in making
6076 translations for any language see the squid wiki for details.
6077 http://wiki.squid-cache.org/Translations
6080 NAME: error_log_languages
6081 IFDEF: USE_ERR_LOCALES
6083 LOC: Config.errorLogMissingLanguages
6086 Log to cache.log what languages users are attempting to
6087 auto-negotiate for translations.
6089 Successful negotiations are not logged. Only failures
6090 have meaning to indicate that Squid may need an upgrade
6091 of its error page translations.
6094 NAME: err_page_stylesheet
6096 LOC: Config.errorStylesheet
6097 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6099 CSS Stylesheet to pattern the display of Squid default error pages.
6101 For information on CSS see http://www.w3.org/Style/CSS/
6106 LOC: Config.errHtmlText
6109 HTML text to include in error messages. Make this a "mailto"
6110 URL to your admin address, or maybe just a link to your
6111 organizations Web page.
6113 To include this in your error messages, you must rewrite
6114 the error template files (found in the "errors" directory).
6115 Wherever you want the 'err_html_text' line to appear,
6116 insert a %L tag in the error template file.
6119 NAME: email_err_data
6122 LOC: Config.onoff.emailErrData
6125 If enabled, information about the occurred error will be
6126 included in the mailto links of the ERR pages (if %W is set)
6127 so that the email body contains the data.
6128 Syntax is <A HREF="mailto:%w%W">%w</A>
6133 LOC: Config.denyInfoList
6136 Usage: deny_info err_page_name acl
6137 or deny_info http://... acl
6138 or deny_info TCP_RESET acl
6140 This can be used to return a ERR_ page for requests which
6141 do not pass the 'http_access' rules. Squid remembers the last
6142 acl it evaluated in http_access, and if a 'deny_info' line exists
6143 for that ACL Squid returns a corresponding error page.
6145 The acl is typically the last acl on the http_access deny line which
6146 denied access. The exceptions to this rule are:
6147 - When Squid needs to request authentication credentials. It's then
6148 the first authentication related acl encountered
6149 - When none of the http_access lines matches. It's then the last
6150 acl processed on the last http_access line.
6151 - When the decision to deny access was made by an adaptation service,
6152 the acl name is the corresponding eCAP or ICAP service_name.
6154 NP: If providing your own custom error pages with error_directory
6155 you may also specify them by your custom file name:
6156 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6158 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6159 may be specified by prefixing the file name with the code and a colon.
6160 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6162 Alternatively you can tell Squid to reset the TCP connection
6163 by specifying TCP_RESET.
6165 Or you can specify an error URL or URL pattern. The browsers will
6166 get redirected to the specified URL after formatting tags have
6167 been replaced. Redirect will be done with 302 or 307 according to
6168 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6169 the URL. e.g. 303:http://example.com/
6172 %a - username (if available. Password NOT included)
6175 %E - Error description
6177 %H - Request domain name
6178 %i - Client IP Address
6180 %o - Message result from external ACL helper
6181 %p - Request Port number
6182 %P - Request Protocol name
6183 %R - Request URL path
6184 %T - Timestamp in RFC 1123 format
6185 %U - Full canonical URL from client
6186 (HTTPS URLs terminate with *)
6187 %u - Full canonical URL from client
6188 %w - Admin email from squid.conf
6190 %% - Literal percent (%) code
6195 OPTIONS INFLUENCING REQUEST FORWARDING
6196 -----------------------------------------------------------------------------
6199 NAME: nonhierarchical_direct
6201 LOC: Config.onoff.nonhierarchical_direct
6204 By default, Squid will send any non-hierarchical requests
6205 (matching hierarchy_stoplist or not cacheable request type) direct
6208 If you set this to off, Squid will prefer to send these
6209 requests to parents.
6211 Note that in most configurations, by turning this off you will only
6212 add latency to these request without any improvement in global hit
6215 If you are inside an firewall see never_direct instead of
6221 LOC: Config.onoff.prefer_direct
6224 Normally Squid tries to use parents for most requests. If you for some
6225 reason like it to first try going direct and only use a parent if
6226 going direct fails set this to on.
6228 By combining nonhierarchical_direct off and prefer_direct on you
6229 can set up Squid to use a parent as a backup path if going direct
6232 Note: If you want Squid to use parents for all requests see
6233 the never_direct directive. prefer_direct only modifies how Squid
6234 acts on cacheable requests.
6239 LOC: Config.accessList.AlwaysDirect
6242 Usage: always_direct allow|deny [!]aclname ...
6244 Here you can use ACL elements to specify requests which should
6245 ALWAYS be forwarded by Squid to the origin servers without using
6246 any peers. For example, to always directly forward requests for
6247 local servers ignoring any parents or siblings you may have use
6250 acl local-servers dstdomain my.domain.net
6251 always_direct allow local-servers
6253 To always forward FTP requests directly, use
6256 always_direct allow FTP
6258 NOTE: There is a similar, but opposite option named
6259 'never_direct'. You need to be aware that "always_direct deny
6260 foo" is NOT the same thing as "never_direct allow foo". You
6261 may need to use a deny rule to exclude a more-specific case of
6262 some other rule. Example:
6264 acl local-external dstdomain external.foo.net
6265 acl local-servers dstdomain .foo.net
6266 always_direct deny local-external
6267 always_direct allow local-servers
6269 NOTE: If your goal is to make the client forward the request
6270 directly to the origin server bypassing Squid then this needs
6271 to be done in the client configuration. Squid configuration
6272 can only tell Squid how Squid should fetch the object.
6274 NOTE: This directive is not related to caching. The replies
6275 is cached as usual even if you use always_direct. To not cache
6276 the replies see the 'cache' directive.
6278 This clause supports both fast and slow acl types.
6279 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6284 LOC: Config.accessList.NeverDirect
6287 Usage: never_direct allow|deny [!]aclname ...
6289 never_direct is the opposite of always_direct. Please read
6290 the description for always_direct if you have not already.
6292 With 'never_direct' you can use ACL elements to specify
6293 requests which should NEVER be forwarded directly to origin
6294 servers. For example, to force the use of a proxy for all
6295 requests, except those in your local domain use something like:
6297 acl local-servers dstdomain .foo.net
6298 never_direct deny local-servers
6299 never_direct allow all
6301 or if Squid is inside a firewall and there are local intranet
6302 servers inside the firewall use something like:
6304 acl local-intranet dstdomain .foo.net
6305 acl local-external dstdomain external.foo.net
6306 always_direct deny local-external
6307 always_direct allow local-intranet
6308 never_direct allow all
6310 This clause supports both fast and slow acl types.
6311 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6315 ADVANCED NETWORKING OPTIONS
6316 -----------------------------------------------------------------------------
6319 NAME: incoming_icp_average
6322 LOC: Config.comm_incoming.icp_average
6325 NAME: incoming_http_average
6328 LOC: Config.comm_incoming.http_average
6331 NAME: incoming_dns_average
6334 LOC: Config.comm_incoming.dns_average
6337 NAME: min_icp_poll_cnt
6340 LOC: Config.comm_incoming.icp_min_poll
6343 NAME: min_dns_poll_cnt
6346 LOC: Config.comm_incoming.dns_min_poll
6349 NAME: min_http_poll_cnt
6352 LOC: Config.comm_incoming.http_min_poll
6354 Heavy voodoo here. I can't even believe you are reading this.
6355 Are you crazy? Don't even think about adjusting these unless
6356 you understand the algorithms in comm_select.c first!
6362 LOC: Config.accept_filter
6366 The name of an accept(2) filter to install on Squid's
6367 listen socket(s). This feature is perhaps specific to
6368 FreeBSD and requires support in the kernel.
6370 The 'httpready' filter delays delivering new connections
6371 to Squid until a full HTTP request has been received.
6372 See the accf_http(9) man page for details.
6374 The 'dataready' filter delays delivering new connections
6375 to Squid until there is some data to process.
6376 See the accf_dataready(9) man page for details.
6380 The 'data' filter delays delivering of new connections
6381 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
6382 You may optionally specify a number of seconds to wait by
6383 'data=N' where N is the number of seconds. Defaults to 30
6384 if not specified. See the tcp(7) man page for details.
6387 accept_filter httpready
6392 NAME: client_ip_max_connections
6394 LOC: Config.client_ip_max_connections
6397 Set an absolute limit on the number of connections a single
6398 client IP can use. Any more than this and Squid will begin to drop
6399 new connections from the client until it closes some links.
6401 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
6402 connections from the client. For finer control use the ACL access controls.
6404 Requires client_db to be enabled (the default).
6406 WARNING: This may noticably slow down traffic received via external proxies
6407 or NAT devices and cause them to rebound error messages back to their clients.
6410 NAME: tcp_recv_bufsize
6414 LOC: Config.tcpRcvBufsz
6416 Size of receive buffer to set for TCP sockets. Probably just
6417 as easy to change your kernel's default. Set to zero to use
6418 the default buffer size.
6423 -----------------------------------------------------------------------------
6430 LOC: Adaptation::Icap::TheConfig.onoff
6433 If you want to enable the ICAP module support, set this to on.
6436 NAME: icap_connect_timeout
6439 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
6442 This parameter specifies how long to wait for the TCP connect to
6443 the requested ICAP server to complete before giving up and either
6444 terminating the HTTP transaction or bypassing the failure.
6446 The default for optional services is peer_connect_timeout.
6447 The default for essential services is connect_timeout.
6448 If this option is explicitly set, its value applies to all services.
6451 NAME: icap_io_timeout
6455 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
6458 This parameter specifies how long to wait for an I/O activity on
6459 an established, active ICAP connection before giving up and
6460 either terminating the HTTP transaction or bypassing the
6463 The default is read_timeout.
6466 NAME: icap_service_failure_limit
6467 COMMENT: limit [in memory-depth time-units]
6468 TYPE: icap_service_failure_limit
6470 LOC: Adaptation::Icap::TheConfig
6473 The limit specifies the number of failures that Squid tolerates
6474 when establishing a new TCP connection with an ICAP service. If
6475 the number of failures exceeds the limit, the ICAP service is
6476 not used for new ICAP requests until it is time to refresh its
6479 A negative value disables the limit. Without the limit, an ICAP
6480 service will not be considered down due to connectivity failures
6481 between ICAP OPTIONS requests.
6483 Squid forgets ICAP service failures older than the specified
6484 value of memory-depth. The memory fading algorithm
6485 is approximate because Squid does not remember individual
6486 errors but groups them instead, splitting the option
6487 value into ten time slots of equal length.
6489 When memory-depth is 0 and by default this option has no
6490 effect on service failure expiration.
6492 Squid always forgets failures when updating service settings
6493 using an ICAP OPTIONS transaction, regardless of this option
6497 # suspend service usage after 10 failures in 5 seconds:
6498 icap_service_failure_limit 10 in 5 seconds
6501 NAME: icap_service_revival_delay
6504 LOC: Adaptation::Icap::TheConfig.service_revival_delay
6507 The delay specifies the number of seconds to wait after an ICAP
6508 OPTIONS request failure before requesting the options again. The
6509 failed ICAP service is considered "down" until fresh OPTIONS are
6512 The actual delay cannot be smaller than the hardcoded minimum
6513 delay of 30 seconds.
6516 NAME: icap_preview_enable
6520 LOC: Adaptation::Icap::TheConfig.preview_enable
6523 The ICAP Preview feature allows the ICAP server to handle the
6524 HTTP message by looking only at the beginning of the message body
6525 or even without receiving the body at all. In some environments,
6526 previews greatly speedup ICAP processing.
6528 During an ICAP OPTIONS transaction, the server may tell Squid what
6529 HTTP messages should be previewed and how big the preview should be.
6530 Squid will not use Preview if the server did not request one.
6532 To disable ICAP Preview for all ICAP services, regardless of
6533 individual ICAP server OPTIONS responses, set this option to "off".
6535 icap_preview_enable off
6538 NAME: icap_preview_size
6541 LOC: Adaptation::Icap::TheConfig.preview_size
6544 The default size of preview data to be sent to the ICAP server.
6545 -1 means no preview. This value might be overwritten on a per server
6546 basis by OPTIONS requests.
6549 NAME: icap_206_enable
6553 LOC: Adaptation::Icap::TheConfig.allow206_enable
6556 206 (Partial Content) responses is an ICAP extension that allows the
6557 ICAP agents to optionally combine adapted and original HTTP message
6558 content. The decision to combine is postponed until the end of the
6559 ICAP response. Squid supports Partial Content extension by default.
6561 Activation of the Partial Content extension is negotiated with each
6562 ICAP service during OPTIONS exchange. Most ICAP servers should handle
6563 negotation correctly even if they do not support the extension, but
6564 some might fail. To disable Partial Content support for all ICAP
6565 services and to avoid any negotiation, set this option to "off".
6571 NAME: icap_default_options_ttl
6574 LOC: Adaptation::Icap::TheConfig.default_options_ttl
6577 The default TTL value for ICAP OPTIONS responses that don't have
6578 an Options-TTL header.
6581 NAME: icap_persistent_connections
6585 LOC: Adaptation::Icap::TheConfig.reuse_connections
6588 Whether or not Squid should use persistent connections to
6592 NAME: adaptation_send_client_ip icap_send_client_ip
6594 IFDEF: USE_ADAPTATION
6596 LOC: Adaptation::Config::send_client_ip
6599 If enabled, Squid shares HTTP client IP information with adaptation
6600 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
6601 For eCAP, Squid sets the libecap::metaClientIp transaction option.
6603 See also: adaptation_uses_indirect_client
6606 NAME: adaptation_send_username icap_send_client_username
6608 IFDEF: USE_ADAPTATION
6610 LOC: Adaptation::Config::send_username
6613 This sends authenticated HTTP client username (if available) to
6614 the adaptation service.
6616 For ICAP, the username value is encoded based on the
6617 icap_client_username_encode option and is sent using the header
6618 specified by the icap_client_username_header option.
6621 NAME: icap_client_username_header
6624 LOC: Adaptation::Icap::TheConfig.client_username_header
6625 DEFAULT: X-Client-Username
6627 ICAP request header name to use for adaptation_send_username.
6630 NAME: icap_client_username_encode
6634 LOC: Adaptation::Icap::TheConfig.client_username_encode
6637 Whether to base64 encode the authenticated client username.
6641 TYPE: icap_service_type
6643 LOC: Adaptation::Icap::TheConfig
6646 Defines a single ICAP service using the following format:
6648 icap_service id vectoring_point uri [option ...]
6651 an opaque identifier or name which is used to direct traffic to
6652 this specific service. Must be unique among all adaptation
6653 services in squid.conf.
6655 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
6656 This specifies at which point of transaction processing the
6657 ICAP service should be activated. *_postcache vectoring points
6658 are not yet supported.
6660 uri: icap://servername:port/servicepath
6661 ICAP server and service location.
6663 ICAP does not allow a single service to handle both REQMOD and RESPMOD
6664 transactions. Squid does not enforce that requirement. You can specify
6665 services with the same service_url and different vectoring_points. You
6666 can even specify multiple identical services as long as their
6667 service_names differ.
6670 Service options are separated by white space. ICAP services support
6671 the following name=value options:
6674 If set to 'on' or '1', the ICAP service is treated as
6675 optional. If the service cannot be reached or malfunctions,
6676 Squid will try to ignore any errors and process the message as
6677 if the service was not enabled. No all ICAP errors can be
6678 bypassed. If set to 0, the ICAP service is treated as
6679 essential and all ICAP errors will result in an error page
6680 returned to the HTTP client.
6682 Bypass is off by default: services are treated as essential.
6685 If set to 'on' or '1', the ICAP service is allowed to
6686 dynamically change the current message adaptation plan by
6687 returning a chain of services to be used next. The services
6688 are specified using the X-Next-Services ICAP response header
6689 value, formatted as a comma-separated list of service names.
6690 Each named service should be configured in squid.conf. Other
6691 services are ignored. An empty X-Next-Services value results
6692 in an empty plan which ends the current adaptation.
6694 Dynamic adaptation plan may cross or cover multiple supported
6695 vectoring points in their natural processing order.
6697 Routing is not allowed by default: the ICAP X-Next-Services
6698 response header is ignored.
6701 Only has effect on split-stack systems. The default on those systems
6702 is to use IPv4-only connections. When set to 'on' this option will
6703 make Squid use IPv6-only connections to contact this ICAP service.
6705 on-overload=block|bypass|wait|force
6706 If the service Max-Connections limit has been reached, do
6707 one of the following for each new ICAP transaction:
6708 * block: send an HTTP error response to the client
6709 * bypass: ignore the "over-connected" ICAP service
6710 * wait: wait (in a FIFO queue) for an ICAP connection slot
6711 * force: proceed, ignoring the Max-Connections limit
6713 In SMP mode with N workers, each worker assumes the service
6714 connection limit is Max-Connections/N, even though not all
6715 workers may use a given service.
6717 The default value is "bypass" if service is bypassable,
6718 otherwise it is set to "wait".
6722 Use the given number as the Max-Connections limit, regardless
6723 of the Max-Connections value given by the service, if any.
6725 Older icap_service format without optional named parameters is
6726 deprecated but supported for backward compatibility.
6729 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
6730 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
6734 TYPE: icap_class_type
6739 This deprecated option was documented to define an ICAP service
6740 chain, even though it actually defined a set of similar, redundant
6741 services, and the chains were not supported.
6743 To define a set of redundant services, please use the
6744 adaptation_service_set directive. For service chains, use
6745 adaptation_service_chain.
6749 TYPE: icap_access_type
6754 This option is deprecated. Please use adaptation_access, which
6755 has the same ICAP functionality, but comes with better
6756 documentation, and eCAP support.
6761 -----------------------------------------------------------------------------
6768 LOC: Adaptation::Ecap::TheConfig.onoff
6771 Controls whether eCAP support is enabled.
6775 TYPE: ecap_service_type
6777 LOC: Adaptation::Ecap::TheConfig
6780 Defines a single eCAP service
6782 ecap_service id vectoring_point uri [option ...]
6785 an opaque identifier or name which is used to direct traffic to
6786 this specific service. Must be unique among all adaptation
6787 services in squid.conf.
6789 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
6790 This specifies at which point of transaction processing the
6791 eCAP service should be activated. *_postcache vectoring points
6792 are not yet supported.
6794 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
6795 Squid uses the eCAP service URI to match this configuration
6796 line with one of the dynamically loaded services. Each loaded
6797 eCAP service must have a unique URI. Obtain the right URI from
6798 the service provider.
6801 Service options are separated by white space. eCAP services support
6802 the following name=value options:
6805 If set to 'on' or '1', the eCAP service is treated as optional.
6806 If the service cannot be reached or malfunctions, Squid will try
6807 to ignore any errors and process the message as if the service
6808 was not enabled. No all eCAP errors can be bypassed.
6809 If set to 'off' or '0', the eCAP service is treated as essential
6810 and all eCAP errors will result in an error page returned to the
6813 Bypass is off by default: services are treated as essential.
6816 If set to 'on' or '1', the eCAP service is allowed to
6817 dynamically change the current message adaptation plan by
6818 returning a chain of services to be used next.
6820 Dynamic adaptation plan may cross or cover multiple supported
6821 vectoring points in their natural processing order.
6823 Routing is not allowed by default.
6825 Older ecap_service format without optional named parameters is
6826 deprecated but supported for backward compatibility.
6830 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
6831 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
6834 NAME: loadable_modules
6836 IFDEF: USE_LOADABLE_MODULES
6837 LOC: Config.loadable_module_names
6840 Instructs Squid to load the specified dynamic module(s) or activate
6841 preloaded module(s).
6843 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
6847 MESSAGE ADAPTATION OPTIONS
6848 -----------------------------------------------------------------------------
6851 NAME: adaptation_service_set
6852 TYPE: adaptation_service_set_type
6853 IFDEF: USE_ADAPTATION
6858 Configures an ordered set of similar, redundant services. This is
6859 useful when hot standby or backup adaptation servers are available.
6861 adaptation_service_set set_name service_name1 service_name2 ...
6863 The named services are used in the set declaration order. The first
6864 applicable adaptation service from the set is used first. The next
6865 applicable service is tried if and only if the transaction with the
6866 previous service fails and the message waiting to be adapted is still
6869 When adaptation starts, broken services are ignored as if they were
6870 not a part of the set. A broken service is a down optional service.
6872 The services in a set must be attached to the same vectoring point
6873 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
6875 If all services in a set are optional then adaptation failures are
6876 bypassable. If all services in the set are essential, then a
6877 transaction failure with one service may still be retried using
6878 another service from the set, but when all services fail, the master
6879 transaction fails as well.
6881 A set may contain a mix of optional and essential services, but that
6882 is likely to lead to surprising results because broken services become
6883 ignored (see above), making previously bypassable failures fatal.
6884 Technically, it is the bypassability of the last failed service that
6887 See also: adaptation_access adaptation_service_chain
6890 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
6891 adaptation service_set svcLogger loggerLocal loggerRemote
6894 NAME: adaptation_service_chain
6895 TYPE: adaptation_service_chain_type
6896 IFDEF: USE_ADAPTATION
6901 Configures a list of complementary services that will be applied
6902 one-by-one, forming an adaptation chain or pipeline. This is useful
6903 when Squid must perform different adaptations on the same message.
6905 adaptation_service_chain chain_name service_name1 svc_name2 ...
6907 The named services are used in the chain declaration order. The first
6908 applicable adaptation service from the chain is used first. The next
6909 applicable service is applied to the successful adaptation results of
6910 the previous service in the chain.
6912 When adaptation starts, broken services are ignored as if they were
6913 not a part of the chain. A broken service is a down optional service.
6915 Request satisfaction terminates the adaptation chain because Squid
6916 does not currently allow declaration of RESPMOD services at the
6917 "reqmod_precache" vectoring point (see icap_service or ecap_service).
6919 The services in a chain must be attached to the same vectoring point
6920 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
6922 A chain may contain a mix of optional and essential services. If an
6923 essential adaptation fails (or the failure cannot be bypassed for
6924 other reasons), the master transaction fails. Otherwise, the failure
6925 is bypassed as if the failed adaptation service was not in the chain.
6927 See also: adaptation_access adaptation_service_set
6930 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
6933 NAME: adaptation_access
6934 TYPE: adaptation_access_type
6935 IFDEF: USE_ADAPTATION
6939 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
6941 adaptation_access service_name allow|deny [!]aclname...
6942 adaptation_access set_name allow|deny [!]aclname...
6944 At each supported vectoring point, the adaptation_access
6945 statements are processed in the order they appear in this
6946 configuration file. Statements pointing to the following services
6947 are ignored (i.e., skipped without checking their ACL):
6949 - services serving different vectoring points
6950 - "broken-but-bypassable" services
6951 - "up" services configured to ignore such transactions
6952 (e.g., based on the ICAP Transfer-Ignore header).
6954 When a set_name is used, all services in the set are checked
6955 using the same rules, to find the first applicable one. See
6956 adaptation_service_set for details.
6958 If an access list is checked and there is a match, the
6959 processing stops: For an "allow" rule, the corresponding
6960 adaptation service is used for the transaction. For a "deny"
6961 rule, no adaptation service is activated.
6963 It is currently not possible to apply more than one adaptation
6964 service at the same vectoring point to the same HTTP transaction.
6966 See also: icap_service and ecap_service
6969 adaptation_access service_1 allow all
6972 NAME: adaptation_service_iteration_limit
6974 IFDEF: USE_ADAPTATION
6975 LOC: Adaptation::Config::service_iteration_limit
6978 Limits the number of iterations allowed when applying adaptation
6979 services to a message. If your longest adaptation set or chain
6980 may have more than 16 services, increase the limit beyond its
6981 default value of 16. If detecting infinite iteration loops sooner
6982 is critical, make the iteration limit match the actual number
6983 of services in your longest adaptation set or chain.
6985 Infinite adaptation loops are most likely with routing services.
6987 See also: icap_service routing=1
6990 NAME: adaptation_masterx_shared_names
6992 IFDEF: USE_ADAPTATION
6993 LOC: Adaptation::Config::masterx_shared_name
6996 For each master transaction (i.e., the HTTP request and response
6997 sequence, including all related ICAP and eCAP exchanges), Squid
6998 maintains a table of metadata. The table entries are (name, value)
6999 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7000 with the master transaction.
7002 This option specifies the table entry names that Squid must accept
7003 from and forward to the adaptation transactions.
7005 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7006 shared table by returning an ICAP header field with a name
7007 specified in adaptation_masterx_shared_names.
7009 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7010 shared table by implementing the libecap::visitEachOption() API
7011 to provide an option with a name specified in
7012 adaptation_masterx_shared_names.
7014 Squid will store and forward the set entry to subsequent adaptation
7015 transactions within the same master transaction scope.
7017 Only one shared entry name is supported at this time.
7020 # share authentication information among ICAP services
7021 adaptation_masterx_shared_names X-Subscriber-ID
7024 NAME: adaptation_meta
7025 TYPE: adaptation_meta_type
7026 IFDEF: USE_ADAPTATION
7027 LOC: Adaptation::Config::metaHeaders
7030 This option allows Squid administrator to add custom ICAP request
7031 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7032 Use it to pass custom authentication tokens and other
7033 transaction-state related meta information to an ICAP/eCAP service.
7035 The addition of a meta header is ACL-driven:
7036 adaptation_meta name value [!]aclname ...
7038 Processing for a given header name stops after the first ACL list match.
7039 Thus, it is impossible to add two headers with the same name. If no ACL
7040 lists match for a given header name, no such header is added. For
7043 # do not debug transactions except for those that need debugging
7044 adaptation_meta X-Debug 1 needs_debugging
7046 # log all transactions except for those that must remain secret
7047 adaptation_meta X-Log 1 !keep_secret
7049 # mark transactions from users in the "G 1" group
7050 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7052 The "value" parameter may be a regular squid.conf token or a "double
7053 quoted string". Within the quoted string, use backslash (\) to escape
7054 any character, which is currently only useful for escaping backslashes
7055 and double quotes. For example,
7056 "this string has one backslash (\\) and two \"quotes\""
7062 LOC: Adaptation::Icap::TheConfig.repeat
7063 DEFAULT_IF_NONE: deny all
7065 This ACL determines which retriable ICAP transactions are
7066 retried. Transactions that received a complete ICAP response
7067 and did not have to consume or produce HTTP bodies to receive
7068 that response are usually retriable.
7070 icap_retry allow|deny [!]aclname ...
7072 Squid automatically retries some ICAP I/O timeouts and errors
7073 due to persistent connection race conditions.
7075 See also: icap_retry_limit
7078 NAME: icap_retry_limit
7081 LOC: Adaptation::Icap::TheConfig.repeat_limit
7084 Limits the number of retries allowed. When set to zero (default),
7085 no retries are allowed.
7087 Communication errors due to persistent connection race
7088 conditions are unavoidable, automatically retried, and do not
7089 count against this limit.
7091 See also: icap_retry
7097 -----------------------------------------------------------------------------
7100 NAME: check_hostnames
7103 LOC: Config.onoff.check_hostnames
7105 For security and stability reasons Squid can check
7106 hostnames for Internet standard RFC compliance. If you want
7107 Squid to perform these checks turn this directive on.
7110 NAME: allow_underscore
7113 LOC: Config.onoff.allow_underscore
7115 Underscore characters is not strictly allowed in Internet hostnames
7116 but nevertheless used by many sites. Set this to off if you want
7117 Squid to be strict about the standard.
7118 This check is performed only when check_hostnames is set to on.
7121 NAME: cache_dns_program
7123 IFDEF: USE_DNSSERVERS
7124 DEFAULT: @DEFAULT_DNSSERVER@
7125 LOC: Config.Program.dnsserver
7127 Specify the location of the executable for dnslookup process.
7131 TYPE: HelperChildConfig
7132 IFDEF: USE_DNSSERVERS
7133 DEFAULT: 32 startup=1 idle=1
7134 LOC: Config.dnsChildren
7136 The maximum number of processes spawn to service DNS name lookups.
7137 If you limit it too few Squid will have to wait for them to process
7138 a backlog of requests, slowing it down. If you allow too many they
7139 will use RAM and other system resources noticably.
7140 The maximum this may be safely set to is 32.
7142 The startup= and idle= options allow some measure of skew in your
7147 Sets a minimum of how many processes are to be spawned when Squid
7148 starts or reconfigures. When set to zero the first request will
7149 cause spawning of the first child process to handle it.
7151 Starting too few will cause an initial slowdown in traffic as Squid
7152 attempts to simultaneously spawn enough processes to cope.
7156 Sets a minimum of how many processes Squid is to try and keep available
7157 at all times. When traffic begins to rise above what the existing
7158 processes can handle this many more will be spawned up to the maximum
7159 configured. A minimum setting of 1 is required.
7162 NAME: dns_retransmit_interval
7165 LOC: Config.Timeout.idns_retransmit
7166 IFDEF: !USE_DNSSERVERS
7168 Initial retransmit interval for DNS queries. The interval is
7169 doubled each time all configured DNS servers have been tried.
7175 LOC: Config.Timeout.idns_query
7176 IFDEF: !USE_DNSSERVERS
7178 DNS Query timeout. If no response is received to a DNS query
7179 within this time all DNS servers for the queried domain
7180 are assumed to be unavailable.
7183 NAME: dns_packet_max
7186 LOC: Config.dns.packet_max
7187 IFDEF: !USE_DNSSERVERS
7189 Maximum number of bytes packet size to advertise via EDNS.
7190 Set to "none" to disable EDNS large packet support.
7192 For legacy reasons DNS UDP replies will default to 512 bytes which
7193 is too small for many responses. EDNS provides a means for Squid to
7194 negotiate receiving larger responses back immediately without having
7195 to failover with repeat requests. Responses larger than this limit
7196 will retain the old behaviour of failover to TCP DNS.
7198 Squid has no real fixed limit internally, but allowing packet sizes
7199 over 1500 bytes requires network jumbogram support and is usually not
7202 WARNING: The RFC also indicates that some older resolvers will reply
7203 with failure of the whole request if the extension is added. Some
7204 resolvers have already been identified which will reply with mangled
7205 EDNS response on occasion. Usually in response to many-KB jumbogram
7206 sizes being advertised by Squid.
7207 Squid will currently treat these both as an unable-to-resolve domain
7208 even if it would be resolvable without EDNS.
7215 LOC: Config.onoff.res_defnames
7217 Normally the RES_DEFNAMES resolver option is disabled
7218 (see res_init(3)). This prevents caches in a hierarchy
7219 from interpreting single-component hostnames locally. To allow
7220 Squid to handle single-component names, enable this option.
7223 NAME: dns_nameservers
7226 LOC: Config.dns_nameservers
7228 Use this if you want to specify a list of DNS name servers
7229 (IP addresses) to use instead of those given in your
7230 /etc/resolv.conf file.
7231 On Windows platforms, if no value is specified here or in
7232 the /etc/resolv.conf file, the list of DNS name servers are
7233 taken from the Windows registry, both static and dynamic DHCP
7234 configurations are supported.
7236 Example: dns_nameservers 10.0.0.1 192.172.0.4
7241 DEFAULT: @DEFAULT_HOSTS@
7242 LOC: Config.etcHostsPath
7244 Location of the host-local IP name-address associations
7245 database. Most Operating Systems have such a file on different
7247 - Un*X & Linux: /etc/hosts
7248 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7249 (%SystemRoot% value install default is c:\winnt)
7250 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7251 (%SystemRoot% value install default is c:\windows)
7252 - Windows 9x/Me: %windir%\hosts
7253 (%windir% value is usually c:\windows)
7254 - Cygwin: /etc/hosts
7256 The file contains newline-separated definitions, in the
7257 form ip_address_in_dotted_form name [name ...] names are
7258 whitespace-separated. Lines beginning with an hash (#)
7259 character are comments.
7261 The file is checked at startup and upon configuration.
7262 If set to 'none', it won't be checked.
7263 If append_domain is used, that domain will be added to
7264 domain-local (i.e. not containing any dot character) host
7270 LOC: Config.appendDomain
7273 Appends local domain name to hostnames without any dots in
7274 them. append_domain must begin with a period.
7276 Be warned there are now Internet names with no dots in
7277 them using only top-domain names, so setting this may
7278 cause some Internet sites to become unavailable.
7281 append_domain .yourdomain.com
7284 NAME: ignore_unknown_nameservers
7286 LOC: Config.onoff.ignore_unknown_nameservers
7288 IFDEF: !USE_DNSSERVERS
7290 By default Squid checks that DNS responses are received
7291 from the same IP addresses they are sent to. If they
7292 don't match, Squid ignores the response and writes a warning
7293 message to cache.log. You can allow responses from unknown
7294 nameservers by setting this option to 'off'.
7297 NAME: dns_v4_fallback
7300 LOC: Config.onoff.dns_require_A
7301 IFDEF: !USE_DNSSERVERS
7303 Standard practice with DNS is to lookup either A or AAAA records
7304 and use the results if it succeeds. Only looking up the other if
7305 the first attempt fails or otherwise produces no results.
7307 That policy however will cause squid to produce error pages for some
7308 servers that advertise AAAA but are unreachable over IPv6.
7310 If this is ON squid will always lookup both AAAA and A, using both.
7311 If this is OFF squid will lookup AAAA and only try A if none found.
7313 WARNING: There are some possibly unwanted side-effects with this on:
7314 *) Doubles the load placed by squid on the DNS network.
7315 *) May negatively impact connection delay times.
7321 LOC: Config.dns.v4_first
7322 IFDEF: !USE_DNSSERVERS
7324 With the IPv6 Internet being as fast or faster than IPv4 Internet
7325 for most networks Squid prefers to contact websites over IPv6.
7327 This option reverses the order of preference to make Squid contact
7328 dual-stack websites over IPv4 first. Squid will still perform both
7329 IPv6 and IPv4 DNS lookups before connecting.
7332 This option will restrict the situations under which IPv6
7333 connectivity is used (and tested). Hiding network problems
7334 which would otherwise be detected and warned about.
7338 COMMENT: (number of entries)
7341 LOC: Config.ipcache.size
7348 LOC: Config.ipcache.low
7355 LOC: Config.ipcache.high
7357 The size, low-, and high-water marks for the IP cache.
7360 NAME: fqdncache_size
7361 COMMENT: (number of entries)
7364 LOC: Config.fqdncache.size
7366 Maximum number of FQDN cache entries.
7371 -----------------------------------------------------------------------------
7378 LOC: Config.onoff.mem_pools
7380 If set, Squid will keep pools of allocated (but unused) memory
7381 available for future use. If memory is a premium on your
7382 system and you believe your malloc library outperforms Squid
7383 routines, disable this.
7386 NAME: memory_pools_limit
7390 LOC: Config.MemPools.limit
7392 Used only with memory_pools on:
7393 memory_pools_limit 50 MB
7395 If set to a non-zero value, Squid will keep at most the specified
7396 limit of allocated (but unused) memory in memory pools. All free()
7397 requests that exceed this limit will be handled by your malloc
7398 library. Squid does not pre-allocate any memory, just safe-keeps
7399 objects that otherwise would be free()d. Thus, it is safe to set
7400 memory_pools_limit to a reasonably high value even if your
7401 configuration will use less memory.
7403 If set to none, Squid will keep all memory it can. That is, there
7404 will be no limit on the total amount of memory used for safe-keeping.
7406 To disable memory allocation optimization, do not set
7407 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
7409 An overhead for maintaining memory pools is not taken into account
7410 when the limit is checked. This overhead is close to four bytes per
7411 object kept. However, pools may actually _save_ memory because of
7412 reduced memory thrashing in your malloc library.
7416 COMMENT: on|off|transparent|truncate|delete
7419 LOC: opt_forwarded_for
7421 If set to "on", Squid will append your client's IP address
7422 in the HTTP requests it forwards. By default it looks like:
7424 X-Forwarded-For: 192.1.2.3
7426 If set to "off", it will appear as
7428 X-Forwarded-For: unknown
7430 If set to "transparent", Squid will not alter the
7431 X-Forwarded-For header in any way.
7433 If set to "delete", Squid will delete the entire
7434 X-Forwarded-For header.
7436 If set to "truncate", Squid will remove all existing
7437 X-Forwarded-For entries, and place the client IP as the sole entry.
7440 NAME: cachemgr_passwd
7441 TYPE: cachemgrpasswd
7443 LOC: Config.passwd_list
7445 Specify passwords for cachemgr operations.
7447 Usage: cachemgr_passwd password action action ...
7449 Some valid actions are (see cache manager menu for a full list):
7489 * Indicates actions which will not be performed without a
7490 valid password, others can be performed if not listed here.
7492 To disable an action, set the password to "disable".
7493 To allow performing an action without a password, set the
7496 Use the keyword "all" to set the same password for all actions.
7499 cachemgr_passwd secret shutdown
7500 cachemgr_passwd lesssssssecret info stats/objects
7501 cachemgr_passwd disable all
7508 LOC: Config.onoff.client_db
7510 If you want to disable collecting per-client statistics,
7511 turn off client_db here.
7514 NAME: refresh_all_ims
7518 LOC: Config.onoff.refresh_all_ims
7520 When you enable this option, squid will always check
7521 the origin server for an update when a client sends an
7522 If-Modified-Since request. Many browsers use IMS
7523 requests when the user requests a reload, and this
7524 ensures those clients receive the latest version.
7526 By default (off), squid may return a Not Modified response
7527 based on the age of the cached version.
7530 NAME: reload_into_ims
7531 IFDEF: USE_HTTP_VIOLATIONS
7535 LOC: Config.onoff.reload_into_ims
7537 When you enable this option, client no-cache or ``reload''
7538 requests will be changed to If-Modified-Since requests.
7539 Doing this VIOLATES the HTTP standard. Enabling this
7540 feature could make you liable for problems which it
7543 see also refresh_pattern for a more selective approach.
7546 NAME: connect_retries
7548 LOC: Config.connect_retries
7551 This sets the maximum number of connection attempts made for each
7552 TCP connection. The connect_retries attempts must all still
7553 complete within the connection timeout period.
7555 The default is not to re-try if the first connection attempt fails.
7556 The (not recommended) maximum is 10 tries.
7558 A warning message will be generated if it is set to a too-high
7559 value and the configured value will be over-ridden.
7561 Note: These re-tries are in addition to forward_max_tries
7562 which limit how many different addresses may be tried to find
7566 NAME: retry_on_error
7568 LOC: Config.retry.onerror
7571 If set to ON Squid will automatically retry requests when
7572 receiving an error response with status 403 (Forbidden),
7573 500 (Internal Error), 501 or 503 (Service not available).
7574 Status 502 and 504 (Gateway errors) are always retried.
7576 This is mainly useful if you are in a complex cache hierarchy to
7577 work around access control errors.
7579 NOTE: This retry will attempt to find another working destination.
7580 Which is different from the server which just failed.
7583 NAME: as_whois_server
7585 LOC: Config.as_whois_server
7586 DEFAULT: whois.ra.net
7588 WHOIS server to query for AS numbers. NOTE: AS numbers are
7589 queried only when Squid starts up, not for every request.
7594 LOC: Config.onoff.offline
7597 Enable this option and Squid will never try to validate cached
7601 NAME: uri_whitespace
7602 TYPE: uri_whitespace
7603 LOC: Config.uri_whitespace
7606 What to do with requests that have whitespace characters in the
7609 strip: The whitespace characters are stripped out of the URL.
7610 This is the behavior recommended by RFC2396.
7611 deny: The request is denied. The user receives an "Invalid
7613 allow: The request is allowed and the URI is not changed. The
7614 whitespace characters remain in the URI. Note the
7615 whitespace is passed to redirector processes if they
7617 encode: The request is allowed and the whitespace characters are
7618 encoded according to RFC1738. This could be considered
7619 a violation of the HTTP/1.1
7620 RFC because proxies are not allowed to rewrite URI's.
7621 chop: The request is allowed and the URI is chopped at the
7622 first whitespace. This might also be considered a
7628 LOC: Config.chroot_dir
7631 Specifies a directory where Squid should do a chroot() while
7632 initializing. This also causes Squid to fully drop root
7633 privileges after initializing. This means, for example, if you
7634 use a HTTP port less than 1024 and try to reconfigure, you may
7635 get an error saying that Squid can not open the port.
7638 NAME: balance_on_multiple_ip
7640 LOC: Config.onoff.balance_on_multiple_ip
7643 Modern IP resolvers in squid sort lookup results by preferred access.
7644 By default squid will use these IP in order and only rotates to
7645 the next listed when the most preffered fails.
7647 Some load balancing servers based on round robin DNS have been
7648 found not to preserve user session state across requests
7649 to different IP addresses.
7651 Enabling this directive Squid rotates IP's per request.
7654 NAME: pipeline_prefetch
7656 LOC: Config.onoff.pipeline_prefetch
7659 To boost the performance of pipelined requests to closer
7660 match that of a non-proxied environment Squid can try to fetch
7661 up to two requests in parallel from a pipeline.
7663 Defaults to off for bandwidth management and access logging
7666 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
7669 NAME: high_response_time_warning
7672 LOC: Config.warnings.high_rptm
7675 If the one-minute median response time exceeds this value,
7676 Squid prints a WARNING with debug level 0 to get the
7677 administrators attention. The value is in milliseconds.
7680 NAME: high_page_fault_warning
7682 LOC: Config.warnings.high_pf
7685 If the one-minute average page fault rate exceeds this
7686 value, Squid prints a WARNING with debug level 0 to get
7687 the administrators attention. The value is in page faults
7691 NAME: high_memory_warning
7693 LOC: Config.warnings.high_memory
7696 If the memory usage (as determined by mallinfo) exceeds
7697 this amount, Squid prints a WARNING with debug level 0 to get
7698 the administrators attention.
7701 NAME: sleep_after_fork
7702 COMMENT: (microseconds)
7704 LOC: Config.sleep_after_fork
7707 When this is set to a non-zero value, the main Squid process
7708 sleeps the specified number of microseconds after a fork()
7709 system call. This sleep may help the situation where your
7710 system reports fork() failures due to lack of (virtual)
7711 memory. Note, however, if you have a lot of child
7712 processes, these sleep delays will add up and your
7713 Squid will not service requests for some amount of time
7714 until all the child processes have been started.
7715 On Windows value less then 1000 (1 milliseconds) are
7719 NAME: windows_ipaddrchangemonitor
7720 IFDEF: _SQUID_MSWIN_
7724 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
7726 On Windows Squid by default will monitor IP address changes and will
7727 reconfigure itself after any detected event. This is very useful for
7728 proxies connected to internet with dial-up interfaces.
7729 In some cases (a Proxy server acting as VPN gateway is one) it could be
7730 desiderable to disable this behaviour setting this to 'off'.
7731 Note: after changing this, Squid service must be restarted.
7736 IFDEF: USE_SQUID_EUI
7738 LOC: Eui::TheConfig.euiLookup
7740 Whether to lookup the EUI or MAC address of a connected client.
7743 NAME: max_filedescriptors max_filedesc
7746 LOC: Config.max_filedescriptors
7748 The maximum number of filedescriptors supported.
7750 The default "0" means Squid inherits the current ulimit setting.
7752 Note: Changing this requires a restart of Squid. Also
7753 not all comm loops supports large values.
7761 Number of main Squid processes or "workers" to fork and maintain.
7762 0: "no daemon" mode, like running "squid -N ..."
7763 1: "no SMP" mode, start one main Squid process daemon (default)
7764 N: start N main Squid process daemons (i.e., SMP mode)
7766 In SMP mode, each worker does nearly all what a single Squid daemon
7767 does (e.g., listen on http_port and forward HTTP requests).
7770 NAME: cpu_affinity_map
7771 TYPE: CpuAffinityMap
7772 LOC: Config.cpuAffinityMap
7775 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
7777 Sets 1:1 mapping between Squid processes and CPU cores. For example,
7779 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
7781 affects processes 1 through 4 only and places them on the first
7782 four even cores, starting with core #1.
7784 CPU cores are numbered starting from 1. Requires support for
7785 sched_getaffinity(2) and sched_setaffinity(2) system calls.
7787 Multiple cpu_affinity_map options are merged.