2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the documentation for the Squid configuration file.
34 This documentation can also be found online at:
35 http://www.squid-cache.org/Doc/config/
37 You may wish to look at the Squid home page and wiki for the
38 FAQ and other documentation:
39 http://www.squid-cache.org/
40 http://wiki.squid-cache.org/SquidFaq
41 http://wiki.squid-cache.org/ConfigExamples
43 This documentation shows what the defaults for various directives
44 happen to be. If you don't need to change the default, you should
45 leave the line out of your squid.conf in most cases.
47 In some cases "none" refers to no default setting at all,
48 while in other cases it refers to the value of the option
49 - the comments for that keyword indicate if this is the case.
54 Configuration options can be included using the "include" directive.
55 Include takes a list of files to include. Quoting and wildcards are
60 include /path/to/included/file/squid.acl.config
62 Includes can be nested up to a hard-coded depth of 16 levels.
63 This arbitrary restriction is to prevent recursive include references
64 from causing Squid entering an infinite loop whilst trying to load
68 Conditional configuration
70 If-statements can be used to make configuration directives
74 ... regular configuration directives ...
76 ... regular configuration directives ...]
79 The else part is optional. The keywords "if", "else", and "endif"
80 must be typed on their own lines, as if they were regular
81 configuration directives.
83 NOTE: An else-if condition is not supported.
85 These individual conditions types are supported:
88 Always evaluates to true.
90 Always evaluates to false.
92 Equality comparison of two integer numbers.
97 The following SMP-related preprocessor macros can be used.
99 ${process_name} expands to the current Squid process "name"
100 (e.g., squid1, squid2, or cache1).
102 ${process_number} expands to the current Squid process
103 identifier, which is an integer number (e.g., 1, 2, 3) unique
104 across all Squid processes.
107 # Options Removed in 3.2
108 NAME: ignore_expect_100
111 Remove this line. The HTTP/1.1 feature is now fully supported by default.
114 NAME: dns_v4_fallback
123 Remove this line. Configure FTP page display using the CSS controls in errorpages.css instead.
126 NAME: maximum_single_addr_tries
129 Replaced by connect_retries. The behaviour has changed, please read the documentation before altering.
132 NAME: url_rewrite_concurrency
135 Remove this line. Set the 'concurrency=' option of url_rewrite_children instead.
138 # Options Removed in 3.1
142 Remove this line. DNS is no longer tested on startup.
145 NAME: extension_methods
148 Remove this line. All valid methods for HTTP are accepted by default.
151 # 2.7 Options Removed/Replaced in 3.2
156 # 2.7 Options Removed/Replaced in 3.1
164 Remove this line. HTTP/1.1 is supported by default.
167 NAME: upgrade_http0.9
170 Remove this line. ICY/1.0 streaming protocol is supported by default.
173 NAME: zph_local zph_mode zph_option zph_parent zph_sibling
176 Alter these entries. Use the qos_flows directive instead.
179 # Options Removed in 3.0
183 Since squid-3.0 replace with request_header_access or reply_header_access
184 depending on whether you wish to match client requests or server replies.
187 NAME: httpd_accel_no_pmtu_disc
190 Since squid-3.0 use the 'disable-pmtu-discovery' flag on http_port instead.
194 OPTIONS FOR AUTHENTICATION
195 -----------------------------------------------------------------------------
204 This is used to define parameters for the various authentication
205 schemes supported by Squid.
207 format: auth_param scheme parameter [setting]
209 The order in which authentication schemes are presented to the client is
210 dependent on the order the scheme first appears in config file. IE
211 has a bug (it's not RFC 2617 compliant) in that it will use the basic
212 scheme if basic is the first entry presented, even if more secure
213 schemes are presented. For now use the order in the recommended
214 settings section below. If other browsers have difficulties (don't
215 recognize the schemes offered even if you are using basic) either
216 put basic first, or disable the other schemes (by commenting out their
219 Once an authentication scheme is fully configured, it can only be
220 shutdown by shutting squid down and restarting. Changes can be made on
221 the fly and activated with a reconfigure. I.E. You can change to a
222 different helper, but not unconfigure the helper completely.
224 Please note that while this directive defines how Squid processes
225 authentication it does not automatically activate authentication.
226 To use authentication you must in addition make use of ACLs based
227 on login name in http_access (proxy_auth, proxy_auth_regex or
228 external with %LOGIN used in the format tag). The browser will be
229 challenged for authentication on the first such acl encountered
230 in http_access processing and will also be re-challenged for new
231 login credentials if the request is being denied by a proxy_auth
234 WARNING: authentication can't be used in a transparently intercepting
235 proxy as the client then thinks it is talking to an origin server and
236 not the proxy. This is a limitation of bending the TCP/IP protocol to
237 transparently intercepting port 80, not a limitation in Squid.
238 Ports flagged 'transparent', 'intercept', or 'tproxy' have
239 authentication disabled.
241 === Parameters for the basic scheme follow. ===
244 Specify the command for the external authenticator. Such a program
245 reads a line containing "username password" and replies with one of
252 the user does not exist.
255 An internal error occured in the helper, preventing
256 a result being identified.
258 "ERR" and "BH" results may optionally be followed by message="..."
259 containing a description available as %m in the returned error page.
261 If you use an authenticator, make sure you have 1 acl of type
264 By default, the basic authentication scheme is not used unless a
265 program is specified.
267 If you want to use the traditional NCSA proxy authentication, set
268 this line to something like
270 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
273 HTTP uses iso-latin-1 as characterset, while some authentication
274 backends such as LDAP expects UTF-8. If this is set to on Squid will
275 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
276 username & password to the helper.
278 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
279 The maximum number of authenticator processes to spawn. If you start too few
280 Squid will have to wait for them to process a backlog of credential
281 verifications, slowing it down. When password verifications are
282 done via a (slow) network you are likely to need lots of
283 authenticator processes.
285 The startup= and idle= options permit some skew in the exact amount
286 run. A minimum of startup=N will begin during startup and reconfigure.
287 Squid will start more in groups of up to idle=N in an attempt to meet
288 traffic needs and to keep idle=N free above those traffic needs up to
291 The concurrency= option sets the number of concurrent requests the
292 helper can process. The default of 0 is used for helpers who only
293 supports one request at a time. Setting this to a number greater than
294 0 changes the protocol used to include a channel number first on the
295 request/response line, allowing multiple requests to be sent to the
296 same helper in parallell without wating for the response.
297 Must not be set unless it's known the helper supports this.
299 auth_param basic children 20 startup=0 idle=1
302 Specifies the realm name which is to be reported to the
303 client for the basic proxy authentication scheme (part of
304 the text the user will see when prompted their username and
305 password). There is no default.
306 auth_param basic realm Squid proxy-caching web server
308 "credentialsttl" timetolive
309 Specifies how long squid assumes an externally validated
310 username:password pair is valid for - in other words how
311 often the helper program is called for that user. Set this
312 low to force revalidation with short lived passwords. Note
313 setting this high does not impact your susceptibility
314 to replay attacks unless you are using an one-time password
315 system (such as SecureID). If you are using such a system,
316 you will be vulnerable to replay attacks unless you also
317 use the max_user_ip ACL in an http_access rule.
319 "casesensitive" on|off
320 Specifies if usernames are case sensitive. Most user databases are
321 case insensitive allowing the same username to be spelled using both
322 lower and upper case letters, but some are case sensitive. This
323 makes a big difference for user_max_ip ACL processing and similar.
324 auth_param basic casesensitive off
326 === Parameters for the digest scheme follow ===
329 Specify the command for the external authenticator. Such
330 a program reads a line containing "username":"realm" and
331 replies with one of three results:
334 the user exists. The ha1= key is mandatory and
335 contains the appropriate H(A1) value, hex encoded.
336 See rfc 2616 for the definition of H(A1).
339 the user does not exist.
342 An internal error occured in the helper, preventing
343 a result being identified.
345 "ERR" and "BH" results may optionally be followed by message="..."
346 containing a description available as %m in the returned error page.
348 By default, the digest authentication scheme is not used unless a
349 program is specified.
351 If you want to use a digest authenticator, set this line to
354 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
357 HTTP uses iso-latin-1 as characterset, while some authentication
358 backends such as LDAP expects UTF-8. If this is set to on Squid will
359 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
360 username & password to the helper.
362 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
363 The maximum number of authenticator processes to spawn (default 5).
364 If you start too few Squid will have to wait for them to
365 process a backlog of H(A1) calculations, slowing it down.
366 When the H(A1) calculations are done via a (slow) network
367 you are likely to need lots of authenticator processes.
369 The startup= and idle= options permit some skew in the exact amount
370 run. A minimum of startup=N will begin during startup and reconfigure.
371 Squid will start more in groups of up to idle=N in an attempt to meet
372 traffic needs and to keep idle=N free above those traffic needs up to
375 The concurrency= option sets the number of concurrent requests the
376 helper can process. The default of 0 is used for helpers who only
377 supports one request at a time. Setting this to a number greater than
378 0 changes the protocol used to include a channel number first on the
379 request/response line, allowing multiple requests to be sent to the
380 same helper in parallell without wating for the response.
381 Must not be set unless it's known the helper supports this.
383 auth_param digest children 20 startup=0 idle=1
386 Specifies the realm name which is to be reported to the
387 client for the digest proxy authentication scheme (part of
388 the text the user will see when prompted their username and
389 password). There is no default.
390 auth_param digest realm Squid proxy-caching web server
392 "nonce_garbage_interval" timeinterval
393 Specifies the interval that nonces that have been issued
394 to client_agent's are checked for validity.
396 "nonce_max_duration" timeinterval
397 Specifies the maximum length of time a given nonce will be
400 "nonce_max_count" number
401 Specifies the maximum number of times a given nonce can be
404 "nonce_strictness" on|off
405 Determines if squid requires strict increment-by-1 behavior
406 for nonce counts, or just incrementing (off - for use when
407 useragents generate nonce counts that occasionally miss 1
408 (ie, 1,2,4,6)). Default off.
410 "check_nonce_count" on|off
411 This directive if set to off can disable the nonce count check
412 completely to work around buggy digest qop implementations in
413 certain mainstream browser versions. Default on to check the
414 nonce count to protect from authentication replay attacks.
416 "post_workaround" on|off
417 This is a workaround to certain buggy browsers who sends
418 an incorrect request digest in POST requests when reusing
419 the same nonce as acquired earlier on a GET request.
421 === NTLM scheme options follow ===
424 Specify the command for the external NTLM authenticator.
425 Such a program reads exchanged NTLMSSP packets with
426 the browser via Squid until authentication is completed.
427 If you use an NTLM authenticator, make sure you have 1 acl
428 of type proxy_auth. By default, the NTLM authenticator program
431 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
433 "children" numberofchildren [startup=N] [idle=N]
434 The maximum number of authenticator processes to spawn (default 5).
435 If you start too few Squid will have to wait for them to
436 process a backlog of credential verifications, slowing it
437 down. When credential verifications are done via a (slow)
438 network you are likely to need lots of authenticator
441 The startup= and idle= options permit some skew in the exact amount
442 run. A minimum of startup=N will begin during startup and reconfigure.
443 Squid will start more in groups of up to idle=N in an attempt to meet
444 traffic needs and to keep idle=N free above those traffic needs up to
447 auth_param ntlm children 20 startup=0 idle=1
450 If you experience problems with PUT/POST requests when using the
451 Negotiate authentication scheme then you can try setting this to
452 off. This will cause Squid to forcibly close the connection on
453 the initial requests where the browser asks which schemes are
454 supported by the proxy.
456 auth_param ntlm keep_alive on
458 === Options for configuring the NEGOTIATE auth-scheme follow ===
461 Specify the command for the external Negotiate authenticator.
462 This protocol is used in Microsoft Active-Directory enabled setups with
463 the Microsoft Internet Explorer or Mozilla Firefox browsers.
464 Its main purpose is to exchange credentials with the Squid proxy
465 using the Kerberos mechanisms.
466 If you use a Negotiate authenticator, make sure you have at least
467 one acl of type proxy_auth active. By default, the negotiate
468 authenticator program is not used.
469 The only supported program for this role is the ntlm_auth
470 program distributed as part of Samba, version 4 or later.
472 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
474 "children" numberofchildren [startup=N] [idle=N]
475 The maximum number of authenticator processes to spawn (default 5).
476 If you start too few Squid will have to wait for them to
477 process a backlog of credential verifications, slowing it
478 down. When crendential verifications are done via a (slow)
479 network you are likely to need lots of authenticator
482 The startup= and idle= options permit some skew in the exact amount
483 run. A minimum of startup=N will begin during startup and reconfigure.
484 Squid will start more in groups of up to idle=N in an attempt to meet
485 traffic needs and to keep idle=N free above those traffic needs up to
488 auth_param negotiate children 20 startup=0 idle=1
491 If you experience problems with PUT/POST requests when using the
492 Negotiate authentication scheme then you can try setting this to
493 off. This will cause Squid to forcibly close the connection on
494 the initial requests where the browser asks which schemes are
495 supported by the proxy.
497 auth_param negotiate keep_alive on
502 #Recommended minimum configuration per scheme:
503 #auth_param negotiate program <uncomment and complete this line to activate>
504 #auth_param negotiate children 20 startup=0 idle=1
505 #auth_param negotiate keep_alive on
507 #auth_param ntlm program <uncomment and complete this line to activate>
508 #auth_param ntlm children 20 startup=0 idle=1
509 #auth_param ntlm keep_alive on
511 #auth_param digest program <uncomment and complete this line>
512 #auth_param digest children 20 startup=0 idle=1
513 #auth_param digest realm Squid proxy-caching web server
514 #auth_param digest nonce_garbage_interval 5 minutes
515 #auth_param digest nonce_max_duration 30 minutes
516 #auth_param digest nonce_max_count 50
518 #auth_param basic program <uncomment and complete this line>
519 #auth_param basic children 5 startup=5 idle=1
520 #auth_param basic realm Squid proxy-caching web server
521 #auth_param basic credentialsttl 2 hours
524 NAME: authenticate_cache_garbage_interval
527 LOC: Config.authenticateGCInterval
529 The time period between garbage collection across the username cache.
530 This is a tradeoff between memory utilization (long intervals - say
531 2 days) and CPU (short intervals - say 1 minute). Only change if you
535 NAME: authenticate_ttl
538 LOC: Config.authenticateTTL
540 The time a user & their credentials stay in the logged in
541 user cache since their last request. When the garbage
542 interval passes, all user credentials that have passed their
543 TTL are removed from memory.
546 NAME: authenticate_ip_ttl
548 LOC: Config.authenticateIpTTL
551 If you use proxy authentication and the 'max_user_ip' ACL,
552 this directive controls how long Squid remembers the IP
553 addresses associated with each user. Use a small value
554 (e.g., 60 seconds) if your users might change addresses
555 quickly, as is the case with dialups. You might be safe
556 using a larger value (e.g., 2 hours) in a corporate LAN
557 environment with relatively static address assignments.
562 -----------------------------------------------------------------------------
565 NAME: external_acl_type
566 TYPE: externalAclHelper
567 LOC: Config.externalAclHelperList
570 This option defines external acl classes using a helper program
571 to look up the status
573 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
577 ttl=n TTL in seconds for cached results (defaults to 3600
580 TTL for cached negative lookups (default same
583 Maximum number of acl helper processes spawned to service
584 external acl lookups of this type. (default 20)
586 Minimum number of acl helper processes to spawn during
587 startup and reconfigure to service external acl lookups
588 of this type. (default 0)
590 Number of acl helper processes to keep ahead of traffic
591 loads. Squid will spawn this many at once whenever load
592 rises above the capabilities of existing processes.
593 Up to the value of children-max. (default 1)
594 concurrency=n concurrency level per process. Only used with helpers
595 capable of processing more than one query at a time.
596 cache=n limit the result cache size, default is unbounded.
597 grace=n Percentage remaining of TTL where a refresh of a
598 cached entry should be initiated without needing to
599 wait for a new reply. (default is for no grace period)
600 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
601 ipv4 / ipv6 IP protocol used to communicate with this helper.
602 The default is to auto-detect IPv6 and use it when available.
604 FORMAT specifications
606 %LOGIN Authenticated user login name
607 %EXT_USER Username from previous external acl
608 %EXT_LOG Log details from previous external acl
609 %EXT_TAG Tag from previous external acl
610 %IDENT Ident user name
612 %SRCPORT Client source port
615 %PROTO Requested protocol
617 %PATH Requested URL path
618 %METHOD Request method
619 %MYADDR Squid interface address
620 %MYPORT Squid http_port number
621 %PATH Requested URL-path (including query-string if any)
622 %USER_CERT SSL User certificate in PEM format
623 %USER_CERTCHAIN SSL User certificate chain in PEM format
624 %USER_CERT_xx SSL User certificate subject attribute xx
625 %USER_CA_xx SSL User certificate issuer attribute xx
627 %>{Header} HTTP request header "Header"
629 HTTP request header "Hdr" list member "member"
631 HTTP request header list member using ; as
632 list separator. ; can be any non-alphanumeric
635 %<{Header} HTTP reply header "Header"
637 HTTP reply header "Hdr" list member "member"
639 HTTP reply header list member using ; as
640 list separator. ; can be any non-alphanumeric
643 %% The percent sign. Useful for helpers which need
644 an unchanging input format.
647 General request syntax:
649 [channel-ID] FORMAT-values [acl-values ...]
652 FORMAT-values consists of transaction details expanded with
653 whitespace separation per the config file FORMAT specification
654 using the FORMAT macros listed above.
656 acl-values consists of any string specified in the referencing
657 config 'acl ... external' line. see the "acl external" directive.
659 Request values sent to the helper are URL escaped to protect
660 each value in requests against whitespaces.
662 If using protocol=2.5 then the request sent to the helper is not
663 URL escaped to protect against whitespace.
665 NOTE: protocol=3.0 is deprecated as no longer necessary.
667 When using the concurrency= option the protocol is changed by
668 introducing a query channel tag in front of the request/response.
669 The query channel tag is a number between 0 and concurrency-1.
670 This value must be echoed back unchanged to Squid as the first part
671 of the response relating to its request.
674 The helper receives lines expanded per the above format specification
675 and for each input line returns 1 line starting with OK/ERR/BH result
676 code and optionally followed by additional keywords with more details.
679 General result syntax:
681 [channel-ID] result keyword=value ...
683 Result consists of one of the codes:
686 the ACL test produced a match.
689 the ACL test does not produce a match.
692 An internal error occured in the helper, preventing
693 a result being identified.
695 The meaning of 'a match' is determined by your squid.conf
696 access control configuration. See the Squid wiki for details.
700 user= The users name (login)
702 password= The users password (for login= cache_peer option)
704 message= Message describing the reason for this response.
705 Available as %o in error pages.
706 Useful on (ERR and BH results).
708 tag= Apply a tag to a request. Only sets a tag once,
709 does not alter existing tags.
711 log= String to be logged in access.log. Available as
712 %ea in logformat specifications.
714 Any keywords may be sent on any response whether OK, ERR or BH.
716 All response keyword values need to be a single token with URL
717 escaping, or enclosed in double quotes (") and escaped using \ on
718 any double quotes or \ characters within the value. The wrapping
719 double quotes are removed before the value is interpreted by Squid.
720 \r and \n are also replace by CR and LF.
722 Some example key values:
726 user="J. \"Bob\" Smith"
733 DEFAULT: ssl::certHasExpired ssl_error X509_V_ERR_CERT_HAS_EXPIRED
734 DEFAULT: ssl::certNotYetValid ssl_error X509_V_ERR_CERT_NOT_YET_VALID
735 DEFAULT: ssl::certDomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
736 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
737 DEFAULT: ssl::certSelfSigned ssl_error X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT
740 DEFAULT: manager url_regex -i ^cache_object:// +i ^https?://[^/]+/squid-internal-mgr/
741 DEFAULT: localhost src 127.0.0.1/32 ::1
742 DEFAULT: to_localhost dst 127.0.0.0/8 0.0.0.0/32 ::1
743 DEFAULT_DOC: ACLs all, manager, localhost, and to_localhost are predefined.
745 Defining an Access List
747 Every access list definition must begin with an aclname and acltype,
748 followed by either type-specific arguments or a quoted filename that
751 acl aclname acltype argument ...
752 acl aclname acltype "file" ...
754 When using "file", the file should contain one item per line.
756 By default, regular expressions are CASE-SENSITIVE.
757 To make them case-insensitive, use the -i option. To return case-sensitive
758 use the +i option between patterns, or make a new ACL line without -i.
760 Some acl types require suspending the current request in order
761 to access some external data source.
762 Those which do are marked with the tag [slow], those which
763 don't are marked as [fast].
764 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
765 for further information
767 ***** ACL TYPES AVAILABLE *****
769 acl aclname src ip-address/mask ... # clients IP address [fast]
770 acl aclname src addr1-addr2/mask ... # range of addresses [fast]
771 acl aclname dst ip-address/mask ... # URL host's IP address [slow]
772 acl aclname localip ip-address/mask ... # IP address the client connected to [fast]
774 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
775 # The arp ACL requires the special configure option --enable-arp-acl.
776 # Furthermore, the ARP ACL code is not portable to all operating systems.
777 # It works on Linux, Solaris, Windows, FreeBSD, and some
778 # other *BSD variants.
781 # NOTE: Squid can only determine the MAC address for clients that are on
782 # the same subnet. If the client is on a different subnet,
783 # then Squid cannot find out its MAC address.
785 acl aclname srcdomain .foo.com ...
786 # reverse lookup, from client IP [slow]
787 acl aclname dstdomain .foo.com ...
788 # Destination server from URL [fast]
789 acl aclname srcdom_regex [-i] \.foo\.com ...
790 # regex matching client name [slow]
791 acl aclname dstdom_regex [-i] \.foo\.com ...
792 # regex matching server [fast]
794 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
795 # based URL is used and no match is found. The name "none" is used
796 # if the reverse lookup fails.
798 acl aclname src_as number ...
799 acl aclname dst_as number ...
801 # Except for access control, AS numbers can be used for
802 # routing of requests to specific caches. Here's an
803 # example for routing all requests for AS#1241 and only
804 # those to mycache.mydomain.net:
805 # acl asexample dst_as 1241
806 # cache_peer_access mycache.mydomain.net allow asexample
807 # cache_peer_access mycache_mydomain.net deny all
809 acl aclname peername myPeer ...
811 # match against a named cache_peer entry
812 # set unique name= on cache_peer lines for reliable use.
814 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
824 # h1:m1 must be less than h2:m2
826 acl aclname url_regex [-i] ^http:// ...
827 # regex matching on whole URL [fast]
828 acl aclname urllogin [-i] [^a-zA-Z0-9] ...
829 # regex matching on URL login field
830 acl aclname urlpath_regex [-i] \.gif$ ...
831 # regex matching on URL path [fast]
833 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
835 acl aclname localport 3128 ... # TCP port the client connected to [fast]
836 # NP: for interception mode this is usually '80'
838 acl aclname myportname 3128 ... # http(s)_port name [fast]
840 acl aclname proto HTTP FTP ... # request protocol [fast]
842 acl aclname method GET POST ... # HTTP request method [fast]
844 acl aclname http_status 200 301 500- 400-403 ...
845 # status code in reply [fast]
847 acl aclname browser [-i] regexp ...
848 # pattern match on User-Agent header (see also req_header below) [fast]
850 acl aclname referer_regex [-i] regexp ...
851 # pattern match on Referer header [fast]
852 # Referer is highly unreliable, so use with care
854 acl aclname ident username ...
855 acl aclname ident_regex [-i] pattern ...
856 # string match on ident output [slow]
857 # use REQUIRED to accept any non-null ident.
859 acl aclname proxy_auth [-i] username ...
860 acl aclname proxy_auth_regex [-i] pattern ...
861 # perform http authentication challenge to the client and match against
862 # supplied credentials [slow]
864 # takes a list of allowed usernames.
865 # use REQUIRED to accept any valid username.
867 # Will use proxy authentication in forward-proxy scenarios, and plain
868 # http authenticaiton in reverse-proxy scenarios
870 # NOTE: when a Proxy-Authentication header is sent but it is not
871 # needed during ACL checking the username is NOT logged
874 # NOTE: proxy_auth requires a EXTERNAL authentication program
875 # to check username/password combinations (see
876 # auth_param directive).
878 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
879 # as the browser needs to be configured for using a proxy in order
880 # to respond to proxy authentication.
882 acl aclname snmp_community string ...
883 # A community string to limit access to your SNMP Agent [fast]
886 # acl snmppublic snmp_community public
888 acl aclname maxconn number
889 # This will be matched when the client's IP address has
890 # more than <number> TCP connections established. [fast]
891 # NOTE: This only measures direct TCP links so X-Forwarded-For
892 # indirect clients are not counted.
894 acl aclname max_user_ip [-s] number
895 # This will be matched when the user attempts to log in from more
896 # than <number> different ip addresses. The authenticate_ip_ttl
897 # parameter controls the timeout on the ip entries. [fast]
898 # If -s is specified the limit is strict, denying browsing
899 # from any further IP addresses until the ttl has expired. Without
900 # -s Squid will just annoy the user by "randomly" denying requests.
901 # (the counter is reset each time the limit is reached and a
903 # NOTE: in acceleration mode or where there is mesh of child proxies,
904 # clients may appear to come from multiple addresses if they are
905 # going through proxy farms, so a limit of 1 may cause user problems.
907 acl aclname random probability
908 # Pseudo-randomly match requests. Based on the probability given.
909 # Probability may be written as a decimal (0.333), fraction (1/3)
910 # or ratio of matches:non-matches (3:5).
912 acl aclname req_mime_type [-i] mime-type ...
913 # regex match against the mime type of the request generated
914 # by the client. Can be used to detect file upload or some
915 # types HTTP tunneling requests [fast]
916 # NOTE: This does NOT match the reply. You cannot use this
917 # to match the returned file type.
919 acl aclname req_header header-name [-i] any\.regex\.here
920 # regex match against any of the known request headers. May be
921 # thought of as a superset of "browser", "referer" and "mime-type"
924 acl aclname rep_mime_type [-i] mime-type ...
925 # regex match against the mime type of the reply received by
926 # squid. Can be used to detect file download or some
927 # types HTTP tunneling requests. [fast]
928 # NOTE: This has no effect in http_access rules. It only has
929 # effect in rules that affect the reply data stream such as
932 acl aclname rep_header header-name [-i] any\.regex\.here
933 # regex match against any of the known reply headers. May be
934 # thought of as a superset of "browser", "referer" and "mime-type"
937 acl aclname external class_name [arguments...]
938 # external ACL lookup via a helper class defined by the
939 # external_acl_type directive [slow]
941 acl aclname user_cert attribute values...
942 # match against attributes in a user SSL certificate
943 # attribute is one of DN/C/O/CN/L/ST [fast]
945 acl aclname ca_cert attribute values...
946 # match against attributes a users issuing CA SSL certificate
947 # attribute is one of DN/C/O/CN/L/ST [fast]
949 acl aclname ext_user username ...
950 acl aclname ext_user_regex [-i] pattern ...
951 # string match on username returned by external acl helper [slow]
952 # use REQUIRED to accept any non-null user name.
954 acl aclname tag tagvalue ...
955 # string match on tag returned by external acl helper [slow]
957 acl aclname hier_code codename ...
958 # string match against squid hierarchy code(s); [fast]
959 # e.g., DIRECT, PARENT_HIT, NONE, etc.
961 # NOTE: This has no effect in http_access rules. It only has
962 # effect in rules that affect the reply data stream such as
966 acl aclname ssl_error errorname
967 # match against SSL certificate validation error [fast]
969 # For valid error names see in @DEFAULT_ERROR_DIR@/templates/error-details.txt
972 # The following can be used as shortcuts for certificate properties:
973 # [ssl::]certHasExpired: the "not after" field is in the past
974 # [ssl::]certNotYetValid: the "not before" field is in the future
975 # [ssl::]certUntrusted: The certificate issuer is not to be trusted.
976 # [ssl::]certSelfSigned: The certificate is self signed.
977 # [ssl::]certDomainMismatch: The certificate CN domain does not
978 # match the name the name of the host we are connecting to.
980 # The ssl::certHasExpired, ssl::certNotYetValid, ssl::certDomainMismatch,
981 # ssl::certUntrusted, and ssl::certSelfSigned can also be used as
982 # predefined ACLs, just like the 'all' ACL.
984 # NOTE: The ssl_error ACL is only supported with sslproxy_cert_error,
985 # sslproxy_cert_sign, and sslproxy_cert_adapt options.
987 acl aclname server_cert_fingerprint [-sha1] fingerprint
988 # match against server SSL certificate fingerprint [fast]
990 # The fingerprint is the digest of the DER encoded version
991 # of the whole certificate. The user should use the form: XX:XX:...
992 # Optional argument specifies the digest algorithm to use.
993 # The SHA1 digest algorithm is the default and is currently
994 # the only algorithm supported (-sha1).
998 acl macaddress arp 09:00:2b:23:45:67
999 acl myexample dst_as 1241
1000 acl password proxy_auth REQUIRED
1001 acl fileupload req_mime_type -i ^multipart/form-data$
1002 acl javascript rep_mime_type -i ^application/x-javascript$
1006 # Recommended minimum configuration:
1009 # Example rule allowing access from your local networks.
1010 # Adapt to list your (internal) IP networks from where browsing
1012 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
1013 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
1014 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
1015 acl localnet src fc00::/7 # RFC 4193 local private network range
1016 acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
1018 acl SSL_ports port 443
1019 acl Safe_ports port 80 # http
1020 acl Safe_ports port 21 # ftp
1021 acl Safe_ports port 443 # https
1022 acl Safe_ports port 70 # gopher
1023 acl Safe_ports port 210 # wais
1024 acl Safe_ports port 1025-65535 # unregistered ports
1025 acl Safe_ports port 280 # http-mgmt
1026 acl Safe_ports port 488 # gss-http
1027 acl Safe_ports port 591 # filemaker
1028 acl Safe_ports port 777 # multiling http
1029 acl CONNECT method CONNECT
1033 NAME: follow_x_forwarded_for
1035 IFDEF: FOLLOW_X_FORWARDED_FOR
1036 LOC: Config.accessList.followXFF
1037 DEFAULT_IF_NONE: deny all
1039 Allowing or Denying the X-Forwarded-For header to be followed to
1040 find the original source of a request.
1042 Requests may pass through a chain of several other proxies
1043 before reaching us. The X-Forwarded-For header will contain a
1044 comma-separated list of the IP addresses in the chain, with the
1045 rightmost address being the most recent.
1047 If a request reaches us from a source that is allowed by this
1048 configuration item, then we consult the X-Forwarded-For header
1049 to see where that host received the request from. If the
1050 X-Forwarded-For header contains multiple addresses, we continue
1051 backtracking until we reach an address for which we are not allowed
1052 to follow the X-Forwarded-For header, or until we reach the first
1053 address in the list. For the purpose of ACL used in the
1054 follow_x_forwarded_for directive the src ACL type always matches
1055 the address we are testing and srcdomain matches its rDNS.
1057 The end result of this process is an IP address that we will
1058 refer to as the indirect client address. This address may
1059 be treated as the client address for access control, ICAP, delay
1060 pools and logging, depending on the acl_uses_indirect_client,
1061 icap_uses_indirect_client, delay_pool_uses_indirect_client,
1062 log_uses_indirect_client and tproxy_uses_indirect_client options.
1064 This clause only supports fast acl types.
1065 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1067 SECURITY CONSIDERATIONS:
1069 Any host for which we follow the X-Forwarded-For header
1070 can place incorrect information in the header, and Squid
1071 will use the incorrect information as if it were the
1072 source address of the request. This may enable remote
1073 hosts to bypass any access control restrictions that are
1074 based on the client's source addresses.
1078 acl localhost src 127.0.0.1
1079 acl my_other_proxy srcdomain .proxy.example.com
1080 follow_x_forwarded_for allow localhost
1081 follow_x_forwarded_for allow my_other_proxy
1084 NAME: acl_uses_indirect_client
1087 IFDEF: FOLLOW_X_FORWARDED_FOR
1089 LOC: Config.onoff.acl_uses_indirect_client
1091 Controls whether the indirect client address
1092 (see follow_x_forwarded_for) is used instead of the
1093 direct client address in acl matching.
1095 NOTE: maxconn ACL considers direct TCP links and indirect
1096 clients will always have zero. So no match.
1099 NAME: delay_pool_uses_indirect_client
1102 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_DELAY_POOLS
1104 LOC: Config.onoff.delay_pool_uses_indirect_client
1106 Controls whether the indirect client address
1107 (see follow_x_forwarded_for) is used instead of the
1108 direct client address in delay pools.
1111 NAME: log_uses_indirect_client
1114 IFDEF: FOLLOW_X_FORWARDED_FOR
1116 LOC: Config.onoff.log_uses_indirect_client
1118 Controls whether the indirect client address
1119 (see follow_x_forwarded_for) is used instead of the
1120 direct client address in the access log.
1123 NAME: tproxy_uses_indirect_client
1126 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
1128 LOC: Config.onoff.tproxy_uses_indirect_client
1130 Controls whether the indirect client address
1131 (see follow_x_forwarded_for) is used instead of the
1132 direct client address when spoofing the outgoing client.
1134 This has no effect on requests arriving in non-tproxy
1137 SECURITY WARNING: Usage of this option is dangerous
1138 and should not be used trivially. Correct configuration
1139 of follow_x_forewarded_for with a limited set of trusted
1140 sources is required to prevent abuse of your proxy.
1145 LOC: Config.accessList.http
1146 DEFAULT_IF_NONE: deny all
1148 Allowing or Denying access based on defined access lists
1150 Access to the HTTP port:
1151 http_access allow|deny [!]aclname ...
1153 NOTE on default values:
1155 If there are no "access" lines present, the default is to deny
1158 If none of the "access" lines cause a match, the default is the
1159 opposite of the last line in the list. If the last line was
1160 deny, the default is allow. Conversely, if the last line
1161 is allow, the default will be deny. For these reasons, it is a
1162 good idea to have an "deny all" entry at the end of your access
1163 lists to avoid potential confusion.
1165 This clause supports both fast and slow acl types.
1166 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1171 # Recommended minimum Access Permission configuration:
1173 # Only allow cachemgr access from localhost
1174 http_access allow localhost manager
1175 http_access deny manager
1177 # Deny requests to certain unsafe ports
1178 http_access deny !Safe_ports
1180 # Deny CONNECT to other than secure SSL ports
1181 http_access deny CONNECT !SSL_ports
1183 # We strongly recommend the following be uncommented to protect innocent
1184 # web applications running on the proxy server who think the only
1185 # one who can access services on "localhost" is a local user
1186 #http_access deny to_localhost
1189 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
1192 # Example rule allowing access from your local networks.
1193 # Adapt localnet in the ACL section to list your (internal) IP networks
1194 # from where browsing should be allowed
1195 http_access allow localnet
1196 http_access allow localhost
1198 # And finally deny all other access to this proxy
1199 http_access deny all
1203 NAME: adapted_http_access http_access2
1205 LOC: Config.accessList.adapted_http
1208 Allowing or Denying access based on defined access lists
1210 Essentially identical to http_access, but runs after redirectors
1211 and ICAP/eCAP adaptation. Allowing access control based on their
1214 If not set then only http_access is used.
1217 NAME: http_reply_access
1219 LOC: Config.accessList.reply
1222 Allow replies to client requests. This is complementary to http_access.
1224 http_reply_access allow|deny [!] aclname ...
1226 NOTE: if there are no access lines present, the default is to allow
1229 If none of the access lines cause a match the opposite of the
1230 last line will apply. Thus it is good practice to end the rules
1231 with an "allow all" or "deny all" entry.
1233 This clause supports both fast and slow acl types.
1234 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1239 LOC: Config.accessList.icp
1240 DEFAULT_IF_NONE: deny all
1242 Allowing or Denying access to the ICP port based on defined
1245 icp_access allow|deny [!]aclname ...
1247 See http_access for details
1249 This clause only supports fast acl types.
1250 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1252 # Allow ICP queries from local networks only
1253 #icp_access allow localnet
1254 #icp_access deny all
1260 LOC: Config.accessList.htcp
1261 DEFAULT_IF_NONE: deny all
1263 Allowing or Denying access to the HTCP port based on defined
1266 htcp_access allow|deny [!]aclname ...
1268 See http_access for details
1270 NOTE: The default if no htcp_access lines are present is to
1271 deny all traffic. This default may cause problems with peers
1272 using the htcp option.
1274 This clause only supports fast acl types.
1275 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1277 # Allow HTCP queries from local networks only
1278 #htcp_access allow localnet
1279 #htcp_access deny all
1282 NAME: htcp_clr_access
1285 LOC: Config.accessList.htcp_clr
1286 DEFAULT_IF_NONE: deny all
1288 Allowing or Denying access to purge content using HTCP based
1289 on defined access lists
1291 htcp_clr_access allow|deny [!]aclname ...
1293 See http_access for details
1295 This clause only supports fast acl types.
1296 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1298 # Allow HTCP CLR requests from trusted peers
1299 acl htcp_clr_peer src 172.16.1.2
1300 htcp_clr_access allow htcp_clr_peer
1305 LOC: Config.accessList.miss
1308 Determins whether network access is permitted when satisfying a request.
1311 to force your neighbors to use you as a sibling instead of
1314 acl localclients src 172.16.0.0/16
1315 miss_access allow localclients
1316 miss_access deny !localclients
1318 This means only your local clients are allowed to fetch relayed/MISS
1319 replies from the network and all other clients can only fetch cached
1323 The default for this setting allows all clients who passed the
1324 http_access rules to relay via this proxy.
1326 This clause only supports fast acl types.
1327 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1330 NAME: ident_lookup_access
1333 DEFAULT_IF_NONE: deny all
1334 LOC: Ident::TheConfig.identLookup
1336 A list of ACL elements which, if matched, cause an ident
1337 (RFC 931) lookup to be performed for this request. For
1338 example, you might choose to always perform ident lookups
1339 for your main multi-user Unix boxes, but not for your Macs
1340 and PCs. By default, ident lookups are not performed for
1343 To enable ident lookups for specific client addresses, you
1344 can follow this example:
1346 acl ident_aware_hosts src 198.168.1.0/24
1347 ident_lookup_access allow ident_aware_hosts
1348 ident_lookup_access deny all
1350 Only src type ACL checks are fully supported. A srcdomain
1351 ACL might work at times, but it will not always provide
1354 This clause only supports fast acl types.
1355 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1358 NAME: reply_body_max_size
1359 COMMENT: size [acl acl...]
1362 LOC: Config.ReplyBodySize
1364 This option specifies the maximum size of a reply body. It can be
1365 used to prevent users from downloading very large files, such as
1366 MP3's and movies. When the reply headers are received, the
1367 reply_body_max_size lines are processed, and the first line where
1368 all (if any) listed ACLs are true is used as the maximum body size
1371 This size is checked twice. First when we get the reply headers,
1372 we check the content-length value. If the content length value exists
1373 and is larger than the allowed size, the request is denied and the
1374 user receives an error message that says "the request or reply
1375 is too large." If there is no content-length, and the reply
1376 size exceeds this limit, the client's connection is just closed
1377 and they will receive a partial reply.
1379 WARNING: downstream caches probably can not detect a partial reply
1380 if there is no content-length header, so they will cache
1381 partial responses and give them out as hits. You should NOT
1382 use this option if you have downstream caches.
1384 WARNING: A maximum size smaller than the size of squid's error messages
1385 will cause an infinite loop and crash squid. Ensure that the smallest
1386 non-zero value you use is greater that the maximum header size plus
1387 the size of your largest error page.
1389 If you set this parameter none (the default), there will be
1392 Configuration Format is:
1393 reply_body_max_size SIZE UNITS [acl ...]
1395 reply_body_max_size 10 MB
1401 -----------------------------------------------------------------------------
1404 NAME: http_port ascii_port
1407 LOC: Config.Sockaddr.http
1409 Usage: port [mode] [options]
1410 hostname:port [mode] [options]
1411 1.2.3.4:port [mode] [options]
1413 The socket addresses where Squid will listen for HTTP client
1414 requests. You may specify multiple socket addresses.
1415 There are three forms: port alone, hostname with port, and
1416 IP address with port. If you specify a hostname or IP
1417 address, Squid binds the socket to that specific
1418 address. Most likely, you do not need to bind to a specific
1419 address, so you can use the port number alone.
1421 If you are running Squid in accelerator mode, you
1422 probably want to listen on port 80 also, or instead.
1424 The -a command line option may be used to specify additional
1425 port(s) where Squid listens for proxy request. Such ports will
1426 be plain proxy ports with no options.
1428 You may specify multiple socket addresses on multiple lines.
1432 intercept Support for IP-Layer interception of
1433 outgoing requests without browser settings.
1434 NP: disables authentication and IPv6 on the port.
1436 tproxy Support Linux TPROXY for spoofing outgoing
1437 connections using the client IP address.
1438 NP: disables authentication and maybe IPv6 on the port.
1440 accel Accelerator / reverse proxy mode
1442 ssl-bump For each CONNECT request allowed by ssl_bump ACLs,
1443 establish secure connection with the client and with
1444 the server, decrypt HTTPS messages as they pass through
1445 Squid, and treat them as unencrypted HTTP messages,
1446 becoming the man-in-the-middle.
1448 The ssl_bump option is required to fully enable
1449 bumping of CONNECT requests.
1451 Omitting the mode flag causes default forward proxy mode to be used.
1454 Accelerator Mode Options:
1456 defaultsite=domainname
1457 What to use for the Host: header if it is not present
1458 in a request. Determines what site (not origin server)
1459 accelerators should consider the default.
1461 no-vhost Disable using HTTP/1.1 Host header for virtual domain support.
1463 protocol= Protocol to reconstruct accelerated requests with.
1464 Defaults to http for http_port and https for
1467 vport Virtual host port support. Using the http_port number
1468 instead of the port passed on Host: headers.
1470 vport=NN Virtual host port support. Using the specified port
1471 number instead of the port passed on Host: headers.
1474 Act as if this Squid is the origin server.
1475 This currently means generate new Date: and Expires:
1476 headers on HIT instead of adding Age:.
1478 ignore-cc Ignore request Cache-Control headers.
1480 WARNING: This option violates HTTP specifications if
1481 used in non-accelerator setups.
1483 allow-direct Allow direct forwarding in accelerator mode. Normally
1484 accelerated requests are denied direct forwarding as if
1485 never_direct was used.
1487 WARNING: this option opens accelerator mode to security
1488 vulnerabilities usually only affecting in interception
1489 mode. Make sure to protect forwarding with suitable
1490 http_access rules when using this.
1493 SSL Bump Mode Options:
1494 In addition to these options ssl-bump requires TLS/SSL options.
1496 generate-host-certificates[=<on|off>]
1497 Dynamically create SSL server certificates for the
1498 destination hosts of bumped CONNECT requests.When
1499 enabled, the cert and key options are used to sign
1500 generated certificates. Otherwise generated
1501 certificate will be selfsigned.
1502 If there is a CA certificate lifetime of the generated
1503 certificate equals lifetime of the CA certificate. If
1504 generated certificate is selfsigned lifetime is three
1506 This option is enabled by default when ssl-bump is used.
1507 See the ssl-bump option above for more information.
1509 dynamic_cert_mem_cache_size=SIZE
1510 Approximate total RAM size spent on cached generated
1511 certificates. If set to zero, caching is disabled. The
1512 default value is 4MB.
1516 cert= Path to SSL certificate (PEM format).
1518 key= Path to SSL private key file (PEM format)
1519 if not specified, the certificate file is
1520 assumed to be a combined certificate and
1523 version= The version of SSL/TLS supported
1524 1 automatic (default)
1531 cipher= Colon separated list of supported ciphers.
1532 NOTE: some ciphers such as EDH ciphers depend on
1533 additional settings. If those settings are
1534 omitted the ciphers may be silently ignored
1535 by the OpenSSL library.
1537 options= Various SSL implementation options. The most important
1539 NO_SSLv2 Disallow the use of SSLv2
1540 NO_SSLv3 Disallow the use of SSLv3
1541 NO_TLSv1 Disallow the use of TLSv1.0
1542 NO_TLSv1_1 Disallow the use of TLSv1.1
1543 NO_TLSv1_2 Disallow the use of TLSv1.2
1544 SINGLE_DH_USE Always create a new key when using
1545 temporary/ephemeral DH key exchanges
1546 ALL Enable various bug workarounds
1547 suggested as "harmless" by OpenSSL
1548 Be warned that this reduces SSL/TLS
1549 strength to some attacks.
1550 See OpenSSL SSL_CTX_set_options documentation for a
1551 complete list of options.
1553 clientca= File containing the list of CAs to use when
1554 requesting a client certificate.
1556 cafile= File containing additional CA certificates to
1557 use when verifying client certificates. If unset
1558 clientca will be used.
1560 capath= Directory containing additional CA certificates
1561 and CRL lists to use when verifying client certificates.
1563 crlfile= File of additional CRL lists to use when verifying
1564 the client certificate, in addition to CRLs stored in
1565 the capath. Implies VERIFY_CRL flag below.
1567 dhparams= File containing DH parameters for temporary/ephemeral
1568 DH key exchanges. See OpenSSL documentation for details
1569 on how to create this file.
1570 WARNING: EDH ciphers will be silently disabled if this
1573 sslflags= Various flags modifying the use of SSL:
1575 Don't request client certificates
1576 immediately, but wait until acl processing
1577 requires a certificate (not yet implemented).
1579 Don't use the default CA lists built in
1582 Don't allow for session reuse. Each connection
1583 will result in a new SSL session.
1585 Verify CRL lists when accepting client
1588 Verify CRL lists for all certificates in the
1589 client certificate chain.
1591 sslcontext= SSL session ID context identifier.
1595 connection-auth[=on|off]
1596 use connection-auth=off to tell Squid to prevent
1597 forwarding Microsoft connection oriented authentication
1598 (NTLM, Negotiate and Kerberos)
1600 disable-pmtu-discovery=
1601 Control Path-MTU discovery usage:
1602 off lets OS decide on what to do (default).
1603 transparent disable PMTU discovery when transparent
1605 always disable always PMTU discovery.
1607 In many setups of transparently intercepting proxies
1608 Path-MTU discovery can not work on traffic towards the
1609 clients. This is the case when the intercepting device
1610 does not fully track connections and fails to forward
1611 ICMP must fragment messages to the cache server. If you
1612 have such setup and experience that certain clients
1613 sporadically hang or never complete requests set
1614 disable-pmtu-discovery option to 'transparent'.
1616 name= Specifies a internal name for the port. Defaults to
1617 the port specification (port or addr:port)
1619 tcpkeepalive[=idle,interval,timeout]
1620 Enable TCP keepalive probes of idle connections.
1621 In seconds; idle is the initial time before TCP starts
1622 probing the connection, interval how often to probe, and
1623 timeout the time before giving up.
1625 If you run Squid on a dual-homed machine with an internal
1626 and an external interface we recommend you to specify the
1627 internal address:port in http_port. This way Squid will only be
1628 visible on the internal address.
1632 # Squid normally listens to port 3128
1633 http_port @DEFAULT_HTTP_PORT@
1641 LOC: Config.Sockaddr.https
1643 Usage: [ip:]port cert=certificate.pem [key=key.pem] [mode] [options...]
1645 The socket address where Squid will listen for client requests made
1646 over TLS or SSL connections. Commonly referred to as HTTPS.
1648 This is most useful for situations where you are running squid in
1649 accelerator mode and you want to do the SSL work at the accelerator level.
1651 You may specify multiple socket addresses on multiple lines,
1652 each with their own SSL certificate and/or options.
1656 accel Accelerator / reverse proxy mode
1658 intercept Support for IP-Layer interception of
1659 outgoing requests without browser settings.
1660 NP: disables authentication and IPv6 on the port.
1662 tproxy Support Linux TPROXY for spoofing outgoing
1663 connections using the client IP address.
1664 NP: disables authentication and maybe IPv6 on the port.
1666 ssl-bump For each intercepted connection allowed by ssl_bump
1667 ACLs, establish a secure connection with the client and with
1668 the server, decrypt HTTPS messages as they pass through
1669 Squid, and treat them as unencrypted HTTP messages,
1670 becoming the man-in-the-middle.
1672 An "ssl_bump server-first" match is required to
1673 fully enable bumping of intercepted SSL connections.
1675 Requires tproxy or intercept.
1677 Omitting the mode flag causes default forward proxy mode to be used.
1680 See http_port for a list of generic options
1685 cert= Path to SSL certificate (PEM format).
1687 key= Path to SSL private key file (PEM format)
1688 if not specified, the certificate file is
1689 assumed to be a combined certificate and
1692 version= The version of SSL/TLS supported
1693 1 automatic (default)
1698 cipher= Colon separated list of supported ciphers.
1700 options= Various SSL engine options. The most important
1702 NO_SSLv2 Disallow the use of SSLv2
1703 NO_SSLv3 Disallow the use of SSLv3
1704 NO_TLSv1 Disallow the use of TLSv1
1705 SINGLE_DH_USE Always create a new key when using
1706 temporary/ephemeral DH key exchanges
1707 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1708 documentation for a complete list of options.
1710 clientca= File containing the list of CAs to use when
1711 requesting a client certificate.
1713 cafile= File containing additional CA certificates to
1714 use when verifying client certificates. If unset
1715 clientca will be used.
1717 capath= Directory containing additional CA certificates
1718 and CRL lists to use when verifying client certificates.
1720 crlfile= File of additional CRL lists to use when verifying
1721 the client certificate, in addition to CRLs stored in
1722 the capath. Implies VERIFY_CRL flag below.
1724 dhparams= File containing DH parameters for temporary/ephemeral
1727 sslflags= Various flags modifying the use of SSL:
1729 Don't request client certificates
1730 immediately, but wait until acl processing
1731 requires a certificate (not yet implemented).
1733 Don't use the default CA lists built in
1736 Don't allow for session reuse. Each connection
1737 will result in a new SSL session.
1739 Verify CRL lists when accepting client
1742 Verify CRL lists for all certificates in the
1743 client certificate chain.
1745 sslcontext= SSL session ID context identifier.
1747 generate-host-certificates[=<on|off>]
1748 Dynamically create SSL server certificates for the
1749 destination hosts of bumped SSL requests.When
1750 enabled, the cert and key options are used to sign
1751 generated certificates. Otherwise generated
1752 certificate will be selfsigned.
1753 If there is CA certificate life time of generated
1754 certificate equals lifetime of CA certificate. If
1755 generated certificate is selfsigned lifetime is three
1757 This option is enabled by default when SslBump is used.
1758 See the sslBump option above for more information.
1760 dynamic_cert_mem_cache_size=SIZE
1761 Approximate total RAM size spent on cached generated
1762 certificates. If set to zero, caching is disabled. The
1763 default value is 4MB.
1765 See http_port for a list of available options.
1768 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1771 LOC: Ip::Qos::TheConfig.tosToServer
1773 Allows you to select a TOS/Diffserv value for packets outgoing
1774 on the server side, based on an ACL.
1776 tcp_outgoing_tos ds-field [!]aclname ...
1778 Example where normal_service_net uses the TOS value 0x00
1779 and good_service_net uses 0x20
1781 acl normal_service_net src 10.0.0.0/24
1782 acl good_service_net src 10.0.1.0/24
1783 tcp_outgoing_tos 0x00 normal_service_net
1784 tcp_outgoing_tos 0x20 good_service_net
1786 TOS/DSCP values really only have local significance - so you should
1787 know what you're specifying. For more information, see RFC2474,
1788 RFC2475, and RFC3260.
1790 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1791 "default" to use whatever default your host has. Note that in
1792 practice often only multiples of 4 is usable as the two rightmost bits
1793 have been redefined for use by ECN (RFC 3168 section 23.1).
1795 Processing proceeds in the order specified, and stops at first fully
1799 NAME: clientside_tos
1802 LOC: Ip::Qos::TheConfig.tosToClient
1804 Allows you to select a TOS/Diffserv value for packets being transmitted
1805 on the client-side, based on an ACL.
1807 clientside_tos ds-field [!]aclname ...
1809 Example where normal_service_net uses the TOS value 0x00
1810 and good_service_net uses 0x20
1812 acl normal_service_net src 10.0.0.0/24
1813 acl good_service_net src 10.0.1.0/24
1814 clientside_tos 0x00 normal_service_net
1815 clientside_tos 0x20 good_service_net
1817 Note: This feature is incompatible with qos_flows. Any TOS values set here
1818 will be overwritten by TOS values in qos_flows.
1821 NAME: tcp_outgoing_mark
1823 IFDEF: SO_MARK&&USE_LIBCAP
1825 LOC: Ip::Qos::TheConfig.nfmarkToServer
1827 Allows you to apply a Netfilter mark value to outgoing packets
1828 on the server side, based on an ACL.
1830 tcp_outgoing_mark mark-value [!]aclname ...
1832 Example where normal_service_net uses the mark value 0x00
1833 and good_service_net uses 0x20
1835 acl normal_service_net src 10.0.0.0/24
1836 acl good_service_net src 10.0.1.0/24
1837 tcp_outgoing_mark 0x00 normal_service_net
1838 tcp_outgoing_mark 0x20 good_service_net
1841 NAME: clientside_mark
1843 IFDEF: SO_MARK&&USE_LIBCAP
1845 LOC: Ip::Qos::TheConfig.nfmarkToClient
1847 Allows you to apply a Netfilter mark value to packets being transmitted
1848 on the client-side, based on an ACL.
1850 clientside_mark mark-value [!]aclname ...
1852 Example where normal_service_net uses the mark value 0x00
1853 and good_service_net uses 0x20
1855 acl normal_service_net src 10.0.0.0/24
1856 acl good_service_net src 10.0.1.0/24
1857 clientside_mark 0x00 normal_service_net
1858 clientside_mark 0x20 good_service_net
1860 Note: This feature is incompatible with qos_flows. Any mark values set here
1861 will be overwritten by mark values in qos_flows.
1868 LOC: Ip::Qos::TheConfig
1870 Allows you to select a TOS/DSCP value to mark outgoing
1871 connections with, based on where the reply was sourced. For
1872 platforms using netfilter, allows you to set a netfilter mark
1873 value instead of, or in addition to, a TOS value.
1875 TOS values really only have local significance - so you should
1876 know what you're specifying. For more information, see RFC2474,
1877 RFC2475, and RFC3260.
1879 The TOS/DSCP byte must be exactly that - a octet value 0 - 255. Note that
1880 in practice often only multiples of 4 is usable as the two rightmost bits
1881 have been redefined for use by ECN (RFC 3168 section 23.1).
1883 Mark values can be any unsigned 32-bit integer value.
1885 This setting is configured by setting the following values:
1887 tos|mark Whether to set TOS or netfilter mark values
1889 local-hit=0xFF Value to mark local cache hits.
1891 sibling-hit=0xFF Value to mark hits from sibling peers.
1893 parent-hit=0xFF Value to mark hits from parent peers.
1895 miss=0xFF[/mask] Value to mark cache misses. Takes precedence
1896 over the preserve-miss feature (see below), unless
1897 mask is specified, in which case only the bits
1898 specified in the mask are written.
1900 The TOS variant of the following features are only possible on Linux
1901 and require your kernel to be patched with the TOS preserving ZPH
1902 patch, available from http://zph.bratcheda.org
1903 No patch is needed to preserve the netfilter mark, which will work
1904 with all variants of netfilter.
1906 disable-preserve-miss
1907 This option disables the preservation of the TOS or netfilter
1908 mark. By default, the existing TOS or netfilter mark value of
1909 the response coming from the remote server will be retained
1910 and masked with miss-mark.
1911 NOTE: in the case of a netfilter mark, the mark must be set on
1912 the connection (using the CONNMARK target) not on the packet
1916 Allows you to mask certain bits in the TOS or mark value
1917 received from the remote server, before copying the value to
1918 the TOS sent towards clients.
1919 Default for tos: 0xFF (TOS from server is not changed).
1920 Default for mark: 0xFFFFFFFF (mark from server is not changed).
1922 All of these features require the --enable-zph-qos compilation flag
1923 (enabled by default). Netfilter marking also requires the
1924 libnetfilter_conntrack libraries (--with-netfilter-conntrack) and
1925 libcap 2.09+ (--with-libcap).
1929 NAME: tcp_outgoing_address
1932 LOC: Config.accessList.outgoing_address
1934 Allows you to map requests to different outgoing IP addresses
1935 based on the username or source address of the user making
1938 tcp_outgoing_address ipaddr [[!]aclname] ...
1941 Forwarding clients with dedicated IPs for certain subnets.
1943 acl normal_service_net src 10.0.0.0/24
1944 acl good_service_net src 10.0.2.0/24
1946 tcp_outgoing_address 2001:db8::c001 good_service_net
1947 tcp_outgoing_address 10.1.0.2 good_service_net
1949 tcp_outgoing_address 2001:db8::beef normal_service_net
1950 tcp_outgoing_address 10.1.0.1 normal_service_net
1952 tcp_outgoing_address 2001:db8::1
1953 tcp_outgoing_address 10.1.0.3
1955 Processing proceeds in the order specified, and stops at first fully
1958 Squid will add an implicit IP version test to each line.
1959 Requests going to IPv4 websites will use the outgoing 10.1.0.* addresses.
1960 Requests going to IPv6 websites will use the outgoing 2001:db8:* addresses.
1963 NOTE: The use of this directive using client dependent ACLs is
1964 incompatible with the use of server side persistent connections. To
1965 ensure correct results it is best to set server_persistent_connections
1966 to off when using this directive in such configurations.
1968 NOTE: The use of this directive to set a local IP on outgoing TCP links
1969 is incompatible with using TPROXY to set client IP out outbound TCP links.
1970 When needing to contact peers use the no-tproxy cache_peer option and the
1971 client_dst_passthru directive re-enable normal forwarding such as this.
1975 NAME: host_verify_strict
1978 LOC: Config.onoff.hostStrictVerify
1980 Regardless of this option setting, when dealing with intercepted
1981 traffic, Squid always verifies that the destination IP address matches
1982 the Host header domain or IP (called 'authority form URL').
1984 This enforcement is performed to satisfy a MUST-level requirement in
1985 RFC 2616 section 14.23: "The Host field value MUST represent the naming
1986 authority of the origin server or gateway given by the original URL".
1989 Squid always responds with an HTTP 409 (Conflict) error
1990 page and logs a security warning if there is no match.
1992 Squid verifies that the destination IP address matches
1993 the Host header for forward-proxy and reverse-proxy traffic
1994 as well. For those traffic types, Squid also enables the
1995 following checks, comparing the corresponding Host header
1996 and Request-URI components:
1998 * The host names (domain or IP) must be identical,
1999 but valueless or missing Host header disables all checks.
2000 For the two host names to match, both must be either IP
2003 * Port numbers must be identical, but if a port is missing
2004 the scheme-default port is assumed.
2007 When set to OFF (the default):
2008 Squid allows suspicious requests to continue but logs a
2009 security warning and blocks caching of the response.
2011 * Forward-proxy traffic is not checked at all.
2013 * Reverse-proxy traffic is not checked at all.
2015 * Intercepted traffic which passes verification is handled
2016 according to client_dst_passthru.
2018 * Intercepted requests which fail verification are sent
2019 to the client original destination instead of DIRECT.
2020 This overrides 'client_dst_passthru off'.
2022 For now suspicious intercepted CONNECT requests are always
2023 responded to with an HTTP 409 (Conflict) error page.
2028 As described in CVE-2009-0801 when the Host: header alone is used
2029 to determine the destination of a request it becomes trivial for
2030 malicious scripts on remote websites to bypass browser same-origin
2031 security policy and sandboxing protections.
2033 The cause of this is that such applets are allowed to perform their
2034 own HTTP stack, in which case the same-origin policy of the browser
2035 sandbox only verifies that the applet tries to contact the same IP
2036 as from where it was loaded at the IP level. The Host: header may
2037 be different from the connected IP and approved origin.
2041 NAME: client_dst_passthru
2044 LOC: Config.onoff.client_dst_passthru
2046 With NAT or TPROXY intercepted traffic Squid may pass the request
2047 directly to the original client destination IP or seek a faster
2048 source using the HTTP Host header.
2050 Using Host to locate alternative servers can provide faster
2051 connectivity with a range of failure recovery options.
2052 But can also lead to connectivity trouble when the client and
2053 server are attempting stateful interactions unaware of the proxy.
2055 This option (on by default) prevents alternative DNS entries being
2056 located to send intercepted traffic DIRECT to an origin server.
2057 The clients original destination IP and port will be used instead.
2059 Regardless of this option setting, when dealing with intercepted
2060 traffic Squid will verify the Host: header and any traffic which
2061 fails Host verification will be treated as if this option were ON.
2063 see host_verify_strict for details on the verification process.
2068 -----------------------------------------------------------------------------
2071 NAME: ssl_unclean_shutdown
2075 LOC: Config.SSL.unclean_shutdown
2077 Some browsers (especially MSIE) bugs out on SSL shutdown
2084 LOC: Config.SSL.ssl_engine
2087 The OpenSSL engine to use. You will need to set this if you
2088 would like to use hardware SSL acceleration for example.
2091 NAME: sslproxy_client_certificate
2094 LOC: Config.ssl_client.cert
2097 Client SSL Certificate to use when proxying https:// URLs
2100 NAME: sslproxy_client_key
2103 LOC: Config.ssl_client.key
2106 Client SSL Key to use when proxying https:// URLs
2109 NAME: sslproxy_version
2112 LOC: Config.ssl_client.version
2115 SSL version level to use when proxying https:// URLs
2117 The versions of SSL/TLS supported:
2119 1 automatic (default)
2127 NAME: sslproxy_options
2130 LOC: Config.ssl_client.options
2133 SSL implementation options to use when proxying https:// URLs
2135 The most important being:
2137 NO_SSLv2 Disallow the use of SSLv2
2138 NO_SSLv3 Disallow the use of SSLv3
2139 NO_TLSv1 Disallow the use of TLSv1.0
2140 NO_TLSv1_1 Disallow the use of TLSv1.1
2141 NO_TLSv1_2 Disallow the use of TLSv1.2
2143 Always create a new key when using temporary/ephemeral
2146 Disable use of RFC5077 session tickets. Some servers
2147 may have problems understanding the TLS extension due
2148 to ambiguous specification in RFC4507.
2149 ALL Enable various bug workarounds suggested as "harmless"
2150 by OpenSSL. Be warned that this may reduce SSL/TLS
2151 strength to some attacks.
2153 See the OpenSSL SSL_CTX_set_options documentation for a
2154 complete list of possible options.
2157 NAME: sslproxy_cipher
2160 LOC: Config.ssl_client.cipher
2163 SSL cipher list to use when proxying https:// URLs
2165 Colon separated list of supported ciphers.
2168 NAME: sslproxy_cafile
2171 LOC: Config.ssl_client.cafile
2174 file containing CA certificates to use when verifying server
2175 certificates while proxying https:// URLs
2178 NAME: sslproxy_capath
2181 LOC: Config.ssl_client.capath
2184 directory containing CA certificates to use when verifying
2185 server certificates while proxying https:// URLs
2190 TYPE: sslproxy_ssl_bump
2191 LOC: Config.accessList.ssl_bump
2194 This option is consulted when a CONNECT request is received on
2195 an http_port (or a new connection is intercepted at an
2196 https_port), provided that port was configured with an ssl-bump
2197 flag. The subsequent data on the connection is either treated as
2198 HTTPS and decrypted OR tunneled at TCP level without decryption,
2199 depending on the first bumping "mode" which ACLs match.
2201 ssl_bump <mode> [!]acl ...
2203 The following bumping modes are supported:
2206 Allow bumping of the connection. Establish a secure connection
2207 with the client first, then connect to the server. This old mode
2208 does not allow Squid to mimic server SSL certificate and does
2209 not work with intercepted SSL connections.
2212 Allow bumping of the connection. Establish a secure connection
2213 with the server first, then establish a secure connection with
2214 the client, using a mimicked server certificate. Works with both
2215 CONNECT requests and intercepted SSL connections.
2218 Become a TCP tunnel without decoding the connection.
2219 Works with both CONNECT requests and intercepted SSL
2220 connections. This is the default behavior when no
2221 ssl_bump option is given or no ssl_bump ACLs match.
2223 By default, no connections are bumped.
2225 The first matching ssl_bump option wins. If no ACLs match, the
2226 connection is not bumped. Unlike most allow/deny ACL lists, ssl_bump
2227 does not have an implicit "negate the last given option" rule. You
2228 must make that rule explicit if you convert old ssl_bump allow/deny
2229 rules that rely on such an implicit rule.
2231 This clause supports both fast and slow acl types.
2232 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2234 See also: http_port ssl-bump, https_port ssl-bump
2237 # Example: Bump all requests except those originating from
2238 # localhost and those going to example.com.
2240 acl broken_sites dstdomain .example.com
2241 ssl_bump none localhost
2242 ssl_bump none broken_sites
2243 ssl_bump server-first all
2246 NAME: sslproxy_flags
2249 LOC: Config.ssl_client.flags
2252 Various flags modifying the use of SSL while proxying https:// URLs:
2253 DONT_VERIFY_PEER Accept certificates that fail verification.
2254 For refined control, see sslproxy_cert_error.
2255 NO_DEFAULT_CA Don't use the default CA list built in
2259 NAME: sslproxy_cert_error
2262 LOC: Config.ssl_client.cert_error
2265 Use this ACL to bypass server certificate validation errors.
2267 For example, the following lines will bypass all validation errors
2268 when talking to servers for example.com. All other
2269 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
2271 acl BrokenButTrustedServers dstdomain example.com
2272 sslproxy_cert_error allow BrokenButTrustedServers
2273 sslproxy_cert_error deny all
2275 This clause only supports fast acl types.
2276 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2277 Using slow acl types may result in server crashes
2279 Without this option, all server certificate validation errors
2280 terminate the transaction. Bypassing validation errors is dangerous
2281 because an error usually implies that the server cannot be trusted and
2282 the connection may be insecure.
2284 See also: sslproxy_flags and DONT_VERIFY_PEER.
2286 Default setting: sslproxy_cert_error deny all
2289 NAME: sslproxy_cert_sign
2292 POSTSCRIPTUM: signUntrusted ssl::certUntrusted
2293 POSTSCRIPTUM: signSelf ssl::certSelfSigned
2294 POSTSCRIPTUM: signTrusted all
2295 TYPE: sslproxy_cert_sign
2296 LOC: Config.ssl_client.cert_sign
2299 sslproxy_cert_sign <signing algorithm> acl ...
2301 The following certificate signing algorithms are supported:
2303 Sign using the configured CA certificate which is usually
2304 placed in and trusted by end-user browsers. This is the
2305 default for trusted origin server certificates.
2307 Sign to guarantee an X509_V_ERR_CERT_UNTRUSTED browser error.
2308 This is the default for untrusted origin server certificates
2309 that are not self-signed (see ssl::certUntrusted).
2311 Sign using a self-signed certificate with the right CN to
2312 generate a X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT error in the
2313 browser. This is the default for self-signed origin server
2314 certificates (see ssl::certSelfSigned).
2316 This clause only supports fast acl types.
2318 When sslproxy_cert_sign acl(s) match, Squid uses the corresponding
2319 signing algorithm to generate the certificate and ignores all
2320 subsequent sslproxy_cert_sign options (the first match wins). If no
2321 acl(s) match, the default signing algorithm is determined by errors
2322 detected when obtaining and validating the origin server certificate.
2324 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2325 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2326 CONNECT request that carries a domain name. In all other cases (CONNECT
2327 to an IP address or an intercepted SSL connection), Squid cannot detect
2328 the domain mismatch at certificate generation time when
2329 bump-server-first is used.
2332 NAME: sslproxy_cert_adapt
2335 TYPE: sslproxy_cert_adapt
2336 LOC: Config.ssl_client.cert_adapt
2339 sslproxy_cert_adapt <adaptation algorithm> acl ...
2341 The following certificate adaptation algorithms are supported:
2343 Sets the "Not After" property to the "Not After" property of
2344 the CA certificate used to sign generated certificates.
2346 Sets the "Not Before" property to the "Not Before" property of
2347 the CA certificate used to sign generated certificates.
2348 setCommonName or setCommonName{CN}
2349 Sets Subject.CN property to the host name specified as a
2350 CN parameter or, if no explicit CN parameter was specified,
2351 extracted from the CONNECT request. It is a misconfiguration
2352 to use setCommonName without an explicit parameter for
2353 intercepted or tproxied SSL connections.
2355 This clause only supports fast acl types.
2357 Squid first groups sslproxy_cert_adapt options by adaptation algorithm.
2358 Within a group, when sslproxy_cert_adapt acl(s) match, Squid uses the
2359 corresponding adaptation algorithm to generate the certificate and
2360 ignores all subsequent sslproxy_cert_adapt options in that algorithm's
2361 group (i.e., the first match wins within each algorithm group). If no
2362 acl(s) match, the default mimicking action takes place.
2364 WARNING: SQUID_X509_V_ERR_DOMAIN_MISMATCH and ssl:certDomainMismatch can
2365 be used with sslproxy_cert_adapt, but if and only if Squid is bumping a
2366 CONNECT request that carries a domain name. In all other cases (CONNECT
2367 to an IP address or an intercepted SSL connection), Squid cannot detect
2368 the domain mismatch at certificate generation time when
2369 bump-server-first is used.
2372 NAME: sslpassword_program
2375 LOC: Config.Program.ssl_password
2378 Specify a program used for entering SSL key passphrases
2379 when using encrypted SSL certificate keys. If not specified
2380 keys must either be unencrypted, or Squid started with the -N
2381 option to allow it to query interactively for the passphrase.
2383 The key file name is given as argument to the program allowing
2384 selection of the right password if you have multiple encrypted
2389 OPTIONS RELATING TO EXTERNAL SSL_CRTD
2390 -----------------------------------------------------------------------------
2393 NAME: sslcrtd_program
2396 DEFAULT: @DEFAULT_SSL_CRTD@ -s @DEFAULT_SSL_DB_DIR@ -M 4MB
2397 LOC: Ssl::TheConfig.ssl_crtd
2399 Specify the location and options of the executable for ssl_crtd process.
2400 @DEFAULT_SSL_CRTD@ program requires -s and -M parameters
2401 For more information use:
2402 @DEFAULT_SSL_CRTD@ -h
2405 NAME: sslcrtd_children
2406 TYPE: HelperChildConfig
2408 DEFAULT: 32 startup=5 idle=1
2409 LOC: Ssl::TheConfig.ssl_crtdChildren
2411 The maximum number of processes spawn to service ssl server.
2412 The maximum this may be safely set to is 32.
2414 The startup= and idle= options allow some measure of skew in your
2419 Sets the minimum number of processes to spawn when Squid
2420 starts or reconfigures. When set to zero the first request will
2421 cause spawning of the first child process to handle it.
2423 Starting too few children temporary slows Squid under load while it
2424 tries to spawn enough additional processes to cope with traffic.
2428 Sets a minimum of how many processes Squid is to try and keep available
2429 at all times. When traffic begins to rise above what the existing
2430 processes can handle this many more will be spawned up to the maximum
2431 configured. A minimum setting of 1 is required.
2433 You must have at least one ssl_crtd process.
2436 NAME: sslcrtvalidator_program
2440 LOC: Ssl::TheConfig.ssl_crt_validator
2442 Specify the location and options of the executable for ssl_crt_validator
2446 NAME: sslcrtvalidator_children
2447 TYPE: HelperChildConfig
2449 DEFAULT: 32 startup=5 idle=1 concurrency=1
2450 LOC: Ssl::TheConfig.ssl_crt_validator_Children
2452 The maximum number of processes spawn to service ssl server.
2453 The maximum this may be safely set to is 32.
2455 The startup= and idle= options allow some measure of skew in your
2460 Sets the minimum number of processes to spawn when Squid
2461 starts or reconfigures. When set to zero the first request will
2462 cause spawning of the first child process to handle it.
2464 Starting too few children temporary slows Squid under load while it
2465 tries to spawn enough additional processes to cope with traffic.
2469 Sets a minimum of how many processes Squid is to try and keep available
2470 at all times. When traffic begins to rise above what the existing
2471 processes can handle this many more will be spawned up to the maximum
2472 configured. A minimum setting of 1 is required.
2476 The number of requests each certificate validator helper can handle in
2477 parallel. Defaults to 0 which indicates the certficate validator
2478 is a old-style single threaded redirector.
2480 When this directive is set to a value >= 1 then the protocol
2481 used to communicate with the helper is modified to include
2482 a request ID in front of the request/response. The request
2483 ID from the request must be echoed back with the response
2486 You must have at least one ssl_crt_validator process.
2490 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
2491 -----------------------------------------------------------------------------
2499 To specify other caches in a hierarchy, use the format:
2501 cache_peer hostname type http-port icp-port [options]
2506 # hostname type port port options
2507 # -------------------- -------- ----- ----- -----------
2508 cache_peer parent.foo.net parent 3128 3130 default
2509 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
2510 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
2511 cache_peer example.com parent 80 0 default
2512 cache_peer cdn.example.com sibling 3128 0
2514 type: either 'parent', 'sibling', or 'multicast'.
2516 proxy-port: The port number where the peer accept HTTP requests.
2517 For other Squid proxies this is usually 3128
2518 For web servers this is usually 80
2520 icp-port: Used for querying neighbor caches about objects.
2521 Set to 0 if the peer does not support ICP or HTCP.
2522 See ICP and HTCP options below for additional details.
2525 ==== ICP OPTIONS ====
2527 You MUST also set icp_port and icp_access explicitly when using these options.
2528 The defaults will prevent peer traffic using ICP.
2531 no-query Disable ICP queries to this neighbor.
2534 Indicates the named peer is a member of a multicast group.
2535 ICP queries will not be sent directly to the peer, but ICP
2536 replies will be accepted from it.
2538 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
2539 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
2542 To only send ICP queries to this neighbor infrequently.
2543 This is used to keep the neighbor round trip time updated
2544 and is usually used in conjunction with weighted-round-robin.
2547 ==== HTCP OPTIONS ====
2549 You MUST also set htcp_port and htcp_access explicitly when using these options.
2550 The defaults will prevent peer traffic using HTCP.
2553 htcp Send HTCP, instead of ICP, queries to the neighbor.
2554 You probably also want to set the "icp-port" to 4827
2555 instead of 3130. This directive accepts a comma separated
2556 list of options described below.
2558 htcp=oldsquid Send HTCP to old Squid versions (2.5 or earlier).
2560 htcp=no-clr Send HTCP to the neighbor but without
2561 sending any CLR requests. This cannot be used with
2564 htcp=only-clr Send HTCP to the neighbor but ONLY CLR requests.
2565 This cannot be used with no-clr.
2568 Send HTCP to the neighbor including CLRs but only when
2569 they do not result from PURGE requests.
2572 Forward any HTCP CLR requests this proxy receives to the peer.
2575 ==== PEER SELECTION METHODS ====
2577 The default peer selection method is ICP, with the first responding peer
2578 being used as source. These options can be used for better load balancing.
2581 default This is a parent cache which can be used as a "last-resort"
2582 if a peer cannot be located by any of the peer-selection methods.
2583 If specified more than once, only the first is used.
2585 round-robin Load-Balance parents which should be used in a round-robin
2586 fashion in the absence of any ICP queries.
2587 weight=N can be used to add bias.
2589 weighted-round-robin
2590 Load-Balance parents which should be used in a round-robin
2591 fashion with the frequency of each parent being based on the
2592 round trip time. Closer parents are used more often.
2593 Usually used for background-ping parents.
2594 weight=N can be used to add bias.
2596 carp Load-Balance parents which should be used as a CARP array.
2597 The requests will be distributed among the parents based on the
2598 CARP load balancing hash function based on their weight.
2600 userhash Load-balance parents based on the client proxy_auth or ident username.
2602 sourcehash Load-balance parents based on the client source IP.
2605 To be used only for cache peers of type "multicast".
2606 ALL members of this multicast group have "sibling"
2607 relationship with it, not "parent". This is to a multicast
2608 group when the requested object would be fetched only from
2609 a "parent" cache, anyway. It's useful, e.g., when
2610 configuring a pool of redundant Squid proxies, being
2611 members of the same multicast group.
2614 ==== PEER SELECTION OPTIONS ====
2616 weight=N use to affect the selection of a peer during any weighted
2617 peer-selection mechanisms.
2618 The weight must be an integer; default is 1,
2619 larger weights are favored more.
2620 This option does not affect parent selection if a peering
2621 protocol is not in use.
2623 basetime=N Specify a base amount to be subtracted from round trip
2625 It is subtracted before division by weight in calculating
2626 which parent to fectch from. If the rtt is less than the
2627 base time the rtt is set to a minimal value.
2629 ttl=N Specify a TTL to use when sending multicast ICP queries
2631 Only useful when sending to a multicast group.
2632 Because we don't accept ICP replies from random
2633 hosts, you must configure other group members as
2634 peers with the 'multicast-responder' option.
2636 no-delay To prevent access to this neighbor from influencing the
2639 digest-url=URL Tell Squid to fetch the cache digest (if digests are
2640 enabled) for this host from the specified URL rather
2641 than the Squid default location.
2644 ==== CARP OPTIONS ====
2646 carp-key=key-specification
2647 use a different key than the full URL to hash against the peer.
2648 the key-specification is a comma-separated list of the keywords
2649 scheme, host, port, path, params
2650 Order is not important.
2652 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
2654 originserver Causes this parent to be contacted as an origin server.
2655 Meant to be used in accelerator setups when the peer
2659 Set the Host header of requests forwarded to this peer.
2660 Useful in accelerator setups where the server (peer)
2661 expects a certain domain name but clients may request
2662 others. ie example.com or www.example.com
2664 no-digest Disable request of cache digests.
2667 Disables requesting ICMP RTT database (NetDB).
2670 ==== AUTHENTICATION OPTIONS ====
2673 If this is a personal/workgroup proxy and your parent
2674 requires proxy authentication.
2676 Note: The string can include URL escapes (i.e. %20 for
2677 spaces). This also means % must be written as %%.
2680 Send login details received from client to this peer.
2681 Both Proxy- and WWW-Authorization headers are passed
2682 without alteration to the peer.
2683 Authentication is not required by Squid for this to work.
2685 Note: This will pass any form of authentication but
2686 only Basic auth will work through a proxy unless the
2687 connection-auth options are also used.
2689 login=PASS Send login details received from client to this peer.
2690 Authentication is not required by this option.
2692 If there are no client-provided authentication headers
2693 to pass on, but username and password are available
2694 from an external ACL user= and password= result tags
2695 they may be sent instead.
2697 Note: To combine this with proxy_auth both proxies must
2698 share the same user database as HTTP only allows for
2699 a single login (one for proxy, one for origin server).
2700 Also be warned this will expose your users proxy
2701 password to the peer. USE WITH CAUTION
2704 Send the username to the upstream cache, but with a
2705 fixed password. This is meant to be used when the peer
2706 is in another administrative domain, but it is still
2707 needed to identify each user.
2708 The star can optionally be followed by some extra
2709 information which is added to the username. This can
2710 be used to identify this proxy to the peer, similar to
2711 the login=username:password option above.
2714 If this is a personal/workgroup proxy and your parent
2715 requires a secure proxy authentication.
2716 The first principal from the default keytab or defined by
2717 the environment variable KRB5_KTNAME will be used.
2719 WARNING: The connection may transmit requests from multiple
2720 clients. Negotiate often assumes end-to-end authentication
2721 and a single-client. Which is not strictly true here.
2723 login=NEGOTIATE:principal_name
2724 If this is a personal/workgroup proxy and your parent
2725 requires a secure proxy authentication.
2726 The principal principal_name from the default keytab or
2727 defined by the environment variable KRB5_KTNAME will be
2730 WARNING: The connection may transmit requests from multiple
2731 clients. Negotiate often assumes end-to-end authentication
2732 and a single-client. Which is not strictly true here.
2734 connection-auth=on|off
2735 Tell Squid that this peer does or not support Microsoft
2736 connection oriented authentication, and any such
2737 challenges received from there should be ignored.
2738 Default is auto to automatically determine the status
2742 ==== SSL / HTTPS / TLS OPTIONS ====
2744 ssl Encrypt connections to this peer with SSL/TLS.
2746 sslcert=/path/to/ssl/certificate
2747 A client SSL certificate to use when connecting to
2750 sslkey=/path/to/ssl/key
2751 The private SSL key corresponding to sslcert above.
2752 If 'sslkey' is not specified 'sslcert' is assumed to
2753 reference a combined file containing both the
2754 certificate and the key.
2756 sslversion=1|2|3|4|5|6
2757 The SSL version to use when connecting to this peer
2758 1 = automatic (default)
2765 sslcipher=... The list of valid SSL ciphers to use when connecting
2768 ssloptions=... Specify various SSL implementation options:
2770 NO_SSLv2 Disallow the use of SSLv2
2771 NO_SSLv3 Disallow the use of SSLv3
2772 NO_TLSv1 Disallow the use of TLSv1.0
2773 NO_TLSv1_1 Disallow the use of TLSv1.1
2774 NO_TLSv1_2 Disallow the use of TLSv1.2
2776 Always create a new key when using
2777 temporary/ephemeral DH key exchanges
2778 ALL Enable various bug workarounds
2779 suggested as "harmless" by OpenSSL
2780 Be warned that this reduces SSL/TLS
2781 strength to some attacks.
2783 See the OpenSSL SSL_CTX_set_options documentation for a
2786 sslcafile=... A file containing additional CA certificates to use
2787 when verifying the peer certificate.
2789 sslcapath=... A directory containing additional CA certificates to
2790 use when verifying the peer certificate.
2792 sslcrlfile=... A certificate revocation list file to use when
2793 verifying the peer certificate.
2795 sslflags=... Specify various flags modifying the SSL implementation:
2798 Accept certificates even if they fail to
2801 Don't use the default CA list built in
2804 Don't verify the peer certificate
2805 matches the server name
2807 ssldomain= The peer name as advertised in it's certificate.
2808 Used for verifying the correctness of the received peer
2809 certificate. If not specified the peer hostname will be
2813 Enable the "Front-End-Https: On" header needed when
2814 using Squid as a SSL frontend in front of Microsoft OWA.
2815 See MS KB document Q307347 for details on this header.
2816 If set to auto the header will only be added if the
2817 request is forwarded as a https:// URL.
2820 ==== GENERAL OPTIONS ====
2823 A peer-specific connect timeout.
2824 Also see the peer_connect_timeout directive.
2826 connect-fail-limit=N
2827 How many times connecting to a peer must fail before
2828 it is marked as down. Default is 10.
2830 allow-miss Disable Squid's use of only-if-cached when forwarding
2831 requests to siblings. This is primarily useful when
2832 icp_hit_stale is used by the sibling. To extensive use
2833 of this option may result in forwarding loops, and you
2834 should avoid having two-way peerings with this option.
2835 For example to deny peer usage on requests from peer
2836 by denying cache_peer_access if the source is a peer.
2838 max-conn=N Limit the amount of connections Squid may open to this
2841 name=xxx Unique name for the peer.
2842 Required if you have multiple peers on the same host
2843 but different ports.
2844 This name can be used in cache_peer_access and similar
2845 directives to dentify the peer.
2846 Can be used by outgoing access controls through the
2849 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2850 requests to this peer. Use normal address selection instead.
2852 proxy-only objects fetched from the peer will not be stored locally.
2856 NAME: cache_peer_domain cache_host_domain
2861 Use to limit the domains for which a neighbor cache will be
2864 cache_peer_domain cache-host domain [domain ...]
2865 cache_peer_domain cache-host !domain
2867 For example, specifying
2869 cache_peer_domain parent.foo.net .edu
2871 has the effect such that UDP query packets are sent to
2872 'bigserver' only when the requested object exists on a
2873 server in the .edu domain. Prefixing the domainname
2874 with '!' means the cache will be queried for objects
2877 NOTE: * Any number of domains may be given for a cache-host,
2878 either on the same or separate lines.
2879 * When multiple domains are given for a particular
2880 cache-host, the first matched domain is applied.
2881 * Cache hosts with no domain restrictions are queried
2883 * There are no defaults.
2884 * There is also a 'cache_peer_access' tag in the ACL
2888 NAME: cache_peer_access
2893 Similar to 'cache_peer_domain' but provides more flexibility by
2896 cache_peer_access cache-host allow|deny [!]aclname ...
2898 The syntax is identical to 'http_access' and the other lists of
2899 ACL elements. See the comments for 'http_access' below, or
2900 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2903 NAME: neighbor_type_domain
2904 TYPE: hostdomaintype
2908 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2910 Modifying the neighbor type for specific domains is now
2911 possible. You can treat some domains differently than the
2912 default neighbor type specified on the 'cache_peer' line.
2913 Normally it should only be necessary to list domains which
2914 should be treated differently because the default neighbor type
2915 applies for hostnames which do not match domains listed here.
2918 cache_peer cache.foo.org parent 3128 3130
2919 neighbor_type_domain cache.foo.org sibling .com .net
2920 neighbor_type_domain cache.foo.org sibling .au .de
2923 NAME: dead_peer_timeout
2927 LOC: Config.Timeout.deadPeer
2929 This controls how long Squid waits to declare a peer cache
2930 as "dead." If there are no ICP replies received in this
2931 amount of time, Squid will declare the peer dead and not
2932 expect to receive any further ICP replies. However, it
2933 continues to send ICP queries, and will mark the peer as
2934 alive upon receipt of the first subsequent ICP reply.
2936 This timeout also affects when Squid expects to receive ICP
2937 replies from peers. If more than 'dead_peer' seconds have
2938 passed since the last ICP reply was received, Squid will not
2939 expect to receive an ICP reply on the next query. Thus, if
2940 your time between requests is greater than this timeout, you
2941 will see a lot of requests sent DIRECT to origin servers
2942 instead of to your parents.
2945 NAME: forward_max_tries
2948 LOC: Config.forward_max_tries
2950 Controls how many different forward paths Squid will try
2951 before giving up. See also forward_timeout.
2953 NOTE: connect_retries (default: none) can make each of these
2954 possible forwarding paths be tried multiple times.
2957 NAME: hierarchy_stoplist
2960 LOC: Config.hierarchy_stoplist
2962 A list of words which, if found in a URL, cause the object to
2963 be handled directly by this cache. In other words, use this
2964 to not query neighbor caches for certain objects. You may
2965 list this option multiple times.
2968 hierarchy_stoplist cgi-bin ?
2970 Note: never_direct overrides this option.
2974 MEMORY CACHE OPTIONS
2975 -----------------------------------------------------------------------------
2982 LOC: Config.memMaxSize
2984 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2985 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2986 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2987 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2989 'cache_mem' specifies the ideal amount of memory to be used
2991 * In-Transit objects
2993 * Negative-Cached objects
2995 Data for these objects are stored in 4 KB blocks. This
2996 parameter specifies the ideal upper limit on the total size of
2997 4 KB blocks allocated. In-Transit objects take the highest
3000 In-transit objects have priority over the others. When
3001 additional space is needed for incoming data, negative-cached
3002 and hot objects will be released. In other words, the
3003 negative-cached and hot objects will fill up any unused space
3004 not needed for in-transit objects.
3006 If circumstances require, this limit will be exceeded.
3007 Specifically, if your incoming request rate requires more than
3008 'cache_mem' of memory to hold in-transit objects, Squid will
3009 exceed this limit to satisfy the new requests. When the load
3010 decreases, blocks will be freed until the high-water mark is
3011 reached. Thereafter, blocks will be used to store hot
3014 If shared memory caching is enabled, Squid does not use the shared
3015 cache space for in-transit objects, but they still consume as much
3016 local memory as they need. For more details about the shared memory
3017 cache, see memory_cache_shared.
3020 NAME: maximum_object_size_in_memory
3024 LOC: Config.Store.maxInMemObjSize
3026 Objects greater than this size will not be attempted to kept in
3027 the memory cache. This should be set high enough to keep objects
3028 accessed frequently in memory to improve performance whilst low
3029 enough to keep larger objects from hoarding cache_mem.
3032 NAME: memory_cache_shared
3035 LOC: Config.memShared
3037 DEFAULT_DOC: "on" where supported if doing memory caching with multiple SMP workers.
3039 Controls whether the memory cache is shared among SMP workers.
3041 The shared memory cache is meant to occupy cache_mem bytes and replace
3042 the non-shared memory cache, although some entities may still be
3043 cached locally by workers for now (e.g., internal and in-transit
3044 objects may be served from a local memory cache even if shared memory
3045 caching is enabled).
3047 By default, the memory cache is shared if and only if all of the
3048 following conditions are satisfied: Squid runs in SMP mode with
3049 multiple workers, cache_mem is positive, and Squid environment
3050 supports required IPC primitives (e.g., POSIX shared memory segments
3051 and GCC-style atomic operations).
3053 To avoid blocking locks, shared memory uses opportunistic algorithms
3054 that do not guarantee that every cachable entity that could have been
3055 shared among SMP workers will actually be shared.
3057 Currently, entities exceeding 32KB in size cannot be shared.
3060 NAME: memory_cache_mode
3065 Controls which objects to keep in the memory cache (cache_mem)
3067 always Keep most recently fetched objects in memory (default)
3069 disk Only disk cache hits are kept in memory, which means
3070 an object must first be cached on disk and then hit
3071 a second time before cached in memory.
3073 network Only objects fetched from network is kept in memory
3076 NAME: memory_replacement_policy
3078 LOC: Config.memPolicy
3081 The memory replacement policy parameter determines which
3082 objects are purged from memory when memory space is needed.
3084 See cache_replacement_policy for details.
3089 -----------------------------------------------------------------------------
3092 NAME: cache_replacement_policy
3094 LOC: Config.replPolicy
3097 The cache replacement policy parameter determines which
3098 objects are evicted (replaced) when disk space is needed.
3100 lru : Squid's original list based LRU policy
3101 heap GDSF : Greedy-Dual Size Frequency
3102 heap LFUDA: Least Frequently Used with Dynamic Aging
3103 heap LRU : LRU policy implemented using a heap
3105 Applies to any cache_dir lines listed below this.
3107 The LRU policies keeps recently referenced objects.
3109 The heap GDSF policy optimizes object hit rate by keeping smaller
3110 popular objects in cache so it has a better chance of getting a
3111 hit. It achieves a lower byte hit rate than LFUDA though since
3112 it evicts larger (possibly popular) objects.
3114 The heap LFUDA policy keeps popular objects in cache regardless of
3115 their size and thus optimizes byte hit rate at the expense of
3116 hit rate since one large, popular object will prevent many
3117 smaller, slightly less popular objects from being cached.
3119 Both policies utilize a dynamic aging mechanism that prevents
3120 cache pollution that can otherwise occur with frequency-based
3121 replacement policies.
3123 NOTE: if using the LFUDA replacement policy you should increase
3124 the value of maximum_object_size above its default of 4096 KB to
3125 to maximize the potential byte hit rate improvement of LFUDA.
3127 For more information about the GDSF and LFUDA cache replacement
3128 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
3129 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
3135 LOC: Config.cacheSwap
3139 cache_dir Type Directory-Name Fs-specific-data [options]
3141 You can specify multiple cache_dir lines to spread the
3142 cache among different disk partitions.
3144 Type specifies the kind of storage system to use. Only "ufs"
3145 is built by default. To enable any of the other storage systems
3146 see the --enable-storeio configure option.
3148 'Directory' is a top-level directory where cache swap
3149 files will be stored. If you want to use an entire disk
3150 for caching, this can be the mount-point directory.
3151 The directory must exist and be writable by the Squid
3152 process. Squid will NOT create this directory for you.
3154 In SMP configurations, cache_dir must not precede the workers option
3155 and should use configuration macros or conditionals to give each
3156 worker interested in disk caching a dedicated cache directory.
3160 "ufs" is the old well-known Squid storage format that has always
3163 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
3165 'Mbytes' is the amount of disk space (MB) to use under this
3166 directory. The default is 100 MB. Change this to suit your
3167 configuration. Do NOT put the size of your disk drive here.
3168 Instead, if you want Squid to use the entire disk drive,
3169 subtract 20% and use that value.
3171 'L1' is the number of first-level subdirectories which
3172 will be created under the 'Directory'. The default is 16.
3174 'L2' is the number of second-level subdirectories which
3175 will be created under each first-level directory. The default
3178 The aufs store type:
3180 "aufs" uses the same storage format as "ufs", utilizing
3181 POSIX-threads to avoid blocking the main Squid process on
3182 disk-I/O. This was formerly known in Squid as async-io.
3184 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
3186 see argument descriptions under ufs above
3188 The diskd store type:
3190 "diskd" uses the same storage format as "ufs", utilizing a
3191 separate process to avoid blocking the main Squid process on
3194 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
3196 see argument descriptions under ufs above
3198 Q1 specifies the number of unacknowledged I/O requests when Squid
3199 stops opening new files. If this many messages are in the queues,
3200 Squid won't open new files. Default is 64
3202 Q2 specifies the number of unacknowledged messages when Squid
3203 starts blocking. If this many messages are in the queues,
3204 Squid blocks until it receives some replies. Default is 72
3206 When Q1 < Q2 (the default), the cache directory is optimized
3207 for lower response time at the expense of a decrease in hit
3208 ratio. If Q1 > Q2, the cache directory is optimized for
3209 higher hit ratio at the expense of an increase in response
3212 The rock store type:
3214 cache_dir rock Directory-Name Mbytes <max-size=bytes> [options]
3216 The Rock Store type is a database-style storage. All cached
3217 entries are stored in a "database" file, using fixed-size slots,
3218 one entry per slot. The database size is specified in MB. The
3219 slot size is specified in bytes using the max-size option. See
3220 below for more info on the max-size option.
3222 swap-timeout=msec: Squid will not start writing a miss to or
3223 reading a hit from disk if it estimates that the swap operation
3224 will take more than the specified number of milliseconds. By
3225 default and when set to zero, disables the disk I/O time limit
3226 enforcement. Ignored when using blocking I/O module because
3227 blocking synchronous I/O does not allow Squid to estimate the
3228 expected swap wait time.
3230 max-swap-rate=swaps/sec: Artificially limits disk access using
3231 the specified I/O rate limit. Swap out requests that
3232 would cause the average I/O rate to exceed the limit are
3233 delayed. Individual swap in requests (i.e., hits or reads) are
3234 not delayed, but they do contribute to measured swap rate and
3235 since they are placed in the same FIFO queue as swap out
3236 requests, they may wait longer if max-swap-rate is smaller.
3237 This is necessary on file systems that buffer "too
3238 many" writes and then start blocking Squid and other processes
3239 while committing those writes to disk. Usually used together
3240 with swap-timeout to avoid excessive delays and queue overflows
3241 when disk demand exceeds available disk "bandwidth". By default
3242 and when set to zero, disables the disk I/O rate limit
3243 enforcement. Currently supported by IpcIo module only.
3246 The coss store type:
3248 NP: COSS filesystem in Squid-3 has been deemed too unstable for
3249 production use and has thus been removed from this release.
3250 We hope that it can be made usable again soon.
3252 block-size=n defines the "block size" for COSS cache_dir's.
3253 Squid uses file numbers as block numbers. Since file numbers
3254 are limited to 24 bits, the block size determines the maximum
3255 size of the COSS partition. The default is 512 bytes, which
3256 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
3257 you should not change the coss block size after Squid
3258 has written some objects to the cache_dir.
3260 The coss file store has changed from 2.5. Now it uses a file
3261 called 'stripe' in the directory names in the config - and
3262 this will be created by squid -z.
3266 no-store, no new objects should be stored to this cache_dir
3268 min-size=n, refers to the min object size in bytes this cache_dir
3269 will accept. It's used to restrict a cache_dir to only store
3270 large objects (e.g. aufs) while other storedirs are optimized
3271 for smaller objects (e.g. COSS). Defaults to 0.
3273 max-size=n, refers to the max object size in bytes this cache_dir
3274 supports. It is used to select the cache_dir to store the object.
3275 Note: To make optimal use of the max-size limits you should order
3276 the cache_dir lines with the smallest max-size value first and the
3277 ones with no max-size specification last.
3279 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
3280 which can be changed with the --with-coss-membuf-size=N configure
3284 # Uncomment and adjust the following to add a disk cache directory.
3285 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
3289 NAME: store_dir_select_algorithm
3291 LOC: Config.store_dir_select_algorithm
3294 Set this to 'round-robin' as an alternative.
3297 NAME: max_open_disk_fds
3299 LOC: Config.max_open_disk_fds
3302 To avoid having disk as the I/O bottleneck Squid can optionally
3303 bypass the on-disk cache if more than this amount of disk file
3304 descriptors are open.
3306 A value of 0 indicates no limit.
3309 NAME: minimum_object_size
3313 LOC: Config.Store.minObjectSize
3315 Objects smaller than this size will NOT be saved on disk. The
3316 value is specified in kilobytes, and the default is 0 KB, which
3317 means there is no minimum.
3320 NAME: maximum_object_size
3324 LOC: Config.Store.maxObjectSize
3326 Objects larger than this size will NOT be saved on disk. The
3327 value is specified in kilobytes, and the default is 4MB. If
3328 you wish to get a high BYTES hit ratio, you should probably
3329 increase this (one 32 MB object hit counts for 3200 10KB
3330 hits). If you wish to increase speed more than your want to
3331 save bandwidth you should leave this low.
3333 NOTE: if using the LFUDA replacement policy you should increase
3334 this value to maximize the byte hit rate improvement of LFUDA!
3335 See replacement_policy below for a discussion of this policy.
3338 NAME: cache_swap_low
3339 COMMENT: (percent, 0-100)
3342 LOC: Config.Swap.lowWaterMark
3345 NAME: cache_swap_high
3346 COMMENT: (percent, 0-100)
3349 LOC: Config.Swap.highWaterMark
3352 The low- and high-water marks for cache object replacement.
3353 Replacement begins when the swap (disk) usage is above the
3354 low-water mark and attempts to maintain utilization near the
3355 low-water mark. As swap utilization gets close to high-water
3356 mark object eviction becomes more aggressive. If utilization is
3357 close to the low-water mark less replacement is done each time.
3359 Defaults are 90% and 95%. If you have a large cache, 5% could be
3360 hundreds of MB. If this is the case you may wish to set these
3361 numbers closer together.
3366 -----------------------------------------------------------------------------
3376 logformat <name> <format specification>
3378 Defines an access log format.
3380 The <format specification> is a string with embedded % format codes
3382 % format codes all follow the same basic structure where all but
3383 the formatcode is optional. Output strings are automatically escaped
3384 as required according to their context and the output format
3385 modifiers are usually not needed, but can be specified if an explicit
3386 output format is desired.
3388 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
3390 " output in quoted string format
3391 [ output in squid text log format as used by log_mime_hdrs
3392 # output in URL quoted format
3397 width minimum and/or maximum field width:
3398 [width_min][.width_max]
3399 When minimum starts with 0, the field is zero-padded.
3400 String values exceeding maximum width are truncated.
3402 {arg} argument such as header name etc
3406 % a literal % character
3407 sn Unique sequence number per log line entry
3408 err_code The ID of an error response served by Squid or
3409 a similar internal error identifier.
3410 err_detail Additional err_code-dependent error information.
3411 note The meta header specified by the argument. Also
3412 logs the adaptation meta headers set by the
3413 adaptation_meta configuration parameter.
3414 If no argument given all meta headers logged.
3416 Connection related format codes:
3418 >a Client source IP address
3420 >p Client source port
3421 >eui Client source EUI (MAC address, EUI-48 or EUI-64 identifier)
3422 >la Local IP address the client connected to
3423 >lp Local port number the client connected to
3425 la Local listening IP address the client connection was connected to.
3426 lp Local listening port number the client connection was connected to.
3428 <a Server IP address of the last server or peer connection
3429 <A Server FQDN or peer name
3430 <p Server port number of the last server or peer connection
3431 <la Local IP address of the last server or peer connection
3432 <lp Local port number of the last server or peer connection
3434 Time related format codes:
3436 ts Seconds since epoch
3437 tu subsecond time (milliseconds)
3438 tl Local time. Optional strftime format argument
3439 default %d/%b/%Y:%H:%M:%S %z
3440 tg GMT time. Optional strftime format argument
3441 default %d/%b/%Y:%H:%M:%S %z
3442 tr Response time (milliseconds)
3443 dt Total time spent making DNS lookups (milliseconds)
3445 Access Control related format codes:
3447 et Tag returned by external acl
3448 ea Log string returned by external acl
3449 un User name (any available)
3450 ul User name from authentication
3451 ue User name from external acl helper
3452 ui User name from ident
3453 us User name from SSL
3455 HTTP related format codes:
3457 [http::]>h Original request header. Optional header name argument
3458 on the format header[:[separator]element]
3459 [http::]>ha The HTTP request headers after adaptation and redirection.
3460 Optional header name argument as for >h
3461 [http::]<h Reply header. Optional header name argument
3463 [http::]>Hs HTTP status code sent to the client
3464 [http::]<Hs HTTP status code received from the next hop
3465 [http::]<bs Number of HTTP-equivalent message body bytes
3466 received from the next hop, excluding chunked
3467 transfer encoding and control messages.
3468 Generated FTP/Gopher listings are treated as
3470 [http::]mt MIME content type
3471 [http::]rm Request method (GET/POST etc)
3472 [http::]>rm Request method from client
3473 [http::]<rm Request method sent to server or peer
3474 [http::]ru Request URL from client (historic, filtered for logging)
3475 [http::]>ru Request URL from client
3476 [http::]<ru Request URL sent to server or peer
3477 [http::]rp Request URL-Path excluding hostname
3478 [http::]>rp Request URL-Path excluding hostname from client
3479 [http::]<rp Request URL-Path excluding hostname sento to server or peer
3480 [http::]rv Request protocol version
3481 [http::]>rv Request protocol version from client
3482 [http::]<rv Request protocol version sent to server or peer
3483 [http::]<st Sent reply size including HTTP headers
3484 [http::]>st Received request size including HTTP headers. In the
3485 case of chunked requests the chunked encoding metadata
3487 [http::]>sh Received HTTP request headers size
3488 [http::]<sh Sent HTTP reply headers size
3489 [http::]st Request+Reply size including HTTP headers
3490 [http::]<sH Reply high offset sent
3491 [http::]<sS Upstream object size
3492 [http::]<pt Peer response time in milliseconds. The timer starts
3493 when the last request byte is sent to the next hop
3494 and stops when the last response byte is received.
3495 [http::]<tt Total server-side time in milliseconds. The timer
3496 starts with the first connect request (or write I/O)
3497 sent to the first selected peer. The timer stops
3498 with the last I/O with the last peer.
3500 Squid handling related format codes:
3502 Ss Squid request status (TCP_MISS etc)
3503 Sh Squid hierarchy status (DEFAULT_PARENT etc)
3505 SSL-related format codes:
3507 ssl::bump_mode SslBump decision for the transaction:
3509 For CONNECT requests that initiated bumping of
3510 a connection and for any request received on
3511 an already bumped connection, Squid logs the
3512 corresponding SslBump mode ("server-first" or
3513 "client-first"). See the ssl_bump option for
3514 more information about these modes.
3516 A "none" token is logged for requests that
3517 triggered "ssl_bump" ACL evaluation matching
3518 either a "none" rule or no rules at all.
3520 In all other cases, a single dash ("-") is
3523 If ICAP is enabled, the following code becomes available (as
3524 well as ICAP log codes documented with the icap_log option):
3526 icap::tt Total ICAP processing time for the HTTP
3527 transaction. The timer ticks when ICAP
3528 ACLs are checked and when ICAP
3529 transaction is in progress.
3531 If adaptation is enabled the following three codes become available:
3533 adapt::<last_h The header of the last ICAP response or
3534 meta-information from the last eCAP
3535 transaction related to the HTTP transaction.
3536 Like <h, accepts an optional header name
3539 adapt::sum_trs Summed adaptation transaction response
3540 times recorded as a comma-separated list in
3541 the order of transaction start time. Each time
3542 value is recorded as an integer number,
3543 representing response time of one or more
3544 adaptation (ICAP or eCAP) transaction in
3545 milliseconds. When a failed transaction is
3546 being retried or repeated, its time is not
3547 logged individually but added to the
3548 replacement (next) transaction. See also:
3551 adapt::all_trs All adaptation transaction response times.
3552 Same as adaptation_strs but response times of
3553 individual transactions are never added
3554 together. Instead, all transaction response
3555 times are recorded individually.
3557 You can prefix adapt::*_trs format codes with adaptation
3558 service name in curly braces to record response time(s) specific
3559 to that service. For example: %{my_service}adapt::sum_trs
3561 If SSL is enabled, the following formating codes become available:
3563 %ssl::>cert_subject The Subject field of the received client
3564 SSL certificate or a dash ('-') if Squid has
3565 received an invalid/malformed certificate or
3566 no certificate at all. Consider encoding the
3567 logged value because Subject often has spaces.
3569 %ssl::>cert_issuer The Issuer field of the received client
3570 SSL certificate or a dash ('-') if Squid has
3571 received an invalid/malformed certificate or
3572 no certificate at all. Consider encoding the
3573 logged value because Issuer often has spaces.
3575 The default formats available (which do not need re-defining) are:
3577 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %[un %Sh/%<a %mt
3578 logformat common %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
3579 logformat combined %>a %[ui %[un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
3580 logformat referrer %ts.%03tu %>a %{Referer}>h %ru
3581 logformat useragent %>a [%tl] "%{User-Agent}>h"
3583 NOTE: When the log_mime_hdrs directive is set to ON.
3584 The squid, common and combined formats have a safely encoded copy
3585 of the mime headers appended to each line within a pair of brackets.
3587 NOTE: The common and combined formats are not quite true to the Apache definition.
3588 The logs from Squid contain an extra status and hierarchy code appended.
3592 NAME: access_log cache_access_log
3594 LOC: Config.Log.accesslogs
3595 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
3597 These files log client request activities. Has a line every HTTP or
3598 ICP request. The format is:
3599 access_log <module>:<place> [<logformat name> [acl acl ...]]
3600 access_log none [acl acl ...]]
3602 Will log to the specified module:place using the specified format (which
3603 must be defined in a logformat directive) those entries which match
3604 ALL the acl's specified (which must be defined in acl clauses).
3605 If no acl is specified, all requests will be logged to this destination.
3607 ===== Modules Currently available =====
3609 none Do not log any requests matching these ACL.
3610 Do not specify Place or logformat name.
3612 stdio Write each log line to disk immediately at the completion of
3614 Place: the filename and path to be written.
3616 daemon Very similar to stdio. But instead of writing to disk the log
3617 line is passed to a daemon helper for asychronous handling instead.
3618 Place: varies depending on the daemon.
3620 log_file_daemon Place: the file name and path to be written.
3622 syslog To log each request via syslog facility.
3623 Place: The syslog facility and priority level for these entries.
3624 Place Format: facility.priority
3626 where facility could be any of:
3627 authpriv, daemon, local0 ... local7 or user.
3629 And priority could be any of:
3630 err, warning, notice, info, debug.
3632 udp To send each log line as text data to a UDP receiver.
3633 Place: The destination host name or IP and port.
3634 Place Format: //host:port
3636 tcp To send each log line as text data to a TCP receiver.
3637 Place: The destination host name or IP and port.
3638 Place Format: //host:port
3641 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
3647 LOC: Config.Log.icaplogs
3650 ICAP log files record ICAP transaction summaries, one line per
3653 The icap_log option format is:
3654 icap_log <filepath> [<logformat name> [acl acl ...]]
3655 icap_log none [acl acl ...]]
3657 Please see access_log option documentation for details. The two
3658 kinds of logs share the overall configuration approach and many
3661 ICAP processing of a single HTTP message or transaction may
3662 require multiple ICAP transactions. In such cases, multiple
3663 ICAP transaction log lines will correspond to a single access
3666 ICAP log uses logformat codes that make sense for an ICAP
3667 transaction. Header-related codes are applied to the HTTP header
3668 embedded in an ICAP server response, with the following caveats:
3669 For REQMOD, there is no HTTP response header unless the ICAP
3670 server performed request satisfaction. For RESPMOD, the HTTP
3671 request header is the header sent to the ICAP server. For
3672 OPTIONS, there are no HTTP headers.
3674 The following format codes are also available for ICAP logs:
3676 icap::<A ICAP server IP address. Similar to <A.
3678 icap::<service_name ICAP service name from the icap_service
3679 option in Squid configuration file.
3681 icap::ru ICAP Request-URI. Similar to ru.
3683 icap::rm ICAP request method (REQMOD, RESPMOD, or
3684 OPTIONS). Similar to existing rm.
3686 icap::>st Bytes sent to the ICAP server (TCP payload
3687 only; i.e., what Squid writes to the socket).
3689 icap::<st Bytes received from the ICAP server (TCP
3690 payload only; i.e., what Squid reads from
3693 icap::<bs Number of message body bytes received from the
3694 ICAP server. ICAP message body, if any, usually
3695 includes encapsulated HTTP message headers and
3696 possibly encapsulated HTTP message body. The
3697 HTTP body part is dechunked before its size is
3700 icap::tr Transaction response time (in
3701 milliseconds). The timer starts when
3702 the ICAP transaction is created and
3703 stops when the transaction is completed.
3706 icap::tio Transaction I/O time (in milliseconds). The
3707 timer starts when the first ICAP request
3708 byte is scheduled for sending. The timers
3709 stops when the last byte of the ICAP response
3712 icap::to Transaction outcome: ICAP_ERR* for all
3713 transaction errors, ICAP_OPT for OPTION
3714 transactions, ICAP_ECHO for 204
3715 responses, ICAP_MOD for message
3716 modification, and ICAP_SAT for request
3717 satisfaction. Similar to Ss.
3719 icap::Hs ICAP response status code. Similar to Hs.
3721 icap::>h ICAP request header(s). Similar to >h.
3723 icap::<h ICAP response header(s). Similar to <h.
3725 The default ICAP log format, which can be used without an explicit
3726 definition, is called icap_squid:
3728 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
3730 See also: logformat, log_icap, and %adapt::<last_h
3733 NAME: logfile_daemon
3735 DEFAULT: @DEFAULT_LOGFILED@
3736 LOC: Log::TheConfig.logfile_daemon
3738 Specify the path to the logfile-writing daemon. This daemon is
3739 used to write the access and store logs, if configured.
3741 Squid sends a number of commands to the log daemon:
3742 L<data>\n - logfile data
3747 r<n>\n - set rotate count to <n>
3748 b<n>\n - 1 = buffer output, 0 = don't buffer output
3750 No responses is expected.
3755 LOC: Config.accessList.log
3757 COMMENT: allow|deny acl acl...
3759 This options allows you to control which requests gets logged
3760 to access.log (see access_log directive). Requests denied for
3761 logging will also not be accounted for in performance counters.
3763 This clause only supports fast acl types.
3764 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3770 LOC: Config.accessList.icap
3773 This options allows you to control which requests get logged
3774 to icap.log. See the icap_log directive for ICAP log details.
3777 NAME: cache_store_log
3780 LOC: Config.Log.store
3782 Logs the activities of the storage manager. Shows which
3783 objects are ejected from the cache, and which objects are
3784 saved and for how long.
3785 There are not really utilities to analyze this data, so you can safely
3786 disable it (the default).
3788 Store log uses modular logging outputs. See access_log for the list
3789 of modules supported.
3792 cache_store_log stdio:@DEFAULT_STORE_LOG@
3793 cache_store_log daemon:@DEFAULT_STORE_LOG@
3796 NAME: cache_swap_state cache_swap_log
3798 LOC: Config.Log.swap
3801 Location for the cache "swap.state" file. This index file holds
3802 the metadata of objects saved on disk. It is used to rebuild
3803 the cache during startup. Normally this file resides in each
3804 'cache_dir' directory, but you may specify an alternate
3805 pathname here. Note you must give a full filename, not just
3806 a directory. Since this is the index for the whole object
3807 list you CANNOT periodically rotate it!
3809 If %s can be used in the file name it will be replaced with a
3810 a representation of the cache_dir name where each / is replaced
3811 with '.'. This is needed to allow adding/removing cache_dir
3812 lines when cache_swap_log is being used.
3814 If have more than one 'cache_dir', and %s is not used in the name
3815 these swap logs will have names such as:
3821 The numbered extension (which is added automatically)
3822 corresponds to the order of the 'cache_dir' lines in this
3823 configuration file. If you change the order of the 'cache_dir'
3824 lines in this file, these index files will NOT correspond to
3825 the correct 'cache_dir' entry (unless you manually rename
3826 them). We recommend you do NOT use this option. It is
3827 better to keep these index files in each 'cache_dir' directory.
3830 NAME: logfile_rotate
3833 LOC: Config.Log.rotateNumber
3835 Specifies the number of logfile rotations to make when you
3836 type 'squid -k rotate'. The default is 10, which will rotate
3837 with extensions 0 through 9. Setting logfile_rotate to 0 will
3838 disable the file name rotation, but the logfiles are still closed
3839 and re-opened. This will enable you to rename the logfiles
3840 yourself just before sending the rotate signal.
3842 Note, the 'squid -k rotate' command normally sends a USR1
3843 signal to the running squid process. In certain situations
3844 (e.g. on Linux with Async I/O), USR1 is used for other
3845 purposes, so -k rotate uses another signal. It is best to get
3846 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3849 Note, from Squid-3.1 this option has no effect on the cache.log,
3850 that log can be rotated separately by using debug_options
3853 NAME: emulate_httpd_log
3856 Replace this with an access_log directive using the format 'common' or 'combined'.
3859 NAME: log_ip_on_direct
3862 Remove this option from your config. To log server or peer names use %<A in the log format.
3867 DEFAULT: @DEFAULT_MIME_TABLE@
3868 LOC: Config.mimeTablePathname
3870 Pathname to Squid's MIME table. You shouldn't need to change
3871 this, but the default file contains examples and formatting
3872 information if you do.
3878 LOC: Config.onoff.log_mime_hdrs
3881 The Cache can record both the request and the response MIME
3882 headers for each HTTP transaction. The headers are encoded
3883 safely and will appear as two bracketed fields at the end of
3884 the access log (for either the native or httpd-emulated log
3885 formats). To enable this logging set log_mime_hdrs to 'on'.
3891 Replace this with an access_log directive using the format 'useragent'.
3894 NAME: referer_log referrer_log
3897 Replace this with an access_log directive using the format 'referrer'.
3902 DEFAULT: @DEFAULT_PID_FILE@
3903 LOC: Config.pidFilename
3905 A filename to write the process-id to. To disable, enter "none".
3911 Remove this option from your config. To log FQDN use %>A in the log format.
3914 NAME: client_netmask
3916 LOC: Config.Addrs.client_netmask
3919 A netmask for client addresses in logfiles and cachemgr output.
3920 Change this to protect the privacy of your cache clients.
3921 A netmask of 255.255.255.0 will log all IP's in that range with
3922 the last digit set to '0'.
3928 Use a regular access.log with ACL limiting it to MISS events.
3931 NAME: strip_query_terms
3933 LOC: Config.onoff.strip_query_terms
3936 By default, Squid strips query terms from requested URLs before
3937 logging. This protects your user's privacy.
3944 LOC: Config.onoff.buffered_logs
3946 cache.log log file is written with stdio functions, and as such
3947 it can be buffered or unbuffered. By default it will be unbuffered.
3948 Buffering it can speed up the writing slightly (though you are
3949 unlikely to need to worry unless you run with tons of debugging
3950 enabled in which case performance will suffer badly anyway..).
3953 NAME: netdb_filename
3955 DEFAULT: stdio:@DEFAULT_NETDB_FILE@
3956 LOC: Config.netdbFilename
3959 A filename where Squid stores it's netdb state between restarts.
3960 To disable, enter "none".
3964 OPTIONS FOR TROUBLESHOOTING
3965 -----------------------------------------------------------------------------
3970 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3971 LOC: Debug::cache_log
3973 Cache logging file. This is where general information about
3974 your cache's behavior goes. You can increase the amount of data
3975 logged to this file and how often its rotated with "debug_options"
3981 LOC: Debug::debugOptions
3983 Logging options are set as section,level where each source file
3984 is assigned a unique section. Lower levels result in less
3985 output, Full debugging (level 9) can result in a very large
3986 log file, so be careful.
3988 The magic word "ALL" sets debugging levels for all sections.
3989 We recommend normally running with "ALL,1".
3991 The rotate=N option can be used to keep more or less of these logs
3992 than would otherwise be kept by logfile_rotate.
3993 For most uses a single log should be enough to monitor current
3994 events affecting Squid.
3999 LOC: Config.coredump_dir
4000 DEFAULT_IF_NONE: none
4002 By default Squid leaves core files in the directory from where
4003 it was started. If you set 'coredump_dir' to a directory
4004 that exists, Squid will chdir() to that directory at startup
4005 and coredump files will be left there.
4009 # Leave coredumps in the first cache dir
4010 coredump_dir @DEFAULT_SWAP_DIR@
4016 OPTIONS FOR FTP GATEWAYING
4017 -----------------------------------------------------------------------------
4023 LOC: Config.Ftp.anon_user
4025 If you want the anonymous login password to be more informative
4026 (and enable the use of picky ftp servers), set this to something
4027 reasonable for your domain, like wwwuser@somewhere.net
4029 The reason why this is domainless by default is the
4030 request can be made on the behalf of a user in any domain,
4031 depending on how the cache is used.
4032 Some ftp server also validate the email address is valid
4033 (for example perl.com).
4039 LOC: Config.Ftp.passive
4041 If your firewall does not allow Squid to use passive
4042 connections, turn off this option.
4044 Use of ftp_epsv_all option requires this to be ON.
4050 LOC: Config.Ftp.epsv_all
4052 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
4054 NATs may be able to put the connection on a "fast path" through the
4055 translator, as the EPRT command will never be used and therefore,
4056 translation of the data portion of the segments will never be needed.
4058 When a client only expects to do two-way FTP transfers this may be
4060 If squid finds that it must do a three-way FTP transfer after issuing
4061 an EPSV ALL command, the FTP session will fail.
4063 If you have any doubts about this option do not use it.
4064 Squid will nicely attempt all other connection methods.
4066 Requires ftp_passive to be ON (default) for any effect.
4072 LOC: Config.Ftp.epsv
4074 FTP Protocol extensions permit the use of a special "EPSV" command.
4076 NATs may be able to put the connection on a "fast path" through the
4077 translator using EPSV, as the EPRT command will never be used
4078 and therefore, translation of the data portion of the segments
4079 will never be needed.
4081 Turning this OFF will prevent EPSV being attempted.
4082 WARNING: Doing so will convert Squid back to the old behavior with all
4083 the related problems with external NAT devices/layers.
4085 Requires ftp_passive to be ON (default) for any effect.
4091 LOC: Config.Ftp.eprt
4093 FTP Protocol extensions permit the use of a special "EPRT" command.
4095 This extension provides a protocol neutral alternative to the
4096 IPv4-only PORT command. When supported it enables active FTP data
4097 channels over IPv6 and efficient NAT handling.
4099 Turning this OFF will prevent EPRT being attempted and will skip
4100 straight to using PORT for IPv4 servers.
4102 Some devices are known to not handle this extension correctly and
4103 may result in crashes. Devices which suport EPRT enough to fail
4104 cleanly will result in Squid attempting PORT anyway. This directive
4105 should only be disabled when EPRT results in device failures.
4107 WARNING: Doing so will convert Squid back to the old behavior with all
4108 the related problems with external NAT devices/layers and IPv4-only FTP.
4111 NAME: ftp_sanitycheck
4114 LOC: Config.Ftp.sanitycheck
4116 For security and data integrity reasons Squid by default performs
4117 sanity checks of the addresses of FTP data connections ensure the
4118 data connection is to the requested server. If you need to allow
4119 FTP connections to servers using another IP address for the data
4120 connection turn this off.
4123 NAME: ftp_telnet_protocol
4126 LOC: Config.Ftp.telnet
4128 The FTP protocol is officially defined to use the telnet protocol
4129 as transport channel for the control connection. However, many
4130 implementations are broken and does not respect this aspect of
4133 If you have trouble accessing files with ASCII code 255 in the
4134 path or similar problems involving this ASCII code you can
4135 try setting this directive to off. If that helps, report to the
4136 operator of the FTP server in question that their FTP server
4137 is broken and does not follow the FTP standard.
4141 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
4142 -----------------------------------------------------------------------------
4147 DEFAULT: @DEFAULT_DISKD@
4148 LOC: Config.Program.diskd
4150 Specify the location of the diskd executable.
4151 Note this is only useful if you have compiled in
4152 diskd as one of the store io modules.
4155 NAME: unlinkd_program
4158 DEFAULT: @DEFAULT_UNLINKD@
4159 LOC: Config.Program.unlinkd
4161 Specify the location of the executable for file deletion process.
4164 NAME: pinger_program
4166 DEFAULT: @DEFAULT_PINGER@
4167 LOC: Config.pinger.program
4170 Specify the location of the executable for the pinger process.
4176 LOC: Config.pinger.enable
4179 Control whether the pinger is active at run-time.
4180 Enables turning ICMP pinger on and off with a simple
4181 squid -k reconfigure.
4186 OPTIONS FOR URL REWRITING
4187 -----------------------------------------------------------------------------
4190 NAME: url_rewrite_program redirect_program
4192 LOC: Config.Program.redirect
4195 Specify the location of the executable URL rewriter to use.
4196 Since they can perform almost any function there isn't one included.
4198 For each requested URL, the rewriter will receive on line with the format
4200 [channel-ID <SP>] URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kv-pairs]<NL>
4203 After processing the request the helper must reply using the following format:
4205 [channel-ID <SP>] result [<SP> kv-pairs]
4207 The result code can be:
4209 OK status=30N url="..."
4210 Redirect the URL to the one supplied in 'url='.
4211 'status=' is optional and contains the status code to send
4212 the client in Squids HTTP response. It must be one of the
4213 HTTP redirect status codes: 301, 302, 303, 307, 308.
4214 When no status is given Squid will use 302.
4216 OK rewrite-url="..."
4217 Rewrite the URL to the one supplied in 'rewrite-url='.
4218 The new URL is fetched directly by Squid and returned to
4219 the client as the response to its request.
4222 Do not change the URL.
4225 An internal error occured in the helper, preventing
4226 a result being identified.
4229 In the future, the interface protocol will be extended with
4230 key=value pairs ("kv-pairs" shown above). Helper programs
4231 should be prepared to receive and possibly ignore additional
4232 whitespace-separated tokens on each input line.
4234 When using the concurrency= option the protocol is changed by
4235 introducing a query channel tag in front of the request/response.
4236 The query channel tag is a number between 0 and concurrency-1.
4237 This value must be echoed back unchanged to Squid as the first part
4238 of the response relating to its request.
4240 WARNING: URL re-writing ability should be avoided whenever possible.
4241 Use the URL redirect form of response instead.
4243 Re-write creates a difference in the state held by the client
4244 and server. Possibly causing confusion when the server response
4245 contains snippets of its view state. Embeded URLs, response
4246 and content Location headers, etc. are not re-written by this
4249 By default, a URL rewriter is not used.
4252 NAME: url_rewrite_children redirect_children
4253 TYPE: HelperChildConfig
4254 DEFAULT: 20 startup=0 idle=1 concurrency=0
4255 LOC: Config.redirectChildren
4257 The maximum number of redirector processes to spawn. If you limit
4258 it too few Squid will have to wait for them to process a backlog of
4259 URLs, slowing it down. If you allow too many they will use RAM
4260 and other system resources noticably.
4262 The startup= and idle= options allow some measure of skew in your
4267 Sets a minimum of how many processes are to be spawned when Squid
4268 starts or reconfigures. When set to zero the first request will
4269 cause spawning of the first child process to handle it.
4271 Starting too few will cause an initial slowdown in traffic as Squid
4272 attempts to simultaneously spawn enough processes to cope.
4276 Sets a minimum of how many processes Squid is to try and keep available
4277 at all times. When traffic begins to rise above what the existing
4278 processes can handle this many more will be spawned up to the maximum
4279 configured. A minimum setting of 1 is required.
4283 The number of requests each redirector helper can handle in
4284 parallel. Defaults to 0 which indicates the redirector
4285 is a old-style single threaded redirector.
4287 When this directive is set to a value >= 1 then the protocol
4288 used to communicate with the helper is modified to include
4289 a request ID in front of the request/response. The request
4290 ID from the request must be echoed back with the response
4294 NAME: url_rewrite_host_header redirect_rewrites_host_header
4297 LOC: Config.onoff.redir_rewrites_host
4299 To preserve same-origin security policies in browsers and
4300 prevent Host: header forgery by redirectors Squid rewrites
4301 any Host: header in redirected requests.
4303 If you are running an accelerator this may not be a wanted
4304 effect of a redirector. This directive enables you disable
4305 Host: alteration in reverse-proxy traffic.
4307 WARNING: Entries are cached on the result of the URL rewriting
4308 process, so be careful if you have domain-virtual hosts.
4310 WARNING: Squid and other software verifies the URL and Host
4311 are matching, so be careful not to relay through other proxies
4312 or inspecting firewalls with this disabled.
4315 NAME: url_rewrite_access redirector_access
4318 LOC: Config.accessList.redirector
4320 If defined, this access list specifies which requests are
4321 sent to the redirector processes. By default all requests
4324 This clause supports both fast and slow acl types.
4325 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4328 NAME: url_rewrite_bypass redirector_bypass
4330 LOC: Config.onoff.redirector_bypass
4333 When this is 'on', a request will not go through the
4334 redirector if all redirectors are busy. If this is 'off'
4335 and the redirector queue grows too large, Squid will exit
4336 with a FATAL error and ask you to increase the number of
4337 redirectors. You should only enable this if the redirectors
4338 are not critical to your caching system. If you use
4339 redirectors for access control, and you enable this option,
4340 users may have access to pages they should not
4341 be allowed to request.
4345 OPTIONS FOR TUNING THE CACHE
4346 -----------------------------------------------------------------------------
4349 NAME: cache no_cache
4352 LOC: Config.accessList.noCache
4354 A list of ACL elements which, if matched and denied, cause the request to
4355 not be satisfied from the cache and the reply to not be cached.
4356 In other words, use this to force certain objects to never be cached.
4358 You must use the words 'allow' or 'deny' to indicate whether items
4359 matching the ACL should be allowed or denied into the cache.
4361 Default is to allow all to be cached.
4363 This clause supports both fast and slow acl types.
4364 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4370 LOC: Config.maxStale
4373 This option puts an upper limit on how stale content Squid
4374 will serve from the cache if cache validation fails.
4375 Can be overriden by the refresh_pattern max-stale option.
4378 NAME: refresh_pattern
4379 TYPE: refreshpattern
4383 usage: refresh_pattern [-i] regex min percent max [options]
4385 By default, regular expressions are CASE-SENSITIVE. To make
4386 them case-insensitive, use the -i option.
4388 'Min' is the time (in minutes) an object without an explicit
4389 expiry time should be considered fresh. The recommended
4390 value is 0, any higher values may cause dynamic applications
4391 to be erroneously cached unless the application designer
4392 has taken the appropriate actions.
4394 'Percent' is a percentage of the objects age (time since last
4395 modification age) an object without explicit expiry time
4396 will be considered fresh.
4398 'Max' is an upper limit on how long objects without an explicit
4399 expiry time will be considered fresh.
4401 options: override-expire
4406 ignore-must-revalidate
4413 override-expire enforces min age even if the server
4414 sent an explicit expiry time (e.g., with the
4415 Expires: header or Cache-Control: max-age). Doing this
4416 VIOLATES the HTTP standard. Enabling this feature
4417 could make you liable for problems which it causes.
4419 Note: override-expire does not enforce staleness - it only extends
4420 freshness / min. If the server returns a Expires time which
4421 is longer than your max time, Squid will still consider
4422 the object fresh for that period of time.
4424 override-lastmod enforces min age even on objects
4425 that were modified recently.
4427 reload-into-ims changes client no-cache or ``reload''
4428 to If-Modified-Since requests. Doing this VIOLATES the
4429 HTTP standard. Enabling this feature could make you
4430 liable for problems which it causes.
4432 ignore-reload ignores a client no-cache or ``reload''
4433 header. Doing this VIOLATES the HTTP standard. Enabling
4434 this feature could make you liable for problems which
4437 ignore-no-store ignores any ``Cache-control: no-store''
4438 headers received from a server. Doing this VIOLATES
4439 the HTTP standard. Enabling this feature could make you
4440 liable for problems which it causes.
4442 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
4443 headers received from a server. Doing this VIOLATES
4444 the HTTP standard. Enabling this feature could make you
4445 liable for problems which it causes.
4447 ignore-private ignores any ``Cache-control: private''
4448 headers received from a server. Doing this VIOLATES
4449 the HTTP standard. Enabling this feature could make you
4450 liable for problems which it causes.
4452 ignore-auth caches responses to requests with authorization,
4453 as if the originserver had sent ``Cache-control: public''
4454 in the response header. Doing this VIOLATES the HTTP standard.
4455 Enabling this feature could make you liable for problems which
4458 refresh-ims causes squid to contact the origin server
4459 when a client issues an If-Modified-Since request. This
4460 ensures that the client will receive an updated version
4461 if one is available.
4463 store-stale stores responses even if they don't have explicit
4464 freshness or a validator (i.e., Last-Modified or an ETag)
4465 present, or if they're already stale. By default, Squid will
4466 not cache such responses because they usually can't be
4467 reused. Note that such responses will be stale by default.
4469 max-stale=NN provide a maximum staleness factor. Squid won't
4470 serve objects more stale than this even if it failed to
4471 validate the object. Default: use the max_stale global limit.
4473 Basically a cached object is:
4475 FRESH if expires < now, else STALE
4477 FRESH if lm-factor < percent, else STALE
4481 The refresh_pattern lines are checked in the order listed here.
4482 The first entry which matches is used. If none of the entries
4483 match the default will be used.
4485 Note, you must uncomment all the default lines if you want
4486 to change one. The default setting is only active if none is
4491 # Add any of your own refresh_pattern entries above these.
4492 refresh_pattern ^ftp: 1440 20% 10080
4493 refresh_pattern ^gopher: 1440 0% 1440
4494 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
4495 refresh_pattern . 0 20% 4320
4499 NAME: quick_abort_min
4503 LOC: Config.quickAbort.min
4506 NAME: quick_abort_max
4510 LOC: Config.quickAbort.max
4513 NAME: quick_abort_pct
4517 LOC: Config.quickAbort.pct
4519 The cache by default continues downloading aborted requests
4520 which are almost completed (less than 16 KB remaining). This
4521 may be undesirable on slow (e.g. SLIP) links and/or very busy
4522 caches. Impatient users may tie up file descriptors and
4523 bandwidth by repeatedly requesting and immediately aborting
4526 When the user aborts a request, Squid will check the
4527 quick_abort values to the amount of data transfered until
4530 If the transfer has less than 'quick_abort_min' KB remaining,
4531 it will finish the retrieval.
4533 If the transfer has more than 'quick_abort_max' KB remaining,
4534 it will abort the retrieval.
4536 If more than 'quick_abort_pct' of the transfer has completed,
4537 it will finish the retrieval.
4539 If you do not want any retrieval to continue after the client
4540 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
4543 If you want retrievals to always continue if they are being
4544 cached set 'quick_abort_min' to '-1 KB'.
4547 NAME: read_ahead_gap
4548 COMMENT: buffer-size
4550 LOC: Config.readAheadGap
4553 The amount of data the cache will buffer ahead of what has been
4554 sent to the client when retrieving an object from another server.
4558 IFDEF: USE_HTTP_VIOLATIONS
4561 LOC: Config.negativeTtl
4564 Set the Default Time-to-Live (TTL) for failed requests.
4565 Certain types of failures (such as "connection refused" and
4566 "404 Not Found") are able to be negatively-cached for a short time.
4567 Modern web servers should provide Expires: header, however if they
4568 do not this can provide a minimum TTL.
4569 The default is not to cache errors with unknown expiry details.
4571 Note that this is different from negative caching of DNS lookups.
4573 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4574 this feature could make you liable for problems which it
4578 NAME: positive_dns_ttl
4581 LOC: Config.positiveDnsTtl
4584 Upper limit on how long Squid will cache positive DNS responses.
4585 Default is 6 hours (360 minutes). This directive must be set
4586 larger than negative_dns_ttl.
4589 NAME: negative_dns_ttl
4592 LOC: Config.negativeDnsTtl
4595 Time-to-Live (TTL) for negative caching of failed DNS lookups.
4596 This also sets the lower cache limit on positive lookups.
4597 Minimum value is 1 second, and it is not recommendable to go
4598 much below 10 seconds.
4601 NAME: range_offset_limit
4602 COMMENT: size [acl acl...]
4604 LOC: Config.rangeOffsetLimit
4607 usage: (size) [units] [[!]aclname]
4609 Sets an upper limit on how far (number of bytes) into the file
4610 a Range request may be to cause Squid to prefetch the whole file.
4611 If beyond this limit, Squid forwards the Range request as it is and
4612 the result is NOT cached.
4614 This is to stop a far ahead range request (lets say start at 17MB)
4615 from making Squid fetch the whole object up to that point before
4616 sending anything to the client.
4618 Multiple range_offset_limit lines may be specified, and they will
4619 be searched from top to bottom on each request until a match is found.
4620 The first match found will be used. If no line matches a request, the
4621 default limit of 0 bytes will be used.
4623 'size' is the limit specified as a number of units.
4625 'units' specifies whether to use bytes, KB, MB, etc.
4626 If no units are specified bytes are assumed.
4628 A size of 0 causes Squid to never fetch more than the
4629 client requested. (default)
4631 A size of 'none' causes Squid to always fetch the object from the
4632 beginning so it may cache the result. (2.0 style)
4634 'aclname' is the name of a defined ACL.
4636 NP: Using 'none' as the byte value here will override any quick_abort settings
4637 that may otherwise apply to the range request. The range request will
4638 be fully fetched from start to finish regardless of the client
4639 actions. This affects bandwidth usage.
4642 NAME: minimum_expiry_time
4645 LOC: Config.minimum_expiry_time
4648 The minimum caching time according to (Expires - Date)
4649 Headers Squid honors if the object can't be revalidated
4650 defaults to 60 seconds. In reverse proxy environments it
4651 might be desirable to honor shorter object lifetimes. It
4652 is most likely better to make your server return a
4653 meaningful Last-Modified header however. In ESI environments
4654 where page fragments often have short lifetimes, this will
4655 often be best set to 0.
4658 NAME: store_avg_object_size
4662 LOC: Config.Store.avgObjectSize
4664 Average object size, used to estimate number of objects your
4665 cache can hold. The default is 13 KB.
4668 NAME: store_objects_per_bucket
4671 LOC: Config.Store.objectsPerBucket
4673 Target number of objects per bucket in the store hash table.
4674 Lowering this value increases the total number of buckets and
4675 also the storage maintenance rate. The default is 20.
4680 -----------------------------------------------------------------------------
4683 NAME: request_header_max_size
4687 LOC: Config.maxRequestHeaderSize
4689 This specifies the maximum size for HTTP headers in a request.
4690 Request headers are usually relatively small (about 512 bytes).
4691 Placing a limit on the request header size will catch certain
4692 bugs (for example with persistent connections) and possibly
4693 buffer-overflow or denial-of-service attacks.
4696 NAME: reply_header_max_size
4700 LOC: Config.maxReplyHeaderSize
4702 This specifies the maximum size for HTTP headers in a reply.
4703 Reply headers are usually relatively small (about 512 bytes).
4704 Placing a limit on the reply header size will catch certain
4705 bugs (for example with persistent connections) and possibly
4706 buffer-overflow or denial-of-service attacks.
4709 NAME: request_body_max_size
4713 LOC: Config.maxRequestBodySize
4715 This specifies the maximum size for an HTTP request body.
4716 In other words, the maximum size of a PUT/POST request.
4717 A user who attempts to send a request with a body larger
4718 than this limit receives an "Invalid Request" error message.
4719 If you set this parameter to a zero (the default), there will
4720 be no limit imposed.
4723 NAME: client_request_buffer_max_size
4727 LOC: Config.maxRequestBufferSize
4729 This specifies the maximum buffer size of a client request.
4730 It prevents squid eating too much memory when somebody uploads
4734 NAME: chunked_request_body_max_size
4738 LOC: Config.maxChunkedRequestBodySize
4740 A broken or confused HTTP/1.1 client may send a chunked HTTP
4741 request to Squid. Squid does not have full support for that
4742 feature yet. To cope with such requests, Squid buffers the
4743 entire request and then dechunks request body to create a
4744 plain HTTP/1.0 request with a known content length. The plain
4745 request is then used by the rest of Squid code as usual.
4747 The option value specifies the maximum size of the buffer used
4748 to hold the request before the conversion. If the chunked
4749 request size exceeds the specified limit, the conversion
4750 fails, and the client receives an "unsupported request" error,
4751 as if dechunking was disabled.
4753 Dechunking is enabled by default. To disable conversion of
4754 chunked requests, set the maximum to zero.
4756 Request dechunking feature and this option in particular are a
4757 temporary hack. When chunking requests and responses are fully
4758 supported, there will be no need to buffer a chunked request.
4762 IFDEF: USE_HTTP_VIOLATIONS
4765 LOC: Config.accessList.brokenPosts
4767 A list of ACL elements which, if matched, causes Squid to send
4768 an extra CRLF pair after the body of a PUT/POST request.
4770 Some HTTP servers has broken implementations of PUT/POST,
4771 and rely on an extra CRLF pair sent by some WWW clients.
4773 Quote from RFC2616 section 4.1 on this matter:
4775 Note: certain buggy HTTP/1.0 client implementations generate an
4776 extra CRLF's after a POST request. To restate what is explicitly
4777 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4778 a request with an extra CRLF.
4780 This clause only supports fast acl types.
4781 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4784 acl buggy_server url_regex ^http://....
4785 broken_posts allow buggy_server
4788 NAME: adaptation_uses_indirect_client icap_uses_indirect_client
4791 IFDEF: FOLLOW_X_FORWARDED_FOR&&USE_ADAPTATION
4793 LOC: Adaptation::Config::use_indirect_client
4795 Controls whether the indirect client IP address (instead of the direct
4796 client IP address) is passed to adaptation services.
4798 See also: follow_x_forwarded_for adaptation_send_client_ip
4802 IFDEF: USE_HTTP_VIOLATIONS
4806 LOC: Config.onoff.via
4808 If set (default), Squid will include a Via header in requests and
4809 replies as required by RFC2616.
4815 LOC: Config.onoff.ie_refresh
4818 Microsoft Internet Explorer up until version 5.5 Service
4819 Pack 1 has an issue with transparent proxies, wherein it
4820 is impossible to force a refresh. Turning this on provides
4821 a partial fix to the problem, by causing all IMS-REFRESH
4822 requests from older IE versions to check the origin server
4823 for fresh content. This reduces hit ratio by some amount
4824 (~10% in my experience), but allows users to actually get
4825 fresh content when they want it. Note because Squid
4826 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4827 of 5.5 is unchanged from old versions of Squid (i.e. a
4828 forced refresh is impossible). Newer versions of IE will,
4829 hopefully, continue to have the new behavior and will be
4830 handled based on that assumption. This option defaults to
4831 the old Squid behavior, which is better for hit ratios but
4832 worse for clients using IE, if they need to be able to
4833 force fresh content.
4836 NAME: vary_ignore_expire
4839 LOC: Config.onoff.vary_ignore_expire
4842 Many HTTP servers supporting Vary gives such objects
4843 immediate expiry time with no cache-control header
4844 when requested by a HTTP/1.0 client. This option
4845 enables Squid to ignore such expiry times until
4846 HTTP/1.1 is fully implemented.
4848 WARNING: If turned on this may eventually cause some
4849 varying objects not intended for caching to get cached.
4852 NAME: request_entities
4854 LOC: Config.onoff.request_entities
4857 Squid defaults to deny GET and HEAD requests with request entities,
4858 as the meaning of such requests are undefined in the HTTP standard
4859 even if not explicitly forbidden.
4861 Set this directive to on if you have clients which insists
4862 on sending request entities in GET or HEAD requests. But be warned
4863 that there is server software (both proxies and web servers) which
4864 can fail to properly process this kind of request which may make you
4865 vulnerable to cache pollution attacks if enabled.
4868 NAME: request_header_access
4869 IFDEF: USE_HTTP_VIOLATIONS
4870 TYPE: http_header_access
4871 LOC: Config.request_header_access
4874 Usage: request_header_access header_name allow|deny [!]aclname ...
4876 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4877 this feature could make you liable for problems which it
4880 This option replaces the old 'anonymize_headers' and the
4881 older 'http_anonymizer' option with something that is much
4882 more configurable. A list of ACLs for each header name allows
4883 removal of specific header fields under specific conditions.
4885 This option only applies to outgoing HTTP request headers (i.e.,
4886 headers sent by Squid to the next HTTP hop such as a cache peer
4887 or an origin server). The option has no effect during cache hit
4888 detection. The equivalent adaptation vectoring point in ICAP
4889 terminology is post-cache REQMOD.
4891 The option is applied to individual outgoing request header
4892 fields. For each request header field F, Squid uses the first
4893 qualifying sets of request_header_access rules:
4895 1. Rules with header_name equal to F's name.
4896 2. Rules with header_name 'Other', provided F's name is not
4897 on the hard-coded list of commonly used HTTP header names.
4898 3. Rules with header_name 'All'.
4900 Within that qualifying rule set, rule ACLs are checked as usual.
4901 If ACLs of an "allow" rule match, the header field is allowed to
4902 go through as is. If ACLs of a "deny" rule match, the header is
4903 removed and request_header_replace is then checked to identify
4904 if the removed header has a replacement. If no rules within the
4905 set have matching ACLs, the header field is left as is.
4907 For example, to achieve the same behavior as the old
4908 'http_anonymizer standard' option, you should use:
4910 request_header_access From deny all
4911 request_header_access Referer deny all
4912 request_header_access Server deny all
4913 request_header_access User-Agent deny all
4914 request_header_access WWW-Authenticate deny all
4915 request_header_access Link deny all
4917 Or, to reproduce the old 'http_anonymizer paranoid' feature
4920 request_header_access Allow allow all
4921 request_header_access Authorization allow all
4922 request_header_access WWW-Authenticate allow all
4923 request_header_access Proxy-Authorization allow all
4924 request_header_access Proxy-Authenticate allow all
4925 request_header_access Cache-Control allow all
4926 request_header_access Content-Encoding allow all
4927 request_header_access Content-Length allow all
4928 request_header_access Content-Type allow all
4929 request_header_access Date allow all
4930 request_header_access Expires allow all
4931 request_header_access Host allow all
4932 request_header_access If-Modified-Since allow all
4933 request_header_access Last-Modified allow all
4934 request_header_access Location allow all
4935 request_header_access Pragma allow all
4936 request_header_access Accept allow all
4937 request_header_access Accept-Charset allow all
4938 request_header_access Accept-Encoding allow all
4939 request_header_access Accept-Language allow all
4940 request_header_access Content-Language allow all
4941 request_header_access Mime-Version allow all
4942 request_header_access Retry-After allow all
4943 request_header_access Title allow all
4944 request_header_access Connection allow all
4945 request_header_access All deny all
4947 although many of those are HTTP reply headers, and so should be
4948 controlled with the reply_header_access directive.
4950 By default, all headers are allowed (no anonymizing is
4954 NAME: reply_header_access
4955 IFDEF: USE_HTTP_VIOLATIONS
4956 TYPE: http_header_access
4957 LOC: Config.reply_header_access
4960 Usage: reply_header_access header_name allow|deny [!]aclname ...
4962 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4963 this feature could make you liable for problems which it
4966 This option only applies to reply headers, i.e., from the
4967 server to the client.
4969 This is the same as request_header_access, but in the other
4970 direction. Please see request_header_access for detailed
4973 For example, to achieve the same behavior as the old
4974 'http_anonymizer standard' option, you should use:
4976 reply_header_access From deny all
4977 reply_header_access Referer deny all
4978 reply_header_access Server deny all
4979 reply_header_access User-Agent deny all
4980 reply_header_access WWW-Authenticate deny all
4981 reply_header_access Link deny all
4983 Or, to reproduce the old 'http_anonymizer paranoid' feature
4986 reply_header_access Allow allow all
4987 reply_header_access Authorization allow all
4988 reply_header_access WWW-Authenticate allow all
4989 reply_header_access Proxy-Authorization allow all
4990 reply_header_access Proxy-Authenticate allow all
4991 reply_header_access Cache-Control allow all
4992 reply_header_access Content-Encoding allow all
4993 reply_header_access Content-Length allow all
4994 reply_header_access Content-Type allow all
4995 reply_header_access Date allow all
4996 reply_header_access Expires allow all
4997 reply_header_access Host allow all
4998 reply_header_access If-Modified-Since allow all
4999 reply_header_access Last-Modified allow all
5000 reply_header_access Location allow all
5001 reply_header_access Pragma allow all
5002 reply_header_access Accept allow all
5003 reply_header_access Accept-Charset allow all
5004 reply_header_access Accept-Encoding allow all
5005 reply_header_access Accept-Language allow all
5006 reply_header_access Content-Language allow all
5007 reply_header_access Mime-Version allow all
5008 reply_header_access Retry-After allow all
5009 reply_header_access Title allow all
5010 reply_header_access Connection allow all
5011 reply_header_access All deny all
5013 although the HTTP request headers won't be usefully controlled
5014 by this directive -- see request_header_access for details.
5016 By default, all headers are allowed (no anonymizing is
5020 NAME: request_header_replace header_replace
5021 IFDEF: USE_HTTP_VIOLATIONS
5022 TYPE: http_header_replace
5023 LOC: Config.request_header_access
5026 Usage: request_header_replace header_name message
5027 Example: request_header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
5029 This option allows you to change the contents of headers
5030 denied with request_header_access above, by replacing them
5031 with some fixed string. This replaces the old fake_user_agent
5034 This only applies to request headers, not reply headers.
5036 By default, headers are removed if denied.
5039 NAME: reply_header_replace
5040 IFDEF: USE_HTTP_VIOLATIONS
5041 TYPE: http_header_replace
5042 LOC: Config.reply_header_access
5045 Usage: reply_header_replace header_name message
5046 Example: reply_header_replace Server Foo/1.0
5048 This option allows you to change the contents of headers
5049 denied with reply_header_access above, by replacing them
5050 with some fixed string.
5052 This only applies to reply headers, not request headers.
5054 By default, headers are removed if denied.
5057 NAME: request_header_add
5058 TYPE: HeaderWithAclList
5059 LOC: Config.request_header_add
5062 Usage: request_header_add field-name field-value acl1 [acl2] ...
5063 Example: request_header_add X-Client-CA "CA=%ssl::>cert_issuer" all
5065 This option adds header fields to outgoing HTTP requests (i.e.,
5066 request headers sent by Squid to the next HTTP hop such as a
5067 cache peer or an origin server). The option has no effect during
5068 cache hit detection. The equivalent adaptation vectoring point
5069 in ICAP terminology is post-cache REQMOD.
5071 Field-name is a token specifying an HTTP header name. If a
5072 standard HTTP header name is used, Squid does not check whether
5073 the new header conflicts with any existing headers or violates
5074 HTTP rules. If the request to be modified already contains a
5075 field with the same name, the old field is preserved but the
5076 header field values are not merged.
5078 Field-value is either a token or a quoted string. If quoted
5079 string format is used, then the surrounding quotes are removed
5080 while escape sequences and %macros are processed.
5082 In theory, all of the logformat codes can be used as %macros.
5083 However, unlike logging (which happens at the very end of
5084 transaction lifetime), the transaction may not yet have enough
5085 information to expand a macro when the new header value is needed.
5086 And some information may already be available to Squid but not yet
5087 committed where the macro expansion code can access it (report
5088 such instances!). The macro will be expanded into a single dash
5089 ('-') in such cases. Not all macros have been tested.
5091 One or more Squid ACLs may be specified to restrict header
5092 injection to matching requests. As always in squid.conf, all
5093 ACLs in an option ACL list must be satisfied for the insertion
5094 to happen. The request_header_add option supports fast ACLs
5103 This option used to log custom information about the master
5104 transaction. For example, an admin may configure Squid to log
5105 which "user group" the transaction belongs to, where "user group"
5106 will be determined based on a set of ACLs and not [just]
5107 authentication information.
5108 Values of key/value pairs can be logged using %{key}note macros:
5110 note key value acl ...
5111 logformat myFormat ... %{key}note ...
5114 NAME: relaxed_header_parser
5115 COMMENT: on|off|warn
5117 LOC: Config.onoff.relaxed_header_parser
5120 In the default "on" setting Squid accepts certain forms
5121 of non-compliant HTTP messages where it is unambiguous
5122 what the sending application intended even if the message
5123 is not correctly formatted. The messages is then normalized
5124 to the correct form when forwarded by Squid.
5126 If set to "warn" then a warning will be emitted in cache.log
5127 each time such HTTP error is encountered.
5129 If set to "off" then such HTTP errors will cause the request
5130 or response to be rejected.
5135 -----------------------------------------------------------------------------
5138 NAME: forward_timeout
5141 LOC: Config.Timeout.forward
5144 This parameter specifies how long Squid should at most attempt in
5145 finding a forwarding path for the request before giving up.
5148 NAME: connect_timeout
5151 LOC: Config.Timeout.connect
5154 This parameter specifies how long to wait for the TCP connect to
5155 the requested server or peer to complete before Squid should
5156 attempt to find another path where to forward the request.
5159 NAME: peer_connect_timeout
5162 LOC: Config.Timeout.peer_connect
5165 This parameter specifies how long to wait for a pending TCP
5166 connection to a peer cache. The default is 30 seconds. You
5167 may also set different timeout values for individual neighbors
5168 with the 'connect-timeout' option on a 'cache_peer' line.
5174 LOC: Config.Timeout.read
5177 The read_timeout is applied on server-side connections. After
5178 each successful read(), the timeout will be extended by this
5179 amount. If no data is read again after this amount of time,
5180 the request is aborted and logged with ERR_READ_TIMEOUT. The
5181 default is 15 minutes.
5187 LOC: Config.Timeout.write
5190 This timeout is tracked for all connections that have data
5191 available for writing and are waiting for the socket to become
5192 ready. After each successful write, the timeout is extended by
5193 the configured amount. If Squid has data to write but the
5194 connection is not ready for the configured duration, the
5195 transaction associated with the connection is terminated. The
5196 default is 15 minutes.
5199 NAME: request_timeout
5201 LOC: Config.Timeout.request
5204 How long to wait for complete HTTP request headers after initial
5205 connection establishment.
5208 NAME: client_idle_pconn_timeout persistent_request_timeout
5210 LOC: Config.Timeout.clientIdlePconn
5213 How long to wait for the next HTTP request on a persistent
5214 client connection after the previous request completes.
5217 NAME: client_lifetime
5220 LOC: Config.Timeout.lifetime
5223 The maximum amount of time a client (browser) is allowed to
5224 remain connected to the cache process. This protects the Cache
5225 from having a lot of sockets (and hence file descriptors) tied up
5226 in a CLOSE_WAIT state from remote clients that go away without
5227 properly shutting down (either because of a network failure or
5228 because of a poor client implementation). The default is one
5231 NOTE: The default value is intended to be much larger than any
5232 client would ever need to be connected to your cache. You
5233 should probably change client_lifetime only as a last resort.
5234 If you seem to have many client connections tying up
5235 filedescriptors, we recommend first tuning the read_timeout,
5236 request_timeout, persistent_request_timeout and quick_abort values.
5239 NAME: half_closed_clients
5241 LOC: Config.onoff.half_closed_clients
5244 Some clients may shutdown the sending side of their TCP
5245 connections, while leaving their receiving sides open. Sometimes,
5246 Squid can not tell the difference between a half-closed and a
5247 fully-closed TCP connection.
5249 By default, Squid will immediately close client connections when
5250 read(2) returns "no more data to read."
5252 Change this option to 'on' and Squid will keep open connections
5253 until a read(2) or write(2) on the socket returns an error.
5254 This may show some benefits for reverse proxies. But if not
5255 it is recommended to leave OFF.
5258 NAME: server_idle_pconn_timeout pconn_timeout
5260 LOC: Config.Timeout.serverIdlePconn
5263 Timeout for idle persistent connections to servers and other
5270 LOC: Ident::TheConfig.timeout
5273 Maximum time to wait for IDENT lookups to complete.
5275 If this is too high, and you enabled IDENT lookups from untrusted
5276 users, you might be susceptible to denial-of-service by having
5277 many ident requests going at once.
5280 NAME: shutdown_lifetime
5283 LOC: Config.shutdownLifetime
5286 When SIGTERM or SIGHUP is received, the cache is put into
5287 "shutdown pending" mode until all active sockets are closed.
5288 This value is the lifetime to set for all open descriptors
5289 during shutdown mode. Any active clients after this many
5290 seconds will receive a 'timeout' message.
5294 ADMINISTRATIVE PARAMETERS
5295 -----------------------------------------------------------------------------
5301 LOC: Config.adminEmail
5303 Email-address of local cache manager who will receive
5304 mail if the cache dies. The default is "webmaster."
5310 LOC: Config.EmailFrom
5312 From: email-address for mail sent when the cache dies.
5313 The default is to use 'appname@unique_hostname'.
5314 Default appname value is "squid", can be changed into
5315 src/globals.h before building squid.
5321 LOC: Config.EmailProgram
5323 Email program used to send mail if the cache dies.
5324 The default is "mail". The specified program must comply
5325 with the standard Unix mail syntax:
5326 mail-program recipient < mailfile
5328 Optional command line options can be specified.
5331 NAME: cache_effective_user
5333 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
5334 LOC: Config.effectiveUser
5336 If you start Squid as root, it will change its effective/real
5337 UID/GID to the user specified below. The default is to change
5338 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
5339 see also; cache_effective_group
5342 NAME: cache_effective_group
5345 LOC: Config.effectiveGroup
5347 Squid sets the GID to the effective user's default group ID
5348 (taken from the password file) and supplementary group list
5349 from the groups membership.
5351 If you want Squid to run with a specific GID regardless of
5352 the group memberships of the effective user then set this
5353 to the group (or GID) you want Squid to run as. When set
5354 all other group privileges of the effective user are ignored
5355 and only this GID is effective. If Squid is not started as
5356 root the user starting Squid MUST be member of the specified
5359 This option is not recommended by the Squid Team.
5360 Our preference is for administrators to configure a secure
5361 user account for squid with UID/GID matching system policies.
5364 NAME: httpd_suppress_version_string
5368 LOC: Config.onoff.httpd_suppress_version_string
5370 Suppress Squid version string info in HTTP headers and HTML error pages.
5373 NAME: visible_hostname
5375 LOC: Config.visibleHostname
5378 If you want to present a special hostname in error messages, etc,
5379 define this. Otherwise, the return value of gethostname()
5380 will be used. If you have multiple caches in a cluster and
5381 get errors about IP-forwarding you must set them to have individual
5382 names with this setting.
5385 NAME: unique_hostname
5387 LOC: Config.uniqueHostname
5390 If you want to have multiple machines with the same
5391 'visible_hostname' you must give each machine a different
5392 'unique_hostname' so forwarding loops can be detected.
5395 NAME: hostname_aliases
5397 LOC: Config.hostnameAliases
5400 A list of other DNS names your cache has.
5408 Minimum umask which should be enforced while the proxy
5409 is running, in addition to the umask set at startup.
5411 For a traditional octal representation of umasks, start
5416 OPTIONS FOR THE CACHE REGISTRATION SERVICE
5417 -----------------------------------------------------------------------------
5419 This section contains parameters for the (optional) cache
5420 announcement service. This service is provided to help
5421 cache administrators locate one another in order to join or
5422 create cache hierarchies.
5424 An 'announcement' message is sent (via UDP) to the registration
5425 service by Squid. By default, the announcement message is NOT
5426 SENT unless you enable it with 'announce_period' below.
5428 The announcement message includes your hostname, plus the
5429 following information from this configuration file:
5435 All current information is processed regularly and made
5436 available on the Web at http://www.ircache.net/Cache/Tracker/.
5439 NAME: announce_period
5441 LOC: Config.Announce.period
5444 This is how frequently to send cache announcements. The
5445 default is `0' which disables sending the announcement
5448 To enable announcing your cache, just set an announce period.
5451 announce_period 1 day
5456 DEFAULT: tracker.ircache.net
5457 LOC: Config.Announce.host
5463 LOC: Config.Announce.file
5469 LOC: Config.Announce.port
5471 announce_host and announce_port set the hostname and port
5472 number where the registration message will be sent.
5474 Hostname will default to 'tracker.ircache.net' and port will
5475 default default to 3131. If the 'filename' argument is given,
5476 the contents of that file will be included in the announce
5481 HTTPD-ACCELERATOR OPTIONS
5482 -----------------------------------------------------------------------------
5485 NAME: httpd_accel_surrogate_id
5488 LOC: Config.Accel.surrogate_id
5490 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
5491 need an identification token to allow control targeting. Because
5492 a farm of surrogates may all perform the same tasks, they may share
5493 an identification token.
5495 The default ID is the visible_hostname
5498 NAME: http_accel_surrogate_remote
5502 LOC: Config.onoff.surrogate_is_remote
5504 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
5505 Set this to on to have squid behave as a remote surrogate.
5509 IFDEF: USE_SQUID_ESI
5510 COMMENT: libxml2|expat|custom
5512 LOC: ESIParser::Type
5515 ESI markup is not strictly XML compatible. The custom ESI parser
5516 will give higher performance, but cannot handle non ASCII character
5521 DELAY POOL PARAMETERS
5522 -----------------------------------------------------------------------------
5526 TYPE: delay_pool_count
5528 IFDEF: USE_DELAY_POOLS
5531 This represents the number of delay pools to be used. For example,
5532 if you have one class 2 delay pool and one class 3 delays pool, you
5533 have a total of 2 delay pools.
5537 TYPE: delay_pool_class
5539 IFDEF: USE_DELAY_POOLS
5542 This defines the class of each delay pool. There must be exactly one
5543 delay_class line for each delay pool. For example, to define two
5544 delay pools, one of class 2 and one of class 3, the settings above
5548 delay_pools 4 # 4 delay pools
5549 delay_class 1 2 # pool 1 is a class 2 pool
5550 delay_class 2 3 # pool 2 is a class 3 pool
5551 delay_class 3 4 # pool 3 is a class 4 pool
5552 delay_class 4 5 # pool 4 is a class 5 pool
5554 The delay pool classes are:
5556 class 1 Everything is limited by a single aggregate
5559 class 2 Everything is limited by a single aggregate
5560 bucket as well as an "individual" bucket chosen
5561 from bits 25 through 32 of the IPv4 address.
5563 class 3 Everything is limited by a single aggregate
5564 bucket as well as a "network" bucket chosen
5565 from bits 17 through 24 of the IP address and a
5566 "individual" bucket chosen from bits 17 through
5567 32 of the IPv4 address.
5569 class 4 Everything in a class 3 delay pool, with an
5570 additional limit on a per user basis. This
5571 only takes effect if the username is established
5572 in advance - by forcing authentication in your
5575 class 5 Requests are grouped according their tag (see
5576 external_acl's tag= reply).
5579 Each pool also requires a delay_parameters directive to configure the pool size
5580 and speed limits used whenever the pool is applied to a request. Along with
5581 a set of delay_access directives to determine when it is used.
5583 NOTE: If an IP address is a.b.c.d
5584 -> bits 25 through 32 are "d"
5585 -> bits 17 through 24 are "c"
5586 -> bits 17 through 32 are "c * 256 + d"
5588 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
5589 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
5593 TYPE: delay_pool_access
5595 IFDEF: USE_DELAY_POOLS
5598 This is used to determine which delay pool a request falls into.
5600 delay_access is sorted per pool and the matching starts with pool 1,
5601 then pool 2, ..., and finally pool N. The first delay pool where the
5602 request is allowed is selected for the request. If it does not allow
5603 the request to any pool then the request is not delayed (default).
5605 For example, if you want some_big_clients in delay
5606 pool 1 and lotsa_little_clients in delay pool 2:
5609 delay_access 1 allow some_big_clients
5610 delay_access 1 deny all
5611 delay_access 2 allow lotsa_little_clients
5612 delay_access 2 deny all
5613 delay_access 3 allow authenticated_clients
5616 NAME: delay_parameters
5617 TYPE: delay_pool_rates
5619 IFDEF: USE_DELAY_POOLS
5622 This defines the parameters for a delay pool. Each delay pool has
5623 a number of "buckets" associated with it, as explained in the
5624 description of delay_class.
5626 For a class 1 delay pool, the syntax is:
5628 delay_parameters pool aggregate
5630 For a class 2 delay pool:
5632 delay_parameters pool aggregate individual
5634 For a class 3 delay pool:
5636 delay_parameters pool aggregate network individual
5638 For a class 4 delay pool:
5640 delay_parameters pool aggregate network individual user
5642 For a class 5 delay pool:
5644 delay_parameters pool tagrate
5646 The option variables are:
5648 pool a pool number - ie, a number between 1 and the
5649 number specified in delay_pools as used in
5652 aggregate the speed limit parameters for the aggregate bucket
5655 individual the speed limit parameters for the individual
5656 buckets (class 2, 3).
5658 network the speed limit parameters for the network buckets
5661 user the speed limit parameters for the user buckets
5664 tagrate the speed limit parameters for the tag buckets
5667 A pair of delay parameters is written restore/maximum, where restore is
5668 the number of bytes (not bits - modem and network speeds are usually
5669 quoted in bits) per second placed into the bucket, and maximum is the
5670 maximum number of bytes which can be in the bucket at any time.
5672 There must be one delay_parameters line for each delay pool.
5675 For example, if delay pool number 1 is a class 2 delay pool as in the
5676 above example, and is being used to strictly limit each host to 64Kbit/sec
5677 (plus overheads), with no overall limit, the line is:
5679 delay_parameters 1 -1/-1 8000/8000
5681 Note that 8 x 8000 KByte/sec -> 64Kbit/sec.
5683 Note that the figure -1 is used to represent "unlimited".
5686 And, if delay pool number 2 is a class 3 delay pool as in the above
5687 example, and you want to limit it to a total of 256Kbit/sec (strict limit)
5688 with each 8-bit network permitted 64Kbit/sec (strict limit) and each
5689 individual host permitted 4800bit/sec with a bucket maximum size of 64Kbits
5690 to permit a decent web page to be downloaded at a decent speed
5691 (if the network is not being limited due to overuse) but slow down
5692 large downloads more significantly:
5694 delay_parameters 2 32000/32000 8000/8000 600/8000
5696 Note that 8 x 32000 KByte/sec -> 256Kbit/sec.
5697 8 x 8000 KByte/sec -> 64Kbit/sec.
5698 8 x 600 Byte/sec -> 4800bit/sec.
5701 Finally, for a class 4 delay pool as in the example - each user will
5702 be limited to 128Kbits/sec no matter how many workstations they are logged into.:
5704 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
5707 NAME: delay_initial_bucket_level
5708 COMMENT: (percent, 0-100)
5711 IFDEF: USE_DELAY_POOLS
5712 LOC: Config.Delay.initial
5714 The initial bucket percentage is used to determine how much is put
5715 in each bucket when squid starts, is reconfigured, or first notices
5716 a host accessing it (in class 2 and class 3, individual hosts and
5717 networks only have buckets associated with them once they have been
5722 CLIENT DELAY POOL PARAMETERS
5723 -----------------------------------------------------------------------------
5726 NAME: client_delay_pools
5727 TYPE: client_delay_pool_count
5729 IFDEF: USE_DELAY_POOLS
5730 LOC: Config.ClientDelay
5732 This option specifies the number of client delay pools used. It must
5733 preceed other client_delay_* options.
5736 client_delay_pools 2
5739 NAME: client_delay_initial_bucket_level
5740 COMMENT: (percent, 0-no_limit)
5743 IFDEF: USE_DELAY_POOLS
5744 LOC: Config.ClientDelay.initial
5746 This option determines the initial bucket size as a percentage of
5747 max_bucket_size from client_delay_parameters. Buckets are created
5748 at the time of the "first" connection from the matching IP. Idle
5749 buckets are periodically deleted up.
5751 You can specify more than 100 percent but note that such "oversized"
5752 buckets are not refilled until their size goes down to max_bucket_size
5753 from client_delay_parameters.
5756 client_delay_initial_bucket_level 50
5759 NAME: client_delay_parameters
5760 TYPE: client_delay_pool_rates
5762 IFDEF: USE_DELAY_POOLS
5763 LOC: Config.ClientDelay
5766 This option configures client-side bandwidth limits using the
5769 client_delay_parameters pool speed_limit max_bucket_size
5771 pool is an integer ID used for client_delay_access matching.
5773 speed_limit is bytes added to the bucket per second.
5775 max_bucket_size is the maximum size of a bucket, enforced after any
5776 speed_limit additions.
5778 Please see the delay_parameters option for more information and
5782 client_delay_parameters 1 1024 2048
5783 client_delay_parameters 2 51200 16384
5786 NAME: client_delay_access
5787 TYPE: client_delay_pool_access
5789 IFDEF: USE_DELAY_POOLS
5790 LOC: Config.ClientDelay
5793 This option determines the client-side delay pool for the
5796 client_delay_access pool_ID allow|deny acl_name
5798 All client_delay_access options are checked in their pool ID
5799 order, starting with pool 1. The first checked pool with allowed
5800 request is selected for the request. If no ACL matches or there
5801 are no client_delay_access options, the request bandwidth is not
5804 The ACL-selected pool is then used to find the
5805 client_delay_parameters for the request. Client-side pools are
5806 not used to aggregate clients. Clients are always aggregated
5807 based on their source IP addresses (one bucket per source IP).
5809 Please see delay_access for more examples.
5812 client_delay_access 1 allow low_rate_network
5813 client_delay_access 2 allow vips_network
5817 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
5818 -----------------------------------------------------------------------------
5823 LOC: Config.Wccp.router
5827 Use this option to define your WCCP ``home'' router for
5830 wccp_router supports a single WCCP(v1) router
5832 wccp2_router supports multiple WCCPv2 routers
5834 only one of the two may be used at the same time and defines
5835 which version of WCCP to use.
5839 TYPE: IpAddress_list
5840 LOC: Config.Wccp2.router
5844 Use this option to define your WCCP ``home'' router for
5847 wccp_router supports a single WCCP(v1) router
5849 wccp2_router supports multiple WCCPv2 routers
5851 only one of the two may be used at the same time and defines
5852 which version of WCCP to use.
5857 LOC: Config.Wccp.version
5861 This directive is only relevant if you need to set up WCCP(v1)
5862 to some very old and end-of-life Cisco routers. In all other
5863 setups it must be left unset or at the default setting.
5864 It defines an internal version in the WCCP(v1) protocol,
5865 with version 4 being the officially documented protocol.
5867 According to some users, Cisco IOS 11.2 and earlier only
5868 support WCCP version 3. If you're using that or an earlier
5869 version of IOS, you may need to change this value to 3, otherwise
5870 do not specify this parameter.
5873 NAME: wccp2_rebuild_wait
5875 LOC: Config.Wccp2.rebuildwait
5879 If this is enabled Squid will wait for the cache dir rebuild to finish
5880 before sending the first wccp2 HereIAm packet
5883 NAME: wccp2_forwarding_method
5885 LOC: Config.Wccp2.forwarding_method
5889 WCCP2 allows the setting of forwarding methods between the
5890 router/switch and the cache. Valid values are as follows:
5892 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5893 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5895 Currently (as of IOS 12.4) cisco routers only support GRE.
5896 Cisco switches only support the L2 redirect assignment method.
5899 NAME: wccp2_return_method
5901 LOC: Config.Wccp2.return_method
5905 WCCP2 allows the setting of return methods between the
5906 router/switch and the cache for packets that the cache
5907 decides not to handle. Valid values are as follows:
5909 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
5910 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
5912 Currently (as of IOS 12.4) cisco routers only support GRE.
5913 Cisco switches only support the L2 redirect assignment.
5915 If the "ip wccp redirect exclude in" command has been
5916 enabled on the cache interface, then it is still safe for
5917 the proxy server to use a l2 redirect method even if this
5918 option is set to GRE.
5921 NAME: wccp2_assignment_method
5923 LOC: Config.Wccp2.assignment_method
5927 WCCP2 allows the setting of methods to assign the WCCP hash
5928 Valid values are as follows:
5930 hash - Hash assignment
5931 mask - Mask assignment
5933 As a general rule, cisco routers support the hash assignment method
5934 and cisco switches support the mask assignment method.
5939 LOC: Config.Wccp2.info
5940 DEFAULT_IF_NONE: standard 0
5943 WCCP2 allows for multiple traffic services. There are two
5944 types: "standard" and "dynamic". The standard type defines
5945 one service id - http (id 0). The dynamic service ids can be from
5946 51 to 255 inclusive. In order to use a dynamic service id
5947 one must define the type of traffic to be redirected; this is done
5948 using the wccp2_service_info option.
5950 The "standard" type does not require a wccp2_service_info option,
5951 just specifying the service id will suffice.
5953 MD5 service authentication can be enabled by adding
5954 "password=<password>" to the end of this service declaration.
5958 wccp2_service standard 0 # for the 'web-cache' standard service
5959 wccp2_service dynamic 80 # a dynamic service type which will be
5960 # fleshed out with subsequent options.
5961 wccp2_service standard 0 password=foo
5964 NAME: wccp2_service_info
5965 TYPE: wccp2_service_info
5966 LOC: Config.Wccp2.info
5970 Dynamic WCCPv2 services require further information to define the
5971 traffic you wish to have diverted.
5975 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
5976 priority=<priority> ports=<port>,<port>..
5978 The relevant WCCPv2 flags:
5979 + src_ip_hash, dst_ip_hash
5980 + source_port_hash, dst_port_hash
5981 + src_ip_alt_hash, dst_ip_alt_hash
5982 + src_port_alt_hash, dst_port_alt_hash
5985 The port list can be one to eight entries.
5989 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
5990 priority=240 ports=80
5992 Note: the service id must have been defined by a previous
5993 'wccp2_service dynamic <id>' entry.
5998 LOC: Config.Wccp2.weight
6002 Each cache server gets assigned a set of the destination
6003 hash proportional to their weight.
6008 LOC: Config.Wccp.address
6015 LOC: Config.Wccp2.address
6019 Use this option if you require WCCP to use a specific
6022 The default behavior is to not bind to any specific address.
6026 PERSISTENT CONNECTION HANDLING
6027 -----------------------------------------------------------------------------
6029 Also see "pconn_timeout" in the TIMEOUTS section
6032 NAME: client_persistent_connections
6034 LOC: Config.onoff.client_pconns
6038 NAME: server_persistent_connections
6040 LOC: Config.onoff.server_pconns
6043 Persistent connection support for clients and servers. By
6044 default, Squid uses persistent connections (when allowed)
6045 with its clients and servers. You can use these options to
6046 disable persistent connections with clients and/or servers.
6049 NAME: persistent_connection_after_error
6051 LOC: Config.onoff.error_pconns
6054 With this directive the use of persistent connections after
6055 HTTP errors can be disabled. Useful if you have clients
6056 who fail to handle errors on persistent connections proper.
6059 NAME: detect_broken_pconn
6061 LOC: Config.onoff.detect_broken_server_pconns
6064 Some servers have been found to incorrectly signal the use
6065 of HTTP/1.0 persistent connections even on replies not
6066 compatible, causing significant delays. This server problem
6067 has mostly been seen on redirects.
6069 By enabling this directive Squid attempts to detect such
6070 broken replies and automatically assume the reply is finished
6071 after 10 seconds timeout.
6075 CACHE DIGEST OPTIONS
6076 -----------------------------------------------------------------------------
6079 NAME: digest_generation
6080 IFDEF: USE_CACHE_DIGESTS
6082 LOC: Config.onoff.digest_generation
6085 This controls whether the server will generate a Cache Digest
6086 of its contents. By default, Cache Digest generation is
6087 enabled if Squid is compiled with --enable-cache-digests defined.
6090 NAME: digest_bits_per_entry
6091 IFDEF: USE_CACHE_DIGESTS
6093 LOC: Config.digest.bits_per_entry
6096 This is the number of bits of the server's Cache Digest which
6097 will be associated with the Digest entry for a given HTTP
6098 Method and URL (public key) combination. The default is 5.
6101 NAME: digest_rebuild_period
6102 IFDEF: USE_CACHE_DIGESTS
6105 LOC: Config.digest.rebuild_period
6108 This is the wait time between Cache Digest rebuilds.
6111 NAME: digest_rewrite_period
6113 IFDEF: USE_CACHE_DIGESTS
6115 LOC: Config.digest.rewrite_period
6118 This is the wait time between Cache Digest writes to
6122 NAME: digest_swapout_chunk_size
6125 IFDEF: USE_CACHE_DIGESTS
6126 LOC: Config.digest.swapout_chunk_size
6129 This is the number of bytes of the Cache Digest to write to
6130 disk at a time. It defaults to 4096 bytes (4KB), the Squid
6134 NAME: digest_rebuild_chunk_percentage
6135 COMMENT: (percent, 0-100)
6136 IFDEF: USE_CACHE_DIGESTS
6138 LOC: Config.digest.rebuild_chunk_percentage
6141 This is the percentage of the Cache Digest to be scanned at a
6142 time. By default it is set to 10% of the Cache Digest.
6147 -----------------------------------------------------------------------------
6152 LOC: Config.Port.snmp
6156 The port number where Squid listens for SNMP requests. To enable
6157 SNMP support set this to a suitable port number. Port number
6158 3401 is often used for the Squid SNMP agent. By default it's
6159 set to "0" (disabled)
6167 LOC: Config.accessList.snmp
6168 DEFAULT_IF_NONE: deny all
6171 Allowing or denying access to the SNMP port.
6173 All access to the agent is denied by default.
6176 snmp_access allow|deny [!]aclname ...
6178 This clause only supports fast acl types.
6179 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6181 snmp_access allow snmppublic localhost
6182 snmp_access deny all
6185 NAME: snmp_incoming_address
6187 LOC: Config.Addrs.snmp_incoming
6192 NAME: snmp_outgoing_address
6194 LOC: Config.Addrs.snmp_outgoing
6198 Just like 'udp_incoming_address', but for the SNMP port.
6200 snmp_incoming_address is used for the SNMP socket receiving
6201 messages from SNMP agents.
6202 snmp_outgoing_address is used for SNMP packets returned to SNMP
6205 The default snmp_incoming_address is to listen on all
6206 available network interfaces.
6208 If snmp_outgoing_address is not set it will use the same socket
6209 as snmp_incoming_address. Only change this if you want to have
6210 SNMP replies sent using another address than where this Squid
6211 listens for SNMP queries.
6213 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
6214 the same value since they both use port 3401.
6219 -----------------------------------------------------------------------------
6222 NAME: icp_port udp_port
6225 LOC: Config.Port.icp
6227 The port number where Squid sends and receives ICP queries to
6228 and from neighbor caches. The standard UDP port for ICP is 3130.
6229 Default is disabled (0).
6232 icp_port @DEFAULT_ICP_PORT@
6239 LOC: Config.Port.htcp
6241 The port number where Squid sends and receives HTCP queries to
6242 and from neighbor caches. To turn it on you want to set it to
6243 4827. By default it is set to "0" (disabled).
6249 NAME: log_icp_queries
6253 LOC: Config.onoff.log_udp
6255 If set, ICP queries are logged to access.log. You may wish
6256 do disable this if your ICP load is VERY high to speed things
6257 up or to simplify log analysis.
6260 NAME: udp_incoming_address
6262 LOC:Config.Addrs.udp_incoming
6265 udp_incoming_address is used for UDP packets received from other
6268 The default behavior is to not bind to any specific address.
6270 Only change this if you want to have all UDP queries received on
6271 a specific interface/address.
6273 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
6274 modules. Altering it will affect all of them in the same manner.
6276 see also; udp_outgoing_address
6278 NOTE, udp_incoming_address and udp_outgoing_address can not
6279 have the same value since they both use the same port.
6282 NAME: udp_outgoing_address
6284 LOC: Config.Addrs.udp_outgoing
6287 udp_outgoing_address is used for UDP packets sent out to other
6290 The default behavior is to not bind to any specific address.
6292 Instead it will use the same socket as udp_incoming_address.
6293 Only change this if you want to have UDP queries sent using another
6294 address than where this Squid listens for UDP queries from other
6297 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
6298 modules. Altering it will affect all of them in the same manner.
6300 see also; udp_incoming_address
6302 NOTE, udp_incoming_address and udp_outgoing_address can not
6303 have the same value since they both use the same port.
6310 LOC: Config.onoff.icp_hit_stale
6312 If you want to return ICP_HIT for stale cache objects, set this
6313 option to 'on'. If you have sibling relationships with caches
6314 in other administrative domains, this should be 'off'. If you only
6315 have sibling relationships with caches under your control,
6316 it is probably okay to set this to 'on'.
6317 If set to 'on', your siblings should use the option "allow-miss"
6318 on their cache_peer lines for connecting to you.
6321 NAME: minimum_direct_hops
6324 LOC: Config.minDirectHops
6326 If using the ICMP pinging stuff, do direct fetches for sites
6327 which are no more than this many hops away.
6330 NAME: minimum_direct_rtt
6333 LOC: Config.minDirectRtt
6335 If using the ICMP pinging stuff, do direct fetches for sites
6336 which are no more than this many rtt milliseconds away.
6342 LOC: Config.Netdb.low
6348 LOC: Config.Netdb.high
6350 The low and high water marks for the ICMP measurement
6351 database. These are counts, not percents. The defaults are
6352 900 and 1000. When the high water mark is reached, database
6353 entries will be deleted until the low mark is reached.
6356 NAME: netdb_ping_period
6358 LOC: Config.Netdb.period
6361 The minimum period for measuring a site. There will be at
6362 least this much delay between successive pings to the same
6363 network. The default is five minutes.
6370 LOC: Config.onoff.query_icmp
6372 If you want to ask your peers to include ICMP data in their ICP
6373 replies, enable this option.
6375 If your peer has configured Squid (during compilation) with
6376 '--enable-icmp' that peer will send ICMP pings to origin server
6377 sites of the URLs it receives. If you enable this option the
6378 ICP replies from that peer will include the ICMP data (if available).
6379 Then, when choosing a parent cache, Squid will choose the parent with
6380 the minimal RTT to the origin server. When this happens, the
6381 hierarchy field of the access.log will be
6382 "CLOSEST_PARENT_MISS". This option is off by default.
6385 NAME: test_reachability
6389 LOC: Config.onoff.test_reachability
6391 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
6392 instead of ICP_MISS if the target host is NOT in the ICMP
6393 database, or has a zero RTT.
6396 NAME: icp_query_timeout
6400 LOC: Config.Timeout.icp_query
6402 Normally Squid will automatically determine an optimal ICP
6403 query timeout value based on the round-trip-time of recent ICP
6404 queries. If you want to override the value determined by
6405 Squid, set this 'icp_query_timeout' to a non-zero value. This
6406 value is specified in MILLISECONDS, so, to use a 2-second
6407 timeout (the old default), you would write:
6409 icp_query_timeout 2000
6412 NAME: maximum_icp_query_timeout
6416 LOC: Config.Timeout.icp_query_max
6418 Normally the ICP query timeout is determined dynamically. But
6419 sometimes it can lead to very large values (say 5 seconds).
6420 Use this option to put an upper limit on the dynamic timeout
6421 value. Do NOT use this option to always use a fixed (instead
6422 of a dynamic) timeout value. To set a fixed timeout see the
6423 'icp_query_timeout' directive.
6426 NAME: minimum_icp_query_timeout
6430 LOC: Config.Timeout.icp_query_min
6432 Normally the ICP query timeout is determined dynamically. But
6433 sometimes it can lead to very small timeouts, even lower than
6434 the normal latency variance on your link due to traffic.
6435 Use this option to put an lower limit on the dynamic timeout
6436 value. Do NOT use this option to always use a fixed (instead
6437 of a dynamic) timeout value. To set a fixed timeout see the
6438 'icp_query_timeout' directive.
6441 NAME: background_ping_rate
6445 LOC: Config.backgroundPingRate
6447 Controls how often the ICP pings are sent to siblings that
6448 have background-ping set.
6452 MULTICAST ICP OPTIONS
6453 -----------------------------------------------------------------------------
6458 LOC: Config.mcast_group_list
6461 This tag specifies a list of multicast groups which your server
6462 should join to receive multicasted ICP queries.
6464 NOTE! Be very careful what you put here! Be sure you
6465 understand the difference between an ICP _query_ and an ICP
6466 _reply_. This option is to be set only if you want to RECEIVE
6467 multicast queries. Do NOT set this option to SEND multicast
6468 ICP (use cache_peer for that). ICP replies are always sent via
6469 unicast, so this option does not affect whether or not you will
6470 receive replies from multicast group members.
6472 You must be very careful to NOT use a multicast address which
6473 is already in use by another group of caches.
6475 If you are unsure about multicast, please read the Multicast
6476 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
6478 Usage: mcast_groups 239.128.16.128 224.0.1.20
6480 By default, Squid doesn't listen on any multicast groups.
6483 NAME: mcast_miss_addr
6484 IFDEF: MULTICAST_MISS_STREAM
6486 LOC: Config.mcast_miss.addr
6489 If you enable this option, every "cache miss" URL will
6490 be sent out on the specified multicast address.
6492 Do not enable this option unless you are are absolutely
6493 certain you understand what you are doing.
6496 NAME: mcast_miss_ttl
6497 IFDEF: MULTICAST_MISS_STREAM
6499 LOC: Config.mcast_miss.ttl
6502 This is the time-to-live value for packets multicasted
6503 when multicasting off cache miss URLs is enabled. By
6504 default this is set to 'site scope', i.e. 16.
6507 NAME: mcast_miss_port
6508 IFDEF: MULTICAST_MISS_STREAM
6510 LOC: Config.mcast_miss.port
6513 This is the port number to be used in conjunction with
6517 NAME: mcast_miss_encode_key
6518 IFDEF: MULTICAST_MISS_STREAM
6520 LOC: Config.mcast_miss.encode_key
6521 DEFAULT: XXXXXXXXXXXXXXXX
6523 The URLs that are sent in the multicast miss stream are
6524 encrypted. This is the encryption key.
6527 NAME: mcast_icp_query_timeout
6531 LOC: Config.Timeout.mcast_icp_query
6533 For multicast peers, Squid regularly sends out ICP "probes" to
6534 count how many other peers are listening on the given multicast
6535 address. This value specifies how long Squid should wait to
6536 count all the replies. The default is 2000 msec, or 2
6541 INTERNAL ICON OPTIONS
6542 -----------------------------------------------------------------------------
6545 NAME: icon_directory
6547 LOC: Config.icons.directory
6548 DEFAULT: @DEFAULT_ICON_DIR@
6550 Where the icons are stored. These are normally kept in
6554 NAME: global_internal_static
6556 LOC: Config.onoff.global_internal_static
6559 This directive controls is Squid should intercept all requests for
6560 /squid-internal-static/ no matter which host the URL is requesting
6561 (default on setting), or if nothing special should be done for
6562 such URLs (off setting). The purpose of this directive is to make
6563 icons etc work better in complex cache hierarchies where it may
6564 not always be possible for all corners in the cache mesh to reach
6565 the server generating a directory listing.
6568 NAME: short_icon_urls
6570 LOC: Config.icons.use_short_names
6573 If this is enabled Squid will use short URLs for icons.
6574 If disabled it will revert to the old behavior of including
6575 it's own name and port in the URL.
6577 If you run a complex cache hierarchy with a mix of Squid and
6578 other proxies you may need to disable this directive.
6583 -----------------------------------------------------------------------------
6586 NAME: error_directory
6588 LOC: Config.errorDirectory
6591 If you wish to create your own versions of the default
6592 error files to customize them to suit your company copy
6593 the error/template files to another directory and point
6596 WARNING: This option will disable multi-language support
6597 on error pages if used.
6599 The squid developers are interested in making squid available in
6600 a wide variety of languages. If you are making translations for a
6601 language that Squid does not currently provide please consider
6602 contributing your translation back to the project.
6603 http://wiki.squid-cache.org/Translations
6605 The squid developers working on translations are happy to supply drop-in
6606 translated error files in exchange for any new language contributions.
6609 NAME: error_default_language
6610 IFDEF: USE_ERR_LOCALES
6612 LOC: Config.errorDefaultLanguage
6615 Set the default language which squid will send error pages in
6616 if no existing translation matches the clients language
6619 If unset (default) generic English will be used.
6621 The squid developers are interested in making squid available in
6622 a wide variety of languages. If you are interested in making
6623 translations for any language see the squid wiki for details.
6624 http://wiki.squid-cache.org/Translations
6627 NAME: error_log_languages
6628 IFDEF: USE_ERR_LOCALES
6630 LOC: Config.errorLogMissingLanguages
6633 Log to cache.log what languages users are attempting to
6634 auto-negotiate for translations.
6636 Successful negotiations are not logged. Only failures
6637 have meaning to indicate that Squid may need an upgrade
6638 of its error page translations.
6641 NAME: err_page_stylesheet
6643 LOC: Config.errorStylesheet
6644 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
6646 CSS Stylesheet to pattern the display of Squid default error pages.
6648 For information on CSS see http://www.w3.org/Style/CSS/
6653 LOC: Config.errHtmlText
6656 HTML text to include in error messages. Make this a "mailto"
6657 URL to your admin address, or maybe just a link to your
6658 organizations Web page.
6660 To include this in your error messages, you must rewrite
6661 the error template files (found in the "errors" directory).
6662 Wherever you want the 'err_html_text' line to appear,
6663 insert a %L tag in the error template file.
6666 NAME: email_err_data
6669 LOC: Config.onoff.emailErrData
6672 If enabled, information about the occurred error will be
6673 included in the mailto links of the ERR pages (if %W is set)
6674 so that the email body contains the data.
6675 Syntax is <A HREF="mailto:%w%W">%w</A>
6680 LOC: Config.denyInfoList
6683 Usage: deny_info err_page_name acl
6684 or deny_info http://... acl
6685 or deny_info TCP_RESET acl
6687 This can be used to return a ERR_ page for requests which
6688 do not pass the 'http_access' rules. Squid remembers the last
6689 acl it evaluated in http_access, and if a 'deny_info' line exists
6690 for that ACL Squid returns a corresponding error page.
6692 The acl is typically the last acl on the http_access deny line which
6693 denied access. The exceptions to this rule are:
6694 - When Squid needs to request authentication credentials. It's then
6695 the first authentication related acl encountered
6696 - When none of the http_access lines matches. It's then the last
6697 acl processed on the last http_access line.
6698 - When the decision to deny access was made by an adaptation service,
6699 the acl name is the corresponding eCAP or ICAP service_name.
6701 NP: If providing your own custom error pages with error_directory
6702 you may also specify them by your custom file name:
6703 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
6705 By defaut Squid will send "403 Forbidden". A different 4xx or 5xx
6706 may be specified by prefixing the file name with the code and a colon.
6707 e.g. 404:ERR_CUSTOM_ACCESS_DENIED
6709 Alternatively you can tell Squid to reset the TCP connection
6710 by specifying TCP_RESET.
6712 Or you can specify an error URL or URL pattern. The browsers will
6713 get redirected to the specified URL after formatting tags have
6714 been replaced. Redirect will be done with 302 or 307 according to
6715 HTTP/1.1 specs. A different 3xx code may be specified by prefixing
6716 the URL. e.g. 303:http://example.com/
6719 %a - username (if available. Password NOT included)
6722 %E - Error description
6724 %H - Request domain name
6725 %i - Client IP Address
6727 %o - Message result from external ACL helper
6728 %p - Request Port number
6729 %P - Request Protocol name
6730 %R - Request URL path
6731 %T - Timestamp in RFC 1123 format
6732 %U - Full canonical URL from client
6733 (HTTPS URLs terminate with *)
6734 %u - Full canonical URL from client
6735 %w - Admin email from squid.conf
6737 %% - Literal percent (%) code
6742 OPTIONS INFLUENCING REQUEST FORWARDING
6743 -----------------------------------------------------------------------------
6746 NAME: nonhierarchical_direct
6748 LOC: Config.onoff.nonhierarchical_direct
6751 By default, Squid will send any non-hierarchical requests
6752 (matching hierarchy_stoplist or not cacheable request type) direct
6755 If you set this to off, Squid will prefer to send these
6756 requests to parents.
6758 Note that in most configurations, by turning this off you will only
6759 add latency to these request without any improvement in global hit
6762 If you are inside an firewall see never_direct instead of
6768 LOC: Config.onoff.prefer_direct
6771 Normally Squid tries to use parents for most requests. If you for some
6772 reason like it to first try going direct and only use a parent if
6773 going direct fails set this to on.
6775 By combining nonhierarchical_direct off and prefer_direct on you
6776 can set up Squid to use a parent as a backup path if going direct
6779 Note: If you want Squid to use parents for all requests see
6780 the never_direct directive. prefer_direct only modifies how Squid
6781 acts on cacheable requests.
6786 LOC: Config.accessList.AlwaysDirect
6789 Usage: always_direct allow|deny [!]aclname ...
6791 Here you can use ACL elements to specify requests which should
6792 ALWAYS be forwarded by Squid to the origin servers without using
6793 any peers. For example, to always directly forward requests for
6794 local servers ignoring any parents or siblings you may have use
6797 acl local-servers dstdomain my.domain.net
6798 always_direct allow local-servers
6800 To always forward FTP requests directly, use
6803 always_direct allow FTP
6805 NOTE: There is a similar, but opposite option named
6806 'never_direct'. You need to be aware that "always_direct deny
6807 foo" is NOT the same thing as "never_direct allow foo". You
6808 may need to use a deny rule to exclude a more-specific case of
6809 some other rule. Example:
6811 acl local-external dstdomain external.foo.net
6812 acl local-servers dstdomain .foo.net
6813 always_direct deny local-external
6814 always_direct allow local-servers
6816 NOTE: If your goal is to make the client forward the request
6817 directly to the origin server bypassing Squid then this needs
6818 to be done in the client configuration. Squid configuration
6819 can only tell Squid how Squid should fetch the object.
6821 NOTE: This directive is not related to caching. The replies
6822 is cached as usual even if you use always_direct. To not cache
6823 the replies see the 'cache' directive.
6825 This clause supports both fast and slow acl types.
6826 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6831 LOC: Config.accessList.NeverDirect
6834 Usage: never_direct allow|deny [!]aclname ...
6836 never_direct is the opposite of always_direct. Please read
6837 the description for always_direct if you have not already.
6839 With 'never_direct' you can use ACL elements to specify
6840 requests which should NEVER be forwarded directly to origin
6841 servers. For example, to force the use of a proxy for all
6842 requests, except those in your local domain use something like:
6844 acl local-servers dstdomain .foo.net
6845 never_direct deny local-servers
6846 never_direct allow all
6848 or if Squid is inside a firewall and there are local intranet
6849 servers inside the firewall use something like:
6851 acl local-intranet dstdomain .foo.net
6852 acl local-external dstdomain external.foo.net
6853 always_direct deny local-external
6854 always_direct allow local-intranet
6855 never_direct allow all
6857 This clause supports both fast and slow acl types.
6858 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
6862 ADVANCED NETWORKING OPTIONS
6863 -----------------------------------------------------------------------------
6866 NAME: incoming_udp_average incoming_icp_average
6869 LOC: Config.comm_incoming.udp.average
6871 Heavy voodoo here. I can't even believe you are reading this.
6872 Are you crazy? Don't even think about adjusting these unless
6873 you understand the algorithms in comm_select.c first!
6876 NAME: incoming_tcp_average incoming_http_average
6879 LOC: Config.comm_incoming.tcp.average
6881 Heavy voodoo here. I can't even believe you are reading this.
6882 Are you crazy? Don't even think about adjusting these unless
6883 you understand the algorithms in comm_select.c first!
6886 NAME: incoming_dns_average
6889 LOC: Config.comm_incoming.dns.average
6891 Heavy voodoo here. I can't even believe you are reading this.
6892 Are you crazy? Don't even think about adjusting these unless
6893 you understand the algorithms in comm_select.c first!
6896 NAME: min_udp_poll_cnt min_icp_poll_cnt
6899 LOC: Config.comm_incoming.udp.min_poll
6901 Heavy voodoo here. I can't even believe you are reading this.
6902 Are you crazy? Don't even think about adjusting these unless
6903 you understand the algorithms in comm_select.c first!
6906 NAME: min_dns_poll_cnt
6909 LOC: Config.comm_incoming.dns.min_poll
6911 Heavy voodoo here. I can't even believe you are reading this.
6912 Are you crazy? Don't even think about adjusting these unless
6913 you understand the algorithms in comm_select.c first!
6916 NAME: min_tcp_poll_cnt min_http_poll_cnt
6919 LOC: Config.comm_incoming.tcp.min_poll
6921 Heavy voodoo here. I can't even believe you are reading this.
6922 Are you crazy? Don't even think about adjusting these unless
6923 you understand the algorithms in comm_select.c first!
6929 LOC: Config.accept_filter
6933 The name of an accept(2) filter to install on Squid's
6934 listen socket(s). This feature is perhaps specific to
6935 FreeBSD and requires support in the kernel.
6937 The 'httpready' filter delays delivering new connections
6938 to Squid until a full HTTP request has been received.
6939 See the accf_http(9) man page for details.
6941 The 'dataready' filter delays delivering new connections
6942 to Squid until there is some data to process.
6943 See the accf_dataready(9) man page for details.
6947 The 'data' filter delays delivering of new connections
6948 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
6949 You may optionally specify a number of seconds to wait by
6950 'data=N' where N is the number of seconds. Defaults to 30
6951 if not specified. See the tcp(7) man page for details.
6954 accept_filter httpready
6959 NAME: client_ip_max_connections
6961 LOC: Config.client_ip_max_connections
6964 Set an absolute limit on the number of connections a single
6965 client IP can use. Any more than this and Squid will begin to drop
6966 new connections from the client until it closes some links.
6968 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
6969 connections from the client. For finer control use the ACL access controls.
6971 Requires client_db to be enabled (the default).
6973 WARNING: This may noticably slow down traffic received via external proxies
6974 or NAT devices and cause them to rebound error messages back to their clients.
6977 NAME: tcp_recv_bufsize
6981 LOC: Config.tcpRcvBufsz
6983 Size of receive buffer to set for TCP sockets. Probably just
6984 as easy to change your kernel's default. Set to zero to use
6985 the default buffer size.
6990 -----------------------------------------------------------------------------
6997 LOC: Adaptation::Icap::TheConfig.onoff
7000 If you want to enable the ICAP module support, set this to on.
7003 NAME: icap_connect_timeout
7006 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
7009 This parameter specifies how long to wait for the TCP connect to
7010 the requested ICAP server to complete before giving up and either
7011 terminating the HTTP transaction or bypassing the failure.
7013 The default for optional services is peer_connect_timeout.
7014 The default for essential services is connect_timeout.
7015 If this option is explicitly set, its value applies to all services.
7018 NAME: icap_io_timeout
7022 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
7025 This parameter specifies how long to wait for an I/O activity on
7026 an established, active ICAP connection before giving up and
7027 either terminating the HTTP transaction or bypassing the
7030 The default is read_timeout.
7033 NAME: icap_service_failure_limit
7034 COMMENT: limit [in memory-depth time-units]
7035 TYPE: icap_service_failure_limit
7037 LOC: Adaptation::Icap::TheConfig
7040 The limit specifies the number of failures that Squid tolerates
7041 when establishing a new TCP connection with an ICAP service. If
7042 the number of failures exceeds the limit, the ICAP service is
7043 not used for new ICAP requests until it is time to refresh its
7046 A negative value disables the limit. Without the limit, an ICAP
7047 service will not be considered down due to connectivity failures
7048 between ICAP OPTIONS requests.
7050 Squid forgets ICAP service failures older than the specified
7051 value of memory-depth. The memory fading algorithm
7052 is approximate because Squid does not remember individual
7053 errors but groups them instead, splitting the option
7054 value into ten time slots of equal length.
7056 When memory-depth is 0 and by default this option has no
7057 effect on service failure expiration.
7059 Squid always forgets failures when updating service settings
7060 using an ICAP OPTIONS transaction, regardless of this option
7064 # suspend service usage after 10 failures in 5 seconds:
7065 icap_service_failure_limit 10 in 5 seconds
7068 NAME: icap_service_revival_delay
7071 LOC: Adaptation::Icap::TheConfig.service_revival_delay
7074 The delay specifies the number of seconds to wait after an ICAP
7075 OPTIONS request failure before requesting the options again. The
7076 failed ICAP service is considered "down" until fresh OPTIONS are
7079 The actual delay cannot be smaller than the hardcoded minimum
7080 delay of 30 seconds.
7083 NAME: icap_preview_enable
7087 LOC: Adaptation::Icap::TheConfig.preview_enable
7090 The ICAP Preview feature allows the ICAP server to handle the
7091 HTTP message by looking only at the beginning of the message body
7092 or even without receiving the body at all. In some environments,
7093 previews greatly speedup ICAP processing.
7095 During an ICAP OPTIONS transaction, the server may tell Squid what
7096 HTTP messages should be previewed and how big the preview should be.
7097 Squid will not use Preview if the server did not request one.
7099 To disable ICAP Preview for all ICAP services, regardless of
7100 individual ICAP server OPTIONS responses, set this option to "off".
7102 icap_preview_enable off
7105 NAME: icap_preview_size
7108 LOC: Adaptation::Icap::TheConfig.preview_size
7111 The default size of preview data to be sent to the ICAP server.
7112 -1 means no preview. This value might be overwritten on a per server
7113 basis by OPTIONS requests.
7116 NAME: icap_206_enable
7120 LOC: Adaptation::Icap::TheConfig.allow206_enable
7123 206 (Partial Content) responses is an ICAP extension that allows the
7124 ICAP agents to optionally combine adapted and original HTTP message
7125 content. The decision to combine is postponed until the end of the
7126 ICAP response. Squid supports Partial Content extension by default.
7128 Activation of the Partial Content extension is negotiated with each
7129 ICAP service during OPTIONS exchange. Most ICAP servers should handle
7130 negotation correctly even if they do not support the extension, but
7131 some might fail. To disable Partial Content support for all ICAP
7132 services and to avoid any negotiation, set this option to "off".
7138 NAME: icap_default_options_ttl
7141 LOC: Adaptation::Icap::TheConfig.default_options_ttl
7144 The default TTL value for ICAP OPTIONS responses that don't have
7145 an Options-TTL header.
7148 NAME: icap_persistent_connections
7152 LOC: Adaptation::Icap::TheConfig.reuse_connections
7155 Whether or not Squid should use persistent connections to
7159 NAME: adaptation_send_client_ip icap_send_client_ip
7161 IFDEF: USE_ADAPTATION
7163 LOC: Adaptation::Config::send_client_ip
7166 If enabled, Squid shares HTTP client IP information with adaptation
7167 services. For ICAP, Squid adds the X-Client-IP header to ICAP requests.
7168 For eCAP, Squid sets the libecap::metaClientIp transaction option.
7170 See also: adaptation_uses_indirect_client
7173 NAME: adaptation_send_username icap_send_client_username
7175 IFDEF: USE_ADAPTATION
7177 LOC: Adaptation::Config::send_username
7180 This sends authenticated HTTP client username (if available) to
7181 the adaptation service.
7183 For ICAP, the username value is encoded based on the
7184 icap_client_username_encode option and is sent using the header
7185 specified by the icap_client_username_header option.
7188 NAME: icap_client_username_header
7191 LOC: Adaptation::Icap::TheConfig.client_username_header
7192 DEFAULT: X-Client-Username
7194 ICAP request header name to use for adaptation_send_username.
7197 NAME: icap_client_username_encode
7201 LOC: Adaptation::Icap::TheConfig.client_username_encode
7204 Whether to base64 encode the authenticated client username.
7208 TYPE: icap_service_type
7210 LOC: Adaptation::Icap::TheConfig
7213 Defines a single ICAP service using the following format:
7215 icap_service id vectoring_point uri [option ...]
7218 an opaque identifier or name which is used to direct traffic to
7219 this specific service. Must be unique among all adaptation
7220 services in squid.conf.
7222 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7223 This specifies at which point of transaction processing the
7224 ICAP service should be activated. *_postcache vectoring points
7225 are not yet supported.
7227 uri: icap://servername:port/servicepath
7228 ICAP server and service location.
7230 ICAP does not allow a single service to handle both REQMOD and RESPMOD
7231 transactions. Squid does not enforce that requirement. You can specify
7232 services with the same service_url and different vectoring_points. You
7233 can even specify multiple identical services as long as their
7234 service_names differ.
7237 Service options are separated by white space. ICAP services support
7238 the following name=value options:
7241 If set to 'on' or '1', the ICAP service is treated as
7242 optional. If the service cannot be reached or malfunctions,
7243 Squid will try to ignore any errors and process the message as
7244 if the service was not enabled. No all ICAP errors can be
7245 bypassed. If set to 0, the ICAP service is treated as
7246 essential and all ICAP errors will result in an error page
7247 returned to the HTTP client.
7249 Bypass is off by default: services are treated as essential.
7252 If set to 'on' or '1', the ICAP service is allowed to
7253 dynamically change the current message adaptation plan by
7254 returning a chain of services to be used next. The services
7255 are specified using the X-Next-Services ICAP response header
7256 value, formatted as a comma-separated list of service names.
7257 Each named service should be configured in squid.conf. Other
7258 services are ignored. An empty X-Next-Services value results
7259 in an empty plan which ends the current adaptation.
7261 Dynamic adaptation plan may cross or cover multiple supported
7262 vectoring points in their natural processing order.
7264 Routing is not allowed by default: the ICAP X-Next-Services
7265 response header is ignored.
7268 Only has effect on split-stack systems. The default on those systems
7269 is to use IPv4-only connections. When set to 'on' this option will
7270 make Squid use IPv6-only connections to contact this ICAP service.
7272 on-overload=block|bypass|wait|force
7273 If the service Max-Connections limit has been reached, do
7274 one of the following for each new ICAP transaction:
7275 * block: send an HTTP error response to the client
7276 * bypass: ignore the "over-connected" ICAP service
7277 * wait: wait (in a FIFO queue) for an ICAP connection slot
7278 * force: proceed, ignoring the Max-Connections limit
7280 In SMP mode with N workers, each worker assumes the service
7281 connection limit is Max-Connections/N, even though not all
7282 workers may use a given service.
7284 The default value is "bypass" if service is bypassable,
7285 otherwise it is set to "wait".
7289 Use the given number as the Max-Connections limit, regardless
7290 of the Max-Connections value given by the service, if any.
7292 Older icap_service format without optional named parameters is
7293 deprecated but supported for backward compatibility.
7296 icap_service svcBlocker reqmod_precache icap://icap1.mydomain.net:1344/reqmod bypass=0
7297 icap_service svcLogger reqmod_precache icap://icap2.mydomain.net:1344/respmod routing=on
7301 TYPE: icap_class_type
7306 This deprecated option was documented to define an ICAP service
7307 chain, even though it actually defined a set of similar, redundant
7308 services, and the chains were not supported.
7310 To define a set of redundant services, please use the
7311 adaptation_service_set directive. For service chains, use
7312 adaptation_service_chain.
7316 TYPE: icap_access_type
7321 This option is deprecated. Please use adaptation_access, which
7322 has the same ICAP functionality, but comes with better
7323 documentation, and eCAP support.
7328 -----------------------------------------------------------------------------
7335 LOC: Adaptation::Ecap::TheConfig.onoff
7338 Controls whether eCAP support is enabled.
7342 TYPE: ecap_service_type
7344 LOC: Adaptation::Ecap::TheConfig
7347 Defines a single eCAP service
7349 ecap_service id vectoring_point uri [option ...]
7352 an opaque identifier or name which is used to direct traffic to
7353 this specific service. Must be unique among all adaptation
7354 services in squid.conf.
7356 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
7357 This specifies at which point of transaction processing the
7358 eCAP service should be activated. *_postcache vectoring points
7359 are not yet supported.
7361 uri: ecap://vendor/service_name?custom&cgi=style¶meters=optional
7362 Squid uses the eCAP service URI to match this configuration
7363 line with one of the dynamically loaded services. Each loaded
7364 eCAP service must have a unique URI. Obtain the right URI from
7365 the service provider.
7368 Service options are separated by white space. eCAP services support
7369 the following name=value options:
7372 If set to 'on' or '1', the eCAP service is treated as optional.
7373 If the service cannot be reached or malfunctions, Squid will try
7374 to ignore any errors and process the message as if the service
7375 was not enabled. No all eCAP errors can be bypassed.
7376 If set to 'off' or '0', the eCAP service is treated as essential
7377 and all eCAP errors will result in an error page returned to the
7380 Bypass is off by default: services are treated as essential.
7383 If set to 'on' or '1', the eCAP service is allowed to
7384 dynamically change the current message adaptation plan by
7385 returning a chain of services to be used next.
7387 Dynamic adaptation plan may cross or cover multiple supported
7388 vectoring points in their natural processing order.
7390 Routing is not allowed by default.
7392 Older ecap_service format without optional named parameters is
7393 deprecated but supported for backward compatibility.
7397 ecap_service s1 reqmod_precache ecap://filters.R.us/leakDetector?on_error=block bypass=off
7398 ecap_service s2 respmod_precache ecap://filters.R.us/virusFilter config=/etc/vf.cfg bypass=on
7401 NAME: loadable_modules
7403 IFDEF: USE_LOADABLE_MODULES
7404 LOC: Config.loadable_module_names
7407 Instructs Squid to load the specified dynamic module(s) or activate
7408 preloaded module(s).
7410 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
7414 MESSAGE ADAPTATION OPTIONS
7415 -----------------------------------------------------------------------------
7418 NAME: adaptation_service_set
7419 TYPE: adaptation_service_set_type
7420 IFDEF: USE_ADAPTATION
7425 Configures an ordered set of similar, redundant services. This is
7426 useful when hot standby or backup adaptation servers are available.
7428 adaptation_service_set set_name service_name1 service_name2 ...
7430 The named services are used in the set declaration order. The first
7431 applicable adaptation service from the set is used first. The next
7432 applicable service is tried if and only if the transaction with the
7433 previous service fails and the message waiting to be adapted is still
7436 When adaptation starts, broken services are ignored as if they were
7437 not a part of the set. A broken service is a down optional service.
7439 The services in a set must be attached to the same vectoring point
7440 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7442 If all services in a set are optional then adaptation failures are
7443 bypassable. If all services in the set are essential, then a
7444 transaction failure with one service may still be retried using
7445 another service from the set, but when all services fail, the master
7446 transaction fails as well.
7448 A set may contain a mix of optional and essential services, but that
7449 is likely to lead to surprising results because broken services become
7450 ignored (see above), making previously bypassable failures fatal.
7451 Technically, it is the bypassability of the last failed service that
7454 See also: adaptation_access adaptation_service_chain
7457 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
7458 adaptation service_set svcLogger loggerLocal loggerRemote
7461 NAME: adaptation_service_chain
7462 TYPE: adaptation_service_chain_type
7463 IFDEF: USE_ADAPTATION
7468 Configures a list of complementary services that will be applied
7469 one-by-one, forming an adaptation chain or pipeline. This is useful
7470 when Squid must perform different adaptations on the same message.
7472 adaptation_service_chain chain_name service_name1 svc_name2 ...
7474 The named services are used in the chain declaration order. The first
7475 applicable adaptation service from the chain is used first. The next
7476 applicable service is applied to the successful adaptation results of
7477 the previous service in the chain.
7479 When adaptation starts, broken services are ignored as if they were
7480 not a part of the chain. A broken service is a down optional service.
7482 Request satisfaction terminates the adaptation chain because Squid
7483 does not currently allow declaration of RESPMOD services at the
7484 "reqmod_precache" vectoring point (see icap_service or ecap_service).
7486 The services in a chain must be attached to the same vectoring point
7487 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
7489 A chain may contain a mix of optional and essential services. If an
7490 essential adaptation fails (or the failure cannot be bypassed for
7491 other reasons), the master transaction fails. Otherwise, the failure
7492 is bypassed as if the failed adaptation service was not in the chain.
7494 See also: adaptation_access adaptation_service_set
7497 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
7500 NAME: adaptation_access
7501 TYPE: adaptation_access_type
7502 IFDEF: USE_ADAPTATION
7506 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
7508 adaptation_access service_name allow|deny [!]aclname...
7509 adaptation_access set_name allow|deny [!]aclname...
7511 At each supported vectoring point, the adaptation_access
7512 statements are processed in the order they appear in this
7513 configuration file. Statements pointing to the following services
7514 are ignored (i.e., skipped without checking their ACL):
7516 - services serving different vectoring points
7517 - "broken-but-bypassable" services
7518 - "up" services configured to ignore such transactions
7519 (e.g., based on the ICAP Transfer-Ignore header).
7521 When a set_name is used, all services in the set are checked
7522 using the same rules, to find the first applicable one. See
7523 adaptation_service_set for details.
7525 If an access list is checked and there is a match, the
7526 processing stops: For an "allow" rule, the corresponding
7527 adaptation service is used for the transaction. For a "deny"
7528 rule, no adaptation service is activated.
7530 It is currently not possible to apply more than one adaptation
7531 service at the same vectoring point to the same HTTP transaction.
7533 See also: icap_service and ecap_service
7536 adaptation_access service_1 allow all
7539 NAME: adaptation_service_iteration_limit
7541 IFDEF: USE_ADAPTATION
7542 LOC: Adaptation::Config::service_iteration_limit
7545 Limits the number of iterations allowed when applying adaptation
7546 services to a message. If your longest adaptation set or chain
7547 may have more than 16 services, increase the limit beyond its
7548 default value of 16. If detecting infinite iteration loops sooner
7549 is critical, make the iteration limit match the actual number
7550 of services in your longest adaptation set or chain.
7552 Infinite adaptation loops are most likely with routing services.
7554 See also: icap_service routing=1
7557 NAME: adaptation_masterx_shared_names
7559 IFDEF: USE_ADAPTATION
7560 LOC: Adaptation::Config::masterx_shared_name
7563 For each master transaction (i.e., the HTTP request and response
7564 sequence, including all related ICAP and eCAP exchanges), Squid
7565 maintains a table of metadata. The table entries are (name, value)
7566 pairs shared among eCAP and ICAP exchanges. The table is destroyed
7567 with the master transaction.
7569 This option specifies the table entry names that Squid must accept
7570 from and forward to the adaptation transactions.
7572 An ICAP REQMOD or RESPMOD transaction may set an entry in the
7573 shared table by returning an ICAP header field with a name
7574 specified in adaptation_masterx_shared_names.
7576 An eCAP REQMOD or RESPMOD transaction may set an entry in the
7577 shared table by implementing the libecap::visitEachOption() API
7578 to provide an option with a name specified in
7579 adaptation_masterx_shared_names.
7581 Squid will store and forward the set entry to subsequent adaptation
7582 transactions within the same master transaction scope.
7584 Only one shared entry name is supported at this time.
7587 # share authentication information among ICAP services
7588 adaptation_masterx_shared_names X-Subscriber-ID
7591 NAME: adaptation_meta
7593 IFDEF: USE_ADAPTATION
7594 LOC: Adaptation::Config::metaHeaders
7597 This option allows Squid administrator to add custom ICAP request
7598 headers or eCAP options to Squid ICAP requests or eCAP transactions.
7599 Use it to pass custom authentication tokens and other
7600 transaction-state related meta information to an ICAP/eCAP service.
7602 The addition of a meta header is ACL-driven:
7603 adaptation_meta name value [!]aclname ...
7605 Processing for a given header name stops after the first ACL list match.
7606 Thus, it is impossible to add two headers with the same name. If no ACL
7607 lists match for a given header name, no such header is added. For
7610 # do not debug transactions except for those that need debugging
7611 adaptation_meta X-Debug 1 needs_debugging
7613 # log all transactions except for those that must remain secret
7614 adaptation_meta X-Log 1 !keep_secret
7616 # mark transactions from users in the "G 1" group
7617 adaptation_meta X-Authenticated-Groups "G 1" authed_as_G1
7619 The "value" parameter may be a regular squid.conf token or a "double
7620 quoted string". Within the quoted string, use backslash (\) to escape
7621 any character, which is currently only useful for escaping backslashes
7622 and double quotes. For example,
7623 "this string has one backslash (\\) and two \"quotes\""
7625 Used adaptation_meta header values may be logged via %note
7626 logformat code. If multiple adaptation_meta headers with the same name
7627 are used during master transaction lifetime, the header values are
7628 logged in the order they were used and duplicate values are ignored
7629 (only the first repeated value will be logged).
7635 LOC: Adaptation::Icap::TheConfig.repeat
7636 DEFAULT_IF_NONE: deny all
7638 This ACL determines which retriable ICAP transactions are
7639 retried. Transactions that received a complete ICAP response
7640 and did not have to consume or produce HTTP bodies to receive
7641 that response are usually retriable.
7643 icap_retry allow|deny [!]aclname ...
7645 Squid automatically retries some ICAP I/O timeouts and errors
7646 due to persistent connection race conditions.
7648 See also: icap_retry_limit
7651 NAME: icap_retry_limit
7654 LOC: Adaptation::Icap::TheConfig.repeat_limit
7657 Limits the number of retries allowed. When set to zero (default),
7658 no retries are allowed.
7660 Communication errors due to persistent connection race
7661 conditions are unavoidable, automatically retried, and do not
7662 count against this limit.
7664 See also: icap_retry
7670 -----------------------------------------------------------------------------
7673 NAME: check_hostnames
7676 LOC: Config.onoff.check_hostnames
7678 For security and stability reasons Squid can check
7679 hostnames for Internet standard RFC compliance. If you want
7680 Squid to perform these checks turn this directive on.
7683 NAME: allow_underscore
7686 LOC: Config.onoff.allow_underscore
7688 Underscore characters is not strictly allowed in Internet hostnames
7689 but nevertheless used by many sites. Set this to off if you want
7690 Squid to be strict about the standard.
7691 This check is performed only when check_hostnames is set to on.
7694 NAME: cache_dns_program
7696 IFDEF: USE_DNSHELPER
7697 DEFAULT: @DEFAULT_DNSSERVER@
7698 LOC: Config.Program.dnsserver
7700 Specify the location of the executable for dnslookup process.
7704 TYPE: HelperChildConfig
7705 IFDEF: USE_DNSHELPER
7706 DEFAULT: 32 startup=1 idle=1
7707 LOC: Config.dnsChildren
7709 The maximum number of processes spawn to service DNS name lookups.
7710 If you limit it too few Squid will have to wait for them to process
7711 a backlog of requests, slowing it down. If you allow too many they
7712 will use RAM and other system resources noticably.
7713 The maximum this may be safely set to is 32.
7715 The startup= and idle= options allow some measure of skew in your
7720 Sets a minimum of how many processes are to be spawned when Squid
7721 starts or reconfigures. When set to zero the first request will
7722 cause spawning of the first child process to handle it.
7724 Starting too few will cause an initial slowdown in traffic as Squid
7725 attempts to simultaneously spawn enough processes to cope.
7729 Sets a minimum of how many processes Squid is to try and keep available
7730 at all times. When traffic begins to rise above what the existing
7731 processes can handle this many more will be spawned up to the maximum
7732 configured. A minimum setting of 1 is required.
7735 NAME: dns_retransmit_interval
7738 LOC: Config.Timeout.idns_retransmit
7739 IFDEF: !USE_DNSHELPER
7741 Initial retransmit interval for DNS queries. The interval is
7742 doubled each time all configured DNS servers have been tried.
7748 LOC: Config.Timeout.idns_query
7749 IFDEF: !USE_DNSHELPER
7751 DNS Query timeout. If no response is received to a DNS query
7752 within this time all DNS servers for the queried domain
7753 are assumed to be unavailable.
7756 NAME: dns_packet_max
7759 LOC: Config.dns.packet_max
7760 IFDEF: !USE_DNSHELPER
7762 Maximum number of bytes packet size to advertise via EDNS.
7763 Set to "none" to disable EDNS large packet support.
7765 For legacy reasons DNS UDP replies will default to 512 bytes which
7766 is too small for many responses. EDNS provides a means for Squid to
7767 negotiate receiving larger responses back immediately without having
7768 to failover with repeat requests. Responses larger than this limit
7769 will retain the old behaviour of failover to TCP DNS.
7771 Squid has no real fixed limit internally, but allowing packet sizes
7772 over 1500 bytes requires network jumbogram support and is usually not
7775 WARNING: The RFC also indicates that some older resolvers will reply
7776 with failure of the whole request if the extension is added. Some
7777 resolvers have already been identified which will reply with mangled
7778 EDNS response on occasion. Usually in response to many-KB jumbogram
7779 sizes being advertised by Squid.
7780 Squid will currently treat these both as an unable-to-resolve domain
7781 even if it would be resolvable without EDNS.
7788 LOC: Config.onoff.res_defnames
7790 Normally the RES_DEFNAMES resolver option is disabled
7791 (see res_init(3)). This prevents caches in a hierarchy
7792 from interpreting single-component hostnames locally. To allow
7793 Squid to handle single-component names, enable this option.
7796 NAME: dns_nameservers
7799 LOC: Config.dns_nameservers
7801 Use this if you want to specify a list of DNS name servers
7802 (IP addresses) to use instead of those given in your
7803 /etc/resolv.conf file.
7804 On Windows platforms, if no value is specified here or in
7805 the /etc/resolv.conf file, the list of DNS name servers are
7806 taken from the Windows registry, both static and dynamic DHCP
7807 configurations are supported.
7809 Example: dns_nameservers 10.0.0.1 192.172.0.4
7814 DEFAULT: @DEFAULT_HOSTS@
7815 LOC: Config.etcHostsPath
7817 Location of the host-local IP name-address associations
7818 database. Most Operating Systems have such a file on different
7820 - Un*X & Linux: /etc/hosts
7821 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
7822 (%SystemRoot% value install default is c:\winnt)
7823 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
7824 (%SystemRoot% value install default is c:\windows)
7825 - Windows 9x/Me: %windir%\hosts
7826 (%windir% value is usually c:\windows)
7827 - Cygwin: /etc/hosts
7829 The file contains newline-separated definitions, in the
7830 form ip_address_in_dotted_form name [name ...] names are
7831 whitespace-separated. Lines beginning with an hash (#)
7832 character are comments.
7834 The file is checked at startup and upon configuration.
7835 If set to 'none', it won't be checked.
7836 If append_domain is used, that domain will be added to
7837 domain-local (i.e. not containing any dot character) host
7843 LOC: Config.appendDomain
7846 Appends local domain name to hostnames without any dots in
7847 them. append_domain must begin with a period.
7849 Be warned there are now Internet names with no dots in
7850 them using only top-domain names, so setting this may
7851 cause some Internet sites to become unavailable.
7854 append_domain .yourdomain.com
7857 NAME: ignore_unknown_nameservers
7859 LOC: Config.onoff.ignore_unknown_nameservers
7861 IFDEF: !USE_DNSHELPER
7863 By default Squid checks that DNS responses are received
7864 from the same IP addresses they are sent to. If they
7865 don't match, Squid ignores the response and writes a warning
7866 message to cache.log. You can allow responses from unknown
7867 nameservers by setting this option to 'off'.
7873 LOC: Config.dns.v4_first
7874 IFDEF: !USE_DNSHELPER
7876 With the IPv6 Internet being as fast or faster than IPv4 Internet
7877 for most networks Squid prefers to contact websites over IPv6.
7879 This option reverses the order of preference to make Squid contact
7880 dual-stack websites over IPv4 first. Squid will still perform both
7881 IPv6 and IPv4 DNS lookups before connecting.
7884 This option will restrict the situations under which IPv6
7885 connectivity is used (and tested). Hiding network problems
7886 which would otherwise be detected and warned about.
7890 COMMENT: (number of entries)
7893 LOC: Config.ipcache.size
7900 LOC: Config.ipcache.low
7907 LOC: Config.ipcache.high
7909 The size, low-, and high-water marks for the IP cache.
7912 NAME: fqdncache_size
7913 COMMENT: (number of entries)
7916 LOC: Config.fqdncache.size
7918 Maximum number of FQDN cache entries.
7923 -----------------------------------------------------------------------------
7930 LOC: Config.onoff.mem_pools
7932 If set, Squid will keep pools of allocated (but unused) memory
7933 available for future use. If memory is a premium on your
7934 system and you believe your malloc library outperforms Squid
7935 routines, disable this.
7938 NAME: memory_pools_limit
7942 LOC: Config.MemPools.limit
7944 Used only with memory_pools on:
7945 memory_pools_limit 50 MB
7947 If set to a non-zero value, Squid will keep at most the specified
7948 limit of allocated (but unused) memory in memory pools. All free()
7949 requests that exceed this limit will be handled by your malloc
7950 library. Squid does not pre-allocate any memory, just safe-keeps
7951 objects that otherwise would be free()d. Thus, it is safe to set
7952 memory_pools_limit to a reasonably high value even if your
7953 configuration will use less memory.
7955 If set to none, Squid will keep all memory it can. That is, there
7956 will be no limit on the total amount of memory used for safe-keeping.
7958 To disable memory allocation optimization, do not set
7959 memory_pools_limit to 0 or none. Set memory_pools to "off" instead.
7961 An overhead for maintaining memory pools is not taken into account
7962 when the limit is checked. This overhead is close to four bytes per
7963 object kept. However, pools may actually _save_ memory because of
7964 reduced memory thrashing in your malloc library.
7968 COMMENT: on|off|transparent|truncate|delete
7971 LOC: opt_forwarded_for
7973 If set to "on", Squid will append your client's IP address
7974 in the HTTP requests it forwards. By default it looks like:
7976 X-Forwarded-For: 192.1.2.3
7978 If set to "off", it will appear as
7980 X-Forwarded-For: unknown
7982 If set to "transparent", Squid will not alter the
7983 X-Forwarded-For header in any way.
7985 If set to "delete", Squid will delete the entire
7986 X-Forwarded-For header.
7988 If set to "truncate", Squid will remove all existing
7989 X-Forwarded-For entries, and place the client IP as the sole entry.
7992 NAME: cachemgr_passwd
7993 TYPE: cachemgrpasswd
7995 LOC: Config.passwd_list
7997 Specify passwords for cachemgr operations.
7999 Usage: cachemgr_passwd password action action ...
8001 Some valid actions are (see cache manager menu for a full list):
8041 * Indicates actions which will not be performed without a
8042 valid password, others can be performed if not listed here.
8044 To disable an action, set the password to "disable".
8045 To allow performing an action without a password, set the
8048 Use the keyword "all" to set the same password for all actions.
8051 cachemgr_passwd secret shutdown
8052 cachemgr_passwd lesssssssecret info stats/objects
8053 cachemgr_passwd disable all
8060 LOC: Config.onoff.client_db
8062 If you want to disable collecting per-client statistics,
8063 turn off client_db here.
8066 NAME: refresh_all_ims
8070 LOC: Config.onoff.refresh_all_ims
8072 When you enable this option, squid will always check
8073 the origin server for an update when a client sends an
8074 If-Modified-Since request. Many browsers use IMS
8075 requests when the user requests a reload, and this
8076 ensures those clients receive the latest version.
8078 By default (off), squid may return a Not Modified response
8079 based on the age of the cached version.
8082 NAME: reload_into_ims
8083 IFDEF: USE_HTTP_VIOLATIONS
8087 LOC: Config.onoff.reload_into_ims
8089 When you enable this option, client no-cache or ``reload''
8090 requests will be changed to If-Modified-Since requests.
8091 Doing this VIOLATES the HTTP standard. Enabling this
8092 feature could make you liable for problems which it
8095 see also refresh_pattern for a more selective approach.
8098 NAME: connect_retries
8100 LOC: Config.connect_retries
8103 This sets the maximum number of connection attempts made for each
8104 TCP connection. The connect_retries attempts must all still
8105 complete within the connection timeout period.
8107 The default is not to re-try if the first connection attempt fails.
8108 The (not recommended) maximum is 10 tries.
8110 A warning message will be generated if it is set to a too-high
8111 value and the configured value will be over-ridden.
8113 Note: These re-tries are in addition to forward_max_tries
8114 which limit how many different addresses may be tried to find
8118 NAME: retry_on_error
8120 LOC: Config.retry.onerror
8123 If set to ON Squid will automatically retry requests when
8124 receiving an error response with status 403 (Forbidden),
8125 500 (Internal Error), 501 or 503 (Service not available).
8126 Status 502 and 504 (Gateway errors) are always retried.
8128 This is mainly useful if you are in a complex cache hierarchy to
8129 work around access control errors.
8131 NOTE: This retry will attempt to find another working destination.
8132 Which is different from the server which just failed.
8135 NAME: as_whois_server
8137 LOC: Config.as_whois_server
8138 DEFAULT: whois.ra.net
8140 WHOIS server to query for AS numbers. NOTE: AS numbers are
8141 queried only when Squid starts up, not for every request.
8146 LOC: Config.onoff.offline
8149 Enable this option and Squid will never try to validate cached
8153 NAME: uri_whitespace
8154 TYPE: uri_whitespace
8155 LOC: Config.uri_whitespace
8158 What to do with requests that have whitespace characters in the
8161 strip: The whitespace characters are stripped out of the URL.
8162 This is the behavior recommended by RFC2396.
8163 deny: The request is denied. The user receives an "Invalid
8165 allow: The request is allowed and the URI is not changed. The
8166 whitespace characters remain in the URI. Note the
8167 whitespace is passed to redirector processes if they
8169 encode: The request is allowed and the whitespace characters are
8170 encoded according to RFC1738. This could be considered
8171 a violation of the HTTP/1.1
8172 RFC because proxies are not allowed to rewrite URI's.
8173 chop: The request is allowed and the URI is chopped at the
8174 first whitespace. This might also be considered a
8180 LOC: Config.chroot_dir
8183 Specifies a directory where Squid should do a chroot() while
8184 initializing. This also causes Squid to fully drop root
8185 privileges after initializing. This means, for example, if you
8186 use a HTTP port less than 1024 and try to reconfigure, you may
8187 get an error saying that Squid can not open the port.
8190 NAME: balance_on_multiple_ip
8192 LOC: Config.onoff.balance_on_multiple_ip
8195 Modern IP resolvers in squid sort lookup results by preferred access.
8196 By default squid will use these IP in order and only rotates to
8197 the next listed when the most preffered fails.
8199 Some load balancing servers based on round robin DNS have been
8200 found not to preserve user session state across requests
8201 to different IP addresses.
8203 Enabling this directive Squid rotates IP's per request.
8206 NAME: pipeline_prefetch
8208 LOC: Config.onoff.pipeline_prefetch
8211 To boost the performance of pipelined requests to closer
8212 match that of a non-proxied environment Squid can try to fetch
8213 up to two requests in parallel from a pipeline.
8215 Defaults to off for bandwidth management and access logging
8218 WARNING: pipelining breaks NTLM and Negotiate/Kerberos authentication.
8221 NAME: high_response_time_warning
8224 LOC: Config.warnings.high_rptm
8227 If the one-minute median response time exceeds this value,
8228 Squid prints a WARNING with debug level 0 to get the
8229 administrators attention. The value is in milliseconds.
8232 NAME: high_page_fault_warning
8234 LOC: Config.warnings.high_pf
8237 If the one-minute average page fault rate exceeds this
8238 value, Squid prints a WARNING with debug level 0 to get
8239 the administrators attention. The value is in page faults
8243 NAME: high_memory_warning
8245 LOC: Config.warnings.high_memory
8248 If the memory usage (as determined by mallinfo) exceeds
8249 this amount, Squid prints a WARNING with debug level 0 to get
8250 the administrators attention.
8253 NAME: sleep_after_fork
8254 COMMENT: (microseconds)
8256 LOC: Config.sleep_after_fork
8259 When this is set to a non-zero value, the main Squid process
8260 sleeps the specified number of microseconds after a fork()
8261 system call. This sleep may help the situation where your
8262 system reports fork() failures due to lack of (virtual)
8263 memory. Note, however, if you have a lot of child
8264 processes, these sleep delays will add up and your
8265 Squid will not service requests for some amount of time
8266 until all the child processes have been started.
8267 On Windows value less then 1000 (1 milliseconds) are
8271 NAME: windows_ipaddrchangemonitor
8272 IFDEF: _SQUID_WINDOWS_
8276 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
8278 On Windows Squid by default will monitor IP address changes and will
8279 reconfigure itself after any detected event. This is very useful for
8280 proxies connected to internet with dial-up interfaces.
8281 In some cases (a Proxy server acting as VPN gateway is one) it could be
8282 desiderable to disable this behaviour setting this to 'off'.
8283 Note: after changing this, Squid service must be restarted.
8288 IFDEF: USE_SQUID_EUI
8290 LOC: Eui::TheConfig.euiLookup
8292 Whether to lookup the EUI or MAC address of a connected client.
8295 NAME: max_filedescriptors max_filedesc
8298 LOC: Config.max_filedescriptors
8300 The maximum number of filedescriptors supported.
8302 The default "0" means Squid inherits the current ulimit setting.
8304 Note: Changing this requires a restart of Squid. Also
8305 not all comm loops supports large values.
8313 Number of main Squid processes or "workers" to fork and maintain.
8314 0: "no daemon" mode, like running "squid -N ..."
8315 1: "no SMP" mode, start one main Squid process daemon (default)
8316 N: start N main Squid process daemons (i.e., SMP mode)
8318 In SMP mode, each worker does nearly all what a single Squid daemon
8319 does (e.g., listen on http_port and forward HTTP requests).
8322 NAME: cpu_affinity_map
8323 TYPE: CpuAffinityMap
8324 LOC: Config.cpuAffinityMap
8327 Usage: cpu_affinity_map process_numbers=P1,P2,... cores=C1,C2,...
8329 Sets 1:1 mapping between Squid processes and CPU cores. For example,
8331 cpu_affinity_map process_numbers=1,2,3,4 cores=1,3,5,7
8333 affects processes 1 through 4 only and places them on the first
8334 four even cores, starting with core #1.
8336 CPU cores are numbered starting from 1. Requires support for
8337 sched_getaffinity(2) and sched_setaffinity(2) system calls.
8339 Multiple cpu_affinity_map options are merged.