2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the documentation for the Squid configuration file.
34 This documentation can also be found online at:
35 http://www.squid-cache.org/Doc/config/
37 You may wish to look at the Squid home page and wiki for the
38 FAQ and other documentation:
39 http://www.squid-cache.org/
40 http://wiki.squid-cache.org/SquidFaq
41 http://wiki.squid-cache.org/ConfigExamples
43 This documentation shows what the defaults for various directives
44 happen to be. If you don't need to change the default, you should
45 leave the line out of your squid.conf in most cases.
47 In some cases "none" refers to no default setting at all,
48 while in other cases it refers to the value of the option
49 - the comments for that keyword indicate if this is the case.
54 Configuration options can be included using the "include" directive.
55 Include takes a list of files to include. Quoting and wildcards are
60 include /path/to/included/file/squid.acl.config
62 Includes can be nested up to a hard-coded depth of 16 levels.
63 This arbitrary restriction is to prevent recursive include references
64 from causing Squid entering an infinite loop whilst trying to load
68 Conditional configuration
70 If-statements can be used to make configuration directives
74 ... regular configuration directives ...
76 ... regular configuration directives ...]
79 The else part is optional. The keywords "if", "else", and "endif"
80 must be typed on their own lines, as if they were regular
81 configuration directives.
83 NOTE: An else-if condition is not supported.
85 These individual conditions types are supported:
88 Always evaluates to true.
90 Always evaluates to false.
92 Equality comparison of two integer numbers.
97 The following SMP-related preprocessor macros can be used.
99 ${process_name} expands to the current Squid process "name"
100 (e.g., squid1, squid2, or cache1).
102 ${process_number} expands to the current Squid process
103 identifier, which is an integer number (e.g., 1, 2, 3) unique
104 across all Squid processes.
107 # no Options Removed in 3.3
109 # Options Removed in 3.2
110 NAME: ignore_expect_100
113 Remove this line. The HTTP/1.1 feature is now fully supported by default.
116 NAME: dns_v4_fallback
119 Remove this line. Squid performs a 'Happy Eyeballs' algorithm, the 'fallback' algorithm is no longer relevant.
125 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
128 NAME: maximum_single_addr_tries
131 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
137 Remove this line. The feature is supported by default in storage types where update is implemented.
140 NAME: url_rewrite_concurrency
143 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
146 # Options Removed in 3.1
150 Remove this line. DNS is no longer tested on startup.
153 NAME: extension_methods
156 Remove this line. All valid methods for HTTP are accepted by default.
159 # 2.7 Options Removed/Replaced in 3.2
164 # 2.7 Options Removed/Replaced in 3.1
172 Remove this line. HTTP/1.1 is supported by default.
175 NAME: upgrade_http0.9
178 Remove this line. ICY/1.0 streaming protocol is supported by default.
181 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
184 Alter these entries. Use the qos_flows directive instead.
187 # Options Removed in 3.0
191 Since squid-3.0 replace with request_header_access or reply_header_access
192 depending on whether you wish to match client requests or server replies.
195 NAME: httpd_accel_no_pmtu_disc
198 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
202 OPTIONS FOR AUTHENTICATION
203 -----------------------------------------------------------------------------
212 This is used to define parameters for the various authentication
213 schemes supported by Squid.
215 format: auth_param scheme parameter [setting]
217 The order in which authentication schemes are presented to the client is
218 dependent on the order the scheme first appears in config file. IE
219 has a bug (it's not RFC 2617 compliant) in that it will use the basic
220 scheme if basic is the first entry presented, even if more secure
221 schemes are presented. For now use the order in the recommended
222 settings section below. If other browsers have difficulties (don't
223 recognize the schemes offered even if you are using basic) either
224 put basic first, or disable the other schemes (by commenting out their
227 Once an authentication scheme is fully configured, it can only be
228 shutdown by shutting squid down and restarting. Changes can be made on
229 the fly and activated with a reconfigure. I.E. You can change to a
230 different helper, but not unconfigure the helper completely.
232 Please note that while this directive defines how Squid processes
233 authentication it does not automatically activate authentication.
234 To use authentication you must in addition make use of ACLs based
235 on login name in http_access (proxy_auth, proxy_auth_regex or
236 external with %LOGIN used in the format tag). The browser will be
237 challenged for authentication on the first such acl encountered
238 in http_access processing and will also be re-challenged for new
239 login credentials if the request is being denied by a proxy_auth
242 WARNING: authentication can't be used in a transparently intercepting
243 proxy as the client then thinks it is talking to an origin server and
244 not the proxy. This is a limitation of bending the TCP/IP protocol to
245 transparently intercepting port 80, not a limitation in Squid.
246 Ports flagged 'transparent', 'intercept', or 'tproxy' have
247 authentication disabled.
249 === Parameters for the basic scheme follow. ===
252 Specify the command for the external authenticator. Such a program
253 reads a line containing "username password" and replies with one of
260 the user does not exist.
263 An internal error occurred in the helper, preventing
264 a result being identified.
266 "ERR" and "BH" results may optionally be followed by message="..."
267 containing a description available as %m in the returned error page.
269 If you use an authenticator, make sure you have 1 acl of type
272 By default, the basic authentication scheme is not used unless a
273 program is specified.
275 If you want to use the traditional NCSA proxy authentication, set
276 this line to something like
278 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
281 HTTP uses iso-latin-1 as character set, while some authentication
282 backends such as LDAP expects UTF-8. If this is set to on Squid will
283 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
284 username & password to the helper.
286 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
287 The maximum number of authenticator processes to spawn. If you start too few
288 Squid will have to wait for them to process a backlog of credential
289 verifications, slowing it down. When password verifications are
290 done via a (slow) network you are likely to need lots of
291 authenticator processes.
293 The startup= and idle= options permit some skew in the exact amount
294 run. A minimum of startup=N will begin during startup and reconfigure.
295 Squid will start more in groups of up to idle=N in an attempt to meet
296 traffic needs and to keep idle=N free above those traffic needs up to
299 The concurrency= option sets the number of concurrent requests the
300 helper can process. The default of 0 is used for helpers who only
301 supports one request at a time. Setting this to a number greater than
302 0 changes the protocol used to include a channel number first on the
303 request/response line, allowing multiple requests to be sent to the
304 same helper in parallel without waiting for the response.
305 Must not be set unless it's known the helper supports this.
307 auth_param basic children 20 startup=0 idle=1
310 Specifies the realm name which is to be reported to the
311 client for the basic proxy authentication scheme (part of
312 the text the user will see when prompted their username and
313 password). There is no default.
314 auth_param basic realm Squid proxy-caching web server
316 "credentialsttl" timetolive
317 Specifies how long squid assumes an externally validated
318 username:password pair is valid for - in other words how
319 often the helper program is called for that user. Set this
320 low to force revalidation with short lived passwords. Note
321 setting this high does not impact your susceptibility
322 to replay attacks unless you are using an one-time password
323 system (such as SecureID). If you are using such a system,
324 you will be vulnerable to replay attacks unless you also
325 use the max_user_ip ACL in an http_access rule.
327 "casesensitive" on|off
328 Specifies if usernames are case sensitive. Most user databases are
329 case insensitive allowing the same username to be spelled using both
330 lower and upper case letters, but some are case sensitive. This
331 makes a big difference for user_max_ip ACL processing and similar.
332 auth_param basic casesensitive off
334 === Parameters for the digest scheme follow ===
337 Specify the command for the external authenticator. Such
338 a program reads a line containing "username":"realm" and
339 replies with one of three results:
342 the user exists. The ha1= key is mandatory and
343 contains the appropriate H(A1) value, hex encoded.
344 See rfc 2616 for the definition of H(A1).
347 the user does not exist.
350 An internal error occurred in the helper, preventing
351 a result being identified.
353 "ERR" and "BH" results may optionally be followed by message="..."
354 containing a description available as %m in the returned error page.
356 By default, the digest authentication scheme is not used unless a
357 program is specified.
359 If you want to use a digest authenticator, set this line to
362 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
365 HTTP uses iso-latin-1 as character set, while some authentication
366 backends such as LDAP expects UTF-8. If this is set to on Squid will
367 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
368 username & password to the helper.
370 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
371 The maximum number of authenticator processes to spawn (default 5).
372 If you start too few Squid will have to wait for them to
373 process a backlog of H(A1) calculations, slowing it down.
374 When the H(A1) calculations are done via a (slow) network
375 you are likely to need lots of authenticator processes.
377 The startup= and idle= options permit some skew in the exact amount
378 run. A minimum of startup=N will begin during startup and reconfigure.
379 Squid will start more in groups of up to idle=N in an attempt to meet
380 traffic needs and to keep idle=N free above those traffic needs up to
383 The concurrency= option sets the number of concurrent requests the
384 helper can process. The default of 0 is used for helpers who only
385 supports one request at a time. Setting this to a number greater than
386 0 changes the protocol used to include a channel number first on the
387 request/response line, allowing multiple requests to be sent to the
388 same helper in parallel without waiting for the response.
389 Must not be set unless it's known the helper supports this.
391 auth_param digest children 20 startup=0 idle=1
394 Specifies the realm name which is to be reported to the
395 client for the digest proxy authentication scheme (part of
396 the text the user will see when prompted their username and
397 password). There is no default.
398 auth_param digest realm Squid proxy-caching web server
400 "nonce_garbage_interval" timeinterval
401 Specifies the interval that nonces that have been issued
402 to client_agent's are checked for validity.
404 "nonce_max_duration" timeinterval
405 Specifies the maximum length of time a given nonce will be
408 "nonce_max_count" number
409 Specifies the maximum number of times a given nonce can be
412 "nonce_strictness" on|off
413 Determines if squid requires strict increment-by-1 behavior
414 for nonce counts, or just incrementing (off - for use when
415 user agents generate nonce counts that occasionally miss 1
416 (ie, 1,2,4,6)). Default off.
418 "check_nonce_count" on|off
419 This directive if set to off can disable the nonce count check
420 completely to work around buggy digest qop implementations in
421 certain mainstream browser versions. Default on to check the
422 nonce count to protect from authentication replay attacks.
424 "post_workaround" on|off
425 This is a workaround to certain buggy browsers who sends
426 an incorrect request digest in POST requests when reusing
427 the same nonce as acquired earlier on a GET request.
429 === NTLM scheme options follow ===
432 Specify the command for the external NTLM authenticator.
433 Such a program reads exchanged NTLMSSP packets with
434 the browser via Squid until authentication is completed.
435 If you use an NTLM authenticator, make sure you have 1 acl
436 of type proxy_auth. By default, the NTLM authenticator program
439 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
441 "children" numberofchildren [startup=N] [idle=N]
442 The maximum number of authenticator processes to spawn (default 5).
443 If you start too few Squid will have to wait for them to
444 process a backlog of credential verifications, slowing it
445 down. When credential verifications are done via a (slow)
446 network you are likely to need lots of authenticator
449 The startup= and idle= options permit some skew in the exact amount
450 run. A minimum of startup=N will begin during startup and reconfigure.
451 Squid will start more in groups of up to idle=N in an attempt to meet
452 traffic needs and to keep idle=N free above those traffic needs up to
455 auth_param ntlm children 20 startup=0 idle=1
458 If you experience problems with PUT/POST requests when using the
459 Negotiate authentication scheme then you can try setting this to
460 off. This will cause Squid to forcibly close the connection on
461 the initial requests where the browser asks which schemes are
462 supported by the proxy.
464 auth_param ntlm keep_alive on
466 === Options for configuring the NEGOTIATE auth-scheme follow ===
469 Specify the command for the external Negotiate authenticator.
470 This protocol is used in Microsoft Active-Directory enabled setups with
471 the Microsoft Internet Explorer or Mozilla Firefox browsers.
472 Its main purpose is to exchange credentials with the Squid proxy
473 using the Kerberos mechanisms.
474 If you use a Negotiate authenticator, make sure you have at least
475 one acl of type proxy_auth active. By default, the negotiate
476 authenticator program is not used.
477 The only supported program for this role is the ntlm_auth
478 program distributed as part of Samba, version 4 or later.
480 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
482 "children" numberofchildren [startup=N] [idle=N]
483 The maximum number of authenticator processes to spawn (default 5).
484 If you start too few Squid will have to wait for them to
485 process a backlog of credential verifications, slowing it
486 down. When credential verifications are done via a (slow)
487 network you are likely to need lots of authenticator
490 The startup= and idle= options permit some skew in the exact amount
491 run. A minimum of startup=N will begin during startup and reconfigure.
492 Squid will start more in groups of up to idle=N in an attempt to meet
493 traffic needs and to keep idle=N free above those traffic needs up to
496 auth_param negotiate children 20 startup=0 idle=1
499 If you experience problems with PUT/POST requests when using the
500 Negotiate authentication scheme then you can try setting this to
501 off. This will cause Squid to forcibly close the connection on
502 the initial requests where the browser asks which schemes are
503 supported by the proxy.
505 auth_param negotiate keep_alive on
510 #Recommended minimum configuration per scheme:
511 #auth_param negotiate program <uncomment and complete this line to activate>
512 #auth_param negotiate children 20 startup=0 idle=1
513 #auth_param negotiate keep_alive on
515 #auth_param ntlm program <uncomment and complete this line to activate>
516 #auth_param ntlm children 20 startup=0 idle=1
517 #auth_param ntlm keep_alive on
519 #auth_param digest program <uncomment and complete this line>
520 #auth_param digest children 20 startup=0 idle=1
521 #auth_param digest realm Squid proxy-caching web server
522 #auth_param digest nonce_garbage_interval 5 minutes
523 #auth_param digest nonce_max_duration 30 minutes
524 #auth_param digest nonce_max_count 50
526 #auth_param basic program <uncomment and complete this line>
527 #auth_param basic children 5 startup=5 idle=1
528 #auth_param basic realm Squid proxy-caching web server
529 #auth_param basic credentialsttl 2 hours
532 NAME: authenticate_cache_garbage_interval
535 LOC: Config.authenticateGCInterval
537 The time period between garbage collection across the username cache.
538 This is a trade-off between memory utilization (long intervals - say
539 2 days) and CPU (short intervals - say 1 minute). Only change if you
543 NAME: authenticate_ttl
546 LOC: Config.authenticateTTL
548 The time a user & their credentials stay in the logged in
549 user cache since their last request. When the garbage
550 interval passes, all user credentials that have passed their
551 TTL are removed from memory.
554 NAME: authenticate_ip_ttl
556 LOC: Config.authenticateIpTTL
559 If you use proxy authentication and the 'max_user_ip' ACL,
560 this directive controls how long Squid remembers the IP
561 addresses associated with each user. Use a small value
562 (e.g., 60 seconds) if your users might change addresses
563 quickly, as is the case with dialup. You might be safe
564 using a larger value (e.g., 2 hours) in a corporate LAN
565 environment with relatively static address assignments.
570 -----------------------------------------------------------------------------
573 NAME: external_acl_type
574 TYPE: externalAclHelper
575 LOC: Config.externalAclHelperList
578 This option defines external acl classes using a helper program
579 to look up the status
581 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
585 ttl=n TTL in seconds for cached results (defaults to 3600
588 TTL for cached negative lookups (default same
591 Maximum number of acl helper processes spawned to service
592 external acl lookups of this type. (default 20)
594 Minimum number of acl helper processes to spawn during
595 startup and reconfigure to service external acl lookups
596 of this type. (default 0)
598 Number of acl helper processes to keep ahead of traffic
599 loads. Squid will spawn this many at once whenever load
600 rises above the capabilities of existing processes.
601 Up to the value of children-max. (default 1)
602 concurrency=n concurrency level per process. Only used with helpers
603 capable of processing more than one query at a time.
604 cache=n limit the result cache size, default is unbounded.
605 grace=n Percentage remaining of TTL where a refresh of a
606 cached entry should be initiated without needing to
607 wait for a new reply. (default is for no grace period)
608 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
609 ipv4 / ipv6 IP protocol used to communicate with this helper.
610 The default is to auto-detect IPv6 and use it when available.
612 FORMAT specifications
614 %LOGIN Authenticated user login name
615 %EXT_USER Username from previous external acl
616 %EXT_LOG Log details from previous external acl
617 %EXT_TAG Tag from previous external acl
618 %IDENT Ident user name
620 %SRCPORT Client source port
623 %PROTO Requested protocol
625 %PATH Requested URL path
626 %METHOD Request method
627 %MYADDR Squid interface address
628 %MYPORT Squid http_port number
629 %PATH Requested URL-path (including query-string if any)
630 %USER_CERT SSL User certificate in PEM format
631 %USER_CERTCHAIN SSL User certificate chain in PEM format
632 %USER_CERT_xx SSL User certificate subject attribute xx
633 %USER_CA_xx SSL User certificate issuer attribute xx
635 %>{Header} HTTP request header "Header"
637 HTTP request header "Hdr" list member "member"
639 HTTP request header list member using ; as
640 list separator. ; can be any non-alphanumeric
643 %<{Header} HTTP reply header "Header"
645 HTTP reply header "Hdr" list member "member"
647 HTTP reply header list member using ; as
648 list separator. ; can be any non-alphanumeric
651 %% The percent sign. Useful for helpers which need
652 an unchanging input format.
655 General request syntax:
657 [channel-ID] FORMAT-values [acl-values ...]
660 FORMAT-values consists of transaction details expanded with
661 whitespace separation per the config file FORMAT specification
662 using the FORMAT macros listed above.
664 acl-values consists of any string specified in the referencing
665 config 'acl ... external' line. see the "acl external" directive.
667 Request values sent to the helper are URL escaped to protect
668 each value in requests against whitespaces.
670 If using protocol=2.5 then the request sent to the helper is not
671 URL escaped to protect against whitespace.
673 NOTE: protocol=3.0 is deprecated as no longer necessary.
675 When using the concurrency= option the protocol is changed by
676 introducing a query channel tag in front of the request/response.
677 The query channel tag is a number between 0 and concurrency-1.
678 This value must be echoed back unchanged to Squid as the first part
679 of the response relating to its request.
682 The helper receives lines expanded per the above format specification
683 and for each input line returns 1 line starting with OK/ERR/BH result
684 code and optionally followed by additional keywords with more details.
687 General result syntax:
689 [channel-ID] result keyword=value ...
691 Result consists of one of the codes:
694 the ACL test produced a match.
697 the ACL test does not produce a match.
700 An internal error occurred in the helper, preventing
701 a result being identified.
703 The meaning of 'a match' is determined by your squid.conf
704 access control configuration. See the Squid wiki for details.
708 user= The users name (login)
710 password= The users password (for login= cache_peer option)
712 message= Message describing the reason for this response.
713 Available as %o in error pages.
714 Useful on (ERR and BH results).
716 tag= Apply a tag to a request. Only sets a tag once,
717 does not alter existing tags.
719 log= String to be logged in access.log. Available as
720 %ea in logformat specifications.
722 Any keywords may be sent on any response whether OK, ERR or BH.
724 All response keyword values need to be a single token with URL
725 escaping, or enclosed in double quotes (") and escaped using \ on
726 any double quotes or \ characters within the value. The wrapping
727 double quotes are removed before the value is interpreted by Squid.
728 \r and \n are also replace by CR and LF.
730 Some example key values:
734 user="J. \"Bob\" Smith"
741 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
742 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
743 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
744 DEFAULT: ssl::certUntrusted ssl_error X509_V_ERR_INVALID_CA X509_V_ERR_SELF_SIGNED_CERT_IN_CHAIN X509_V_ERR_UNABLE_TO_VERIFY_LEAF_SIGNATURE X509_V_ERR_UNABLE_TO_GET_ISSUER_CERT X509_V_ERR_UNABLE_TO_GET_ISSUER_CERT_LOCALLY X509_V_ERR_CERT_UNTRUSTED
745 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
748 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
749 DEFAULT: localhost src 127.0.0.1/32 ::1
750 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
751 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
753 Defining an Access List
755 Every access list definition must begin with an aclname and acltype,
756 followed by either type-specific arguments or a quoted filename that
759 acl aclname acltype argument ...
760 acl aclname acltype "file" ...
762 When using "file", the file should contain one item per line.
764 By default, regular expressions are CASE-SENSITIVE.
765 To make them case-insensitive, use the -i option. To return case-sensitive
766 use the +i option between patterns, or make a new ACL line without -i.
768 Some acl types require suspending the current request in order
769 to access some external data source.
770 Those which do are marked with the tag [slow], those which
771 don't are marked as [fast].
772 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
773 for further information
775 ***** ACL TYPES AVAILABLE *****
777 acl aclname src ip-address/mask ... # clients IP address [fast]
778 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
779 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
780 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
782 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
783 # The arp ACL requires the special configure option --enable-arp-acl.
784 # Furthermore, the ARP ACL code is not portable to all operating systems.
785 # It works on Linux, Solaris, Windows, FreeBSD, and some
786 # other *BSD variants.
789 # NOTE: Squid can only determine the MAC address for clients that are on
790 # the same subnet. If the client is on a different subnet,
791 # then Squid cannot find out its MAC address.
793 acl aclname srcdomain .foo.com ...
794 # reverse lookup, from client IP [slow]
795 acl aclname dstdomain .foo.com ...
796 # Destination server from URL [fast]
797 acl aclname srcdom_regex [-i] \.foo\.com ...
798 # regex matching client name [slow]
799 acl aclname dstdom_regex [-i] \.foo\.com ...
800 # regex matching server [fast]
802 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
803 # based URL is used and no match is found. The name "none" is used
804 # if the reverse lookup fails.
806 acl aclname src_as number ...
807 acl aclname dst_as number ...
809 # Except for access control, AS numbers can be used for
810 # routing of requests to specific caches. Here's an
811 # example for routing all requests for AS#1241 and only
812 # those to mycache.mydomain.net:
813 # acl asexample dst_as 1241
814 # cache_peer_access mycache.mydomain.net allow asexample
815 # cache_peer_access mycache_mydomain.net deny all
817 acl aclname peername myPeer ...
819 # match against a named cache_peer entry
820 # set unique name= on cache_peer lines for reliable use.
822 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
832 # h1:m1 must be less than h2:m2
834 acl aclname url_regex [-i] ^http:// ...
835 # regex matching on whole URL [fast]
836 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
837 # regex matching on URL login field
838 acl aclname urlpath_regex [-i] \.gif$ ...
839 # regex matching on URL path [fast]
841 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
843 acl aclname localport 3128 ... # TCP port the client connected to [fast]
844 # NP: for interception mode this is usually '80'
846 acl aclname myportname 3128 ... # http(s)_port name [fast]
848 acl aclname proto HTTP FTP ... # request protocol [fast]
850 acl aclname method GET POST ... # HTTP request method [fast]
852 acl aclname http_status 200 301 500- 400-403 ...
853 # status code in reply [fast]
855 acl aclname browser [-i] regexp ...
856 # pattern match on User-Agent header (see also req_header below) [fast]
858 acl aclname referer_regex [-i] regexp ...
859 # pattern match on Referer header [fast]
860 # Referer is highly unreliable, so use with care
862 acl aclname ident username ...
863 acl aclname ident_regex [-i] pattern ...
864 # string match on ident output [slow]
865 # use REQUIRED to accept any non-null ident.
867 acl aclname proxy_auth [-i] username ...
868 acl aclname proxy_auth_regex [-i] pattern ...
869 # perform http authentication challenge to the client and match against
870 # supplied credentials [slow]
872 # takes a list of allowed usernames.
873 # use REQUIRED to accept any valid username.
875 # Will use proxy authentication in forward-proxy scenarios, and plain
876 # http authenticaiton in reverse-proxy scenarios
878 # NOTE: when a Proxy-Authentication header is sent but it is not
879 # needed during ACL checking the username is NOT logged
882 # NOTE: proxy_auth requires a EXTERNAL authentication program
883 # to check username/password combinations (see
884 # auth_param directive).
886 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
887 # as the browser needs to be configured for using a proxy in order
888 # to respond to proxy authentication.
890 acl aclname snmp_community string ...
891 # A community string to limit access to your SNMP Agent [fast]
894 # acl snmppublic snmp_community public
896 acl aclname maxconn number
897 # This will be matched when the client's IP address has
898 # more than <number> TCP connections established. [fast]
899 # NOTE: This only measures direct TCP links so X-Forwarded-For
900 # indirect clients are not counted.
902 acl aclname max_user_ip [-s] number
903 # This will be matched when the user attempts to log in from more
904 # than <number> different ip addresses. The authenticate_ip_ttl
905 # parameter controls the timeout on the ip entries. [fast]
906 # If -s is specified the limit is strict, denying browsing
907 # from any further IP addresses until the ttl has expired. Without
908 # -s Squid will just annoy the user by "randomly" denying requests.
909 # (the counter is reset each time the limit is reached and a
911 # NOTE: in acceleration mode or where there is mesh of child proxies,
912 # clients may appear to come from multiple addresses if they are
913 # going through proxy farms, so a limit of 1 may cause user problems.
915 acl aclname random probability
916 # Pseudo-randomly match requests. Based on the probability given.
917 # Probability may be written as a decimal (0.333), fraction (1/3)
918 # or ratio of matches:non-matches (3:5).
920 acl aclname req_mime_type [-i] mime-type ...
921 # regex match against the mime type of the request generated
922 # by the client. Can be used to detect file upload or some
923 # types HTTP tunneling requests [fast]
924 # NOTE: This does NOT match the reply. You cannot use this
925 # to match the returned file type.
927 acl aclname req_header header-name [-i] any\.regex\.here
928 # regex match against any of the known request headers. May be
929 # thought of as a superset of "browser", "referer" and "mime-type"
932 acl aclname rep_mime_type [-i] mime-type ...
933 # regex match against the mime type of the reply received by
934 # squid. Can be used to detect file download or some
935 # types HTTP tunneling requests. [fast]
936 # NOTE: This has no effect in http_access rules. It only has
937 # effect in rules that affect the reply data stream such as
940 acl aclname rep_header header-name [-i] any\.regex\.here
941 # regex match against any of the known reply headers. May be
942 # thought of as a superset of "browser", "referer" and "mime-type"
945 acl aclname external class_name [arguments...]
946 # external ACL lookup via a helper class defined by the
947 # external_acl_type directive [slow]
949 acl aclname user_cert attribute values...
950 # match against attributes in a user SSL certificate
951 # attribute is one of DN/C/O/CN/L/ST [fast]
953 acl aclname ca_cert attribute values...
954 # match against attributes a users issuing CA SSL certificate
955 # attribute is one of DN/C/O/CN/L/ST [fast]
957 acl aclname ext_user username ...
958 acl aclname ext_user_regex [-i] pattern ...
959 # string match on username returned by external acl helper [slow]
960 # use REQUIRED to accept any non-null user name.
962 acl aclname tag tagvalue ...
963 # string match on tag returned by external acl helper [slow]
965 acl aclname hier_code codename ...
966 # string match against squid hierarchy code(s); [fast]
967 # e.g., DIRECT, PARENT_HIT, NONE, etc.
969 # NOTE: This has no effect in http_access rules. It only has
970 # effect in rules that affect the reply data stream such as
974 acl aclname ssl_error errorname
975 # match against SSL certificate validation error [fast]
977 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
980 # The following can be used as shortcuts for certificate properties:
981 # [ssl::]certHasExpired: the "not after" field is in the past
982 # [ssl::]certNotYetValid: the "not before" field is in the future
983 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
984 # [ssl::]certSelfSigned: The certificate is self signed.
985 # [ssl::]certDomainMismatch: The certificate CN domain does not
986 # match the name the name of the host we are connecting to.
988 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
989 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
990 # predefined ACLs, just like the 'all' ACL.
992 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
993 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
995 acl aclname server_cert_fingerprint [-sha1] fingerprint
996 # match against server SSL certificate fingerprint [fast]
998 # The fingerprint is the digest of the DER encoded version
999 # of the whole certificate. The user should use the form: XX:XX:...
1000 # Optional argument specifies the digest algorithm to use.
1001 # The SHA1 digest algorithm is the default and is currently
1002 # the only algorithm supported (-sha1).
1006 acl macaddress arp 09:00:2b:23:45:67
1007 acl myexample dst_as 1241
1008 acl password proxy_auth REQUIRED
1009 acl fileupload req_mime_type -i ^multipart/form-data$
1010 acl javascript rep_mime_type -i ^application/x-javascript$
1014 # Recommended minimum configuration:
1017 # Example rule allowing access from your local networks.
1018 # Adapt to list your (internal) IP networks from where browsing
1020 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1021 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1022 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1023 acl localnet src fc00::/7 # RFC 4193 local private network range
1024 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1026 acl SSL_ports port 443
1027 acl Safe_ports port 80 # http
1028 acl Safe_ports port 21 # ftp
1029 acl Safe_ports port 443 # https
1030 acl Safe_ports port 70 # gopher
1031 acl Safe_ports port 210 # wais
1032 acl Safe_ports port 1025-65535 # unregistered ports
1033 acl Safe_ports port 280 # http-mgmt
1034 acl Safe_ports port 488 # gss-http
1035 acl Safe_ports port 591 # filemaker
1036 acl Safe_ports port 777 # multiling http
1037 acl CONNECT method CONNECT
1041 NAME: follow_x_forwarded_for
1043 IFDEF: FOLLOW_X_FORWARDED_FOR
1044 LOC: Config.accessList.followXFF
1045 DEFAULT_IF_NONE: deny all
1047 Allowing or Denying the X-Forwarded-For header to be followed to
1048 find the original source of a request.
1050 Requests may pass through a chain of several other proxies
1051 before reaching us. The X-Forwarded-For header will contain a
1052 comma-separated list of the IP addresses in the chain, with the
1053 rightmost address being the most recent.
1055 If a request reaches us from a source that is allowed by this
1056 configuration item, then we consult the X-Forwarded-For header
1057 to see where that host received the request from. If the
1058 X-Forwarded-For header contains multiple addresses, we continue
1059 backtracking until we reach an address for which we are not allowed
1060 to follow the X-Forwarded-For header, or until we reach the first
1061 address in the list. For the purpose of ACL used in the
1062 follow_x_forwarded_for directive the src ACL type always matches
1063 the address we are testing and srcdomain matches its rDNS.
1065 The end result of this process is an IP address that we will
1066 refer to as the indirect client address. This address may
1067 be treated as the client address for access control, ICAP, delay
1068 pools and logging, depending on the acl_uses_indirect_client,
1069 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1070 log_uses_indirect_client and tproxy_uses_indirect_client options.
1072 This clause only supports fast acl types.
1073 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1075 SECURITY CONSIDERATIONS:
1077 Any host for which we follow the X-Forwarded-For header
1078 can place incorrect information in the header, and Squid
1079 will use the incorrect information as if it were the
1080 source address of the request. This may enable remote
1081 hosts to bypass any access control restrictions that are
1082 based on the client's source addresses.
1086 acl localhost src 127.0.0.1
1087 acl my_other_proxy srcdomain .proxy.example.com
1088 follow_x_forwarded_for allow localhost
1089 follow_x_forwarded_for allow my_other_proxy
1092 NAME: acl_uses_indirect_client
1095 IFDEF: FOLLOW_X_FORWARDED_FOR
1097 LOC: Config.onoff.acl_uses_indirect_client
1099 Controls whether the indirect client address
1100 (see follow_x_forwarded_for) is used instead of the
1101 direct client address in acl matching.
1103 NOTE: maxconn ACL considers direct TCP links and indirect
1104 clients will always have zero. So no match.
1107 NAME: delay_pool_uses_indirect_client
1110 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1112 LOC: Config.onoff.delay_pool_uses_indirect_client
1114 Controls whether the indirect client address
1115 (see follow_x_forwarded_for) is used instead of the
1116 direct client address in delay pools.
1119 NAME: log_uses_indirect_client
1122 IFDEF: FOLLOW_X_FORWARDED_FOR
1124 LOC: Config.onoff.log_uses_indirect_client
1126 Controls whether the indirect client address
1127 (see follow_x_forwarded_for) is used instead of the
1128 direct client address in the access log.
1131 NAME: tproxy_uses_indirect_client
1134 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1136 LOC: Config.onoff.tproxy_uses_indirect_client
1138 Controls whether the indirect client address
1139 (see follow_x_forwarded_for) is used instead of the
1140 direct client address when spoofing the outgoing client.
1142 This has no effect on requests arriving in non-tproxy
1145 SECURITY WARNING: Usage of this option is dangerous
1146 and should not be used trivially. Correct configuration
1147 of follow_x_forewarded_for with a limited set of trusted
1148 sources is required to prevent abuse of your proxy.
1153 LOC: Config.accessList.http
1154 DEFAULT_IF_NONE: deny all
1156 Allowing or Denying access based on defined access lists
1158 Access to the HTTP port:
1159 http_access allow|deny [!]aclname ...
1161 NOTE on default values:
1163 If there are no "access" lines present, the default is to deny
1166 If none of the "access" lines cause a match, the default is the
1167 opposite of the last line in the list. If the last line was
1168 deny, the default is allow. Conversely, if the last line
1169 is allow, the default will be deny. For these reasons, it is a
1170 good idea to have an "deny all" entry at the end of your access
1171 lists to avoid potential confusion.
1173 This clause supports both fast and slow acl types.
1174 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1179 # Recommended minimum Access Permission configuration:
1181 # Only allow cachemgr access from localhost
1182 http_access allow localhost manager
1183 http_access deny manager
1185 # Deny requests to certain unsafe ports
1186 http_access deny !Safe_ports
1188 # Deny CONNECT to other than secure SSL ports
1189 http_access deny CONNECT !SSL_ports
1191 # We strongly recommend the following be uncommented to protect innocent
1192 # web applications running on the proxy server who think the only
1193 # one who can access services on "localhost" is a local user
1194 #http_access deny to_localhost
1197 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1200 # Example rule allowing access from your local networks.
1201 # Adapt localnet in the ACL section to list your (internal) IP networks
1202 # from where browsing should be allowed
1203 http_access allow localnet
1204 http_access allow localhost
1206 # And finally deny all other access to this proxy
1207 http_access deny all
1211 NAME: adapted_http_access http_access2
1213 LOC: Config.accessList.adapted_http
1216 Allowing or Denying access based on defined access lists
1218 Essentially identical to http_access, but runs after redirectors
1219 and ICAP/eCAP adaptation. Allowing access control based on their
1222 If not set then only http_access is used.
1225 NAME: http_reply_access
1227 LOC: Config.accessList.reply
1230 Allow replies to client requests. This is complementary to http_access.
1232 http_reply_access allow|deny [!] aclname ...
1234 NOTE: if there are no access lines present, the default is to allow
1237 If none of the access lines cause a match the opposite of the
1238 last line will apply. Thus it is good practice to end the rules
1239 with an "allow all" or "deny all" entry.
1241 This clause supports both fast and slow acl types.
1242 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1247 LOC: Config.accessList.icp
1248 DEFAULT_IF_NONE: deny all
1250 Allowing or Denying access to the ICP port based on defined
1253 icp_access allow|deny [!]aclname ...
1255 See http_access for details
1257 This clause only supports fast acl types.
1258 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1260 # Allow ICP queries from local networks only
1261 #icp_access allow localnet
1262 #icp_access deny all
1268 LOC: Config.accessList.htcp
1269 DEFAULT_IF_NONE: deny all
1271 Allowing or Denying access to the HTCP port based on defined
1274 htcp_access allow|deny [!]aclname ...
1276 See http_access for details
1278 NOTE: The default if no htcp_access lines are present is to
1279 deny all traffic. This default may cause problems with peers
1280 using the htcp option.
1282 This clause only supports fast acl types.
1283 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1285 # Allow HTCP queries from local networks only
1286 #htcp_access allow localnet
1287 #htcp_access deny all
1290 NAME: htcp_clr_access
1293 LOC: Config.accessList.htcp_clr
1294 DEFAULT_IF_NONE: deny all
1296 Allowing or Denying access to purge content using HTCP based
1297 on defined access lists
1299 htcp_clr_access allow|deny [!]aclname ...
1301 See http_access for details
1303 This clause only supports fast acl types.
1304 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1306 # Allow HTCP CLR requests from trusted peers
1307 acl htcp_clr_peer src 172.16.1.2
1308 htcp_clr_access allow htcp_clr_peer
1313 LOC: Config.accessList.miss
1316 Determins whether network access is permitted when satisfying a request.
1319 to force your neighbors to use you as a sibling instead of
1322 acl localclients src 172.16.0.0/16
1323 miss_access allow localclients
1324 miss_access deny !localclients
1326 This means only your local clients are allowed to fetch relayed/MISS
1327 replies from the network and all other clients can only fetch cached
1331 The default for this setting allows all clients who passed the
1332 http_access rules to relay via this proxy.
1334 This clause only supports fast acl types.
1335 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1338 NAME: ident_lookup_access
1341 DEFAULT_IF_NONE: deny all
1342 LOC: Ident::TheConfig.identLookup
1344 A list of ACL elements which, if matched, cause an ident
1345 (RFC 931) lookup to be performed for this request. For
1346 example, you might choose to always perform ident lookups
1347 for your main multi-user Unix boxes, but not for your Macs
1348 and PCs. By default, ident lookups are not performed for
1351 To enable ident lookups for specific client addresses, you
1352 can follow this example:
1354 acl ident_aware_hosts src 198.168.1.0/24
1355 ident_lookup_access allow ident_aware_hosts
1356 ident_lookup_access deny all
1358 Only src type ACL checks are fully supported. A srcdomain
1359 ACL might work at times, but it will not always provide
1362 This clause only supports fast acl types.
1363 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1366 NAME: reply_body_max_size
1367 COMMENT: size [acl acl...]
1370 LOC: Config.ReplyBodySize
1372 This option specifies the maximum size of a reply body. It can be
1373 used to prevent users from downloading very large files, such as
1374 MP3's and movies. When the reply headers are received, the
1375 reply_body_max_size lines are processed, and the first line where
1376 all (if any) listed ACLs are true is used as the maximum body size
1379 This size is checked twice. First when we get the reply headers,
1380 we check the content-length value. If the content length value exists
1381 and is larger than the allowed size, the request is denied and the
1382 user receives an error message that says "the request or reply
1383 is too large." If there is no content-length, and the reply
1384 size exceeds this limit, the client's connection is just closed
1385 and they will receive a partial reply.
1387 WARNING: downstream caches probably can not detect a partial reply
1388 if there is no content-length header, so they will cache
1389 partial responses and give them out as hits. You should NOT
1390 use this option if you have downstream caches.
1392 WARNING: A maximum size smaller than the size of squid's error messages
1393 will cause an infinite loop and crash squid. Ensure that the smallest
1394 non-zero value you use is greater that the maximum header size plus
1395 the size of your largest error page.
1397 If you set this parameter none (the default), there will be
1400 Configuration Format is:
1401 reply_body_max_size SIZE UNITS [acl ...]
1403 reply_body_max_size 10 MB
1409 -----------------------------------------------------------------------------
1412 NAME: http_port ascii_port
1415 LOC: Config.Sockaddr.http
1417 Usage: port [mode] [options]
1418 hostname:port [mode] [options]
1419 1.2.3.4:port [mode] [options]
1421 The socket addresses where Squid will listen for HTTP client
1422 requests. You may specify multiple socket addresses.
1423 There are three forms: port alone, hostname with port, and
1424 IP address with port. If you specify a hostname or IP
1425 address, Squid binds the socket to that specific
1426 address. Most likely, you do not need to bind to a specific
1427 address, so you can use the port number alone.
1429 If you are running Squid in accelerator mode, you
1430 probably want to listen on port 80 also, or instead.
1432 The -a command line option may be used to specify additional
1433 port(s) where Squid listens for proxy request. Such ports will
1434 be plain proxy ports with no options.
1436 You may specify multiple socket addresses on multiple lines.
1440 intercept Support for IP-Layer interception of
1441 outgoing requests without browser settings.
1442 NP: disables authentication and IPv6 on the port.
1444 tproxy Support Linux TPROXY for spoofing outgoing
1445 connections using the client IP address.
1446 NP: disables authentication and maybe IPv6 on the port.
1448 accel Accelerator / reverse proxy mode
1450 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1451 establish secure connection with the client and with
1452 the server, decrypt HTTPS messages as they pass through
1453 Squid, and treat them as unencrypted HTTP messages,
1454 becoming the man-in-the-middle.
1456 The ssl_bump option is required to fully enable
1457 bumping of CONNECT requests.
1459 Omitting the mode flag causes default forward proxy mode to be used.
1462 Accelerator Mode Options:
1464 defaultsite=domainname
1465 What to use for the Host: header if it is not present
1466 in a request. Determines what site (not origin server)
1467 accelerators should consider the default.
1469 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1471 protocol= Protocol to reconstruct accelerated requests with.
1472 Defaults to http for http_port and https for
1475 vport Virtual host port support. Using the http_port number
1476 instead of the port passed on Host: headers.
1478 vport=NN Virtual host port support. Using the specified port
1479 number instead of the port passed on Host: headers.
1482 Act as if this Squid is the origin server.
1483 This currently means generate new Date: and Expires:
1484 headers on HIT instead of adding Age:.
1486 ignore-cc Ignore request Cache-Control headers.
1488 WARNING: This option violates HTTP specifications if
1489 used in non-accelerator setups.
1491 allow-direct Allow direct forwarding in accelerator mode. Normally
1492 accelerated requests are denied direct forwarding as if
1493 never_direct was used.
1495 WARNING: this option opens accelerator mode to security
1496 vulnerabilities usually only affecting in interception
1497 mode. Make sure to protect forwarding with suitable
1498 http_access rules when using this.
1501 SSL Bump Mode Options:
1502 In addition to these options ssl-bump requires TLS/SSL options.
1504 generate-host-certificates[=<on|off>]
1505 Dynamically create SSL server certificates for the
1506 destination hosts of bumped CONNECT requests.When
1507 enabled, the cert and key options are used to sign
1508 generated certificates. Otherwise generated
1509 certificate will be selfsigned.
1510 If there is a CA certificate lifetime of the generated
1511 certificate equals lifetime of the CA certificate. If
1512 generated certificate is selfsigned lifetime is three
1514 This option is enabled by default when ssl-bump is used.
1515 See the ssl-bump option above for more information.
1517 dynamic_cert_mem_cache_size=SIZE
1518 Approximate total RAM size spent on cached generated
1519 certificates. If set to zero, caching is disabled. The
1520 default value is 4MB.
1524 cert= Path to SSL certificate (PEM format).
1526 key= Path to SSL private key file (PEM format)
1527 if not specified, the certificate file is
1528 assumed to be a combined certificate and
1531 version= The version of SSL/TLS supported
1532 1 automatic (default)
1539 cipher= Colon separated list of supported ciphers.
1540 NOTE: some ciphers such as EDH ciphers depend on
1541 additional settings. If those settings are
1542 omitted the ciphers may be silently ignored
1543 by the OpenSSL library.
1545 options= Various SSL implementation options. The most important
1547 NO_SSLv2 Disallow the use of SSLv2
1548 NO_SSLv3 Disallow the use of SSLv3
1549 NO_TLSv1 Disallow the use of TLSv1.0
1550 NO_TLSv1_1 Disallow the use of TLSv1.1
1551 NO_TLSv1_2 Disallow the use of TLSv1.2
1552 SINGLE_DH_USE Always create a new key when using
1553 temporary/ephemeral DH key exchanges
1554 ALL Enable various bug workarounds
1555 suggested as "harmless" by OpenSSL
1556 Be warned that this reduces SSL/TLS
1557 strength to some attacks.
1558 See OpenSSL SSL_CTX_set_options documentation for a
1559 complete list of options.
1561 clientca= File containing the list of CAs to use when
1562 requesting a client certificate.
1564 cafile= File containing additional CA certificates to
1565 use when verifying client certificates. If unset
1566 clientca will be used.
1568 capath= Directory containing additional CA certificates
1569 and CRL lists to use when verifying client certificates.
1571 crlfile= File of additional CRL lists to use when verifying
1572 the client certificate, in addition to CRLs stored in
1573 the capath. Implies VERIFY_CRL flag below.
1575 dhparams= File containing DH parameters for temporary/ephemeral
1576 DH key exchanges. See OpenSSL documentation for details
1577 on how to create this file.
1578 WARNING: EDH ciphers will be silently disabled if this
1581 sslflags= Various flags modifying the use of SSL:
1583 Don't request client certificates
1584 immediately, but wait until acl processing
1585 requires a certificate (not yet implemented).
1587 Don't use the default CA lists built in
1590 Don't allow for session reuse. Each connection
1591 will result in a new SSL session.
1593 Verify CRL lists when accepting client
1596 Verify CRL lists for all certificates in the
1597 client certificate chain.
1599 sslcontext= SSL session ID context identifier.
1603 connection-auth[=on|off]
1604 use connection-auth=off to tell Squid to prevent
1605 forwarding Microsoft connection oriented authentication
1606 (NTLM, Negotiate and Kerberos)
1608 disable-pmtu-discovery=
1609 Control Path-MTU discovery usage:
1610 off lets OS decide on what to do (default).
1611 transparent disable PMTU discovery when transparent
1613 always disable always PMTU discovery.
1615 In many setups of transparently intercepting proxies
1616 Path-MTU discovery can not work on traffic towards the
1617 clients. This is the case when the intercepting device
1618 does not fully track connections and fails to forward
1619 ICMP must fragment messages to the cache server. If you
1620 have such setup and experience that certain clients
1621 sporadically hang or never complete requests set
1622 disable-pmtu-discovery option to 'transparent'.
1624 name= Specifies a internal name for the port. Defaults to
1625 the port specification (port or addr:port)
1627 tcpkeepalive[=idle,interval,timeout]
1628 Enable TCP keepalive probes of idle connections.
1629 In seconds; idle is the initial time before TCP starts
1630 probing the connection, interval how often to probe, and
1631 timeout the time before giving up.
1633 If you run Squid on a dual-homed machine with an internal
1634 and an external interface we recommend you to specify the
1635 internal address:port in http_port. This way Squid will only be
1636 visible on the internal address.
1640 # Squid normally listens to port 3128
1641 http_port @DEFAULT_HTTP_PORT@
1649 LOC: Config.Sockaddr.https
1651 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1653 The socket address where Squid will listen for client requests made
1654 over TLS or SSL connections. Commonly referred to as HTTPS.
1656 This is most useful for situations where you are running squid in
1657 accelerator mode and you want to do the SSL work at the accelerator level.
1659 You may specify multiple socket addresses on multiple lines,
1660 each with their own SSL certificate and/or options.
1664 accel Accelerator / reverse proxy mode
1666 intercept Support for IP-Layer interception of
1667 outgoing requests without browser settings.
1668 NP: disables authentication and IPv6 on the port.
1670 tproxy Support Linux TPROXY for spoofing outgoing
1671 connections using the client IP address.
1672 NP: disables authentication and maybe IPv6 on the port.
1674 ssl-bump For each intercepted connection allowed by ssl_bump
1675 ACLs, establish a secure connection with the client and with
1676 the server, decrypt HTTPS messages as they pass through
1677 Squid, and treat them as unencrypted HTTP messages,
1678 becoming the man-in-the-middle.
1680 An "ssl_bump server-first" match is required to
1681 fully enable bumping of intercepted SSL connections.
1683 Requires tproxy or intercept.
1685 Omitting the mode flag causes default forward proxy mode to be used.
1688 See http_port for a list of generic options
1693 cert= Path to SSL certificate (PEM format).
1695 key= Path to SSL private key file (PEM format)
1696 if not specified, the certificate file is
1697 assumed to be a combined certificate and
1700 version= The version of SSL/TLS supported
1701 1 automatic (default)
1706 cipher= Colon separated list of supported ciphers.
1708 options= Various SSL engine options. The most important
1710 NO_SSLv2 Disallow the use of SSLv2
1711 NO_SSLv3 Disallow the use of SSLv3
1712 NO_TLSv1 Disallow the use of TLSv1
1713 SINGLE_DH_USE Always create a new key when using
1714 temporary/ephemeral DH key exchanges
1715 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1716 documentation for a complete list of options.
1718 clientca= File containing the list of CAs to use when
1719 requesting a client certificate.
1721 cafile= File containing additional CA certificates to
1722 use when verifying client certificates. If unset
1723 clientca will be used.
1725 capath= Directory containing additional CA certificates
1726 and CRL lists to use when verifying client certificates.
1728 crlfile= File of additional CRL lists to use when verifying
1729 the client certificate, in addition to CRLs stored in
1730 the capath. Implies VERIFY_CRL flag below.
1732 dhparams= File containing DH parameters for temporary/ephemeral
1735 sslflags= Various flags modifying the use of SSL:
1737 Don't request client certificates
1738 immediately, but wait until acl processing
1739 requires a certificate (not yet implemented).
1741 Don't use the default CA lists built in
1744 Don't allow for session reuse. Each connection
1745 will result in a new SSL session.
1747 Verify CRL lists when accepting client
1750 Verify CRL lists for all certificates in the
1751 client certificate chain.
1753 sslcontext= SSL session ID context identifier.
1755 generate-host-certificates[=<on|off>]
1756 Dynamically create SSL server certificates for the
1757 destination hosts of bumped SSL requests.When
1758 enabled, the cert and key options are used to sign
1759 generated certificates. Otherwise generated
1760 certificate will be selfsigned.
1761 If there is CA certificate life time of generated
1762 certificate equals lifetime of CA certificate. If
1763 generated certificate is selfsigned lifetime is three
1765 This option is enabled by default when SslBump is used.
1766 See the sslBump option above for more information.
1768 dynamic_cert_mem_cache_size=SIZE
1769 Approximate total RAM size spent on cached generated
1770 certificates. If set to zero, caching is disabled. The
1771 default value is 4MB.
1773 See http_port for a list of available options.
1776 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1779 LOC: Ip::Qos::TheConfig.tosToServer
1781 Allows you to select a TOS/Diffserv value for packets outgoing
1782 on the server side, based on an ACL.
1784 tcp_outgoing_tos ds-field [!]aclname ...
1786 Example where normal_service_net uses the TOS value 0x00
1787 and good_service_net uses 0x20
1789 acl normal_service_net src 10.0.0.0/24
1790 acl good_service_net src 10.0.1.0/24
1791 tcp_outgoing_tos 0x00 normal_service_net
1792 tcp_outgoing_tos 0x20 good_service_net
1794 TOS/DSCP values really only have local significance - so you should
1795 know what you're specifying. For more information, see RFC2474,
1796 RFC2475, and RFC3260.
1798 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1799 "default" to use whatever default your host has. Note that in
1800 practice often only multiples of 4 is usable as the two rightmost bits
1801 have been redefined for use by ECN (RFC 3168 section 23.1).
1803 Processing proceeds in the order specified, and stops at first fully
1807 NAME: clientside_tos
1810 LOC: Ip::Qos::TheConfig.tosToClient
1812 Allows you to select a TOS/Diffserv value for packets being transmitted
1813 on the client-side, based on an ACL.
1815 clientside_tos ds-field [!]aclname ...
1817 Example where normal_service_net uses the TOS value 0x00
1818 and good_service_net uses 0x20
1820 acl normal_service_net src 10.0.0.0/24
1821 acl good_service_net src 10.0.1.0/24
1822 clientside_tos 0x00 normal_service_net
1823 clientside_tos 0x20 good_service_net
1825 Note: This feature is incompatible with qos_flows. Any TOS values set here
1826 will be overwritten by TOS values in qos_flows.
1829 NAME: tcp_outgoing_mark
1831 IFDEF: SO_MARK&&USE_LIBCAP
1833 LOC: Ip::Qos::TheConfig.nfmarkToServer
1835 Allows you to apply a Netfilter mark value to outgoing packets
1836 on the server side, based on an ACL.
1838 tcp_outgoing_mark mark-value [!]aclname ...
1840 Example where normal_service_net uses the mark value 0x00
1841 and good_service_net uses 0x20
1843 acl normal_service_net src 10.0.0.0/24
1844 acl good_service_net src 10.0.1.0/24
1845 tcp_outgoing_mark 0x00 normal_service_net
1846 tcp_outgoing_mark 0x20 good_service_net
1849 NAME: clientside_mark
1851 IFDEF: SO_MARK&&USE_LIBCAP
1853 LOC: Ip::Qos::TheConfig.nfmarkToClient
1855 Allows you to apply a Netfilter mark value to packets being transmitted
1856 on the client-side, based on an ACL.
1858 clientside_mark mark-value [!]aclname ...
1860 Example where normal_service_net uses the mark value 0x00
1861 and good_service_net uses 0x20
1863 acl normal_service_net src 10.0.0.0/24
1864 acl good_service_net src 10.0.1.0/24
1865 clientside_mark 0x00 normal_service_net
1866 clientside_mark 0x20 good_service_net
1868 Note: This feature is incompatible with qos_flows. Any mark values set here
1869 will be overwritten by mark values in qos_flows.
1876 LOC: Ip::Qos::TheConfig
1878 Allows you to select a TOS/DSCP value to mark outgoing
1879 connections with, based on where the reply was sourced. For
1880 platforms using netfilter, allows you to set a netfilter mark
1881 value instead of, or in addition to, a TOS value.
1883 TOS values really only have local significance - so you should
1884 know what you're specifying. For more information, see RFC2474,
1885 RFC2475, and RFC3260.
1887 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1888 in practice often only multiples of 4 is usable as the two rightmost bits
1889 have been redefined for use by ECN (RFC 3168 section 23.1).
1891 Mark values can be any unsigned 32-bit integer value.
1893 This setting is configured by setting the following values:
1895 tos|mark Whether to set TOS or netfilter mark values
1897 local-hit=0xFF Value to mark local cache hits.
1899 sibling-hit=0xFF Value to mark hits from sibling peers.
1901 parent-hit=0xFF Value to mark hits from parent peers.
1903 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1904 over the preserve-miss feature (see below), unless
1905 mask is specified, in which case only the bits
1906 specified in the mask are written.
1908 The TOS variant of the following features are only possible on Linux
1909 and require your kernel to be patched with the TOS preserving ZPH
1910 patch, available from http://zph.bratcheda.org
1911 No patch is needed to preserve the netfilter mark, which will work
1912 with all variants of netfilter.
1914 disable-preserve-miss
1915 This option disables the preservation of the TOS or netfilter
1916 mark. By default, the existing TOS or netfilter mark value of
1917 the response coming from the remote server will be retained
1918 and masked with miss-mark.
1919 NOTE: in the case of a netfilter mark, the mark must be set on
1920 the connection (using the CONNMARK target) not on the packet
1924 Allows you to mask certain bits in the TOS or mark value
1925 received from the remote server, before copying the value to
1926 the TOS sent towards clients.
1927 Default for tos: 0xFF (TOS from server is not changed).
1928 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1930 All of these features require the --enable-zph-qos compilation flag
1931 (enabled by default). Netfilter marking also requires the
1932 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1933 libcap 2.09+ (--with-libcap).
1937 NAME: tcp_outgoing_address
1940 LOC: Config.accessList.outgoing_address
1942 Allows you to map requests to different outgoing IP addresses
1943 based on the username or source address of the user making
1946 tcp_outgoing_address ipaddr [[!]aclname] ...
1949 Forwarding clients with dedicated IPs for certain subnets.
1951 acl normal_service_net src 10.0.0.0/24
1952 acl good_service_net src 10.0.2.0/24
1954 tcp_outgoing_address 2001:db8::c001 good_service_net
1955 tcp_outgoing_address 10.1.0.2 good_service_net
1957 tcp_outgoing_address 2001:db8::beef normal_service_net
1958 tcp_outgoing_address 10.1.0.1 normal_service_net
1960 tcp_outgoing_address 2001:db8::1
1961 tcp_outgoing_address 10.1.0.3
1963 Processing proceeds in the order specified, and stops at first fully
1966 Squid will add an implicit IP version test to each line.
1967 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1968 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1971 NOTE: The use of this directive using client dependent ACLs is
1972 incompatible with the use of server side persistent connections. To
1973 ensure correct results it is best to set server_persistent_connections
1974 to off when using this directive in such configurations.
1976 NOTE: The use of this directive to set a local IP on outgoing TCP links
1977 is incompatible with using TPROXY to set client IP out outbound TCP links.
1978 When needing to contact peers use the no-tproxy cache_peer option and the
1979 client_dst_passthru directive re-enable normal forwarding such as this.
1983 NAME: host_verify_strict
1986 LOC: Config.onoff.hostStrictVerify
1988 Regardless of this option setting, when dealing with intercepted
1989 traffic, Squid always verifies that the destination IP address matches
1990 the Host header domain or IP (called 'authority form URL').
1992 This enforcement is performed to satisfy a MUST-level requirement in
1993 RFC 2616 section 14.23: "The Host field value MUST represent the naming
1994 authority of the origin server or gateway given by the original URL".
1997 Squid always responds with an HTTP 409 (Conflict) error
1998 page and logs a security warning if there is no match.
2000 Squid verifies that the destination IP address matches
2001 the Host header for forward-proxy and reverse-proxy traffic
2002 as well. For those traffic types, Squid also enables the
2003 following checks, comparing the corresponding Host header
2004 and Request-URI components:
2006 * The host names (domain or IP) must be identical,
2007 but valueless or missing Host header disables all checks.
2008 For the two host names to match, both must be either IP
2011 * Port numbers must be identical, but if a port is missing
2012 the scheme-default port is assumed.
2015 When set to OFF (the default):
2016 Squid allows suspicious requests to continue but logs a
2017 security warning and blocks caching of the response.
2019 * Forward-proxy traffic is not checked at all.
2021 * Reverse-proxy traffic is not checked at all.
2023 * Intercepted traffic which passes verification is handled
2024 according to client_dst_passthru.
2026 * Intercepted requests which fail verification are sent
2027 to the client original destination instead of DIRECT.
2028 This overrides 'client_dst_passthru off'.
2030 For now suspicious intercepted CONNECT requests are always
2031 responded to with an HTTP 409 (Conflict) error page.
2036 As described in CVE-2009-0801 when the Host: header alone is used
2037 to determine the destination of a request it becomes trivial for
2038 malicious scripts on remote websites to bypass browser same-origin
2039 security policy and sandboxing protections.
2041 The cause of this is that such applets are allowed to perform their
2042 own HTTP stack, in which case the same-origin policy of the browser
2043 sandbox only verifies that the applet tries to contact the same IP
2044 as from where it was loaded at the IP level. The Host: header may
2045 be different from the connected IP and approved origin.
2049 NAME: client_dst_passthru
2052 LOC: Config.onoff.client_dst_passthru
2054 With NAT or TPROXY intercepted traffic Squid may pass the request
2055 directly to the original client destination IP or seek a faster
2056 source using the HTTP Host header.
2058 Using Host to locate alternative servers can provide faster
2059 connectivity with a range of failure recovery options.
2060 But can also lead to connectivity trouble when the client and
2061 server are attempting stateful interactions unaware of the proxy.
2063 This option (on by default) prevents alternative DNS entries being
2064 located to send intercepted traffic DIRECT to an origin server.
2065 The clients original destination IP and port will be used instead.
2067 Regardless of this option setting, when dealing with intercepted
2068 traffic Squid will verify the Host: header and any traffic which
2069 fails Host verification will be treated as if this option were ON.
2071 see host_verify_strict for details on the verification process.
2076 -----------------------------------------------------------------------------
2079 NAME: ssl_unclean_shutdown
2083 LOC: Config.SSL.unclean_shutdown
2085 Some browsers (especially MSIE) bugs out on SSL shutdown
2092 LOC: Config.SSL.ssl_engine
2095 The OpenSSL engine to use. You will need to set this if you
2096 would like to use hardware SSL acceleration for example.
2099 NAME: sslproxy_client_certificate
2102 LOC: Config.ssl_client.cert
2105 Client SSL Certificate to use when proxying https:// URLs
2108 NAME: sslproxy_client_key
2111 LOC: Config.ssl_client.key
2114 Client SSL Key to use when proxying https:// URLs
2117 NAME: sslproxy_version
2120 LOC: Config.ssl_client.version
2123 SSL version level to use when proxying https:// URLs
2125 The versions of SSL/TLS supported:
2127 1 automatic (default)
2135 NAME: sslproxy_options
2138 LOC: Config.ssl_client.options
2141 SSL implementation options to use when proxying https:// URLs
2143 The most important being:
2145 NO_SSLv2 Disallow the use of SSLv2
2146 NO_SSLv3 Disallow the use of SSLv3
2147 NO_TLSv1 Disallow the use of TLSv1.0
2148 NO_TLSv1_1 Disallow the use of TLSv1.1
2149 NO_TLSv1_2 Disallow the use of TLSv1.2
2151 Always create a new key when using temporary/ephemeral
2154 Disable use of RFC5077 session tickets. Some servers
2155 may have problems understanding the TLS extension due
2156 to ambiguous specification in RFC4507.
2157 ALL Enable various bug workarounds suggested as "harmless"
2158 by OpenSSL. Be warned that this may reduce SSL/TLS
2159 strength to some attacks.
2161 See the OpenSSL SSL_CTX_set_options documentation for a
2162 complete list of possible options.
2165 NAME: sslproxy_cipher
2168 LOC: Config.ssl_client.cipher
2171 SSL cipher list to use when proxying https:// URLs
2173 Colon separated list of supported ciphers.
2176 NAME: sslproxy_cafile
2179 LOC: Config.ssl_client.cafile
2182 file containing CA certificates to use when verifying server
2183 certificates while proxying https:// URLs
2186 NAME: sslproxy_capath
2189 LOC: Config.ssl_client.capath
2192 directory containing CA certificates to use when verifying
2193 server certificates while proxying https:// URLs
2198 TYPE: sslproxy_ssl_bump
2199 LOC: Config.accessList.ssl_bump
2202 This option is consulted when a CONNECT request is received on
2203 an http_port (or a new connection is intercepted at an
2204 https_port), provided that port was configured with an ssl-bump
2205 flag. The subsequent data on the connection is either treated as
2206 HTTPS and decrypted OR tunneled at TCP level without decryption,
2207 depending on the first bumping "mode" which ACLs match.
2209 ssl_bump <mode> [!]acl ...
2211 The following bumping modes are supported:
2214 Allow bumping of the connection. Establish a secure connection
2215 with the client first, then connect to the server. This old mode
2216 does not allow Squid to mimic server SSL certificate and does
2217 not work with intercepted SSL connections.
2220 Allow bumping of the connection. Establish a secure connection
2221 with the server first, then establish a secure connection with
2222 the client, using a mimicked server certificate. Works with both
2223 CONNECT requests and intercepted SSL connections.
2226 Become a TCP tunnel without decoding the connection.
2227 Works with both CONNECT requests and intercepted SSL
2228 connections. This is the default behavior when no
2229 ssl_bump option is given or no ssl_bump ACLs match.
2231 By default, no connections are bumped.
2233 The first matching ssl_bump option wins. If no ACLs match, the
2234 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2235 does not have an implicit "negate the last given option" rule. You
2236 must make that rule explicit if you convert old ssl_bump allow/deny
2237 rules that rely on such an implicit rule.
2239 This clause supports both fast and slow acl types.
2240 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2242 See also: http_port ssl-bump, https_port ssl-bump
2245 # Example: Bump all requests except those originating from
2246 # localhost and those going to example.com.
2248 acl broken_sites dstdomain .example.com
2249 ssl_bump none localhost
2250 ssl_bump none broken_sites
2251 ssl_bump server-first all
2254 NAME: sslproxy_flags
2257 LOC: Config.ssl_client.flags
2260 Various flags modifying the use of SSL while proxying https:// URLs:
2261 DONT_VERIFY_PEER Accept certificates that fail verification.
2262 For refined control, see sslproxy_cert_error.
2263 NO_DEFAULT_CA Don't use the default CA list built in
2267 NAME: sslproxy_cert_error
2270 LOC: Config.ssl_client.cert_error
2273 Use this ACL to bypass server certificate validation errors.
2275 For example, the following lines will bypass all validation errors
2276 when talking to servers for example.com. All other
2277 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2279 acl BrokenButTrustedServers dstdomain example.com
2280 sslproxy_cert_error allow BrokenButTrustedServers
2281 sslproxy_cert_error deny all
2283 This clause only supports fast acl types.
2284 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2285 Using slow acl types may result in server crashes
2287 Without this option, all server certificate validation errors
2288 terminate the transaction. Bypassing validation errors is dangerous
2289 because an error usually implies that the server cannot be trusted and
2290 the connection may be insecure.
2292 See also: sslproxy_flags and DONT_VERIFY_PEER.
2294 Default setting: sslproxy_cert_error deny all
2297 NAME: sslproxy_cert_sign
2300 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2301 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2302 POSTSCRIPTUM: signTrusted all
2303 TYPE: sslproxy_cert_sign
2304 LOC: Config.ssl_client.cert_sign
2307 sslproxy_cert_sign <signing algorithm> acl ...
2309 The following certificate signing algorithms are supported:
2311 Sign using the configured CA certificate which is usually
2312 placed in and trusted by end-user browsers. This is the
2313 default for trusted origin server certificates.
2315 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2316 This is the default for untrusted origin server certificates
2317 that are not self-signed (see ssl::certUntrusted).
2319 Sign using a self-signed certificate with the right CN to
2320 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2321 browser. This is the default for self-signed origin server
2322 certificates (see ssl::certSelfSigned).
2324 This clause only supports fast acl types.
2326 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2327 signing algorithm to generate the certificate and ignores all
2328 subsequent sslproxy_cert_sign options (the first match wins). If no
2329 acl(s) match, the default signing algorithm is determined by errors
2330 detected when obtaining and validating the origin server certificate.
2332 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2333 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2334 CONNECT request that carries a domain name. In all other cases (CONNECT
2335 to an IP address or an intercepted SSL connection), Squid cannot detect
2336 the domain mismatch at certificate generation time when
2337 bump-server-first is used.
2340 NAME: sslproxy_cert_adapt
2343 TYPE: sslproxy_cert_adapt
2344 LOC: Config.ssl_client.cert_adapt
2347 sslproxy_cert_adapt <adaptation algorithm> acl ...
2349 The following certificate adaptation algorithms are supported:
2351 Sets the "Not After" property to the "Not After" property of
2352 the CA certificate used to sign generated certificates.
2354 Sets the "Not Before" property to the "Not Before" property of
2355 the CA certificate used to sign generated certificates.
2356 setCommonName or setCommonName{CN}
2357 Sets Subject.CN property to the host name specified as a
2358 CN parameter or, if no explicit CN parameter was specified,
2359 extracted from the CONNECT request. It is a misconfiguration
2360 to use setCommonName without an explicit parameter for
2361 intercepted or tproxied SSL connections.
2363 This clause only supports fast acl types.
2365 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2366 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2367 corresponding adaptation algorithm to generate the certificate and
2368 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2369 group (i.e., the first match wins within each algorithm group). If no
2370 acl(s) match, the default mimicking action takes place.
2372 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2373 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2374 CONNECT request that carries a domain name. In all other cases (CONNECT
2375 to an IP address or an intercepted SSL connection), Squid cannot detect
2376 the domain mismatch at certificate generation time when
2377 bump-server-first is used.
2380 NAME: sslpassword_program
2383 LOC: Config.Program.ssl_password
2386 Specify a program used for entering SSL key passphrases
2387 when using encrypted SSL certificate keys. If not specified
2388 keys must either be unencrypted, or Squid started with the -N
2389 option to allow it to query interactively for the passphrase.
2391 The key file name is given as argument to the program allowing
2392 selection of the right password if you have multiple encrypted
2397 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2398 -----------------------------------------------------------------------------
2401 NAME: sslcrtd_program
2404 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2405 LOC: Ssl::TheConfig.ssl_crtd
2407 Specify the location and options of the executable for ssl_crtd process.
2408 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2409 For more information use:
2410 @DEFAULT_SSL_CRTD@ -h
2413 NAME: sslcrtd_children
2414 TYPE: HelperChildConfig
2416 DEFAULT: 32 startup=5 idle=1
2417 LOC: Ssl::TheConfig.ssl_crtdChildren
2419 The maximum number of processes spawn to service ssl server.
2420 The maximum this may be safely set to is 32.
2422 The startup= and idle= options allow some measure of skew in your
2427 Sets the minimum number of processes to spawn when Squid
2428 starts or reconfigures. When set to zero the first request will
2429 cause spawning of the first child process to handle it.
2431 Starting too few children temporary slows Squid under load while it
2432 tries to spawn enough additional processes to cope with traffic.
2436 Sets a minimum of how many processes Squid is to try and keep available
2437 at all times. When traffic begins to rise above what the existing
2438 processes can handle this many more will be spawned up to the maximum
2439 configured. A minimum setting of 1 is required.
2441 You must have at least one ssl_crtd process.
2444 NAME: sslcrtvalidator_program
2448 LOC: Ssl::TheConfig.ssl_crt_validator
2450 Specify the location and options of the executable for ssl_crt_validator
2452 sslcrtvalidator_program [ttl=n] [cache=n] path ...
2455 ttl=n TTL in seconds for cached results.The default is 60 secs
2456 cache=n limit the result cache size. The default value is 2048
2459 NAME: sslcrtvalidator_children
2460 TYPE: HelperChildConfig
2462 DEFAULT: 32 startup=5 idle=1 concurrency=1
2463 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2465 The maximum number of processes spawn to service ssl server.
2466 The maximum this may be safely set to is 32.
2468 The startup= and idle= options allow some measure of skew in your
2473 Sets the minimum number of processes to spawn when Squid
2474 starts or reconfigures. When set to zero the first request will
2475 cause spawning of the first child process to handle it.
2477 Starting too few children temporary slows Squid under load while it
2478 tries to spawn enough additional processes to cope with traffic.
2482 Sets a minimum of how many processes Squid is to try and keep available
2483 at all times. When traffic begins to rise above what the existing
2484 processes can handle this many more will be spawned up to the maximum
2485 configured. A minimum setting of 1 is required.
2489 The number of requests each certificate validator helper can handle in
2490 parallel. Defaults to 0 which indicates the certficate validator
2491 is a old-style single threaded redirector.
2493 When this directive is set to a value >= 1 then the protocol
2494 used to communicate with the helper is modified to include
2495 a request ID in front of the request/response. The request
2496 ID from the request must be echoed back with the response
2499 You must have at least one ssl_crt_validator process.
2503 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2504 -----------------------------------------------------------------------------
2512 To specify other caches in a hierarchy, use the format:
2514 cache_peer hostname type http-port icp-port [options]
2519 # hostname type port port options
2520 # -------------------- -------- ----- ----- -----------
2521 cache_peer parent.foo.net parent 3128 3130 default
2522 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2523 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2524 cache_peer example.com parent 80 0 default
2525 cache_peer cdn.example.com sibling 3128 0
2527 type: either 'parent', 'sibling', or 'multicast'.
2529 proxy-port: The port number where the peer accept HTTP requests.
2530 For other Squid proxies this is usually 3128
2531 For web servers this is usually 80
2533 icp-port: Used for querying neighbor caches about objects.
2534 Set to 0 if the peer does not support ICP or HTCP.
2535 See ICP and HTCP options below for additional details.
2538 ==== ICP OPTIONS ====
2540 You MUST also set icp_port and icp_access explicitly when using these options.
2541 The defaults will prevent peer traffic using ICP.
2544 no-query Disable ICP queries to this neighbor.
2547 Indicates the named peer is a member of a multicast group.
2548 ICP queries will not be sent directly to the peer, but ICP
2549 replies will be accepted from it.
2551 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2552 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2555 To only send ICP queries to this neighbor infrequently.
2556 This is used to keep the neighbor round trip time updated
2557 and is usually used in conjunction with weighted-round-robin.
2560 ==== HTCP OPTIONS ====
2562 You MUST also set htcp_port and htcp_access explicitly when using these options.
2563 The defaults will prevent peer traffic using HTCP.
2566 htcp Send HTCP, instead of ICP, queries to the neighbor.
2567 You probably also want to set the "icp-port" to 4827
2568 instead of 3130. This directive accepts a comma separated
2569 list of options described below.
2571 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2573 htcp=no-clr Send HTCP to the neighbor but without
2574 sending any CLR requests. This cannot be used with
2577 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2578 This cannot be used with no-clr.
2581 Send HTCP to the neighbor including CLRs but only when
2582 they do not result from PURGE requests.
2585 Forward any HTCP CLR requests this proxy receives to the peer.
2588 ==== PEER SELECTION METHODS ====
2590 The default peer selection method is ICP, with the first responding peer
2591 being used as source. These options can be used for better load balancing.
2594 default This is a parent cache which can be used as a "last-resort"
2595 if a peer cannot be located by any of the peer-selection methods.
2596 If specified more than once, only the first is used.
2598 round-robin Load-Balance parents which should be used in a round-robin
2599 fashion in the absence of any ICP queries.
2600 weight=N can be used to add bias.
2602 weighted-round-robin
2603 Load-Balance parents which should be used in a round-robin
2604 fashion with the frequency of each parent being based on the
2605 round trip time. Closer parents are used more often.
2606 Usually used for background-ping parents.
2607 weight=N can be used to add bias.
2609 carp Load-Balance parents which should be used as a CARP array.
2610 The requests will be distributed among the parents based on the
2611 CARP load balancing hash function based on their weight.
2613 userhash Load-balance parents based on the client proxy_auth or ident username.
2615 sourcehash Load-balance parents based on the client source IP.
2618 To be used only for cache peers of type "multicast".
2619 ALL members of this multicast group have "sibling"
2620 relationship with it, not "parent". This is to a multicast
2621 group when the requested object would be fetched only from
2622 a "parent" cache, anyway. It's useful, e.g., when
2623 configuring a pool of redundant Squid proxies, being
2624 members of the same multicast group.
2627 ==== PEER SELECTION OPTIONS ====
2629 weight=N use to affect the selection of a peer during any weighted
2630 peer-selection mechanisms.
2631 The weight must be an integer; default is 1,
2632 larger weights are favored more.
2633 This option does not affect parent selection if a peering
2634 protocol is not in use.
2636 basetime=N Specify a base amount to be subtracted from round trip
2638 It is subtracted before division by weight in calculating
2639 which parent to fectch from. If the rtt is less than the
2640 base time the rtt is set to a minimal value.
2642 ttl=N Specify a TTL to use when sending multicast ICP queries
2644 Only useful when sending to a multicast group.
2645 Because we don't accept ICP replies from random
2646 hosts, you must configure other group members as
2647 peers with the 'multicast-responder' option.
2649 no-delay To prevent access to this neighbor from influencing the
2652 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2653 enabled) for this host from the specified URL rather
2654 than the Squid default location.
2657 ==== CARP OPTIONS ====
2659 carp-key=key-specification
2660 use a different key than the full URL to hash against the peer.
2661 the key-specification is a comma-separated list of the keywords
2662 scheme, host, port, path, params
2663 Order is not important.
2665 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2667 originserver Causes this parent to be contacted as an origin server.
2668 Meant to be used in accelerator setups when the peer
2672 Set the Host header of requests forwarded to this peer.
2673 Useful in accelerator setups where the server (peer)
2674 expects a certain domain name but clients may request
2675 others. ie example.com or www.example.com
2677 no-digest Disable request of cache digests.
2680 Disables requesting ICMP RTT database (NetDB).
2683 ==== AUTHENTICATION OPTIONS ====
2686 If this is a personal/workgroup proxy and your parent
2687 requires proxy authentication.
2689 Note: The string can include URL escapes (i.e. %20 for
2690 spaces). This also means % must be written as %%.
2693 Send login details received from client to this peer.
2694 Both Proxy- and WWW-Authorization headers are passed
2695 without alteration to the peer.
2696 Authentication is not required by Squid for this to work.
2698 Note: This will pass any form of authentication but
2699 only Basic auth will work through a proxy unless the
2700 connection-auth options are also used.
2702 login=PASS Send login details received from client to this peer.
2703 Authentication is not required by this option.
2705 If there are no client-provided authentication headers
2706 to pass on, but username and password are available
2707 from an external ACL user= and password= result tags
2708 they may be sent instead.
2710 Note: To combine this with proxy_auth both proxies must
2711 share the same user database as HTTP only allows for
2712 a single login (one for proxy, one for origin server).
2713 Also be warned this will expose your users proxy
2714 password to the peer. USE WITH CAUTION
2717 Send the username to the upstream cache, but with a
2718 fixed password. This is meant to be used when the peer
2719 is in another administrative domain, but it is still
2720 needed to identify each user.
2721 The star can optionally be followed by some extra
2722 information which is added to the username. This can
2723 be used to identify this proxy to the peer, similar to
2724 the login=username:password option above.
2727 If this is a personal/workgroup proxy and your parent
2728 requires a secure proxy authentication.
2729 The first principal from the default keytab or defined by
2730 the environment variable KRB5_KTNAME will be used.
2732 WARNING: The connection may transmit requests from multiple
2733 clients. Negotiate often assumes end-to-end authentication
2734 and a single-client. Which is not strictly true here.
2736 login=NEGOTIATE:principal_name
2737 If this is a personal/workgroup proxy and your parent
2738 requires a secure proxy authentication.
2739 The principal principal_name from the default keytab or
2740 defined by the environment variable KRB5_KTNAME will be
2743 WARNING: The connection may transmit requests from multiple
2744 clients. Negotiate often assumes end-to-end authentication
2745 and a single-client. Which is not strictly true here.
2747 connection-auth=on|off
2748 Tell Squid that this peer does or not support Microsoft
2749 connection oriented authentication, and any such
2750 challenges received from there should be ignored.
2751 Default is auto to automatically determine the status
2755 ==== SSL / HTTPS / TLS OPTIONS ====
2757 ssl Encrypt connections to this peer with SSL/TLS.
2759 sslcert=/path/to/ssl/certificate
2760 A client SSL certificate to use when connecting to
2763 sslkey=/path/to/ssl/key
2764 The private SSL key corresponding to sslcert above.
2765 If 'sslkey' is not specified 'sslcert' is assumed to
2766 reference a combined file containing both the
2767 certificate and the key.
2769 sslversion=1|2|3|4|5|6
2770 The SSL version to use when connecting to this peer
2771 1 = automatic (default)
2778 sslcipher=... The list of valid SSL ciphers to use when connecting
2781 ssloptions=... Specify various SSL implementation options:
2783 NO_SSLv2 Disallow the use of SSLv2
2784 NO_SSLv3 Disallow the use of SSLv3
2785 NO_TLSv1 Disallow the use of TLSv1.0
2786 NO_TLSv1_1 Disallow the use of TLSv1.1
2787 NO_TLSv1_2 Disallow the use of TLSv1.2
2789 Always create a new key when using
2790 temporary/ephemeral DH key exchanges
2791 ALL Enable various bug workarounds
2792 suggested as "harmless" by OpenSSL
2793 Be warned that this reduces SSL/TLS
2794 strength to some attacks.
2796 See the OpenSSL SSL_CTX_set_options documentation for a
2799 sslcafile=... A file containing additional CA certificates to use
2800 when verifying the peer certificate.
2802 sslcapath=... A directory containing additional CA certificates to
2803 use when verifying the peer certificate.
2805 sslcrlfile=... A certificate revocation list file to use when
2806 verifying the peer certificate.
2808 sslflags=... Specify various flags modifying the SSL implementation:
2811 Accept certificates even if they fail to
2814 Don't use the default CA list built in
2817 Don't verify the peer certificate
2818 matches the server name
2820 ssldomain= The peer name as advertised in it's certificate.
2821 Used for verifying the correctness of the received peer
2822 certificate. If not specified the peer hostname will be
2826 Enable the "Front-End-Https: On" header needed when
2827 using Squid as a SSL frontend in front of Microsoft OWA.
2828 See MS KB document Q307347 for details on this header.
2829 If set to auto the header will only be added if the
2830 request is forwarded as a https:// URL.
2833 ==== GENERAL OPTIONS ====
2836 A peer-specific connect timeout.
2837 Also see the peer_connect_timeout directive.
2839 connect-fail-limit=N
2840 How many times connecting to a peer must fail before
2841 it is marked as down. Default is 10.
2843 allow-miss Disable Squid's use of only-if-cached when forwarding
2844 requests to siblings. This is primarily useful when
2845 icp_hit_stale is used by the sibling. To extensive use
2846 of this option may result in forwarding loops, and you
2847 should avoid having two-way peerings with this option.
2848 For example to deny peer usage on requests from peer
2849 by denying cache_peer_access if the source is a peer.
2851 max-conn=N Limit the amount of connections Squid may open to this
2854 name=xxx Unique name for the peer.
2855 Required if you have multiple peers on the same host
2856 but different ports.
2857 This name can be used in cache_peer_access and similar
2858 directives to dentify the peer.
2859 Can be used by outgoing access controls through the
2862 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2863 requests to this peer. Use normal address selection instead.
2865 proxy-only objects fetched from the peer will not be stored locally.
2869 NAME: cache_peer_domain cache_host_domain
2874 Use to limit the domains for which a neighbor cache will be
2877 cache_peer_domain cache-host domain [domain ...]
2878 cache_peer_domain cache-host !domain
2880 For example, specifying
2882 cache_peer_domain parent.foo.net .edu
2884 has the effect such that UDP query packets are sent to
2885 'bigserver' only when the requested object exists on a
2886 server in the .edu domain. Prefixing the domainname
2887 with '!' means the cache will be queried for objects
2890 NOTE: * Any number of domains may be given for a cache-host,
2891 either on the same or separate lines.
2892 * When multiple domains are given for a particular
2893 cache-host, the first matched domain is applied.
2894 * Cache hosts with no domain restrictions are queried
2896 * There are no defaults.
2897 * There is also a 'cache_peer_access' tag in the ACL
2901 NAME: cache_peer_access
2906 Similar to 'cache_peer_domain' but provides more flexibility by
2909 cache_peer_access cache-host allow|deny [!]aclname ...
2911 The syntax is identical to 'http_access' and the other lists of
2912 ACL elements. See the comments for 'http_access' below, or
2913 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2916 NAME: neighbor_type_domain
2917 TYPE: hostdomaintype
2921 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2923 Modifying the neighbor type for specific domains is now
2924 possible. You can treat some domains differently than the
2925 default neighbor type specified on the 'cache_peer' line.
2926 Normally it should only be necessary to list domains which
2927 should be treated differently because the default neighbor type
2928 applies for hostnames which do not match domains listed here.
2931 cache_peer cache.foo.org parent 3128 3130
2932 neighbor_type_domain cache.foo.org sibling .com .net
2933 neighbor_type_domain cache.foo.org sibling .au .de
2936 NAME: dead_peer_timeout
2940 LOC: Config.Timeout.deadPeer
2942 This controls how long Squid waits to declare a peer cache
2943 as "dead." If there are no ICP replies received in this
2944 amount of time, Squid will declare the peer dead and not
2945 expect to receive any further ICP replies. However, it
2946 continues to send ICP queries, and will mark the peer as
2947 alive upon receipt of the first subsequent ICP reply.
2949 This timeout also affects when Squid expects to receive ICP
2950 replies from peers. If more than 'dead_peer' seconds have
2951 passed since the last ICP reply was received, Squid will not
2952 expect to receive an ICP reply on the next query. Thus, if
2953 your time between requests is greater than this timeout, you
2954 will see a lot of requests sent DIRECT to origin servers
2955 instead of to your parents.
2958 NAME: forward_max_tries
2961 LOC: Config.forward_max_tries
2963 Controls how many different forward paths Squid will try
2964 before giving up. See also forward_timeout.
2966 NOTE: connect_retries (default: none) can make each of these
2967 possible forwarding paths be tried multiple times.
2970 NAME: hierarchy_stoplist
2973 LOC: Config.hierarchy_stoplist
2975 A list of words which, if found in a URL, cause the object to
2976 be handled directly by this cache. In other words, use this
2977 to not query neighbor caches for certain objects. You may
2978 list this option multiple times.
2981 hierarchy_stoplist cgi-bin ?
2983 Note: never_direct overrides this option.
2987 MEMORY CACHE OPTIONS
2988 -----------------------------------------------------------------------------
2995 LOC: Config.memMaxSize
2997 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2998 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2999 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
3000 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
3002 'cache_mem' specifies the ideal amount of memory to be used
3004 * In-Transit objects
3006 * Negative-Cached objects
3008 Data for these objects are stored in 4 KB blocks. This
3009 parameter specifies the ideal upper limit on the total size of
3010 4 KB blocks allocated. In-Transit objects take the highest
3013 In-transit objects have priority over the others. When
3014 additional space is needed for incoming data, negative-cached
3015 and hot objects will be released. In other words, the
3016 negative-cached and hot objects will fill up any unused space
3017 not needed for in-transit objects.
3019 If circumstances require, this limit will be exceeded.
3020 Specifically, if your incoming request rate requires more than
3021 'cache_mem' of memory to hold in-transit objects, Squid will
3022 exceed this limit to satisfy the new requests. When the load
3023 decreases, blocks will be freed until the high-water mark is
3024 reached. Thereafter, blocks will be used to store hot
3027 If shared memory caching is enabled, Squid does not use the shared
3028 cache space for in-transit objects, but they still consume as much
3029 local memory as they need. For more details about the shared memory
3030 cache, see memory_cache_shared.
3033 NAME: maximum_object_size_in_memory
3037 LOC: Config.Store.maxInMemObjSize
3039 Objects greater than this size will not be attempted to kept in
3040 the memory cache. This should be set high enough to keep objects
3041 accessed frequently in memory to improve performance whilst low
3042 enough to keep larger objects from hoarding cache_mem.
3045 NAME: memory_cache_shared
3048 LOC: Config.memShared
3050 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3052 Controls whether the memory cache is shared among SMP workers.
3054 The shared memory cache is meant to occupy cache_mem bytes and replace
3055 the non-shared memory cache, although some entities may still be
3056 cached locally by workers for now (e.g., internal and in-transit
3057 objects may be served from a local memory cache even if shared memory
3058 caching is enabled).
3060 By default, the memory cache is shared if and only if all of the
3061 following conditions are satisfied: Squid runs in SMP mode with
3062 multiple workers, cache_mem is positive, and Squid environment
3063 supports required IPC primitives (e.g., POSIX shared memory segments
3064 and GCC-style atomic operations).
3066 To avoid blocking locks, shared memory uses opportunistic algorithms
3067 that do not guarantee that every cachable entity that could have been
3068 shared among SMP workers will actually be shared.
3070 Currently, entities exceeding 32KB in size cannot be shared.
3073 NAME: memory_cache_mode
3078 Controls which objects to keep in the memory cache (cache_mem)
3080 always Keep most recently fetched objects in memory (default)
3082 disk Only disk cache hits are kept in memory, which means
3083 an object must first be cached on disk and then hit
3084 a second time before cached in memory.
3086 network Only objects fetched from network is kept in memory
3089 NAME: memory_replacement_policy
3091 LOC: Config.memPolicy
3094 The memory replacement policy parameter determines which
3095 objects are purged from memory when memory space is needed.
3097 See cache_replacement_policy for details.
3102 -----------------------------------------------------------------------------
3105 NAME: cache_replacement_policy
3107 LOC: Config.replPolicy
3110 The cache replacement policy parameter determines which
3111 objects are evicted (replaced) when disk space is needed.
3113 lru : Squid's original list based LRU policy
3114 heap GDSF : Greedy-Dual Size Frequency
3115 heap LFUDA: Least Frequently Used with Dynamic Aging
3116 heap LRU : LRU policy implemented using a heap
3118 Applies to any cache_dir lines listed below this.
3120 The LRU policies keeps recently referenced objects.
3122 The heap GDSF policy optimizes object hit rate by keeping smaller
3123 popular objects in cache so it has a better chance of getting a
3124 hit. It achieves a lower byte hit rate than LFUDA though since
3125 it evicts larger (possibly popular) objects.
3127 The heap LFUDA policy keeps popular objects in cache regardless of
3128 their size and thus optimizes byte hit rate at the expense of
3129 hit rate since one large, popular object will prevent many
3130 smaller, slightly less popular objects from being cached.
3132 Both policies utilize a dynamic aging mechanism that prevents
3133 cache pollution that can otherwise occur with frequency-based
3134 replacement policies.
3136 NOTE: if using the LFUDA replacement policy you should increase
3137 the value of maximum_object_size above its default of 4096 KB to
3138 to maximize the potential byte hit rate improvement of LFUDA.
3140 For more information about the GDSF and LFUDA cache replacement
3141 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3142 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3148 LOC: Config.cacheSwap
3152 cache_dir Type Directory-Name Fs-specific-data [options]
3154 You can specify multiple cache_dir lines to spread the
3155 cache among different disk partitions.
3157 Type specifies the kind of storage system to use. Only "ufs"
3158 is built by default. To enable any of the other storage systems
3159 see the --enable-storeio configure option.
3161 'Directory' is a top-level directory where cache swap
3162 files will be stored. If you want to use an entire disk
3163 for caching, this can be the mount-point directory.
3164 The directory must exist and be writable by the Squid
3165 process. Squid will NOT create this directory for you.
3167 In SMP configurations, cache_dir must not precede the workers option
3168 and should use configuration macros or conditionals to give each
3169 worker interested in disk caching a dedicated cache directory.
3173 "ufs" is the old well-known Squid storage format that has always
3176 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3178 'Mbytes' is the amount of disk space (MB) to use under this
3179 directory. The default is 100 MB. Change this to suit your
3180 configuration. Do NOT put the size of your disk drive here.
3181 Instead, if you want Squid to use the entire disk drive,
3182 subtract 20% and use that value.
3184 'L1' is the number of first-level subdirectories which
3185 will be created under the 'Directory'. The default is 16.
3187 'L2' is the number of second-level subdirectories which
3188 will be created under each first-level directory. The default
3191 The aufs store type:
3193 "aufs" uses the same storage format as "ufs", utilizing
3194 POSIX-threads to avoid blocking the main Squid process on
3195 disk-I/O. This was formerly known in Squid as async-io.
3197 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3199 see argument descriptions under ufs above
3201 The diskd store type:
3203 "diskd" uses the same storage format as "ufs", utilizing a
3204 separate process to avoid blocking the main Squid process on
3207 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3209 see argument descriptions under ufs above
3211 Q1 specifies the number of unacknowledged I/O requests when Squid
3212 stops opening new files. If this many messages are in the queues,
3213 Squid won't open new files. Default is 64
3215 Q2 specifies the number of unacknowledged messages when Squid
3216 starts blocking. If this many messages are in the queues,
3217 Squid blocks until it receives some replies. Default is 72
3219 When Q1 < Q2 (the default), the cache directory is optimized
3220 for lower response time at the expense of a decrease in hit
3221 ratio. If Q1 > Q2, the cache directory is optimized for
3222 higher hit ratio at the expense of an increase in response
3225 The rock store type:
3227 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3229 The Rock Store type is a database-style storage. All cached
3230 entries are stored in a "database" file, using fixed-size slots,
3231 one entry per slot. The database size is specified in MB. The
3232 slot size is specified in bytes using the max-size option. See
3233 below for more info on the max-size option.
3235 swap-timeout=msec: Squid will not start writing a miss to or
3236 reading a hit from disk if it estimates that the swap operation
3237 will take more than the specified number of milliseconds. By
3238 default and when set to zero, disables the disk I/O time limit
3239 enforcement. Ignored when using blocking I/O module because
3240 blocking synchronous I/O does not allow Squid to estimate the
3241 expected swap wait time.
3243 max-swap-rate=swaps/sec: Artificially limits disk access using
3244 the specified I/O rate limit. Swap out requests that
3245 would cause the average I/O rate to exceed the limit are
3246 delayed. Individual swap in requests (i.e., hits or reads) are
3247 not delayed, but they do contribute to measured swap rate and
3248 since they are placed in the same FIFO queue as swap out
3249 requests, they may wait longer if max-swap-rate is smaller.
3250 This is necessary on file systems that buffer "too
3251 many" writes and then start blocking Squid and other processes
3252 while committing those writes to disk. Usually used together
3253 with swap-timeout to avoid excessive delays and queue overflows
3254 when disk demand exceeds available disk "bandwidth". By default
3255 and when set to zero, disables the disk I/O rate limit
3256 enforcement. Currently supported by IpcIo module only.
3259 The coss store type:
3261 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3262 production use and has thus been removed from this release.
3263 We hope that it can be made usable again soon.
3265 block-size=n defines the "block size" for COSS cache_dir's.
3266 Squid uses file numbers as block numbers. Since file numbers
3267 are limited to 24 bits, the block size determines the maximum
3268 size of the COSS partition. The default is 512 bytes, which
3269 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3270 you should not change the coss block size after Squid
3271 has written some objects to the cache_dir.
3273 The coss file store has changed from 2.5. Now it uses a file
3274 called 'stripe' in the directory names in the config - and
3275 this will be created by squid -z.
3279 no-store, no new objects should be stored to this cache_dir
3281 min-size=n, refers to the min object size in bytes this cache_dir
3282 will accept. It's used to restrict a cache_dir to only store
3283 large objects (e.g. aufs) while other storedirs are optimized
3284 for smaller objects (e.g. COSS). Defaults to 0.
3286 max-size=n, refers to the max object size in bytes this cache_dir
3287 supports. It is used to select the cache_dir to store the object.
3288 Note: To make optimal use of the max-size limits you should order
3289 the cache_dir lines with the smallest max-size value first and the
3290 ones with no max-size specification last.
3292 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3293 which can be changed with the --with-coss-membuf-size=N configure
3297 # Uncomment and adjust the following to add a disk cache directory.
3298 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3302 NAME: store_dir_select_algorithm
3304 LOC: Config.store_dir_select_algorithm
3307 Set this to 'round-robin' as an alternative.
3310 NAME: max_open_disk_fds
3312 LOC: Config.max_open_disk_fds
3315 To avoid having disk as the I/O bottleneck Squid can optionally
3316 bypass the on-disk cache if more than this amount of disk file
3317 descriptors are open.
3319 A value of 0 indicates no limit.
3322 NAME: minimum_object_size
3326 LOC: Config.Store.minObjectSize
3328 Objects smaller than this size will NOT be saved on disk. The
3329 value is specified in kilobytes, and the default is 0 KB, which
3330 means there is no minimum.
3333 NAME: maximum_object_size
3337 LOC: Config.Store.maxObjectSize
3339 Objects larger than this size will NOT be saved on disk. The
3340 value is specified in kilobytes, and the default is 4MB. If
3341 you wish to get a high BYTES hit ratio, you should probably
3342 increase this (one 32 MB object hit counts for 3200 10KB
3343 hits). If you wish to increase speed more than your want to
3344 save bandwidth you should leave this low.
3346 NOTE: if using the LFUDA replacement policy you should increase
3347 this value to maximize the byte hit rate improvement of LFUDA!
3348 See replacement_policy below for a discussion of this policy.
3351 NAME: cache_swap_low
3352 COMMENT: (percent, 0-100)
3355 LOC: Config.Swap.lowWaterMark
3358 NAME: cache_swap_high
3359 COMMENT: (percent, 0-100)
3362 LOC: Config.Swap.highWaterMark
3365 The low- and high-water marks for cache object replacement.
3366 Replacement begins when the swap (disk) usage is above the
3367 low-water mark and attempts to maintain utilization near the
3368 low-water mark. As swap utilization gets close to high-water
3369 mark object eviction becomes more aggressive. If utilization is
3370 close to the low-water mark less replacement is done each time.
3372 Defaults are 90% and 95%. If you have a large cache, 5% could be
3373 hundreds of MB. If this is the case you may wish to set these
3374 numbers closer together.
3379 -----------------------------------------------------------------------------
3389 logformat <name> <format specification>
3391 Defines an access log format.
3393 The <format specification> is a string with embedded % format codes
3395 % format codes all follow the same basic structure where all but
3396 the formatcode is optional. Output strings are automatically escaped
3397 as required according to their context and the output format
3398 modifiers are usually not needed, but can be specified if an explicit
3399 output format is desired.
3401 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3403 " output in quoted string format
3404 [ output in squid text log format as used by log_mime_hdrs
3405 # output in URL quoted format
3410 width minimum and/or maximum field width:
3411 [width_min][.width_max]
3412 When minimum starts with 0, the field is zero-padded.
3413 String values exceeding maximum width are truncated.
3415 {arg} argument such as header name etc
3419 % a literal % character
3420 sn Unique sequence number per log line entry
3421 err_code The ID of an error response served by Squid or
3422 a similar internal error identifier.
3423 err_detail Additional err_code-dependent error information.
3424 note The meta header specified by the argument. Also
3425 logs the adaptation meta headers set by the
3426 adaptation_meta configuration parameter.
3427 If no argument given all meta headers logged.
3429 Connection related format codes:
3431 >a Client source IP address
3433 >p Client source port
3434 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3435 >la Local IP address the client connected to
3436 >lp Local port number the client connected to
3438 la Local listening IP address the client connection was connected to.
3439 lp Local listening port number the client connection was connected to.
3441 <a Server IP address of the last server or peer connection
3442 <A Server FQDN or peer name
3443 <p Server port number of the last server or peer connection
3444 <la Local IP address of the last server or peer connection
3445 <lp Local port number of the last server or peer connection
3447 Time related format codes:
3449 ts Seconds since epoch
3450 tu subsecond time (milliseconds)
3451 tl Local time. Optional strftime format argument
3452 default %d/%b/%Y:%H:%M:%S %z
3453 tg GMT time. Optional strftime format argument
3454 default %d/%b/%Y:%H:%M:%S %z
3455 tr Response time (milliseconds)
3456 dt Total time spent making DNS lookups (milliseconds)
3458 Access Control related format codes:
3460 et Tag returned by external acl
3461 ea Log string returned by external acl
3462 un User name (any available)
3463 ul User name from authentication
3464 ue User name from external acl helper
3465 ui User name from ident
3466 us User name from SSL
3468 HTTP related format codes:
3470 [http::]>h Original request header. Optional header name argument
3471 on the format header[:[separator]element]
3472 [http::]>ha The HTTP request headers after adaptation and redirection.
3473 Optional header name argument as for >h
3474 [http::]<h Reply header. Optional header name argument
3476 [http::]>Hs HTTP status code sent to the client
3477 [http::]<Hs HTTP status code received from the next hop
3478 [http::]<bs Number of HTTP-equivalent message body bytes
3479 received from the next hop, excluding chunked
3480 transfer encoding and control messages.
3481 Generated FTP/Gopher listings are treated as
3483 [http::]mt MIME content type
3484 [http::]rm Request method (GET/POST etc)
3485 [http::]>rm Request method from client
3486 [http::]<rm Request method sent to server or peer
3487 [http::]ru Request URL from client (historic, filtered for logging)
3488 [http::]>ru Request URL from client
3489 [http::]<ru Request URL sent to server or peer
3490 [http::]rp Request URL-Path excluding hostname
3491 [http::]>rp Request URL-Path excluding hostname from client
3492 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3493 [http::]rv Request protocol version
3494 [http::]>rv Request protocol version from client
3495 [http::]<rv Request protocol version sent to server or peer
3496 [http::]<st Sent reply size including HTTP headers
3497 [http::]>st Received request size including HTTP headers. In the
3498 case of chunked requests the chunked encoding metadata
3500 [http::]>sh Received HTTP request headers size
3501 [http::]<sh Sent HTTP reply headers size
3502 [http::]st Request+Reply size including HTTP headers
3503 [http::]<sH Reply high offset sent
3504 [http::]<sS Upstream object size
3505 [http::]<pt Peer response time in milliseconds. The timer starts
3506 when the last request byte is sent to the next hop
3507 and stops when the last response byte is received.
3508 [http::]<tt Total server-side time in milliseconds. The timer
3509 starts with the first connect request (or write I/O)
3510 sent to the first selected peer. The timer stops
3511 with the last I/O with the last peer.
3513 Squid handling related format codes:
3515 Ss Squid request status (TCP_MISS etc)
3516 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3518 SSL-related format codes:
3520 ssl::bump_mode SslBump decision for the transaction:
3522 For CONNECT requests that initiated bumping of
3523 a connection and for any request received on
3524 an already bumped connection, Squid logs the
3525 corresponding SslBump mode ("server-first" or
3526 "client-first"). See the ssl_bump option for
3527 more information about these modes.
3529 A "none" token is logged for requests that
3530 triggered "ssl_bump" ACL evaluation matching
3531 either a "none" rule or no rules at all.
3533 In all other cases, a single dash ("-") is
3536 If ICAP is enabled, the following code becomes available (as
3537 well as ICAP log codes documented with the icap_log option):
3539 icap::tt Total ICAP processing time for the HTTP
3540 transaction. The timer ticks when ICAP
3541 ACLs are checked and when ICAP
3542 transaction is in progress.
3544 If adaptation is enabled the following three codes become available:
3546 adapt::<last_h The header of the last ICAP response or
3547 meta-information from the last eCAP
3548 transaction related to the HTTP transaction.
3549 Like <h, accepts an optional header name
3552 adapt::sum_trs Summed adaptation transaction response
3553 times recorded as a comma-separated list in
3554 the order of transaction start time. Each time
3555 value is recorded as an integer number,
3556 representing response time of one or more
3557 adaptation (ICAP or eCAP) transaction in
3558 milliseconds. When a failed transaction is
3559 being retried or repeated, its time is not
3560 logged individually but added to the
3561 replacement (next) transaction. See also:
3564 adapt::all_trs All adaptation transaction response times.
3565 Same as adaptation_strs but response times of
3566 individual transactions are never added
3567 together. Instead, all transaction response
3568 times are recorded individually.
3570 You can prefix adapt::*_trs format codes with adaptation
3571 service name in curly braces to record response time(s) specific
3572 to that service. For example: %{my_service}adapt::sum_trs
3574 If SSL is enabled, the following formating codes become available:
3576 %ssl::>cert_subject The Subject field of the received client
3577 SSL certificate or a dash ('-') if Squid has
3578 received an invalid/malformed certificate or
3579 no certificate at all. Consider encoding the
3580 logged value because Subject often has spaces.
3582 %ssl::>cert_issuer The Issuer field of the received client
3583 SSL certificate or a dash ('-') if Squid has
3584 received an invalid/malformed certificate or
3585 no certificate at all. Consider encoding the
3586 logged value because Issuer often has spaces.
3588 The default formats available (which do not need re-defining) are:
3590 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3591 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3592 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3593 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3594 logformat useragent %>a [%tl] "%{User-Agent}>h"
3596 NOTE: When the log_mime_hdrs directive is set to ON.
3597 The squid, common and combined formats have a safely encoded copy
3598 of the mime headers appended to each line within a pair of brackets.
3600 NOTE: The common and combined formats are not quite true to the Apache definition.
3601 The logs from Squid contain an extra status and hierarchy code appended.
3605 NAME: access_log cache_access_log
3607 LOC: Config.Log.accesslogs
3608 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3610 These files log client request activities. Has a line every HTTP or
3611 ICP request. The format is:
3612 access_log <module>:<place> [<logformat name> [acl acl ...]]
3613 access_log none [acl acl ...]]
3615 Will log to the specified module:place using the specified format (which
3616 must be defined in a logformat directive) those entries which match
3617 ALL the acl's specified (which must be defined in acl clauses).
3618 If no acl is specified, all requests will be logged to this destination.
3620 ===== Modules Currently available =====
3622 none Do not log any requests matching these ACL.
3623 Do not specify Place or logformat name.
3625 stdio Write each log line to disk immediately at the completion of
3627 Place: the filename and path to be written.
3629 daemon Very similar to stdio. But instead of writing to disk the log
3630 line is passed to a daemon helper for asychronous handling instead.
3631 Place: varies depending on the daemon.
3633 log_file_daemon Place: the file name and path to be written.
3635 syslog To log each request via syslog facility.
3636 Place: The syslog facility and priority level for these entries.
3637 Place Format: facility.priority
3639 where facility could be any of:
3640 authpriv, daemon, local0 ... local7 or user.
3642 And priority could be any of:
3643 err, warning, notice, info, debug.
3645 udp To send each log line as text data to a UDP receiver.
3646 Place: The destination host name or IP and port.
3647 Place Format: //host:port
3649 tcp To send each log line as text data to a TCP receiver.
3650 Place: The destination host name or IP and port.
3651 Place Format: //host:port
3654 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3660 LOC: Config.Log.icaplogs
3663 ICAP log files record ICAP transaction summaries, one line per
3666 The icap_log option format is:
3667 icap_log <filepath> [<logformat name> [acl acl ...]]
3668 icap_log none [acl acl ...]]
3670 Please see access_log option documentation for details. The two
3671 kinds of logs share the overall configuration approach and many
3674 ICAP processing of a single HTTP message or transaction may
3675 require multiple ICAP transactions. In such cases, multiple
3676 ICAP transaction log lines will correspond to a single access
3679 ICAP log uses logformat codes that make sense for an ICAP
3680 transaction. Header-related codes are applied to the HTTP header
3681 embedded in an ICAP server response, with the following caveats:
3682 For REQMOD, there is no HTTP response header unless the ICAP
3683 server performed request satisfaction. For RESPMOD, the HTTP
3684 request header is the header sent to the ICAP server. For
3685 OPTIONS, there are no HTTP headers.
3687 The following format codes are also available for ICAP logs:
3689 icap::<A ICAP server IP address. Similar to <A.
3691 icap::<service_name ICAP service name from the icap_service
3692 option in Squid configuration file.
3694 icap::ru ICAP Request-URI. Similar to ru.
3696 icap::rm ICAP request method (REQMOD, RESPMOD, or
3697 OPTIONS). Similar to existing rm.
3699 icap::>st Bytes sent to the ICAP server (TCP payload
3700 only; i.e., what Squid writes to the socket).
3702 icap::<st Bytes received from the ICAP server (TCP
3703 payload only; i.e., what Squid reads from
3706 icap::<bs Number of message body bytes received from the
3707 ICAP server. ICAP message body, if any, usually
3708 includes encapsulated HTTP message headers and
3709 possibly encapsulated HTTP message body. The
3710 HTTP body part is dechunked before its size is
3713 icap::tr Transaction response time (in
3714 milliseconds). The timer starts when
3715 the ICAP transaction is created and
3716 stops when the transaction is completed.
3719 icap::tio Transaction I/O time (in milliseconds). The
3720 timer starts when the first ICAP request
3721 byte is scheduled for sending. The timers
3722 stops when the last byte of the ICAP response
3725 icap::to Transaction outcome: ICAP_ERR* for all
3726 transaction errors, ICAP_OPT for OPTION
3727 transactions, ICAP_ECHO for 204
3728 responses, ICAP_MOD for message
3729 modification, and ICAP_SAT for request
3730 satisfaction. Similar to Ss.
3732 icap::Hs ICAP response status code. Similar to Hs.
3734 icap::>h ICAP request header(s). Similar to >h.
3736 icap::<h ICAP response header(s). Similar to <h.
3738 The default ICAP log format, which can be used without an explicit
3739 definition, is called icap_squid:
3741 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3743 See also: logformat, log_icap, and %adapt::<last_h
3746 NAME: logfile_daemon
3748 DEFAULT: @DEFAULT_LOGFILED@
3749 LOC: Log::TheConfig.logfile_daemon
3751 Specify the path to the logfile-writing daemon. This daemon is
3752 used to write the access and store logs, if configured.
3754 Squid sends a number of commands to the log daemon:
3755 L<data>\n - logfile data
3760 r<n>\n - set rotate count to <n>
3761 b<n>\n - 1 = buffer output, 0 = don't buffer output
3763 No responses is expected.
3768 LOC: Config.accessList.log
3770 COMMENT: allow|deny acl acl...
3772 This options allows you to control which requests gets logged
3773 to access.log (see access_log directive). Requests denied for
3774 logging will also not be accounted for in performance counters.
3776 This clause only supports fast acl types.
3777 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3783 LOC: Config.accessList.icap
3786 This options allows you to control which requests get logged
3787 to icap.log. See the icap_log directive for ICAP log details.
3790 NAME: cache_store_log
3793 LOC: Config.Log.store
3795 Logs the activities of the storage manager. Shows which
3796 objects are ejected from the cache, and which objects are
3797 saved and for how long.
3798 There are not really utilities to analyze this data, so you can safely
3799 disable it (the default).
3801 Store log uses modular logging outputs. See access_log for the list
3802 of modules supported.
3805 cache_store_log stdio:@DEFAULT_STORE_LOG@
3806 cache_store_log daemon:@DEFAULT_STORE_LOG@
3809 NAME: cache_swap_state cache_swap_log
3811 LOC: Config.Log.swap
3814 Location for the cache "swap.state" file. This index file holds
3815 the metadata of objects saved on disk. It is used to rebuild
3816 the cache during startup. Normally this file resides in each
3817 'cache_dir' directory, but you may specify an alternate
3818 pathname here. Note you must give a full filename, not just
3819 a directory. Since this is the index for the whole object
3820 list you CANNOT periodically rotate it!
3822 If %s can be used in the file name it will be replaced with a
3823 a representation of the cache_dir name where each / is replaced
3824 with '.'. This is needed to allow adding/removing cache_dir
3825 lines when cache_swap_log is being used.
3827 If have more than one 'cache_dir', and %s is not used in the name
3828 these swap logs will have names such as:
3834 The numbered extension (which is added automatically)
3835 corresponds to the order of the 'cache_dir' lines in this
3836 configuration file. If you change the order of the 'cache_dir'
3837 lines in this file, these index files will NOT correspond to
3838 the correct 'cache_dir' entry (unless you manually rename
3839 them). We recommend you do NOT use this option. It is
3840 better to keep these index files in each 'cache_dir' directory.
3843 NAME: logfile_rotate
3846 LOC: Config.Log.rotateNumber
3848 Specifies the number of logfile rotations to make when you
3849 type 'squid -k rotate'. The default is 10, which will rotate
3850 with extensions 0 through 9. Setting logfile_rotate to 0 will
3851 disable the file name rotation, but the logfiles are still closed
3852 and re-opened. This will enable you to rename the logfiles
3853 yourself just before sending the rotate signal.
3855 Note, the 'squid -k rotate' command normally sends a USR1
3856 signal to the running squid process. In certain situations
3857 (e.g. on Linux with Async I/O), USR1 is used for other
3858 purposes, so -k rotate uses another signal. It is best to get
3859 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3862 Note, from Squid-3.1 this option has no effect on the cache.log,
3863 that log can be rotated separately by using debug_options
3866 NAME: emulate_httpd_log
3869 Replace this with an access_log directive using the format 'common' or 'combined'.
3872 NAME: log_ip_on_direct
3875 Remove this option from your config. To log server or peer names use %<A in the log format.
3880 DEFAULT: @DEFAULT_MIME_TABLE@
3881 LOC: Config.mimeTablePathname
3883 Pathname to Squid's MIME table. You shouldn't need to change
3884 this, but the default file contains examples and formatting
3885 information if you do.
3891 LOC: Config.onoff.log_mime_hdrs
3894 The Cache can record both the request and the response MIME
3895 headers for each HTTP transaction. The headers are encoded
3896 safely and will appear as two bracketed fields at the end of
3897 the access log (for either the native or httpd-emulated log
3898 formats). To enable this logging set log_mime_hdrs to 'on'.
3904 Replace this with an access_log directive using the format 'useragent'.
3907 NAME: referer_log referrer_log
3910 Replace this with an access_log directive using the format 'referrer'.
3915 DEFAULT: @DEFAULT_PID_FILE@
3916 LOC: Config.pidFilename
3918 A filename to write the process-id to. To disable, enter "none".
3924 Remove this option from your config. To log FQDN use %>A in the log format.
3927 NAME: client_netmask
3929 LOC: Config.Addrs.client_netmask
3932 A netmask for client addresses in logfiles and cachemgr output.
3933 Change this to protect the privacy of your cache clients.
3934 A netmask of 255.255.255.0 will log all IP's in that range with
3935 the last digit set to '0'.
3941 Use a regular access.log with ACL limiting it to MISS events.
3944 NAME: strip_query_terms
3946 LOC: Config.onoff.strip_query_terms
3949 By default, Squid strips query terms from requested URLs before
3950 logging. This protects your user's privacy.
3957 LOC: Config.onoff.buffered_logs
3959 cache.log log file is written with stdio functions, and as such
3960 it can be buffered or unbuffered. By default it will be unbuffered.
3961 Buffering it can speed up the writing slightly (though you are
3962 unlikely to need to worry unless you run with tons of debugging
3963 enabled in which case performance will suffer badly anyway..).
3966 NAME: netdb_filename
3968 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3969 LOC: Config.netdbFilename
3972 A filename where Squid stores it's netdb state between restarts.
3973 To disable, enter "none".
3977 OPTIONS FOR TROUBLESHOOTING
3978 -----------------------------------------------------------------------------
3983 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3984 LOC: Debug::cache_log
3986 Cache logging file. This is where general information about
3987 your cache's behavior goes. You can increase the amount of data
3988 logged to this file and how often its rotated with "debug_options"
3994 LOC: Debug::debugOptions
3996 Logging options are set as section,level where each source file
3997 is assigned a unique section. Lower levels result in less
3998 output, Full debugging (level 9) can result in a very large
3999 log file, so be careful.
4001 The magic word "ALL" sets debugging levels for all sections.
4002 We recommend normally running with "ALL,1".
4004 The rotate=N option can be used to keep more or less of these logs
4005 than would otherwise be kept by logfile_rotate.
4006 For most uses a single log should be enough to monitor current
4007 events affecting Squid.
4012 LOC: Config.coredump_dir
4013 DEFAULT_IF_NONE: none
4015 By default Squid leaves core files in the directory from where
4016 it was started. If you set 'coredump_dir' to a directory
4017 that exists, Squid will chdir() to that directory at startup
4018 and coredump files will be left there.
4022 # Leave coredumps in the first cache dir
4023 coredump_dir @DEFAULT_SWAP_DIR@
4029 OPTIONS FOR FTP GATEWAYING
4030 -----------------------------------------------------------------------------
4036 LOC: Config.Ftp.anon_user
4038 If you want the anonymous login password to be more informative
4039 (and enable the use of picky ftp servers), set this to something
4040 reasonable for your domain, like wwwuser@somewhere.net
4042 The reason why this is domainless by default is the
4043 request can be made on the behalf of a user in any domain,
4044 depending on how the cache is used.
4045 Some ftp server also validate the email address is valid
4046 (for example perl.com).
4052 LOC: Config.Ftp.passive
4054 If your firewall does not allow Squid to use passive
4055 connections, turn off this option.
4057 Use of ftp_epsv_all option requires this to be ON.
4063 LOC: Config.Ftp.epsv_all
4065 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4067 NATs may be able to put the connection on a "fast path" through the
4068 translator, as the EPRT command will never be used and therefore,
4069 translation of the data portion of the segments will never be needed.
4071 When a client only expects to do two-way FTP transfers this may be
4073 If squid finds that it must do a three-way FTP transfer after issuing
4074 an EPSV ALL command, the FTP session will fail.
4076 If you have any doubts about this option do not use it.
4077 Squid will nicely attempt all other connection methods.
4079 Requires ftp_passive to be ON (default) for any effect.
4085 LOC: Config.Ftp.epsv
4087 FTP Protocol extensions permit the use of a special "EPSV" command.
4089 NATs may be able to put the connection on a "fast path" through the
4090 translator using EPSV, as the EPRT command will never be used
4091 and therefore, translation of the data portion of the segments
4092 will never be needed.
4094 Turning this OFF will prevent EPSV being attempted.
4095 WARNING: Doing so will convert Squid back to the old behavior with all
4096 the related problems with external NAT devices/layers.
4098 Requires ftp_passive to be ON (default) for any effect.
4104 LOC: Config.Ftp.eprt
4106 FTP Protocol extensions permit the use of a special "EPRT" command.
4108 This extension provides a protocol neutral alternative to the
4109 IPv4-only PORT command. When supported it enables active FTP data
4110 channels over IPv6 and efficient NAT handling.
4112 Turning this OFF will prevent EPRT being attempted and will skip
4113 straight to using PORT for IPv4 servers.
4115 Some devices are known to not handle this extension correctly and
4116 may result in crashes. Devices which suport EPRT enough to fail
4117 cleanly will result in Squid attempting PORT anyway. This directive
4118 should only be disabled when EPRT results in device failures.
4120 WARNING: Doing so will convert Squid back to the old behavior with all
4121 the related problems with external NAT devices/layers and IPv4-only FTP.
4124 NAME: ftp_sanitycheck
4127 LOC: Config.Ftp.sanitycheck
4129 For security and data integrity reasons Squid by default performs
4130 sanity checks of the addresses of FTP data connections ensure the
4131 data connection is to the requested server. If you need to allow
4132 FTP connections to servers using another IP address for the data
4133 connection turn this off.
4136 NAME: ftp_telnet_protocol
4139 LOC: Config.Ftp.telnet
4141 The FTP protocol is officially defined to use the telnet protocol
4142 as transport channel for the control connection. However, many
4143 implementations are broken and does not respect this aspect of
4146 If you have trouble accessing files with ASCII code 255 in the
4147 path or similar problems involving this ASCII code you can
4148 try setting this directive to off. If that helps, report to the
4149 operator of the FTP server in question that their FTP server
4150 is broken and does not follow the FTP standard.
4154 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4155 -----------------------------------------------------------------------------
4160 DEFAULT: @DEFAULT_DISKD@
4161 LOC: Config.Program.diskd
4163 Specify the location of the diskd executable.
4164 Note this is only useful if you have compiled in
4165 diskd as one of the store io modules.
4168 NAME: unlinkd_program
4171 DEFAULT: @DEFAULT_UNLINKD@
4172 LOC: Config.Program.unlinkd
4174 Specify the location of the executable for file deletion process.
4177 NAME: pinger_program
4179 DEFAULT: @DEFAULT_PINGER@
4180 LOC: Config.pinger.program
4183 Specify the location of the executable for the pinger process.
4189 LOC: Config.pinger.enable
4192 Control whether the pinger is active at run-time.
4193 Enables turning ICMP pinger on and off with a simple
4194 squid -k reconfigure.
4199 OPTIONS FOR URL REWRITING
4200 -----------------------------------------------------------------------------
4203 NAME: url_rewrite_program redirect_program
4205 LOC: Config.Program.redirect
4208 Specify the location of the executable URL rewriter to use.
4209 Since they can perform almost any function there isn't one included.
4211 For each requested URL, the rewriter will receive on line with the format
4213 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4216 After processing the request the helper must reply using the following format:
4218 [channel-ID <SP>] result [<SP> kv-pairs]
4220 The result code can be:
4222 OK status=30N url="..."
4223 Redirect the URL to the one supplied in 'url='.
4224 'status=' is optional and contains the status code to send
4225 the client in Squids HTTP response. It must be one of the
4226 HTTP redirect status codes: 301, 302, 303, 307, 308.
4227 When no status is given Squid will use 302.
4229 OK rewrite-url="..."
4230 Rewrite the URL to the one supplied in 'rewrite-url='.
4231 The new URL is fetched directly by Squid and returned to
4232 the client as the response to its request.
4235 Do not change the URL.
4238 An internal error occurred in the helper, preventing
4239 a result being identified.
4242 In the future, the interface protocol will be extended with
4243 key=value pairs ("kv-pairs" shown above). Helper programs
4244 should be prepared to receive and possibly ignore additional
4245 whitespace-separated tokens on each input line.
4247 When using the concurrency= option the protocol is changed by
4248 introducing a query channel tag in front of the request/response.
4249 The query channel tag is a number between 0 and concurrency-1.
4250 This value must be echoed back unchanged to Squid as the first part
4251 of the response relating to its request.
4253 WARNING: URL re-writing ability should be avoided whenever possible.
4254 Use the URL redirect form of response instead.
4256 Re-write creates a difference in the state held by the client
4257 and server. Possibly causing confusion when the server response
4258 contains snippets of its view state. Embeded URLs, response
4259 and content Location headers, etc. are not re-written by this
4262 By default, a URL rewriter is not used.
4265 NAME: url_rewrite_children redirect_children
4266 TYPE: HelperChildConfig
4267 DEFAULT: 20 startup=0 idle=1 concurrency=0
4268 LOC: Config.redirectChildren
4270 The maximum number of redirector processes to spawn. If you limit
4271 it too few Squid will have to wait for them to process a backlog of
4272 URLs, slowing it down. If you allow too many they will use RAM
4273 and other system resources noticably.
4275 The startup= and idle= options allow some measure of skew in your
4280 Sets a minimum of how many processes are to be spawned when Squid
4281 starts or reconfigures. When set to zero the first request will
4282 cause spawning of the first child process to handle it.
4284 Starting too few will cause an initial slowdown in traffic as Squid
4285 attempts to simultaneously spawn enough processes to cope.
4289 Sets a minimum of how many processes Squid is to try and keep available
4290 at all times. When traffic begins to rise above what the existing
4291 processes can handle this many more will be spawned up to the maximum
4292 configured. A minimum setting of 1 is required.
4296 The number of requests each redirector helper can handle in
4297 parallel. Defaults to 0 which indicates the redirector
4298 is a old-style single threaded redirector.
4300 When this directive is set to a value >= 1 then the protocol
4301 used to communicate with the helper is modified to include
4302 an ID in front of the request/response. The ID from the request
4303 must be echoed back with the response to that request.
4306 NAME: url_rewrite_host_header redirect_rewrites_host_header
4309 LOC: Config.onoff.redir_rewrites_host
4311 To preserve same-origin security policies in browsers and
4312 prevent Host: header forgery by redirectors Squid rewrites
4313 any Host: header in redirected requests.
4315 If you are running an accelerator this may not be a wanted
4316 effect of a redirector. This directive enables you disable
4317 Host: alteration in reverse-proxy traffic.
4319 WARNING: Entries are cached on the result of the URL rewriting
4320 process, so be careful if you have domain-virtual hosts.
4322 WARNING: Squid and other software verifies the URL and Host
4323 are matching, so be careful not to relay through other proxies
4324 or inspecting firewalls with this disabled.
4327 NAME: url_rewrite_access redirector_access
4330 LOC: Config.accessList.redirector
4332 If defined, this access list specifies which requests are
4333 sent to the redirector processes. By default all requests
4336 This clause supports both fast and slow acl types.
4337 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4340 NAME: url_rewrite_bypass redirector_bypass
4342 LOC: Config.onoff.redirector_bypass
4345 When this is 'on', a request will not go through the
4346 redirector if all redirectors are busy. If this is 'off'
4347 and the redirector queue grows too large, Squid will exit
4348 with a FATAL error and ask you to increase the number of
4349 redirectors. You should only enable this if the redirectors
4350 are not critical to your caching system. If you use
4351 redirectors for access control, and you enable this option,
4352 users may have access to pages they should not
4353 be allowed to request.
4357 OPTIONS FOR TUNING THE CACHE
4358 -----------------------------------------------------------------------------
4361 NAME: cache no_cache
4364 LOC: Config.accessList.noCache
4366 A list of ACL elements which, if matched and denied, cause the request to
4367 not be satisfied from the cache and the reply to not be cached.
4368 In other words, use this to force certain objects to never be cached.
4370 You must use the words 'allow' or 'deny' to indicate whether items
4371 matching the ACL should be allowed or denied into the cache.
4373 Default is to allow all to be cached.
4375 This clause supports both fast and slow acl types.
4376 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4382 LOC: Config.maxStale
4385 This option puts an upper limit on how stale content Squid
4386 will serve from the cache if cache validation fails.
4387 Can be overriden by the refresh_pattern max-stale option.
4390 NAME: refresh_pattern
4391 TYPE: refreshpattern
4395 usage: refresh_pattern [-i] regex min percent max [options]
4397 By default, regular expressions are CASE-SENSITIVE. To make
4398 them case-insensitive, use the -i option.
4400 'Min' is the time (in minutes) an object without an explicit
4401 expiry time should be considered fresh. The recommended
4402 value is 0, any higher values may cause dynamic applications
4403 to be erroneously cached unless the application designer
4404 has taken the appropriate actions.
4406 'Percent' is a percentage of the objects age (time since last
4407 modification age) an object without explicit expiry time
4408 will be considered fresh.
4410 'Max' is an upper limit on how long objects without an explicit
4411 expiry time will be considered fresh.
4413 options: override-expire
4418 ignore-must-revalidate
4425 override-expire enforces min age even if the server
4426 sent an explicit expiry time (e.g., with the
4427 Expires: header or Cache-Control: max-age). Doing this
4428 VIOLATES the HTTP standard. Enabling this feature
4429 could make you liable for problems which it causes.
4431 Note: override-expire does not enforce staleness - it only extends
4432 freshness / min. If the server returns a Expires time which
4433 is longer than your max time, Squid will still consider
4434 the object fresh for that period of time.
4436 override-lastmod enforces min age even on objects
4437 that were modified recently.
4439 reload-into-ims changes client no-cache or ``reload''
4440 to If-Modified-Since requests. Doing this VIOLATES the
4441 HTTP standard. Enabling this feature could make you
4442 liable for problems which it causes.
4444 ignore-reload ignores a client no-cache or ``reload''
4445 header. Doing this VIOLATES the HTTP standard. Enabling
4446 this feature could make you liable for problems which
4449 ignore-no-store ignores any ``Cache-control: no-store''
4450 headers received from a server. Doing this VIOLATES
4451 the HTTP standard. Enabling this feature could make you
4452 liable for problems which it causes.
4454 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4455 headers received from a server. Doing this VIOLATES
4456 the HTTP standard. Enabling this feature could make you
4457 liable for problems which it causes.
4459 ignore-private ignores any ``Cache-control: private''
4460 headers received from a server. Doing this VIOLATES
4461 the HTTP standard. Enabling this feature could make you
4462 liable for problems which it causes.
4464 ignore-auth caches responses to requests with authorization,
4465 as if the originserver had sent ``Cache-control: public''
4466 in the response header. Doing this VIOLATES the HTTP standard.
4467 Enabling this feature could make you liable for problems which
4470 refresh-ims causes squid to contact the origin server
4471 when a client issues an If-Modified-Since request. This
4472 ensures that the client will receive an updated version
4473 if one is available.
4475 store-stale stores responses even if they don't have explicit
4476 freshness or a validator (i.e., Last-Modified or an ETag)
4477 present, or if they're already stale. By default, Squid will
4478 not cache such responses because they usually can't be
4479 reused. Note that such responses will be stale by default.
4481 max-stale=NN provide a maximum staleness factor. Squid won't
4482 serve objects more stale than this even if it failed to
4483 validate the object. Default: use the max_stale global limit.
4485 Basically a cached object is:
4487 FRESH if expires < now, else STALE
4489 FRESH if lm-factor < percent, else STALE
4493 The refresh_pattern lines are checked in the order listed here.
4494 The first entry which matches is used. If none of the entries
4495 match the default will be used.
4497 Note, you must uncomment all the default lines if you want
4498 to change one. The default setting is only active if none is
4503 # Add any of your own refresh_pattern entries above these.
4504 refresh_pattern ^ftp: 1440 20% 10080
4505 refresh_pattern ^gopher: 1440 0% 1440
4506 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4507 refresh_pattern . 0 20% 4320
4511 NAME: quick_abort_min
4515 LOC: Config.quickAbort.min
4518 NAME: quick_abort_max
4522 LOC: Config.quickAbort.max
4525 NAME: quick_abort_pct
4529 LOC: Config.quickAbort.pct
4531 The cache by default continues downloading aborted requests
4532 which are almost completed (less than 16 KB remaining). This
4533 may be undesirable on slow (e.g. SLIP) links and/or very busy
4534 caches. Impatient users may tie up file descriptors and
4535 bandwidth by repeatedly requesting and immediately aborting
4538 When the user aborts a request, Squid will check the
4539 quick_abort values to the amount of data transfered until
4542 If the transfer has less than 'quick_abort_min' KB remaining,
4543 it will finish the retrieval.
4545 If the transfer has more than 'quick_abort_max' KB remaining,
4546 it will abort the retrieval.
4548 If more than 'quick_abort_pct' of the transfer has completed,
4549 it will finish the retrieval.
4551 If you do not want any retrieval to continue after the client
4552 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4555 If you want retrievals to always continue if they are being
4556 cached set 'quick_abort_min' to '-1 KB'.
4559 NAME: read_ahead_gap
4560 COMMENT: buffer-size
4562 LOC: Config.readAheadGap
4565 The amount of data the cache will buffer ahead of what has been
4566 sent to the client when retrieving an object from another server.
4570 IFDEF: USE_HTTP_VIOLATIONS
4573 LOC: Config.negativeTtl
4576 Set the Default Time-to-Live (TTL) for failed requests.
4577 Certain types of failures (such as "connection refused" and
4578 "404 Not Found") are able to be negatively-cached for a short time.
4579 Modern web servers should provide Expires: header, however if they
4580 do not this can provide a minimum TTL.
4581 The default is not to cache errors with unknown expiry details.
4583 Note that this is different from negative caching of DNS lookups.
4585 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4586 this feature could make you liable for problems which it
4590 NAME: positive_dns_ttl
4593 LOC: Config.positiveDnsTtl
4596 Upper limit on how long Squid will cache positive DNS responses.
4597 Default is 6 hours (360 minutes). This directive must be set
4598 larger than negative_dns_ttl.
4601 NAME: negative_dns_ttl
4604 LOC: Config.negativeDnsTtl
4607 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4608 This also sets the lower cache limit on positive lookups.
4609 Minimum value is 1 second, and it is not recommendable to go
4610 much below 10 seconds.
4613 NAME: range_offset_limit
4614 COMMENT: size [acl acl...]
4616 LOC: Config.rangeOffsetLimit
4619 usage: (size) [units] [[!]aclname]
4621 Sets an upper limit on how far (number of bytes) into the file
4622 a Range request may be to cause Squid to prefetch the whole file.
4623 If beyond this limit, Squid forwards the Range request as it is and
4624 the result is NOT cached.
4626 This is to stop a far ahead range request (lets say start at 17MB)
4627 from making Squid fetch the whole object up to that point before
4628 sending anything to the client.
4630 Multiple range_offset_limit lines may be specified, and they will
4631 be searched from top to bottom on each request until a match is found.
4632 The first match found will be used. If no line matches a request, the
4633 default limit of 0 bytes will be used.
4635 'size' is the limit specified as a number of units.
4637 'units' specifies whether to use bytes, KB, MB, etc.
4638 If no units are specified bytes are assumed.
4640 A size of 0 causes Squid to never fetch more than the
4641 client requested. (default)
4643 A size of 'none' causes Squid to always fetch the object from the
4644 beginning so it may cache the result. (2.0 style)
4646 'aclname' is the name of a defined ACL.
4648 NP: Using 'none' as the byte value here will override any quick_abort settings
4649 that may otherwise apply to the range request. The range request will
4650 be fully fetched from start to finish regardless of the client
4651 actions. This affects bandwidth usage.
4654 NAME: minimum_expiry_time
4657 LOC: Config.minimum_expiry_time
4660 The minimum caching time according to (Expires - Date)
4661 Headers Squid honors if the object can't be revalidated
4662 defaults to 60 seconds. In reverse proxy environments it
4663 might be desirable to honor shorter object lifetimes. It
4664 is most likely better to make your server return a
4665 meaningful Last-Modified header however. In ESI environments
4666 where page fragments often have short lifetimes, this will
4667 often be best set to 0.
4670 NAME: store_avg_object_size
4674 LOC: Config.Store.avgObjectSize
4676 Average object size, used to estimate number of objects your
4677 cache can hold. The default is 13 KB.
4680 NAME: store_objects_per_bucket
4683 LOC: Config.Store.objectsPerBucket
4685 Target number of objects per bucket in the store hash table.
4686 Lowering this value increases the total number of buckets and
4687 also the storage maintenance rate. The default is 20.
4692 -----------------------------------------------------------------------------
4695 NAME: request_header_max_size
4699 LOC: Config.maxRequestHeaderSize
4701 This specifies the maximum size for HTTP headers in a request.
4702 Request headers are usually relatively small (about 512 bytes).
4703 Placing a limit on the request header size will catch certain
4704 bugs (for example with persistent connections) and possibly
4705 buffer-overflow or denial-of-service attacks.
4708 NAME: reply_header_max_size
4712 LOC: Config.maxReplyHeaderSize
4714 This specifies the maximum size for HTTP headers in a reply.
4715 Reply headers are usually relatively small (about 512 bytes).
4716 Placing a limit on the reply header size will catch certain
4717 bugs (for example with persistent connections) and possibly
4718 buffer-overflow or denial-of-service attacks.
4721 NAME: request_body_max_size
4725 LOC: Config.maxRequestBodySize
4727 This specifies the maximum size for an HTTP request body.
4728 In other words, the maximum size of a PUT/POST request.
4729 A user who attempts to send a request with a body larger
4730 than this limit receives an "Invalid Request" error message.
4731 If you set this parameter to a zero (the default), there will
4732 be no limit imposed.
4735 NAME: client_request_buffer_max_size
4739 LOC: Config.maxRequestBufferSize
4741 This specifies the maximum buffer size of a client request.
4742 It prevents squid eating too much memory when somebody uploads
4746 NAME: chunked_request_body_max_size
4750 LOC: Config.maxChunkedRequestBodySize
4752 A broken or confused HTTP/1.1 client may send a chunked HTTP
4753 request to Squid. Squid does not have full support for that
4754 feature yet. To cope with such requests, Squid buffers the
4755 entire request and then dechunks request body to create a
4756 plain HTTP/1.0 request with a known content length. The plain
4757 request is then used by the rest of Squid code as usual.
4759 The option value specifies the maximum size of the buffer used
4760 to hold the request before the conversion. If the chunked
4761 request size exceeds the specified limit, the conversion
4762 fails, and the client receives an "unsupported request" error,
4763 as if dechunking was disabled.
4765 Dechunking is enabled by default. To disable conversion of
4766 chunked requests, set the maximum to zero.
4768 Request dechunking feature and this option in particular are a
4769 temporary hack. When chunking requests and responses are fully
4770 supported, there will be no need to buffer a chunked request.
4774 IFDEF: USE_HTTP_VIOLATIONS
4777 LOC: Config.accessList.brokenPosts
4779 A list of ACL elements which, if matched, causes Squid to send
4780 an extra CRLF pair after the body of a PUT/POST request.
4782 Some HTTP servers has broken implementations of PUT/POST,
4783 and rely on an extra CRLF pair sent by some WWW clients.
4785 Quote from RFC2616 section 4.1 on this matter:
4787 Note: certain buggy HTTP/1.0 client implementations generate an
4788 extra CRLF's after a POST request. To restate what is explicitly
4789 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4790 a request with an extra CRLF.
4792 This clause only supports fast acl types.
4793 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4796 acl buggy_server url_regex ^http://....
4797 broken_posts allow buggy_server
4800 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4803 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4805 LOC: Adaptation::Config::use_indirect_client
4807 Controls whether the indirect client IP address (instead of the direct
4808 client IP address) is passed to adaptation services.
4810 See also: follow_x_forwarded_for adaptation_send_client_ip
4814 IFDEF: USE_HTTP_VIOLATIONS
4818 LOC: Config.onoff.via
4820 If set (default), Squid will include a Via header in requests and
4821 replies as required by RFC2616.
4827 LOC: Config.onoff.ie_refresh
4830 Microsoft Internet Explorer up until version 5.5 Service
4831 Pack 1 has an issue with transparent proxies, wherein it
4832 is impossible to force a refresh. Turning this on provides
4833 a partial fix to the problem, by causing all IMS-REFRESH
4834 requests from older IE versions to check the origin server
4835 for fresh content. This reduces hit ratio by some amount
4836 (~10% in my experience), but allows users to actually get
4837 fresh content when they want it. Note because Squid
4838 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4839 of 5.5 is unchanged from old versions of Squid (i.e. a
4840 forced refresh is impossible). Newer versions of IE will,
4841 hopefully, continue to have the new behavior and will be
4842 handled based on that assumption. This option defaults to
4843 the old Squid behavior, which is better for hit ratios but
4844 worse for clients using IE, if they need to be able to
4845 force fresh content.
4848 NAME: vary_ignore_expire
4851 LOC: Config.onoff.vary_ignore_expire
4854 Many HTTP servers supporting Vary gives such objects
4855 immediate expiry time with no cache-control header
4856 when requested by a HTTP/1.0 client. This option
4857 enables Squid to ignore such expiry times until
4858 HTTP/1.1 is fully implemented.
4860 WARNING: If turned on this may eventually cause some
4861 varying objects not intended for caching to get cached.
4864 NAME: request_entities
4866 LOC: Config.onoff.request_entities
4869 Squid defaults to deny GET and HEAD requests with request entities,
4870 as the meaning of such requests are undefined in the HTTP standard
4871 even if not explicitly forbidden.
4873 Set this directive to on if you have clients which insists
4874 on sending request entities in GET or HEAD requests. But be warned
4875 that there is server software (both proxies and web servers) which
4876 can fail to properly process this kind of request which may make you
4877 vulnerable to cache pollution attacks if enabled.
4880 NAME: request_header_access
4881 IFDEF: USE_HTTP_VIOLATIONS
4882 TYPE: http_header_access
4883 LOC: Config.request_header_access
4886 Usage: request_header_access header_name allow|deny [!]aclname ...
4888 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4889 this feature could make you liable for problems which it
4892 This option replaces the old 'anonymize_headers' and the
4893 older 'http_anonymizer' option with something that is much
4894 more configurable. A list of ACLs for each header name allows
4895 removal of specific header fields under specific conditions.
4897 This option only applies to outgoing HTTP request headers (i.e.,
4898 headers sent by Squid to the next HTTP hop such as a cache peer
4899 or an origin server). The option has no effect during cache hit
4900 detection. The equivalent adaptation vectoring point in ICAP
4901 terminology is post-cache REQMOD.
4903 The option is applied to individual outgoing request header
4904 fields. For each request header field F, Squid uses the first
4905 qualifying sets of request_header_access rules:
4907 1. Rules with header_name equal to F's name.
4908 2. Rules with header_name 'Other', provided F's name is not
4909 on the hard-coded list of commonly used HTTP header names.
4910 3. Rules with header_name 'All'.
4912 Within that qualifying rule set, rule ACLs are checked as usual.
4913 If ACLs of an "allow" rule match, the header field is allowed to
4914 go through as is. If ACLs of a "deny" rule match, the header is
4915 removed and request_header_replace is then checked to identify
4916 if the removed header has a replacement. If no rules within the
4917 set have matching ACLs, the header field is left as is.
4919 For example, to achieve the same behavior as the old
4920 'http_anonymizer standard' option, you should use:
4922 request_header_access From deny all
4923 request_header_access Referer deny all
4924 request_header_access Server deny all
4925 request_header_access User-Agent deny all
4926 request_header_access WWW-Authenticate deny all
4927 request_header_access Link deny all
4929 Or, to reproduce the old 'http_anonymizer paranoid' feature
4932 request_header_access Allow allow all
4933 request_header_access Authorization allow all
4934 request_header_access WWW-Authenticate allow all
4935 request_header_access Proxy-Authorization allow all
4936 request_header_access Proxy-Authenticate allow all
4937 request_header_access Cache-Control allow all
4938 request_header_access Content-Encoding allow all
4939 request_header_access Content-Length allow all
4940 request_header_access Content-Type allow all
4941 request_header_access Date allow all
4942 request_header_access Expires allow all
4943 request_header_access Host allow all
4944 request_header_access If-Modified-Since allow all
4945 request_header_access Last-Modified allow all
4946 request_header_access Location allow all
4947 request_header_access Pragma allow all
4948 request_header_access Accept allow all
4949 request_header_access Accept-Charset allow all
4950 request_header_access Accept-Encoding allow all
4951 request_header_access Accept-Language allow all
4952 request_header_access Content-Language allow all
4953 request_header_access Mime-Version allow all
4954 request_header_access Retry-After allow all
4955 request_header_access Title allow all
4956 request_header_access Connection allow all
4957 request_header_access All deny all
4959 although many of those are HTTP reply headers, and so should be
4960 controlled with the reply_header_access directive.
4962 By default, all headers are allowed (no anonymizing is
4966 NAME: reply_header_access
4967 IFDEF: USE_HTTP_VIOLATIONS
4968 TYPE: http_header_access
4969 LOC: Config.reply_header_access
4972 Usage: reply_header_access header_name allow|deny [!]aclname ...
4974 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4975 this feature could make you liable for problems which it
4978 This option only applies to reply headers, i.e., from the
4979 server to the client.
4981 This is the same as request_header_access, but in the other
4982 direction. Please see request_header_access for detailed
4985 For example, to achieve the same behavior as the old
4986 'http_anonymizer standard' option, you should use:
4988 reply_header_access From deny all
4989 reply_header_access Referer deny all
4990 reply_header_access Server deny all
4991 reply_header_access User-Agent deny all
4992 reply_header_access WWW-Authenticate deny all
4993 reply_header_access Link deny all
4995 Or, to reproduce the old 'http_anonymizer paranoid' feature
4998 reply_header_access Allow allow all
4999 reply_header_access Authorization allow all
5000 reply_header_access WWW-Authenticate allow all
5001 reply_header_access Proxy-Authorization allow all
5002 reply_header_access Proxy-Authenticate allow all
5003 reply_header_access Cache-Control allow all
5004 reply_header_access Content-Encoding allow all
5005 reply_header_access Content-Length allow all
5006 reply_header_access Content-Type allow all
5007 reply_header_access Date allow all
5008 reply_header_access Expires allow all
5009 reply_header_access Host allow all
5010 reply_header_access If-Modified-Since allow all
5011 reply_header_access Last-Modified allow all
5012 reply_header_access Location allow all
5013 reply_header_access Pragma allow all
5014 reply_header_access Accept allow all
5015 reply_header_access Accept-Charset allow all
5016 reply_header_access Accept-Encoding allow all
5017 reply_header_access Accept-Language allow all
5018 reply_header_access Content-Language allow all
5019 reply_header_access Mime-Version allow all
5020 reply_header_access Retry-After allow all
5021 reply_header_access Title allow all
5022 reply_header_access Connection allow all
5023 reply_header_access All deny all
5025 although the HTTP request headers won't be usefully controlled
5026 by this directive -- see request_header_access for details.
5028 By default, all headers are allowed (no anonymizing is
5032 NAME: request_header_replace header_replace
5033 IFDEF: USE_HTTP_VIOLATIONS
5034 TYPE: http_header_replace
5035 LOC: Config.request_header_access
5038 Usage: request_header_replace header_name message
5039 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5041 This option allows you to change the contents of headers
5042 denied with request_header_access above, by replacing them
5043 with some fixed string. This replaces the old fake_user_agent
5046 This only applies to request headers, not reply headers.
5048 By default, headers are removed if denied.
5051 NAME: reply_header_replace
5052 IFDEF: USE_HTTP_VIOLATIONS
5053 TYPE: http_header_replace
5054 LOC: Config.reply_header_access
5057 Usage: reply_header_replace header_name message
5058 Example: reply_header_replace Server Foo/1.0
5060 This option allows you to change the contents of headers
5061 denied with reply_header_access above, by replacing them
5062 with some fixed string.
5064 This only applies to reply headers, not request headers.
5066 By default, headers are removed if denied.
5069 NAME: request_header_add
5070 TYPE: HeaderWithAclList
5071 LOC: Config.request_header_add
5074 Usage: request_header_add field-name field-value acl1 [acl2] ...
5075 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5077 This option adds header fields to outgoing HTTP requests (i.e.,
5078 request headers sent by Squid to the next HTTP hop such as a
5079 cache peer or an origin server). The option has no effect during
5080 cache hit detection. The equivalent adaptation vectoring point
5081 in ICAP terminology is post-cache REQMOD.
5083 Field-name is a token specifying an HTTP header name. If a
5084 standard HTTP header name is used, Squid does not check whether
5085 the new header conflicts with any existing headers or violates
5086 HTTP rules. If the request to be modified already contains a
5087 field with the same name, the old field is preserved but the
5088 header field values are not merged.
5090 Field-value is either a token or a quoted string. If quoted
5091 string format is used, then the surrounding quotes are removed
5092 while escape sequences and %macros are processed.
5094 In theory, all of the logformat codes can be used as %macros.
5095 However, unlike logging (which happens at the very end of
5096 transaction lifetime), the transaction may not yet have enough
5097 information to expand a macro when the new header value is needed.
5098 And some information may already be available to Squid but not yet
5099 committed where the macro expansion code can access it (report
5100 such instances!). The macro will be expanded into a single dash
5101 ('-') in such cases. Not all macros have been tested.
5103 One or more Squid ACLs may be specified to restrict header
5104 injection to matching requests. As always in squid.conf, all
5105 ACLs in an option ACL list must be satisfied for the insertion
5106 to happen. The request_header_add option supports fast ACLs
5115 This option used to log custom information about the master
5116 transaction. For example, an admin may configure Squid to log
5117 which "user group" the transaction belongs to, where "user group"
5118 will be determined based on a set of ACLs and not [just]
5119 authentication information.
5120 Values of key/value pairs can be logged using %{key}note macros:
5122 note key value acl ...
5123 logformat myFormat ... %{key}note ...
5126 NAME: relaxed_header_parser
5127 COMMENT: on|off|warn
5129 LOC: Config.onoff.relaxed_header_parser
5132 In the default "on" setting Squid accepts certain forms
5133 of non-compliant HTTP messages where it is unambiguous
5134 what the sending application intended even if the message
5135 is not correctly formatted. The messages is then normalized
5136 to the correct form when forwarded by Squid.
5138 If set to "warn" then a warning will be emitted in cache.log
5139 each time such HTTP error is encountered.
5141 If set to "off" then such HTTP errors will cause the request
5142 or response to be rejected.
5147 -----------------------------------------------------------------------------
5150 NAME: forward_timeout
5153 LOC: Config.Timeout.forward
5156 This parameter specifies how long Squid should at most attempt in
5157 finding a forwarding path for the request before giving up.
5160 NAME: connect_timeout
5163 LOC: Config.Timeout.connect
5166 This parameter specifies how long to wait for the TCP connect to
5167 the requested server or peer to complete before Squid should
5168 attempt to find another path where to forward the request.
5171 NAME: peer_connect_timeout
5174 LOC: Config.Timeout.peer_connect
5177 This parameter specifies how long to wait for a pending TCP
5178 connection to a peer cache. The default is 30 seconds. You
5179 may also set different timeout values for individual neighbors
5180 with the 'connect-timeout' option on a 'cache_peer' line.
5186 LOC: Config.Timeout.read
5189 The read_timeout is applied on server-side connections. After
5190 each successful read(), the timeout will be extended by this
5191 amount. If no data is read again after this amount of time,
5192 the request is aborted and logged with ERR_READ_TIMEOUT. The
5193 default is 15 minutes.
5199 LOC: Config.Timeout.write
5202 This timeout is tracked for all connections that have data
5203 available for writing and are waiting for the socket to become
5204 ready. After each successful write, the timeout is extended by
5205 the configured amount. If Squid has data to write but the
5206 connection is not ready for the configured duration, the
5207 transaction associated with the connection is terminated. The
5208 default is 15 minutes.
5211 NAME: request_timeout
5213 LOC: Config.Timeout.request
5216 How long to wait for complete HTTP request headers after initial
5217 connection establishment.
5220 NAME: client_idle_pconn_timeout persistent_request_timeout
5222 LOC: Config.Timeout.clientIdlePconn
5225 How long to wait for the next HTTP request on a persistent
5226 client connection after the previous request completes.
5229 NAME: client_lifetime
5232 LOC: Config.Timeout.lifetime
5235 The maximum amount of time a client (browser) is allowed to
5236 remain connected to the cache process. This protects the Cache
5237 from having a lot of sockets (and hence file descriptors) tied up
5238 in a CLOSE_WAIT state from remote clients that go away without
5239 properly shutting down (either because of a network failure or
5240 because of a poor client implementation). The default is one
5243 NOTE: The default value is intended to be much larger than any
5244 client would ever need to be connected to your cache. You
5245 should probably change client_lifetime only as a last resort.
5246 If you seem to have many client connections tying up
5247 filedescriptors, we recommend first tuning the read_timeout,
5248 request_timeout, persistent_request_timeout and quick_abort values.
5251 NAME: half_closed_clients
5253 LOC: Config.onoff.half_closed_clients
5256 Some clients may shutdown the sending side of their TCP
5257 connections, while leaving their receiving sides open. Sometimes,
5258 Squid can not tell the difference between a half-closed and a
5259 fully-closed TCP connection.
5261 By default, Squid will immediately close client connections when
5262 read(2) returns "no more data to read."
5264 Change this option to 'on' and Squid will keep open connections
5265 until a read(2) or write(2) on the socket returns an error.
5266 This may show some benefits for reverse proxies. But if not
5267 it is recommended to leave OFF.
5270 NAME: server_idle_pconn_timeout pconn_timeout
5272 LOC: Config.Timeout.serverIdlePconn
5275 Timeout for idle persistent connections to servers and other
5282 LOC: Ident::TheConfig.timeout
5285 Maximum time to wait for IDENT lookups to complete.
5287 If this is too high, and you enabled IDENT lookups from untrusted
5288 users, you might be susceptible to denial-of-service by having
5289 many ident requests going at once.
5292 NAME: shutdown_lifetime
5295 LOC: Config.shutdownLifetime
5298 When SIGTERM or SIGHUP is received, the cache is put into
5299 "shutdown pending" mode until all active sockets are closed.
5300 This value is the lifetime to set for all open descriptors
5301 during shutdown mode. Any active clients after this many
5302 seconds will receive a 'timeout' message.
5306 ADMINISTRATIVE PARAMETERS
5307 -----------------------------------------------------------------------------
5313 LOC: Config.adminEmail
5315 Email-address of local cache manager who will receive
5316 mail if the cache dies. The default is "webmaster."
5322 LOC: Config.EmailFrom
5324 From: email-address for mail sent when the cache dies.
5325 The default is to use 'appname@unique_hostname'.
5326 Default appname value is "squid", can be changed into
5327 src/globals.h before building squid.
5333 LOC: Config.EmailProgram
5335 Email program used to send mail if the cache dies.
5336 The default is "mail". The specified program must comply
5337 with the standard Unix mail syntax:
5338 mail-program recipient < mailfile
5340 Optional command line options can be specified.
5343 NAME: cache_effective_user
5345 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5346 LOC: Config.effectiveUser
5348 If you start Squid as root, it will change its effective/real
5349 UID/GID to the user specified below. The default is to change
5350 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5351 see also; cache_effective_group
5354 NAME: cache_effective_group
5357 LOC: Config.effectiveGroup
5359 Squid sets the GID to the effective user's default group ID
5360 (taken from the password file) and supplementary group list
5361 from the groups membership.
5363 If you want Squid to run with a specific GID regardless of
5364 the group memberships of the effective user then set this
5365 to the group (or GID) you want Squid to run as. When set
5366 all other group privileges of the effective user are ignored
5367 and only this GID is effective. If Squid is not started as
5368 root the user starting Squid MUST be member of the specified
5371 This option is not recommended by the Squid Team.
5372 Our preference is for administrators to configure a secure
5373 user account for squid with UID/GID matching system policies.
5376 NAME: httpd_suppress_version_string
5380 LOC: Config.onoff.httpd_suppress_version_string
5382 Suppress Squid version string info in HTTP headers and HTML error pages.
5385 NAME: visible_hostname
5387 LOC: Config.visibleHostname
5390 If you want to present a special hostname in error messages, etc,
5391 define this. Otherwise, the return value of gethostname()
5392 will be used. If you have multiple caches in a cluster and
5393 get errors about IP-forwarding you must set them to have individual
5394 names with this setting.
5397 NAME: unique_hostname
5399 LOC: Config.uniqueHostname
5402 If you want to have multiple machines with the same
5403 'visible_hostname' you must give each machine a different
5404 'unique_hostname' so forwarding loops can be detected.
5407 NAME: hostname_aliases
5409 LOC: Config.hostnameAliases
5412 A list of other DNS names your cache has.
5420 Minimum umask which should be enforced while the proxy
5421 is running, in addition to the umask set at startup.
5423 For a traditional octal representation of umasks, start
5428 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5429 -----------------------------------------------------------------------------
5431 This section contains parameters for the (optional) cache
5432 announcement service. This service is provided to help
5433 cache administrators locate one another in order to join or
5434 create cache hierarchies.
5436 An 'announcement' message is sent (via UDP) to the registration
5437 service by Squid. By default, the announcement message is NOT
5438 SENT unless you enable it with 'announce_period' below.
5440 The announcement message includes your hostname, plus the
5441 following information from this configuration file:
5447 All current information is processed regularly and made
5448 available on the Web at http://www.ircache.net/Cache/Tracker/.
5451 NAME: announce_period
5453 LOC: Config.Announce.period
5456 This is how frequently to send cache announcements. The
5457 default is `0' which disables sending the announcement
5460 To enable announcing your cache, just set an announce period.
5463 announce_period 1 day
5468 DEFAULT: tracker.ircache.net
5469 LOC: Config.Announce.host
5475 LOC: Config.Announce.file
5481 LOC: Config.Announce.port
5483 announce_host and announce_port set the hostname and port
5484 number where the registration message will be sent.
5486 Hostname will default to 'tracker.ircache.net' and port will
5487 default default to 3131. If the 'filename' argument is given,
5488 the contents of that file will be included in the announce
5493 HTTPD-ACCELERATOR OPTIONS
5494 -----------------------------------------------------------------------------
5497 NAME: httpd_accel_surrogate_id
5500 LOC: Config.Accel.surrogate_id
5502 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5503 need an identification token to allow control targeting. Because
5504 a farm of surrogates may all perform the same tasks, they may share
5505 an identification token.
5507 The default ID is the visible_hostname
5510 NAME: http_accel_surrogate_remote
5514 LOC: Config.onoff.surrogate_is_remote
5516 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
5517 Set this to on to have squid behave as a remote surrogate.
5521 IFDEF: USE_SQUID_ESI
5522 COMMENT: libxml2|expat|custom
5524 LOC: ESIParser::Type
5527 ESI markup is not strictly XML compatible. The custom ESI parser
5528 will give higher performance, but cannot handle non ASCII character
5533 DELAY POOL PARAMETERS
5534 -----------------------------------------------------------------------------
5538 TYPE: delay_pool_count
5540 IFDEF: USE_DELAY_POOLS
5543 This represents the number of delay pools to be used. For example,
5544 if you have one class 2 delay pool and one class 3 delays pool, you
5545 have a total of 2 delay pools.
5549 TYPE: delay_pool_class
5551 IFDEF: USE_DELAY_POOLS
5554 This defines the class of each delay pool. There must be exactly one
5555 delay_class line for each delay pool. For example, to define two
5556 delay pools, one of class 2 and one of class 3, the settings above
5560 delay_pools 4 # 4 delay pools
5561 delay_class 1 2 # pool 1 is a class 2 pool
5562 delay_class 2 3 # pool 2 is a class 3 pool
5563 delay_class 3 4 # pool 3 is a class 4 pool
5564 delay_class 4 5 # pool 4 is a class 5 pool
5566 The delay pool classes are:
5568 class 1 Everything is limited by a single aggregate
5571 class 2 Everything is limited by a single aggregate
5572 bucket as well as an "individual" bucket chosen
5573 from bits 25 through 32 of the IPv4 address.
5575 class 3 Everything is limited by a single aggregate
5576 bucket as well as a "network" bucket chosen
5577 from bits 17 through 24 of the IP address and a
5578 "individual" bucket chosen from bits 17 through
5579 32 of the IPv4 address.
5581 class 4 Everything in a class 3 delay pool, with an
5582 additional limit on a per user basis. This
5583 only takes effect if the username is established
5584 in advance - by forcing authentication in your
5587 class 5 Requests are grouped according their tag (see
5588 external_acl's tag= reply).
5591 Each pool also requires a delay_parameters directive to configure the pool size
5592 and speed limits used whenever the pool is applied to a request. Along with
5593 a set of delay_access directives to determine when it is used.
5595 NOTE: If an IP address is a.b.c.d
5596 -> bits 25 through 32 are "d"
5597 -> bits 17 through 24 are "c"
5598 -> bits 17 through 32 are "c * 256 + d"
5600 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5601 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5605 TYPE: delay_pool_access
5607 IFDEF: USE_DELAY_POOLS
5610 This is used to determine which delay pool a request falls into.
5612 delay_access is sorted per pool and the matching starts with pool 1,
5613 then pool 2, ..., and finally pool N. The first delay pool where the
5614 request is allowed is selected for the request. If it does not allow
5615 the request to any pool then the request is not delayed (default).
5617 For example, if you want some_big_clients in delay
5618 pool 1 and lotsa_little_clients in delay pool 2:
5621 delay_access 1 allow some_big_clients
5622 delay_access 1 deny all
5623 delay_access 2 allow lotsa_little_clients
5624 delay_access 2 deny all
5625 delay_access 3 allow authenticated_clients
5628 NAME: delay_parameters
5629 TYPE: delay_pool_rates
5631 IFDEF: USE_DELAY_POOLS
5634 This defines the parameters for a delay pool. Each delay pool has
5635 a number of "buckets" associated with it, as explained in the
5636 description of delay_class.
5638 For a class 1 delay pool, the syntax is:
5640 delay_parameters pool aggregate
5642 For a class 2 delay pool:
5644 delay_parameters pool aggregate individual
5646 For a class 3 delay pool:
5648 delay_parameters pool aggregate network individual
5650 For a class 4 delay pool:
5652 delay_parameters pool aggregate network individual user
5654 For a class 5 delay pool:
5656 delay_parameters pool tagrate
5658 The option variables are:
5660 pool a pool number - ie, a number between 1 and the
5661 number specified in delay_pools as used in
5664 aggregate the speed limit parameters for the aggregate bucket
5667 individual the speed limit parameters for the individual
5668 buckets (class 2, 3).
5670 network the speed limit parameters for the network buckets
5673 user the speed limit parameters for the user buckets
5676 tagrate the speed limit parameters for the tag buckets
5679 A pair of delay parameters is written restore/maximum, where restore is
5680 the number of bytes (not bits - modem and network speeds are usually
5681 quoted in bits) per second placed into the bucket, and maximum is the
5682 maximum number of bytes which can be in the bucket at any time.
5684 There must be one delay_parameters line for each delay pool.
5687 For example, if delay pool number 1 is a class 2 delay pool as in the
5688 above example, and is being used to strictly limit each host to 64Kbit/sec
5689 (plus overheads), with no overall limit, the line is:
5691 delay_parameters 1 -1/-1 8000/8000
5693 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5695 Note that the figure -1 is used to represent "unlimited".
5698 And, if delay pool number 2 is a class 3 delay pool as in the above
5699 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5700 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5701 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5702 to permit a decent web page to be downloaded at a decent speed
5703 (if the network is not being limited due to overuse) but slow down
5704 large downloads more significantly:
5706 delay_parameters 2 32000/32000 8000/8000 600/8000
5708 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5709 8 x 8000 KByte/sec -> 64Kbit/sec.
5710 8 x 600 Byte/sec -> 4800bit/sec.
5713 Finally, for a class 4 delay pool as in the example - each user will
5714 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5716 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5719 NAME: delay_initial_bucket_level
5720 COMMENT: (percent, 0-100)
5723 IFDEF: USE_DELAY_POOLS
5724 LOC: Config.Delay.initial
5726 The initial bucket percentage is used to determine how much is put
5727 in each bucket when squid starts, is reconfigured, or first notices
5728 a host accessing it (in class 2 and class 3, individual hosts and
5729 networks only have buckets associated with them once they have been
5734 CLIENT DELAY POOL PARAMETERS
5735 -----------------------------------------------------------------------------
5738 NAME: client_delay_pools
5739 TYPE: client_delay_pool_count
5741 IFDEF: USE_DELAY_POOLS
5742 LOC: Config.ClientDelay
5744 This option specifies the number of client delay pools used. It must
5745 preceed other client_delay_* options.
5748 client_delay_pools 2
5751 NAME: client_delay_initial_bucket_level
5752 COMMENT: (percent, 0-no_limit)
5755 IFDEF: USE_DELAY_POOLS
5756 LOC: Config.ClientDelay.initial
5758 This option determines the initial bucket size as a percentage of
5759 max_bucket_size from client_delay_parameters. Buckets are created
5760 at the time of the "first" connection from the matching IP. Idle
5761 buckets are periodically deleted up.
5763 You can specify more than 100 percent but note that such "oversized"
5764 buckets are not refilled until their size goes down to max_bucket_size
5765 from client_delay_parameters.
5768 client_delay_initial_bucket_level 50
5771 NAME: client_delay_parameters
5772 TYPE: client_delay_pool_rates
5774 IFDEF: USE_DELAY_POOLS
5775 LOC: Config.ClientDelay
5778 This option configures client-side bandwidth limits using the
5781 client_delay_parameters pool speed_limit max_bucket_size
5783 pool is an integer ID used for client_delay_access matching.
5785 speed_limit is bytes added to the bucket per second.
5787 max_bucket_size is the maximum size of a bucket, enforced after any
5788 speed_limit additions.
5790 Please see the delay_parameters option for more information and
5794 client_delay_parameters 1 1024 2048
5795 client_delay_parameters 2 51200 16384
5798 NAME: client_delay_access
5799 TYPE: client_delay_pool_access
5801 IFDEF: USE_DELAY_POOLS
5802 LOC: Config.ClientDelay
5805 This option determines the client-side delay pool for the
5808 client_delay_access pool_ID allow|deny acl_name
5810 All client_delay_access options are checked in their pool ID
5811 order, starting with pool 1. The first checked pool with allowed
5812 request is selected for the request. If no ACL matches or there
5813 are no client_delay_access options, the request bandwidth is not
5816 The ACL-selected pool is then used to find the
5817 client_delay_parameters for the request. Client-side pools are
5818 not used to aggregate clients. Clients are always aggregated
5819 based on their source IP addresses (one bucket per source IP).
5821 Please see delay_access for more examples.
5824 client_delay_access 1 allow low_rate_network
5825 client_delay_access 2 allow vips_network
5829 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5830 -----------------------------------------------------------------------------
5835 LOC: Config.Wccp.router
5839 Use this option to define your WCCP ``home'' router for
5842 wccp_router supports a single WCCP(v1) router
5844 wccp2_router supports multiple WCCPv2 routers
5846 only one of the two may be used at the same time and defines
5847 which version of WCCP to use.
5851 TYPE: IpAddress_list
5852 LOC: Config.Wccp2.router
5856 Use this option to define your WCCP ``home'' router for
5859 wccp_router supports a single WCCP(v1) router
5861 wccp2_router supports multiple WCCPv2 routers
5863 only one of the two may be used at the same time and defines
5864 which version of WCCP to use.
5869 LOC: Config.Wccp.version
5873 This directive is only relevant if you need to set up WCCP(v1)
5874 to some very old and end-of-life Cisco routers. In all other
5875 setups it must be left unset or at the default setting.
5876 It defines an internal version in the WCCP(v1) protocol,
5877 with version 4 being the officially documented protocol.
5879 According to some users, Cisco IOS 11.2 and earlier only
5880 support WCCP version 3. If you're using that or an earlier
5881 version of IOS, you may need to change this value to 3, otherwise
5882 do not specify this parameter.
5885 NAME: wccp2_rebuild_wait
5887 LOC: Config.Wccp2.rebuildwait
5891 If this is enabled Squid will wait for the cache dir rebuild to finish
5892 before sending the first wccp2 HereIAm packet
5895 NAME: wccp2_forwarding_method
5897 LOC: Config.Wccp2.forwarding_method
5901 WCCP2 allows the setting of forwarding methods between the
5902 router/switch and the cache. Valid values are as follows:
5904 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5905 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5907 Currently (as of IOS 12.4) cisco routers only support GRE.
5908 Cisco switches only support the L2 redirect assignment method.
5911 NAME: wccp2_return_method
5913 LOC: Config.Wccp2.return_method
5917 WCCP2 allows the setting of return methods between the
5918 router/switch and the cache for packets that the cache
5919 decides not to handle. Valid values are as follows:
5921 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5922 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5924 Currently (as of IOS 12.4) cisco routers only support GRE.
5925 Cisco switches only support the L2 redirect assignment.
5927 If the "ip wccp redirect exclude in" command has been
5928 enabled on the cache interface, then it is still safe for
5929 the proxy server to use a l2 redirect method even if this
5930 option is set to GRE.
5933 NAME: wccp2_assignment_method
5935 LOC: Config.Wccp2.assignment_method
5939 WCCP2 allows the setting of methods to assign the WCCP hash
5940 Valid values are as follows:
5942 hash - Hash assignment
5943 mask - Mask assignment
5945 As a general rule, cisco routers support the hash assignment method
5946 and cisco switches support the mask assignment method.
5951 LOC: Config.Wccp2.info
5952 DEFAULT_IF_NONE: standard 0
5955 WCCP2 allows for multiple traffic services. There are two
5956 types: "standard" and "dynamic". The standard type defines
5957 one service id - http (id 0). The dynamic service ids can be from
5958 51 to 255 inclusive. In order to use a dynamic service id
5959 one must define the type of traffic to be redirected; this is done
5960 using the wccp2_service_info option.
5962 The "standard" type does not require a wccp2_service_info option,
5963 just specifying the service id will suffice.
5965 MD5 service authentication can be enabled by adding
5966 "password=<password>" to the end of this service declaration.
5970 wccp2_service standard 0 # for the 'web-cache' standard service
5971 wccp2_service dynamic 80 # a dynamic service type which will be
5972 # fleshed out with subsequent options.
5973 wccp2_service standard 0 password=foo
5976 NAME: wccp2_service_info
5977 TYPE: wccp2_service_info
5978 LOC: Config.Wccp2.info
5982 Dynamic WCCPv2 services require further information to define the
5983 traffic you wish to have diverted.
5987 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
5988 priority=<priority> ports=<port>,<port>..
5990 The relevant WCCPv2 flags:
5991 + src_ip_hash, dst_ip_hash
5992 + source_port_hash, dst_port_hash
5993 + src_ip_alt_hash, dst_ip_alt_hash
5994 + src_port_alt_hash, dst_port_alt_hash
5997 The port list can be one to eight entries.
6001 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
6002 priority=240 ports=80
6004 Note: the service id must have been defined by a previous
6005 'wccp2_service dynamic <id>' entry.
6010 LOC: Config.Wccp2.weight
6014 Each cache server gets assigned a set of the destination
6015 hash proportional to their weight.
6020 LOC: Config.Wccp.address
6027 LOC: Config.Wccp2.address
6031 Use this option if you require WCCP to use a specific
6034 The default behavior is to not bind to any specific address.
6038 PERSISTENT CONNECTION HANDLING
6039 -----------------------------------------------------------------------------
6041 Also see "pconn_timeout" in the TIMEOUTS section
6044 NAME: client_persistent_connections
6046 LOC: Config.onoff.client_pconns
6050 NAME: server_persistent_connections
6052 LOC: Config.onoff.server_pconns
6055 Persistent connection support for clients and servers. By
6056 default, Squid uses persistent connections (when allowed)
6057 with its clients and servers. You can use these options to
6058 disable persistent connections with clients and/or servers.
6061 NAME: persistent_connection_after_error
6063 LOC: Config.onoff.error_pconns
6066 With this directive the use of persistent connections after
6067 HTTP errors can be disabled. Useful if you have clients
6068 who fail to handle errors on persistent connections proper.
6071 NAME: detect_broken_pconn
6073 LOC: Config.onoff.detect_broken_server_pconns
6076 Some servers have been found to incorrectly signal the use
6077 of HTTP/1.0 persistent connections even on replies not
6078 compatible, causing significant delays. This server problem
6079 has mostly been seen on redirects.
6081 By enabling this directive Squid attempts to detect such
6082 broken replies and automatically assume the reply is finished
6083 after 10 seconds timeout.
6087 CACHE DIGEST OPTIONS
6088 -----------------------------------------------------------------------------
6091 NAME: digest_generation
6092 IFDEF: USE_CACHE_DIGESTS
6094 LOC: Config.onoff.digest_generation
6097 This controls whether the server will generate a Cache Digest
6098 of its contents. By default, Cache Digest generation is
6099 enabled if Squid is compiled with --enable-cache-digests defined.
6102 NAME: digest_bits_per_entry
6103 IFDEF: USE_CACHE_DIGESTS
6105 LOC: Config.digest.bits_per_entry
6108 This is the number of bits of the server's Cache Digest which
6109 will be associated with the Digest entry for a given HTTP
6110 Method and URL (public key) combination. The default is 5.
6113 NAME: digest_rebuild_period
6114 IFDEF: USE_CACHE_DIGESTS
6117 LOC: Config.digest.rebuild_period
6120 This is the wait time between Cache Digest rebuilds.
6123 NAME: digest_rewrite_period
6125 IFDEF: USE_CACHE_DIGESTS
6127 LOC: Config.digest.rewrite_period
6130 This is the wait time between Cache Digest writes to
6134 NAME: digest_swapout_chunk_size
6137 IFDEF: USE_CACHE_DIGESTS
6138 LOC: Config.digest.swapout_chunk_size
6141 This is the number of bytes of the Cache Digest to write to
6142 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6146 NAME: digest_rebuild_chunk_percentage
6147 COMMENT: (percent, 0-100)
6148 IFDEF: USE_CACHE_DIGESTS
6150 LOC: Config.digest.rebuild_chunk_percentage
6153 This is the percentage of the Cache Digest to be scanned at a
6154 time. By default it is set to 10% of the Cache Digest.
6159 -----------------------------------------------------------------------------
6164 LOC: Config.Port.snmp
6168 The port number where Squid listens for SNMP requests. To enable
6169 SNMP support set this to a suitable port number. Port number
6170 3401 is often used for the Squid SNMP agent. By default it's
6171 set to "0" (disabled)
6179 LOC: Config.accessList.snmp
6180 DEFAULT_IF_NONE: deny all
6183 Allowing or denying access to the SNMP port.
6185 All access to the agent is denied by default.
6188 snmp_access allow|deny [!]aclname ...
6190 This clause only supports fast acl types.
6191 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6193 snmp_access allow snmppublic localhost
6194 snmp_access deny all
6197 NAME: snmp_incoming_address
6199 LOC: Config.Addrs.snmp_incoming
6204 NAME: snmp_outgoing_address
6206 LOC: Config.Addrs.snmp_outgoing
6210 Just like 'udp_incoming_address', but for the SNMP port.
6212 snmp_incoming_address is used for the SNMP socket receiving
6213 messages from SNMP agents.
6214 snmp_outgoing_address is used for SNMP packets returned to SNMP
6217 The default snmp_incoming_address is to listen on all
6218 available network interfaces.
6220 If snmp_outgoing_address is not set it will use the same socket
6221 as snmp_incoming_address. Only change this if you want to have
6222 SNMP replies sent using another address than where this Squid
6223 listens for SNMP queries.
6225 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6226 the same value since they both use port 3401.
6231 -----------------------------------------------------------------------------
6234 NAME: icp_port udp_port
6237 LOC: Config.Port.icp
6239 The port number where Squid sends and receives ICP queries to
6240 and from neighbor caches. The standard UDP port for ICP is 3130.
6241 Default is disabled (0).
6244 icp_port @DEFAULT_ICP_PORT@
6251 LOC: Config.Port.htcp
6253 The port number where Squid sends and receives HTCP queries to
6254 and from neighbor caches. To turn it on you want to set it to
6255 4827. By default it is set to "0" (disabled).
6261 NAME: log_icp_queries
6265 LOC: Config.onoff.log_udp
6267 If set, ICP queries are logged to access.log. You may wish
6268 do disable this if your ICP load is VERY high to speed things
6269 up or to simplify log analysis.
6272 NAME: udp_incoming_address
6274 LOC:Config.Addrs.udp_incoming
6277 udp_incoming_address is used for UDP packets received from other
6280 The default behavior is to not bind to any specific address.
6282 Only change this if you want to have all UDP queries received on
6283 a specific interface/address.
6285 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6286 modules. Altering it will affect all of them in the same manner.
6288 see also; udp_outgoing_address
6290 NOTE, udp_incoming_address and udp_outgoing_address can not
6291 have the same value since they both use the same port.
6294 NAME: udp_outgoing_address
6296 LOC: Config.Addrs.udp_outgoing
6299 udp_outgoing_address is used for UDP packets sent out to other
6302 The default behavior is to not bind to any specific address.
6304 Instead it will use the same socket as udp_incoming_address.
6305 Only change this if you want to have UDP queries sent using another
6306 address than where this Squid listens for UDP queries from other
6309 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6310 modules. Altering it will affect all of them in the same manner.
6312 see also; udp_incoming_address
6314 NOTE, udp_incoming_address and udp_outgoing_address can not
6315 have the same value since they both use the same port.
6322 LOC: Config.onoff.icp_hit_stale
6324 If you want to return ICP_HIT for stale cache objects, set this
6325 option to 'on'. If you have sibling relationships with caches
6326 in other administrative domains, this should be 'off'. If you only
6327 have sibling relationships with caches under your control,
6328 it is probably okay to set this to 'on'.
6329 If set to 'on', your siblings should use the option "allow-miss"
6330 on their cache_peer lines for connecting to you.
6333 NAME: minimum_direct_hops
6336 LOC: Config.minDirectHops
6338 If using the ICMP pinging stuff, do direct fetches for sites
6339 which are no more than this many hops away.
6342 NAME: minimum_direct_rtt
6345 LOC: Config.minDirectRtt
6347 If using the ICMP pinging stuff, do direct fetches for sites
6348 which are no more than this many rtt milliseconds away.
6354 LOC: Config.Netdb.low
6360 LOC: Config.Netdb.high
6362 The low and high water marks for the ICMP measurement
6363 database. These are counts, not percents. The defaults are
6364 900 and 1000. When the high water mark is reached, database
6365 entries will be deleted until the low mark is reached.
6368 NAME: netdb_ping_period
6370 LOC: Config.Netdb.period
6373 The minimum period for measuring a site. There will be at
6374 least this much delay between successive pings to the same
6375 network. The default is five minutes.
6382 LOC: Config.onoff.query_icmp
6384 If you want to ask your peers to include ICMP data in their ICP
6385 replies, enable this option.
6387 If your peer has configured Squid (during compilation) with
6388 '--enable-icmp' that peer will send ICMP pings to origin server
6389 sites of the URLs it receives. If you enable this option the
6390 ICP replies from that peer will include the ICMP data (if available).
6391 Then, when choosing a parent cache, Squid will choose the parent with
6392 the minimal RTT to the origin server. When this happens, the
6393 hierarchy field of the access.log will be
6394 "CLOSEST_PARENT_MISS". This option is off by default.
6397 NAME: test_reachability
6401 LOC: Config.onoff.test_reachability
6403 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6404 instead of ICP_MISS if the target host is NOT in the ICMP
6405 database, or has a zero RTT.
6408 NAME: icp_query_timeout
6412 LOC: Config.Timeout.icp_query
6414 Normally Squid will automatically determine an optimal ICP
6415 query timeout value based on the round-trip-time of recent ICP
6416 queries. If you want to override the value determined by
6417 Squid, set this 'icp_query_timeout' to a non-zero value. This
6418 value is specified in MILLISECONDS, so, to use a 2-second
6419 timeout (the old default), you would write:
6421 icp_query_timeout 2000
6424 NAME: maximum_icp_query_timeout
6428 LOC: Config.Timeout.icp_query_max
6430 Normally the ICP query timeout is determined dynamically. But
6431 sometimes it can lead to very large values (say 5 seconds).
6432 Use this option to put an upper limit on the dynamic timeout
6433 value. Do NOT use this option to always use a fixed (instead
6434 of a dynamic) timeout value. To set a fixed timeout see the
6435 'icp_query_timeout' directive.
6438 NAME: minimum_icp_query_timeout
6442 LOC: Config.Timeout.icp_query_min
6444 Normally the ICP query timeout is determined dynamically. But
6445 sometimes it can lead to very small timeouts, even lower than
6446 the normal latency variance on your link due to traffic.
6447 Use this option to put an lower limit on the dynamic timeout
6448 value. Do NOT use this option to always use a fixed (instead
6449 of a dynamic) timeout value. To set a fixed timeout see the
6450 'icp_query_timeout' directive.
6453 NAME: background_ping_rate
6457 LOC: Config.backgroundPingRate
6459 Controls how often the ICP pings are sent to siblings that
6460 have background-ping set.
6464 MULTICAST ICP OPTIONS
6465 -----------------------------------------------------------------------------
6470 LOC: Config.mcast_group_list
6473 This tag specifies a list of multicast groups which your server
6474 should join to receive multicasted ICP queries.
6476 NOTE! Be very careful what you put here! Be sure you
6477 understand the difference between an ICP _query_ and an ICP
6478 _reply_. This option is to be set only if you want to RECEIVE
6479 multicast queries. Do NOT set this option to SEND multicast
6480 ICP (use cache_peer for that). ICP replies are always sent via
6481 unicast, so this option does not affect whether or not you will
6482 receive replies from multicast group members.
6484 You must be very careful to NOT use a multicast address which
6485 is already in use by another group of caches.
6487 If you are unsure about multicast, please read the Multicast
6488 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6490 Usage: mcast_groups 239.128.16.128 224.0.1.20
6492 By default, Squid doesn't listen on any multicast groups.
6495 NAME: mcast_miss_addr
6496 IFDEF: MULTICAST_MISS_STREAM
6498 LOC: Config.mcast_miss.addr
6501 If you enable this option, every "cache miss" URL will
6502 be sent out on the specified multicast address.
6504 Do not enable this option unless you are are absolutely
6505 certain you understand what you are doing.
6508 NAME: mcast_miss_ttl
6509 IFDEF: MULTICAST_MISS_STREAM
6511 LOC: Config.mcast_miss.ttl
6514 This is the time-to-live value for packets multicasted
6515 when multicasting off cache miss URLs is enabled. By
6516 default this is set to 'site scope', i.e. 16.
6519 NAME: mcast_miss_port
6520 IFDEF: MULTICAST_MISS_STREAM
6522 LOC: Config.mcast_miss.port
6525 This is the port number to be used in conjunction with
6529 NAME: mcast_miss_encode_key
6530 IFDEF: MULTICAST_MISS_STREAM
6532 LOC: Config.mcast_miss.encode_key
6533 DEFAULT: XXXXXXXXXXXXXXXX
6535 The URLs that are sent in the multicast miss stream are
6536 encrypted. This is the encryption key.
6539 NAME: mcast_icp_query_timeout
6543 LOC: Config.Timeout.mcast_icp_query
6545 For multicast peers, Squid regularly sends out ICP "probes" to
6546 count how many other peers are listening on the given multicast
6547 address. This value specifies how long Squid should wait to
6548 count all the replies. The default is 2000 msec, or 2
6553 INTERNAL ICON OPTIONS
6554 -----------------------------------------------------------------------------
6557 NAME: icon_directory
6559 LOC: Config.icons.directory
6560 DEFAULT: @DEFAULT_ICON_DIR@
6562 Where the icons are stored. These are normally kept in
6566 NAME: global_internal_static
6568 LOC: Config.onoff.global_internal_static
6571 This directive controls is Squid should intercept all requests for
6572 /squid-internal-static/ no matter which host the URL is requesting
6573 (default on setting), or if nothing special should be done for
6574 such URLs (off setting). The purpose of this directive is to make
6575 icons etc work better in complex cache hierarchies where it may
6576 not always be possible for all corners in the cache mesh to reach
6577 the server generating a directory listing.
6580 NAME: short_icon_urls
6582 LOC: Config.icons.use_short_names
6585 If this is enabled Squid will use short URLs for icons.
6586 If disabled it will revert to the old behavior of including
6587 it's own name and port in the URL.
6589 If you run a complex cache hierarchy with a mix of Squid and
6590 other proxies you may need to disable this directive.
6595 -----------------------------------------------------------------------------
6598 NAME: error_directory
6600 LOC: Config.errorDirectory
6603 If you wish to create your own versions of the default
6604 error files to customize them to suit your company copy
6605 the error/template files to another directory and point
6608 WARNING: This option will disable multi-language support
6609 on error pages if used.
6611 The squid developers are interested in making squid available in
6612 a wide variety of languages. If you are making translations for a
6613 language that Squid does not currently provide please consider
6614 contributing your translation back to the project.
6615 http://wiki.squid-cache.org/Translations
6617 The squid developers working on translations are happy to supply drop-in
6618 translated error files in exchange for any new language contributions.
6621 NAME: error_default_language
6622 IFDEF: USE_ERR_LOCALES
6624 LOC: Config.errorDefaultLanguage
6627 Set the default language which squid will send error pages in
6628 if no existing translation matches the clients language
6631 If unset (default) generic English will be used.
6633 The squid developers are interested in making squid available in
6634 a wide variety of languages. If you are interested in making
6635 translations for any language see the squid wiki for details.
6636 http://wiki.squid-cache.org/Translations
6639 NAME: error_log_languages
6640 IFDEF: USE_ERR_LOCALES
6642 LOC: Config.errorLogMissingLanguages
6645 Log to cache.log what languages users are attempting to
6646 auto-negotiate for translations.
6648 Successful negotiations are not logged. Only failures
6649 have meaning to indicate that Squid may need an upgrade
6650 of its error page translations.
6653 NAME: err_page_stylesheet
6655 LOC: Config.errorStylesheet
6656 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6658 CSS Stylesheet to pattern the display of Squid default error pages.
6660 For information on CSS see http://www.w3.org/Style/CSS/
6665 LOC: Config.errHtmlText
6668 HTML text to include in error messages. Make this a "mailto"
6669 URL to your admin address, or maybe just a link to your
6670 organizations Web page.
6672 To include this in your error messages, you must rewrite
6673 the error template files (found in the "errors" directory).
6674 Wherever you want the 'err_html_text' line to appear,
6675 insert a %L tag in the error template file.
6678 NAME: email_err_data
6681 LOC: Config.onoff.emailErrData
6684 If enabled, information about the occurred error will be
6685 included in the mailto links of the ERR pages (if %W is set)
6686 so that the email body contains the data.
6687 Syntax is <A HREF="mailto:%w%W">%w</A>
6692 LOC: Config.denyInfoList
6695 Usage: deny_info err_page_name acl
6696 or deny_info http://... acl
6697 or deny_info TCP_RESET acl
6699 This can be used to return a ERR_ page for requests which
6700 do not pass the 'http_access' rules. Squid remembers the last
6701 acl it evaluated in http_access, and if a 'deny_info' line exists
6702 for that ACL Squid returns a corresponding error page.
6704 The acl is typically the last acl on the http_access deny line which
6705 denied access. The exceptions to this rule are:
6706 - When Squid needs to request authentication credentials. It's then
6707 the first authentication related acl encountered
6708 - When none of the http_access lines matches. It's then the last
6709 acl processed on the last http_access line.
6710 - When the decision to deny access was made by an adaptation service,
6711 the acl name is the corresponding eCAP or ICAP service_name.
6713 NP: If providing your own custom error pages with error_directory
6714 you may also specify them by your custom file name:
6715 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6717 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6718 may be specified by prefixing the file name with the code and a colon.
6719 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6721 Alternatively you can tell Squid to reset the TCP connection
6722 by specifying TCP_RESET.
6724 Or you can specify an error URL or URL pattern. The browsers will
6725 get redirected to the specified URL after formatting tags have
6726 been replaced. Redirect will be done with 302 or 307 according to
6727 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6728 the URL. e.g. 303:http://example.com/
6731 %a - username (if available. Password NOT included)
6734 %E - Error description
6736 %H - Request domain name
6737 %i - Client IP Address
6739 %o - Message result from external ACL helper
6740 %p - Request Port number
6741 %P - Request Protocol name
6742 %R - Request URL path
6743 %T - Timestamp in RFC 1123 format
6744 %U - Full canonical URL from client
6745 (HTTPS URLs terminate with *)
6746 %u - Full canonical URL from client
6747 %w - Admin email from squid.conf
6749 %% - Literal percent (%) code
6754 OPTIONS INFLUENCING REQUEST FORWARDING
6755 -----------------------------------------------------------------------------
6758 NAME: nonhierarchical_direct
6760 LOC: Config.onoff.nonhierarchical_direct
6763 By default, Squid will send any non-hierarchical requests
6764 (matching hierarchy_stoplist or not cacheable request type) direct
6767 If you set this to off, Squid will prefer to send these
6768 requests to parents.
6770 Note that in most configurations, by turning this off you will only
6771 add latency to these request without any improvement in global hit
6774 If you are inside an firewall see never_direct instead of
6780 LOC: Config.onoff.prefer_direct
6783 Normally Squid tries to use parents for most requests. If you for some
6784 reason like it to first try going direct and only use a parent if
6785 going direct fails set this to on.
6787 By combining nonhierarchical_direct off and prefer_direct on you
6788 can set up Squid to use a parent as a backup path if going direct
6791 Note: If you want Squid to use parents for all requests see
6792 the never_direct directive. prefer_direct only modifies how Squid
6793 acts on cacheable requests.
6798 LOC: Config.accessList.AlwaysDirect
6801 Usage: always_direct allow|deny [!]aclname ...
6803 Here you can use ACL elements to specify requests which should
6804 ALWAYS be forwarded by Squid to the origin servers without using
6805 any peers. For example, to always directly forward requests for
6806 local servers ignoring any parents or siblings you may have use
6809 acl local-servers dstdomain my.domain.net
6810 always_direct allow local-servers
6812 To always forward FTP requests directly, use
6815 always_direct allow FTP
6817 NOTE: There is a similar, but opposite option named
6818 'never_direct'. You need to be aware that "always_direct deny
6819 foo" is NOT the same thing as "never_direct allow foo". You
6820 may need to use a deny rule to exclude a more-specific case of
6821 some other rule. Example:
6823 acl local-external dstdomain external.foo.net
6824 acl local-servers dstdomain .foo.net
6825 always_direct deny local-external
6826 always_direct allow local-servers
6828 NOTE: If your goal is to make the client forward the request
6829 directly to the origin server bypassing Squid then this needs
6830 to be done in the client configuration. Squid configuration
6831 can only tell Squid how Squid should fetch the object.
6833 NOTE: This directive is not related to caching. The replies
6834 is cached as usual even if you use always_direct. To not cache
6835 the replies see the 'cache' directive.
6837 This clause supports both fast and slow acl types.
6838 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6843 LOC: Config.accessList.NeverDirect
6846 Usage: never_direct allow|deny [!]aclname ...
6848 never_direct is the opposite of always_direct. Please read
6849 the description for always_direct if you have not already.
6851 With 'never_direct' you can use ACL elements to specify
6852 requests which should NEVER be forwarded directly to origin
6853 servers. For example, to force the use of a proxy for all
6854 requests, except those in your local domain use something like:
6856 acl local-servers dstdomain .foo.net
6857 never_direct deny local-servers
6858 never_direct allow all
6860 or if Squid is inside a firewall and there are local intranet
6861 servers inside the firewall use something like:
6863 acl local-intranet dstdomain .foo.net
6864 acl local-external dstdomain external.foo.net
6865 always_direct deny local-external
6866 always_direct allow local-intranet
6867 never_direct allow all
6869 This clause supports both fast and slow acl types.
6870 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6874 ADVANCED NETWORKING OPTIONS
6875 -----------------------------------------------------------------------------
6878 NAME: incoming_udp_average incoming_icp_average
6881 LOC: Config.comm_incoming.udp.average
6883 Heavy voodoo here. I can't even believe you are reading this.
6884 Are you crazy? Don't even think about adjusting these unless
6885 you understand the algorithms in comm_select.c first!
6888 NAME: incoming_tcp_average incoming_http_average
6891 LOC: Config.comm_incoming.tcp.average
6893 Heavy voodoo here. I can't even believe you are reading this.
6894 Are you crazy? Don't even think about adjusting these unless
6895 you understand the algorithms in comm_select.c first!
6898 NAME: incoming_dns_average
6901 LOC: Config.comm_incoming.dns.average
6903 Heavy voodoo here. I can't even believe you are reading this.
6904 Are you crazy? Don't even think about adjusting these unless
6905 you understand the algorithms in comm_select.c first!
6908 NAME: min_udp_poll_cnt min_icp_poll_cnt
6911 LOC: Config.comm_incoming.udp.min_poll
6913 Heavy voodoo here. I can't even believe you are reading this.
6914 Are you crazy? Don't even think about adjusting these unless
6915 you understand the algorithms in comm_select.c first!
6918 NAME: min_dns_poll_cnt
6921 LOC: Config.comm_incoming.dns.min_poll
6923 Heavy voodoo here. I can't even believe you are reading this.
6924 Are you crazy? Don't even think about adjusting these unless
6925 you understand the algorithms in comm_select.c first!
6928 NAME: min_tcp_poll_cnt min_http_poll_cnt
6931 LOC: Config.comm_incoming.tcp.min_poll
6933 Heavy voodoo here. I can't even believe you are reading this.
6934 Are you crazy? Don't even think about adjusting these unless
6935 you understand the algorithms in comm_select.c first!
6941 LOC: Config.accept_filter
6945 The name of an accept(2) filter to install on Squid's
6946 listen socket(s). This feature is perhaps specific to
6947 FreeBSD and requires support in the kernel.
6949 The 'httpready' filter delays delivering new connections
6950 to Squid until a full HTTP request has been received.
6951 See the accf_http(9) man page for details.
6953 The 'dataready' filter delays delivering new connections
6954 to Squid until there is some data to process.
6955 See the accf_dataready(9) man page for details.
6959 The 'data' filter delays delivering of new connections
6960 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
6961 You may optionally specify a number of seconds to wait by
6962 'data=N' where N is the number of seconds. Defaults to 30
6963 if not specified. See the tcp(7) man page for details.
6966 accept_filter httpready
6971 NAME: client_ip_max_connections
6973 LOC: Config.client_ip_max_connections
6976 Set an absolute limit on the number of connections a single
6977 client IP can use. Any more than this and Squid will begin to drop
6978 new connections from the client until it closes some links.
6980 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
6981 connections from the client. For finer control use the ACL access controls.
6983 Requires client_db to be enabled (the default).
6985 WARNING: This may noticably slow down traffic received via external proxies
6986 or NAT devices and cause them to rebound error messages back to their clients.
6989 NAME: tcp_recv_bufsize
6993 LOC: Config.tcpRcvBufsz
6995 Size of receive buffer to set for TCP sockets. Probably just
6996 as easy to change your kernel's default. Set to zero to use
6997 the default buffer size.
7002 -----------------------------------------------------------------------------
7009 LOC: Adaptation::Icap::TheConfig.onoff
7012 If you want to enable the ICAP module support, set this to on.
7015 NAME: icap_connect_timeout
7018 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7021 This parameter specifies how long to wait for the TCP connect to
7022 the requested ICAP server to complete before giving up and either
7023 terminating the HTTP transaction or bypassing the failure.
7025 The default for optional services is peer_connect_timeout.
7026 The default for essential services is connect_timeout.
7027 If this option is explicitly set, its value applies to all services.
7030 NAME: icap_io_timeout
7034 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7037 This parameter specifies how long to wait for an I/O activity on
7038 an established, active ICAP connection before giving up and
7039 either terminating the HTTP transaction or bypassing the
7042 The default is read_timeout.
7045 NAME: icap_service_failure_limit
7046 COMMENT: limit [in memory-depth time-units]
7047 TYPE: icap_service_failure_limit
7049 LOC: Adaptation::Icap::TheConfig
7052 The limit specifies the number of failures that Squid tolerates
7053 when establishing a new TCP connection with an ICAP service. If
7054 the number of failures exceeds the limit, the ICAP service is
7055 not used for new ICAP requests until it is time to refresh its
7058 A negative value disables the limit. Without the limit, an ICAP
7059 service will not be considered down due to connectivity failures
7060 between ICAP OPTIONS requests.
7062 Squid forgets ICAP service failures older than the specified
7063 value of memory-depth. The memory fading algorithm
7064 is approximate because Squid does not remember individual
7065 errors but groups them instead, splitting the option
7066 value into ten time slots of equal length.
7068 When memory-depth is 0 and by default this option has no
7069 effect on service failure expiration.
7071 Squid always forgets failures when updating service settings
7072 using an ICAP OPTIONS transaction, regardless of this option
7076 # suspend service usage after 10 failures in 5 seconds:
7077 icap_service_failure_limit 10 in 5 seconds
7080 NAME: icap_service_revival_delay
7083 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7086 The delay specifies the number of seconds to wait after an ICAP
7087 OPTIONS request failure before requesting the options again. The
7088 failed ICAP service is considered "down" until fresh OPTIONS are
7091 The actual delay cannot be smaller than the hardcoded minimum
7092 delay of 30 seconds.
7095 NAME: icap_preview_enable
7099 LOC: Adaptation::Icap::TheConfig.preview_enable
7102 The ICAP Preview feature allows the ICAP server to handle the
7103 HTTP message by looking only at the beginning of the message body
7104 or even without receiving the body at all. In some environments,
7105 previews greatly speedup ICAP processing.
7107 During an ICAP OPTIONS transaction, the server may tell Squid what
7108 HTTP messages should be previewed and how big the preview should be.
7109 Squid will not use Preview if the server did not request one.
7111 To disable ICAP Preview for all ICAP services, regardless of
7112 individual ICAP server OPTIONS responses, set this option to "off".
7114 icap_preview_enable off
7117 NAME: icap_preview_size
7120 LOC: Adaptation::Icap::TheConfig.preview_size
7123 The default size of preview data to be sent to the ICAP server.
7124 -1 means no preview. This value might be overwritten on a per server
7125 basis by OPTIONS requests.
7128 NAME: icap_206_enable
7132 LOC: Adaptation::Icap::TheConfig.allow206_enable
7135 206 (Partial Content) responses is an ICAP extension that allows the
7136 ICAP agents to optionally combine adapted and original HTTP message
7137 content. The decision to combine is postponed until the end of the
7138 ICAP response. Squid supports Partial Content extension by default.
7140 Activation of the Partial Content extension is negotiated with each
7141 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7142 negotation correctly even if they do not support the extension, but
7143 some might fail. To disable Partial Content support for all ICAP
7144 services and to avoid any negotiation, set this option to "off".
7150 NAME: icap_default_options_ttl
7153 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7156 The default TTL value for ICAP OPTIONS responses that don't have
7157 an Options-TTL header.
7160 NAME: icap_persistent_connections
7164 LOC: Adaptation::Icap::TheConfig.reuse_connections
7167 Whether or not Squid should use persistent connections to
7171 NAME: adaptation_send_client_ip icap_send_client_ip
7173 IFDEF: USE_ADAPTATION
7175 LOC: Adaptation::Config::send_client_ip
7178 If enabled, Squid shares HTTP client IP information with adaptation
7179 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7180 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7182 See also: adaptation_uses_indirect_client
7185 NAME: adaptation_send_username icap_send_client_username
7187 IFDEF: USE_ADAPTATION
7189 LOC: Adaptation::Config::send_username
7192 This sends authenticated HTTP client username (if available) to
7193 the adaptation service.
7195 For ICAP, the username value is encoded based on the
7196 icap_client_username_encode option and is sent using the header
7197 specified by the icap_client_username_header option.
7200 NAME: icap_client_username_header
7203 LOC: Adaptation::Icap::TheConfig.client_username_header
7204 DEFAULT: X-Client-Username
7206 ICAP request header name to use for adaptation_send_username.
7209 NAME: icap_client_username_encode
7213 LOC: Adaptation::Icap::TheConfig.client_username_encode
7216 Whether to base64 encode the authenticated client username.
7220 TYPE: icap_service_type
7222 LOC: Adaptation::Icap::TheConfig
7225 Defines a single ICAP service using the following format:
7227 icap_service id vectoring_point uri [option ...]
7230 an opaque identifier or name which is used to direct traffic to
7231 this specific service. Must be unique among all adaptation
7232 services in squid.conf.
7234 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7235 This specifies at which point of transaction processing the
7236 ICAP service should be activated. *_postcache vectoring points
7237 are not yet supported.
7239 uri: icap://servername:port/servicepath
7240 ICAP server and service location.
7242 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7243 transactions. Squid does not enforce that requirement. You can specify
7244 services with the same service_url and different vectoring_points. You
7245 can even specify multiple identical services as long as their
7246 service_names differ.
7249 Service options are separated by white space. ICAP services support
7250 the following name=value options:
7253 If set to 'on' or '1', the ICAP service is treated as
7254 optional. If the service cannot be reached or malfunctions,
7255 Squid will try to ignore any errors and process the message as
7256 if the service was not enabled. No all ICAP errors can be
7257 bypassed. If set to 0, the ICAP service is treated as
7258 essential and all ICAP errors will result in an error page
7259 returned to the HTTP client.
7261 Bypass is off by default: services are treated as essential.
7264 If set to 'on' or '1', the ICAP service is allowed to
7265 dynamically change the current message adaptation plan by
7266 returning a chain of services to be used next. The services
7267 are specified using the X-Next-Services ICAP response header
7268 value, formatted as a comma-separated list of service names.
7269 Each named service should be configured in squid.conf. Other
7270 services are ignored. An empty X-Next-Services value results
7271 in an empty plan which ends the current adaptation.
7273 Dynamic adaptation plan may cross or cover multiple supported
7274 vectoring points in their natural processing order.
7276 Routing is not allowed by default: the ICAP X-Next-Services
7277 response header is ignored.
7280 Only has effect on split-stack systems. The default on those systems
7281 is to use IPv4-only connections. When set to 'on' this option will
7282 make Squid use IPv6-only connections to contact this ICAP service.
7284 on-overload=block|bypass|wait|force
7285 If the service Max-Connections limit has been reached, do
7286 one of the following for each new ICAP transaction:
7287 * block: send an HTTP error response to the client
7288 * bypass: ignore the "over-connected" ICAP service
7289 * wait: wait (in a FIFO queue) for an ICAP connection slot
7290 * force: proceed, ignoring the Max-Connections limit
7292 In SMP mode with N workers, each worker assumes the service
7293 connection limit is Max-Connections/N, even though not all
7294 workers may use a given service.
7296 The default value is "bypass" if service is bypassable,
7297 otherwise it is set to "wait".
7301 Use the given number as the Max-Connections limit, regardless
7302 of the Max-Connections value given by the service, if any.
7304 Older icap_service format without optional named parameters is
7305 deprecated but supported for backward compatibility.
7308 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7309 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7313 TYPE: icap_class_type
7318 This deprecated option was documented to define an ICAP service
7319 chain, even though it actually defined a set of similar, redundant
7320 services, and the chains were not supported.
7322 To define a set of redundant services, please use the
7323 adaptation_service_set directive. For service chains, use
7324 adaptation_service_chain.
7328 TYPE: icap_access_type
7333 This option is deprecated. Please use adaptation_access, which
7334 has the same ICAP functionality, but comes with better
7335 documentation, and eCAP support.
7340 -----------------------------------------------------------------------------
7347 LOC: Adaptation::Ecap::TheConfig.onoff
7350 Controls whether eCAP support is enabled.
7354 TYPE: ecap_service_type
7356 LOC: Adaptation::Ecap::TheConfig
7359 Defines a single eCAP service
7361 ecap_service id vectoring_point uri [option ...]
7364 an opaque identifier or name which is used to direct traffic to
7365 this specific service. Must be unique among all adaptation
7366 services in squid.conf.
7368 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7369 This specifies at which point of transaction processing the
7370 eCAP service should be activated. *_postcache vectoring points
7371 are not yet supported.
7373 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7374 Squid uses the eCAP service URI to match this configuration
7375 line with one of the dynamically loaded services. Each loaded
7376 eCAP service must have a unique URI. Obtain the right URI from
7377 the service provider.
7380 Service options are separated by white space. eCAP services support
7381 the following name=value options:
7384 If set to 'on' or '1', the eCAP service is treated as optional.
7385 If the service cannot be reached or malfunctions, Squid will try
7386 to ignore any errors and process the message as if the service
7387 was not enabled. No all eCAP errors can be bypassed.
7388 If set to 'off' or '0', the eCAP service is treated as essential
7389 and all eCAP errors will result in an error page returned to the
7392 Bypass is off by default: services are treated as essential.
7395 If set to 'on' or '1', the eCAP service is allowed to
7396 dynamically change the current message adaptation plan by
7397 returning a chain of services to be used next.
7399 Dynamic adaptation plan may cross or cover multiple supported
7400 vectoring points in their natural processing order.
7402 Routing is not allowed by default.
7404 Older ecap_service format without optional named parameters is
7405 deprecated but supported for backward compatibility.
7409 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7410 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7413 NAME: loadable_modules
7415 IFDEF: USE_LOADABLE_MODULES
7416 LOC: Config.loadable_module_names
7419 Instructs Squid to load the specified dynamic module(s) or activate
7420 preloaded module(s).
7422 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7426 MESSAGE ADAPTATION OPTIONS
7427 -----------------------------------------------------------------------------
7430 NAME: adaptation_service_set
7431 TYPE: adaptation_service_set_type
7432 IFDEF: USE_ADAPTATION
7437 Configures an ordered set of similar, redundant services. This is
7438 useful when hot standby or backup adaptation servers are available.
7440 adaptation_service_set set_name service_name1 service_name2 ...
7442 The named services are used in the set declaration order. The first
7443 applicable adaptation service from the set is used first. The next
7444 applicable service is tried if and only if the transaction with the
7445 previous service fails and the message waiting to be adapted is still
7448 When adaptation starts, broken services are ignored as if they were
7449 not a part of the set. A broken service is a down optional service.
7451 The services in a set must be attached to the same vectoring point
7452 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7454 If all services in a set are optional then adaptation failures are
7455 bypassable. If all services in the set are essential, then a
7456 transaction failure with one service may still be retried using
7457 another service from the set, but when all services fail, the master
7458 transaction fails as well.
7460 A set may contain a mix of optional and essential services, but that
7461 is likely to lead to surprising results because broken services become
7462 ignored (see above), making previously bypassable failures fatal.
7463 Technically, it is the bypassability of the last failed service that
7466 See also: adaptation_access adaptation_service_chain
7469 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7470 adaptation service_set svcLogger loggerLocal loggerRemote
7473 NAME: adaptation_service_chain
7474 TYPE: adaptation_service_chain_type
7475 IFDEF: USE_ADAPTATION
7480 Configures a list of complementary services that will be applied
7481 one-by-one, forming an adaptation chain or pipeline. This is useful
7482 when Squid must perform different adaptations on the same message.
7484 adaptation_service_chain chain_name service_name1 svc_name2 ...
7486 The named services are used in the chain declaration order. The first
7487 applicable adaptation service from the chain is used first. The next
7488 applicable service is applied to the successful adaptation results of
7489 the previous service in the chain.
7491 When adaptation starts, broken services are ignored as if they were
7492 not a part of the chain. A broken service is a down optional service.
7494 Request satisfaction terminates the adaptation chain because Squid
7495 does not currently allow declaration of RESPMOD services at the
7496 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7498 The services in a chain must be attached to the same vectoring point
7499 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7501 A chain may contain a mix of optional and essential services. If an
7502 essential adaptation fails (or the failure cannot be bypassed for
7503 other reasons), the master transaction fails. Otherwise, the failure
7504 is bypassed as if the failed adaptation service was not in the chain.
7506 See also: adaptation_access adaptation_service_set
7509 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
7512 NAME: adaptation_access
7513 TYPE: adaptation_access_type
7514 IFDEF: USE_ADAPTATION
7518 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
7520 adaptation_access service_name allow|deny [!]aclname...
7521 adaptation_access set_name allow|deny [!]aclname...
7523 At each supported vectoring point, the adaptation_access
7524 statements are processed in the order they appear in this
7525 configuration file. Statements pointing to the following services
7526 are ignored (i.e., skipped without checking their ACL):
7528 - services serving different vectoring points
7529 - "broken-but-bypassable" services
7530 - "up" services configured to ignore such transactions
7531 (e.g., based on the ICAP Transfer-Ignore header).
7533 When a set_name is used, all services in the set are checked
7534 using the same rules, to find the first applicable one. See
7535 adaptation_service_set for details.
7537 If an access list is checked and there is a match, the
7538 processing stops: For an "allow" rule, the corresponding
7539 adaptation service is used for the transaction. For a "deny"
7540 rule, no adaptation service is activated.
7542 It is currently not possible to apply more than one adaptation
7543 service at the same vectoring point to the same HTTP transaction.
7545 See also: icap_service and ecap_service
7548 adaptation_access service_1 allow all
7551 NAME: adaptation_service_iteration_limit
7553 IFDEF: USE_ADAPTATION
7554 LOC: Adaptation::Config::service_iteration_limit
7557 Limits the number of iterations allowed when applying adaptation
7558 services to a message. If your longest adaptation set or chain
7559 may have more than 16 services, increase the limit beyond its
7560 default value of 16. If detecting infinite iteration loops sooner
7561 is critical, make the iteration limit match the actual number
7562 of services in your longest adaptation set or chain.
7564 Infinite adaptation loops are most likely with routing services.
7566 See also: icap_service routing=1
7569 NAME: adaptation_masterx_shared_names
7571 IFDEF: USE_ADAPTATION
7572 LOC: Adaptation::Config::masterx_shared_name
7575 For each master transaction (i.e., the HTTP request and response
7576 sequence, including all related ICAP and eCAP exchanges), Squid
7577 maintains a table of metadata. The table entries are (name, value)
7578 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7579 with the master transaction.
7581 This option specifies the table entry names that Squid must accept
7582 from and forward to the adaptation transactions.
7584 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7585 shared table by returning an ICAP header field with a name
7586 specified in adaptation_masterx_shared_names.
7588 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7589 shared table by implementing the libecap::visitEachOption() API
7590 to provide an option with a name specified in
7591 adaptation_masterx_shared_names.
7593 Squid will store and forward the set entry to subsequent adaptation
7594 transactions within the same master transaction scope.
7596 Only one shared entry name is supported at this time.
7599 # share authentication information among ICAP services
7600 adaptation_masterx_shared_names X-Subscriber-ID
7603 NAME: adaptation_meta
7605 IFDEF: USE_ADAPTATION
7606 LOC: Adaptation::Config::metaHeaders
7609 This option allows Squid administrator to add custom ICAP request
7610 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7611 Use it to pass custom authentication tokens and other
7612 transaction-state related meta information to an ICAP/eCAP service.
7614 The addition of a meta header is ACL-driven:
7615 adaptation_meta name value [!]aclname ...
7617 Processing for a given header name stops after the first ACL list match.
7618 Thus, it is impossible to add two headers with the same name. If no ACL
7619 lists match for a given header name, no such header is added. For
7622 # do not debug transactions except for those that need debugging
7623 adaptation_meta X-Debug 1 needs_debugging
7625 # log all transactions except for those that must remain secret
7626 adaptation_meta X-Log 1 !keep_secret
7628 # mark transactions from users in the "G 1" group
7629 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7631 The "value" parameter may be a regular squid.conf token or a "double
7632 quoted string". Within the quoted string, use backslash (\) to escape
7633 any character, which is currently only useful for escaping backslashes
7634 and double quotes. For example,
7635 "this string has one backslash (\\) and two \"quotes\""
7637 Used adaptation_meta header values may be logged via %note
7638 logformat code. If multiple adaptation_meta headers with the same name
7639 are used during master transaction lifetime, the header values are
7640 logged in the order they were used and duplicate values are ignored
7641 (only the first repeated value will be logged).
7647 LOC: Adaptation::Icap::TheConfig.repeat
7648 DEFAULT_IF_NONE: deny all
7650 This ACL determines which retriable ICAP transactions are
7651 retried. Transactions that received a complete ICAP response
7652 and did not have to consume or produce HTTP bodies to receive
7653 that response are usually retriable.
7655 icap_retry allow|deny [!]aclname ...
7657 Squid automatically retries some ICAP I/O timeouts and errors
7658 due to persistent connection race conditions.
7660 See also: icap_retry_limit
7663 NAME: icap_retry_limit
7666 LOC: Adaptation::Icap::TheConfig.repeat_limit
7669 Limits the number of retries allowed. When set to zero (default),
7670 no retries are allowed.
7672 Communication errors due to persistent connection race
7673 conditions are unavoidable, automatically retried, and do not
7674 count against this limit.
7676 See also: icap_retry
7682 -----------------------------------------------------------------------------
7685 NAME: check_hostnames
7688 LOC: Config.onoff.check_hostnames
7690 For security and stability reasons Squid can check
7691 hostnames for Internet standard RFC compliance. If you want
7692 Squid to perform these checks turn this directive on.
7695 NAME: allow_underscore
7698 LOC: Config.onoff.allow_underscore
7700 Underscore characters is not strictly allowed in Internet hostnames
7701 but nevertheless used by many sites. Set this to off if you want
7702 Squid to be strict about the standard.
7703 This check is performed only when check_hostnames is set to on.
7706 NAME: cache_dns_program
7708 IFDEF: USE_DNSHELPER
7709 DEFAULT: @DEFAULT_DNSSERVER@
7710 LOC: Config.Program.dnsserver
7712 Specify the location of the executable for dnslookup process.
7716 TYPE: HelperChildConfig
7717 IFDEF: USE_DNSHELPER
7718 DEFAULT: 32 startup=1 idle=1
7719 LOC: Config.dnsChildren
7721 The maximum number of processes spawn to service DNS name lookups.
7722 If you limit it too few Squid will have to wait for them to process
7723 a backlog of requests, slowing it down. If you allow too many they
7724 will use RAM and other system resources noticably.
7725 The maximum this may be safely set to is 32.
7727 The startup= and idle= options allow some measure of skew in your
7732 Sets a minimum of how many processes are to be spawned when Squid
7733 starts or reconfigures. When set to zero the first request will
7734 cause spawning of the first child process to handle it.
7736 Starting too few will cause an initial slowdown in traffic as Squid
7737 attempts to simultaneously spawn enough processes to cope.
7741 Sets a minimum of how many processes Squid is to try and keep available
7742 at all times. When traffic begins to rise above what the existing
7743 processes can handle this many more will be spawned up to the maximum
7744 configured. A minimum setting of 1 is required.
7747 NAME: dns_retransmit_interval
7750 LOC: Config.Timeout.idns_retransmit
7751 IFDEF: !USE_DNSHELPER
7753 Initial retransmit interval for DNS queries. The interval is
7754 doubled each time all configured DNS servers have been tried.
7760 LOC: Config.Timeout.idns_query
7761 IFDEF: !USE_DNSHELPER
7763 DNS Query timeout. If no response is received to a DNS query
7764 within this time all DNS servers for the queried domain
7765 are assumed to be unavailable.
7768 NAME: dns_packet_max
7771 LOC: Config.dns.packet_max
7772 IFDEF: !USE_DNSHELPER
7774 Maximum number of bytes packet size to advertise via EDNS.
7775 Set to "none" to disable EDNS large packet support.
7777 For legacy reasons DNS UDP replies will default to 512 bytes which
7778 is too small for many responses. EDNS provides a means for Squid to
7779 negotiate receiving larger responses back immediately without having
7780 to failover with repeat requests. Responses larger than this limit
7781 will retain the old behaviour of failover to TCP DNS.
7783 Squid has no real fixed limit internally, but allowing packet sizes
7784 over 1500 bytes requires network jumbogram support and is usually not
7787 WARNING: The RFC also indicates that some older resolvers will reply
7788 with failure of the whole request if the extension is added. Some
7789 resolvers have already been identified which will reply with mangled
7790 EDNS response on occasion. Usually in response to many-KB jumbogram
7791 sizes being advertised by Squid.
7792 Squid will currently treat these both as an unable-to-resolve domain
7793 even if it would be resolvable without EDNS.
7800 LOC: Config.onoff.res_defnames
7802 Normally the RES_DEFNAMES resolver option is disabled
7803 (see res_init(3)). This prevents caches in a hierarchy
7804 from interpreting single-component hostnames locally. To allow
7805 Squid to handle single-component names, enable this option.
7808 NAME: dns_nameservers
7811 LOC: Config.dns_nameservers
7813 Use this if you want to specify a list of DNS name servers
7814 (IP addresses) to use instead of those given in your
7815 /etc/resolv.conf file.
7816 On Windows platforms, if no value is specified here or in
7817 the /etc/resolv.conf file, the list of DNS name servers are
7818 taken from the Windows registry, both static and dynamic DHCP
7819 configurations are supported.
7821 Example: dns_nameservers 10.0.0.1 192.172.0.4
7826 DEFAULT: @DEFAULT_HOSTS@
7827 LOC: Config.etcHostsPath
7829 Location of the host-local IP name-address associations
7830 database. Most Operating Systems have such a file on different
7832 - Un*X & Linux: /etc/hosts
7833 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7834 (%SystemRoot% value install default is c:\winnt)
7835 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7836 (%SystemRoot% value install default is c:\windows)
7837 - Windows 9x/Me: %windir%\hosts
7838 (%windir% value is usually c:\windows)
7839 - Cygwin: /etc/hosts
7841 The file contains newline-separated definitions, in the
7842 form ip_address_in_dotted_form name [name ...] names are
7843 whitespace-separated. Lines beginning with an hash (#)
7844 character are comments.
7846 The file is checked at startup and upon configuration.
7847 If set to 'none', it won't be checked.
7848 If append_domain is used, that domain will be added to
7849 domain-local (i.e. not containing any dot character) host
7855 LOC: Config.appendDomain
7858 Appends local domain name to hostnames without any dots in
7859 them. append_domain must begin with a period.
7861 Be warned there are now Internet names with no dots in
7862 them using only top-domain names, so setting this may
7863 cause some Internet sites to become unavailable.
7866 append_domain .yourdomain.com
7869 NAME: ignore_unknown_nameservers
7871 LOC: Config.onoff.ignore_unknown_nameservers
7873 IFDEF: !USE_DNSHELPER
7875 By default Squid checks that DNS responses are received
7876 from the same IP addresses they are sent to. If they
7877 don't match, Squid ignores the response and writes a warning
7878 message to cache.log. You can allow responses from unknown
7879 nameservers by setting this option to 'off'.
7885 LOC: Config.dns.v4_first
7886 IFDEF: !USE_DNSHELPER
7888 With the IPv6 Internet being as fast or faster than IPv4 Internet
7889 for most networks Squid prefers to contact websites over IPv6.
7891 This option reverses the order of preference to make Squid contact
7892 dual-stack websites over IPv4 first. Squid will still perform both
7893 IPv6 and IPv4 DNS lookups before connecting.
7896 This option will restrict the situations under which IPv6
7897 connectivity is used (and tested). Hiding network problems
7898 which would otherwise be detected and warned about.
7902 COMMENT: (number of entries)
7905 LOC: Config.ipcache.size
7912 LOC: Config.ipcache.low
7919 LOC: Config.ipcache.high
7921 The size, low-, and high-water marks for the IP cache.
7924 NAME: fqdncache_size
7925 COMMENT: (number of entries)
7928 LOC: Config.fqdncache.size
7930 Maximum number of FQDN cache entries.
7935 -----------------------------------------------------------------------------
7942 LOC: Config.onoff.mem_pools
7944 If set, Squid will keep pools of allocated (but unused) memory
7945 available for future use. If memory is a premium on your
7946 system and you believe your malloc library outperforms Squid
7947 routines, disable this.
7950 NAME: memory_pools_limit
7954 LOC: Config.MemPools.limit
7956 Used only with memory_pools on:
7957 memory_pools_limit 50 MB
7959 If set to a non-zero value, Squid will keep at most the specified
7960 limit of allocated (but unused) memory in memory pools. All free()
7961 requests that exceed this limit will be handled by your malloc
7962 library. Squid does not pre-allocate any memory, just safe-keeps
7963 objects that otherwise would be free()d. Thus, it is safe to set
7964 memory_pools_limit to a reasonably high value even if your
7965 configuration will use less memory.
7967 If set to none, Squid will keep all memory it can. That is, there
7968 will be no limit on the total amount of memory used for safe-keeping.
7970 To disable memory allocation optimization, do not set
7971 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
7973 An overhead for maintaining memory pools is not taken into account
7974 when the limit is checked. This overhead is close to four bytes per
7975 object kept. However, pools may actually _save_ memory because of
7976 reduced memory thrashing in your malloc library.
7980 COMMENT: on|off|transparent|truncate|delete
7983 LOC: opt_forwarded_for
7985 If set to "on", Squid will append your client's IP address
7986 in the HTTP requests it forwards. By default it looks like:
7988 X-Forwarded-For: 192.1.2.3
7990 If set to "off", it will appear as
7992 X-Forwarded-For: unknown
7994 If set to "transparent", Squid will not alter the
7995 X-Forwarded-For header in any way.
7997 If set to "delete", Squid will delete the entire
7998 X-Forwarded-For header.
8000 If set to "truncate", Squid will remove all existing
8001 X-Forwarded-For entries, and place the client IP as the sole entry.
8004 NAME: cachemgr_passwd
8005 TYPE: cachemgrpasswd
8007 LOC: Config.passwd_list
8009 Specify passwords for cachemgr operations.
8011 Usage: cachemgr_passwd password action action ...
8013 Some valid actions are (see cache manager menu for a full list):
8053 * Indicates actions which will not be performed without a
8054 valid password, others can be performed if not listed here.
8056 To disable an action, set the password to "disable".
8057 To allow performing an action without a password, set the
8060 Use the keyword "all" to set the same password for all actions.
8063 cachemgr_passwd secret shutdown
8064 cachemgr_passwd lesssssssecret info stats/objects
8065 cachemgr_passwd disable all
8072 LOC: Config.onoff.client_db
8074 If you want to disable collecting per-client statistics,
8075 turn off client_db here.
8078 NAME: refresh_all_ims
8082 LOC: Config.onoff.refresh_all_ims
8084 When you enable this option, squid will always check
8085 the origin server for an update when a client sends an
8086 If-Modified-Since request. Many browsers use IMS
8087 requests when the user requests a reload, and this
8088 ensures those clients receive the latest version.
8090 By default (off), squid may return a Not Modified response
8091 based on the age of the cached version.
8094 NAME: reload_into_ims
8095 IFDEF: USE_HTTP_VIOLATIONS
8099 LOC: Config.onoff.reload_into_ims
8101 When you enable this option, client no-cache or ``reload''
8102 requests will be changed to If-Modified-Since requests.
8103 Doing this VIOLATES the HTTP standard. Enabling this
8104 feature could make you liable for problems which it
8107 see also refresh_pattern for a more selective approach.
8110 NAME: connect_retries
8112 LOC: Config.connect_retries
8115 This sets the maximum number of connection attempts made for each
8116 TCP connection. The connect_retries attempts must all still
8117 complete within the connection timeout period.
8119 The default is not to re-try if the first connection attempt fails.
8120 The (not recommended) maximum is 10 tries.
8122 A warning message will be generated if it is set to a too-high
8123 value and the configured value will be over-ridden.
8125 Note: These re-tries are in addition to forward_max_tries
8126 which limit how many different addresses may be tried to find
8130 NAME: retry_on_error
8132 LOC: Config.retry.onerror
8135 If set to ON Squid will automatically retry requests when
8136 receiving an error response with status 403 (Forbidden),
8137 500 (Internal Error), 501 or 503 (Service not available).
8138 Status 502 and 504 (Gateway errors) are always retried.
8140 This is mainly useful if you are in a complex cache hierarchy to
8141 work around access control errors.
8143 NOTE: This retry will attempt to find another working destination.
8144 Which is different from the server which just failed.
8147 NAME: as_whois_server
8149 LOC: Config.as_whois_server
8150 DEFAULT: whois.ra.net
8152 WHOIS server to query for AS numbers. NOTE: AS numbers are
8153 queried only when Squid starts up, not for every request.
8158 LOC: Config.onoff.offline
8161 Enable this option and Squid will never try to validate cached
8165 NAME: uri_whitespace
8166 TYPE: uri_whitespace
8167 LOC: Config.uri_whitespace
8170 What to do with requests that have whitespace characters in the
8173 strip: The whitespace characters are stripped out of the URL.
8174 This is the behavior recommended by RFC2396.
8175 deny: The request is denied. The user receives an "Invalid
8177 allow: The request is allowed and the URI is not changed. The
8178 whitespace characters remain in the URI. Note the
8179 whitespace is passed to redirector processes if they
8181 encode: The request is allowed and the whitespace characters are
8182 encoded according to RFC1738. This could be considered
8183 a violation of the HTTP/1.1
8184 RFC because proxies are not allowed to rewrite URI's.
8185 chop: The request is allowed and the URI is chopped at the
8186 first whitespace. This might also be considered a
8192 LOC: Config.chroot_dir
8195 Specifies a directory where Squid should do a chroot() while
8196 initializing. This also causes Squid to fully drop root
8197 privileges after initializing. This means, for example, if you
8198 use a HTTP port less than 1024 and try to reconfigure, you may
8199 get an error saying that Squid can not open the port.
8202 NAME: balance_on_multiple_ip
8204 LOC: Config.onoff.balance_on_multiple_ip
8207 Modern IP resolvers in squid sort lookup results by preferred access.
8208 By default squid will use these IP in order and only rotates to
8209 the next listed when the most preffered fails.
8211 Some load balancing servers based on round robin DNS have been
8212 found not to preserve user session state across requests
8213 to different IP addresses.
8215 Enabling this directive Squid rotates IP's per request.
8218 NAME: pipeline_prefetch
8220 LOC: Config.onoff.pipeline_prefetch
8223 To boost the performance of pipelined requests to closer
8224 match that of a non-proxied environment Squid can try to fetch
8225 up to two requests in parallel from a pipeline.
8227 Defaults to off for bandwidth management and access logging
8230 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8233 NAME: high_response_time_warning
8236 LOC: Config.warnings.high_rptm
8239 If the one-minute median response time exceeds this value,
8240 Squid prints a WARNING with debug level 0 to get the
8241 administrators attention. The value is in milliseconds.
8244 NAME: high_page_fault_warning
8246 LOC: Config.warnings.high_pf
8249 If the one-minute average page fault rate exceeds this
8250 value, Squid prints a WARNING with debug level 0 to get
8251 the administrators attention. The value is in page faults
8255 NAME: high_memory_warning
8257 LOC: Config.warnings.high_memory
8260 If the memory usage (as determined by mallinfo) exceeds
8261 this amount, Squid prints a WARNING with debug level 0 to get
8262 the administrators attention.
8265 NAME: sleep_after_fork
8266 COMMENT: (microseconds)
8268 LOC: Config.sleep_after_fork
8271 When this is set to a non-zero value, the main Squid process
8272 sleeps the specified number of microseconds after a fork()
8273 system call. This sleep may help the situation where your
8274 system reports fork() failures due to lack of (virtual)
8275 memory. Note, however, if you have a lot of child
8276 processes, these sleep delays will add up and your
8277 Squid will not service requests for some amount of time
8278 until all the child processes have been started.
8279 On Windows value less then 1000 (1 milliseconds) are
8283 NAME: windows_ipaddrchangemonitor
8284 IFDEF: _SQUID_WINDOWS_
8288 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8290 On Windows Squid by default will monitor IP address changes and will
8291 reconfigure itself after any detected event. This is very useful for
8292 proxies connected to internet with dial-up interfaces.
8293 In some cases (a Proxy server acting as VPN gateway is one) it could be
8294 desiderable to disable this behaviour setting this to 'off'.
8295 Note: after changing this, Squid service must be restarted.
8300 IFDEF: USE_SQUID_EUI
8302 LOC: Eui::TheConfig.euiLookup
8304 Whether to lookup the EUI or MAC address of a connected client.
8307 NAME: max_filedescriptors max_filedesc
8310 LOC: Config.max_filedescriptors
8312 The maximum number of filedescriptors supported.
8314 The default "0" means Squid inherits the current ulimit setting.
8316 Note: Changing this requires a restart of Squid. Also
8317 not all comm loops supports large values.
8325 Number of main Squid processes or "workers" to fork and maintain.
8326 0: "no daemon" mode, like running "squid -N ..."
8327 1: "no SMP" mode, start one main Squid process daemon (default)
8328 N: start N main Squid process daemons (i.e., SMP mode)
8330 In SMP mode, each worker does nearly all what a single Squid daemon
8331 does (e.g., listen on http_port and forward HTTP requests).
8334 NAME: cpu_affinity_map
8335 TYPE: CpuAffinityMap
8336 LOC: Config.cpuAffinityMap
8339 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8341 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8343 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8345 affects processes 1 through 4 only and places them on the first
8346 four even cores, starting with core #1.
8348 CPU cores are numbered starting from 1. Requires support for
8349 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8351 Multiple cpu_affinity_map options are merged.