2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
30 WELCOME TO SQUID @VERSION@
31 ----------------------------
33 This is the default Squid configuration file. You may wish
34 to look at the Squid home page (http://www.squid-cache.org/)
35 for the FAQ and other documentation.
37 The default Squid config file shows what the defaults for
38 various options happen to be. If you don't need to change the
39 default, you shouldn't uncomment the line. Doing so may cause
40 run-time problems. In some cases "none" refers to no default
41 setting at all, while in other cases it refers to a valid
42 option - the comments for that keyword indicate if this is the
48 Configuration options can be included using the "include" directive.
49 Include takes a list of files to include. Quoting and wildcards is
54 include /path/to/included/file/squid.acl.config
56 Includes can be nested up to a hard-coded depth of 16 levels.
57 This arbitrary restriction is to prevent recursive include references
58 from causing Squid entering an infinite loop whilst trying to load
63 OPTIONS FOR AUTHENTICATION
64 -----------------------------------------------------------------------------
69 LOC: Config.authConfiguration
72 This is used to define parameters for the various authentication
73 schemes supported by Squid.
75 format: auth_param scheme parameter [setting]
77 The order in which authentication schemes are presented to the client is
78 dependent on the order the scheme first appears in config file. IE
79 has a bug (it's not RFC 2617 compliant) in that it will use the basic
80 scheme if basic is the first entry presented, even if more secure
81 schemes are presented. For now use the order in the recommended
82 settings section below. If other browsers have difficulties (don't
83 recognize the schemes offered even if you are using basic) either
84 put basic first, or disable the other schemes (by commenting out their
87 Once an authentication scheme is fully configured, it can only be
88 shutdown by shutting squid down and restarting. Changes can be made on
89 the fly and activated with a reconfigure. I.E. You can change to a
90 different helper, but not unconfigure the helper completely.
92 Please note that while this directive defines how Squid processes
93 authentication it does not automatically activate authentication.
94 To use authentication you must in addition make use of ACLs based
95 on login name in http_access (proxy_auth, proxy_auth_regex or
96 external with %LOGIN used in the format tag). The browser will be
97 challenged for authentication on the first such acl encountered
98 in http_access processing and will also be re-challenged for new
99 login credentials if the request is being denied by a proxy_auth
102 WARNING: authentication can't be used in a transparently intercepting
103 proxy as the client then thinks it is talking to an origin server and
104 not the proxy. This is a limitation of bending the TCP/IP protocol to
105 transparently intercepting port 80, not a limitation in Squid.
106 Ports flagged 'transparent', 'intercept', or 'tproxy' have
107 authentication disabled.
109 === Parameters for the basic scheme follow. ===
112 Specify the command for the external authenticator. Such a program
113 reads a line containing "username password" and replies "OK" or
114 "ERR" in an endless loop. "ERR" responses may optionally be followed
115 by a error description available as %m in the returned error page.
116 If you use an authenticator, make sure you have 1 acl of type
119 By default, the basic authentication scheme is not used unless a
120 program is specified.
122 If you want to use the traditional NCSA proxy authentication, set
123 this line to something like
125 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
128 HTTP uses iso-latin-1 as characterset, while some authentication
129 backends such as LDAP expects UTF-8. If this is set to on Squid will
130 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
131 username & password to the helper.
133 "children" numberofchildren
134 The number of authenticator processes to spawn. If you start too few
135 Squid will have to wait for them to process a backlog of credential
136 verifications, slowing it down. When password verifications are
137 done via a (slow) network you are likely to need lots of
138 authenticator processes.
139 auth_param basic children 5
141 "concurrency" concurrency
142 The number of concurrent requests the helper can process.
143 The default of 0 is used for helpers who only supports
144 one request at a time. Setting this changes the protocol used to
145 include a channel number first on the request/response line, allowing
146 multiple requests to be sent to the same helper in parallell without
147 wating for the response.
148 Must not be set unless it's known the helper supports this.
149 auth_param basic concurrency 0
152 Specifies the realm name which is to be reported to the
153 client for the basic proxy authentication scheme (part of
154 the text the user will see when prompted their username and
155 password). There is no default.
156 auth_param basic realm Squid proxy-caching web server
158 "credentialsttl" timetolive
159 Specifies how long squid assumes an externally validated
160 username:password pair is valid for - in other words how
161 often the helper program is called for that user. Set this
162 low to force revalidation with short lived passwords. Note
163 setting this high does not impact your susceptibility
164 to replay attacks unless you are using an one-time password
165 system (such as SecureID). If you are using such a system,
166 you will be vulnerable to replay attacks unless you also
167 use the max_user_ip ACL in an http_access rule.
169 "casesensitive" on|off
170 Specifies if usernames are case sensitive. Most user databases are
171 case insensitive allowing the same username to be spelled using both
172 lower and upper case letters, but some are case sensitive. This
173 makes a big difference for user_max_ip ACL processing and similar.
174 auth_param basic casesensitive off
176 === Parameters for the digest scheme follow ===
179 Specify the command for the external authenticator. Such
180 a program reads a line containing "username":"realm" and
181 replies with the appropriate H(A1) value hex encoded or
182 ERR if the user (or his H(A1) hash) does not exists.
183 See rfc 2616 for the definition of H(A1).
184 "ERR" responses may optionally be followed by a error description
185 available as %m in the returned error page.
187 By default, the digest authentication scheme is not used unless a
188 program is specified.
190 If you want to use a digest authenticator, set this line to
193 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
196 HTTP uses iso-latin-1 as characterset, while some authentication
197 backends such as LDAP expects UTF-8. If this is set to on Squid will
198 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
199 username & password to the helper.
201 "children" numberofchildren
202 The number of authenticator processes to spawn (no default).
203 If you start too few Squid will have to wait for them to
204 process a backlog of H(A1) calculations, slowing it down.
205 When the H(A1) calculations are done via a (slow) network
206 you are likely to need lots of authenticator processes.
207 auth_param digest children 5
210 Specifies the realm name which is to be reported to the
211 client for the digest proxy authentication scheme (part of
212 the text the user will see when prompted their username and
213 password). There is no default.
214 auth_param digest realm Squid proxy-caching web server
216 "nonce_garbage_interval" timeinterval
217 Specifies the interval that nonces that have been issued
218 to client_agent's are checked for validity.
220 "nonce_max_duration" timeinterval
221 Specifies the maximum length of time a given nonce will be
224 "nonce_max_count" number
225 Specifies the maximum number of times a given nonce can be
228 "nonce_strictness" on|off
229 Determines if squid requires strict increment-by-1 behavior
230 for nonce counts, or just incrementing (off - for use when
231 useragents generate nonce counts that occasionally miss 1
232 (ie, 1,2,4,6)). Default off.
234 "check_nonce_count" on|off
235 This directive if set to off can disable the nonce count check
236 completely to work around buggy digest qop implementations in
237 certain mainstream browser versions. Default on to check the
238 nonce count to protect from authentication replay attacks.
240 "post_workaround" on|off
241 This is a workaround to certain buggy browsers who sends
242 an incorrect request digest in POST requests when reusing
243 the same nonce as acquired earlier on a GET request.
245 === NTLM scheme options follow ===
248 Specify the command for the external NTLM authenticator.
249 Such a program reads exchanged NTLMSSP packets with
250 the browser via Squid until authentication is completed.
251 If you use an NTLM authenticator, make sure you have 1 acl
252 of type proxy_auth. By default, the NTLM authenticator_program
255 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
257 "children" numberofchildren
258 The number of authenticator processes to spawn (no default).
259 If you start too few Squid will have to wait for them to
260 process a backlog of credential verifications, slowing it
261 down. When credential verifications are done via a (slow)
262 network you are likely to need lots of authenticator
265 auth_param ntlm children 5
268 If you experience problems with PUT/POST requests when using the
269 Negotiate authentication scheme then you can try setting this to
270 off. This will cause Squid to forcibly close the connection on
271 the initial requests where the browser asks which schemes are
272 supported by the proxy.
274 auth_param ntlm keep_alive on
276 === Options for configuring the NEGOTIATE auth-scheme follow ===
279 Specify the command for the external Negotiate authenticator.
280 This protocol is used in Microsoft Active-Directory enabled setups with
281 the Microsoft Internet Explorer or Mozilla Firefox browsers.
282 Its main purpose is to exchange credentials with the Squid proxy
283 using the Kerberos mechanisms.
284 If you use a Negotiate authenticator, make sure you have at least
285 one acl of type proxy_auth active. By default, the negotiate
286 authenticator_program is not used.
287 The only supported program for this role is the ntlm_auth
288 program distributed as part of Samba, version 4 or later.
290 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
292 "children" numberofchildren
293 The number of authenticator processes to spawn (no default).
294 If you start too few Squid will have to wait for them to
295 process a backlog of credential verifications, slowing it
296 down. When crendential verifications are done via a (slow)
297 network you are likely to need lots of authenticator
299 auth_param negotiate children 5
302 If you experience problems with PUT/POST requests when using the
303 Negotiate authentication scheme then you can try setting this to
304 off. This will cause Squid to forcibly close the connection on
305 the initial requests where the browser asks which schemes are
306 supported by the proxy.
308 auth_param negotiate keep_alive on
311 #Recommended minimum configuration per scheme:
312 #auth_param negotiate program <uncomment and complete this line to activate>
313 #auth_param negotiate children 5
314 #auth_param negotiate keep_alive on
315 #auth_param ntlm program <uncomment and complete this line to activate>
316 #auth_param ntlm children 5
317 #auth_param ntlm keep_alive on
318 #auth_param digest program <uncomment and complete this line>
319 #auth_param digest children 5
320 #auth_param digest realm Squid proxy-caching web server
321 #auth_param digest nonce_garbage_interval 5 minutes
322 #auth_param digest nonce_max_duration 30 minutes
323 #auth_param digest nonce_max_count 50
324 #auth_param basic program <uncomment and complete this line>
325 #auth_param basic children 5
326 #auth_param basic realm Squid proxy-caching web server
327 #auth_param basic credentialsttl 2 hours
331 NAME: authenticate_cache_garbage_interval
334 LOC: Config.authenticateGCInterval
336 The time period between garbage collection across the username cache.
337 This is a tradeoff between memory utilization (long intervals - say
338 2 days) and CPU (short intervals - say 1 minute). Only change if you
342 NAME: authenticate_ttl
345 LOC: Config.authenticateTTL
347 The time a user & their credentials stay in the logged in
348 user cache since their last request. When the garbage
349 interval passes, all user credentials that have passed their
350 TTL are removed from memory.
353 NAME: authenticate_ip_ttl
355 LOC: Config.authenticateIpTTL
358 If you use proxy authentication and the 'max_user_ip' ACL,
359 this directive controls how long Squid remembers the IP
360 addresses associated with each user. Use a small value
361 (e.g., 60 seconds) if your users might change addresses
362 quickly, as is the case with dialups. You might be safe
363 using a larger value (e.g., 2 hours) in a corporate LAN
364 environment with relatively static address assignments.
369 -----------------------------------------------------------------------------
372 NAME: external_acl_type
373 TYPE: externalAclHelper
374 LOC: Config.externalAclHelperList
377 This option defines external acl classes using a helper program
378 to look up the status
380 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
384 ttl=n TTL in seconds for cached results (defaults to 3600
387 TTL for cached negative lookups (default same
389 children=n Number of acl helper processes spawn to service
390 external acl lookups of this type. (default 5)
391 concurrency=n concurrency level per process. Only used with helpers
392 capable of processing more than one query at a time.
393 cache=n result cache size, 0 is unbounded (default)
394 grace=n Percentage remaining of TTL where a refresh of a
395 cached entry should be initiated without needing to
396 wait for a new reply. (default 0 for no grace period)
397 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
398 ipv4 / ipv6 IP-mode used to communicate to this helper.
399 For compatability with older configurations and helpers
400 the default is 'ipv4'.
402 FORMAT specifications
404 %LOGIN Authenticated user login name
405 %EXT_USER Username from external acl
406 %IDENT Ident user name
408 %SRCPORT Client source port
411 %PROTO Requested protocol
413 %PATH Requested URL path
414 %METHOD Request method
415 %MYADDR Squid interface address
416 %MYPORT Squid http_port number
417 %PATH Requested URL-path (including query-string if any)
418 %USER_CERT SSL User certificate in PEM format
419 %USER_CERTCHAIN SSL User certificate chain in PEM format
420 %USER_CERT_xx SSL User certificate subject attribute xx
421 %USER_CA_xx SSL User certificate issuer attribute xx
423 %>{Header} HTTP request header "Header"
425 HTTP request header "Hdr" list member "member"
427 HTTP request header list member using ; as
428 list separator. ; can be any non-alphanumeric
431 %<{Header} HTTP reply header "Header"
433 HTTP reply header "Hdr" list member "member"
435 HTTP reply header list member using ; as
436 list separator. ; can be any non-alphanumeric
439 In addition to the above, any string specified in the referencing
440 acl will also be included in the helper request line, after the
441 specified formats (see the "acl external" directive)
443 The helper receives lines per the above format specification,
444 and returns lines starting with OK or ERR indicating the validity
445 of the request and optionally followed by additional keywords with
448 General result syntax:
450 OK/ERR keyword=value ...
454 user= The users name (login)
455 password= The users password (for login= cache_peer option)
456 message= Message describing the reason. Available as %o
458 tag= Apply a tag to a request (for both ERR and OK results)
459 Only sets a tag, does not alter existing tags.
460 log= String to be logged in access.log. Available as
461 %ea in logformat specifications
463 If protocol=3.0 (the default) then URL escaping is used to protect
464 each value in both requests and responses.
466 If using protocol=2.5 then all values need to be enclosed in quotes
467 if they may contain whitespace, or the whitespace escaped using \.
468 And quotes or \ characters within the keyword value must be \ escaped.
470 When using the concurrency= option the protocol is changed by
471 introducing a query channel tag infront of the request/response.
472 The query channel tag is a number between 0 and concurrency-1.
480 Defining an Access List
482 Every access list definition must begin with an aclname and acltype,
483 followed by either type-specific arguments or a quoted filename that
486 acl aclname acltype argument ...
487 acl aclname acltype "file" ...
489 When using "file", the file should contain one item per line.
491 By default, regular expressions are CASE-SENSITIVE. To make
492 them case-insensitive, use the -i option.
494 Some acl types require suspending the current request in order
495 to access some external data source.
496 Those which do are marked with the tag [slow], those which
497 don't are marked as [fast].
498 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
499 for further information
501 ***** ACL TYPES AVAILABLE *****
503 acl aclname src ip-address/netmask ... # clients IP address [fast]
504 acl aclname src addr1-addr2/netmask ... # range of addresses [fast]
505 acl aclname dst ip-address/netmask ... # URL host's IP address [slow]
506 acl aclname myip ip-address/netmask ... # local socket IP address [fast]
508 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
509 # The arp ACL requires the special configure option --enable-arp-acl.
510 # Furthermore, the ARP ACL code is not portable to all operating systems.
511 # It works on Linux, Solaris, Windows, FreeBSD, and some
512 # other *BSD variants.
515 # NOTE: Squid can only determine the MAC address for clients that are on
516 # the same subnet. If the client is on a different subnet,
517 # then Squid cannot find out its MAC address.
519 acl aclname srcdomain .foo.com ...
520 # reverse lookup, from client IP [slow]
521 acl aclname dstdomain .foo.com ...
522 # Destination server from URL [fast]
523 acl aclname srcdom_regex [-i] \.foo\.com ...
524 # regex matching client name [slow]
525 acl aclname dstdom_regex [-i] \.foo\.com ...
526 # regex matching server [fast]
528 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
529 # based URL is used and no match is found. The name "none" is used
530 # if the reverse lookup fails.
532 acl aclname src_as number ...
533 acl aclname dst_as number ...
535 # Except for access control, AS numbers can be used for
536 # routing of requests to specific caches. Here's an
537 # example for routing all requests for AS#1241 and only
538 # those to mycache.mydomain.net:
539 # acl asexample dst_as 1241
540 # cache_peer_access mycache.mydomain.net allow asexample
541 # cache_peer_access mycache_mydomain.net deny all
543 acl aclname peername myPeer ...
545 # match against a named cache_peer entry
546 # set unique name= on cache_peer lines for reliable use.
548 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
558 # h1:m1 must be less than h2:m2
560 acl aclname url_regex [-i] ^http:// ...
561 # regex matching on whole URL [fast]
562 acl aclname urlpath_regex [-i] \.gif$ ...
563 # regex matching on URL path [fast]
565 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
567 acl aclname myport 3128 ... # local socket TCP port [fast]
568 acl aclname myportname 3128 ... # http(s)_port name [fast]
570 acl aclname proto HTTP FTP ... # request protocol [fast]
572 acl aclname method GET POST ... # HTTP request method [fast]
574 acl aclname http_status 200 301 500- 400-403 ...
575 # status code in reply [fast]
577 acl aclname browser [-i] regexp ...
578 # pattern match on User-Agent header (see also req_header below) [fast]
580 acl aclname referer_regex [-i] regexp ...
581 # pattern match on Referer header [fast]
582 # Referer is highly unreliable, so use with care
584 acl aclname ident username ...
585 acl aclname ident_regex [-i] pattern ...
586 # string match on ident output [slow]
587 # use REQUIRED to accept any non-null ident.
589 acl aclname proxy_auth [-i] username ...
590 acl aclname proxy_auth_regex [-i] pattern ...
591 # perform http authentication challenge to the client and match against
592 # supplied credentials [slow]
594 # takes a list of allowed usernames.
595 # use REQUIRED to accept any valid username.
597 # Will use proxy authentication in forward-proxy scenarios, and plain
598 # http authenticaiton in reverse-proxy scenarios
600 # NOTE: when a Proxy-Authentication header is sent but it is not
601 # needed during ACL checking the username is NOT logged
604 # NOTE: proxy_auth requires a EXTERNAL authentication program
605 # to check username/password combinations (see
606 # auth_param directive).
608 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
609 # as the browser needs to be configured for using a proxy in order
610 # to respond to proxy authentication.
612 acl aclname snmp_community string ...
613 # A community string to limit access to your SNMP Agent [fast]
616 # acl snmppublic snmp_community public
618 acl aclname maxconn number
619 # This will be matched when the client's IP address has
620 # more than <number> HTTP connections established. [fast]
622 acl aclname max_user_ip [-s] number
623 # This will be matched when the user attempts to log in from more
624 # than <number> different ip addresses. The authenticate_ip_ttl
625 # parameter controls the timeout on the ip entries. [fast]
626 # If -s is specified the limit is strict, denying browsing
627 # from any further IP addresses until the ttl has expired. Without
628 # -s Squid will just annoy the user by "randomly" denying requests.
629 # (the counter is reset each time the limit is reached and a
631 # NOTE: in acceleration mode or where there is mesh of child proxies,
632 # clients may appear to come from multiple addresses if they are
633 # going through proxy farms, so a limit of 1 may cause user problems.
635 acl aclname req_mime_type [-i] mime-type ...
636 # regex match against the mime type of the request generated
637 # by the client. Can be used to detect file upload or some
638 # types HTTP tunneling requests [fast]
639 # NOTE: This does NOT match the reply. You cannot use this
640 # to match the returned file type.
642 acl aclname req_header header-name [-i] any\.regex\.here
643 # regex match against any of the known request headers. May be
644 # thought of as a superset of "browser", "referer" and "mime-type"
647 acl aclname rep_mime_type [-i] mime-type ...
648 # regex match against the mime type of the reply received by
649 # squid. Can be used to detect file download or some
650 # types HTTP tunneling requests. [fast]
651 # NOTE: This has no effect in http_access rules. It only has
652 # effect in rules that affect the reply data stream such as
655 acl aclname rep_header header-name [-i] any\.regex\.here
656 # regex match against any of the known reply headers. May be
657 # thought of as a superset of "browser", "referer" and "mime-type"
660 acl aclname external class_name [arguments...]
661 # external ACL lookup via a helper class defined by the
662 # external_acl_type directive [slow]
664 acl aclname user_cert attribute values...
665 # match against attributes in a user SSL certificate
666 # attribute is one of DN/C/O/CN/L/ST [fast]
668 acl aclname ca_cert attribute values...
669 # match against attributes a users issuing CA SSL certificate
670 # attribute is one of DN/C/O/CN/L/ST [fast]
672 acl aclname ext_user username ...
673 acl aclname ext_user_regex [-i] pattern ...
674 # string match on username returned by external acl helper [slow]
675 # use REQUIRED to accept any non-null user name.
677 acl aclname tag tagvalue ...
678 # string match on tag returned by external acl helper [slow]
680 acl aclname hier_code codename ...
681 # string match against squid hierarchy code(s); [fast]
682 # e.g., DIRECT, PARENT_HIT, NONE, etc.
684 # NOTE: This has no effect in http_access rules. It only has
685 # effect in rules that affect the reply data stream such as
689 acl macaddress arp 09:00:2b:23:45:67
690 acl myexample dst_as 1241
691 acl password proxy_auth REQUIRED
692 acl fileupload req_mime_type -i ^multipart/form-data$
693 acl javascript rep_mime_type -i ^application/x-javascript$
696 #Recommended minimum configuration:
697 acl manager proto cache_object
698 acl localhost src 127.0.0.1/32
699 acl to_localhost dst 127.0.0.0/8 0.0.0.0/32
701 # Example rule allowing access from your local networks.
702 # Adapt to list your (internal) IP networks from where browsing
704 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
705 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
706 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
708 acl SSL_ports port 443
709 acl Safe_ports port 80 # http
710 acl Safe_ports port 21 # ftp
711 acl Safe_ports port 443 # https
712 acl Safe_ports port 70 # gopher
713 acl Safe_ports port 210 # wais
714 acl Safe_ports port 1025-65535 # unregistered ports
715 acl Safe_ports port 280 # http-mgmt
716 acl Safe_ports port 488 # gss-http
717 acl Safe_ports port 591 # filemaker
718 acl Safe_ports port 777 # multiling http
719 acl CONNECT method CONNECT
723 NAME: follow_x_forwarded_for
725 IFDEF: FOLLOW_X_FORWARDED_FOR
726 LOC: Config.accessList.followXFF
728 DEFAULT_IF_NONE: deny all
730 Allowing or Denying the X-Forwarded-For header to be followed to
731 find the original source of a request.
733 Requests may pass through a chain of several other proxies
734 before reaching us. The X-Forwarded-For header will contain a
735 comma-separated list of the IP addresses in the chain, with the
736 rightmost address being the most recent.
738 If a request reaches us from a source that is allowed by this
739 configuration item, then we consult the X-Forwarded-For header
740 to see where that host received the request from. If the
741 X-Forwarded-For header contains multiple addresses, and if
742 acl_uses_indirect_client is on, then we continue backtracking
743 until we reach an address for which we are not allowed to
744 follow the X-Forwarded-For header, or until we reach the first
745 address in the list. (If acl_uses_indirect_client is off, then
746 it's impossible to backtrack through more than one level of
747 X-Forwarded-For addresses.)
749 The end result of this process is an IP address that we will
750 refer to as the indirect client address. This address may
751 be treated as the client address for access control, delay
752 pools and logging, depending on the acl_uses_indirect_client,
753 delay_pool_uses_indirect_client and log_uses_indirect_client
756 This clause only supports fast acl types.
757 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
759 SECURITY CONSIDERATIONS:
761 Any host for which we follow the X-Forwarded-For header
762 can place incorrect information in the header, and Squid
763 will use the incorrect information as if it were the
764 source address of the request. This may enable remote
765 hosts to bypass any access control restrictions that are
766 based on the client's source addresses.
770 acl localhost src 127.0.0.1
771 acl my_other_proxy srcdomain .proxy.example.com
772 follow_x_forwarded_for allow localhost
773 follow_x_forwarded_for allow my_other_proxy
776 NAME: acl_uses_indirect_client
779 IFDEF: FOLLOW_X_FORWARDED_FOR
781 LOC: Config.onoff.acl_uses_indirect_client
783 Controls whether the indirect client address
784 (see follow_x_forwarded_for) is used instead of the
785 direct client address in acl matching.
788 NAME: delay_pool_uses_indirect_client
791 IFDEF: FOLLOW_X_FORWARDED_FOR DELAY_POOLS
793 LOC: Config.onoff.delay_pool_uses_indirect_client
795 Controls whether the indirect client address
796 (see follow_x_forwarded_for) is used instead of the
797 direct client address in delay pools.
800 NAME: log_uses_indirect_client
803 IFDEF: FOLLOW_X_FORWARDED_FOR
805 LOC: Config.onoff.log_uses_indirect_client
807 Controls whether the indirect client address
808 (see follow_x_forwarded_for) is used instead of the
809 direct client address in the access log.
814 LOC: Config.accessList.http
816 DEFAULT_IF_NONE: deny all
818 Allowing or Denying access based on defined access lists
820 Access to the HTTP port:
821 http_access allow|deny [!]aclname ...
823 NOTE on default values:
825 If there are no "access" lines present, the default is to deny
828 If none of the "access" lines cause a match, the default is the
829 opposite of the last line in the list. If the last line was
830 deny, the default is allow. Conversely, if the last line
831 is allow, the default will be deny. For these reasons, it is a
832 good idea to have an "deny all" or "allow all" entry at the end
833 of your access lists to avoid potential confusion.
835 This clause supports both fast and slow acl types.
836 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
839 #Recommended minimum configuration:
841 # Only allow cachemgr access from localhost
842 http_access allow manager localhost
843 http_access deny manager
844 # Deny requests to unknown ports
845 http_access deny !Safe_ports
846 # Deny CONNECT to other than SSL ports
847 http_access deny CONNECT !SSL_ports
849 # We strongly recommend the following be uncommented to protect innocent
850 # web applications running on the proxy server who think the only
851 # one who can access services on "localhost" is a local user
852 #http_access deny to_localhost
854 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
856 # Example rule allowing access from your local networks.
857 # Adapt localnet in the ACL section to list your (internal) IP networks
858 # from where browsing should be allowed
859 http_access allow localnet
860 http_access allow localhost
862 # And finally deny all other access to this proxy
867 NAME: http_reply_access
869 LOC: Config.accessList.reply
872 Allow replies to client requests. This is complementary to http_access.
874 http_reply_access allow|deny [!] aclname ...
876 NOTE: if there are no access lines present, the default is to allow
879 If none of the access lines cause a match the opposite of the
880 last line will apply. Thus it is good practice to end the rules
881 with an "allow all" or "deny all" entry.
883 This clause supports both fast and slow acl types.
884 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
889 LOC: Config.accessList.icp
891 DEFAULT_IF_NONE: deny all
893 Allowing or Denying access to the ICP port based on defined
896 icp_access allow|deny [!]aclname ...
898 See http_access for details
900 This clause only supports fast acl types.
901 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
903 #Allow ICP queries from local networks only
904 #icp_access allow localnet
912 LOC: Config.accessList.htcp
914 DEFAULT_IF_NONE: deny all
916 Allowing or Denying access to the HTCP port based on defined
919 htcp_access allow|deny [!]aclname ...
921 See http_access for details
923 NOTE: The default if no htcp_access lines are present is to
924 deny all traffic. This default may cause problems with peers
925 using the htcp or htcp-oldsquid options.
927 This clause only supports fast acl types.
928 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
930 #Allow HTCP queries from local networks only
931 #htcp_access allow localnet
932 #htcp_access deny all
936 NAME: htcp_clr_access
939 LOC: Config.accessList.htcp_clr
941 DEFAULT_IF_NONE: deny all
943 Allowing or Denying access to purge content using HTCP based
944 on defined access lists
946 htcp_clr_access allow|deny [!]aclname ...
948 See http_access for details
950 This clause only supports fast acl types.
951 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
952 #Allow HTCP CLR requests from trusted peers
953 acl htcp_clr_peer src 172.16.1.2
954 htcp_clr_access allow htcp_clr_peer
959 LOC: Config.accessList.miss
962 Use to force your neighbors to use you as a sibling instead of
963 a parent. For example:
965 acl localclients src 172.16.0.0/16
966 miss_access allow localclients
967 miss_access deny !localclients
969 This means only your local clients are allowed to fetch
970 MISSES and all other clients can only fetch HITS.
972 By default, allow all clients who passed the http_access rules
973 to fetch MISSES from us.
975 This clause only supports fast acl types.
976 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
979 NAME: ident_lookup_access
983 DEFAULT_IF_NONE: deny all
984 LOC: Ident::TheConfig.identLookup
986 A list of ACL elements which, if matched, cause an ident
987 (RFC 931) lookup to be performed for this request. For
988 example, you might choose to always perform ident lookups
989 for your main multi-user Unix boxes, but not for your Macs
990 and PCs. By default, ident lookups are not performed for
993 To enable ident lookups for specific client addresses, you
994 can follow this example:
996 acl ident_aware_hosts src 198.168.1.0/24
997 ident_lookup_access allow ident_aware_hosts
998 ident_lookup_access deny all
1000 Only src type ACL checks are fully supported. A srcdomain
1001 ACL might work at times, but it will not always provide
1004 This clause only supports fast acl types.
1005 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1008 NAME: reply_body_max_size
1009 COMMENT: size [acl acl...]
1012 LOC: Config.ReplyBodySize
1014 This option specifies the maximum size of a reply body. It can be
1015 used to prevent users from downloading very large files, such as
1016 MP3's and movies. When the reply headers are received, the
1017 reply_body_max_size lines are processed, and the first line where
1018 all (if any) listed ACLs are true is used as the maximum body size
1021 This size is checked twice. First when we get the reply headers,
1022 we check the content-length value. If the content length value exists
1023 and is larger than the allowed size, the request is denied and the
1024 user receives an error message that says "the request or reply
1025 is too large." If there is no content-length, and the reply
1026 size exceeds this limit, the client's connection is just closed
1027 and they will receive a partial reply.
1029 WARNING: downstream caches probably can not detect a partial reply
1030 if there is no content-length header, so they will cache
1031 partial responses and give them out as hits. You should NOT
1032 use this option if you have downstream caches.
1034 WARNING: A maximum size smaller than the size of squid's error messages
1035 will cause an infinite loop and crash squid. Ensure that the smallest
1036 non-zero value you use is greater that the maximum header size plus
1037 the size of your largest error page.
1039 If you set this parameter none (the default), there will be
1042 Configuration Format is:
1043 reply_body_max_size SIZE UNITS [acl ...]
1045 reply_body_max_size 10 MB
1051 -----------------------------------------------------------------------------
1054 NAME: http_port ascii_port
1055 TYPE: http_port_list
1057 LOC: Config.Sockaddr.http
1059 Usage: port [options]
1060 hostname:port [options]
1061 1.2.3.4:port [options]
1063 The socket addresses where Squid will listen for HTTP client
1064 requests. You may specify multiple socket addresses.
1065 There are three forms: port alone, hostname with port, and
1066 IP address with port. If you specify a hostname or IP
1067 address, Squid binds the socket to that specific
1068 address. This replaces the old 'tcp_incoming_address'
1069 option. Most likely, you do not need to bind to a specific
1070 address, so you can use the port number alone.
1072 If you are running Squid in accelerator mode, you
1073 probably want to listen on port 80 also, or instead.
1075 The -a command line option may be used to specify additional
1076 port(s) where Squid listens for proxy request. Such ports will
1077 be plain proxy ports with no options.
1079 You may specify multiple socket addresses on multiple lines.
1083 intercept Support for IP-Layer interception of
1084 outgoing requests without browser settings.
1085 NP: disables authentication and IPv6 on the port.
1087 tproxy Support Linux TPROXY for spoofing outgoing
1088 connections using the client IP address.
1089 NP: disables authentication and IPv6 on the port.
1091 accel Accelerator mode. Also needs at least one of
1092 vhost / vport / defaultsite.
1094 allow-direct Allow direct forwarding in accelerator mode. Normally
1095 accelerated requests are denied direct forwarding as if
1096 never_direct was used.
1098 defaultsite=domainname
1099 What to use for the Host: header if it is not present
1100 in a request. Determines what site (not origin server)
1101 accelerators should consider the default.
1104 vhost Accelerator mode using Host header for virtual
1105 domain support. Implies accel.
1107 vport Accelerator with IP based virtual host support.
1110 vport=NN As above, but uses specified port number rather
1111 than the http_port number. Implies accel.
1113 protocol= Protocol to reconstruct accelerated requests with.
1116 connection-auth[=on|off]
1117 use connection-auth=off to tell Squid to prevent
1118 forwarding Microsoft connection oriented authentication
1119 (NTLM, Negotiate and Kerberos)
1121 disable-pmtu-discovery=
1122 Control Path-MTU discovery usage:
1123 off lets OS decide on what to do (default).
1124 transparent disable PMTU discovery when transparent
1126 always disable always PMTU discovery.
1128 In many setups of transparently intercepting proxies
1129 Path-MTU discovery can not work on traffic towards the
1130 clients. This is the case when the intercepting device
1131 does not fully track connections and fails to forward
1132 ICMP must fragment messages to the cache server. If you
1133 have such setup and experience that certain clients
1134 sporadically hang or never complete requests set
1135 disable-pmtu-discovery option to 'transparent'.
1137 sslBump Intercept each CONNECT request matching ssl_bump ACL,
1138 establish secure connection with the client and with
1139 the server, decrypt HTTP messages as they pass through
1140 Squid, and treat them as unencrypted HTTP messages,
1141 becoming the man-in-the-middle.
1143 When this option is enabled, additional options become
1144 available to specify SSL-related properties of the
1145 client-side connection: cert, key, version, cipher,
1146 options, clientca, cafile, capath, crlfile, dhparams,
1147 sslflags, and sslcontext. See the https_port directive
1148 for more information on these options.
1150 The ssl_bump option is required to fully enable
1151 the SslBump feature.
1153 name= Specifies a internal name for the port. Defaults to
1154 the port specification (port or addr:port)
1156 tcpkeepalive[=idle,interval,timeout]
1157 Enable TCP keepalive probes of idle connections
1158 idle is the initial time before TCP starts probing
1159 the connection, interval how often to probe, and
1160 timeout the time before giving up.
1162 If you run Squid on a dual-homed machine with an internal
1163 and an external interface we recommend you to specify the
1164 internal address:port in http_port. This way Squid will only be
1165 visible on the internal address.
1168 # Squid normally listens to port 3128
1169 http_port @DEFAULT_HTTP_PORT@
1175 TYPE: https_port_list
1177 LOC: Config.Sockaddr.https
1179 Usage: [ip:]port cert=certificate.pem [key=key.pem] [options...]
1181 The socket address where Squid will listen for HTTPS client
1184 This is really only useful for situations where you are running
1185 squid in accelerator mode and you want to do the SSL work at the
1188 You may specify multiple socket addresses on multiple lines,
1189 each with their own SSL certificate and/or options.
1193 accel Accelerator mode. Also needs at least one of
1194 defaultsite or vhost.
1196 defaultsite= The name of the https site presented on
1197 this port. Implies accel.
1199 vhost Accelerator mode using Host header for virtual
1200 domain support. Requires a wildcard certificate
1201 or other certificate valid for more than one domain.
1204 protocol= Protocol to reconstruct accelerated requests with.
1207 cert= Path to SSL certificate (PEM format).
1209 key= Path to SSL private key file (PEM format)
1210 if not specified, the certificate file is
1211 assumed to be a combined certificate and
1214 version= The version of SSL/TLS supported
1215 1 automatic (default)
1220 cipher= Colon separated list of supported ciphers.
1222 options= Various SSL engine options. The most important
1224 NO_SSLv2 Disallow the use of SSLv2
1225 NO_SSLv3 Disallow the use of SSLv3
1226 NO_TLSv1 Disallow the use of TLSv1
1227 SINGLE_DH_USE Always create a new key when using
1228 temporary/ephemeral DH key exchanges
1229 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1230 documentation for a complete list of options.
1232 clientca= File containing the list of CAs to use when
1233 requesting a client certificate.
1235 cafile= File containing additional CA certificates to
1236 use when verifying client certificates. If unset
1237 clientca will be used.
1239 capath= Directory containing additional CA certificates
1240 and CRL lists to use when verifying client certificates.
1242 crlfile= File of additional CRL lists to use when verifying
1243 the client certificate, in addition to CRLs stored in
1244 the capath. Implies VERIFY_CRL flag below.
1246 dhparams= File containing DH parameters for temporary/ephemeral
1249 sslflags= Various flags modifying the use of SSL:
1251 Don't request client certificates
1252 immediately, but wait until acl processing
1253 requires a certificate (not yet implemented).
1255 Don't use the default CA lists built in
1258 Don't allow for session reuse. Each connection
1259 will result in a new SSL session.
1261 Verify CRL lists when accepting client
1264 Verify CRL lists for all certificates in the
1265 client certificate chain.
1267 sslcontext= SSL session ID context identifier.
1269 vport Accelerator with IP based virtual host support.
1271 vport=NN As above, but uses specified port number rather
1272 than the https_port number. Implies accel.
1274 name= Specifies a internal name for the port. Defaults to
1275 the port specification (port or addr:port)
1279 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1282 LOC: Config.accessList.outgoing_tos
1284 Allows you to select a TOS/Diffserv value to mark outgoing
1285 connections with, based on the username or source address
1288 tcp_outgoing_tos ds-field [!]aclname ...
1290 Example where normal_service_net uses the TOS value 0x00
1291 and good_service_net uses 0x20
1293 acl normal_service_net src 10.0.0.0/255.255.255.0
1294 acl good_service_net src 10.0.1.0/255.255.255.0
1295 tcp_outgoing_tos 0x00 normal_service_net
1296 tcp_outgoing_tos 0x20 good_service_net
1298 TOS/DSCP values really only have local significance - so you should
1299 know what you're specifying. For more information, see RFC2474,
1300 RFC2475, and RFC3260.
1302 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1303 "default" to use whatever default your host has. Note that in
1304 practice often only values 0 - 63 is usable as the two highest bits
1305 have been redefined for use by ECN (RFC3168).
1307 Processing proceeds in the order specified, and stops at first fully
1310 Note: The use of this directive using client dependent ACLs is
1311 incompatible with the use of server side persistent connections. To
1312 ensure correct results it is best to set server_persisten_connections
1313 to off when using this directive in such configurations.
1316 NAME: clientside_tos
1319 LOC: Config.accessList.clientside_tos
1321 Allows you to select a TOS/Diffserv value to mark client-side
1322 connections with, based on the username or source address
1332 Allows you to select a TOS/DSCP value to mark outgoing
1333 connections with, based on where the reply was sourced.
1335 TOS values really only have local significance - so you should
1336 know what you're specifying. For more information, see RFC2474,
1337 RFC2475, and RFC3260.
1339 The TOS/DSCP byte must be exactly that - octet value 0x00-0xFF.
1340 Note that in practice often only values up to 0x3F are usable
1341 as the two highest bits have been redefined for use by ECN
1344 This setting is configured by setting the source TOS values:
1346 local-hit=0xFF Value to mark local cache hits.
1348 sibling-hit=0xFF Value to mark hits from sibling peers.
1350 parent-hit=0xFF Value to mark hits from parent peers.
1353 NOTE: 'miss' preserve feature is only possible on Linux at this time.
1355 For the following to work correctly, you will need to patch your
1356 linux kernel with the TOS preserving ZPH patch.
1357 The kernel patch can be downloaded from http://zph.bratcheda.org
1359 disable-preserve-miss
1360 If set, any HTTP response towards clients will
1361 have the TOS value of the response comming from the
1362 remote server masked with the value of miss-mask.
1365 Allows you to mask certain bits in the TOS received from the
1366 remote server, before copying the value to the TOS sent
1368 Default: 0xFF (TOS from server is not changed).
1372 NAME: tcp_outgoing_address
1375 LOC: Config.accessList.outgoing_address
1377 Allows you to map requests to different outgoing IP addresses
1378 based on the username or source address of the user making
1381 tcp_outgoing_address ipaddr [[!]aclname] ...
1383 Example where requests from 10.0.0.0/24 will be forwarded
1384 with source address 10.1.0.1, 10.0.2.0/24 forwarded with
1385 source address 10.1.0.2 and the rest will be forwarded with
1386 source address 10.1.0.3.
1388 acl normal_service_net src 10.0.0.0/24
1389 acl good_service_net src 10.0.2.0/24
1390 tcp_outgoing_address 10.1.0.1 normal_service_net
1391 tcp_outgoing_address 10.1.0.2 good_service_net
1392 tcp_outgoing_address 10.1.0.3
1394 Processing proceeds in the order specified, and stops at first fully
1397 Note: The use of this directive using client dependent ACLs is
1398 incompatible with the use of server side persistent connections. To
1399 ensure correct results it is best to set server_persistent_connections
1400 to off when using this directive in such configurations.
1405 Squid is built with a capability of bridging the IPv4 and IPv6
1407 tcp_outgoing_address as exampled above breaks this bridging by forcing
1408 all outbound traffic through a certain IPv4 which may be on the wrong
1409 side of the IPv4/IPv6 boundary.
1411 To operate with tcp_outgoing_address and keep the bridging benefits
1412 an additional ACL needs to be used which ensures the IPv6-bound traffic
1413 is never forced or permitted out the IPv4 interface.
1415 acl to_ipv6 dst ipv6
1416 tcp_outgoing_address 2002::c001 good_service_net to_ipv6
1417 tcp_outgoing_address 10.1.0.2 good_service_net !to_ipv6
1419 tcp_outgoing_address 2002::beef normal_service_net to_ipv6
1420 tcp_outgoing_address 10.1.0.1 normal_service_net !to_ipv6
1422 tcp_outgoing_address 2002::1 to_ipv6
1423 tcp_outgoing_address 10.1.0.3 !to_ipv6
1426 'dst ipv6' bases its selection assuming DIRECT access.
1427 If peers are used the peername ACL are needed to select outgoing
1428 address which can link to the peer.
1430 'dst ipv6' is a slow ACL. It will only work here if 'dst' is used
1431 previously in the http_access rules to locate the destination IP.
1432 Some more magic may be needed for that:
1433 http_access allow to_ipv6 !all
1434 (meaning, allow if to IPv6 but not from anywhere ;)
1440 -----------------------------------------------------------------------------
1443 NAME: ssl_unclean_shutdown
1447 LOC: Config.SSL.unclean_shutdown
1449 Some browsers (especially MSIE) bugs out on SSL shutdown
1456 LOC: Config.SSL.ssl_engine
1459 The OpenSSL engine to use. You will need to set this if you
1460 would like to use hardware SSL acceleration for example.
1463 NAME: sslproxy_client_certificate
1466 LOC: Config.ssl_client.cert
1469 Client SSL Certificate to use when proxying https:// URLs
1472 NAME: sslproxy_client_key
1475 LOC: Config.ssl_client.key
1478 Client SSL Key to use when proxying https:// URLs
1481 NAME: sslproxy_version
1484 LOC: Config.ssl_client.version
1487 SSL version level to use when proxying https:// URLs
1490 NAME: sslproxy_options
1493 LOC: Config.ssl_client.options
1496 SSL engine options to use when proxying https:// URLs
1499 NAME: sslproxy_cipher
1502 LOC: Config.ssl_client.cipher
1505 SSL cipher list to use when proxying https:// URLs
1508 NAME: sslproxy_cafile
1511 LOC: Config.ssl_client.cafile
1514 file containing CA certificates to use when verifying server
1515 certificates while proxying https:// URLs
1518 NAME: sslproxy_capath
1521 LOC: Config.ssl_client.capath
1524 directory containing CA certificates to use when verifying
1525 server certificates while proxying https:// URLs
1531 LOC: Config.accessList.ssl_bump
1534 This ACL controls which CONNECT requests to an http_port
1535 marked with an sslBump flag are actually "bumped". Please
1536 see the sslBump flag of an http_port option for more details
1537 about decoding proxied SSL connections.
1539 By default, no requests are bumped.
1541 See also: http_port sslBump
1543 This clause only supports fast acl types.
1544 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1546 # Example: Bump all requests except those originating from localhost and
1547 # those going to webax.com or example.com sites.
1549 # acl localhost src 127.0.0.1/32
1550 # acl broken_sites dstdomain .webax.com
1551 # acl broken_sites dstdomain .example.com
1552 # ssl_bump deny localhost
1553 # ssl_bump deny broken_sites
1554 # ssl_bump allow all
1558 NAME: sslproxy_flags
1561 LOC: Config.ssl_client.flags
1564 Various flags modifying the use of SSL while proxying https:// URLs:
1565 DONT_VERIFY_PEER Accept certificates that fail verification.
1566 For refined control, see sslproxy_cert_error.
1567 NO_DEFAULT_CA Don't use the default CA list built in
1572 NAME: sslproxy_cert_error
1575 LOC: Config.ssl_client.cert_error
1578 Use this ACL to bypass server certificate validation errors.
1580 For example, the following lines will bypass all validation errors
1581 when talking to servers located at 172.16.0.0/16. All other
1582 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
1584 acl BrokenServersAtTrustedIP dst 172.16.0.0/16
1585 sslproxy_cert_error allow BrokenServersAtTrustedIP
1586 sslproxy_cert_error deny all
1588 This clause only supports fast acl types.
1589 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1590 Using slow acl types may result in server crashes
1592 Without this option, all server certificate validation errors
1593 terminate the transaction. Bypassing validation errors is dangerous
1594 because an error usually implies that the server cannot be trusted and
1595 the connection may be insecure.
1597 See also: sslproxy_flags and DONT_VERIFY_PEER.
1601 # sslproxy_cert_error deny all
1607 NAME: sslpassword_program
1610 LOC: Config.Program.ssl_password
1613 Specify a program used for entering SSL key passphrases
1614 when using encrypted SSL certificate keys. If not specified
1615 keys must either be unencrypted, or Squid started with the -N
1616 option to allow it to query interactively for the passphrase.
1620 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
1621 -----------------------------------------------------------------------------
1629 To specify other caches in a hierarchy, use the format:
1631 cache_peer hostname type http-port icp-port [options]
1636 # hostname type port port options
1637 # -------------------- -------- ----- ----- -----------
1638 cache_peer parent.foo.net parent 3128 3130 default
1639 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
1640 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
1641 cache_peer example.com parent 80 0 no-query default
1642 cache_peer cdn.example.com sibling 3128 0
1644 type: either 'parent', 'sibling', or 'multicast'.
1646 proxy-port: The port number where the peer accept HTTP requests.
1647 For other Squid proxies this is usually 3128
1648 For web servers this is usually 80
1650 icp-port: Used for querying neighbor caches about objects.
1651 Set to 0 if the peer does not support ICP or HTCP.
1652 See ICP and HTCP options below for additional details.
1655 ==== ICP OPTIONS ====
1657 You MUST also set icp_port and icp_access explicitly when using these options.
1658 The defaults will prevent peer traffic using ICP.
1661 no-query Disable ICP queries to this neighbor.
1664 Indicates the named peer is a member of a multicast group.
1665 ICP queries will not be sent directly to the peer, but ICP
1666 replies will be accepted from it.
1668 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
1669 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
1672 To only send ICP queries to this neighbor infrequently.
1673 This is used to keep the neighbor round trip time updated
1674 and is usually used in conjunction with weighted-round-robin.
1677 ==== HTCP OPTIONS ====
1679 You MUST also set htcp_port and htcp_access explicitly when using these options.
1680 The defaults will prevent peer traffic using HTCP.
1683 htcp Send HTCP, instead of ICP, queries to the neighbor.
1684 You probably also want to set the "icp-port" to 4827
1687 htcp-oldsquid Send HTCP to old Squid versions.
1689 htcp-no-clr Send HTCP to the neighbor but without
1690 sending any CLR requests. This cannot be used with
1693 htcp-only-clr Send HTCP to the neighbor but ONLY CLR requests.
1694 This cannot be used with htcp-no-clr.
1697 Send HTCP to the neighbor including CLRs but only when
1698 they do not result from PURGE requests.
1701 Forward any HTCP CLR requests this proxy receives to the peer.
1704 ==== PEER SELECTION METHODS ====
1706 The default peer selection method is ICP, with the first responding peer
1707 being used as source. These options can be used for better load balancing.
1710 default This is a parent cache which can be used as a "last-resort"
1711 if a peer cannot be located by any of the peer-selection methods.
1712 If specified more than once, only the first is used.
1714 round-robin Load-Balance parents which should be used in a round-robin
1715 fashion in the absence of any ICP queries.
1716 weight=N can be used to add bias.
1718 weighted-round-robin
1719 Load-Balance parents which should be used in a round-robin
1720 fashion with the frequency of each parent being based on the
1721 round trip time. Closer parents are used more often.
1722 Usually used for background-ping parents.
1723 weight=N can be used to add bias.
1725 carp Load-Balance parents which should be used as a CARP array.
1726 The requests will be distributed among the parents based on the
1727 CARP load balancing hash function based on their weight.
1729 userhash Load-balance parents based on the client proxy_auth or ident username.
1731 sourcehash Load-balance parents based on the client source IP.
1734 ==== PEER SELECTION OPTIONS ====
1736 weight=N use to affect the selection of a peer during any weighted
1737 peer-selection mechanisms.
1738 The weight must be an integer; default is 1,
1739 larger weights are favored more.
1740 This option does not affect parent selection if a peering
1741 protocol is not in use.
1743 basetime=N Specify a base amount to be subtracted from round trip
1745 It is subtracted before division by weight in calculating
1746 which parent to fectch from. If the rtt is less than the
1747 base time the rtt is set to a minimal value.
1749 ttl=N Specify a IP multicast TTL to use when sending an ICP
1750 queries to this address.
1751 Only useful when sending to a multicast group.
1752 Because we don't accept ICP replies from random
1753 hosts, you must configure other group members as
1754 peers with the 'multicast-responder' option.
1756 no-delay To prevent access to this neighbor from influencing the
1759 digest-url=URL Tell Squid to fetch the cache digest (if digests are
1760 enabled) for this host from the specified URL rather
1761 than the Squid default location.
1764 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
1766 originserver Causes this parent to be contacted as an origin server.
1767 Meant to be used in accelerator setups when the peer
1771 Set the Host header of requests forwarded to this peer.
1772 Useful in accelerator setups where the server (peer)
1773 expects a certain domain name but clients may request
1774 others. ie example.com or www.example.com
1776 no-digest Disable request of cache digests.
1779 Disables requesting ICMP RTT database (NetDB).
1782 ==== AUTHENTICATION OPTIONS ====
1785 If this is a personal/workgroup proxy and your parent
1786 requires proxy authentication.
1788 Note: The string can include URL escapes (i.e. %20 for
1789 spaces). This also means % must be written as %%.
1792 Send login details received from client to this peer.
1793 Both Proxy- and WWW-Authorization headers are passed
1794 without alteration to the peer.
1795 Authentication is not required by Squid for this to work.
1797 Note: This will pass any form of authentication but
1798 only Basic auth will work through a proxy unless the
1799 connection-auth options are also used.
1801 login=PASS Send login details received from client to this peer.
1802 Authentication is not required by this option.
1804 If there are no client-provided authentication headers
1805 to pass on, but username and password are available
1806 from an external ACL user= and password= result tags
1807 they may be sent instead.
1809 Note: To combine this with proxy_auth both proxies must
1810 share the same user database as HTTP only allows for
1811 a single login (one for proxy, one for origin server).
1812 Also be warned this will expose your users proxy
1813 password to the peer. USE WITH CAUTION
1816 Send the username to the upstream cache, but with a
1817 fixed password. This is meant to be used when the peer
1818 is in another administrative domain, but it is still
1819 needed to identify each user.
1820 The star can optionally be followed by some extra
1821 information which is added to the username. This can
1822 be used to identify this proxy to the peer, similar to
1823 the login=username:password option above.
1826 If this is a personal/workgroup proxy and your parent
1827 requires a secure proxy authentication.
1828 The first principal from the default keytab or defined by
1829 the environment variable KRB5_KTNAME will be used.
1831 login=NEGOTIATE:principal_name
1832 If this is a personal/workgroup proxy and your parent
1833 requires a secure proxy authentication.
1834 The principal principal_name from the default keytab or
1835 defined by the environment variable KRB5_KTNAME will be
1838 connection-auth=on|off
1839 Tell Squid that this peer does or not support Microsoft
1840 connection oriented authentication, and any such
1841 challenges received from there should be ignored.
1842 Default is auto to automatically determine the status
1846 ==== SSL / HTTPS / TLS OPTIONS ====
1848 ssl Encrypt connections to this peer with SSL/TLS.
1850 sslcert=/path/to/ssl/certificate
1851 A client SSL certificate to use when connecting to
1854 sslkey=/path/to/ssl/key
1855 The private SSL key corresponding to sslcert above.
1856 If 'sslkey' is not specified 'sslcert' is assumed to
1857 reference a combined file containing both the
1858 certificate and the key.
1861 The SSL version to use when connecting to this peer
1862 1 = automatic (default)
1867 sslcipher=... The list of valid SSL ciphers to use when connecting
1870 ssloptions=... Specify various SSL engine options:
1871 NO_SSLv2 Disallow the use of SSLv2
1872 NO_SSLv3 Disallow the use of SSLv3
1873 NO_TLSv1 Disallow the use of TLSv1
1874 See src/ssl_support.c or the OpenSSL documentation for
1875 a more complete list.
1877 sslcafile=... A file containing additional CA certificates to use
1878 when verifying the peer certificate.
1880 sslcapath=... A directory containing additional CA certificates to
1881 use when verifying the peer certificate.
1883 sslcrlfile=... A certificate revocation list file to use when
1884 verifying the peer certificate.
1886 sslflags=... Specify various flags modifying the SSL implementation:
1889 Accept certificates even if they fail to
1892 Don't use the default CA list built in
1895 Don't verify the peer certificate
1896 matches the server name
1898 ssldomain= The peer name as advertised in it's certificate.
1899 Used for verifying the correctness of the received peer
1900 certificate. If not specified the peer hostname will be
1904 Enable the "Front-End-Https: On" header needed when
1905 using Squid as a SSL frontend in front of Microsoft OWA.
1906 See MS KB document Q307347 for details on this header.
1907 If set to auto the header will only be added if the
1908 request is forwarded as a https:// URL.
1911 ==== GENERAL OPTIONS ====
1914 A peer-specific connect timeout.
1915 Also see the peer_connect_timeout directive.
1917 connect-fail-limit=N
1918 How many times connecting to a peer must fail before
1919 it is marked as down. Default is 10.
1921 allow-miss Disable Squid's use of only-if-cached when forwarding
1922 requests to siblings. This is primarily useful when
1923 icp_hit_stale is used by the sibling. To extensive use
1924 of this option may result in forwarding loops, and you
1925 should avoid having two-way peerings with this option.
1926 For example to deny peer usage on requests from peer
1927 by denying cache_peer_access if the source is a peer.
1929 max-conn=N Limit the amount of connections Squid may open to this
1932 name=xxx Unique name for the peer.
1933 Required if you have multiple peers on the same host
1934 but different ports.
1935 This name can be used in cache_peer_access and similar
1936 directives to dentify the peer.
1937 Can be used by outgoing access controls through the
1940 proxy-only objects fetched from the peer will not be stored locally.
1944 NAME: cache_peer_domain cache_host_domain
1949 Use to limit the domains for which a neighbor cache will be
1952 cache_peer_domain cache-host domain [domain ...]
1953 cache_peer_domain cache-host !domain
1955 For example, specifying
1957 cache_peer_domain parent.foo.net .edu
1959 has the effect such that UDP query packets are sent to
1960 'bigserver' only when the requested object exists on a
1961 server in the .edu domain. Prefixing the domainname
1962 with '!' means the cache will be queried for objects
1965 NOTE: * Any number of domains may be given for a cache-host,
1966 either on the same or separate lines.
1967 * When multiple domains are given for a particular
1968 cache-host, the first matched domain is applied.
1969 * Cache hosts with no domain restrictions are queried
1971 * There are no defaults.
1972 * There is also a 'cache_peer_access' tag in the ACL
1976 NAME: cache_peer_access
1981 Similar to 'cache_peer_domain' but provides more flexibility by
1984 cache_peer_access cache-host allow|deny [!]aclname ...
1986 The syntax is identical to 'http_access' and the other lists of
1987 ACL elements. See the comments for 'http_access' below, or
1988 the Squid FAQ (http://www.squid-cache.org/FAQ/FAQ-10.html).
1991 NAME: neighbor_type_domain
1992 TYPE: hostdomaintype
1996 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
1998 Modifying the neighbor type for specific domains is now
1999 possible. You can treat some domains differently than the the
2000 default neighbor type specified on the 'cache_peer' line.
2001 Normally it should only be necessary to list domains which
2002 should be treated differently because the default neighbor type
2003 applies for hostnames which do not match domains listed here.
2006 cache_peer cache.foo.org parent 3128 3130
2007 neighbor_type_domain cache.foo.org sibling .com .net
2008 neighbor_type_domain cache.foo.org sibling .au .de
2011 NAME: dead_peer_timeout
2015 LOC: Config.Timeout.deadPeer
2017 This controls how long Squid waits to declare a peer cache
2018 as "dead." If there are no ICP replies received in this
2019 amount of time, Squid will declare the peer dead and not
2020 expect to receive any further ICP replies. However, it
2021 continues to send ICP queries, and will mark the peer as
2022 alive upon receipt of the first subsequent ICP reply.
2024 This timeout also affects when Squid expects to receive ICP
2025 replies from peers. If more than 'dead_peer' seconds have
2026 passed since the last ICP reply was received, Squid will not
2027 expect to receive an ICP reply on the next query. Thus, if
2028 your time between requests is greater than this timeout, you
2029 will see a lot of requests sent DIRECT to origin servers
2030 instead of to your parents.
2033 NAME: forward_max_tries
2036 LOC: Config.forward_max_tries
2038 Controls how many different forward paths Squid will try
2039 before giving up. See also forward_timeout.
2042 NAME: hierarchy_stoplist
2045 LOC: Config.hierarchy_stoplist
2047 A list of words which, if found in a URL, cause the object to
2048 be handled directly by this cache. In other words, use this
2049 to not query neighbor caches for certain objects. You may
2050 list this option multiple times.
2051 Note: never_direct overrides this option.
2053 #We recommend you to use at least the following line.
2054 hierarchy_stoplist cgi-bin ?
2059 MEMORY CACHE OPTIONS
2060 -----------------------------------------------------------------------------
2067 LOC: Config.memMaxSize
2069 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2070 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2071 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2072 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2074 'cache_mem' specifies the ideal amount of memory to be used
2076 * In-Transit objects
2078 * Negative-Cached objects
2080 Data for these objects are stored in 4 KB blocks. This
2081 parameter specifies the ideal upper limit on the total size of
2082 4 KB blocks allocated. In-Transit objects take the highest
2085 In-transit objects have priority over the others. When
2086 additional space is needed for incoming data, negative-cached
2087 and hot objects will be released. In other words, the
2088 negative-cached and hot objects will fill up any unused space
2089 not needed for in-transit objects.
2091 If circumstances require, this limit will be exceeded.
2092 Specifically, if your incoming request rate requires more than
2093 'cache_mem' of memory to hold in-transit objects, Squid will
2094 exceed this limit to satisfy the new requests. When the load
2095 decreases, blocks will be freed until the high-water mark is
2096 reached. Thereafter, blocks will be used to store hot
2100 NAME: maximum_object_size_in_memory
2104 LOC: Config.Store.maxInMemObjSize
2106 Objects greater than this size will not be attempted to kept in
2107 the memory cache. This should be set high enough to keep objects
2108 accessed frequently in memory to improve performance whilst low
2109 enough to keep larger objects from hoarding cache_mem.
2112 NAME: memory_cache_mode
2117 Controls which objects to keep in the memory cache (cache_mem)
2119 always Keep most recently fetched objects in memory (default)
2121 disk Only disk cache hits are kept in memory, which means
2122 an object must first be cached on disk and then hit
2123 a second time before cached in memory.
2125 network Only objects fetched from network is kept in memory
2128 NAME: memory_replacement_policy
2130 LOC: Config.memPolicy
2133 The memory replacement policy parameter determines which
2134 objects are purged from memory when memory space is needed.
2136 See cache_replacement_policy for details.
2141 -----------------------------------------------------------------------------
2144 NAME: cache_replacement_policy
2146 LOC: Config.replPolicy
2149 The cache replacement policy parameter determines which
2150 objects are evicted (replaced) when disk space is needed.
2152 lru : Squid's original list based LRU policy
2153 heap GDSF : Greedy-Dual Size Frequency
2154 heap LFUDA: Least Frequently Used with Dynamic Aging
2155 heap LRU : LRU policy implemented using a heap
2157 Applies to any cache_dir lines listed below this.
2159 The LRU policies keeps recently referenced objects.
2161 The heap GDSF policy optimizes object hit rate by keeping smaller
2162 popular objects in cache so it has a better chance of getting a
2163 hit. It achieves a lower byte hit rate than LFUDA though since
2164 it evicts larger (possibly popular) objects.
2166 The heap LFUDA policy keeps popular objects in cache regardless of
2167 their size and thus optimizes byte hit rate at the expense of
2168 hit rate since one large, popular object will prevent many
2169 smaller, slightly less popular objects from being cached.
2171 Both policies utilize a dynamic aging mechanism that prevents
2172 cache pollution that can otherwise occur with frequency-based
2173 replacement policies.
2175 NOTE: if using the LFUDA replacement policy you should increase
2176 the value of maximum_object_size above its default of 4096 KB to
2177 to maximize the potential byte hit rate improvement of LFUDA.
2179 For more information about the GDSF and LFUDA cache replacement
2180 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
2181 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
2187 LOC: Config.cacheSwap
2191 cache_dir Type Directory-Name Fs-specific-data [options]
2193 You can specify multiple cache_dir lines to spread the
2194 cache among different disk partitions.
2196 Type specifies the kind of storage system to use. Only "ufs"
2197 is built by default. To enable any of the other storage systems
2198 see the --enable-storeio configure option.
2200 'Directory' is a top-level directory where cache swap
2201 files will be stored. If you want to use an entire disk
2202 for caching, this can be the mount-point directory.
2203 The directory must exist and be writable by the Squid
2204 process. Squid will NOT create this directory for you.
2208 "ufs" is the old well-known Squid storage format that has always
2211 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
2213 'Mbytes' is the amount of disk space (MB) to use under this
2214 directory. The default is 100 MB. Change this to suit your
2215 configuration. Do NOT put the size of your disk drive here.
2216 Instead, if you want Squid to use the entire disk drive,
2217 subtract 20% and use that value.
2219 'Level-1' is the number of first-level subdirectories which
2220 will be created under the 'Directory'. The default is 16.
2222 'Level-2' is the number of second-level subdirectories which
2223 will be created under each first-level directory. The default
2226 The aufs store type:
2228 "aufs" uses the same storage format as "ufs", utilizing
2229 POSIX-threads to avoid blocking the main Squid process on
2230 disk-I/O. This was formerly known in Squid as async-io.
2232 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
2234 see argument descriptions under ufs above
2236 The diskd store type:
2238 "diskd" uses the same storage format as "ufs", utilizing a
2239 separate process to avoid blocking the main Squid process on
2242 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
2244 see argument descriptions under ufs above
2246 Q1 specifies the number of unacknowledged I/O requests when Squid
2247 stops opening new files. If this many messages are in the queues,
2248 Squid won't open new files. Default is 64
2250 Q2 specifies the number of unacknowledged messages when Squid
2251 starts blocking. If this many messages are in the queues,
2252 Squid blocks until it receives some replies. Default is 72
2254 When Q1 < Q2 (the default), the cache directory is optimized
2255 for lower response time at the expense of a decrease in hit
2256 ratio. If Q1 > Q2, the cache directory is optimized for
2257 higher hit ratio at the expense of an increase in response
2260 The coss store type:
2262 NP: COSS filesystem in Squid-3 has been deemed too unstable for
2263 production use and has thus been removed from this release.
2264 We hope that it can be made usable again soon.
2266 block-size=n defines the "block size" for COSS cache_dir's.
2267 Squid uses file numbers as block numbers. Since file numbers
2268 are limited to 24 bits, the block size determines the maximum
2269 size of the COSS partition. The default is 512 bytes, which
2270 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
2271 you should not change the coss block size after Squid
2272 has written some objects to the cache_dir.
2274 The coss file store has changed from 2.5. Now it uses a file
2275 called 'stripe' in the directory names in the config - and
2276 this will be created by squid -z.
2280 no-store, no new objects should be stored to this cache_dir
2282 max-size=n, refers to the max object size this storedir supports.
2283 It is used to initially choose the storedir to dump the object.
2284 Note: To make optimal use of the max-size limits you should order
2285 the cache_dir lines with the smallest max-size value first and the
2286 ones with no max-size specification last.
2288 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
2289 which can be changed with the --with-coss-membuf-size=N configure
2292 # cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
2296 NAME: store_dir_select_algorithm
2298 LOC: Config.store_dir_select_algorithm
2301 Set this to 'round-robin' as an alternative.
2304 NAME: max_open_disk_fds
2306 LOC: Config.max_open_disk_fds
2309 To avoid having disk as the I/O bottleneck Squid can optionally
2310 bypass the on-disk cache if more than this amount of disk file
2311 descriptors are open.
2313 A value of 0 indicates no limit.
2316 NAME: minimum_object_size
2320 LOC: Config.Store.minObjectSize
2322 Objects smaller than this size will NOT be saved on disk. The
2323 value is specified in kilobytes, and the default is 0 KB, which
2324 means there is no minimum.
2327 NAME: maximum_object_size
2331 LOC: Config.Store.maxObjectSize
2333 Objects larger than this size will NOT be saved on disk. The
2334 value is specified in kilobytes, and the default is 4MB. If
2335 you wish to get a high BYTES hit ratio, you should probably
2336 increase this (one 32 MB object hit counts for 3200 10KB
2337 hits). If you wish to increase speed more than your want to
2338 save bandwidth you should leave this low.
2340 NOTE: if using the LFUDA replacement policy you should increase
2341 this value to maximize the byte hit rate improvement of LFUDA!
2342 See replacement_policy below for a discussion of this policy.
2345 NAME: cache_swap_low
2346 COMMENT: (percent, 0-100)
2349 LOC: Config.Swap.lowWaterMark
2352 NAME: cache_swap_high
2353 COMMENT: (percent, 0-100)
2356 LOC: Config.Swap.highWaterMark
2359 The low- and high-water marks for cache object replacement.
2360 Replacement begins when the swap (disk) usage is above the
2361 low-water mark and attempts to maintain utilization near the
2362 low-water mark. As swap utilization gets close to high-water
2363 mark object eviction becomes more aggressive. If utilization is
2364 close to the low-water mark less replacement is done each time.
2366 Defaults are 90% and 95%. If you have a large cache, 5% could be
2367 hundreds of MB. If this is the case you may wish to set these
2368 numbers closer together.
2373 -----------------------------------------------------------------------------
2378 LOC: Config.Log.logformats
2383 logformat <name> <format specification>
2385 Defines an access log format.
2387 The <format specification> is a string with embedded % format codes
2389 % format codes all follow the same basic structure where all but
2390 the formatcode is optional. Output strings are automatically escaped
2391 as required according to their context and the output format
2392 modifiers are usually not needed, but can be specified if an explicit
2393 output format is desired.
2395 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
2397 " output in quoted string format
2398 [ output in squid text log format as used by log_mime_hdrs
2399 # output in URL quoted format
2403 width field width. If starting with 0 the
2404 output is zero padded
2405 {arg} argument such as header name etc
2409 % a literal % character
2410 >a Client source IP address
2412 >p Client source port
2413 <A Server IP address or peer name
2414 la Local IP address (http_port)
2415 lp Local port number (http_port)
2416 sn Unique sequence number per log line entry
2417 ts Seconds since epoch
2418 tu subsecond time (milliseconds)
2419 tl Local time. Optional strftime format argument
2420 default %d/%b/%Y:%H:%M:%S %z
2421 tg GMT time. Optional strftime format argument
2422 default %d/%b/%Y:%H:%M:%S %z
2423 tr Response time (milliseconds)
2424 dt Total time spent making DNS lookups (milliseconds)
2426 HTTP cache related format codes:
2428 [http::]>h Request header. Optional header name argument
2429 on the format header[:[separator]element]
2430 [http::]<h Reply header. Optional header name argument
2432 [http::]un User name
2433 [http::]ul User name from authentication
2434 [http::]ui User name from ident
2435 [http::]us User name from SSL
2436 [http::]ue User name from external acl helper
2437 [http::]>Hs HTTP status code sent to the client
2438 [http::]<Hs HTTP status code received from the next hop
2439 [http::]Ss Squid request status (TCP_MISS etc)
2440 [http::]Sh Squid hierarchy status (DEFAULT_PARENT etc)
2441 [http::]mt MIME content type
2442 [http::]rm Request method (GET/POST etc)
2443 [http::]ru Request URL
2444 [http::]rp Request URL-Path excluding hostname
2445 [http::]rv Request protocol version
2446 [http::]et Tag returned by external acl
2447 [http::]ea Log string returned by external acl
2448 [http::]<st Sent reply size including HTTP headers
2449 [http::]>st Received request size including HTTP headers. In the
2450 case of chunked requests the chunked encoding metadata
2452 [http::]>sh Received HTTP request headers size
2453 [http::]<sh Sent HTTP reply headers size
2454 [http::]st Request+Reply size including HTTP headers
2455 [http::]<sH Reply high offset sent
2456 [http::]<sS Upstream object size
2457 [http::]<pt Peer response time in milliseconds. The timer starts
2458 when the last request byte is sent to the next hop
2459 and stops when the last response byte is received.
2460 [http::]<tt Total server-side time in milliseconds. The timer
2461 starts with the first connect request (or write I/O)
2462 sent to the first selected peer. The timer stops
2463 with the last I/O with the last peer.
2465 If ICAP is enabled, the following two codes become available (as
2466 well as ICAP log codes documented with the icap_log option):
2468 icap::tt Total ICAP processing time for the HTTP
2469 transaction. The timer ticks when ICAP
2470 ACLs are checked and when ICAP
2471 transaction is in progress.
2473 icap::<last_h The header of the last ICAP response
2474 related to the HTTP transaction. Like
2475 <h, accepts an optional header name
2476 argument. Will not change semantics
2477 when multiple ICAP transactions per HTTP
2478 transaction are supported.
2480 If adaptation is enabled the following two codes become available:
2482 adapt::sum_trs Summed adaptation transaction response
2483 times recorded as a comma-separated list in
2484 the order of transaction start time. Each time
2485 value is recorded as an integer number,
2486 representing response time of one or more
2487 adaptation (ICAP or eCAP) transaction in
2488 milliseconds. When a failed transaction is
2489 being retried or repeated, its time is not
2490 logged individually but added to the
2491 replacement (next) transaction. See also:
2494 adapt::all_trs All adaptation transaction response times.
2495 Same as adaptation_strs but response times of
2496 individual transactions are never added
2497 together. Instead, all transaction response
2498 times are recorded individually.
2500 You can prefix adapt::*_trs format codes with adaptation
2501 service name in curly braces to record response time(s) specific
2502 to that service. For example: %{my_service}adapt::sum_trs
2504 The default formats available (which do not need re-defining) are:
2506 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %un %Sh/%<A %mt
2507 logformat squidmime %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %un %Sh/%<A %mt [%>h] [%<h]
2508 logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
2509 logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
2512 NAME: access_log cache_access_log
2514 LOC: Config.Log.accesslogs
2516 DEFAULT_IF_NONE: @DEFAULT_ACCESS_LOG@ squid
2518 These files log client request activities. Has a line every HTTP or
2519 ICP request. The format is:
2520 access_log <filepath> [<logformat name> [acl acl ...]]
2521 access_log none [acl acl ...]]
2523 Will log to the specified file using the specified format (which
2524 must be defined in a logformat directive) those entries which match
2525 ALL the acl's specified (which must be defined in acl clauses).
2526 If no acl is specified, all requests will be logged to this file.
2528 To disable logging of a request use the filepath "none", in which case
2529 a logformat name should not be specified.
2531 To log the request via syslog specify a filepath of "syslog":
2533 access_log syslog[:facility.priority] [format [acl1 [acl2 ....]]]
2534 where facility could be any of:
2535 authpriv, daemon, local0 .. local7 or user.
2537 And priority could be any of:
2538 err, warning, notice, info, debug.
2541 access_log @DEFAULT_ACCESS_LOG@ squid
2547 LOC: Config.Log.icaplogs
2550 ICAP log files record ICAP transaction summaries, one line per
2553 The icap_log option format is:
2554 icap_log <filepath> [<logformat name> [acl acl ...]]
2555 icap_log none [acl acl ...]]
2557 Please see access_log option documentation for details. The two
2558 kinds of logs share the overall configuration approach and many
2561 ICAP processing of a single HTTP message or transaction may
2562 require multiple ICAP transactions. In such cases, multiple
2563 ICAP transaction log lines will correspond to a single access
2566 ICAP log uses logformat codes that make sense for an ICAP
2567 transaction. Header-related codes are applied to the HTTP header
2568 embedded in an ICAP server response, with the following caveats:
2569 For REQMOD, there is no HTTP response header unless the ICAP
2570 server performed request satisfaction. For RESPMOD, the HTTP
2571 request header is the header sent to the ICAP server. For
2572 OPTIONS, there are no HTTP headers.
2574 The following format codes are also available for ICAP logs:
2576 icap::<A ICAP server IP address. Similar to <A.
2578 icap::<service_name ICAP service name from the icap_service
2579 option in Squid configuration file.
2581 icap::ru ICAP Request-URI. Similar to ru.
2583 icap::rm ICAP request method (REQMOD, RESPMOD, or
2584 OPTIONS). Similar to existing rm.
2586 icap::>st Bytes sent to the ICAP server (TCP payload
2587 only; i.e., what Squid writes to the socket).
2589 icap::<st Bytes received from the ICAP server (TCP
2590 payload only; i.e., what Squid reads from
2593 icap::tr Transaction response time (in
2594 milliseconds). The timer starts when
2595 the ICAP transaction is created and
2596 stops when the transaction is completed.
2599 icap::tio Transaction I/O time (in milliseconds). The
2600 timer starts when the first ICAP request
2601 byte is scheduled for sending. The timers
2602 stops when the last byte of the ICAP response
2605 icap::to Transaction outcome: ICAP_ERR* for all
2606 transaction errors, ICAP_OPT for OPTION
2607 transactions, ICAP_ECHO for 204
2608 responses, ICAP_MOD for message
2609 modification, and ICAP_SAT for request
2610 satisfaction. Similar to Ss.
2612 icap::Hs ICAP response status code. Similar to Hs.
2614 icap::>h ICAP request header(s). Similar to >h.
2616 icap::<h ICAP response header(s). Similar to <h.
2618 The default ICAP log format, which can be used without an explicit
2619 definition, is called icap_squid:
2621 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
2623 See also: logformat, log_icap, and %icap::<last_h
2628 LOC: Config.accessList.log
2630 COMMENT: allow|deny acl acl...
2632 This options allows you to control which requests gets logged
2633 to access.log (see access_log directive). Requests denied for
2634 logging will also not be accounted for in performance counters.
2636 This clause only supports fast acl types.
2637 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2643 LOC: Config.accessList.icap
2646 This options allows you to control which requests get logged
2647 to icap.log. See the icap_log directive for ICAP log details.
2650 NAME: cache_store_log
2653 LOC: Config.Log.store
2655 Logs the activities of the storage manager. Shows which
2656 objects are ejected from the cache, and which objects are
2657 saved and for how long. To disable, enter "none" or remove the line.
2658 There are not really utilities to analyze this data, so you can safely
2661 # cache_store_log @DEFAULT_STORE_LOG@
2665 NAME: cache_swap_state cache_swap_log
2667 LOC: Config.Log.swap
2670 Location for the cache "swap.state" file. This index file holds
2671 the metadata of objects saved on disk. It is used to rebuild
2672 the cache during startup. Normally this file resides in each
2673 'cache_dir' directory, but you may specify an alternate
2674 pathname here. Note you must give a full filename, not just
2675 a directory. Since this is the index for the whole object
2676 list you CANNOT periodically rotate it!
2678 If %s can be used in the file name it will be replaced with a
2679 a representation of the cache_dir name where each / is replaced
2680 with '.'. This is needed to allow adding/removing cache_dir
2681 lines when cache_swap_log is being used.
2683 If have more than one 'cache_dir', and %s is not used in the name
2684 these swap logs will have names such as:
2690 The numbered extension (which is added automatically)
2691 corresponds to the order of the 'cache_dir' lines in this
2692 configuration file. If you change the order of the 'cache_dir'
2693 lines in this file, these index files will NOT correspond to
2694 the correct 'cache_dir' entry (unless you manually rename
2695 them). We recommend you do NOT use this option. It is
2696 better to keep these index files in each 'cache_dir' directory.
2699 NAME: logfile_rotate
2702 LOC: Config.Log.rotateNumber
2704 Specifies the number of logfile rotations to make when you
2705 type 'squid -k rotate'. The default is 10, which will rotate
2706 with extensions 0 through 9. Setting logfile_rotate to 0 will
2707 disable the file name rotation, but the logfiles are still closed
2708 and re-opened. This will enable you to rename the logfiles
2709 yourself just before sending the rotate signal.
2711 Note, the 'squid -k rotate' command normally sends a USR1
2712 signal to the running squid process. In certain situations
2713 (e.g. on Linux with Async I/O), USR1 is used for other
2714 purposes, so -k rotate uses another signal. It is best to get
2715 in the habit of using 'squid -k rotate' instead of 'kill -USR1
2718 Note, from Squid-3.1 this option has no effect on the cache.log,
2719 that log can be rotated separately by using debug_options
2722 NAME: emulate_httpd_log
2726 LOC: Config.onoff.common_log
2728 The Cache can emulate the log file format which many 'httpd'
2729 programs use. To disable/enable this emulation, set
2730 emulate_httpd_log to 'off' or 'on'. The default
2731 is to use the native log format since it includes useful
2732 information Squid-specific log analyzers use.
2735 NAME: log_ip_on_direct
2739 LOC: Config.onoff.log_ip_on_direct
2741 Log the destination IP address in the hierarchy log tag when going
2742 direct. Earlier Squid versions logged the hostname here. If you
2743 prefer the old way set this to off.
2748 DEFAULT: @DEFAULT_MIME_TABLE@
2749 LOC: Config.mimeTablePathname
2751 Pathname to Squid's MIME table. You shouldn't need to change
2752 this, but the default file contains examples and formatting
2753 information if you do.
2759 LOC: Config.onoff.log_mime_hdrs
2762 The Cache can record both the request and the response MIME
2763 headers for each HTTP transaction. The headers are encoded
2764 safely and will appear as two bracketed fields at the end of
2765 the access log (for either the native or httpd-emulated log
2766 formats). To enable this logging set log_mime_hdrs to 'on'.
2771 LOC: Config.Log.useragent
2773 IFDEF: USE_USERAGENT_LOG
2775 Squid will write the User-Agent field from HTTP requests
2776 to the filename specified here. By default useragent_log
2780 NAME: referer_log referrer_log
2782 LOC: Config.Log.referer
2784 IFDEF: USE_REFERER_LOG
2786 Squid will write the Referer field from HTTP requests to the
2787 filename specified here. By default referer_log is disabled.
2788 Note that "referer" is actually a misspelling of "referrer"
2789 however the misspelt version has been accepted into the HTTP RFCs
2795 DEFAULT: @DEFAULT_PID_FILE@
2796 LOC: Config.pidFilename
2798 A filename to write the process-id to. To disable, enter "none".
2805 LOC: Config.onoff.log_fqdn
2807 Turn this on if you wish to log fully qualified domain names
2808 in the access.log. To do this Squid does a DNS lookup of all
2809 IP's connecting to it. This can (in some situations) increase
2810 latency, which makes your cache seem slower for interactive
2814 NAME: client_netmask
2816 LOC: Config.Addrs.client_netmask
2817 DEFAULT: 255.255.255.255
2819 A netmask for client addresses in logfiles and cachemgr output.
2820 Change this to protect the privacy of your cache clients.
2821 A netmask of 255.255.255.0 will log all IP's in that range with
2822 the last digit set to '0'.
2829 LOC: Config.Log.forward
2831 Logs the server-side requests.
2833 This is currently work in progress.
2836 NAME: strip_query_terms
2838 LOC: Config.onoff.strip_query_terms
2841 By default, Squid strips query terms from requested URLs before
2842 logging. This protects your user's privacy.
2849 LOC: Config.onoff.buffered_logs
2851 cache.log log file is written with stdio functions, and as such
2852 it can be buffered or unbuffered. By default it will be unbuffered.
2853 Buffering it can speed up the writing slightly (though you are
2854 unlikely to need to worry unless you run with tons of debugging
2855 enabled in which case performance will suffer badly anyway..).
2858 NAME: netdb_filename
2860 DEFAULT: @DEFAULT_NETDB_FILE@
2861 LOC: Config.netdbFilename
2864 A filename where Squid stores it's netdb state between restarts.
2865 To disable, enter "none".
2869 OPTIONS FOR TROUBLESHOOTING
2870 -----------------------------------------------------------------------------
2876 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
2877 LOC: Debug::cache_log
2879 Cache logging file. This is where general information about
2880 your cache's behavior goes. You can increase the amount of data
2881 logged to this file and how often its rotated with "debug_options"
2887 LOC: Debug::debugOptions
2889 Logging options are set as section,level where each source file
2890 is assigned a unique section. Lower levels result in less
2891 output, Full debugging (level 9) can result in a very large
2892 log file, so be careful.
2894 The magic word "ALL" sets debugging levels for all sections.
2895 We recommend normally running with "ALL,1".
2897 The rotate=N option can be used to keep more or less of these logs
2898 than would otherwise be kept by logfile_rotate.
2899 For most uses a single log should be enough to monitor current
2900 events affecting Squid.
2905 LOC: Config.coredump_dir
2907 DEFAULT_IF_NONE: none
2909 By default Squid leaves core files in the directory from where
2910 it was started. If you set 'coredump_dir' to a directory
2911 that exists, Squid will chdir() to that directory at startup
2912 and coredump files will be left there.
2915 # Leave coredumps in the first cache dir
2916 coredump_dir @DEFAULT_SWAP_DIR@
2922 OPTIONS FOR FTP GATEWAYING
2923 -----------------------------------------------------------------------------
2929 LOC: Config.Ftp.anon_user
2931 If you want the anonymous login password to be more informative
2932 (and enable the use of picky ftp servers), set this to something
2933 reasonable for your domain, like wwwuser@somewhere.net
2935 The reason why this is domainless by default is the
2936 request can be made on the behalf of a user in any domain,
2937 depending on how the cache is used.
2938 Some ftp server also validate the email address is valid
2939 (for example perl.com).
2942 NAME: ftp_list_width
2945 LOC: Config.Ftp.list_width
2947 Sets the width of ftp listings. This should be set to fit in
2948 the width of a standard browser. Setting this too small
2949 can cut off long filenames when browsing ftp sites.
2955 LOC: Config.Ftp.passive
2957 If your firewall does not allow Squid to use passive
2958 connections, turn off this option.
2960 Use of ftp_epsv_all option requires this to be ON.
2966 LOC: Config.Ftp.epsv_all
2968 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
2970 NATs may be able to put the connection on a "fast path" through the
2971 translator, as the EPRT command will never be used and therefore,
2972 translation of the data portion of the segments will never be needed.
2974 When a client only expects to do two-way FTP transfers this may be
2976 If squid finds that it must do a three-way FTP transfer after issuing
2977 an EPSV ALL command, the FTP session will fail.
2979 If you have any doubts about this option do not use it.
2980 Squid will nicely attempt all other connection methods.
2982 Requires ftp_passive to be ON (default) for any effect.
2988 LOC: Config.Ftp.epsv
2990 FTP Protocol extensions permit the use of a special "EPSV" command.
2992 NATs may be able to put the connection on a "fast path" through the
2993 translator using EPSV, as the EPRT command will never be used
2994 and therefore, translation of the data portion of the segments
2995 will never be needed.
2997 Turning this OFF will prevent EPSV being attempted.
2998 WARNING: Doing so will convert Squid back to the old behavior with all
2999 the related problems with external NAT devices/layers.
3001 Requires ftp_passive to be ON (default) for any effect.
3004 NAME: ftp_sanitycheck
3007 LOC: Config.Ftp.sanitycheck
3009 For security and data integrity reasons Squid by default performs
3010 sanity checks of the addresses of FTP data connections ensure the
3011 data connection is to the requested server. If you need to allow
3012 FTP connections to servers using another IP address for the data
3013 connection turn this off.
3016 NAME: ftp_telnet_protocol
3019 LOC: Config.Ftp.telnet
3021 The FTP protocol is officially defined to use the telnet protocol
3022 as transport channel for the control connection. However, many
3023 implementations are broken and does not respect this aspect of
3026 If you have trouble accessing files with ASCII code 255 in the
3027 path or similar problems involving this ASCII code you can
3028 try setting this directive to off. If that helps, report to the
3029 operator of the FTP server in question that their FTP server
3030 is broken and does not follow the FTP standard.
3034 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
3035 -----------------------------------------------------------------------------
3040 DEFAULT: @DEFAULT_DISKD@
3041 LOC: Config.Program.diskd
3043 Specify the location of the diskd executable.
3044 Note this is only useful if you have compiled in
3045 diskd as one of the store io modules.
3048 NAME: unlinkd_program
3051 DEFAULT: @DEFAULT_UNLINKD@
3052 LOC: Config.Program.unlinkd
3054 Specify the location of the executable for file deletion process.
3057 NAME: pinger_program
3059 DEFAULT: @DEFAULT_PINGER@
3060 LOC: Config.pinger.program
3063 Specify the location of the executable for the pinger process.
3069 LOC: Config.pinger.enable
3072 Control whether the pinger is active at run-time.
3073 Enables turning ICMP pinger on and off with a simple
3074 squid -k reconfigure.
3079 OPTIONS FOR URL REWRITING
3080 -----------------------------------------------------------------------------
3083 NAME: url_rewrite_program redirect_program
3085 LOC: Config.Program.redirect
3088 Specify the location of the executable for the URL rewriter.
3089 Since they can perform almost any function there isn't one included.
3091 For each requested URL rewriter will receive on line with the format
3093 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
3095 In the future, the rewriter interface will be extended with
3096 key=value pairs ("kvpairs" shown above). Rewriter programs
3097 should be prepared to receive and possibly ignore additional
3098 whitespace-separated tokens on each input line.
3100 And the rewriter may return a rewritten URL. The other components of
3101 the request line does not need to be returned (ignored if they are).
3103 The rewriter can also indicate that a client-side redirect should
3104 be performed to the new URL. This is done by prefixing the returned
3105 URL with "301:" (moved permanently) or 302: (moved temporarily).
3107 By default, a URL rewriter is not used.
3110 NAME: url_rewrite_children redirect_children
3113 LOC: Config.redirectChildren
3115 The number of redirector processes to spawn. If you start
3116 too few Squid will have to wait for them to process a backlog of
3117 URLs, slowing it down. If you start too many they will use RAM
3118 and other system resources.
3121 NAME: url_rewrite_concurrency redirect_concurrency
3124 LOC: Config.redirectConcurrency
3126 The number of requests each redirector helper can handle in
3127 parallel. Defaults to 0 which indicates the redirector
3128 is a old-style single threaded redirector.
3130 When this directive is set to a value >= 1 then the protocol
3131 used to communicate with the helper is modified to include
3132 a request ID in front of the request/response. The request
3133 ID from the request must be echoed back with the response
3137 NAME: url_rewrite_host_header redirect_rewrites_host_header
3140 LOC: Config.onoff.redir_rewrites_host
3142 By default Squid rewrites any Host: header in redirected
3143 requests. If you are running an accelerator this may
3144 not be a wanted effect of a redirector.
3146 WARNING: Entries are cached on the result of the URL rewriting
3147 process, so be careful if you have domain-virtual hosts.
3150 NAME: url_rewrite_access redirector_access
3153 LOC: Config.accessList.redirector
3155 If defined, this access list specifies which requests are
3156 sent to the redirector processes. By default all requests
3159 This clause supports both fast and slow acl types.
3160 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3163 NAME: url_rewrite_bypass redirector_bypass
3165 LOC: Config.onoff.redirector_bypass
3168 When this is 'on', a request will not go through the
3169 redirector if all redirectors are busy. If this is 'off'
3170 and the redirector queue grows too large, Squid will exit
3171 with a FATAL error and ask you to increase the number of
3172 redirectors. You should only enable this if the redirectors
3173 are not critical to your caching system. If you use
3174 redirectors for access control, and you enable this option,
3175 users may have access to pages they should not
3176 be allowed to request.
3180 OPTIONS FOR TUNING THE CACHE
3181 -----------------------------------------------------------------------------
3184 NAME: cache no_cache
3187 LOC: Config.accessList.noCache
3189 A list of ACL elements which, if matched and denied, cause the request to
3190 not be satisfied from the cache and the reply to not be cached.
3191 In other words, use this to force certain objects to never be cached.
3193 You must use the words 'allow' or 'deny' to indicate whether items
3194 matching the ACL should be allowed or denied into the cache.
3196 Default is to allow all to be cached.
3198 This clause supports both fast and slow acl types.
3199 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3202 NAME: refresh_pattern
3203 TYPE: refreshpattern
3207 usage: refresh_pattern [-i] regex min percent max [options]
3209 By default, regular expressions are CASE-SENSITIVE. To make
3210 them case-insensitive, use the -i option.
3212 'Min' is the time (in minutes) an object without an explicit
3213 expiry time should be considered fresh. The recommended
3214 value is 0, any higher values may cause dynamic applications
3215 to be erroneously cached unless the application designer
3216 has taken the appropriate actions.
3218 'Percent' is a percentage of the objects age (time since last
3219 modification age) an object without explicit expiry time
3220 will be considered fresh.
3222 'Max' is an upper limit on how long objects without an explicit
3223 expiry time will be considered fresh.
3225 options: override-expire
3231 ignore-must-revalidate
3236 override-expire enforces min age even if the server
3237 sent an explicit expiry time (e.g., with the
3238 Expires: header or Cache-Control: max-age). Doing this
3239 VIOLATES the HTTP standard. Enabling this feature
3240 could make you liable for problems which it causes.
3242 Note: override-expire does not enforce staleness - it only extends
3243 freshness / min. If the server returns a Expires time which
3244 is longer than your max time, Squid will still consider
3245 the object fresh for that period of time.
3247 override-lastmod enforces min age even on objects
3248 that were modified recently.
3250 reload-into-ims changes client no-cache or ``reload''
3251 to If-Modified-Since requests. Doing this VIOLATES the
3252 HTTP standard. Enabling this feature could make you
3253 liable for problems which it causes.
3255 ignore-reload ignores a client no-cache or ``reload''
3256 header. Doing this VIOLATES the HTTP standard. Enabling
3257 this feature could make you liable for problems which
3260 ignore-no-cache ignores any ``Pragma: no-cache'' and
3261 ``Cache-control: no-cache'' headers received from a server.
3262 The HTTP RFC never allows the use of this (Pragma) header
3263 from a server, only a client, though plenty of servers
3266 ignore-no-store ignores any ``Cache-control: no-store''
3267 headers received from a server. Doing this VIOLATES
3268 the HTTP standard. Enabling this feature could make you
3269 liable for problems which it causes.
3271 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
3272 headers received from a server. Doing this VIOLATES
3273 the HTTP standard. Enabling this feature could make you
3274 liable for problems which it causes.
3276 ignore-private ignores any ``Cache-control: private''
3277 headers received from a server. Doing this VIOLATES
3278 the HTTP standard. Enabling this feature could make you
3279 liable for problems which it causes.
3281 ignore-auth caches responses to requests with authorization,
3282 as if the originserver had sent ``Cache-control: public''
3283 in the response header. Doing this VIOLATES the HTTP standard.
3284 Enabling this feature could make you liable for problems which
3287 refresh-ims causes squid to contact the origin server
3288 when a client issues an If-Modified-Since request. This
3289 ensures that the client will receive an updated version
3290 if one is available.
3292 Basically a cached object is:
3294 FRESH if expires < now, else STALE
3296 FRESH if lm-factor < percent, else STALE
3300 The refresh_pattern lines are checked in the order listed here.
3301 The first entry which matches is used. If none of the entries
3302 match the default will be used.
3304 Note, you must uncomment all the default lines if you want
3305 to change one. The default setting is only active if none is
3310 refresh_pattern ^ftp: 1440 20% 10080
3311 refresh_pattern ^gopher: 1440 0% 1440
3312 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
3313 refresh_pattern . 0 20% 4320
3317 NAME: quick_abort_min
3321 LOC: Config.quickAbort.min
3324 NAME: quick_abort_max
3328 LOC: Config.quickAbort.max
3331 NAME: quick_abort_pct
3335 LOC: Config.quickAbort.pct
3337 The cache by default continues downloading aborted requests
3338 which are almost completed (less than 16 KB remaining). This
3339 may be undesirable on slow (e.g. SLIP) links and/or very busy
3340 caches. Impatient users may tie up file descriptors and
3341 bandwidth by repeatedly requesting and immediately aborting
3344 When the user aborts a request, Squid will check the
3345 quick_abort values to the amount of data transfered until
3348 If the transfer has less than 'quick_abort_min' KB remaining,
3349 it will finish the retrieval.
3351 If the transfer has more than 'quick_abort_max' KB remaining,
3352 it will abort the retrieval.
3354 If more than 'quick_abort_pct' of the transfer has completed,
3355 it will finish the retrieval.
3357 If you do not want any retrieval to continue after the client
3358 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
3361 If you want retrievals to always continue if they are being
3362 cached set 'quick_abort_min' to '-1 KB'.
3365 NAME: read_ahead_gap
3366 COMMENT: buffer-size
3368 LOC: Config.readAheadGap
3371 The amount of data the cache will buffer ahead of what has been
3372 sent to the client when retrieving an object from another server.
3376 IFDEF: HTTP_VIOLATIONS
3379 LOC: Config.negativeTtl
3382 Set the Default Time-to-Live (TTL) for failed requests.
3383 Certain types of failures (such as "connection refused" and
3384 "404 Not Found") are able to be negatively-cached for a short time.
3385 Modern web servers should provide Expires: header, however if they
3386 do not this can provide a minimum TTL.
3387 The default is not to cache errors with unknown expiry details.
3389 Note that this is different from negative caching of DNS lookups.
3391 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3392 this feature could make you liable for problems which it
3396 NAME: positive_dns_ttl
3399 LOC: Config.positiveDnsTtl
3402 Upper limit on how long Squid will cache positive DNS responses.
3403 Default is 6 hours (360 minutes). This directive must be set
3404 larger than negative_dns_ttl.
3407 NAME: negative_dns_ttl
3410 LOC: Config.negativeDnsTtl
3413 Time-to-Live (TTL) for negative caching of failed DNS lookups.
3414 This also sets the lower cache limit on positive lookups.
3415 Minimum value is 1 second, and it is not recommendable to go
3416 much below 10 seconds.
3419 NAME: range_offset_limit
3422 LOC: Config.rangeOffsetLimit
3425 Sets a upper limit on how far into the the file a Range request
3426 may be to cause Squid to prefetch the whole file. If beyond this
3427 limit Squid forwards the Range request as it is and the result
3430 This is to stop a far ahead range request (lets say start at 17MB)
3431 from making Squid fetch the whole object up to that point before
3432 sending anything to the client.
3434 A value of 0 causes Squid to never fetch more than the
3435 client requested. (default)
3437 A value of -1 causes Squid to always fetch the object from the
3438 beginning so it may cache the result. (2.0 style)
3440 NP: Using -1 here will override any quick_abort settings that may
3441 otherwise apply to the range request. The range request will
3442 be fully fetched from start to finish regardless of the client
3443 actions. This affects bandwidth usage.
3446 NAME: minimum_expiry_time
3449 LOC: Config.minimum_expiry_time
3452 The minimum caching time according to (Expires - Date)
3453 Headers Squid honors if the object can't be revalidated
3454 defaults to 60 seconds. In reverse proxy environments it
3455 might be desirable to honor shorter object lifetimes. It
3456 is most likely better to make your server return a
3457 meaningful Last-Modified header however. In ESI environments
3458 where page fragments often have short lifetimes, this will
3459 often be best set to 0.
3462 NAME: store_avg_object_size
3466 LOC: Config.Store.avgObjectSize
3468 Average object size, used to estimate number of objects your
3469 cache can hold. The default is 13 KB.
3472 NAME: store_objects_per_bucket
3475 LOC: Config.Store.objectsPerBucket
3477 Target number of objects per bucket in the store hash table.
3478 Lowering this value increases the total number of buckets and
3479 also the storage maintenance rate. The default is 20.
3484 -----------------------------------------------------------------------------
3487 NAME: request_header_max_size
3491 LOC: Config.maxRequestHeaderSize
3493 This specifies the maximum size for HTTP headers in a request.
3494 Request headers are usually relatively small (about 512 bytes).
3495 Placing a limit on the request header size will catch certain
3496 bugs (for example with persistent connections) and possibly
3497 buffer-overflow or denial-of-service attacks.
3500 NAME: reply_header_max_size
3504 LOC: Config.maxReplyHeaderSize
3506 This specifies the maximum size for HTTP headers in a reply.
3507 Reply headers are usually relatively small (about 512 bytes).
3508 Placing a limit on the reply header size will catch certain
3509 bugs (for example with persistent connections) and possibly
3510 buffer-overflow or denial-of-service attacks.
3513 NAME: request_body_max_size
3517 LOC: Config.maxRequestBodySize
3519 This specifies the maximum size for an HTTP request body.
3520 In other words, the maximum size of a PUT/POST request.
3521 A user who attempts to send a request with a body larger
3522 than this limit receives an "Invalid Request" error message.
3523 If you set this parameter to a zero (the default), there will
3524 be no limit imposed.
3527 NAME: chunked_request_body_max_size
3531 LOC: Config.maxChunkedRequestBodySize
3533 A broken or confused HTTP/1.1 client may send a chunked HTTP
3534 request to Squid. Squid does not have full support for that
3535 feature yet. To cope with such requests, Squid buffers the
3536 entire request and then dechunks request body to create a
3537 plain HTTP/1.0 request with a known content length. The plain
3538 request is then used by the rest of Squid code as usual.
3540 The option value specifies the maximum size of the buffer used
3541 to hold the request before the conversion. If the chunked
3542 request size exceeds the specified limit, the conversion
3543 fails, and the client receives an "unsupported request" error,
3544 as if dechunking was disabled.
3546 Dechunking is enabled by default. To disable conversion of
3547 chunked requests, set the maximum to zero.
3549 Request dechunking feature and this option in particular are a
3550 temporary hack. When chunking requests and responses are fully
3551 supported, there will be no need to buffer a chunked request.
3555 IFDEF: HTTP_VIOLATIONS
3558 LOC: Config.accessList.brokenPosts
3560 A list of ACL elements which, if matched, causes Squid to send
3561 an extra CRLF pair after the body of a PUT/POST request.
3563 Some HTTP servers has broken implementations of PUT/POST,
3564 and rely on an extra CRLF pair sent by some WWW clients.
3566 Quote from RFC2616 section 4.1 on this matter:
3568 Note: certain buggy HTTP/1.0 client implementations generate an
3569 extra CRLF's after a POST request. To restate what is explicitly
3570 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
3571 a request with an extra CRLF.
3573 This clause only supports fast acl types.
3574 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3577 acl buggy_server url_regex ^http://....
3578 broken_posts allow buggy_server
3582 IFDEF: HTTP_VIOLATIONS
3586 LOC: Config.onoff.via
3588 If set (default), Squid will include a Via header in requests and
3589 replies as required by RFC2616.
3595 LOC: Config.onoff.ie_refresh
3598 Microsoft Internet Explorer up until version 5.5 Service
3599 Pack 1 has an issue with transparent proxies, wherein it
3600 is impossible to force a refresh. Turning this on provides
3601 a partial fix to the problem, by causing all IMS-REFRESH
3602 requests from older IE versions to check the origin server
3603 for fresh content. This reduces hit ratio by some amount
3604 (~10% in my experience), but allows users to actually get
3605 fresh content when they want it. Note because Squid
3606 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
3607 of 5.5 is unchanged from old versions of Squid (i.e. a
3608 forced refresh is impossible). Newer versions of IE will,
3609 hopefully, continue to have the new behavior and will be
3610 handled based on that assumption. This option defaults to
3611 the old Squid behavior, which is better for hit ratios but
3612 worse for clients using IE, if they need to be able to
3613 force fresh content.
3616 NAME: vary_ignore_expire
3619 LOC: Config.onoff.vary_ignore_expire
3622 Many HTTP servers supporting Vary gives such objects
3623 immediate expiry time with no cache-control header
3624 when requested by a HTTP/1.0 client. This option
3625 enables Squid to ignore such expiry times until
3626 HTTP/1.1 is fully implemented.
3628 WARNING: If turned on this may eventually cause some
3629 varying objects not intended for caching to get cached.
3632 NAME: request_entities
3634 LOC: Config.onoff.request_entities
3637 Squid defaults to deny GET and HEAD requests with request entities,
3638 as the meaning of such requests are undefined in the HTTP standard
3639 even if not explicitly forbidden.
3641 Set this directive to on if you have clients which insists
3642 on sending request entities in GET or HEAD requests. But be warned
3643 that there is server software (both proxies and web servers) which
3644 can fail to properly process this kind of request which may make you
3645 vulnerable to cache pollution attacks if enabled.
3648 NAME: request_header_access
3649 IFDEF: HTTP_VIOLATIONS
3650 TYPE: http_header_access[]
3651 LOC: Config.request_header_access
3654 Usage: request_header_access header_name allow|deny [!]aclname ...
3656 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3657 this feature could make you liable for problems which it
3660 This option replaces the old 'anonymize_headers' and the
3661 older 'http_anonymizer' option with something that is much
3662 more configurable. This new method creates a list of ACLs
3663 for each header, allowing you very fine-tuned header
3666 This option only applies to request headers, i.e., from the
3667 client to the server.
3669 You can only specify known headers for the header name.
3670 Other headers are reclassified as 'Other'. You can also
3671 refer to all the headers with 'All'.
3673 For example, to achieve the same behavior as the old
3674 'http_anonymizer standard' option, you should use:
3676 request_header_access From deny all
3677 request_header_access Referer deny all
3678 request_header_access Server deny all
3679 request_header_access User-Agent deny all
3680 request_header_access WWW-Authenticate deny all
3681 request_header_access Link deny all
3683 Or, to reproduce the old 'http_anonymizer paranoid' feature
3686 request_header_access Allow allow all
3687 request_header_access Authorization allow all
3688 request_header_access WWW-Authenticate allow all
3689 request_header_access Proxy-Authorization allow all
3690 request_header_access Proxy-Authenticate allow all
3691 request_header_access Cache-Control allow all
3692 request_header_access Content-Encoding allow all
3693 request_header_access Content-Length allow all
3694 request_header_access Content-Type allow all
3695 request_header_access Date allow all
3696 request_header_access Expires allow all
3697 request_header_access Host allow all
3698 request_header_access If-Modified-Since allow all
3699 request_header_access Last-Modified allow all
3700 request_header_access Location allow all
3701 request_header_access Pragma allow all
3702 request_header_access Accept allow all
3703 request_header_access Accept-Charset allow all
3704 request_header_access Accept-Encoding allow all
3705 request_header_access Accept-Language allow all
3706 request_header_access Content-Language allow all
3707 request_header_access Mime-Version allow all
3708 request_header_access Retry-After allow all
3709 request_header_access Title allow all
3710 request_header_access Connection allow all
3711 request_header_access Proxy-Connection allow all
3712 request_header_access All deny all
3714 although many of those are HTTP reply headers, and so should be
3715 controlled with the reply_header_access directive.
3717 By default, all headers are allowed (no anonymizing is
3721 NAME: reply_header_access
3722 IFDEF: HTTP_VIOLATIONS
3723 TYPE: http_header_access[]
3724 LOC: Config.reply_header_access
3727 Usage: reply_header_access header_name allow|deny [!]aclname ...
3729 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3730 this feature could make you liable for problems which it
3733 This option only applies to reply headers, i.e., from the
3734 server to the client.
3736 This is the same as request_header_access, but in the other
3739 This option replaces the old 'anonymize_headers' and the
3740 older 'http_anonymizer' option with something that is much
3741 more configurable. This new method creates a list of ACLs
3742 for each header, allowing you very fine-tuned header
3745 You can only specify known headers for the header name.
3746 Other headers are reclassified as 'Other'. You can also
3747 refer to all the headers with 'All'.
3749 For example, to achieve the same behavior as the old
3750 'http_anonymizer standard' option, you should use:
3752 reply_header_access From deny all
3753 reply_header_access Referer deny all
3754 reply_header_access Server deny all
3755 reply_header_access User-Agent deny all
3756 reply_header_access WWW-Authenticate deny all
3757 reply_header_access Link deny all
3759 Or, to reproduce the old 'http_anonymizer paranoid' feature
3762 reply_header_access Allow allow all
3763 reply_header_access Authorization allow all
3764 reply_header_access WWW-Authenticate allow all
3765 reply_header_access Proxy-Authorization allow all
3766 reply_header_access Proxy-Authenticate allow all
3767 reply_header_access Cache-Control allow all
3768 reply_header_access Content-Encoding allow all
3769 reply_header_access Content-Length allow all
3770 reply_header_access Content-Type allow all
3771 reply_header_access Date allow all
3772 reply_header_access Expires allow all
3773 reply_header_access Host allow all
3774 reply_header_access If-Modified-Since allow all
3775 reply_header_access Last-Modified allow all
3776 reply_header_access Location allow all
3777 reply_header_access Pragma allow all
3778 reply_header_access Accept allow all
3779 reply_header_access Accept-Charset allow all
3780 reply_header_access Accept-Encoding allow all
3781 reply_header_access Accept-Language allow all
3782 reply_header_access Content-Language allow all
3783 reply_header_access Mime-Version allow all
3784 reply_header_access Retry-After allow all
3785 reply_header_access Title allow all
3786 reply_header_access Connection allow all
3787 reply_header_access Proxy-Connection allow all
3788 reply_header_access All deny all
3790 although the HTTP request headers won't be usefully controlled
3791 by this directive -- see request_header_access for details.
3793 By default, all headers are allowed (no anonymizing is
3797 NAME: header_replace
3798 IFDEF: HTTP_VIOLATIONS
3799 TYPE: http_header_replace[]
3800 LOC: Config.request_header_access
3803 Usage: header_replace header_name message
3804 Example: header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
3806 This option allows you to change the contents of headers
3807 denied with header_access above, by replacing them with
3808 some fixed string. This replaces the old fake_user_agent
3811 This only applies to request headers, not reply headers.
3813 By default, headers are removed if denied.
3816 NAME: relaxed_header_parser
3817 COMMENT: on|off|warn
3819 LOC: Config.onoff.relaxed_header_parser
3822 In the default "on" setting Squid accepts certain forms
3823 of non-compliant HTTP messages where it is unambiguous
3824 what the sending application intended even if the message
3825 is not correctly formatted. The messages is then normalized
3826 to the correct form when forwarded by Squid.
3828 If set to "warn" then a warning will be emitted in cache.log
3829 each time such HTTP error is encountered.
3831 If set to "off" then such HTTP errors will cause the request
3832 or response to be rejected.
3837 -----------------------------------------------------------------------------
3840 NAME: forward_timeout
3843 LOC: Config.Timeout.forward
3846 This parameter specifies how long Squid should at most attempt in
3847 finding a forwarding path for the request before giving up.
3850 NAME: connect_timeout
3853 LOC: Config.Timeout.connect
3856 This parameter specifies how long to wait for the TCP connect to
3857 the requested server or peer to complete before Squid should
3858 attempt to find another path where to forward the request.
3861 NAME: peer_connect_timeout
3864 LOC: Config.Timeout.peer_connect
3867 This parameter specifies how long to wait for a pending TCP
3868 connection to a peer cache. The default is 30 seconds. You
3869 may also set different timeout values for individual neighbors
3870 with the 'connect-timeout' option on a 'cache_peer' line.
3876 LOC: Config.Timeout.read
3879 The read_timeout is applied on server-side connections. After
3880 each successful read(), the timeout will be extended by this
3881 amount. If no data is read again after this amount of time,
3882 the request is aborted and logged with ERR_READ_TIMEOUT. The
3883 default is 15 minutes.
3886 NAME: request_timeout
3888 LOC: Config.Timeout.request
3891 How long to wait for an HTTP request after initial
3892 connection establishment.
3895 NAME: persistent_request_timeout
3897 LOC: Config.Timeout.persistent_request
3900 How long to wait for the next HTTP request on a persistent
3901 connection after the previous request completes.
3904 NAME: client_lifetime
3907 LOC: Config.Timeout.lifetime
3910 The maximum amount of time a client (browser) is allowed to
3911 remain connected to the cache process. This protects the Cache
3912 from having a lot of sockets (and hence file descriptors) tied up
3913 in a CLOSE_WAIT state from remote clients that go away without
3914 properly shutting down (either because of a network failure or
3915 because of a poor client implementation). The default is one
3918 NOTE: The default value is intended to be much larger than any
3919 client would ever need to be connected to your cache. You
3920 should probably change client_lifetime only as a last resort.
3921 If you seem to have many client connections tying up
3922 filedescriptors, we recommend first tuning the read_timeout,
3923 request_timeout, persistent_request_timeout and quick_abort values.
3926 NAME: half_closed_clients
3928 LOC: Config.onoff.half_closed_clients
3931 Some clients may shutdown the sending side of their TCP
3932 connections, while leaving their receiving sides open. Sometimes,
3933 Squid can not tell the difference between a half-closed and a
3934 fully-closed TCP connection.
3936 By default, Squid will immediately close client connections when
3937 read(2) returns "no more data to read."
3939 Change this option to 'on' and Squid will keep open connections
3940 until a read(2) or write(2) on the socket returns an error.
3941 This may show some benefits for reverse proxies. But if not
3942 it is recommended to leave OFF.
3947 LOC: Config.Timeout.pconn
3950 Timeout for idle persistent connections to servers and other
3957 LOC: Ident::TheConfig.timeout
3960 Maximum time to wait for IDENT lookups to complete.
3962 If this is too high, and you enabled IDENT lookups from untrusted
3963 users, you might be susceptible to denial-of-service by having
3964 many ident requests going at once.
3967 NAME: shutdown_lifetime
3970 LOC: Config.shutdownLifetime
3973 When SIGTERM or SIGHUP is received, the cache is put into
3974 "shutdown pending" mode until all active sockets are closed.
3975 This value is the lifetime to set for all open descriptors
3976 during shutdown mode. Any active clients after this many
3977 seconds will receive a 'timeout' message.
3981 ADMINISTRATIVE PARAMETERS
3982 -----------------------------------------------------------------------------
3988 LOC: Config.adminEmail
3990 Email-address of local cache manager who will receive
3991 mail if the cache dies. The default is "webmaster."
3997 LOC: Config.EmailFrom
3999 From: email-address for mail sent when the cache dies.
4000 The default is to use 'appname@unique_hostname'.
4001 Default appname value is "squid", can be changed into
4002 src/globals.h before building squid.
4008 LOC: Config.EmailProgram
4010 Email program used to send mail if the cache dies.
4011 The default is "mail". The specified program must comply
4012 with the standard Unix mail syntax:
4013 mail-program recipient < mailfile
4015 Optional command line options can be specified.
4018 NAME: cache_effective_user
4020 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
4021 LOC: Config.effectiveUser
4023 If you start Squid as root, it will change its effective/real
4024 UID/GID to the user specified below. The default is to change
4025 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
4026 see also; cache_effective_group
4029 NAME: cache_effective_group
4032 LOC: Config.effectiveGroup
4034 Squid sets the GID to the effective user's default group ID
4035 (taken from the password file) and supplementary group list
4036 from the groups membership.
4038 If you want Squid to run with a specific GID regardless of
4039 the group memberships of the effective user then set this
4040 to the group (or GID) you want Squid to run as. When set
4041 all other group privileges of the effective user are ignored
4042 and only this GID is effective. If Squid is not started as
4043 root the user starting Squid MUST be member of the specified
4046 This option is not recommended by the Squid Team.
4047 Our preference is for administrators to configure a secure
4048 user account for squid with UID/GID matching system policies.
4051 NAME: httpd_suppress_version_string
4055 LOC: Config.onoff.httpd_suppress_version_string
4057 Suppress Squid version string info in HTTP headers and HTML error pages.
4060 NAME: visible_hostname
4062 LOC: Config.visibleHostname
4065 If you want to present a special hostname in error messages, etc,
4066 define this. Otherwise, the return value of gethostname()
4067 will be used. If you have multiple caches in a cluster and
4068 get errors about IP-forwarding you must set them to have individual
4069 names with this setting.
4072 NAME: unique_hostname
4074 LOC: Config.uniqueHostname
4077 If you want to have multiple machines with the same
4078 'visible_hostname' you must give each machine a different
4079 'unique_hostname' so forwarding loops can be detected.
4082 NAME: hostname_aliases
4084 LOC: Config.hostnameAliases
4087 A list of other DNS names your cache has.
4095 Minimum umask which should be enforced while the proxy
4096 is running, in addition to the umask set at startup.
4098 For a traditional octal representation of umasks, start
4103 OPTIONS FOR THE CACHE REGISTRATION SERVICE
4104 -----------------------------------------------------------------------------
4106 This section contains parameters for the (optional) cache
4107 announcement service. This service is provided to help
4108 cache administrators locate one another in order to join or
4109 create cache hierarchies.
4111 An 'announcement' message is sent (via UDP) to the registration
4112 service by Squid. By default, the announcement message is NOT
4113 SENT unless you enable it with 'announce_period' below.
4115 The announcement message includes your hostname, plus the
4116 following information from this configuration file:
4122 All current information is processed regularly and made
4123 available on the Web at http://www.ircache.net/Cache/Tracker/.
4126 NAME: announce_period
4128 LOC: Config.Announce.period
4131 This is how frequently to send cache announcements. The
4132 default is `0' which disables sending the announcement
4135 To enable announcing your cache, just uncomment the line
4139 #To enable announcing your cache, just uncomment the line below.
4140 #announce_period 1 day
4146 DEFAULT: tracker.ircache.net
4147 LOC: Config.Announce.host
4153 LOC: Config.Announce.file
4159 LOC: Config.Announce.port
4161 announce_host and announce_port set the hostname and port
4162 number where the registration message will be sent.
4164 Hostname will default to 'tracker.ircache.net' and port will
4165 default default to 3131. If the 'filename' argument is given,
4166 the contents of that file will be included in the announce
4171 HTTPD-ACCELERATOR OPTIONS
4172 -----------------------------------------------------------------------------
4175 NAME: httpd_accel_surrogate_id
4176 IFDEF: USE_SQUID_ESI
4178 LOC: Config.Accel.surrogate_id
4181 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
4182 need an identification token to allow control targeting. Because
4183 a farm of surrogates may all perform the same tasks, they may share
4184 an identification token.
4187 NAME: http_accel_surrogate_remote
4188 IFDEF: USE_SQUID_ESI
4192 LOC: Config.onoff.surrogate_is_remote
4194 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
4195 Set this to on to have squid behave as a remote surrogate.
4199 IFDEF: USE_SQUID_ESI
4200 COMMENT: libxml2|expat|custom
4202 LOC: ESIParser::Type
4205 ESI markup is not strictly XML compatible. The custom ESI parser
4206 will give higher performance, but cannot handle non ASCII character
4211 DELAY POOL PARAMETERS
4212 -----------------------------------------------------------------------------
4216 TYPE: delay_pool_count
4221 This represents the number of delay pools to be used. For example,
4222 if you have one class 2 delay pool and one class 3 delays pool, you
4223 have a total of 2 delay pools.
4227 TYPE: delay_pool_class
4232 This defines the class of each delay pool. There must be exactly one
4233 delay_class line for each delay pool. For example, to define two
4234 delay pools, one of class 2 and one of class 3, the settings above
4238 delay_pools 4 # 4 delay pools
4239 delay_class 1 2 # pool 1 is a class 2 pool
4240 delay_class 2 3 # pool 2 is a class 3 pool
4241 delay_class 3 4 # pool 3 is a class 4 pool
4242 delay_class 4 5 # pool 4 is a class 5 pool
4244 The delay pool classes are:
4246 class 1 Everything is limited by a single aggregate
4249 class 2 Everything is limited by a single aggregate
4250 bucket as well as an "individual" bucket chosen
4251 from bits 25 through 32 of the IPv4 address.
4253 class 3 Everything is limited by a single aggregate
4254 bucket as well as a "network" bucket chosen
4255 from bits 17 through 24 of the IP address and a
4256 "individual" bucket chosen from bits 17 through
4257 32 of the IPv4 address.
4259 class 4 Everything in a class 3 delay pool, with an
4260 additional limit on a per user basis. This
4261 only takes effect if the username is established
4262 in advance - by forcing authentication in your
4265 class 5 Requests are grouped according their tag (see
4266 external_acl's tag= reply).
4268 NOTE: If an IP address is a.b.c.d
4269 -> bits 25 through 32 are "d"
4270 -> bits 17 through 24 are "c"
4271 -> bits 17 through 32 are "c * 256 + d"
4273 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
4274 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
4278 TYPE: delay_pool_access
4283 This is used to determine which delay pool a request falls into.
4285 delay_access is sorted per pool and the matching starts with pool 1,
4286 then pool 2, ..., and finally pool N. The first delay pool where the
4287 request is allowed is selected for the request. If it does not allow
4288 the request to any pool then the request is not delayed (default).
4290 For example, if you want some_big_clients in delay
4291 pool 1 and lotsa_little_clients in delay pool 2:
4294 delay_access 1 allow some_big_clients
4295 delay_access 1 deny all
4296 delay_access 2 allow lotsa_little_clients
4297 delay_access 2 deny all
4298 delay_access 3 allow authenticated_clients
4301 NAME: delay_parameters
4302 TYPE: delay_pool_rates
4307 This defines the parameters for a delay pool. Each delay pool has
4308 a number of "buckets" associated with it, as explained in the
4309 description of delay_class. For a class 1 delay pool, the syntax is:
4311 delay_parameters pool aggregate
4313 For a class 2 delay pool:
4315 delay_parameters pool aggregate individual
4317 For a class 3 delay pool:
4319 delay_parameters pool aggregate network individual
4321 For a class 4 delay pool:
4323 delay_parameters pool aggregate network individual user
4325 For a class 5 delay pool:
4327 delay_parameters pool tag
4329 The variables here are:
4331 pool a pool number - ie, a number between 1 and the
4332 number specified in delay_pools as used in
4335 aggregate the "delay parameters" for the aggregate bucket
4338 individual the "delay parameters" for the individual
4339 buckets (class 2, 3).
4341 network the "delay parameters" for the network buckets
4344 user the delay parameters for the user buckets
4347 tag the delay parameters for the tag buckets
4350 A pair of delay parameters is written restore/maximum, where restore is
4351 the number of bytes (not bits - modem and network speeds are usually
4352 quoted in bits) per second placed into the bucket, and maximum is the
4353 maximum number of bytes which can be in the bucket at any time.
4355 For example, if delay pool number 1 is a class 2 delay pool as in the
4356 above example, and is being used to strictly limit each host to 64kbps
4357 (plus overheads), with no overall limit, the line is:
4359 delay_parameters 1 -1/-1 8000/8000
4361 Note that the figure -1 is used to represent "unlimited".
4363 And, if delay pool number 2 is a class 3 delay pool as in the above
4364 example, and you want to limit it to a total of 256kbps (strict limit)
4365 with each 8-bit network permitted 64kbps (strict limit) and each
4366 individual host permitted 4800bps with a bucket maximum size of 64kb
4367 to permit a decent web page to be downloaded at a decent speed
4368 (if the network is not being limited due to overuse) but slow down
4369 large downloads more significantly:
4371 delay_parameters 2 32000/32000 8000/8000 600/8000
4373 There must be one delay_parameters line for each delay pool.
4375 Finally, for a class 4 delay pool as in the example - each user will
4376 be limited to 128Kb no matter how many workstations they are logged into.:
4378 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
4381 NAME: delay_initial_bucket_level
4382 COMMENT: (percent, 0-100)
4386 LOC: Config.Delay.initial
4388 The initial bucket percentage is used to determine how much is put
4389 in each bucket when squid starts, is reconfigured, or first notices
4390 a host accessing it (in class 2 and class 3, individual hosts and
4391 networks only have buckets associated with them once they have been
4396 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
4397 -----------------------------------------------------------------------------
4402 LOC: Config.Wccp.router
4406 Use this option to define your WCCP ``home'' router for
4409 wccp_router supports a single WCCP(v1) router
4411 wccp2_router supports multiple WCCPv2 routers
4413 only one of the two may be used at the same time and defines
4414 which version of WCCP to use.
4418 TYPE: IpAddress_list
4419 LOC: Config.Wccp2.router
4423 Use this option to define your WCCP ``home'' router for
4426 wccp_router supports a single WCCP(v1) router
4428 wccp2_router supports multiple WCCPv2 routers
4430 only one of the two may be used at the same time and defines
4431 which version of WCCP to use.
4436 LOC: Config.Wccp.version
4440 This directive is only relevant if you need to set up WCCP(v1)
4441 to some very old and end-of-life Cisco routers. In all other
4442 setups it must be left unset or at the default setting.
4443 It defines an internal version in the WCCP(v1) protocol,
4444 with version 4 being the officially documented protocol.
4446 According to some users, Cisco IOS 11.2 and earlier only
4447 support WCCP version 3. If you're using that or an earlier
4448 version of IOS, you may need to change this value to 3, otherwise
4449 do not specify this parameter.
4452 NAME: wccp2_rebuild_wait
4454 LOC: Config.Wccp2.rebuildwait
4458 If this is enabled Squid will wait for the cache dir rebuild to finish
4459 before sending the first wccp2 HereIAm packet
4462 NAME: wccp2_forwarding_method
4464 LOC: Config.Wccp2.forwarding_method
4468 WCCP2 allows the setting of forwarding methods between the
4469 router/switch and the cache. Valid values are as follows:
4471 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4472 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4474 Currently (as of IOS 12.4) cisco routers only support GRE.
4475 Cisco switches only support the L2 redirect assignment method.
4478 NAME: wccp2_return_method
4480 LOC: Config.Wccp2.return_method
4484 WCCP2 allows the setting of return methods between the
4485 router/switch and the cache for packets that the cache
4486 decides not to handle. Valid values are as follows:
4488 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4489 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4491 Currently (as of IOS 12.4) cisco routers only support GRE.
4492 Cisco switches only support the L2 redirect assignment.
4494 If the "ip wccp redirect exclude in" command has been
4495 enabled on the cache interface, then it is still safe for
4496 the proxy server to use a l2 redirect method even if this
4497 option is set to GRE.
4500 NAME: wccp2_assignment_method
4502 LOC: Config.Wccp2.assignment_method
4506 WCCP2 allows the setting of methods to assign the WCCP hash
4507 Valid values are as follows:
4509 hash - Hash assignment
4510 mask - Mask assignment
4512 As a general rule, cisco routers support the hash assignment method
4513 and cisco switches support the mask assignment method.
4518 LOC: Config.Wccp2.info
4520 DEFAULT_IF_NONE: standard 0
4523 WCCP2 allows for multiple traffic services. There are two
4524 types: "standard" and "dynamic". The standard type defines
4525 one service id - http (id 0). The dynamic service ids can be from
4526 51 to 255 inclusive. In order to use a dynamic service id
4527 one must define the type of traffic to be redirected; this is done
4528 using the wccp2_service_info option.
4530 The "standard" type does not require a wccp2_service_info option,
4531 just specifying the service id will suffice.
4533 MD5 service authentication can be enabled by adding
4534 "password=<password>" to the end of this service declaration.
4538 wccp2_service standard 0 # for the 'web-cache' standard service
4539 wccp2_service dynamic 80 # a dynamic service type which will be
4540 # fleshed out with subsequent options.
4541 wccp2_service standard 0 password=foo
4544 NAME: wccp2_service_info
4545 TYPE: wccp2_service_info
4546 LOC: Config.Wccp2.info
4550 Dynamic WCCPv2 services require further information to define the
4551 traffic you wish to have diverted.
4555 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
4556 priority=<priority> ports=<port>,<port>..
4558 The relevant WCCPv2 flags:
4559 + src_ip_hash, dst_ip_hash
4560 + source_port_hash, dst_port_hash
4561 + src_ip_alt_hash, dst_ip_alt_hash
4562 + src_port_alt_hash, dst_port_alt_hash
4565 The port list can be one to eight entries.
4569 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
4570 priority=240 ports=80
4572 Note: the service id must have been defined by a previous
4573 'wccp2_service dynamic <id>' entry.
4578 LOC: Config.Wccp2.weight
4582 Each cache server gets assigned a set of the destination
4583 hash proportional to their weight.
4588 LOC: Config.Wccp.address
4595 LOC: Config.Wccp2.address
4599 Use this option if you require WCCP to use a specific
4602 The default behavior is to not bind to any specific address.
4606 PERSISTENT CONNECTION HANDLING
4607 -----------------------------------------------------------------------------
4609 Also see "pconn_timeout" in the TIMEOUTS section
4612 NAME: client_persistent_connections
4614 LOC: Config.onoff.client_pconns
4618 NAME: server_persistent_connections
4620 LOC: Config.onoff.server_pconns
4623 Persistent connection support for clients and servers. By
4624 default, Squid uses persistent connections (when allowed)
4625 with its clients and servers. You can use these options to
4626 disable persistent connections with clients and/or servers.
4629 NAME: persistent_connection_after_error
4631 LOC: Config.onoff.error_pconns
4634 With this directive the use of persistent connections after
4635 HTTP errors can be disabled. Useful if you have clients
4636 who fail to handle errors on persistent connections proper.
4639 NAME: detect_broken_pconn
4641 LOC: Config.onoff.detect_broken_server_pconns
4644 Some servers have been found to incorrectly signal the use
4645 of HTTP/1.0 persistent connections even on replies not
4646 compatible, causing significant delays. This server problem
4647 has mostly been seen on redirects.
4649 By enabling this directive Squid attempts to detect such
4650 broken replies and automatically assume the reply is finished
4651 after 10 seconds timeout.
4655 CACHE DIGEST OPTIONS
4656 -----------------------------------------------------------------------------
4659 NAME: digest_generation
4660 IFDEF: USE_CACHE_DIGESTS
4662 LOC: Config.onoff.digest_generation
4665 This controls whether the server will generate a Cache Digest
4666 of its contents. By default, Cache Digest generation is
4667 enabled if Squid is compiled with --enable-cache-digests defined.
4670 NAME: digest_bits_per_entry
4671 IFDEF: USE_CACHE_DIGESTS
4673 LOC: Config.digest.bits_per_entry
4676 This is the number of bits of the server's Cache Digest which
4677 will be associated with the Digest entry for a given HTTP
4678 Method and URL (public key) combination. The default is 5.
4681 NAME: digest_rebuild_period
4682 IFDEF: USE_CACHE_DIGESTS
4685 LOC: Config.digest.rebuild_period
4688 This is the wait time between Cache Digest rebuilds.
4691 NAME: digest_rewrite_period
4693 IFDEF: USE_CACHE_DIGESTS
4695 LOC: Config.digest.rewrite_period
4698 This is the wait time between Cache Digest writes to
4702 NAME: digest_swapout_chunk_size
4705 IFDEF: USE_CACHE_DIGESTS
4706 LOC: Config.digest.swapout_chunk_size
4709 This is the number of bytes of the Cache Digest to write to
4710 disk at a time. It defaults to 4096 bytes (4KB), the Squid
4714 NAME: digest_rebuild_chunk_percentage
4715 COMMENT: (percent, 0-100)
4716 IFDEF: USE_CACHE_DIGESTS
4718 LOC: Config.digest.rebuild_chunk_percentage
4721 This is the percentage of the Cache Digest to be scanned at a
4722 time. By default it is set to 10% of the Cache Digest.
4727 -----------------------------------------------------------------------------
4732 LOC: Config.Port.snmp
4736 The port number where Squid listens for SNMP requests. To enable
4737 SNMP support set this to a suitable port number. Port number
4738 3401 is often used for the Squid SNMP agent. By default it's
4739 set to "0" (disabled)
4747 LOC: Config.accessList.snmp
4749 DEFAULT_IF_NONE: deny all
4752 Allowing or denying access to the SNMP port.
4754 All access to the agent is denied by default.
4757 snmp_access allow|deny [!]aclname ...
4759 This clause only supports fast acl types.
4760 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
4762 snmp_access allow snmppublic localhost
4763 snmp_access deny all
4766 NAME: snmp_incoming_address
4768 LOC: Config.Addrs.snmp_incoming
4773 NAME: snmp_outgoing_address
4775 LOC: Config.Addrs.snmp_outgoing
4776 DEFAULT: 255.255.255.255
4779 Just like 'udp_incoming_address', but for the SNMP port.
4781 snmp_incoming_address is used for the SNMP socket receiving
4782 messages from SNMP agents.
4783 snmp_outgoing_address is used for SNMP packets returned to SNMP
4786 The default snmp_incoming_address (0.0.0.0) is to listen on all
4787 available network interfaces.
4789 If snmp_outgoing_address is set to 255.255.255.255 (the default)
4790 it will use the same socket as snmp_incoming_address. Only
4791 change this if you want to have SNMP replies sent using another
4792 address than where this Squid listens for SNMP queries.
4794 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
4795 the same value since they both use port 3401.
4800 -----------------------------------------------------------------------------
4803 NAME: icp_port udp_port
4806 LOC: Config.Port.icp
4808 The port number where Squid sends and receives ICP queries to
4809 and from neighbor caches. The standard UDP port for ICP is 3130.
4810 Default is disabled (0).
4812 #icp_port @DEFAULT_ICP_PORT@
4820 LOC: Config.Port.htcp
4822 The port number where Squid sends and receives HTCP queries to
4823 and from neighbor caches. To turn it on you want to set it to
4824 4827. By default it is set to "0" (disabled).
4830 NAME: log_icp_queries
4834 LOC: Config.onoff.log_udp
4836 If set, ICP queries are logged to access.log. You may wish
4837 do disable this if your ICP load is VERY high to speed things
4838 up or to simplify log analysis.
4841 NAME: udp_incoming_address
4843 LOC:Config.Addrs.udp_incoming
4846 udp_incoming_address is used for UDP packets received from other
4849 The default behavior is to not bind to any specific address.
4851 Only change this if you want to have all UDP queries received on
4852 a specific interface/address.
4854 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
4855 modules. Altering it will affect all of them in the same manner.
4857 see also; udp_outgoing_address
4859 NOTE, udp_incoming_address and udp_outgoing_address can not
4860 have the same value since they both use the same port.
4863 NAME: udp_outgoing_address
4865 LOC: Config.Addrs.udp_outgoing
4866 DEFAULT: 255.255.255.255
4868 udp_outgoing_address is used for UDP packets sent out to other
4871 The default behavior is to not bind to any specific address.
4873 Instead it will use the same socket as udp_incoming_address.
4874 Only change this if you want to have UDP queries sent using another
4875 address than where this Squid listens for UDP queries from other
4878 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
4879 modules. Altering it will affect all of them in the same manner.
4881 see also; udp_incoming_address
4883 NOTE, udp_incoming_address and udp_outgoing_address can not
4884 have the same value since they both use the same port.
4891 LOC: Config.onoff.icp_hit_stale
4893 If you want to return ICP_HIT for stale cache objects, set this
4894 option to 'on'. If you have sibling relationships with caches
4895 in other administrative domains, this should be 'off'. If you only
4896 have sibling relationships with caches under your control,
4897 it is probably okay to set this to 'on'.
4898 If set to 'on', your siblings should use the option "allow-miss"
4899 on their cache_peer lines for connecting to you.
4902 NAME: minimum_direct_hops
4905 LOC: Config.minDirectHops
4907 If using the ICMP pinging stuff, do direct fetches for sites
4908 which are no more than this many hops away.
4911 NAME: minimum_direct_rtt
4914 LOC: Config.minDirectRtt
4916 If using the ICMP pinging stuff, do direct fetches for sites
4917 which are no more than this many rtt milliseconds away.
4923 LOC: Config.Netdb.low
4929 LOC: Config.Netdb.high
4931 The low and high water marks for the ICMP measurement
4932 database. These are counts, not percents. The defaults are
4933 900 and 1000. When the high water mark is reached, database
4934 entries will be deleted until the low mark is reached.
4937 NAME: netdb_ping_period
4939 LOC: Config.Netdb.period
4942 The minimum period for measuring a site. There will be at
4943 least this much delay between successive pings to the same
4944 network. The default is five minutes.
4951 LOC: Config.onoff.query_icmp
4953 If you want to ask your peers to include ICMP data in their ICP
4954 replies, enable this option.
4956 If your peer has configured Squid (during compilation) with
4957 '--enable-icmp' that peer will send ICMP pings to origin server
4958 sites of the URLs it receives. If you enable this option the
4959 ICP replies from that peer will include the ICMP data (if available).
4960 Then, when choosing a parent cache, Squid will choose the parent with
4961 the minimal RTT to the origin server. When this happens, the
4962 hierarchy field of the access.log will be
4963 "CLOSEST_PARENT_MISS". This option is off by default.
4966 NAME: test_reachability
4970 LOC: Config.onoff.test_reachability
4972 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
4973 instead of ICP_MISS if the target host is NOT in the ICMP
4974 database, or has a zero RTT.
4977 NAME: icp_query_timeout
4981 LOC: Config.Timeout.icp_query
4983 Normally Squid will automatically determine an optimal ICP
4984 query timeout value based on the round-trip-time of recent ICP
4985 queries. If you want to override the value determined by
4986 Squid, set this 'icp_query_timeout' to a non-zero value. This
4987 value is specified in MILLISECONDS, so, to use a 2-second
4988 timeout (the old default), you would write:
4990 icp_query_timeout 2000
4993 NAME: maximum_icp_query_timeout
4997 LOC: Config.Timeout.icp_query_max
4999 Normally the ICP query timeout is determined dynamically. But
5000 sometimes it can lead to very large values (say 5 seconds).
5001 Use this option to put an upper limit on the dynamic timeout
5002 value. Do NOT use this option to always use a fixed (instead
5003 of a dynamic) timeout value. To set a fixed timeout see the
5004 'icp_query_timeout' directive.
5007 NAME: minimum_icp_query_timeout
5011 LOC: Config.Timeout.icp_query_min
5013 Normally the ICP query timeout is determined dynamically. But
5014 sometimes it can lead to very small timeouts, even lower than
5015 the normal latency variance on your link due to traffic.
5016 Use this option to put an lower limit on the dynamic timeout
5017 value. Do NOT use this option to always use a fixed (instead
5018 of a dynamic) timeout value. To set a fixed timeout see the
5019 'icp_query_timeout' directive.
5022 NAME: background_ping_rate
5026 LOC: Config.backgroundPingRate
5028 Controls how often the ICP pings are sent to siblings that
5029 have background-ping set.
5033 MULTICAST ICP OPTIONS
5034 -----------------------------------------------------------------------------
5039 LOC: Config.mcast_group_list
5042 This tag specifies a list of multicast groups which your server
5043 should join to receive multicasted ICP queries.
5045 NOTE! Be very careful what you put here! Be sure you
5046 understand the difference between an ICP _query_ and an ICP
5047 _reply_. This option is to be set only if you want to RECEIVE
5048 multicast queries. Do NOT set this option to SEND multicast
5049 ICP (use cache_peer for that). ICP replies are always sent via
5050 unicast, so this option does not affect whether or not you will
5051 receive replies from multicast group members.
5053 You must be very careful to NOT use a multicast address which
5054 is already in use by another group of caches.
5056 If you are unsure about multicast, please read the Multicast
5057 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
5059 Usage: mcast_groups 239.128.16.128 224.0.1.20
5061 By default, Squid doesn't listen on any multicast groups.
5064 NAME: mcast_miss_addr
5065 IFDEF: MULTICAST_MISS_STREAM
5067 LOC: Config.mcast_miss.addr
5068 DEFAULT: 255.255.255.255
5070 If you enable this option, every "cache miss" URL will
5071 be sent out on the specified multicast address.
5073 Do not enable this option unless you are are absolutely
5074 certain you understand what you are doing.
5077 NAME: mcast_miss_ttl
5078 IFDEF: MULTICAST_MISS_STREAM
5080 LOC: Config.mcast_miss.ttl
5083 This is the time-to-live value for packets multicasted
5084 when multicasting off cache miss URLs is enabled. By
5085 default this is set to 'site scope', i.e. 16.
5088 NAME: mcast_miss_port
5089 IFDEF: MULTICAST_MISS_STREAM
5091 LOC: Config.mcast_miss.port
5094 This is the port number to be used in conjunction with
5098 NAME: mcast_miss_encode_key
5099 IFDEF: MULTICAST_MISS_STREAM
5101 LOC: Config.mcast_miss.encode_key
5102 DEFAULT: XXXXXXXXXXXXXXXX
5104 The URLs that are sent in the multicast miss stream are
5105 encrypted. This is the encryption key.
5108 NAME: mcast_icp_query_timeout
5112 LOC: Config.Timeout.mcast_icp_query
5114 For multicast peers, Squid regularly sends out ICP "probes" to
5115 count how many other peers are listening on the given multicast
5116 address. This value specifies how long Squid should wait to
5117 count all the replies. The default is 2000 msec, or 2
5122 INTERNAL ICON OPTIONS
5123 -----------------------------------------------------------------------------
5126 NAME: icon_directory
5128 LOC: Config.icons.directory
5129 DEFAULT: @DEFAULT_ICON_DIR@
5131 Where the icons are stored. These are normally kept in
5135 NAME: global_internal_static
5137 LOC: Config.onoff.global_internal_static
5140 This directive controls is Squid should intercept all requests for
5141 /squid-internal-static/ no matter which host the URL is requesting
5142 (default on setting), or if nothing special should be done for
5143 such URLs (off setting). The purpose of this directive is to make
5144 icons etc work better in complex cache hierarchies where it may
5145 not always be possible for all corners in the cache mesh to reach
5146 the server generating a directory listing.
5149 NAME: short_icon_urls
5151 LOC: Config.icons.use_short_names
5154 If this is enabled Squid will use short URLs for icons.
5155 If disabled it will revert to the old behavior of including
5156 it's own name and port in the URL.
5158 If you run a complex cache hierarchy with a mix of Squid and
5159 other proxies you may need to disable this directive.
5164 -----------------------------------------------------------------------------
5167 NAME: error_directory
5169 LOC: Config.errorDirectory
5172 If you wish to create your own versions of the default
5173 error files to customize them to suit your company copy
5174 the error/template files to another directory and point
5177 WARNING: This option will disable multi-language support
5178 on error pages if used.
5180 The squid developers are interested in making squid available in
5181 a wide variety of languages. If you are making translations for a
5182 language that Squid does not currently provide please consider
5183 contributing your translation back to the project.
5184 http://wiki.squid-cache.org/Translations
5186 The squid developers working on translations are happy to supply drop-in
5187 translated error files in exchange for any new language contributions.
5190 NAME: error_default_language
5191 IFDEF: USE_ERR_LOCALES
5193 LOC: Config.errorDefaultLanguage
5196 Set the default language which squid will send error pages in
5197 if no existing translation matches the clients language
5200 If unset (default) generic English will be used.
5202 The squid developers are interested in making squid available in
5203 a wide variety of languages. If you are interested in making
5204 translations for any language see the squid wiki for details.
5205 http://wiki.squid-cache.org/Translations
5208 NAME: error_log_languages
5209 IFDEF: USE_ERR_LOCALES
5211 LOC: Config.errorLogMissingLanguages
5214 Log to cache.log what languages users are attempting to
5215 auto-negotiate for translations.
5217 Successful negotiations are not logged. Only failures
5218 have meaning to indicate that Squid may need an upgrade
5219 of its error page translations.
5222 NAME: err_page_stylesheet
5224 LOC: Config.errorStylesheet
5225 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
5227 CSS Stylesheet to pattern the display of Squid default error pages.
5229 For information on CSS see http://www.w3.org/Style/CSS/
5234 LOC: Config.errHtmlText
5237 HTML text to include in error messages. Make this a "mailto"
5238 URL to your admin address, or maybe just a link to your
5239 organizations Web page.
5241 To include this in your error messages, you must rewrite
5242 the error template files (found in the "errors" directory).
5243 Wherever you want the 'err_html_text' line to appear,
5244 insert a %L tag in the error template file.
5247 NAME: email_err_data
5250 LOC: Config.onoff.emailErrData
5253 If enabled, information about the occurred error will be
5254 included in the mailto links of the ERR pages (if %W is set)
5255 so that the email body contains the data.
5256 Syntax is <A HREF="mailto:%w%W">%w</A>
5261 LOC: Config.denyInfoList
5264 Usage: deny_info err_page_name acl
5265 or deny_info http://... acl
5266 or deny_info TCP_RESET acl
5268 This can be used to return a ERR_ page for requests which
5269 do not pass the 'http_access' rules. Squid remembers the last
5270 acl it evaluated in http_access, and if a 'deny_info' line exists
5271 for that ACL Squid returns a corresponding error page.
5273 The acl is typically the last acl on the http_access deny line which
5274 denied access. The exceptions to this rule are:
5275 - When Squid needs to request authentication credentials. It's then
5276 the first authentication related acl encountered
5277 - When none of the http_access lines matches. It's then the last
5278 acl processed on the last http_access line.
5280 NP: If providing your own custom error pages with error_directory
5281 you may also specify them by your custom file name:
5282 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
5284 Alternatively you can specify an error URL. The browsers will
5285 get redirected (302) to the specified URL. %s in the redirection
5286 URL will be replaced by the requested URL.
5288 Alternatively you can tell Squid to reset the TCP connection
5289 by specifying TCP_RESET.
5293 OPTIONS INFLUENCING REQUEST FORWARDING
5294 -----------------------------------------------------------------------------
5297 NAME: nonhierarchical_direct
5299 LOC: Config.onoff.nonhierarchical_direct
5302 By default, Squid will send any non-hierarchical requests
5303 (matching hierarchy_stoplist or not cacheable request type) direct
5306 If you set this to off, Squid will prefer to send these
5307 requests to parents.
5309 Note that in most configurations, by turning this off you will only
5310 add latency to these request without any improvement in global hit
5313 If you are inside an firewall see never_direct instead of
5319 LOC: Config.onoff.prefer_direct
5322 Normally Squid tries to use parents for most requests. If you for some
5323 reason like it to first try going direct and only use a parent if
5324 going direct fails set this to on.
5326 By combining nonhierarchical_direct off and prefer_direct on you
5327 can set up Squid to use a parent as a backup path if going direct
5330 Note: If you want Squid to use parents for all requests see
5331 the never_direct directive. prefer_direct only modifies how Squid
5332 acts on cacheable requests.
5337 LOC: Config.accessList.AlwaysDirect
5340 Usage: always_direct allow|deny [!]aclname ...
5342 Here you can use ACL elements to specify requests which should
5343 ALWAYS be forwarded by Squid to the origin servers without using
5344 any peers. For example, to always directly forward requests for
5345 local servers ignoring any parents or siblings you may have use
5348 acl local-servers dstdomain my.domain.net
5349 always_direct allow local-servers
5351 To always forward FTP requests directly, use
5354 always_direct allow FTP
5356 NOTE: There is a similar, but opposite option named
5357 'never_direct'. You need to be aware that "always_direct deny
5358 foo" is NOT the same thing as "never_direct allow foo". You
5359 may need to use a deny rule to exclude a more-specific case of
5360 some other rule. Example:
5362 acl local-external dstdomain external.foo.net
5363 acl local-servers dstdomain .foo.net
5364 always_direct deny local-external
5365 always_direct allow local-servers
5367 NOTE: If your goal is to make the client forward the request
5368 directly to the origin server bypassing Squid then this needs
5369 to be done in the client configuration. Squid configuration
5370 can only tell Squid how Squid should fetch the object.
5372 NOTE: This directive is not related to caching. The replies
5373 is cached as usual even if you use always_direct. To not cache
5374 the replies see the 'cache' directive.
5376 This clause supports both fast and slow acl types.
5377 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5382 LOC: Config.accessList.NeverDirect
5385 Usage: never_direct allow|deny [!]aclname ...
5387 never_direct is the opposite of always_direct. Please read
5388 the description for always_direct if you have not already.
5390 With 'never_direct' you can use ACL elements to specify
5391 requests which should NEVER be forwarded directly to origin
5392 servers. For example, to force the use of a proxy for all
5393 requests, except those in your local domain use something like:
5395 acl local-servers dstdomain .foo.net
5396 never_direct deny local-servers
5397 never_direct allow all
5399 or if Squid is inside a firewall and there are local intranet
5400 servers inside the firewall use something like:
5402 acl local-intranet dstdomain .foo.net
5403 acl local-external dstdomain external.foo.net
5404 always_direct deny local-external
5405 always_direct allow local-intranet
5406 never_direct allow all
5408 This clause supports both fast and slow acl types.
5409 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5413 ADVANCED NETWORKING OPTIONS
5414 -----------------------------------------------------------------------------
5417 NAME: incoming_icp_average
5420 LOC: Config.comm_incoming.icp_average
5423 NAME: incoming_http_average
5426 LOC: Config.comm_incoming.http_average
5429 NAME: incoming_dns_average
5432 LOC: Config.comm_incoming.dns_average
5435 NAME: min_icp_poll_cnt
5438 LOC: Config.comm_incoming.icp_min_poll
5441 NAME: min_dns_poll_cnt
5444 LOC: Config.comm_incoming.dns_min_poll
5447 NAME: min_http_poll_cnt
5450 LOC: Config.comm_incoming.http_min_poll
5452 Heavy voodoo here. I can't even believe you are reading this.
5453 Are you crazy? Don't even think about adjusting these unless
5454 you understand the algorithms in comm_select.c first!
5460 LOC: Config.accept_filter
5464 The name of an accept(2) filter to install on Squid's
5465 listen socket(s). This feature is perhaps specific to
5466 FreeBSD and requires support in the kernel.
5468 The 'httpready' filter delays delivering new connections
5469 to Squid until a full HTTP request has been received.
5470 See the accf_http(9) man page for details.
5472 The 'dataready' filter delays delivering new connections
5473 to Squid until there is some data to process.
5474 See the accf_dataready(9) man page for details.
5478 The 'data' filter delays delivering of new connections
5479 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
5480 You may optionally specify a number of seconds to wait by
5481 'data=N' where N is the number of seconds. Defaults to 30
5482 if not specified. See the tcp(7) man page for details.
5485 accept_filter httpready
5490 NAME: tcp_recv_bufsize
5494 LOC: Config.tcpRcvBufsz
5496 Size of receive buffer to set for TCP sockets. Probably just
5497 as easy to change your kernel's default. Set to zero to use
5498 the default buffer size.
5503 -----------------------------------------------------------------------------
5510 LOC: Adaptation::Icap::TheConfig.onoff
5513 If you want to enable the ICAP module support, set this to on.
5516 NAME: icap_connect_timeout
5519 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
5522 This parameter specifies how long to wait for the TCP connect to
5523 the requested ICAP server to complete before giving up and either
5524 terminating the HTTP transaction or bypassing the failure.
5526 The default for optional services is peer_connect_timeout.
5527 The default for essential services is connect_timeout.
5528 If this option is explicitly set, its value applies to all services.
5531 NAME: icap_io_timeout
5535 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
5538 This parameter specifies how long to wait for an I/O activity on
5539 an established, active ICAP connection before giving up and
5540 either terminating the HTTP transaction or bypassing the
5543 The default is read_timeout.
5546 NAME: icap_service_failure_limit
5549 LOC: Adaptation::Icap::TheConfig.service_failure_limit
5552 The limit specifies the number of failures that Squid tolerates
5553 when establishing a new TCP connection with an ICAP service. If
5554 the number of failures exceeds the limit, the ICAP service is
5555 not used for new ICAP requests until it is time to refresh its
5556 OPTIONS. The per-service failure counter is reset to zero each
5557 time Squid fetches new service OPTIONS.
5559 A negative value disables the limit. Without the limit, an ICAP
5560 service will not be considered down due to connectivity failures
5561 between ICAP OPTIONS requests.
5564 NAME: icap_service_revival_delay
5567 LOC: Adaptation::Icap::TheConfig.service_revival_delay
5570 The delay specifies the number of seconds to wait after an ICAP
5571 OPTIONS request failure before requesting the options again. The
5572 failed ICAP service is considered "down" until fresh OPTIONS are
5575 The actual delay cannot be smaller than the hardcoded minimum
5576 delay of 30 seconds.
5579 NAME: icap_preview_enable
5583 LOC: Adaptation::Icap::TheConfig.preview_enable
5586 The ICAP Preview feature allows the ICAP server to handle the
5587 HTTP message by looking only at the beginning of the message body
5588 or even without receiving the body at all. In some environments,
5589 previews greatly speedup ICAP processing.
5591 During an ICAP OPTIONS transaction, the server may tell Squid what
5592 HTTP messages should be previewed and how big the preview should be.
5593 Squid will not use Preview if the server did not request one.
5595 To disable ICAP Preview for all ICAP services, regardless of
5596 individual ICAP server OPTIONS responses, set this option to "off".
5598 icap_preview_enable off
5601 NAME: icap_preview_size
5604 LOC: Adaptation::Icap::TheConfig.preview_size
5607 The default size of preview data to be sent to the ICAP server.
5608 -1 means no preview. This value might be overwritten on a per server
5609 basis by OPTIONS requests.
5612 NAME: icap_default_options_ttl
5615 LOC: Adaptation::Icap::TheConfig.default_options_ttl
5618 The default TTL value for ICAP OPTIONS responses that don't have
5619 an Options-TTL header.
5622 NAME: icap_persistent_connections
5626 LOC: Adaptation::Icap::TheConfig.reuse_connections
5629 Whether or not Squid should use persistent connections to
5633 NAME: icap_send_client_ip
5637 LOC: Adaptation::Icap::TheConfig.send_client_ip
5640 This adds the header "X-Client-IP" to ICAP requests.
5643 NAME: icap_send_client_username
5647 LOC: Adaptation::Icap::TheConfig.send_client_username
5650 This sends authenticated HTTP client username (if available) to
5651 the ICAP service. The username value is encoded based on the
5652 icap_client_username_encode option and is sent using the header
5653 specified by the icap_client_username_header option.
5656 NAME: icap_client_username_header
5659 LOC: Adaptation::Icap::TheConfig.client_username_header
5660 DEFAULT: X-Client-Username
5662 ICAP request header name to use for send_client_username.
5665 NAME: icap_client_username_encode
5669 LOC: Adaptation::Icap::TheConfig.client_username_encode
5672 Whether to base64 encode the authenticated client username.
5676 TYPE: icap_service_type
5678 LOC: Adaptation::Icap::TheConfig
5681 Defines a single ICAP service using the following format:
5683 icap_service service_name vectoring_point [options] service_url
5686 an opaque identifier which must be unique in squid.conf
5688 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
5689 This specifies at which point of transaction processing the
5690 ICAP service should be activated. *_postcache vectoring points
5691 are not yet supported.
5693 service_url: icap://servername:port/servicepath
5694 ICAP server and service location.
5696 ICAP does not allow a single service to handle both REQMOD and RESPMOD
5697 transactions. Squid does not enforce that requirement. You can specify
5698 services with the same service_url and different vectoring_points. You
5699 can even specify multiple identical services as long as their
5700 service_names differ.
5703 Service options are separated by white space. ICAP services support
5704 the following name=value options:
5707 If set to 'on' or '1', the ICAP service is treated as
5708 optional. If the service cannot be reached or malfunctions,
5709 Squid will try to ignore any errors and process the message as
5710 if the service was not enabled. No all ICAP errors can be
5711 bypassed. If set to 0, the ICAP service is treated as
5712 essential and all ICAP errors will result in an error page
5713 returned to the HTTP client.
5715 Bypass is off by default: services are treated as essential.
5718 If set to 'on' or '1', the ICAP service is allowed to
5719 dynamically change the current message adaptation plan by
5720 returning a chain of services to be used next. The services
5721 are specified using the X-Next-Services ICAP response header
5722 value, formatted as a comma-separated list of service names.
5723 Each named service should be configured in squid.conf and
5724 should have the same method and vectoring point as the current
5725 ICAP transaction. Services violating these rules are ignored.
5726 An empty X-Next-Services value results in an empty plan which
5727 ends the current adaptation.
5729 Routing is not allowed by default: the ICAP X-Next-Services
5730 response header is ignored.
5732 Older icap_service format without optional named parameters is
5733 deprecated but supported for backward compatibility.
5736 icap_service svcBlocker reqmod_precache bypass=0 icap://icap1.mydomain.net:1344/reqmod
5737 icap_service svcLogger reqmod_precache routing=on icap://icap2.mydomain.net:1344/respmod
5741 TYPE: icap_class_type
5746 This deprecated option was documented to define an ICAP service
5747 chain, even though it actually defined a set of similar, redundant
5748 services, and the chains were not supported.
5750 To define a set of redundant services, please use the
5751 adaptation_service_set directive. For service chains, use
5752 adaptation_service_chain.
5756 TYPE: icap_access_type
5761 This option is deprecated. Please use adaptation_access, which
5762 has the same ICAP functionality, but comes with better
5763 documentation, and eCAP support.
5768 -----------------------------------------------------------------------------
5775 LOC: Adaptation::Ecap::TheConfig.onoff
5778 Controls whether eCAP support is enabled.
5782 TYPE: ecap_service_type
5784 LOC: Adaptation::Ecap::TheConfig
5787 Defines a single eCAP service
5789 ecap_service servicename vectoring_point bypass service_url
5791 vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
5792 This specifies at which point of transaction processing the
5793 eCAP service should be activated. *_postcache vectoring points
5794 are not yet supported.
5796 If set to 1, the eCAP service is treated as optional. If the
5797 service cannot be reached or malfunctions, Squid will try to
5798 ignore any errors and process the message as if the service
5799 was not enabled. No all eCAP errors can be bypassed.
5800 If set to 0, the eCAP service is treated as essential and all
5801 eCAP errors will result in an error page returned to the
5803 service_url = ecap://vendor/service_name?custom&cgi=style¶meters=optional
5806 ecap_service service_1 reqmod_precache 0 ecap://filters-R-us/leakDetector?on_error=block
5807 ecap_service service_2 respmod_precache 1 icap://filters-R-us/virusFilter?config=/etc/vf.cfg
5810 NAME: loadable_modules
5812 IFDEF: USE_LOADABLE_MODULES
5813 LOC: Config.loadable_module_names
5816 Instructs Squid to load the specified dynamic module(s) or activate
5817 preloaded module(s).
5819 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
5823 MESSAGE ADAPTATION OPTIONS
5824 -----------------------------------------------------------------------------
5827 NAME: adaptation_service_set
5828 TYPE: adaptation_service_set_type
5829 IFDEF: USE_ADAPTATION
5834 Configures an ordered set of similar, redundant services. This is
5835 useful when hot standby or backup adaptation servers are available.
5837 adaptation_service_set set_name service_name1 service_name2 ...
5839 The named services are used in the set declaration order. The first
5840 applicable adaptation service from the set is used first. The next
5841 applicable service is tried if and only if the transaction with the
5842 previous service fails and the message waiting to be adapted is still
5845 When adaptation starts, broken services are ignored as if they were
5846 not a part of the set. A broken service is a down optional service.
5848 The services in a set must be attached to the same vectoring point
5849 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
5851 If all services in a set are optional then adaptation failures are
5852 bypassable. If all services in the set are essential, then a
5853 transaction failure with one service may still be retried using
5854 another service from the set, but when all services fail, the master
5855 transaction fails as well.
5857 A set may contain a mix of optional and essential services, but that
5858 is likely to lead to surprising results because broken services become
5859 ignored (see above), making previously bypassable failures fatal.
5860 Technically, it is the bypassability of the last failed service that
5863 See also: adaptation_access adaptation_service_chain
5866 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
5867 adaptation service_set svcLogger loggerLocal loggerRemote
5870 NAME: adaptation_service_chain
5871 TYPE: adaptation_service_chain_type
5872 IFDEF: USE_ADAPTATION
5877 Configures a list of complementary services that will be applied
5878 one-by-one, forming an adaptation chain or pipeline. This is useful
5879 when Squid must perform different adaptations on the same message.
5881 adaptation_service_chain chain_name service_name1 svc_name2 ...
5883 The named services are used in the chain declaration order. The first
5884 applicable adaptation service from the chain is used first. The next
5885 applicable service is applied to the successful adaptation results of
5886 the previous service in the chain.
5888 When adaptation starts, broken services are ignored as if they were
5889 not a part of the chain. A broken service is a down optional service.
5891 Request satisfaction terminates the adaptation chain because Squid
5892 does not currently allow declaration of RESPMOD services at the
5893 "reqmod_precache" vectoring point (see icap_service or ecap_service).
5895 The services in a chain must be attached to the same vectoring point
5896 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
5898 A chain may contain a mix of optional and essential services. If an
5899 essential adaptation fails (or the failure cannot be bypassed for
5900 other reasons), the master transaction fails. Otherwise, the failure
5901 is bypassed as if the failed adaptation service was not in the chain.
5903 See also: adaptation_access adaptation_service_set
5906 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
5909 NAME: adaptation_access
5910 TYPE: adaptation_access_type
5911 IFDEF: USE_ADAPTATION
5915 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
5917 adaptation_access service_name allow|deny [!]aclname...
5918 adaptation_access set_name allow|deny [!]aclname...
5920 At each supported vectoring point, the adaptation_access
5921 statements are processed in the order they appear in this
5922 configuration file. Statements pointing to the following services
5923 are ignored (i.e., skipped without checking their ACL):
5925 - services serving different vectoring points
5926 - "broken-but-bypassable" services
5927 - "up" services configured to ignore such transactions
5928 (e.g., based on the ICAP Transfer-Ignore header).
5930 When a set_name is used, all services in the set are checked
5931 using the same rules, to find the first applicable one. See
5932 adaptation_service_set for details.
5934 If an access list is checked and there is a match, the
5935 processing stops: For an "allow" rule, the corresponding
5936 adaptation service is used for the transaction. For a "deny"
5937 rule, no adaptation service is activated.
5939 It is currently not possible to apply more than one adaptation
5940 service at the same vectoring point to the same HTTP transaction.
5942 See also: icap_service and ecap_service
5945 adaptation_access service_1 allow all
5948 NAME: adaptation_service_iteration_limit
5950 IFDEF: USE_ADAPTATION
5951 LOC: Adaptation::Config::service_iteration_limit
5954 Limits the number of iterations allowed when applying adaptation
5955 services to a message. If your longest adaptation set or chain
5956 may have more than 16 services, increase the limit beyond its
5957 default value of 16. If detecting infinite iteration loops sooner
5958 is critical, make the iteration limit match the actual number
5959 of services in your longest adaptation set or chain.
5961 Infinite adaptation loops are most likely with routing services.
5963 See also: icap_service routing=1
5966 NAME: adaptation_masterx_shared_names
5968 IFDEF: USE_ADAPTATION
5969 LOC: Adaptation::Config::masterx_shared_name
5972 For each master transaction (i.e., the HTTP request and response
5973 sequence, including all related ICAP and eCAP exchanges), Squid
5974 maintains a table of metadata. The table entries are (name, value)
5975 pairs shared among eCAP and ICAP exchanges. The table is destroyed
5976 with the master transaction.
5978 This option specifies the table entry names that Squid must accept
5979 from and forward to the adaptation transactions.
5981 An ICAP REQMOD or RESPMOD transaction may set an entry in the
5982 shared table by returning an ICAP header field with a name
5983 specified in adaptation_masterx_shared_names. Squid will store
5984 and forward that ICAP header field to subsequent ICAP
5985 transactions within the same master transaction scope.
5987 Only one shared entry name is supported at this time.
5990 # share authentication information among ICAP services
5991 adaptation_masterx_shared_names X-Subscriber-ID
5997 LOC: Adaptation::Icap::TheConfig.repeat
5999 DEFAULT_IF_NONE: deny all
6001 This ACL determines which retriable ICAP transactions are
6002 retried. Transactions that received a complete ICAP response
6003 and did not have to consume or produce HTTP bodies to receive
6004 that response are usually retriable.
6006 icap_retry allow|deny [!]aclname ...
6008 Squid automatically retries some ICAP I/O timeouts and errors
6009 due to persistent connection race conditions.
6011 See also: icap_retry_limit
6014 NAME: icap_retry_limit
6017 LOC: Adaptation::Icap::TheConfig.repeat_limit
6020 Limits the number of retries allowed. When set to zero (default),
6021 no retries are allowed.
6023 Communication errors due to persistent connection race
6024 conditions are unavoidable, automatically retried, and do not
6025 count against this limit.
6027 See also: icap_retry
6033 -----------------------------------------------------------------------------
6036 NAME: check_hostnames
6039 LOC: Config.onoff.check_hostnames
6041 For security and stability reasons Squid can check
6042 hostnames for Internet standard RFC compliance. If you want
6043 Squid to perform these checks turn this directive on.
6046 NAME: allow_underscore
6049 LOC: Config.onoff.allow_underscore
6051 Underscore characters is not strictly allowed in Internet hostnames
6052 but nevertheless used by many sites. Set this to off if you want
6053 Squid to be strict about the standard.
6054 This check is performed only when check_hostnames is set to on.
6057 NAME: cache_dns_program
6059 IFDEF: USE_DNSSERVERS
6060 DEFAULT: @DEFAULT_DNSSERVER@
6061 LOC: Config.Program.dnsserver
6063 Specify the location of the executable for dnslookup process.
6068 IFDEF: USE_DNSSERVERS
6070 LOC: Config.dnsChildren
6072 The number of processes spawn to service DNS name lookups.
6073 For heavily loaded caches on large servers, you should
6074 probably increase this value to at least 10. The maximum
6075 is 32. The default is 5.
6077 You must have at least one dnsserver process.
6080 NAME: dns_retransmit_interval
6083 LOC: Config.Timeout.idns_retransmit
6084 IFDEF: !USE_DNSSERVERS
6086 Initial retransmit interval for DNS queries. The interval is
6087 doubled each time all configured DNS servers have been tried.
6094 LOC: Config.Timeout.idns_query
6095 IFDEF: !USE_DNSSERVERS
6097 DNS Query timeout. If no response is received to a DNS query
6098 within this time all DNS servers for the queried domain
6099 are assumed to be unavailable.
6106 LOC: Config.onoff.res_defnames
6108 Normally the RES_DEFNAMES resolver option is disabled
6109 (see res_init(3)). This prevents caches in a hierarchy
6110 from interpreting single-component hostnames locally. To allow
6111 Squid to handle single-component names, enable this option.
6114 NAME: dns_nameservers
6117 LOC: Config.dns_nameservers
6119 Use this if you want to specify a list of DNS name servers
6120 (IP addresses) to use instead of those given in your
6121 /etc/resolv.conf file.
6122 On Windows platforms, if no value is specified here or in
6123 the /etc/resolv.conf file, the list of DNS name servers are
6124 taken from the Windows registry, both static and dynamic DHCP
6125 configurations are supported.
6127 Example: dns_nameservers 10.0.0.1 192.172.0.4
6132 DEFAULT: @DEFAULT_HOSTS@
6133 LOC: Config.etcHostsPath
6135 Location of the host-local IP name-address associations
6136 database. Most Operating Systems have such a file on different
6138 - Un*X & Linux: /etc/hosts
6139 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
6140 (%SystemRoot% value install default is c:\winnt)
6141 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
6142 (%SystemRoot% value install default is c:\windows)
6143 - Windows 9x/Me: %windir%\hosts
6144 (%windir% value is usually c:\windows)
6145 - Cygwin: /etc/hosts
6147 The file contains newline-separated definitions, in the
6148 form ip_address_in_dotted_form name [name ...] names are
6149 whitespace-separated. Lines beginning with an hash (#)
6150 character are comments.
6152 The file is checked at startup and upon configuration.
6153 If set to 'none', it won't be checked.
6154 If append_domain is used, that domain will be added to
6155 domain-local (i.e. not containing any dot character) host
6161 LOC: Config.appendDomain
6164 Appends local domain name to hostnames without any dots in
6165 them. append_domain must begin with a period.
6167 Be warned there are now Internet names with no dots in
6168 them using only top-domain names, so setting this may
6169 cause some Internet sites to become unavailable.
6172 append_domain .yourdomain.com
6175 NAME: ignore_unknown_nameservers
6177 LOC: Config.onoff.ignore_unknown_nameservers
6180 By default Squid checks that DNS responses are received
6181 from the same IP addresses they are sent to. If they
6182 don't match, Squid ignores the response and writes a warning
6183 message to cache.log. You can allow responses from unknown
6184 nameservers by setting this option to 'off'.
6187 NAME: dns_v4_fallback
6190 LOC: Config.onoff.dns_require_A
6192 Standard practice with DNS is to lookup either A or AAAA records
6193 and use the results if it succeeds. Only looking up the other if
6194 the first attempt fails or otherwise produces no results.
6196 That policy however will cause squid to produce error pages for some
6197 servers that advertise AAAA but are unreachable over IPv6.
6199 If this is ON squid will always lookup both AAAA and A, using both.
6200 If this is OFF squid will lookup AAAA and only try A if none found.
6202 WARNING: There are some possibly unwanted side-effects with this on:
6203 *) Doubles the load placed by squid on the DNS network.
6204 *) May negatively impact connection delay times.
6208 COMMENT: (number of entries)
6211 LOC: Config.ipcache.size
6218 LOC: Config.ipcache.low
6225 LOC: Config.ipcache.high
6227 The size, low-, and high-water marks for the IP cache.
6230 NAME: fqdncache_size
6231 COMMENT: (number of entries)
6234 LOC: Config.fqdncache.size
6236 Maximum number of FQDN cache entries.
6241 -----------------------------------------------------------------------------
6248 LOC: Config.onoff.mem_pools
6250 If set, Squid will keep pools of allocated (but unused) memory
6251 available for future use. If memory is a premium on your
6252 system and you believe your malloc library outperforms Squid
6253 routines, disable this.
6256 NAME: memory_pools_limit
6260 LOC: Config.MemPools.limit
6262 Used only with memory_pools on:
6263 memory_pools_limit 50 MB
6265 If set to a non-zero value, Squid will keep at most the specified
6266 limit of allocated (but unused) memory in memory pools. All free()
6267 requests that exceed this limit will be handled by your malloc
6268 library. Squid does not pre-allocate any memory, just safe-keeps
6269 objects that otherwise would be free()d. Thus, it is safe to set
6270 memory_pools_limit to a reasonably high value even if your
6271 configuration will use less memory.
6273 If set to zero, Squid will keep all memory it can. That is, there
6274 will be no limit on the total amount of memory used for safe-keeping.
6276 To disable memory allocation optimization, do not set
6277 memory_pools_limit to 0. Set memory_pools to "off" instead.
6279 An overhead for maintaining memory pools is not taken into account
6280 when the limit is checked. This overhead is close to four bytes per
6281 object kept. However, pools may actually _save_ memory because of
6282 reduced memory thrashing in your malloc library.
6286 COMMENT: on|off|transparent|truncate|delete
6289 LOC: opt_forwarded_for
6291 If set to "on", Squid will append your client's IP address
6292 in the HTTP requests it forwards. By default it looks like:
6294 X-Forwarded-For: 192.1.2.3
6296 If set to "off", it will appear as
6298 X-Forwarded-For: unknown
6300 If set to "transparent", Squid will not alter the
6301 X-Forwarded-For header in any way.
6303 If set to "delete", Squid will delete the entire
6304 X-Forwarded-For header.
6306 If set to "truncate", Squid will remove all existing
6307 X-Forwarded-For entries, and place itself as the sole entry.
6310 NAME: cachemgr_passwd
6311 TYPE: cachemgrpasswd
6313 LOC: Config.passwd_list
6315 Specify passwords for cachemgr operations.
6317 Usage: cachemgr_passwd password action action ...
6319 Some valid actions are (see cache manager menu for a full list):
6359 * Indicates actions which will not be performed without a
6360 valid password, others can be performed if not listed here.
6362 To disable an action, set the password to "disable".
6363 To allow performing an action without a password, set the
6366 Use the keyword "all" to set the same password for all actions.
6369 cachemgr_passwd secret shutdown
6370 cachemgr_passwd lesssssssecret info stats/objects
6371 cachemgr_passwd disable all
6378 LOC: Config.onoff.client_db
6380 If you want to disable collecting per-client statistics,
6381 turn off client_db here.
6384 NAME: refresh_all_ims
6388 LOC: Config.onoff.refresh_all_ims
6390 When you enable this option, squid will always check
6391 the origin server for an update when a client sends an
6392 If-Modified-Since request. Many browsers use IMS
6393 requests when the user requests a reload, and this
6394 ensures those clients receive the latest version.
6396 By default (off), squid may return a Not Modified response
6397 based on the age of the cached version.
6400 NAME: reload_into_ims
6401 IFDEF: HTTP_VIOLATIONS
6405 LOC: Config.onoff.reload_into_ims
6407 When you enable this option, client no-cache or ``reload''
6408 requests will be changed to If-Modified-Since requests.
6409 Doing this VIOLATES the HTTP standard. Enabling this
6410 feature could make you liable for problems which it
6413 see also refresh_pattern for a more selective approach.
6416 NAME: maximum_single_addr_tries
6418 LOC: Config.retry.maxtries
6421 This sets the maximum number of connection attempts for a
6422 host that only has one address (for multiple-address hosts,
6423 each address is tried once).
6425 The default value is one attempt, the (not recommended)
6426 maximum is 255 tries. A warning message will be generated
6427 if it is set to a value greater than ten.
6429 Note: This is in addition to the request re-forwarding which
6430 takes place if Squid fails to get a satisfying response.
6433 NAME: retry_on_error
6435 LOC: Config.retry.onerror
6438 If set to on Squid will automatically retry requests when
6439 receiving an error response. This is mainly useful if you
6440 are in a complex cache hierarchy to work around access
6444 NAME: as_whois_server
6446 LOC: Config.as_whois_server
6447 DEFAULT: whois.ra.net
6448 DEFAULT_IF_NONE: whois.ra.net
6450 WHOIS server to query for AS numbers. NOTE: AS numbers are
6451 queried only when Squid starts up, not for every request.
6456 LOC: Config.onoff.offline
6459 Enable this option and Squid will never try to validate cached
6463 NAME: uri_whitespace
6464 TYPE: uri_whitespace
6465 LOC: Config.uri_whitespace
6468 What to do with requests that have whitespace characters in the
6471 strip: The whitespace characters are stripped out of the URL.
6472 This is the behavior recommended by RFC2396.
6473 deny: The request is denied. The user receives an "Invalid
6475 allow: The request is allowed and the URI is not changed. The
6476 whitespace characters remain in the URI. Note the
6477 whitespace is passed to redirector processes if they
6479 encode: The request is allowed and the whitespace characters are
6480 encoded according to RFC1738. This could be considered
6481 a violation of the HTTP/1.1
6482 RFC because proxies are not allowed to rewrite URI's.
6483 chop: The request is allowed and the URI is chopped at the
6484 first whitespace. This might also be considered a
6490 LOC: Config.chroot_dir
6493 Specifies a directiry where Squid should do a chroot() while
6494 initializing. This also causes Squid to fully drop root
6495 privileges after initializing. This means, for example, if you
6496 use a HTTP port less than 1024 and try to reconfigure, you may
6497 get an error saying that Squid can not open the port.
6500 NAME: balance_on_multiple_ip
6502 LOC: Config.onoff.balance_on_multiple_ip
6505 Modern IP resolvers in squid sort lookup results by preferred access.
6506 By default squid will use these IP in order and only rotates to
6507 the next listed when the most preffered fails.
6509 Some load balancing servers based on round robin DNS have been
6510 found not to preserve user session state across requests
6511 to different IP addresses.
6513 Enabling this directive Squid rotates IP's per request.
6516 NAME: pipeline_prefetch
6518 LOC: Config.onoff.pipeline_prefetch
6521 To boost the performance of pipelined requests to closer
6522 match that of a non-proxied environment Squid can try to fetch
6523 up to two requests in parallel from a pipeline.
6525 Defaults to off for bandwidth management and access logging
6529 NAME: high_response_time_warning
6532 LOC: Config.warnings.high_rptm
6535 If the one-minute median response time exceeds this value,
6536 Squid prints a WARNING with debug level 0 to get the
6537 administrators attention. The value is in milliseconds.
6540 NAME: high_page_fault_warning
6542 LOC: Config.warnings.high_pf
6545 If the one-minute average page fault rate exceeds this
6546 value, Squid prints a WARNING with debug level 0 to get
6547 the administrators attention. The value is in page faults
6551 NAME: high_memory_warning
6553 LOC: Config.warnings.high_memory
6556 If the memory usage (as determined by mallinfo) exceeds
6557 this amount, Squid prints a WARNING with debug level 0 to get
6558 the administrators attention.
6561 NAME: sleep_after_fork
6562 COMMENT: (microseconds)
6564 LOC: Config.sleep_after_fork
6567 When this is set to a non-zero value, the main Squid process
6568 sleeps the specified number of microseconds after a fork()
6569 system call. This sleep may help the situation where your
6570 system reports fork() failures due to lack of (virtual)
6571 memory. Note, however, if you have a lot of child
6572 processes, these sleep delays will add up and your
6573 Squid will not service requests for some amount of time
6574 until all the child processes have been started.
6575 On Windows value less then 1000 (1 milliseconds) are
6579 NAME: windows_ipaddrchangemonitor
6580 IFDEF: _SQUID_MSWIN_
6584 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
6586 On Windows Squid by default will monitor IP address changes and will
6587 reconfigure itself after any detected event. This is very useful for
6588 proxies connected to internet with dial-up interfaces.
6589 In some cases (a Proxy server acting as VPN gateway is one) it could be
6590 desiderable to disable this behaviour setting this to 'off'.
6591 Note: after changing this, Squid service must be restarted.