3 # $Id: cf.data.pre,v 1.502 2008/02/12 00:05:11 amosjeffries Exp $
5 # SQUID Web Proxy Cache http://www.squid-cache.org/
6 # ----------------------------------------------------------
8 # Squid is the result of efforts by numerous individuals from
9 # the Internet community; see the CONTRIBUTORS file for full
10 # details. Many organizations have provided support for Squid's
11 # development; see the SPONSORS file for full details. Squid is
12 # Copyrighted (C) 2000 by the Regents of the University of
13 # California; see the COPYRIGHT file for full details. Squid
14 # incorporates software developed and/or copyrighted by other
15 # sources; see the CREDITS file for full details.
17 # This program is free software; you can redistribute it and/or modify
18 # it under the terms of the GNU General Public License as published by
19 # the Free Software Foundation; either version 2 of the License, or
20 # (at your option) any later version.
22 # This program is distributed in the hope that it will be useful,
23 # but WITHOUT ANY WARRANTY; without even the implied warranty of
24 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
25 # GNU General Public License for more details.
27 # You should have received a copy of the GNU General Public License
28 # along with this program; if not, write to the Free Software
29 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
33 WELCOME TO SQUID @VERSION@
34 ----------------------------
36 This is the default Squid configuration file. You may wish
37 to look at the Squid home page (http://www.squid-cache.org/)
38 for the FAQ and other documentation.
40 The default Squid config file shows what the defaults for
41 various options happen to be. If you don't need to change the
42 default, you shouldn't uncomment the line. Doing so may cause
43 run-time problems. In some cases "none" refers to no default
44 setting at all, while in other cases it refers to a valid
45 option - the comments for that keyword indicate if this is the
51 Configuration options can be included using the "include" directive.
52 Include takes a list of files to include. Quoting and wildcards is
57 include /path/to/included/file/squid.acl.config
59 Includes can be nested up to a hard-coded depth of 16 levels.
60 This arbitrary restriction is to prevent recursive include references
61 from causing Squid entering an infinite loop whilst trying to load
66 OPTIONS FOR AUTHENTICATION
67 -----------------------------------------------------------------------------
72 LOC: Config.authConfiguration
75 This is used to define parameters for the various authentication
76 schemes supported by Squid.
78 format: auth_param scheme parameter [setting]
80 The order in which authentication schemes are presented to the client is
81 dependent on the order the scheme first appears in config file. IE
82 has a bug (it's not RFC 2617 compliant) in that it will use the basic
83 scheme if basic is the first entry presented, even if more secure
84 schemes are presented. For now use the order in the recommended
85 settings section below. If other browsers have difficulties (don't
86 recognize the schemes offered even if you are using basic) either
87 put basic first, or disable the other schemes (by commenting out their
90 Once an authentication scheme is fully configured, it can only be
91 shutdown by shutting squid down and restarting. Changes can be made on
92 the fly and activated with a reconfigure. I.E. You can change to a
93 different helper, but not unconfigure the helper completely.
95 Please note that while this directive defines how Squid processes
96 authentication it does not automatically activate authentication.
97 To use authentication you must in addition make use of ACLs based
98 on login name in http_access (proxy_auth, proxy_auth_regex or
99 external with %LOGIN used in the format tag). The browser will be
100 challenged for authentication on the first such acl encountered
101 in http_access processing and will also be re-challenged for new
102 login credentials if the request is being denied by a proxy_auth
105 WARNING: authentication can't be used in a transparently intercepting
106 proxy as the client then thinks it is talking to an origin server and
107 not the proxy. This is a limitation of bending the TCP/IP protocol to
108 transparently intercepting port 80, not a limitation in Squid.
109 Ports flagged 'transparent', 'intercept', or 'tproxy' have authentication
112 === Parameters for the basic scheme follow. ===
115 Specify the command for the external authenticator. Such a program
116 reads a line containing "username password" and replies "OK" or
117 "ERR" in an endless loop. "ERR" responses may optionally be followed
118 by a error description available as %m in the returned error page.
119 If you use an authenticator, make sure you have 1 acl of type proxy_auth.
121 By default, the basic authentication scheme is not used unless a
122 program is specified.
124 If you want to use the traditional NCSA proxy authentication, set
125 this line to something like
127 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
129 "children" numberofchildren
130 The number of authenticator processes to spawn. If you start too few
131 Squid will have to wait for them to process a backlog of credential
132 verifications, slowing it down. When password verifications are
133 done via a (slow) network you are likely to need lots of
134 authenticator processes.
135 auth_param basic children 5
137 "concurrency" concurrency
138 The number of concurrent requests the helper can process.
139 The default of 0 is used for helpers who only supports
140 one request at a time. Setting this changes the protocol used to
141 include a channel number first on the request/response line, allowing
142 multiple requests to be sent to the same helper in parallell without
143 wating for the response.
144 Must not be set unless it's known the helper supports this.
145 auth_param basic concurrency 0
148 Specifies the realm name which is to be reported to the
149 client for the basic proxy authentication scheme (part of
150 the text the user will see when prompted their username and
151 password). There is no default.
152 auth_param basic realm Squid proxy-caching web server
154 "credentialsttl" timetolive
155 Specifies how long squid assumes an externally validated
156 username:password pair is valid for - in other words how
157 often the helper program is called for that user. Set this
158 low to force revalidation with short lived passwords. Note
159 setting this high does not impact your susceptibility
160 to replay attacks unless you are using an one-time password
161 system (such as SecureID). If you are using such a system,
162 you will be vulnerable to replay attacks unless you also
163 use the max_user_ip ACL in an http_access rule.
165 "casesensitive" on|off
166 Specifies if usernames are case sensitive. Most user databases are
167 case insensitive allowing the same username to be spelled using both
168 lower and upper case letters, but some are case sensitive. This
169 makes a big difference for user_max_ip ACL processing and similar.
170 auth_param basic casesensitive off
172 === Parameters for the digest scheme follow ===
175 Specify the command for the external authenticator. Such
176 a program reads a line containing "username":"realm" and
177 replies with the appropriate H(A1) value hex encoded or
178 ERR if the user (or his H(A1) hash) does not exists.
179 See rfc 2616 for the definition of H(A1).
180 "ERR" responses may optionally be followed by a error description
181 available as %m in the returned error page.
183 By default, the digest authentication scheme is not used unless a
184 program is specified.
186 If you want to use a digest authenticator, set this line to
189 auth_param digest program @DEFAULT_PREFIX@/bin/digest_auth_pw @DEFAULT_PREFIX@/etc/digpass
191 "children" numberofchildren
192 The number of authenticator processes to spawn (no default).
193 If you start too few Squid will have to wait for them to
194 process a backlog of H(A1) calculations, slowing it down.
195 When the H(A1) calculations are done via a (slow) network
196 you are likely to need lots of authenticator processes.
197 auth_param digest children 5
200 Specifies the realm name which is to be reported to the
201 client for the digest proxy authentication scheme (part of
202 the text the user will see when prompted their username and
203 password). There is no default.
204 auth_param digest realm Squid proxy-caching web server
206 "nonce_garbage_interval" timeinterval
207 Specifies the interval that nonces that have been issued
208 to client_agent's are checked for validity.
210 "nonce_max_duration" timeinterval
211 Specifies the maximum length of time a given nonce will be
214 "nonce_max_count" number
215 Specifies the maximum number of times a given nonce can be
218 "nonce_strictness" on|off
219 Determines if squid requires strict increment-by-1 behavior
220 for nonce counts, or just incrementing (off - for use when
221 useragents generate nonce counts that occasionally miss 1
222 (ie, 1,2,4,6)). Default off.
224 "check_nonce_count" on|off
225 This directive if set to off can disable the nonce count check
226 completely to work around buggy digest qop implementations in
227 certain mainstream browser versions. Default on to check the
228 nonce count to protect from authentication replay attacks.
230 "post_workaround" on|off
231 This is a workaround to certain buggy browsers who sends
232 an incorrect request digest in POST requests when reusing
233 the same nonce as acquired earlier on a GET request.
235 === NTLM scheme options follow ===
238 Specify the command for the external NTLM authenticator.
239 Such a program reads exchanged NTLMSSP packets with
240 the browser via Squid until authentication is completed.
241 If you use an NTLM authenticator, make sure you have 1 acl
242 of type proxy_auth. By default, the NTLM authenticator_program
245 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
247 "children" numberofchildren
248 The number of authenticator processes to spawn (no default).
249 If you start too few Squid will have to wait for them to
250 process a backlog of credential verifications, slowing it
251 down. When credential verifications are done via a (slow)
252 network you are likely to need lots of authenticator
255 auth_param ntlm children 5
258 If you experience problems with PUT/POST requests when using the
259 Negotiate authentication scheme then you can try setting this to
260 off. This will cause Squid to forcibly close the connection on
261 the initial requests where the browser asks which schemes are
262 supported by the proxy.
264 auth_param ntlm keep_alive on
266 === Options for configuring the NEGOTIATE auth-scheme follow ===
269 Specify the command for the external Negotiate authenticator.
270 This protocol is used in Microsoft Active-Directory enabled setups with
271 the Microsoft Internet Explorer or Mozilla Firefox browsers.
272 Its main purpose is to exchange credentials with the Squid proxy
273 using the Kerberos mechanisms.
274 If you use a Negotiate authenticator, make sure you have at least one acl
275 of type proxy_auth active. By default, the negotiate authenticator_program
277 The only supported program for this role is the ntlm_auth
278 program distributed as part of Samba, version 4 or later.
280 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
282 "children" numberofchildren
283 The number of authenticator processes to spawn (no default).
284 If you start too few Squid will have to wait for them to
285 process a backlog of credential verifications, slowing it
286 down. When crendential verifications are done via a (slow)
287 network you are likely to need lots of authenticator
289 auth_param negotiate children 5
292 If you experience problems with PUT/POST requests when using the
293 Negotiate authentication scheme then you can try setting this to
294 off. This will cause Squid to forcibly close the connection on
295 the initial requests where the browser asks which schemes are
296 supported by the proxy.
298 auth_param negotiate keep_alive on
301 #Recommended minimum configuration per scheme:
302 #auth_param negotiate program <uncomment and complete this line to activate>
303 #auth_param negotiate children 5
304 #auth_param negotiate keep_alive on
305 #auth_param ntlm program <uncomment and complete this line to activate>
306 #auth_param ntlm children 5
307 #auth_param ntlm keep_alive on
308 #auth_param digest program <uncomment and complete this line>
309 #auth_param digest children 5
310 #auth_param digest realm Squid proxy-caching web server
311 #auth_param digest nonce_garbage_interval 5 minutes
312 #auth_param digest nonce_max_duration 30 minutes
313 #auth_param digest nonce_max_count 50
314 #auth_param basic program <uncomment and complete this line>
315 #auth_param basic children 5
316 #auth_param basic realm Squid proxy-caching web server
317 #auth_param basic credentialsttl 2 hours
321 NAME: authenticate_cache_garbage_interval
324 LOC: Config.authenticateGCInterval
326 The time period between garbage collection across the username cache.
327 This is a tradeoff between memory utilization (long intervals - say
328 2 days) and CPU (short intervals - say 1 minute). Only change if you
332 NAME: authenticate_ttl
335 LOC: Config.authenticateTTL
337 The time a user & their credentials stay in the logged in
338 user cache since their last request. When the garbage
339 interval passes, all user credentials that have passed their
340 TTL are removed from memory.
343 NAME: authenticate_ip_ttl
345 LOC: Config.authenticateIpTTL
348 If you use proxy authentication and the 'max_user_ip' ACL,
349 this directive controls how long Squid remembers the IP
350 addresses associated with each user. Use a small value
351 (e.g., 60 seconds) if your users might change addresses
352 quickly, as is the case with dialups. You might be safe
353 using a larger value (e.g., 2 hours) in a corporate LAN
354 environment with relatively static address assignments.
359 -----------------------------------------------------------------------------
362 NAME: external_acl_type
363 TYPE: externalAclHelper
364 LOC: Config.externalAclHelperList
367 This option defines external acl classes using a helper program
368 to look up the status
370 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
374 ttl=n TTL in seconds for cached results (defaults to 3600
377 TTL for cached negative lookups (default same
379 children=n Number of acl helper processes spawn to service
380 external acl lookups of this type. (default 5)
381 concurrency=n concurrency level per process. Only used with helpers
382 capable of processing more than one query at a time.
383 cache=n result cache size, 0 is unbounded (default)
384 grace=n Percentage remaining of TTL where a refresh of a
385 cached entry should be initiated without needing to
386 wait for a new reply. (default 0 for no grace period)
387 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
388 ipv4 / ipv6 IP-mode used to communicate to this helper.
389 For compatability with older configurations and helpers
390 'ipv4' is the default unless --with-localhost-ipv6 is used.
391 --with-localhost-ipv6 changes the default to 'ipv6'.
392 SPECIAL NOTE: these options override --with-localhost-ipv6
394 FORMAT specifications
396 %LOGIN Authenticated user login name
397 %EXT_USER Username from external acl
398 %IDENT Ident user name
400 %SRCPORT Client source port
403 %PROTO Requested protocol
405 %PATH Requested URL path
406 %METHOD Request method
407 %MYADDR Squid interface address
408 %MYPORT Squid http_port number
409 %PATH Requested URL-path (including query-string if any)
410 %USER_CERT SSL User certificate in PEM format
411 %USER_CERTCHAIN SSL User certificate chain in PEM format
412 %USER_CERT_xx SSL User certificate subject attribute xx
413 %USER_CA_xx SSL User certificate issuer attribute xx
415 %>{Header} HTTP request header
417 HTTP request header list member
419 HTTP request header list member using ; as
420 list separator. ; can be any non-alphanumeric
423 %<{Header} HTTP reply header
425 HTTP reply header list member
427 HTTP reply header list member using ; as
428 list separator. ; can be any non-alphanumeric
431 In addition to the above, any string specified in the referencing
432 acl will also be included in the helper request line, after the
433 specified formats (see the "acl external" directive)
435 The helper receives lines per the above format specification,
436 and returns lines starting with OK or ERR indicating the validity
437 of the request and optionally followed by additional keywords with
440 General result syntax:
442 OK/ERR keyword=value ...
446 user= The users name (login)
447 password= The users password (for login= cache_peer option)
448 message= Message describing the reason. Available as %o
450 tag= Apply a tag to a request (for both ERR and OK results)
451 Only sets a tag, does not alter existing tags.
452 log= String to be logged in access.log. Available as
453 %ea in logformat specifications
455 If protocol=3.0 (the default) then URL escaping is used to protect
456 each value in both requests and responses.
458 If using protocol=2.5 then all values need to be enclosed in quotes
459 if they may contain whitespace, or the whitespace escaped using \.
460 And quotes or \ characters within the keyword value must be \ escaped.
462 When using the concurrency= option the protocol is changed by
463 introducing a query channel tag infront of the request/response.
464 The query channel tag is a number between 0 and concurrency-1.
472 Defining an Access List
474 Every access list definition must begin with an aclname and acltype,
475 followed by either type-specific arguments or a quoted filename that
478 acl aclname acltype argument ...
479 acl aclname acltype "file" ...
481 When using "file", the file should contain one item per line.
483 By default, regular expressions are CASE-SENSITIVE. To make
484 them case-insensitive, use the -i option.
487 ***** ACL TYPES AVAILABLE *****
489 acl aclname src ip-address/netmask ... # clients IP address
490 acl aclname src addr1-addr2/netmask ... # range of addresses
491 acl aclname dst ip-address/netmask ... # URL host's IP address
492 acl aclname myip ip-address/netmask ... # local socket IP address
494 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
495 # The arp ACL requires the special configure option --enable-arp-acl.
496 # Furthermore, the ARP ACL code is not portable to all operating systems.
497 # It works on Linux, Solaris, Windows, FreeBSD, and some other *BSD variants.
499 # NOTE: Squid can only determine the MAC address for clients that are on
500 # the same subnet. If the client is on a different subnet, then Squid cannot
501 # find out its MAC address.
503 acl aclname srcdomain .foo.com ... # reverse lookup, from client IP
504 acl aclname dstdomain .foo.com ... # Destination server from URL
505 acl aclname srcdom_regex [-i] \.foo\.com ... # regex matching client name
506 acl aclname dstdom_regex [-i] \.foo\.com ... # regex matching server
507 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
508 # based URL is used and no match is found. The name "none" is used
509 # if the reverse lookup fails.
511 acl aclname src_as number ...
512 acl aclname dst_as number ...
513 # Except for access control, AS numbers can be used for
514 # routing of requests to specific caches. Here's an
515 # example for routing all requests for AS#1241 and only
516 # those to mycache.mydomain.net:
517 # acl asexample dst_as 1241
518 # cache_peer_access mycache.mydomain.net allow asexample
519 # cache_peer_access mycache_mydomain.net deny all
521 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
530 # h1:m1 must be less than h2:m2
532 acl aclname url_regex [-i] ^http:// ... # regex matching on whole URL
533 acl aclname urlpath_regex [-i] \.gif$ ... # regex matching on URL path
535 acl aclname port 80 70 21 ...
536 acl aclname port 0-1024 ... # ranges allowed
537 acl aclname myport 3128 ... # (local socket TCP port)
538 acl aclname myportname 3128 ... # http(s)_port name
540 acl aclname proto HTTP FTP ...
542 acl aclname method GET POST ...
544 acl aclname http_status 200 301 500- 400-403 ... # status code in reply
546 acl aclname browser [-i] regexp ...
547 # pattern match on User-Agent header (see also req_header below)
549 acl aclname referer_regex [-i] regexp ...
550 # pattern match on Referer header
551 # Referer is highly unreliable, so use with care
553 acl aclname ident username ...
554 acl aclname ident_regex [-i] pattern ...
555 # string match on ident output.
556 # use REQUIRED to accept any non-null ident.
558 acl aclname proxy_auth [-i] username ...
559 acl aclname proxy_auth_regex [-i] pattern ...
560 # list of valid usernames
561 # use REQUIRED to accept any valid username.
563 # NOTE: when a Proxy-Authentication header is sent but it is not
564 # needed during ACL checking the username is NOT logged
567 # NOTE: proxy_auth requires a EXTERNAL authentication program
568 # to check username/password combinations (see
569 # auth_param directive).
571 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
572 # as the browser needs to be configured for using a proxy in order
573 # to respond to proxy authentication.
575 acl aclname snmp_community string ...
576 # A community string to limit access to your SNMP Agent
579 # acl snmppublic snmp_community public
581 acl aclname maxconn number
582 # This will be matched when the client's IP address has
583 # more than <number> HTTP connections established.
585 acl aclname max_user_ip [-s] number
586 # This will be matched when the user attempts to log in from more
587 # than <number> different ip addresses. The authenticate_ip_ttl
588 # parameter controls the timeout on the ip entries.
589 # If -s is specified the limit is strict, denying browsing
590 # from any further IP addresses until the ttl has expired. Without
591 # -s Squid will just annoy the user by "randomly" denying requests.
592 # (the counter is reset each time the limit is reached and a
594 # NOTE: in acceleration mode or where there is mesh of child proxies,
595 # clients may appear to come from multiple addresses if they are
596 # going through proxy farms, so a limit of 1 may cause user problems.
598 acl aclname req_mime_type [-i] mime-type ...
599 # regex match against the mime type of the request generated
600 # by the client. Can be used to detect file upload or some
601 # types HTTP tunneling requests.
602 # NOTE: This does NOT match the reply. You cannot use this
603 # to match the returned file type.
605 acl aclname req_header header-name [-i] any\.regex\.here
606 # regex match against any of the known request headers. May be
607 # thought of as a superset of "browser", "referer" and "mime-type"
610 acl aclname rep_mime_type [-i] mime-type ...
611 # regex match against the mime type of the reply received by
612 # squid. Can be used to detect file download or some
613 # types HTTP tunneling requests.
614 # NOTE: This has no effect in http_access rules. It only has
615 # effect in rules that affect the reply data stream such as
618 acl aclname rep_header header-name [-i] any\.regex\.here
619 # regex match against any of the known reply headers. May be
620 # thought of as a superset of "browser", "referer" and "mime-type"
623 acl aclname external class_name [arguments...]
624 # external ACL lookup via a helper class defined by the
625 # external_acl_type directive.
627 acl aclname user_cert attribute values...
628 # match against attributes in a user SSL certificate
629 # attribute is one of DN/C/O/CN/L/ST
631 acl aclname ca_cert attribute values...
632 # match against attributes a users issuing CA SSL certificate
633 # attribute is one of DN/C/O/CN/L/ST
635 acl aclname ext_user username ...
636 acl aclname ext_user_regex [-i] pattern ...
637 # string match on username returned by external acl helper
638 # use REQUIRED to accept any non-null user name.
641 acl macaddress arp 09:00:2b:23:45:67
642 acl myexample dst_as 1241
643 acl password proxy_auth REQUIRED
644 acl fileupload req_mime_type -i ^multipart/form-data$
645 acl javascript rep_mime_type -i ^application/x-javascript$
648 #Recommended minimum configuration:
649 acl manager proto cache_object
650 acl localhost src 127.0.0.1/32
651 acl to_localhost dst 127.0.0.0/8
653 # Example rule allowing access from your local networks.
654 # Adapt to list your (internal) IP networks from where browsing
656 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
657 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
658 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
660 acl SSL_ports port 443
661 acl Safe_ports port 80 # http
662 acl Safe_ports port 21 # ftp
663 acl Safe_ports port 443 # https
664 acl Safe_ports port 70 # gopher
665 acl Safe_ports port 210 # wais
666 acl Safe_ports port 1025-65535 # unregistered ports
667 acl Safe_ports port 280 # http-mgmt
668 acl Safe_ports port 488 # gss-http
669 acl Safe_ports port 591 # filemaker
670 acl Safe_ports port 777 # multiling http
671 acl CONNECT method CONNECT
675 NAME: follow_x_forwarded_for
677 IFDEF: FOLLOW_X_FORWARDED_FOR
678 LOC: Config.accessList.followXFF
680 DEFAULT_IF_NONE: deny all
682 Allowing or Denying the X-Forwarded-For header to be followed to
683 find the original source of a request.
685 Requests may pass through a chain of several other proxies
686 before reaching us. The X-Forwarded-For header will contain a
687 comma-separated list of the IP addresses in the chain, with the
688 rightmost address being the most recent.
690 If a request reaches us from a source that is allowed by this
691 configuration item, then we consult the X-Forwarded-For header
692 to see where that host received the request from. If the
693 X-Forwarded-For header contains multiple addresses, and if
694 acl_uses_indirect_client is on, then we continue backtracking
695 until we reach an address for which we are not allowed to
696 follow the X-Forwarded-For header, or until we reach the first
697 address in the list. (If acl_uses_indirect_client is off, then
698 it's impossible to backtrack through more than one level of
699 X-Forwarded-For addresses.)
701 The end result of this process is an IP address that we will
702 refer to as the indirect client address. This address may
703 be treated as the client address for access control, delay
704 pools and logging, depending on the acl_uses_indirect_client,
705 delay_pool_uses_indirect_client and log_uses_indirect_client
708 SECURITY CONSIDERATIONS:
710 Any host for which we follow the X-Forwarded-For header
711 can place incorrect information in the header, and Squid
712 will use the incorrect information as if it were the
713 source address of the request. This may enable remote
714 hosts to bypass any access control restrictions that are
715 based on the client's source addresses.
719 acl localhost src 127.0.0.1
720 acl my_other_proxy srcdomain .proxy.example.com
721 follow_x_forwarded_for allow localhost
722 follow_x_forwarded_for allow my_other_proxy
725 NAME: acl_uses_indirect_client
728 IFDEF: FOLLOW_X_FORWARDED_FOR
730 LOC: Config.onoff.acl_uses_indirect_client
732 Controls whether the indirect client address
733 (see follow_x_forwarded_for) is used instead of the
734 direct client address in acl matching.
737 NAME: delay_pool_uses_indirect_client
740 IFDEF: FOLLOW_X_FORWARDED_FOR && DELAY_POOLS
742 LOC: Config.onoff.delay_pool_uses_indirect_client
744 Controls whether the indirect client address
745 (see follow_x_forwarded_for) is used instead of the
746 direct client address in delay pools.
749 NAME: log_uses_indirect_client
752 IFDEF: FOLLOW_X_FORWARDED_FOR
754 LOC: Config.onoff.log_uses_indirect_client
756 Controls whether the indirect client address
757 (see follow_x_forwarded_for) is used instead of the
758 direct client address in the access log.
763 LOC: Config.accessList.http
765 DEFAULT_IF_NONE: deny all
767 Allowing or Denying access based on defined access lists
769 Access to the HTTP port:
770 http_access allow|deny [!]aclname ...
772 NOTE on default values:
774 If there are no "access" lines present, the default is to deny
777 If none of the "access" lines cause a match, the default is the
778 opposite of the last line in the list. If the last line was
779 deny, the default is allow. Conversely, if the last line
780 is allow, the default will be deny. For these reasons, it is a
781 good idea to have an "deny all" or "allow all" entry at the end
782 of your access lists to avoid potential confusion.
785 #Recommended minimum configuration:
787 # Only allow cachemgr access from localhost
788 http_access allow manager localhost
789 http_access deny manager
790 # Deny requests to unknown ports
791 http_access deny !Safe_ports
792 # Deny CONNECT to other than SSL ports
793 http_access deny CONNECT !SSL_ports
795 # We strongly recommend the following be uncommented to protect innocent
796 # web applications running on the proxy server who think the only
797 # one who can access services on "localhost" is a local user
798 #http_access deny to_localhost
800 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
802 # Example rule allowing access from your local networks.
803 # Adapt localnet in the ACL section to list your (internal) IP networks
804 # from where browsing should be allowed
805 http_access allow localnet
807 # And finally deny all other access to this proxy
812 NAME: http_reply_access
814 LOC: Config.accessList.reply
817 Allow replies to client requests. This is complementary to http_access.
819 http_reply_access allow|deny [!] aclname ...
821 NOTE: if there are no access lines present, the default is to allow
824 If none of the access lines cause a match the opposite of the
825 last line will apply. Thus it is good practice to end the rules
826 with an "allow all" or "deny all" entry.
831 LOC: Config.accessList.icp
833 DEFAULT_IF_NONE: deny all
835 Allowing or Denying access to the ICP port based on defined
838 icp_access allow|deny [!]aclname ...
840 See http_access for details
843 #Allow ICP queries from local networks only
844 #icp_access allow localnet
852 LOC: Config.accessList.htcp
854 DEFAULT_IF_NONE: deny all
856 Allowing or Denying access to the HTCP port based on defined
859 htcp_access allow|deny [!]aclname ...
861 See http_access for details
863 NOTE: The default if no htcp_access lines are present is to
864 deny all traffic. This default may cause problems with peers
865 using the htcp or htcp-oldsquid options.
868 #Allow HTCP queries from local networks only
869 #htcp_access allow localnet
870 #htcp_access deny all
874 NAME: htcp_clr_access
877 LOC: Config.accessList.htcp_clr
879 DEFAULT_IF_NONE: deny all
881 Allowing or Denying access to purge content using HTCP based
882 on defined access lists
884 htcp_clr_access allow|deny [!]aclname ...
886 See http_access for details
888 #Allow HTCP CLR requests from trusted peers
889 acl htcp_clr_peer src 172.16.1.2
890 htcp_clr_access allow htcp_clr_peer
895 LOC: Config.accessList.miss
898 Use to force your neighbors to use you as a sibling instead of
899 a parent. For example:
901 acl localclients src 172.16.0.0/16
902 miss_access allow localclients
903 miss_access deny !localclients
905 This means only your local clients are allowed to fetch
906 MISSES and all other clients can only fetch HITS.
908 By default, allow all clients who passed the http_access rules
909 to fetch MISSES from us.
912 NAME: ident_lookup_access
916 DEFAULT_IF_NONE: deny all
917 LOC: Config.accessList.identLookup
919 A list of ACL elements which, if matched, cause an ident
920 (RFC 931) lookup to be performed for this request. For
921 example, you might choose to always perform ident lookups
922 for your main multi-user Unix boxes, but not for your Macs
923 and PCs. By default, ident lookups are not performed for
926 To enable ident lookups for specific client addresses, you
927 can follow this example:
929 acl ident_aware_hosts src 198.168.1.0/255.255.255.0
930 ident_lookup_access allow ident_aware_hosts
931 ident_lookup_access deny all
933 Only src type ACL checks are fully supported. A src_domain
934 ACL might work at times, but it will not always provide
938 NAME: reply_body_max_size
939 COMMENT: size [acl acl...]
942 LOC: Config.ReplyBodySize
944 This option specifies the maximum size of a reply body. It can be
945 used to prevent users from downloading very large files, such as
946 MP3's and movies. When the reply headers are received, the
947 reply_body_max_size lines are processed, and the first line where
948 all (if any) listed ACLs are true is used as the maximum body size
951 This size is checked twice. First when we get the reply headers,
952 we check the content-length value. If the content length value exists
953 and is larger than the allowed size, the request is denied and the
954 user receives an error message that says "the request or reply
955 is too large." If there is no content-length, and the reply
956 size exceeds this limit, the client's connection is just closed
957 and they will receive a partial reply.
959 WARNING: downstream caches probably can not detect a partial reply
960 if there is no content-length header, so they will cache
961 partial responses and give them out as hits. You should NOT
962 use this option if you have downstream caches.
964 WARNING: A maximum size smaller than the size of squid's error messages
965 will cause an infinite loop and crash squid. Ensure that the smallest
966 non-zero value you use is greater that the maximum header size plus
967 the size of your largest error page.
969 If you set this parameter none (the default), there will be
975 -----------------------------------------------------------------------------
978 NAME: http_port ascii_port
981 LOC: Config.Sockaddr.http
983 Usage: port [options]
984 hostname:port [options]
985 1.2.3.4:port [options]
987 The socket addresses where Squid will listen for HTTP client
988 requests. You may specify multiple socket addresses.
989 There are three forms: port alone, hostname with port, and
990 IP address with port. If you specify a hostname or IP
991 address, Squid binds the socket to that specific
992 address. This replaces the old 'tcp_incoming_address'
993 option. Most likely, you do not need to bind to a specific
994 address, so you can use the port number alone.
996 If you are running Squid in accelerator mode, you
997 probably want to listen on port 80 also, or instead.
999 The -a command line option may be used to specify additional
1000 port(s) where Squid listens for proxy request. Such ports will
1001 be plain proxy ports with no options.
1003 You may specify multiple socket addresses on multiple lines.
1007 intercept Support for IP-Layer interception of
1008 outgoing requests without browser settings.
1009 NP: disables authentication and IPv6 on the port.
1011 tproxy Support Linux TPROXY for spoofing outgoing
1012 connections using the client IP address.
1013 NP: disables authentication and IPv6 on the port.
1015 accel Accelerator mode. Also needs at least one of
1016 vhost / vport / defaultsite.
1018 defaultsite=domainname
1019 What to use for the Host: header if it is not present
1020 in a request. Determines what site (not origin server)
1021 accelerators should consider the default.
1024 vhost Accelerator mode using Host header for virtual
1025 domain support. Implies accel.
1027 vport Accelerator with IP based virtual host support.
1030 vport=NN As above, but uses specified port number rather
1031 than the http_port number. Implies accel.
1033 protocol= Protocol to reconstruct accelerated requests with.
1036 disable-pmtu-discovery=
1037 Control Path-MTU discovery usage:
1038 off lets OS decide on what to do (default).
1039 transparent disable PMTU discovery when transparent
1041 always disable always PMTU discovery.
1043 In many setups of transparently intercepting proxies
1044 Path-MTU discovery can not work on traffic towards the
1045 clients. This is the case when the intercepting device
1046 does not fully track connections and fails to forward
1047 ICMP must fragment messages to the cache server. If you
1048 have such setup and experience that certain clients
1049 sporadically hang or never complete requests set
1050 disable-pmtu-discovery option to 'transparent'.
1052 sslBump Intercept each CONNECT request matching ssl_bump ACL,
1053 establish secure connection with the client and with
1054 the server, decrypt HTTP messages as they pass through
1055 Squid, and treat them as unencrypted HTTP messages,
1056 becoming the man-in-the-middle.
1058 When this option is enabled, additional options become
1059 available to specify SSL-related properties of the
1060 client-side connection: cert, key, version, cipher,
1061 options, clientca, cafile, capath, crlfile, dhparams,
1062 sslflags, and sslcontext. See the https_port directive
1063 for more information on these options.
1065 The ssl_bump option is required to fully enable
1066 the SslBump feature.
1068 name= Specifies a internal name for the port. Defaults to
1069 the port specification (port or addr:port)
1071 keepalive[=idle,interval,timeout]
1072 Enable TCP keepalive probes of idle connections
1073 idle is the initial time before TCP starts probing
1074 the connection, interval how often to probe, and
1075 timeout the time before giving up.
1077 If you run Squid on a dual-homed machine with an internal
1078 and an external interface we recommend you to specify the
1079 internal address:port in http_port. This way Squid will only be
1080 visible on the internal address.
1083 # Squid normally listens to port 3128
1084 http_port @DEFAULT_HTTP_PORT@
1090 TYPE: https_port_list
1092 LOC: Config.Sockaddr.https
1094 Usage: [ip:]port cert=certificate.pem [key=key.pem] [options...]
1096 The socket address where Squid will listen for HTTPS client
1099 This is really only useful for situations where you are running
1100 squid in accelerator mode and you want to do the SSL work at the
1103 You may specify multiple socket addresses on multiple lines,
1104 each with their own SSL certificate and/or options.
1108 accel Accelerator mode. Also needs at least one of
1109 defaultsite or vhost.
1111 defaultsite= The name of the https site presented on
1112 this port. Implies accel.
1114 vhost Accelerator mode using Host header for virtual
1115 domain support. Requires a wildcard certificate
1116 or other certificate valid for more than one domain.
1119 protocol= Protocol to reconstruct accelerated requests with.
1122 cert= Path to SSL certificate (PEM format).
1124 key= Path to SSL private key file (PEM format)
1125 if not specified, the certificate file is
1126 assumed to be a combined certificate and
1129 version= The version of SSL/TLS supported
1130 1 automatic (default)
1135 cipher= Colon separated list of supported ciphers.
1137 options= Various SSL engine options. The most important
1139 NO_SSLv2 Disallow the use of SSLv2
1140 NO_SSLv3 Disallow the use of SSLv3
1141 NO_TLSv1 Disallow the use of TLSv1
1142 SINGLE_DH_USE Always create a new key when using
1143 temporary/ephemeral DH key exchanges
1144 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1145 documentation for a complete list of options.
1147 clientca= File containing the list of CAs to use when
1148 requesting a client certificate.
1150 cafile= File containing additional CA certificates to
1151 use when verifying client certificates. If unset
1152 clientca will be used.
1154 capath= Directory containing additional CA certificates
1155 and CRL lists to use when verifying client certificates.
1157 crlfile= File of additional CRL lists to use when verifying
1158 the client certificate, in addition to CRLs stored in
1159 the capath. Implies VERIFY_CRL flag below.
1161 dhparams= File containing DH parameters for temporary/ephemeral
1164 sslflags= Various flags modifying the use of SSL:
1166 Don't request client certificates
1167 immediately, but wait until acl processing
1168 requires a certificate (not yet implemented).
1170 Don't use the default CA lists built in
1173 Don't allow for session reuse. Each connection
1174 will result in a new SSL session.
1176 Verify CRL lists when accepting client
1179 Verify CRL lists for all certificates in the
1180 client certificate chain.
1182 sslcontext= SSL session ID context identifier.
1184 vport Accelerator with IP based virtual host support.
1186 vport=NN As above, but uses specified port number rather
1187 than the https_port number. Implies accel.
1189 name= Specifies a internal name for the port. Defaults to
1190 the port specification (port or addr:port)
1194 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1197 LOC: Config.accessList.outgoing_tos
1199 Allows you to select a TOS/Diffserv value to mark outgoing
1200 connections with, based on the username or source address
1203 tcp_outgoing_tos ds-field [!]aclname ...
1205 Example where normal_service_net uses the TOS value 0x00
1206 and normal_service_net uses 0x20
1208 acl normal_service_net src 10.0.0.0/255.255.255.0
1209 acl good_service_net src 10.0.1.0/255.255.255.0
1210 tcp_outgoing_tos 0x00 normal_service_net
1211 tcp_outgoing_tos 0x20 good_service_net
1213 TOS/DSCP values really only have local significance - so you should
1214 know what you're specifying. For more information, see RFC2474 and
1217 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1218 "default" to use whatever default your host has. Note that in
1219 practice often only values 0 - 63 is usable as the two highest bits
1220 have been redefined for use by ECN (RFC3168).
1222 Processing proceeds in the order specified, and stops at first fully
1225 Note: The use of this directive using client dependent ACLs is
1226 incompatible with the use of server side persistent connections. To
1227 ensure correct results it is best to set server_persisten_connections
1228 to off when using this directive in such configurations.
1231 NAME: clientside_tos
1234 LOC: Config.accessList.clientside_tos
1236 Allows you to select a TOS/Diffserv value to mark client-side
1237 connections with, based on the username or source address
1245 LOC: Config.zph_tos_local
1247 Allows you to select a TOS/Diffserv value to mark local hits. Read above
1248 (tcp_outgoing_tos) for details/requirements about TOS.
1249 Default: 0 (disabled).
1256 LOC: Config.zph_tos_peer
1258 Allows you to select a TOS/Diffserv value to mark peer hits. Read above
1259 (tcp_outgoing_tos) for details/requirements about TOS.
1260 Default: 0 (disabled).
1263 NAME: zph_tos_parent
1268 LOC: Config.onoff.zph_tos_parent
1270 Set this to off if you want only sibling hits to be marked.
1271 If set to on (default), parent hits are being marked too.
1274 NAME: zph_preserve_miss_tos
1279 LOC: Config.onoff.zph_preserve_miss_tos
1281 If set to on (default), any HTTP response towards clients will
1282 have the TOS value of the response comming from the remote
1283 server masked with the value of zph_preserve_miss_tos_mask.
1284 For this to work correctly, you will need to patch your linux
1285 kernel with the TOS preserving ZPH patch.
1286 The kernel patch can be downloaded from http://zph.bratcheda.org
1289 NAME: zph_preserve_miss_tos_mask
1293 LOC: Config.zph_preserve_miss_tos_mask
1295 Allows you to mask certain bits in the TOS received from the
1296 remote server, before copying the value to the TOS send towards
1298 Default: 255 (TOS from server is not changed).
1301 NAME: tcp_outgoing_address
1304 LOC: Config.accessList.outgoing_address
1306 Allows you to map requests to different outgoing IP addresses
1307 based on the username or source address of the user making
1310 tcp_outgoing_address ipaddr [[!]aclname] ...
1312 Example where requests from 10.0.0.0/24 will be forwarded
1313 with source address 10.1.0.1, 10.0.2.0/24 forwarded with
1314 source address 10.1.0.2 and the rest will be forwarded with
1315 source address 10.1.0.3.
1317 acl normal_service_net src 10.0.0.0/24
1318 acl good_service_net src 10.0.2.0/24
1319 tcp_outgoing_address 10.1.0.1 normal_service_net
1320 tcp_outgoing_address 10.1.0.2 good_service_net
1321 tcp_outgoing_address 10.1.0.3
1323 Processing proceeds in the order specified, and stops at first fully
1326 Note: The use of this directive using client dependent ACLs is
1327 incompatible with the use of server side persistent connections. To
1328 ensure correct results it is best to set server_persistent_connections
1329 to off when using this directive in such configurations.
1334 Squid is built with a capability of bridging the IPv4 and IPv6 internets.
1335 tcp_outgoing_address as exampled above breaks this bridging by forcing
1336 all outbound traffic through a certain IPv4 which may be on the wrong
1337 side of the IPv4/IPv6 boundary.
1339 To operate with tcp_outgoing_address and keep the bridging benefits
1340 an additional ACL needs to be used which ensures the IPv6-bound traffic
1341 is never forced or permitted out the IPv4 interface.
1343 acl to_ipv6 dst ipv6
1344 tcp_outgoing_address 2002::c001 good_service_net to_ipv6
1345 tcp_outgoing_address 10.1.0.2 good_service_net !to_ipv6
1347 tcp_outgoing_address 2002::beef normal_service_net to_ipv6
1348 tcp_outgoing_address 10.1.0.1 normal_service_net !to_ipv6
1350 tcp_outgoing_address 2002::1 to_ipv6
1351 tcp_outgoing_address 10.1.0.3 !to_ipv6
1356 -----------------------------------------------------------------------------
1359 NAME: ssl_unclean_shutdown
1363 LOC: Config.SSL.unclean_shutdown
1365 Some browsers (especially MSIE) bugs out on SSL shutdown
1372 LOC: Config.SSL.ssl_engine
1375 The OpenSSL engine to use. You will need to set this if you
1376 would like to use hardware SSL acceleration for example.
1379 NAME: sslproxy_client_certificate
1382 LOC: Config.ssl_client.cert
1385 Client SSL Certificate to use when proxying https:// URLs
1388 NAME: sslproxy_client_key
1391 LOC: Config.ssl_client.key
1394 Client SSL Key to use when proxying https:// URLs
1397 NAME: sslproxy_version
1400 LOC: Config.ssl_client.version
1403 SSL version level to use when proxying https:// URLs
1406 NAME: sslproxy_options
1409 LOC: Config.ssl_client.options
1412 SSL engine options to use when proxying https:// URLs
1415 NAME: sslproxy_cipher
1418 LOC: Config.ssl_client.cipher
1421 SSL cipher list to use when proxying https:// URLs
1424 NAME: sslproxy_cafile
1427 LOC: Config.ssl_client.cafile
1430 file containing CA certificates to use when verifying server
1431 certificates while proxying https:// URLs
1434 NAME: sslproxy_capath
1437 LOC: Config.ssl_client.capath
1440 directory containing CA certificates to use when verifying
1441 server certificates while proxying https:// URLs
1447 LOC: Config.accessList.ssl_bump
1450 This ACL controls which CONNECT requests to an http_port
1451 marked with an sslBump flag are actually "bumped". Please
1452 see the sslBump flag of an http_port option for more details
1453 about decoding proxied SSL connections.
1455 By default, no requests are bumped.
1457 See also: http_port sslBump
1460 # Example: Bump all requests except those originating from localhost and
1461 # those going to webax.com or example.com sites.
1463 # acl localhost src 127.0.0.1/32
1464 # acl broken_sites dstdomain .webax.com
1465 # acl broken_sites dstdomain .example.com
1466 # ssl_bump deny localhost
1467 # ssl_bump deny broken_sites
1468 # ssl_bump allow all
1472 NAME: sslproxy_flags
1475 LOC: Config.ssl_client.flags
1478 Various flags modifying the use of SSL while proxying https:// URLs:
1479 DONT_VERIFY_PEER Accept certificates that fail verification.
1480 For refined control, see sslproxy_cert_error.
1481 NO_DEFAULT_CA Don't use the default CA list built in
1486 NAME: sslproxy_cert_error
1489 LOC: Config.ssl_client.cert_error
1492 Use this ACL to bypass server certificate validation errors.
1494 For example, the following lines will bypass all validation errors
1495 when talking to servers located at 172.16.0.0/16. All other
1496 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
1498 acl BrokenServersAtTrustedIP dst 172.16.0.0/16
1499 sslproxy_cert_error allow BrokenServersAtTrustedIP
1500 sslproxy_cert_error deny all
1502 This option must use fast ACL expressions only. Expressions that use
1503 external lookups or communication result in unpredictable behavior or
1506 Without this option, all server certificate validation errors
1507 terminate the transaction. Bypassing validation errors is dangerous
1508 because an error usually implies that the server cannot be trusted and
1509 the connection may be insecure.
1511 See also: sslproxy_flags and DONT_VERIFY_PEER.
1515 # sslproxy_cert_error deny all
1521 NAME: sslpassword_program
1524 LOC: Config.Program.ssl_password
1527 Specify a program used for entering SSL key passphrases
1528 when using encrypted SSL certificate keys. If not specified
1529 keys must either be unencrypted, or Squid started with the -N
1530 option to allow it to query interactively for the passphrase.
1534 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
1535 -----------------------------------------------------------------------------
1543 To specify other caches in a hierarchy, use the format:
1545 cache_peer hostname type http-port icp-port [options]
1550 # hostname type port port options
1551 # -------------------- -------- ----- ----- -----------
1552 cache_peer parent.foo.net parent 3128 3130 proxy-only default
1553 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
1554 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
1556 type: either 'parent', 'sibling', or 'multicast'.
1558 proxy-port: The port number where the cache listens for proxy
1561 icp-port: Used for querying neighbor caches about
1562 objects. To have a non-ICP neighbor
1563 specify '0' for the ICP port.
1564 NOTE: Also requires icp_port option enabled to send/receive
1565 requests via this method.
1575 weighted-round-robin
1584 login=user:password | PASS | *:password
1595 sslcert=/path/to/ssl/certificate
1596 sslkey=/path/to/ssl/key
1600 front-end-https[=on|auto]
1602 use 'proxy-only' to specify objects fetched
1603 from this cache should not be saved locally.
1605 use 'weight=n' to affect the selection of a peer
1606 during any weighted peer-selection mechanisms.
1607 The weight must be an integer; default is 1,
1608 larger weights are favored more.
1609 This option does not affect parent selection if a peering
1610 protocol is not in use.
1612 use 'basetime=n' to specify a base amount to
1613 be subtracted from round trip times of parents.
1614 It is subtracted before division by weight in calculating
1615 which parent to fectch from. If the rtt is less than the
1616 base time the rtt is set to a minimal value.
1618 use 'ttl=n' to specify a IP multicast TTL to use
1619 when sending an ICP queries to this address.
1620 Only useful when sending to a multicast group.
1621 Because we don't accept ICP replies from random
1622 hosts, you must configure other group members as
1623 peers with the 'multicast-responder' option below.
1625 use 'no-query' to NOT send ICP queries to this
1628 use 'background-ping' to only send ICP queries to this
1629 neighbor infrequently. This is used to keep the neighbor
1630 round trip time updated and is usually used in
1631 conjunction with weighted-round-robin.
1633 use 'default' if this is a parent cache which can
1634 be used as a "last-resort" if a peer cannot be located
1635 by any of the peer-selection mechanisms.
1636 If specified more than once, only the first is used.
1638 use 'round-robin' to define a set of parents which
1639 should be used in a round-robin fashion in the
1640 absence of any ICP queries.
1642 use 'weighted-round-robin' to define a set of parents
1643 which should be used in a round-robin fashion with the
1644 frequency of each parent being based on the round trip
1645 time. Closer parents are used more often.
1646 Usually used for background-ping parents.
1648 use 'carp' to define a set of parents which should
1649 be used as a CARP array. The requests will be
1650 distributed among the parents based on the CARP load
1651 balancing hash function based on their weight.
1653 use 'userhash' to load-balance amongst a set of parents
1654 based on the client proxy_auth or ident username.
1656 use 'sourcehash' to load-balance amongst a set of parents
1657 based on the client source ip.
1659 'multicast-responder' indicates the named peer
1660 is a member of a multicast group. ICP queries will
1661 not be sent directly to the peer, but ICP replies
1662 will be accepted from it.
1664 'closest-only' indicates that, for ICP_OP_MISS
1665 replies, we'll only forward CLOSEST_PARENT_MISSes
1666 and never FIRST_PARENT_MISSes.
1668 use 'no-digest' to NOT request cache digests from
1671 'no-netdb-exchange' disables requesting ICMP
1672 RTT database (NetDB) from the neighbor.
1674 use 'no-delay' to prevent access to this neighbor
1675 from influencing the delay pools.
1677 use 'login=user:password' if this is a personal/workgroup
1678 proxy and your parent requires proxy authentication.
1679 Note: The string can include URL escapes (i.e. %20 for
1680 spaces). This also means % must be written as %%.
1682 use 'login=PASS' if users must authenticate against
1683 the upstream proxy or in the case of a reverse proxy
1684 configuration, the origin web server. This will pass
1685 the users credentials as they are to the peer.
1686 This only works for the Basic HTTP authentication scheme.
1687 Note: To combine this with proxy_auth both proxies must
1688 share the same user database as HTTP only allows for
1689 a single login (one for proxy, one for origin server).
1690 Also be warned this will expose your users proxy
1691 password to the peer. USE WITH CAUTION
1693 use 'login=*:password' to pass the username to the
1694 upstream cache, but with a fixed password. This is meant
1695 to be used when the peer is in another administrative
1696 domain, but it is still needed to identify each user.
1697 The star can optionally be followed by some extra
1698 information which is added to the username. This can
1699 be used to identify this proxy to the peer, similar to
1700 the login=username:password option above.
1702 use 'connect-timeout=nn' to specify a peer
1703 specific connect timeout (also see the
1704 peer_connect_timeout directive)
1706 use 'digest-url=url' to tell Squid to fetch the cache
1707 digest (if digests are enabled) for this host from
1708 the specified URL rather than the Squid default
1711 use 'allow-miss' to disable Squid's use of only-if-cached
1712 when forwarding requests to siblings. This is primarily
1713 useful when icp_hit_stale is used by the sibling. To
1714 extensive use of this option may result in forwarding
1715 loops, and you should avoid having two-way peerings
1716 with this option. (for example to deny peer usage on
1717 requests from peer by denying cache_peer_access if the
1720 use 'max-conn=n' to limit the amount of connections Squid
1721 may open to this peer.
1723 use 'htcp' to send HTCP, instead of ICP, queries
1724 to the neighbor. You probably also want to
1725 set the "icp port" to 4827 instead of 3130.
1726 You MUST also set htcp_access expicitly. The default of
1727 deny all will prevent peer traffic.
1729 use 'htcp-oldsquid' to send HTCP to old Squid versions
1730 You MUST also set htcp_access expicitly. The default of
1731 deny all will prevent peer traffic.
1733 'originserver' causes this parent peer to be contacted as
1734 a origin server. Meant to be used in accelerator setups.
1736 use 'name=xxx' if you have multiple peers on the same
1737 host but different ports. This name can be used to
1738 differentiate the peers in cache_peer_access and similar
1741 use 'forceddomain=name' to forcibly set the Host header
1742 of requests forwarded to this peer. Useful in accelerator
1743 setups where the server (peer) expects a certain domain
1744 name and using redirectors to feed this domain name
1747 use 'ssl' to indicate connections to this peer should
1748 be SSL/TLS encrypted.
1750 use 'sslcert=/path/to/ssl/certificate' to specify a client
1751 SSL certificate to use when connecting to this peer.
1753 use 'sslkey=/path/to/ssl/key' to specify the private SSL
1754 key corresponding to sslcert above. If 'sslkey' is not
1755 specified 'sslcert' is assumed to reference a
1756 combined file containing both the certificate and the key.
1758 use sslversion=1|2|3|4 to specify the SSL version to use
1759 when connecting to this peer
1760 1 = automatic (default)
1765 use sslcipher=... to specify the list of valid SSL ciphers
1766 to use when connecting to this peer.
1768 use ssloptions=... to specify various SSL engine options:
1769 NO_SSLv2 Disallow the use of SSLv2
1770 NO_SSLv3 Disallow the use of SSLv3
1771 NO_TLSv1 Disallow the use of TLSv1
1772 See src/ssl_support.c or the OpenSSL documentation for
1773 a more complete list.
1775 use sslcafile=... to specify a file containing
1776 additional CA certificates to use when verifying the
1779 use sslcapath=... to specify a directory containing
1780 additional CA certificates to use when verifying the
1783 use sslcrlfile=... to specify a certificate revocation
1784 list file to use when verifying the peer certificate.
1786 use sslflags=... to specify various flags modifying the
1789 Accept certificates even if they fail to
1792 Don't use the default CA list built in
1795 Don't verify the peer certificate
1796 matches the server name
1798 use ssldomain= to specify the peer name as advertised
1799 in it's certificate. Used for verifying the correctness
1800 of the received peer certificate. If not specified the
1801 peer hostname will be used.
1803 use front-end-https to enable the "Front-End-Https: On"
1804 header needed when using Squid as a SSL frontend in front
1805 of Microsoft OWA. See MS KB document Q307347 for details
1806 on this header. If set to auto the header will
1807 only be added if the request is forwarded as a https://
1811 NAME: cache_peer_domain cache_host_domain
1816 Use to limit the domains for which a neighbor cache will be
1819 cache_peer_domain cache-host domain [domain ...]
1820 cache_peer_domain cache-host !domain
1822 For example, specifying
1824 cache_peer_domain parent.foo.net .edu
1826 has the effect such that UDP query packets are sent to
1827 'bigserver' only when the requested object exists on a
1828 server in the .edu domain. Prefixing the domainname
1829 with '!' means the cache will be queried for objects
1832 NOTE: * Any number of domains may be given for a cache-host,
1833 either on the same or separate lines.
1834 * When multiple domains are given for a particular
1835 cache-host, the first matched domain is applied.
1836 * Cache hosts with no domain restrictions are queried
1838 * There are no defaults.
1839 * There is also a 'cache_peer_access' tag in the ACL
1843 NAME: cache_peer_access
1848 Similar to 'cache_peer_domain' but provides more flexibility by
1851 cache_peer_access cache-host allow|deny [!]aclname ...
1853 The syntax is identical to 'http_access' and the other lists of
1854 ACL elements. See the comments for 'http_access' below, or
1855 the Squid FAQ (http://www.squid-cache.org/FAQ/FAQ-10.html).
1858 NAME: neighbor_type_domain
1859 TYPE: hostdomaintype
1863 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
1865 Modifying the neighbor type for specific domains is now
1866 possible. You can treat some domains differently than the the
1867 default neighbor type specified on the 'cache_peer' line.
1868 Normally it should only be necessary to list domains which
1869 should be treated differently because the default neighbor type
1870 applies for hostnames which do not match domains listed here.
1873 cache_peer cache.foo.org parent 3128 3130
1874 neighbor_type_domain cache.foo.org sibling .com .net
1875 neighbor_type_domain cache.foo.org sibling .au .de
1878 NAME: dead_peer_timeout
1882 LOC: Config.Timeout.deadPeer
1884 This controls how long Squid waits to declare a peer cache
1885 as "dead." If there are no ICP replies received in this
1886 amount of time, Squid will declare the peer dead and not
1887 expect to receive any further ICP replies. However, it
1888 continues to send ICP queries, and will mark the peer as
1889 alive upon receipt of the first subsequent ICP reply.
1891 This timeout also affects when Squid expects to receive ICP
1892 replies from peers. If more than 'dead_peer' seconds have
1893 passed since the last ICP reply was received, Squid will not
1894 expect to receive an ICP reply on the next query. Thus, if
1895 your time between requests is greater than this timeout, you
1896 will see a lot of requests sent DIRECT to origin servers
1897 instead of to your parents.
1900 NAME: hierarchy_stoplist
1903 LOC: Config.hierarchy_stoplist
1905 A list of words which, if found in a URL, cause the object to
1906 be handled directly by this cache. In other words, use this
1907 to not query neighbor caches for certain objects. You may
1908 list this option multiple times.
1909 Note: never_direct overrides this option.
1911 #We recommend you to use at least the following line.
1912 hierarchy_stoplist cgi-bin ?
1917 MEMORY CACHE OPTIONS
1918 -----------------------------------------------------------------------------
1925 LOC: Config.memMaxSize
1927 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
1928 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
1929 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
1930 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
1932 'cache_mem' specifies the ideal amount of memory to be used
1934 * In-Transit objects
1936 * Negative-Cached objects
1938 Data for these objects are stored in 4 KB blocks. This
1939 parameter specifies the ideal upper limit on the total size of
1940 4 KB blocks allocated. In-Transit objects take the highest
1943 In-transit objects have priority over the others. When
1944 additional space is needed for incoming data, negative-cached
1945 and hot objects will be released. In other words, the
1946 negative-cached and hot objects will fill up any unused space
1947 not needed for in-transit objects.
1949 If circumstances require, this limit will be exceeded.
1950 Specifically, if your incoming request rate requires more than
1951 'cache_mem' of memory to hold in-transit objects, Squid will
1952 exceed this limit to satisfy the new requests. When the load
1953 decreases, blocks will be freed until the high-water mark is
1954 reached. Thereafter, blocks will be used to store hot
1958 NAME: maximum_object_size_in_memory
1962 LOC: Config.Store.maxInMemObjSize
1964 Objects greater than this size will not be attempted to kept in
1965 the memory cache. This should be set high enough to keep objects
1966 accessed frequently in memory to improve performance whilst low
1967 enough to keep larger objects from hoarding cache_mem.
1970 NAME: memory_replacement_policy
1972 LOC: Config.memPolicy
1975 The memory replacement policy parameter determines which
1976 objects are purged from memory when memory space is needed.
1978 See cache_replacement_policy for details.
1983 -----------------------------------------------------------------------------
1986 NAME: cache_replacement_policy
1988 LOC: Config.replPolicy
1991 The cache replacement policy parameter determines which
1992 objects are evicted (replaced) when disk space is needed.
1994 lru : Squid's original list based LRU policy
1995 heap GDSF : Greedy-Dual Size Frequency
1996 heap LFUDA: Least Frequently Used with Dynamic Aging
1997 heap LRU : LRU policy implemented using a heap
1999 Applies to any cache_dir lines listed below this.
2001 The LRU policies keeps recently referenced objects.
2003 The heap GDSF policy optimizes object hit rate by keeping smaller
2004 popular objects in cache so it has a better chance of getting a
2005 hit. It achieves a lower byte hit rate than LFUDA though since
2006 it evicts larger (possibly popular) objects.
2008 The heap LFUDA policy keeps popular objects in cache regardless of
2009 their size and thus optimizes byte hit rate at the expense of
2010 hit rate since one large, popular object will prevent many
2011 smaller, slightly less popular objects from being cached.
2013 Both policies utilize a dynamic aging mechanism that prevents
2014 cache pollution that can otherwise occur with frequency-based
2015 replacement policies.
2017 NOTE: if using the LFUDA replacement policy you should increase
2018 the value of maximum_object_size above its default of 4096 KB to
2019 to maximize the potential byte hit rate improvement of LFUDA.
2021 For more information about the GDSF and LFUDA cache replacement
2022 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
2023 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
2029 LOC: Config.cacheSwap
2033 cache_dir Type Directory-Name Fs-specific-data [options]
2035 You can specify multiple cache_dir lines to spread the
2036 cache among different disk partitions.
2038 Type specifies the kind of storage system to use. Only "ufs"
2039 is built by default. To enable any of the other storage systems
2040 see the --enable-storeio configure option.
2042 'Directory' is a top-level directory where cache swap
2043 files will be stored. If you want to use an entire disk
2044 for caching, this can be the mount-point directory.
2045 The directory must exist and be writable by the Squid
2046 process. Squid will NOT create this directory for you.
2050 "ufs" is the old well-known Squid storage format that has always
2053 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
2055 'Mbytes' is the amount of disk space (MB) to use under this
2056 directory. The default is 100 MB. Change this to suit your
2057 configuration. Do NOT put the size of your disk drive here.
2058 Instead, if you want Squid to use the entire disk drive,
2059 subtract 20% and use that value.
2061 'Level-1' is the number of first-level subdirectories which
2062 will be created under the 'Directory'. The default is 16.
2064 'Level-2' is the number of second-level subdirectories which
2065 will be created under each first-level directory. The default
2068 The aufs store type:
2070 "aufs" uses the same storage format as "ufs", utilizing
2071 POSIX-threads to avoid blocking the main Squid process on
2072 disk-I/O. This was formerly known in Squid as async-io.
2074 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
2076 see argument descriptions under ufs above
2078 The diskd store type:
2080 "diskd" uses the same storage format as "ufs", utilizing a
2081 separate process to avoid blocking the main Squid process on
2084 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
2086 see argument descriptions under ufs above
2088 Q1 specifies the number of unacknowledged I/O requests when Squid
2089 stops opening new files. If this many messages are in the queues,
2090 Squid won't open new files. Default is 64
2092 Q2 specifies the number of unacknowledged messages when Squid
2093 starts blocking. If this many messages are in the queues,
2094 Squid blocks until it receives some replies. Default is 72
2096 When Q1 < Q2 (the default), the cache directory is optimized
2097 for lower response time at the expense of a decrease in hit
2098 ratio. If Q1 > Q2, the cache directory is optimized for
2099 higher hit ratio at the expense of an increase in response
2102 The coss store type:
2104 block-size=n defines the "block size" for COSS cache_dir's.
2105 Squid uses file numbers as block numbers. Since file numbers
2106 are limited to 24 bits, the block size determines the maximum
2107 size of the COSS partition. The default is 512 bytes, which
2108 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
2109 you should not change the coss block size after Squid
2110 has written some objects to the cache_dir.
2112 The coss file store has changed from 2.5. Now it uses a file
2113 called 'stripe' in the directory names in the config - and
2114 this will be created by squid -z.
2118 no-store, no new objects should be stored to this cache_dir
2120 max-size=n, refers to the max object size this storedir supports.
2121 It is used to initially choose the storedir to dump the object.
2122 Note: To make optimal use of the max-size limits you should order
2123 the cache_dir lines with the smallest max-size value first and the
2124 ones with no max-size specification last.
2126 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
2127 which can be changed with the --with-coss-membuf-size=N configure
2130 # cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
2134 NAME: store_dir_select_algorithm
2136 LOC: Config.store_dir_select_algorithm
2139 Set this to 'round-robin' as an alternative.
2142 NAME: max_open_disk_fds
2144 LOC: Config.max_open_disk_fds
2147 To avoid having disk as the I/O bottleneck Squid can optionally
2148 bypass the on-disk cache if more than this amount of disk file
2149 descriptors are open.
2151 A value of 0 indicates no limit.
2154 NAME: minimum_object_size
2158 LOC: Config.Store.minObjectSize
2160 Objects smaller than this size will NOT be saved on disk. The
2161 value is specified in kilobytes, and the default is 0 KB, which
2162 means there is no minimum.
2165 NAME: maximum_object_size
2169 LOC: Config.Store.maxObjectSize
2171 Objects larger than this size will NOT be saved on disk. The
2172 value is specified in kilobytes, and the default is 4MB. If
2173 you wish to get a high BYTES hit ratio, you should probably
2174 increase this (one 32 MB object hit counts for 3200 10KB
2175 hits). If you wish to increase speed more than your want to
2176 save bandwidth you should leave this low.
2178 NOTE: if using the LFUDA replacement policy you should increase
2179 this value to maximize the byte hit rate improvement of LFUDA!
2180 See replacement_policy below for a discussion of this policy.
2183 NAME: cache_swap_low
2184 COMMENT: (percent, 0-100)
2187 LOC: Config.Swap.lowWaterMark
2190 NAME: cache_swap_high
2191 COMMENT: (percent, 0-100)
2194 LOC: Config.Swap.highWaterMark
2197 The low- and high-water marks for cache object replacement.
2198 Replacement begins when the swap (disk) usage is above the
2199 low-water mark and attempts to maintain utilization near the
2200 low-water mark. As swap utilization gets close to high-water
2201 mark object eviction becomes more aggressive. If utilization is
2202 close to the low-water mark less replacement is done each time.
2204 Defaults are 90% and 95%. If you have a large cache, 5% could be
2205 hundreds of MB. If this is the case you may wish to set these
2206 numbers closer together.
2211 -----------------------------------------------------------------------------
2216 LOC: Config.Log.logformats
2221 logformat <name> <format specification>
2223 Defines an access log format.
2225 The <format specification> is a string with embedded % format codes
2227 % format codes all follow the same basic structure where all but
2228 the formatcode is optional. Output strings are automatically escaped
2229 as required according to their context and the output format
2230 modifiers are usually not needed, but can be specified if an explicit
2231 output format is desired.
2233 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
2235 " output in quoted string format
2236 [ output in squid text log format as used by log_mime_hdrs
2237 # output in URL quoted format
2241 width field width. If starting with 0 the
2242 output is zero padded
2243 {arg} argument such as header name etc
2247 >a Client source IP address
2249 >p Client source port
2250 <A Server IP address or peer name
2251 la Local IP address (http_port)
2252 lp Local port number (http_port)
2253 ts Seconds since epoch
2254 tu subsecond time (milliseconds)
2255 tl Local time. Optional strftime format argument
2256 default %d/%b/%Y:%H:%M:%S %z
2257 tg GMT time. Optional strftime format argument
2258 default %d/%b/%Y:%H:%M:%S %z
2259 tr Response time (milliseconds)
2260 >h Request header. Optional header name argument
2261 on the format header[:[separator]element]
2262 <h Reply header. Optional header name argument
2265 ul User name from authentication
2266 ui User name from ident
2267 us User name from SSL
2268 ue User name from external acl helper
2270 Ss Squid request status (TCP_MISS etc)
2271 Sh Squid hierarchy status (DEFAULT_PARENT etc)
2272 mt MIME content type
2273 rm Request method (GET/POST etc)
2275 rp Request URL-Path excluding hostname
2276 rv Request protocol version
2277 et Tag returned by external acl
2278 ea Log string returned by external acl
2279 <st Reply size including HTTP headers
2280 >st Request size including HTTP headers
2281 st Request+Reply size including HTTP headers
2282 <sH Reply high offset sent
2283 <sS Upstream object size
2284 % a literal % character
2286 The default formats available (which do not need re-defining) are:
2288 logformat squid %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt
2289 logformat squidmime %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt [%>h] [%<h]
2290 logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st %Ss:%Sh
2291 logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
2294 NAME: access_log cache_access_log
2296 LOC: Config.Log.accesslogs
2297 DEFAULT: @DEFAULT_ACCESS_LOG@ squid
2299 These files log client request activities. Has a line every HTTP or
2300 ICP request. The format is:
2301 access_log <filepath> [<logformat name> [acl acl ...]]
2302 access_log none [acl acl ...]]
2304 Will log to the specified file using the specified format (which
2305 must be defined in a logformat directive) those entries which match
2306 ALL the acl's specified (which must be defined in acl clauses).
2307 If no acl is specified, all requests will be logged to this file.
2309 To disable logging of a request use the filepath "none", in which case
2310 a logformat name should not be specified.
2312 To log the request via syslog specify a filepath of "syslog":
2314 access_log syslog[:facility.priority] [format [acl1 [acl2 ....]]]
2315 where facility could be any of:
2316 authpriv, daemon, local0 .. local7 or user.
2318 And priority could be any of:
2319 err, warning, notice, info, debug.
2322 access_log @DEFAULT_ACCESS_LOG@ squid
2327 LOC: Config.accessList.log
2329 COMMENT: allow|deny acl acl...
2331 This options allows you to control which requests gets logged
2332 to access.log (see access_log directive). Requests denied for
2333 logging will also not be accounted for in performance counters.
2338 DEFAULT: @DEFAULT_CACHE_LOG@
2341 Cache logging file. This is where general information about
2342 your cache's behavior goes. You can increase the amount of data
2343 logged to this file with the "debug_options" tag below.
2346 NAME: cache_store_log
2349 LOC: Config.Log.store
2351 Logs the activities of the storage manager. Shows which
2352 objects are ejected from the cache, and which objects are
2353 saved and for how long. To disable, enter "none" or remove the line.
2354 There are not really utilities to analyze this data, so you can safely
2357 # cache_store_log @DEFAULT_STORE_LOG@
2361 NAME: cache_swap_state cache_swap_log
2363 LOC: Config.Log.swap
2366 Location for the cache "swap.state" file. This index file holds
2367 the metadata of objects saved on disk. It is used to rebuild
2368 the cache during startup. Normally this file resides in each
2369 'cache_dir' directory, but you may specify an alternate
2370 pathname here. Note you must give a full filename, not just
2371 a directory. Since this is the index for the whole object
2372 list you CANNOT periodically rotate it!
2374 If %s can be used in the file name it will be replaced with a
2375 a representation of the cache_dir name where each / is replaced
2376 with '.'. This is needed to allow adding/removing cache_dir
2377 lines when cache_swap_log is being used.
2379 If have more than one 'cache_dir', and %s is not used in the name
2380 these swap logs will have names such as:
2386 The numbered extension (which is added automatically)
2387 corresponds to the order of the 'cache_dir' lines in this
2388 configuration file. If you change the order of the 'cache_dir'
2389 lines in this file, these index files will NOT correspond to
2390 the correct 'cache_dir' entry (unless you manually rename
2391 them). We recommend you do NOT use this option. It is
2392 better to keep these index files in each 'cache_dir' directory.
2395 NAME: logfile_rotate
2398 LOC: Config.Log.rotateNumber
2400 Specifies the number of logfile rotations to make when you
2401 type 'squid -k rotate'. The default is 10, which will rotate
2402 with extensions 0 through 9. Setting logfile_rotate to 0 will
2403 disable the file name rotation, but the logfiles are still closed
2404 and re-opened. This will enable you to rename the logfiles
2405 yourself just before sending the rotate signal.
2407 Note, the 'squid -k rotate' command normally sends a USR1
2408 signal to the running squid process. In certain situations
2409 (e.g. on Linux with Async I/O), USR1 is used for other
2410 purposes, so -k rotate uses another signal. It is best to get
2411 in the habit of using 'squid -k rotate' instead of 'kill -USR1
2415 NAME: emulate_httpd_log
2419 LOC: Config.onoff.common_log
2421 The Cache can emulate the log file format which many 'httpd'
2422 programs use. To disable/enable this emulation, set
2423 emulate_httpd_log to 'off' or 'on'. The default
2424 is to use the native log format since it includes useful
2425 information Squid-specific log analyzers use.
2428 NAME: log_ip_on_direct
2432 LOC: Config.onoff.log_ip_on_direct
2434 Log the destination IP address in the hierarchy log tag when going
2435 direct. Earlier Squid versions logged the hostname here. If you
2436 prefer the old way set this to off.
2441 DEFAULT: @DEFAULT_MIME_TABLE@
2442 LOC: Config.mimeTablePathname
2444 Pathname to Squid's MIME table. You shouldn't need to change
2445 this, but the default file contains examples and formatting
2446 information if you do.
2452 LOC: Config.onoff.log_mime_hdrs
2455 The Cache can record both the request and the response MIME
2456 headers for each HTTP transaction. The headers are encoded
2457 safely and will appear as two bracketed fields at the end of
2458 the access log (for either the native or httpd-emulated log
2459 formats). To enable this logging set log_mime_hdrs to 'on'.
2464 LOC: Config.Log.useragent
2466 IFDEF: USE_USERAGENT_LOG
2468 Squid will write the User-Agent field from HTTP requests
2469 to the filename specified here. By default useragent_log
2473 NAME: referer_log referrer_log
2475 LOC: Config.Log.referer
2477 IFDEF: USE_REFERER_LOG
2479 Squid will write the Referer field from HTTP requests to the
2480 filename specified here. By default referer_log is disabled.
2481 Note that "referer" is actually a misspelling of "referrer"
2482 however the misspelt version has been accepted into the HTTP RFCs
2488 DEFAULT: @DEFAULT_PID_FILE@
2489 LOC: Config.pidFilename
2491 A filename to write the process-id to. To disable, enter "none".
2497 LOC: Config.debugOptions
2499 Logging options are set as section,level where each source file
2500 is assigned a unique section. Lower levels result in less
2501 output, Full debugging (level 9) can result in a very large
2502 log file, so be careful. The magic word "ALL" sets debugging
2503 levels for all sections. We recommend normally running with
2511 LOC: Config.onoff.log_fqdn
2513 Turn this on if you wish to log fully qualified domain names
2514 in the access.log. To do this Squid does a DNS lookup of all
2515 IP's connecting to it. This can (in some situations) increase
2516 latency, which makes your cache seem slower for interactive
2520 NAME: client_netmask
2522 LOC: Config.Addrs.client_netmask
2523 DEFAULT: 255.255.255.255
2525 A netmask for client addresses in logfiles and cachemgr output.
2526 Change this to protect the privacy of your cache clients.
2527 A netmask of 255.255.255.0 will log all IP's in that range with
2528 the last digit set to '0'.
2535 LOC: Config.Log.forward
2537 Logs the server-side requests.
2539 This is currently work in progress.
2542 NAME: strip_query_terms
2544 LOC: Config.onoff.strip_query_terms
2547 By default, Squid strips query terms from requested URLs before
2548 logging. This protects your user's privacy.
2555 LOC: Config.onoff.buffered_logs
2557 cache.log log file is written with stdio functions, and as such
2558 it can be buffered or unbuffered. By default it will be unbuffered.
2559 Buffering it can speed up the writing slightly (though you are
2560 unlikely to need to worry unless you run with tons of debugging
2561 enabled in which case performance will suffer badly anyway..).
2564 NAME: netdb_filename
2566 DEFAULT: @DEFAULT_NETDB_FILE@
2567 LOC: Config.netdbFilename
2570 A filename where Squid stores it's netdb state between restarts.
2571 To disable, enter "none".
2575 OPTIONS FOR FTP GATEWAYING
2576 -----------------------------------------------------------------------------
2582 LOC: Config.Ftp.anon_user
2584 If you want the anonymous login password to be more informative
2585 (and enable the use of picky ftp servers), set this to something
2586 reasonable for your domain, like wwwuser@somewhere.net
2588 The reason why this is domainless by default is the
2589 request can be made on the behalf of a user in any domain,
2590 depending on how the cache is used.
2591 Some ftp server also validate the email address is valid
2592 (for example perl.com).
2595 NAME: ftp_list_width
2598 LOC: Config.Ftp.list_width
2600 Sets the width of ftp listings. This should be set to fit in
2601 the width of a standard browser. Setting this too small
2602 can cut off long filenames when browsing ftp sites.
2608 LOC: Config.Ftp.passive
2610 If your firewall does not allow Squid to use passive
2611 connections, turn off this option.
2613 Use of ftp_epsv_all option requires this to be ON.
2619 LOC: Config.Ftp.epsv_all
2621 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
2623 NATs may be able to put the connection on a "fast path" through the
2624 translator, as the EPRT command will never be used and therefore,
2625 translation of the data portion of the segments will never be needed.
2627 When a client only expects to do two-way FTP transfers this may be useful.
2628 If squid finds that it must do a three-way FTP transfer after issuing
2629 an EPSV ALL command, the FTP session will fail.
2631 If you have any doubts about this option do not use it.
2632 Squid will nicely attempt all other connection methods.
2634 Requires ftp_passive to be ON (default)
2637 NAME: ftp_sanitycheck
2640 LOC: Config.Ftp.sanitycheck
2642 For security and data integrity reasons Squid by default performs
2643 sanity checks of the addresses of FTP data connections ensure the
2644 data connection is to the requested server. If you need to allow
2645 FTP connections to servers using another IP address for the data
2646 connection turn this off.
2649 NAME: ftp_telnet_protocol
2652 LOC: Config.Ftp.telnet
2654 The FTP protocol is officially defined to use the telnet protocol
2655 as transport channel for the control connection. However, many
2656 implementations are broken and does not respect this aspect of
2659 If you have trouble accessing files with ASCII code 255 in the
2660 path or similar problems involving this ASCII code you can
2661 try setting this directive to off. If that helps, report to the
2662 operator of the FTP server in question that their FTP server
2663 is broken and does not follow the FTP standard.
2667 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
2668 -----------------------------------------------------------------------------
2673 DEFAULT: @DEFAULT_DISKD@
2674 LOC: Config.Program.diskd
2676 Specify the location of the diskd executable.
2677 Note this is only useful if you have compiled in
2678 diskd as one of the store io modules.
2681 NAME: unlinkd_program
2684 DEFAULT: @DEFAULT_UNLINKD@
2685 LOC: Config.Program.unlinkd
2687 Specify the location of the executable for file deletion process.
2690 NAME: pinger_program
2692 DEFAULT: @DEFAULT_PINGER@
2693 LOC: Config.pinger.program
2696 Specify the location of the executable for the pinger process.
2702 LOC: Config.pinger.enable
2705 Control whether the pinger is active at run-time.
2706 Enables turning ICMP pinger on and off with a simple squid -k reconfigure.
2711 OPTIONS FOR URL REWRITING
2712 -----------------------------------------------------------------------------
2715 NAME: url_rewrite_program redirect_program
2717 LOC: Config.Program.redirect
2720 Specify the location of the executable for the URL rewriter.
2721 Since they can perform almost any function there isn't one included.
2723 For each requested URL rewriter will receive on line with the format
2725 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
2727 In the future, the rewriter interface will be extended with
2728 key=value pairs ("kvpairs" shown above). Rewriter programs
2729 should be prepared to receive and possibly ignore additional
2730 whitespace-separated tokens on each input line.
2732 And the rewriter may return a rewritten URL. The other components of
2733 the request line does not need to be returned (ignored if they are).
2735 The rewriter can also indicate that a client-side redirect should
2736 be performed to the new URL. This is done by prefixing the returned
2737 URL with "301:" (moved permanently) or 302: (moved temporarily).
2739 By default, a URL rewriter is not used.
2742 NAME: url_rewrite_children redirect_children
2745 LOC: Config.redirectChildren
2747 The number of redirector processes to spawn. If you start
2748 too few Squid will have to wait for them to process a backlog of
2749 URLs, slowing it down. If you start too many they will use RAM
2750 and other system resources.
2753 NAME: url_rewrite_concurrency redirect_concurrency
2756 LOC: Config.redirectConcurrency
2758 The number of requests each redirector helper can handle in
2759 parallel. Defaults to 0 which indicates the redirector
2760 is a old-style single threaded redirector.
2763 NAME: url_rewrite_host_header redirect_rewrites_host_header
2766 LOC: Config.onoff.redir_rewrites_host
2768 By default Squid rewrites any Host: header in redirected
2769 requests. If you are running an accelerator this may
2770 not be a wanted effect of a redirector.
2772 WARNING: Entries are cached on the result of the URL rewriting
2773 process, so be careful if you have domain-virtual hosts.
2776 NAME: url_rewrite_access redirector_access
2779 LOC: Config.accessList.redirector
2781 If defined, this access list specifies which requests are
2782 sent to the redirector processes. By default all requests
2786 NAME: url_rewrite_bypass redirector_bypass
2788 LOC: Config.onoff.redirector_bypass
2791 When this is 'on', a request will not go through the
2792 redirector if all redirectors are busy. If this is 'off'
2793 and the redirector queue grows too large, Squid will exit
2794 with a FATAL error and ask you to increase the number of
2795 redirectors. You should only enable this if the redirectors
2796 are not critical to your caching system. If you use
2797 redirectors for access control, and you enable this option,
2798 users may have access to pages they should not
2799 be allowed to request.
2803 OPTIONS FOR TUNING THE CACHE
2804 -----------------------------------------------------------------------------
2807 NAME: cache no_cache
2810 LOC: Config.accessList.noCache
2812 A list of ACL elements which, if matched and denied, cause the request to
2813 not be satisfied from the cache and the reply to not be cached.
2814 In other words, use this to force certain objects to never be cached.
2816 You must use the words 'allow' or 'deny' to indicate whether items
2817 matching the ACL should be allowed or denied into the cache.
2819 Default is to allow all to be cached.
2822 NAME: refresh_pattern
2823 TYPE: refreshpattern
2827 usage: refresh_pattern [-i] regex min percent max [options]
2829 By default, regular expressions are CASE-SENSITIVE. To make
2830 them case-insensitive, use the -i option.
2832 'Min' is the time (in minutes) an object without an explicit
2833 expiry time should be considered fresh. The recommended
2834 value is 0, any higher values may cause dynamic applications
2835 to be erroneously cached unless the application designer
2836 has taken the appropriate actions.
2838 'Percent' is a percentage of the objects age (time since last
2839 modification age) an object without explicit expiry time
2840 will be considered fresh.
2842 'Max' is an upper limit on how long objects without an explicit
2843 expiry time will be considered fresh.
2845 options: override-expire
2855 override-expire enforces min age even if the server
2856 sent an explicit expiry time (e.g., with the
2857 Expires: header or Cache-Control: max-age). Doing this
2858 VIOLATES the HTTP standard. Enabling this feature
2859 could make you liable for problems which it causes.
2861 override-lastmod enforces min age even on objects
2862 that were modified recently.
2864 reload-into-ims changes client no-cache or ``reload''
2865 to If-Modified-Since requests. Doing this VIOLATES the
2866 HTTP standard. Enabling this feature could make you
2867 liable for problems which it causes.
2869 ignore-reload ignores a client no-cache or ``reload''
2870 header. Doing this VIOLATES the HTTP standard. Enabling
2871 this feature could make you liable for problems which
2874 ignore-no-cache ignores any ``Pragma: no-cache'' and
2875 ``Cache-control: no-cache'' headers received from a server.
2876 The HTTP RFC never allows the use of this (Pragma) header
2877 from a server, only a client, though plenty of servers
2880 ignore-no-store ignores any ``Cache-control: no-store''
2881 headers received from a server. Doing this VIOLATES
2882 the HTTP standard. Enabling this feature could make you
2883 liable for problems which it causes.
2885 ignore-private ignores any ``Cache-control: private''
2886 headers received from a server. Doing this VIOLATES
2887 the HTTP standard. Enabling this feature could make you
2888 liable for problems which it causes.
2890 ignore-auth caches responses to requests with authorization,
2891 as if the originserver had sent ``Cache-control: public''
2892 in the response header. Doing this VIOLATES the HTTP standard.
2893 Enabling this feature could make you liable for problems which
2896 refresh-ims causes squid to contact the origin server
2897 when a client issues an If-Modified-Since request. This
2898 ensures that the client will receive an updated version
2899 if one is available.
2901 Basically a cached object is:
2903 FRESH if expires < now, else STALE
2905 FRESH if lm-factor < percent, else STALE
2909 The refresh_pattern lines are checked in the order listed here.
2910 The first entry which matches is used. If none of the entries
2911 match the default will be used.
2913 Note, you must uncomment all the default lines if you want
2914 to change one. The default setting is only active if none is
2919 refresh_pattern ^ftp: 1440 20% 10080
2920 refresh_pattern ^gopher: 1440 0% 1440
2921 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
2922 refresh_pattern . 0 20% 4320
2926 NAME: quick_abort_min
2930 LOC: Config.quickAbort.min
2933 NAME: quick_abort_max
2937 LOC: Config.quickAbort.max
2940 NAME: quick_abort_pct
2944 LOC: Config.quickAbort.pct
2946 The cache by default continues downloading aborted requests
2947 which are almost completed (less than 16 KB remaining). This
2948 may be undesirable on slow (e.g. SLIP) links and/or very busy
2949 caches. Impatient users may tie up file descriptors and
2950 bandwidth by repeatedly requesting and immediately aborting
2953 When the user aborts a request, Squid will check the
2954 quick_abort values to the amount of data transfered until
2957 If the transfer has less than 'quick_abort_min' KB remaining,
2958 it will finish the retrieval.
2960 If the transfer has more than 'quick_abort_max' KB remaining,
2961 it will abort the retrieval.
2963 If more than 'quick_abort_pct' of the transfer has completed,
2964 it will finish the retrieval.
2966 If you do not want any retrieval to continue after the client
2967 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
2970 If you want retrievals to always continue if they are being
2971 cached set 'quick_abort_min' to '-1 KB'.
2974 NAME: read_ahead_gap
2975 COMMENT: buffer-size
2977 LOC: Config.readAheadGap
2980 The amount of data the cache will buffer ahead of what has been
2981 sent to the client when retrieving an object from another server.
2985 IFDEF: HTTP_VIOLATIONS
2988 LOC: Config.negativeTtl
2991 Set the Default Time-to-Live (TTL) for failed requests.
2992 Certain types of failures (such as "connection refused" and
2993 "404 Not Found") are able to be negatively-cached for a short time.
2994 Modern web servers should provide Expires: header, however if they
2995 do not this can provide a minimum TTL.
2996 The default is not to cache errors with unknown expiry details.
2998 Note that this is different from negative caching of DNS lookups.
3000 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3001 this feature could make you liable for problems which it
3005 NAME: positive_dns_ttl
3008 LOC: Config.positiveDnsTtl
3011 Upper limit on how long Squid will cache positive DNS responses.
3012 Default is 6 hours (360 minutes). This directive must be set
3013 larger than negative_dns_ttl.
3016 NAME: negative_dns_ttl
3019 LOC: Config.negativeDnsTtl
3022 Time-to-Live (TTL) for negative caching of failed DNS lookups.
3023 This also sets the lower cache limit on positive lookups.
3024 Minimum value is 1 second, and it is not recommendable to go
3025 much below 10 seconds.
3028 NAME: range_offset_limit
3031 LOC: Config.rangeOffsetLimit
3034 Sets a upper limit on how far into the the file a Range request
3035 may be to cause Squid to prefetch the whole file. If beyond this
3036 limit Squid forwards the Range request as it is and the result
3039 This is to stop a far ahead range request (lets say start at 17MB)
3040 from making Squid fetch the whole object up to that point before
3041 sending anything to the client.
3043 A value of -1 causes Squid to always fetch the object from the
3044 beginning so it may cache the result. (2.0 style)
3046 A value of 0 causes Squid to never fetch more than the
3047 client requested. (default)
3050 NAME: minimum_expiry_time
3053 LOC: Config.minimum_expiry_time
3056 The minimum caching time according to (Expires - Date)
3057 Headers Squid honors if the object can't be revalidated
3058 defaults to 60 seconds. In reverse proxy environments it
3059 might be desirable to honor shorter object lifetimes. It
3060 is most likely better to make your server return a
3061 meaningful Last-Modified header however. In ESI environments
3062 where page fragments often have short lifetimes, this will
3063 often be best set to 0.
3066 NAME: store_avg_object_size
3070 LOC: Config.Store.avgObjectSize
3072 Average object size, used to estimate number of objects your
3073 cache can hold. The default is 13 KB.
3076 NAME: store_objects_per_bucket
3079 LOC: Config.Store.objectsPerBucket
3081 Target number of objects per bucket in the store hash table.
3082 Lowering this value increases the total number of buckets and
3083 also the storage maintenance rate. The default is 20.
3088 -----------------------------------------------------------------------------
3091 NAME: request_header_max_size
3095 LOC: Config.maxRequestHeaderSize
3097 This specifies the maximum size for HTTP headers in a request.
3098 Request headers are usually relatively small (about 512 bytes).
3099 Placing a limit on the request header size will catch certain
3100 bugs (for example with persistent connections) and possibly
3101 buffer-overflow or denial-of-service attacks.
3104 NAME: reply_header_max_size
3108 LOC: Config.maxReplyHeaderSize
3110 This specifies the maximum size for HTTP headers in a reply.
3111 Reply headers are usually relatively small (about 512 bytes).
3112 Placing a limit on the reply header size will catch certain
3113 bugs (for example with persistent connections) and possibly
3114 buffer-overflow or denial-of-service attacks.
3117 NAME: request_body_max_size
3121 LOC: Config.maxRequestBodySize
3123 This specifies the maximum size for an HTTP request body.
3124 In other words, the maximum size of a PUT/POST request.
3125 A user who attempts to send a request with a body larger
3126 than this limit receives an "Invalid Request" error message.
3127 If you set this parameter to a zero (the default), there will
3128 be no limit imposed.
3134 LOC: Config.accessList.brokenPosts
3136 A list of ACL elements which, if matched, causes Squid to send
3137 an extra CRLF pair after the body of a PUT/POST request.
3139 Some HTTP servers has broken implementations of PUT/POST,
3140 and rely on an extra CRLF pair sent by some WWW clients.
3142 Quote from RFC2616 section 4.1 on this matter:
3144 Note: certain buggy HTTP/1.0 client implementations generate an
3145 extra CRLF's after a POST request. To restate what is explicitly
3146 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
3147 a request with an extra CRLF.
3150 acl buggy_server url_regex ^http://....
3151 broken_posts allow buggy_server
3155 IFDEF: HTTP_VIOLATIONS
3159 LOC: Config.onoff.via
3161 If set (default), Squid will include a Via header in requests and
3162 replies as required by RFC2616.
3168 LOC: Config.onoff.ie_refresh
3171 Microsoft Internet Explorer up until version 5.5 Service
3172 Pack 1 has an issue with transparent proxies, wherein it
3173 is impossible to force a refresh. Turning this on provides
3174 a partial fix to the problem, by causing all IMS-REFRESH
3175 requests from older IE versions to check the origin server
3176 for fresh content. This reduces hit ratio by some amount
3177 (~10% in my experience), but allows users to actually get
3178 fresh content when they want it. Note because Squid
3179 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
3180 of 5.5 is unchanged from old versions of Squid (i.e. a
3181 forced refresh is impossible). Newer versions of IE will,
3182 hopefully, continue to have the new behavior and will be
3183 handled based on that assumption. This option defaults to
3184 the old Squid behavior, which is better for hit ratios but
3185 worse for clients using IE, if they need to be able to
3186 force fresh content.
3189 NAME: vary_ignore_expire
3192 LOC: Config.onoff.vary_ignore_expire
3195 Many HTTP servers supporting Vary gives such objects
3196 immediate expiry time with no cache-control header
3197 when requested by a HTTP/1.0 client. This option
3198 enables Squid to ignore such expiry times until
3199 HTTP/1.1 is fully implemented.
3200 WARNING: This may eventually cause some varying
3201 objects not intended for caching to get cached.
3204 NAME: request_entities
3206 LOC: Config.onoff.request_entities
3209 Squid defaults to deny GET and HEAD requests with request entities,
3210 as the meaning of such requests are undefined in the HTTP standard
3211 even if not explicitly forbidden.
3213 Set this directive to on if you have clients which insists
3214 on sending request entities in GET or HEAD requests. But be warned
3215 that there is server software (both proxies and web servers) which
3216 can fail to properly process this kind of request which may make you
3217 vulnerable to cache pollution attacks if enabled.
3220 NAME: request_header_access
3221 IFDEF: HTTP_VIOLATIONS
3222 TYPE: http_header_access[]
3223 LOC: Config.request_header_access
3226 Usage: request_header_access header_name allow|deny [!]aclname ...
3228 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3229 this feature could make you liable for problems which it
3232 This option replaces the old 'anonymize_headers' and the
3233 older 'http_anonymizer' option with something that is much
3234 more configurable. This new method creates a list of ACLs
3235 for each header, allowing you very fine-tuned header
3238 This option only applies to request headers, i.e., from the
3239 client to the server.
3241 You can only specify known headers for the header name.
3242 Other headers are reclassified as 'Other'. You can also
3243 refer to all the headers with 'All'.
3245 For example, to achieve the same behavior as the old
3246 'http_anonymizer standard' option, you should use:
3248 request_header_access From deny all
3249 request_header_access Referer deny all
3250 request_header_access Server deny all
3251 request_header_access User-Agent deny all
3252 request_header_access WWW-Authenticate deny all
3253 request_header_access Link deny all
3255 Or, to reproduce the old 'http_anonymizer paranoid' feature
3258 request_header_access Allow allow all
3259 request_header_access Authorization allow all
3260 request_header_access WWW-Authenticate allow all
3261 request_header_access Proxy-Authorization allow all
3262 request_header_access Proxy-Authenticate allow all
3263 request_header_access Cache-Control allow all
3264 request_header_access Content-Encoding allow all
3265 request_header_access Content-Length allow all
3266 request_header_access Content-Type allow all
3267 request_header_access Date allow all
3268 request_header_access Expires allow all
3269 request_header_access Host allow all
3270 request_header_access If-Modified-Since allow all
3271 request_header_access Last-Modified allow all
3272 request_header_access Location allow all
3273 request_header_access Pragma allow all
3274 request_header_access Accept allow all
3275 request_header_access Accept-Charset allow all
3276 request_header_access Accept-Encoding allow all
3277 request_header_access Accept-Language allow all
3278 request_header_access Content-Language allow all
3279 request_header_access Mime-Version allow all
3280 request_header_access Retry-After allow all
3281 request_header_access Title allow all
3282 request_header_access Connection allow all
3283 request_header_access Proxy-Connection allow all
3284 request_header_access All deny all
3286 although many of those are HTTP reply headers, and so should be
3287 controlled with the reply_header_access directive.
3289 By default, all headers are allowed (no anonymizing is
3293 NAME: reply_header_access
3294 IFDEF: HTTP_VIOLATIONS
3295 TYPE: http_header_access[]
3296 LOC: Config.reply_header_access
3299 Usage: reply_header_access header_name allow|deny [!]aclname ...
3301 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3302 this feature could make you liable for problems which it
3305 This option only applies to reply headers, i.e., from the
3306 server to the client.
3308 This is the same as request_header_access, but in the other
3311 This option replaces the old 'anonymize_headers' and the
3312 older 'http_anonymizer' option with something that is much
3313 more configurable. This new method creates a list of ACLs
3314 for each header, allowing you very fine-tuned header
3317 You can only specify known headers for the header name.
3318 Other headers are reclassified as 'Other'. You can also
3319 refer to all the headers with 'All'.
3321 For example, to achieve the same behavior as the old
3322 'http_anonymizer standard' option, you should use:
3324 reply_header_access From deny all
3325 reply_header_access Referer deny all
3326 reply_header_access Server deny all
3327 reply_header_access User-Agent deny all
3328 reply_header_access WWW-Authenticate deny all
3329 reply_header_access Link deny all
3331 Or, to reproduce the old 'http_anonymizer paranoid' feature
3334 reply_header_access Allow allow all
3335 reply_header_access Authorization allow all
3336 reply_header_access WWW-Authenticate allow all
3337 reply_header_access Proxy-Authorization allow all
3338 reply_header_access Proxy-Authenticate allow all
3339 reply_header_access Cache-Control allow all
3340 reply_header_access Content-Encoding allow all
3341 reply_header_access Content-Length allow all
3342 reply_header_access Content-Type allow all
3343 reply_header_access Date allow all
3344 reply_header_access Expires allow all
3345 reply_header_access Host allow all
3346 reply_header_access If-Modified-Since allow all
3347 reply_header_access Last-Modified allow all
3348 reply_header_access Location allow all
3349 reply_header_access Pragma allow all
3350 reply_header_access Accept allow all
3351 reply_header_access Accept-Charset allow all
3352 reply_header_access Accept-Encoding allow all
3353 reply_header_access Accept-Language allow all
3354 reply_header_access Content-Language allow all
3355 reply_header_access Mime-Version allow all
3356 reply_header_access Retry-After allow all
3357 reply_header_access Title allow all
3358 reply_header_access Connection allow all
3359 reply_header_access Proxy-Connection allow all
3360 reply_header_access All deny all
3362 although the HTTP request headers won't be usefully controlled
3363 by this directive -- see request_header_access for details.
3365 By default, all headers are allowed (no anonymizing is
3369 NAME: header_replace
3370 IFDEF: HTTP_VIOLATIONS
3371 TYPE: http_header_replace[]
3372 LOC: Config.request_header_access
3375 Usage: header_replace header_name message
3376 Example: header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
3378 This option allows you to change the contents of headers
3379 denied with header_access above, by replacing them with
3380 some fixed string. This replaces the old fake_user_agent
3383 This only applies to request headers, not reply headers.
3385 By default, headers are removed if denied.
3388 NAME: relaxed_header_parser
3389 COMMENT: on|off|warn
3391 LOC: Config.onoff.relaxed_header_parser
3394 In the default "on" setting Squid accepts certain forms
3395 of non-compliant HTTP messages where it is unambiguous
3396 what the sending application intended even if the message
3397 is not correctly formatted. The messages is then normalized
3398 to the correct form when forwarded by Squid.
3400 If set to "warn" then a warning will be emitted in cache.log
3401 each time such HTTP error is encountered.
3403 If set to "off" then such HTTP errors will cause the request
3404 or response to be rejected.
3409 -----------------------------------------------------------------------------
3412 NAME: forward_timeout
3415 LOC: Config.Timeout.forward
3418 This parameter specifies how long Squid should at most attempt in
3419 finding a forwarding path for the request before giving up.
3422 NAME: connect_timeout
3425 LOC: Config.Timeout.connect
3428 This parameter specifies how long to wait for the TCP connect to
3429 the requested server or peer to complete before Squid should
3430 attempt to find another path where to forward the request.
3433 NAME: peer_connect_timeout
3436 LOC: Config.Timeout.peer_connect
3439 This parameter specifies how long to wait for a pending TCP
3440 connection to a peer cache. The default is 30 seconds. You
3441 may also set different timeout values for individual neighbors
3442 with the 'connect-timeout' option on a 'cache_peer' line.
3448 LOC: Config.Timeout.read
3451 The read_timeout is applied on server-side connections. After
3452 each successful read(), the timeout will be extended by this
3453 amount. If no data is read again after this amount of time,
3454 the request is aborted and logged with ERR_READ_TIMEOUT. The
3455 default is 15 minutes.
3458 NAME: request_timeout
3460 LOC: Config.Timeout.request
3463 How long to wait for an HTTP request after initial
3464 connection establishment.
3467 NAME: persistent_request_timeout
3469 LOC: Config.Timeout.persistent_request
3472 How long to wait for the next HTTP request on a persistent
3473 connection after the previous request completes.
3476 NAME: client_lifetime
3479 LOC: Config.Timeout.lifetime
3482 The maximum amount of time a client (browser) is allowed to
3483 remain connected to the cache process. This protects the Cache
3484 from having a lot of sockets (and hence file descriptors) tied up
3485 in a CLOSE_WAIT state from remote clients that go away without
3486 properly shutting down (either because of a network failure or
3487 because of a poor client implementation). The default is one
3490 NOTE: The default value is intended to be much larger than any
3491 client would ever need to be connected to your cache. You
3492 should probably change client_lifetime only as a last resort.
3493 If you seem to have many client connections tying up
3494 filedescriptors, we recommend first tuning the read_timeout,
3495 request_timeout, persistent_request_timeout and quick_abort values.
3498 NAME: half_closed_clients
3500 LOC: Config.onoff.half_closed_clients
3503 Some clients may shutdown the sending side of their TCP
3504 connections, while leaving their receiving sides open. Sometimes,
3505 Squid can not tell the difference between a half-closed and a
3506 fully-closed TCP connection. By default, half-closed client
3507 connections are kept open until a read(2) or write(2) on the
3508 socket returns an error. Change this option to 'off' and Squid
3509 will immediately close client connections when read(2) returns
3510 "no more data to read."
3515 LOC: Config.Timeout.pconn
3518 Timeout for idle persistent connections to servers and other
3525 LOC: Config.Timeout.ident
3528 Maximum time to wait for IDENT lookups to complete.
3530 If this is too high, and you enabled IDENT lookups from untrusted
3531 users, you might be susceptible to denial-of-service by having
3532 many ident requests going at once.
3535 NAME: shutdown_lifetime
3538 LOC: Config.shutdownLifetime
3541 When SIGTERM or SIGHUP is received, the cache is put into
3542 "shutdown pending" mode until all active sockets are closed.
3543 This value is the lifetime to set for all open descriptors
3544 during shutdown mode. Any active clients after this many
3545 seconds will receive a 'timeout' message.
3549 ADMINISTRATIVE PARAMETERS
3550 -----------------------------------------------------------------------------
3556 LOC: Config.adminEmail
3558 Email-address of local cache manager who will receive
3559 mail if the cache dies. The default is "webmaster."
3565 LOC: Config.EmailFrom
3567 From: email-address for mail sent when the cache dies.
3568 The default is to use 'appname@unique_hostname'.
3569 Default appname value is "squid", can be changed into
3570 src/globals.h before building squid.
3576 LOC: Config.EmailProgram
3578 Email program used to send mail if the cache dies.
3579 The default is "mail". The specified program must comply
3580 with the standard Unix mail syntax:
3581 mail-program recipient < mailfile
3583 Optional command line options can be specified.
3586 NAME: cache_effective_user
3588 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
3589 LOC: Config.effectiveUser
3591 If you start Squid as root, it will change its effective/real
3592 UID/GID to the user specified below. The default is to change
3593 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
3594 see also; cache_effective_group
3597 NAME: cache_effective_group
3600 LOC: Config.effectiveGroup
3602 Squid sets the GID to the effective user's default group ID
3603 (taken from the password file) and supplementary group list
3604 from the groups membership.
3606 If you want Squid to run with a specific GID regardless of
3607 the group memberships of the effective user then set this
3608 to the group (or GID) you want Squid to run as. When set
3609 all other group privileges of the effective user are ignored
3610 and only this GID is effective. If Squid is not started as
3611 root the user starting Squid MUST be member of the specified
3614 This option is not recommended by the Squid Team.
3615 Our preference is for administrators to configure a secure
3616 user account for squid with UID/GID matching system policies.
3619 NAME: httpd_suppress_version_string
3623 LOC: Config.onoff.httpd_suppress_version_string
3625 Suppress Squid version string info in HTTP headers and HTML error pages.
3628 NAME: visible_hostname
3630 LOC: Config.visibleHostname
3633 If you want to present a special hostname in error messages, etc,
3634 define this. Otherwise, the return value of gethostname()
3635 will be used. If you have multiple caches in a cluster and
3636 get errors about IP-forwarding you must set them to have individual
3637 names with this setting.
3640 NAME: unique_hostname
3642 LOC: Config.uniqueHostname
3645 If you want to have multiple machines with the same
3646 'visible_hostname' you must give each machine a different
3647 'unique_hostname' so forwarding loops can be detected.
3650 NAME: hostname_aliases
3652 LOC: Config.hostnameAliases
3655 A list of other DNS names your cache has.
3663 Minimum umask which should be enforced while the proxy
3664 is running, in addition to the umask set at startup.
3666 For a traditional octal representation of umasks, start
3671 OPTIONS FOR THE CACHE REGISTRATION SERVICE
3672 -----------------------------------------------------------------------------
3674 This section contains parameters for the (optional) cache
3675 announcement service. This service is provided to help
3676 cache administrators locate one another in order to join or
3677 create cache hierarchies.
3679 An 'announcement' message is sent (via UDP) to the registration
3680 service by Squid. By default, the announcement message is NOT
3681 SENT unless you enable it with 'announce_period' below.
3683 The announcement message includes your hostname, plus the
3684 following information from this configuration file:
3690 All current information is processed regularly and made
3691 available on the Web at http://www.ircache.net/Cache/Tracker/.
3694 NAME: announce_period
3696 LOC: Config.Announce.period
3699 This is how frequently to send cache announcements. The
3700 default is `0' which disables sending the announcement
3703 To enable announcing your cache, just uncomment the line
3707 #To enable announcing your cache, just uncomment the line below.
3708 #announce_period 1 day
3714 DEFAULT: tracker.ircache.net
3715 LOC: Config.Announce.host
3721 LOC: Config.Announce.file
3727 LOC: Config.Announce.port
3729 announce_host and announce_port set the hostname and port
3730 number where the registration message will be sent.
3732 Hostname will default to 'tracker.ircache.net' and port will
3733 default default to 3131. If the 'filename' argument is given,
3734 the contents of that file will be included in the announce
3739 HTTPD-ACCELERATOR OPTIONS
3740 -----------------------------------------------------------------------------
3743 NAME: httpd_accel_surrogate_id
3744 IFDEF: USE_SQUID_ESI
3746 LOC: Config.Accel.surrogate_id
3749 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
3750 need an identification token to allow control targeting. Because
3751 a farm of surrogates may all perform the same tasks, they may share
3752 an identification token.
3755 NAME: http_accel_surrogate_remote
3756 IFDEF: USE_SQUID_ESI
3760 LOC: Config.onoff.surrogate_is_remote
3762 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
3763 Set this to on to have squid behave as a remote surrogate.
3767 IFDEF: USE_SQUID_ESI
3768 COMMENT: libxml2|expat|custom
3770 LOC: ESIParser::Type
3773 ESI markup is not strictly XML compatible. The custom ESI parser
3774 will give higher performance, but cannot handle non ASCII character
3779 DELAY POOL PARAMETERS
3780 -----------------------------------------------------------------------------
3784 TYPE: delay_pool_count
3789 This represents the number of delay pools to be used. For example,
3790 if you have one class 2 delay pool and one class 3 delays pool, you
3791 have a total of 2 delay pools.
3795 TYPE: delay_pool_class
3800 This defines the class of each delay pool. There must be exactly one
3801 delay_class line for each delay pool. For example, to define two
3802 delay pools, one of class 2 and one of class 3, the settings above
3806 delay_pools 4 # 4 delay pools
3807 delay_class 1 2 # pool 1 is a class 2 pool
3808 delay_class 2 3 # pool 2 is a class 3 pool
3809 delay_class 3 4 # pool 3 is a class 4 pool
3810 delay_class 4 5 # pool 4 is a class 5 pool
3812 The delay pool classes are:
3814 class 1 Everything is limited by a single aggregate
3817 class 2 Everything is limited by a single aggregate
3818 bucket as well as an "individual" bucket chosen
3819 from bits 25 through 32 of the IP address.
3821 class 3 Everything is limited by a single aggregate
3822 bucket as well as a "network" bucket chosen
3823 from bits 17 through 24 of the IP address and a
3824 "individual" bucket chosen from bits 17 through
3825 32 of the IP address.
3827 class 4 Everything in a class 3 delay pool, with an
3828 additional limit on a per user basis. This
3829 only takes effect if the username is established
3830 in advance - by forcing authentication in your
3833 class 5 Requests are grouped according their tag (see
3834 external_acl's tag= reply).
3836 NOTE: If an IP address is a.b.c.d
3837 -> bits 25 through 32 are "d"
3838 -> bits 17 through 24 are "c"
3839 -> bits 17 through 32 are "c * 256 + d"
3843 TYPE: delay_pool_access
3848 This is used to determine which delay pool a request falls into.
3850 delay_access is sorted per pool and the matching starts with pool 1,
3851 then pool 2, ..., and finally pool N. The first delay pool where the
3852 request is allowed is selected for the request. If it does not allow
3853 the request to any pool then the request is not delayed (default).
3855 For example, if you want some_big_clients in delay
3856 pool 1 and lotsa_little_clients in delay pool 2:
3859 delay_access 1 allow some_big_clients
3860 delay_access 1 deny all
3861 delay_access 2 allow lotsa_little_clients
3862 delay_access 2 deny all
3863 delay_access 3 allow authenticated_clients
3866 NAME: delay_parameters
3867 TYPE: delay_pool_rates
3872 This defines the parameters for a delay pool. Each delay pool has
3873 a number of "buckets" associated with it, as explained in the
3874 description of delay_class. For a class 1 delay pool, the syntax is:
3876 delay_parameters pool aggregate
3878 For a class 2 delay pool:
3880 delay_parameters pool aggregate individual
3882 For a class 3 delay pool:
3884 delay_parameters pool aggregate network individual
3886 For a class 4 delay pool:
3888 delay_parameters pool aggregate network individual user
3890 For a class 5 delay pool:
3892 delay_parameters pool tag
3894 The variables here are:
3896 pool a pool number - ie, a number between 1 and the
3897 number specified in delay_pools as used in
3900 aggregate the "delay parameters" for the aggregate bucket
3903 individual the "delay parameters" for the individual
3904 buckets (class 2, 3).
3906 network the "delay parameters" for the network buckets
3909 user the delay parameters for the user buckets
3912 tag the delay parameters for the tag buckets
3915 A pair of delay parameters is written restore/maximum, where restore is
3916 the number of bytes (not bits - modem and network speeds are usually
3917 quoted in bits) per second placed into the bucket, and maximum is the
3918 maximum number of bytes which can be in the bucket at any time.
3920 For example, if delay pool number 1 is a class 2 delay pool as in the
3921 above example, and is being used to strictly limit each host to 64kbps
3922 (plus overheads), with no overall limit, the line is:
3924 delay_parameters 1 -1/-1 8000/8000
3926 Note that the figure -1 is used to represent "unlimited".
3928 And, if delay pool number 2 is a class 3 delay pool as in the above
3929 example, and you want to limit it to a total of 256kbps (strict limit)
3930 with each 8-bit network permitted 64kbps (strict limit) and each
3931 individual host permitted 4800bps with a bucket maximum size of 64kb
3932 to permit a decent web page to be downloaded at a decent speed
3933 (if the network is not being limited due to overuse) but slow down
3934 large downloads more significantly:
3936 delay_parameters 2 32000/32000 8000/8000 600/8000
3938 There must be one delay_parameters line for each delay pool.
3940 Finally, for a class 4 delay pool as in the example - each user will
3941 be limited to 128Kb no matter how many workstations they are logged into.:
3943 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
3946 NAME: delay_initial_bucket_level
3947 COMMENT: (percent, 0-100)
3951 LOC: Config.Delay.initial
3953 The initial bucket percentage is used to determine how much is put
3954 in each bucket when squid starts, is reconfigured, or first notices
3955 a host accessing it (in class 2 and class 3, individual hosts and
3956 networks only have buckets associated with them once they have been
3961 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
3962 -----------------------------------------------------------------------------
3967 LOC: Config.Wccp.router
3973 TYPE: IPAddress_list
3974 LOC: Config.Wccp2.router
3978 Use this option to define your WCCP ``home'' router for
3981 wccp_router supports a single WCCP(v1) router
3983 wccp2_router supports multiple WCCPv2 routers
3985 only one of the two may be used at the same time and defines
3986 which version of WCCP to use.
3991 LOC: Config.Wccp.version
3995 This directive is only relevant if you need to set up WCCP(v1)
3996 to some very old and end-of-life Cisco routers. In all other
3997 setups it must be left unset or at the default setting.
3998 It defines an internal version in the WCCP(v1) protocol,
3999 with version 4 being the officially documented protocol.
4001 According to some users, Cisco IOS 11.2 and earlier only
4002 support WCCP version 3. If you're using that or an earlier
4003 version of IOS, you may need to change this value to 3, otherwise
4004 do not specify this parameter.
4007 NAME: wccp2_rebuild_wait
4009 LOC: Config.Wccp2.rebuildwait
4013 If this is enabled Squid will wait for the cache dir rebuild to finish
4014 before sending the first wccp2 HereIAm packet
4017 NAME: wccp2_forwarding_method
4019 LOC: Config.Wccp2.forwarding_method
4023 WCCP2 allows the setting of forwarding methods between the
4024 router/switch and the cache. Valid values are as follows:
4026 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4027 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4029 Currently (as of IOS 12.4) cisco routers only support GRE.
4030 Cisco switches only support the L2 redirect assignment method.
4033 NAME: wccp2_return_method
4035 LOC: Config.Wccp2.return_method
4039 WCCP2 allows the setting of return methods between the
4040 router/switch and the cache for packets that the cache
4041 decides not to handle. Valid values are as follows:
4043 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4044 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4046 Currently (as of IOS 12.4) cisco routers only support GRE.
4047 Cisco switches only support the L2 redirect assignment.
4049 If the "ip wccp redirect exclude in" command has been
4050 enabled on the cache interface, then it is still safe for
4051 the proxy server to use a l2 redirect method even if this
4052 option is set to GRE.
4055 NAME: wccp2_assignment_method
4057 LOC: Config.Wccp2.assignment_method
4061 WCCP2 allows the setting of methods to assign the WCCP hash
4062 Valid values are as follows:
4067 As a general rule, cisco routers support the hash assignment method
4068 and cisco switches support the mask assignment method.
4073 LOC: Config.Wccp2.info
4075 DEFAULT_IF_NONE: standard 0
4078 WCCP2 allows for multiple traffic services. There are two
4079 types: "standard" and "dynamic". The standard type defines
4080 one service id - http (id 0). The dynamic service ids can be from
4081 51 to 255 inclusive. In order to use a dynamic service id
4082 one must define the type of traffic to be redirected; this is done
4083 using the wccp2_service_info option.
4085 The "standard" type does not require a wccp2_service_info option,
4086 just specifying the service id will suffice.
4088 MD5 service authentication can be enabled by adding
4089 "password=<password>" to the end of this service declaration.
4093 wccp2_service standard 0 # for the 'web-cache' standard service
4094 wccp2_service dynamic 80 # a dynamic service type which will be
4095 # fleshed out with subsequent options.
4096 wccp2_service standard 0 password=foo
4099 NAME: wccp2_service_info
4100 TYPE: wccp2_service_info
4101 LOC: Config.Wccp2.info
4105 Dynamic WCCPv2 services require further information to define the
4106 traffic you wish to have diverted.
4110 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
4111 priority=<priority> ports=<port>,<port>..
4113 The relevant WCCPv2 flags:
4114 + src_ip_hash, dst_ip_hash
4115 + source_port_hash, dst_port_hash
4116 + src_ip_alt_hash, dst_ip_alt_hash
4117 + src_port_alt_hash, dst_port_alt_hash
4120 The port list can be one to eight entries.
4124 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
4125 priority=240 ports=80
4127 Note: the service id must have been defined by a previous
4128 'wccp2_service dynamic <id>' entry.
4133 LOC: Config.Wccp2.weight
4137 Each cache server gets assigned a set of the destination
4138 hash proportional to their weight.
4143 LOC: Config.Wccp.address
4150 LOC: Config.Wccp2.address
4154 Use this option if you require WCCP to use a specific
4157 The default behavior is to not bind to any specific address.
4161 PERSISTENT CONNECTION HANDLING
4162 -----------------------------------------------------------------------------
4164 Also see "pconn_timeout" in the TIMEOUTS section
4167 NAME: client_persistent_connections
4169 LOC: Config.onoff.client_pconns
4173 NAME: server_persistent_connections
4175 LOC: Config.onoff.server_pconns
4178 Persistent connection support for clients and servers. By
4179 default, Squid uses persistent connections (when allowed)
4180 with its clients and servers. You can use these options to
4181 disable persistent connections with clients and/or servers.
4184 NAME: persistent_connection_after_error
4186 LOC: Config.onoff.error_pconns
4189 With this directive the use of persistent connections after
4190 HTTP errors can be disabled. Useful if you have clients
4191 who fail to handle errors on persistent connections proper.
4194 NAME: detect_broken_pconn
4196 LOC: Config.onoff.detect_broken_server_pconns
4199 Some servers have been found to incorrectly signal the use
4200 of HTTP/1.0 persistent connections even on replies not
4201 compatible, causing significant delays. This server problem
4202 has mostly been seen on redirects.
4204 By enabling this directive Squid attempts to detect such
4205 broken replies and automatically assume the reply is finished
4206 after 10 seconds timeout.
4210 CACHE DIGEST OPTIONS
4211 -----------------------------------------------------------------------------
4214 NAME: digest_generation
4215 IFDEF: USE_CACHE_DIGESTS
4217 LOC: Config.onoff.digest_generation
4220 This controls whether the server will generate a Cache Digest
4221 of its contents. By default, Cache Digest generation is
4222 enabled if Squid is compiled with --enable-cache-digests defined.
4225 NAME: digest_bits_per_entry
4226 IFDEF: USE_CACHE_DIGESTS
4228 LOC: Config.digest.bits_per_entry
4231 This is the number of bits of the server's Cache Digest which
4232 will be associated with the Digest entry for a given HTTP
4233 Method and URL (public key) combination. The default is 5.
4236 NAME: digest_rebuild_period
4237 IFDEF: USE_CACHE_DIGESTS
4240 LOC: Config.digest.rebuild_period
4243 This is the wait time between Cache Digest rebuilds.
4246 NAME: digest_rewrite_period
4248 IFDEF: USE_CACHE_DIGESTS
4250 LOC: Config.digest.rewrite_period
4253 This is the wait time between Cache Digest writes to
4257 NAME: digest_swapout_chunk_size
4260 IFDEF: USE_CACHE_DIGESTS
4261 LOC: Config.digest.swapout_chunk_size
4264 This is the number of bytes of the Cache Digest to write to
4265 disk at a time. It defaults to 4096 bytes (4KB), the Squid
4269 NAME: digest_rebuild_chunk_percentage
4270 COMMENT: (percent, 0-100)
4271 IFDEF: USE_CACHE_DIGESTS
4273 LOC: Config.digest.rebuild_chunk_percentage
4276 This is the percentage of the Cache Digest to be scanned at a
4277 time. By default it is set to 10% of the Cache Digest.
4282 -----------------------------------------------------------------------------
4287 LOC: Config.Port.snmp
4291 The port number where Squid listens for SNMP requests. To enable
4292 SNMP support set this to a suitable port number. Port number
4293 3401 is often used for the Squid SNMP agent. By default it's
4294 set to "0" (disabled)
4302 LOC: Config.accessList.snmp
4304 DEFAULT_IF_NONE: deny all
4307 Allowing or denying access to the SNMP port.
4309 All access to the agent is denied by default.
4312 snmp_access allow|deny [!]aclname ...
4315 snmp_access allow snmppublic localhost
4316 snmp_access deny all
4319 NAME: snmp_incoming_address
4321 LOC: Config.Addrs.snmp_incoming
4326 NAME: snmp_outgoing_address
4328 LOC: Config.Addrs.snmp_outgoing
4329 DEFAULT: 255.255.255.255
4332 Just like 'udp_incoming_address', but for the SNMP port.
4334 snmp_incoming_address is used for the SNMP socket receiving
4335 messages from SNMP agents.
4336 snmp_outgoing_address is used for SNMP packets returned to SNMP
4339 The default snmp_incoming_address (0.0.0.0) is to listen on all
4340 available network interfaces.
4342 If snmp_outgoing_address is set to 255.255.255.255 (the default)
4343 it will use the same socket as snmp_incoming_address. Only
4344 change this if you want to have SNMP replies sent using another
4345 address than where this Squid listens for SNMP queries.
4347 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
4348 the same value since they both use port 3401.
4353 -----------------------------------------------------------------------------
4356 NAME: icp_port udp_port
4359 LOC: Config.Port.icp
4361 The port number where Squid sends and receives ICP queries to
4362 and from neighbor caches. The standard UDP port for ICP is 3130.
4363 Default is disabled (0).
4365 #icp_port @DEFAULT_ICP_PORT@
4373 LOC: Config.Port.htcp
4375 The port number where Squid sends and receives HTCP queries to
4376 and from neighbor caches. To turn it on you want to set it to
4377 4827. By default it is set to "0" (disabled).
4383 NAME: log_icp_queries
4387 LOC: Config.onoff.log_udp
4389 If set, ICP queries are logged to access.log. You may wish
4390 do disable this if your ICP load is VERY high to speed things
4391 up or to simplify log analysis.
4394 NAME: udp_incoming_address
4396 LOC:Config.Addrs.udp_incoming
4399 udp_incoming_address is used for UDP packets received from other
4402 The default behavior is to not bind to any specific address.
4404 Only change this if you want to have all UDP queries received on
4405 a specific interface/address.
4407 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
4408 modules. Altering it will affect all of them in the same manner.
4410 see also; udp_outgoing_address
4412 NOTE, udp_incoming_address and udp_outgoing_address can not
4413 have the same value since they both use the same port.
4416 NAME: udp_outgoing_address
4418 LOC: Config.Addrs.udp_outgoing
4419 DEFAULT: 255.255.255.255
4421 udp_outgoing_address is used for UDP packets sent out to other
4424 The default behavior is to not bind to any specific address.
4426 Instead it will use the same socket as udp_incoming_address.
4427 Only change this if you want to have UDP queries sent using another
4428 address than where this Squid listens for UDP queries from other
4431 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
4432 modules. Altering it will affect all of them in the same manner.
4434 see also; udp_incoming_address
4436 NOTE, udp_incoming_address and udp_outgoing_address can not
4437 have the same value since they both use the same port.
4444 LOC: Config.onoff.icp_hit_stale
4446 If you want to return ICP_HIT for stale cache objects, set this
4447 option to 'on'. If you have sibling relationships with caches
4448 in other administrative domains, this should be 'off'. If you only
4449 have sibling relationships with caches under your control,
4450 it is probably okay to set this to 'on'.
4451 If set to 'on', your siblings should use the option "allow-miss"
4452 on their cache_peer lines for connecting to you.
4455 NAME: minimum_direct_hops
4458 LOC: Config.minDirectHops
4460 If using the ICMP pinging stuff, do direct fetches for sites
4461 which are no more than this many hops away.
4464 NAME: minimum_direct_rtt
4467 LOC: Config.minDirectRtt
4469 If using the ICMP pinging stuff, do direct fetches for sites
4470 which are no more than this many rtt milliseconds away.
4476 LOC: Config.Netdb.low
4482 LOC: Config.Netdb.high
4484 The low and high water marks for the ICMP measurement
4485 database. These are counts, not percents. The defaults are
4486 900 and 1000. When the high water mark is reached, database
4487 entries will be deleted until the low mark is reached.
4490 NAME: netdb_ping_period
4492 LOC: Config.Netdb.period
4495 The minimum period for measuring a site. There will be at
4496 least this much delay between successive pings to the same
4497 network. The default is five minutes.
4504 LOC: Config.onoff.query_icmp
4506 If you want to ask your peers to include ICMP data in their ICP
4507 replies, enable this option.
4509 If your peer has configured Squid (during compilation) with
4510 '--enable-icmp' that peer will send ICMP pings to origin server
4511 sites of the URLs it receives. If you enable this option the
4512 ICP replies from that peer will include the ICMP data (if available).
4513 Then, when choosing a parent cache, Squid will choose the parent with
4514 the minimal RTT to the origin server. When this happens, the
4515 hierarchy field of the access.log will be
4516 "CLOSEST_PARENT_MISS". This option is off by default.
4519 NAME: test_reachability
4523 LOC: Config.onoff.test_reachability
4525 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
4526 instead of ICP_MISS if the target host is NOT in the ICMP
4527 database, or has a zero RTT.
4530 NAME: icp_query_timeout
4534 LOC: Config.Timeout.icp_query
4536 Normally Squid will automatically determine an optimal ICP
4537 query timeout value based on the round-trip-time of recent ICP
4538 queries. If you want to override the value determined by
4539 Squid, set this 'icp_query_timeout' to a non-zero value. This
4540 value is specified in MILLISECONDS, so, to use a 2-second
4541 timeout (the old default), you would write:
4543 icp_query_timeout 2000
4546 NAME: maximum_icp_query_timeout
4550 LOC: Config.Timeout.icp_query_max
4552 Normally the ICP query timeout is determined dynamically. But
4553 sometimes it can lead to very large values (say 5 seconds).
4554 Use this option to put an upper limit on the dynamic timeout
4555 value. Do NOT use this option to always use a fixed (instead
4556 of a dynamic) timeout value. To set a fixed timeout see the
4557 'icp_query_timeout' directive.
4560 NAME: minimum_icp_query_timeout
4564 LOC: Config.Timeout.icp_query_min
4566 Normally the ICP query timeout is determined dynamically. But
4567 sometimes it can lead to very small timeouts, even lower than
4568 the normal latency variance on your link due to traffic.
4569 Use this option to put an lower limit on the dynamic timeout
4570 value. Do NOT use this option to always use a fixed (instead
4571 of a dynamic) timeout value. To set a fixed timeout see the
4572 'icp_query_timeout' directive.
4575 NAME: background_ping_rate
4579 LOC: Config.backgroundPingRate
4581 Controls how often the ICP pings are sent to siblings that
4582 have background-ping set.
4586 MULTICAST ICP OPTIONS
4587 -----------------------------------------------------------------------------
4592 LOC: Config.mcast_group_list
4595 This tag specifies a list of multicast groups which your server
4596 should join to receive multicasted ICP queries.
4598 NOTE! Be very careful what you put here! Be sure you
4599 understand the difference between an ICP _query_ and an ICP
4600 _reply_. This option is to be set only if you want to RECEIVE
4601 multicast queries. Do NOT set this option to SEND multicast
4602 ICP (use cache_peer for that). ICP replies are always sent via
4603 unicast, so this option does not affect whether or not you will
4604 receive replies from multicast group members.
4606 You must be very careful to NOT use a multicast address which
4607 is already in use by another group of caches.
4609 If you are unsure about multicast, please read the Multicast
4610 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
4612 Usage: mcast_groups 239.128.16.128 224.0.1.20
4614 By default, Squid doesn't listen on any multicast groups.
4617 NAME: mcast_miss_addr
4618 IFDEF: MULTICAST_MISS_STREAM
4620 LOC: Config.mcast_miss.addr
4621 DEFAULT: 255.255.255.255
4623 If you enable this option, every "cache miss" URL will
4624 be sent out on the specified multicast address.
4626 Do not enable this option unless you are are absolutely
4627 certain you understand what you are doing.
4630 NAME: mcast_miss_ttl
4631 IFDEF: MULTICAST_MISS_STREAM
4633 LOC: Config.mcast_miss.ttl
4636 This is the time-to-live value for packets multicasted
4637 when multicasting off cache miss URLs is enabled. By
4638 default this is set to 'site scope', i.e. 16.
4641 NAME: mcast_miss_port
4642 IFDEF: MULTICAST_MISS_STREAM
4644 LOC: Config.mcast_miss.port
4647 This is the port number to be used in conjunction with
4651 NAME: mcast_miss_encode_key
4652 IFDEF: MULTICAST_MISS_STREAM
4654 LOC: Config.mcast_miss.encode_key
4655 DEFAULT: XXXXXXXXXXXXXXXX
4657 The URLs that are sent in the multicast miss stream are
4658 encrypted. This is the encryption key.
4661 NAME: mcast_icp_query_timeout
4665 LOC: Config.Timeout.mcast_icp_query
4667 For multicast peers, Squid regularly sends out ICP "probes" to
4668 count how many other peers are listening on the given multicast
4669 address. This value specifies how long Squid should wait to
4670 count all the replies. The default is 2000 msec, or 2
4675 INTERNAL ICON OPTIONS
4676 -----------------------------------------------------------------------------
4679 NAME: icon_directory
4681 LOC: Config.icons.directory
4682 DEFAULT: @DEFAULT_ICON_DIR@
4684 Where the icons are stored. These are normally kept in
4688 NAME: global_internal_static
4690 LOC: Config.onoff.global_internal_static
4693 This directive controls is Squid should intercept all requests for
4694 /squid-internal-static/ no matter which host the URL is requesting
4695 (default on setting), or if nothing special should be done for
4696 such URLs (off setting). The purpose of this directive is to make
4697 icons etc work better in complex cache hierarchies where it may
4698 not always be possible for all corners in the cache mesh to reach
4699 the server generating a directory listing.
4702 NAME: short_icon_urls
4704 LOC: Config.icons.use_short_names
4707 If this is enabled Squid will use short URLs for icons.
4708 If disabled it will revert to the old behavior of including
4709 it's own name and port in the URL.
4711 If you run a complex cache hierarchy with a mix of Squid and
4712 other proxies you may need to disable this directive.
4717 -----------------------------------------------------------------------------
4720 NAME: error_directory
4722 LOC: Config.errorDirectory
4725 If you wish to create your own versions of the default
4726 error files to customize them to suit your company copy
4727 the error/template files to another directory and point
4730 WARNING: This option will disable multi-language support
4731 on error pages if used.
4733 The squid developers are interested in making squid available in
4734 a wide variety of languages. If you are making translations for a
4735 language that Squid does not currently provide please consider
4736 contributing your translation back to the project.
4737 http://wiki.squid-cache.org/Translations
4739 The squid developers working on translations are happy to supply drop-in
4740 translated error files in exchange for any new language contributions.
4743 NAME: error_default_language
4744 IFDEF: USE_ERR_LOCALES
4746 LOC: Config.errorDefaultLanguage
4749 Set the default language which squid will send error pages in
4750 if no existing translation matches the clients language
4753 If unset (default) generic English will be used.
4755 The squid developers are interested in making squid available in
4756 a wide variety of languages. If you are interested in making
4757 translations for any language see the squid wiki for details.
4758 http://wiki.squid-cache.org/Translations
4763 LOC: Config.errHtmlText
4766 HTML text to include in error messages. Make this a "mailto"
4767 URL to your admin address, or maybe just a link to your
4768 organizations Web page.
4770 To include this in your error messages, you must rewrite
4771 the error template files (found in the "errors" directory).
4772 Wherever you want the 'err_html_text' line to appear,
4773 insert a %L tag in the error template file.
4776 NAME: email_err_data
4779 LOC: Config.onoff.emailErrData
4782 If enabled, information about the occurred error will be
4783 included in the mailto links of the ERR pages (if %W is set)
4784 so that the email body contains the data.
4785 Syntax is <A HREF="mailto:%w%W">%w</A>
4790 LOC: Config.denyInfoList
4793 Usage: deny_info err_page_name acl
4794 or deny_info http://... acl
4795 or deny_info TCP_RESET acl
4797 This can be used to return a ERR_ page for requests which
4798 do not pass the 'http_access' rules. Squid remembers the last
4799 acl it evaluated in http_access, and if a 'deny_info' line exists
4800 for that ACL Squid returns a corresponding error page.
4802 The acl is typically the last acl on the http_access deny line which
4803 denied access. The exceptions to this rule are:
4804 - When Squid needs to request authentication credentials. It's then
4805 the first authentication related acl encountered
4806 - When none of the http_access lines matches. It's then the last
4807 acl processed on the last http_access line.
4809 NP: If providing your own custom error pages with error_directory
4810 you may also specify them by your custom file name:
4811 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
4813 Alternatively you can specify an error URL. The browsers will
4814 get redirected (302) to the specified URL. %s in the redirection
4815 URL will be replaced by the requested URL.
4817 Alternatively you can tell Squid to reset the TCP connection
4818 by specifying TCP_RESET.
4822 OPTIONS INFLUENCING REQUEST FORWARDING
4823 -----------------------------------------------------------------------------
4826 NAME: nonhierarchical_direct
4828 LOC: Config.onoff.nonhierarchical_direct
4831 By default, Squid will send any non-hierarchical requests
4832 (matching hierarchy_stoplist or not cacheable request type) direct
4835 If you set this to off, Squid will prefer to send these
4836 requests to parents.
4838 Note that in most configurations, by turning this off you will only
4839 add latency to these request without any improvement in global hit
4842 If you are inside an firewall see never_direct instead of
4848 LOC: Config.onoff.prefer_direct
4851 Normally Squid tries to use parents for most requests. If you for some
4852 reason like it to first try going direct and only use a parent if
4853 going direct fails set this to on.
4855 By combining nonhierarchical_direct off and prefer_direct on you
4856 can set up Squid to use a parent as a backup path if going direct
4859 Note: If you want Squid to use parents for all requests see
4860 the never_direct directive. prefer_direct only modifies how Squid
4861 acts on cacheable requests.
4866 LOC: Config.accessList.AlwaysDirect
4869 Usage: always_direct allow|deny [!]aclname ...
4871 Here you can use ACL elements to specify requests which should
4872 ALWAYS be forwarded by Squid to the origin servers without using
4873 any peers. For example, to always directly forward requests for
4874 local servers ignoring any parents or siblings you may have use
4877 acl local-servers dstdomain my.domain.net
4878 always_direct allow local-servers
4880 To always forward FTP requests directly, use
4883 always_direct allow FTP
4885 NOTE: There is a similar, but opposite option named
4886 'never_direct'. You need to be aware that "always_direct deny
4887 foo" is NOT the same thing as "never_direct allow foo". You
4888 may need to use a deny rule to exclude a more-specific case of
4889 some other rule. Example:
4891 acl local-external dstdomain external.foo.net
4892 acl local-servers dstdomain .foo.net
4893 always_direct deny local-external
4894 always_direct allow local-servers
4896 NOTE: If your goal is to make the client forward the request
4897 directly to the origin server bypassing Squid then this needs
4898 to be done in the client configuration. Squid configuration
4899 can only tell Squid how Squid should fetch the object.
4901 NOTE: This directive is not related to caching. The replies
4902 is cached as usual even if you use always_direct. To not cache
4903 the replies see no_cache.
4905 This option replaces some v1.1 options such as local_domain
4911 LOC: Config.accessList.NeverDirect
4914 Usage: never_direct allow|deny [!]aclname ...
4916 never_direct is the opposite of always_direct. Please read
4917 the description for always_direct if you have not already.
4919 With 'never_direct' you can use ACL elements to specify
4920 requests which should NEVER be forwarded directly to origin
4921 servers. For example, to force the use of a proxy for all
4922 requests, except those in your local domain use something like:
4924 acl local-servers dstdomain .foo.net
4925 never_direct deny local-servers
4926 never_direct allow all
4928 or if Squid is inside a firewall and there are local intranet
4929 servers inside the firewall use something like:
4931 acl local-intranet dstdomain .foo.net
4932 acl local-external dstdomain external.foo.net
4933 always_direct deny local-external
4934 always_direct allow local-intranet
4935 never_direct allow all
4937 This option replaces some v1.1 options such as inside_firewall
4942 ADVANCED NETWORKING OPTIONS
4943 -----------------------------------------------------------------------------
4946 NAME: incoming_icp_average
4949 LOC: Config.comm_incoming.icp_average
4952 NAME: incoming_http_average
4955 LOC: Config.comm_incoming.http_average
4958 NAME: incoming_dns_average
4961 LOC: Config.comm_incoming.dns_average
4964 NAME: min_icp_poll_cnt
4967 LOC: Config.comm_incoming.icp_min_poll
4970 NAME: min_dns_poll_cnt
4973 LOC: Config.comm_incoming.dns_min_poll
4976 NAME: min_http_poll_cnt
4979 LOC: Config.comm_incoming.http_min_poll
4981 Heavy voodoo here. I can't even believe you are reading this.
4982 Are you crazy? Don't even think about adjusting these unless
4983 you understand the algorithms in comm_select.c first!
4989 LOC: Config.accept_filter
4993 The name of an accept(2) filter to install on Squid's
4994 listen socket(s). This feature is perhaps specific to
4995 FreeBSD and requires support in the kernel.
4997 The 'httpready' filter delays delivering new connections
4998 to Squid until a full HTTP request has been received.
4999 See the accf_http(9) man page for details.
5001 The 'dataready' filter delays delivering new connections
5002 to Squid until there is some data to process.
5003 See the accf_dataready(9) man page for details.
5007 The 'data' filter delays delivering of new connections
5008 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
5009 You may optionally specify a number of seconds to wait by
5010 'data=N' where N is the number of seconds. Defaults to 30
5011 if not specified. See the tcp(7) man page for details.
5014 accept_filter httpready
5019 NAME: tcp_recv_bufsize
5023 LOC: Config.tcpRcvBufsz
5025 Size of receive buffer to set for TCP sockets. Probably just
5026 as easy to change your kernel's default. Set to zero to use
5027 the default buffer size.
5032 -----------------------------------------------------------------------------
5039 LOC: TheICAPConfig.onoff
5042 If you want to enable the ICAP module support, set this to on.
5045 NAME: icap_connect_timeout
5048 LOC: TheICAPConfig.connect_timeout_raw
5051 This parameter specifies how long to wait for the TCP connect to
5052 the requested ICAP server to complete before giving up and either
5053 terminating the HTTP transaction or bypassing the failure.
5055 The default for optional services is peer_connect_timeout.
5056 The default for essential services is connect_timeout.
5057 If this option is explicitly set, its value applies to all services.
5060 NAME: icap_io_timeout
5064 LOC: TheICAPConfig.io_timeout_raw
5067 This parameter specifies how long to wait for an I/O activity on
5068 an established, active ICAP connection before giving up and
5069 either terminating the HTTP transaction or bypassing the
5072 The default is read_timeout.
5075 NAME: icap_service_failure_limit
5078 LOC: TheICAPConfig.service_failure_limit
5081 The limit specifies the number of failures that Squid tolerates
5082 when establishing a new TCP connection with an ICAP service. If
5083 the number of failures exceeds the limit, the ICAP service is
5084 not used for new ICAP requests until it is time to refresh its
5085 OPTIONS. The per-service failure counter is reset to zero each
5086 time Squid fetches new service OPTIONS.
5088 A negative value disables the limit. Without the limit, an ICAP
5089 service will not be considered down due to connectivity failures
5090 between ICAP OPTIONS requests.
5093 NAME: icap_service_revival_delay
5096 LOC: TheICAPConfig.service_revival_delay
5099 The delay specifies the number of seconds to wait after an ICAP
5100 OPTIONS request failure before requesting the options again. The
5101 failed ICAP service is considered "down" until fresh OPTIONS are
5104 The actual delay cannot be smaller than the hardcoded minimum
5105 delay of 30 seconds.
5108 NAME: icap_preview_enable
5112 LOC: TheICAPConfig.preview_enable
5115 The ICAP Preview feature allows the ICAP server to handle the
5116 HTTP message by looking only at the beginning of the message body
5117 or even without receiving the body at all. In some environments,
5118 previews greatly speedup ICAP processing.
5120 During an ICAP OPTIONS transaction, the server may tell Squid what
5121 HTTP messages should be previewed and how big the preview should be.
5122 Squid will not use Preview if the server did not request one.
5124 To disable ICAP Preview for all ICAP services, regardless of
5125 individual ICAP server OPTIONS responses, set this option to "off".
5127 icap_preview_enable off
5130 NAME: icap_preview_size
5133 LOC: TheICAPConfig.preview_size
5136 The default size of preview data to be sent to the ICAP server.
5137 -1 means no preview. This value might be overwritten on a per server
5138 basis by OPTIONS requests.
5141 NAME: icap_default_options_ttl
5144 LOC: TheICAPConfig.default_options_ttl
5147 The default TTL value for ICAP OPTIONS responses that don't have
5148 an Options-TTL header.
5151 NAME: icap_persistent_connections
5155 LOC: TheICAPConfig.reuse_connections
5158 Whether or not Squid should use persistent connections to
5162 NAME: icap_send_client_ip
5166 LOC: TheICAPConfig.send_client_ip
5169 This adds the header "X-Client-IP" to ICAP requests.
5172 NAME: icap_send_client_username
5176 LOC: TheICAPConfig.send_client_username
5179 This sends authenticated HTTP client username (if available) to
5180 the ICAP service. The username value is encoded based on the
5181 icap_client_username_encode option and is sent using the header
5182 specified by the icap_client_username_header option.
5185 NAME: icap_client_username_header
5188 LOC: TheICAPConfig.client_username_header
5189 DEFAULT: X-Client-Username
5191 ICAP request header name to use for send_client_username.
5194 NAME: icap_client_username_encode
5198 LOC: TheICAPConfig.client_username_encode
5201 Whether to base64 encode the authenticated client username.
5205 TYPE: icap_service_type
5210 Defines a single ICAP service
5212 icap_service servicename vectoring_point bypass service_url
5214 vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
5215 This specifies at which point of transaction processing the
5216 ICAP service should be activated. *_postcache vectoring points
5217 are not yet supported.
5219 If set to 1, the ICAP service is treated as optional. If the
5220 service cannot be reached or malfunctions, Squid will try to
5221 ignore any errors and process the message as if the service
5222 was not enabled. No all ICAP errors can be bypassed.
5223 If set to 0, the ICAP service is treated as essential and all
5224 ICAP errors will result in an error page returned to the
5226 service_url = icap://servername:port/service
5229 icap_service service_1 reqmod_precache 0 icap://icap1.mydomain.net:1344/reqmod
5230 icap_service service_2 respmod_precache 0 icap://icap2.mydomain.net:1344/respmod
5234 TYPE: icap_class_type
5239 This depricated option was documented to define an ICAP service
5240 chain, even though it actually defined a set of similar, redundant
5241 services, and the chains were not supported.
5243 To define a set of redundant services, please use the
5244 adaptation_service_set directive.
5246 If you need adaptation service chains, patches or sponsorship
5251 TYPE: icap_access_type
5256 This option is depricated. Please use adaptation_access, which
5257 has the same ICAP functionality, but comes with better
5258 documentation, and eCAP support.
5263 -----------------------------------------------------------------------------
5266 NAME: loadable_modules
5268 IFDEF: USE_LOADABLE_MODULES
5269 LOC: Config.loadable_module_names
5272 Instructs Squid to load the specified dynamic module(s) or activate
5273 preloaded module(s).
5275 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
5279 MESSAGE ADAPTATION OPTIONS
5280 -----------------------------------------------------------------------------
5283 NAME: adaptation_service_set
5284 TYPE: adaptation_service_set_type
5285 IFDEF: USE_ADAPTATION
5290 Defines a named adaptation service set. The set is populated in
5291 the order of adaptation_service_set directives in this file.
5292 When adaptation ACLs are processed, the first and only the first
5293 applicable adaptation service from the set will be used. Thus,
5294 the set should group similar, redundant services, rather than a
5295 chain of complementary services.
5297 If you have a single adaptation service, you do not need to
5298 define a set containing it because adaptation_access accepts
5301 See also: adaptation_access
5304 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
5305 adaptation service_set svcLogger loggerLocal loggerRemote
5308 NAME: adaptation_access
5309 TYPE: adaptation_access_type
5310 IFDEF: USE_ADAPTATION
5314 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
5316 adaptation_access service_name allow|deny [!]aclname...
5317 adaptation_access set_name allow|deny [!]aclname...
5319 At each supported vectoring point, the adaptation_access
5320 statements are processed in the order they appear in this
5321 configuration file. Statements pointing to the following services
5322 are ignored (i.e., skipped without checking their ACL):
5324 - services serving different vectoring points
5325 - "broken-but-bypassable" services
5326 - "up" services configured to ignore such transactions
5327 (e.g., based on the ICAP Transfer-Ignore header).
5329 When a set_name is used, all services in the set are checked
5330 using the same rules, to find the first applicable one. See
5331 adaptation_service_set for details.
5333 If an access list is checked and there is a match, the
5334 processing stops: For an "allow" rule, the corresponding
5335 adaptation service is used for the transaction. For a "deny"
5336 rule, no adaptation service is activated.
5338 It is currently not possible to apply more than one adaptation
5339 service at the same vectoring point to the same HTTP transaction.
5341 See also: icap_service and ecap_service
5344 adaptation_access service_1 allow all
5349 -----------------------------------------------------------------------------
5352 NAME: check_hostnames
5355 LOC: Config.onoff.check_hostnames
5357 For security and stability reasons Squid can check
5358 hostnames for Internet standard RFC compliance. If you want
5359 Squid to perform these checks turn this directive on.
5362 NAME: allow_underscore
5365 LOC: Config.onoff.allow_underscore
5367 Underscore characters is not strictly allowed in Internet hostnames
5368 but nevertheless used by many sites. Set this to off if you want
5369 Squid to be strict about the standard.
5370 This check is performed only when check_hostnames is set to on.
5373 NAME: cache_dns_program
5375 IFDEF: USE_DNSSERVERS
5376 DEFAULT: @DEFAULT_DNSSERVER@
5377 LOC: Config.Program.dnsserver
5379 Specify the location of the executable for dnslookup process.
5384 IFDEF: USE_DNSSERVERS
5386 LOC: Config.dnsChildren
5388 The number of processes spawn to service DNS name lookups.
5389 For heavily loaded caches on large servers, you should
5390 probably increase this value to at least 10. The maximum
5391 is 32. The default is 5.
5393 You must have at least one dnsserver process.
5396 NAME: dns_retransmit_interval
5399 LOC: Config.Timeout.idns_retransmit
5400 IFDEF: !USE_DNSSERVERS
5402 Initial retransmit interval for DNS queries. The interval is
5403 doubled each time all configured DNS servers have been tried.
5410 LOC: Config.Timeout.idns_query
5411 IFDEF: !USE_DNSSERVERS
5413 DNS Query timeout. If no response is received to a DNS query
5414 within this time all DNS servers for the queried domain
5415 are assumed to be unavailable.
5422 LOC: Config.onoff.res_defnames
5424 Normally the RES_DEFNAMES resolver option is disabled
5425 (see res_init(3)). This prevents caches in a hierarchy
5426 from interpreting single-component hostnames locally. To allow
5427 Squid to handle single-component names, enable this option.
5430 NAME: dns_nameservers
5433 LOC: Config.dns_nameservers
5435 Use this if you want to specify a list of DNS name servers
5436 (IP addresses) to use instead of those given in your
5437 /etc/resolv.conf file.
5438 On Windows platforms, if no value is specified here or in
5439 the /etc/resolv.conf file, the list of DNS name servers are
5440 taken from the Windows registry, both static and dynamic DHCP
5441 configurations are supported.
5443 Example: dns_nameservers 10.0.0.1 192.172.0.4
5448 DEFAULT: @DEFAULT_HOSTS@
5449 LOC: Config.etcHostsPath
5451 Location of the host-local IP name-address associations
5452 database. Most Operating Systems have such a file on different
5454 - Un*X & Linux: /etc/hosts
5455 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
5456 (%SystemRoot% value install default is c:\winnt)
5457 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
5458 (%SystemRoot% value install default is c:\windows)
5459 - Windows 9x/Me: %windir%\hosts
5460 (%windir% value is usually c:\windows)
5461 - Cygwin: /etc/hosts
5463 The file contains newline-separated definitions, in the
5464 form ip_address_in_dotted_form name [name ...] names are
5465 whitespace-separated. Lines beginning with an hash (#)
5466 character are comments.
5468 The file is checked at startup and upon configuration.
5469 If set to 'none', it won't be checked.
5470 If append_domain is used, that domain will be added to
5471 domain-local (i.e. not containing any dot character) host
5477 LOC: Config.appendDomain
5480 Appends local domain name to hostnames without any dots in
5481 them. append_domain must begin with a period.
5483 Be warned there are now Internet names with no dots in
5484 them using only top-domain names, so setting this may
5485 cause some Internet sites to become unavailable.
5488 append_domain .yourdomain.com
5491 NAME: ignore_unknown_nameservers
5493 LOC: Config.onoff.ignore_unknown_nameservers
5496 By default Squid checks that DNS responses are received
5497 from the same IP addresses they are sent to. If they
5498 don't match, Squid ignores the response and writes a warning
5499 message to cache.log. You can allow responses from unknown
5500 nameservers by setting this option to 'off'.
5503 NAME: dns_v4_fallback
5506 LOC: Config.onoff.dns_require_A
5508 Standard practice with DNS is to lookup either A or AAAA records
5509 and use the results if it succeeds. Only looking up the other if
5510 the first attempt fails or otherwise produces no results.
5512 That policy however will cause squid to produce error pages for some
5513 servers that advertise AAAA but are unreachable over IPv6.
5515 If this is ON squid will always lookup both AAAA and A, using both.
5516 If this is OFF squid will lookup AAAA and only try A if none found.
5518 WARNING: There are some possibly unwanted side-effects with this on:
5519 *) Doubles the load placed by squid on the DNS network.
5520 *) May negatively impact connection delay times.
5524 COMMENT: (number of entries)
5527 LOC: Config.ipcache.size
5534 LOC: Config.ipcache.low
5541 LOC: Config.ipcache.high
5543 The size, low-, and high-water marks for the IP cache.
5546 NAME: fqdncache_size
5547 COMMENT: (number of entries)
5550 LOC: Config.fqdncache.size
5552 Maximum number of FQDN cache entries.
5557 -----------------------------------------------------------------------------
5564 LOC: Config.onoff.mem_pools
5566 If set, Squid will keep pools of allocated (but unused) memory
5567 available for future use. If memory is a premium on your
5568 system and you believe your malloc library outperforms Squid
5569 routines, disable this.
5572 NAME: memory_pools_limit
5576 LOC: Config.MemPools.limit
5578 Used only with memory_pools on:
5579 memory_pools_limit 50 MB
5581 If set to a non-zero value, Squid will keep at most the specified
5582 limit of allocated (but unused) memory in memory pools. All free()
5583 requests that exceed this limit will be handled by your malloc
5584 library. Squid does not pre-allocate any memory, just safe-keeps
5585 objects that otherwise would be free()d. Thus, it is safe to set
5586 memory_pools_limit to a reasonably high value even if your
5587 configuration will use less memory.
5589 If set to zero, Squid will keep all memory it can. That is, there
5590 will be no limit on the total amount of memory used for safe-keeping.
5592 To disable memory allocation optimization, do not set
5593 memory_pools_limit to 0. Set memory_pools to "off" instead.
5595 An overhead for maintaining memory pools is not taken into account
5596 when the limit is checked. This overhead is close to four bytes per
5597 object kept. However, pools may actually _save_ memory because of
5598 reduced memory thrashing in your malloc library.
5602 COMMENT: on|off|transparent|truncate|delete
5605 LOC: opt_forwarded_for
5607 If set to "on", Squid will append your client's IP address
5608 in the HTTP requests it forwards. By default it looks like:
5610 X-Forwarded-For: 192.1.2.3
5612 If set to "off", it will appear as
5614 X-Forwarded-For: unknown
5616 If set to "transparent", Squid will not alter the
5617 X-Forwarded-For header in any way.
5619 If set to "delete", Squid will delete the entire
5620 X-Forwarded-For header.
5622 If set to "truncate", Squid will remove all existing
5623 X-Forwarded-For entries, and place itself as the sole entry.
5626 NAME: cachemgr_passwd
5627 TYPE: cachemgrpasswd
5629 LOC: Config.passwd_list
5631 Specify passwords for cachemgr operations.
5633 Usage: cachemgr_passwd password action action ...
5635 Some valid actions are (see cache manager menu for a full list):
5675 * Indicates actions which will not be performed without a
5676 valid password, others can be performed if not listed here.
5678 To disable an action, set the password to "disable".
5679 To allow performing an action without a password, set the
5682 Use the keyword "all" to set the same password for all actions.
5685 cachemgr_passwd secret shutdown
5686 cachemgr_passwd lesssssssecret info stats/objects
5687 cachemgr_passwd disable all
5694 LOC: Config.onoff.client_db
5696 If you want to disable collecting per-client statistics,
5697 turn off client_db here.
5700 NAME: refresh_all_ims
5704 LOC: Config.onoff.refresh_all_ims
5706 When you enable this option, squid will always check
5707 the origin server for an update when a client sends an
5708 If-Modified-Since request. Many browsers use IMS
5709 requests when the user requests a reload, and this
5710 ensures those clients receive the latest version.
5712 By default (off), squid may return a Not Modified response
5713 based on the age of the cached version.
5716 NAME: reload_into_ims
5717 IFDEF: HTTP_VIOLATIONS
5721 LOC: Config.onoff.reload_into_ims
5723 When you enable this option, client no-cache or ``reload''
5724 requests will be changed to If-Modified-Since requests.
5725 Doing this VIOLATES the HTTP standard. Enabling this
5726 feature could make you liable for problems which it
5729 see also refresh_pattern for a more selective approach.
5732 NAME: maximum_single_addr_tries
5734 LOC: Config.retry.maxtries
5737 This sets the maximum number of connection attempts for a
5738 host that only has one address (for multiple-address hosts,
5739 each address is tried once).
5741 The default value is one attempt, the (not recommended)
5742 maximum is 255 tries. A warning message will be generated
5743 if it is set to a value greater than ten.
5745 Note: This is in addition to the request re-forwarding which
5746 takes place if Squid fails to get a satisfying response.
5749 NAME: retry_on_error
5751 LOC: Config.retry.onerror
5754 If set to on Squid will automatically retry requests when
5755 receiving an error response. This is mainly useful if you
5756 are in a complex cache hierarchy to work around access
5760 NAME: as_whois_server
5762 LOC: Config.as_whois_server
5763 DEFAULT: whois.ra.net
5764 DEFAULT_IF_NONE: whois.ra.net
5766 WHOIS server to query for AS numbers. NOTE: AS numbers are
5767 queried only when Squid starts up, not for every request.
5772 LOC: Config.onoff.offline
5775 Enable this option and Squid will never try to validate cached
5779 NAME: uri_whitespace
5780 TYPE: uri_whitespace
5781 LOC: Config.uri_whitespace
5784 What to do with requests that have whitespace characters in the
5787 strip: The whitespace characters are stripped out of the URL.
5788 This is the behavior recommended by RFC2396.
5789 deny: The request is denied. The user receives an "Invalid
5791 allow: The request is allowed and the URI is not changed. The
5792 whitespace characters remain in the URI. Note the
5793 whitespace is passed to redirector processes if they
5795 encode: The request is allowed and the whitespace characters are
5796 encoded according to RFC1738. This could be considered
5797 a violation of the HTTP/1.1
5798 RFC because proxies are not allowed to rewrite URI's.
5799 chop: The request is allowed and the URI is chopped at the
5800 first whitespace. This might also be considered a
5806 LOC: Config.coredump_dir
5808 DEFAULT_IF_NONE: none
5810 By default Squid leaves core files in the directory from where
5811 it was started. If you set 'coredump_dir' to a directory
5812 that exists, Squid will chdir() to that directory at startup
5813 and coredump files will be left there.
5816 # Leave coredumps in the first cache dir
5817 coredump_dir @DEFAULT_SWAP_DIR@
5823 LOC: Config.chroot_dir
5826 Use this to have Squid do a chroot() while initializing. This
5827 also causes Squid to fully drop root privileges after
5828 initializing. This means, for example, if you use a HTTP
5829 port less than 1024 and try to reconfigure, you will may get an
5830 error saying that Squid can not open the port.
5833 NAME: balance_on_multiple_ip
5835 LOC: Config.onoff.balance_on_multiple_ip
5838 Modern IP resolvers in squid sort lookup results by preferred access.
5839 By default squid will use these IP in order and only rotates to
5840 the next listed when the most preffered fails.
5842 Some load balancing servers based on round robin DNS have been
5843 found not to preserve user session state across requests
5844 to different IP addresses.
5846 Enabling this directive Squid rotates IP's per request.
5849 NAME: pipeline_prefetch
5851 LOC: Config.onoff.pipeline_prefetch
5854 To boost the performance of pipelined requests to closer
5855 match that of a non-proxied environment Squid can try to fetch
5856 up to two requests in parallel from a pipeline.
5858 Defaults to off for bandwidth management and access logging
5862 NAME: high_response_time_warning
5865 LOC: Config.warnings.high_rptm
5868 If the one-minute median response time exceeds this value,
5869 Squid prints a WARNING with debug level 0 to get the
5870 administrators attention. The value is in milliseconds.
5873 NAME: high_page_fault_warning
5875 LOC: Config.warnings.high_pf
5878 If the one-minute average page fault rate exceeds this
5879 value, Squid prints a WARNING with debug level 0 to get
5880 the administrators attention. The value is in page faults
5884 NAME: high_memory_warning
5886 LOC: Config.warnings.high_memory
5889 If the memory usage (as determined by mallinfo) exceeds
5890 this amount, Squid prints a WARNING with debug level 0 to get
5891 the administrators attention.
5894 NAME: sleep_after_fork
5895 COMMENT: (microseconds)
5897 LOC: Config.sleep_after_fork
5900 When this is set to a non-zero value, the main Squid process
5901 sleeps the specified number of microseconds after a fork()
5902 system call. This sleep may help the situation where your
5903 system reports fork() failures due to lack of (virtual)
5904 memory. Note, however, if you have a lot of child
5905 processes, these sleep delays will add up and your
5906 Squid will not service requests for some amount of time
5907 until all the child processes have been started.
5908 On Windows value less then 1000 (1 milliseconds) are
5912 NAME: windows_ipaddrchangemonitor
5916 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
5918 On Windows Squid by default will monitor IP address changes and will
5919 reconfigure itself after any detected event. This is very useful for
5920 proxies connected to internet with dial-up interfaces.
5921 In some cases (a Proxy server acting as VPN gateway is one) it could be
5922 desiderable to disable this behaviour setting this to 'off'.
5923 Note: after changing this, Squid service must be restarted.