3 # $Id: cf.data.pre,v 1.472 2007/09/21 15:16:42 hno Exp $
5 # SQUID Web Proxy Cache http://www.squid-cache.org/
6 # ----------------------------------------------------------
8 # Squid is the result of efforts by numerous individuals from
9 # the Internet community; see the CONTRIBUTORS file for full
10 # details. Many organizations have provided support for Squid's
11 # development; see the SPONSORS file for full details. Squid is
12 # Copyrighted (C) 2000 by the Regents of the University of
13 # California; see the COPYRIGHT file for full details. Squid
14 # incorporates software developed and/or copyrighted by other
15 # sources; see the CREDITS file for full details.
17 # This program is free software; you can redistribute it and/or modify
18 # it under the terms of the GNU General Public License as published by
19 # the Free Software Foundation; either version 2 of the License, or
20 # (at your option) any later version.
22 # This program is distributed in the hope that it will be useful,
23 # but WITHOUT ANY WARRANTY; without even the implied warranty of
24 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
25 # GNU General Public License for more details.
27 # You should have received a copy of the GNU General Public License
28 # along with this program; if not, write to the Free Software
29 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
33 WELCOME TO SQUID @VERSION@
34 ----------------------------
36 This is the default Squid configuration file. You may wish
37 to look at the Squid home page (http://www.squid-cache.org/)
38 for the FAQ and other documentation.
40 The default Squid config file shows what the defaults for
41 various options happen to be. If you don't need to change the
42 default, you shouldn't uncomment the line. Doing so may cause
43 run-time problems. In some cases "none" refers to no default
44 setting at all, while in other cases it refers to a valid
45 option - the comments for that keyword indicate if this is the
51 OPTIONS FOR AUTHENTICATION
52 -----------------------------------------------------------------------------
57 LOC: Config.authConfiguration
60 This is used to define parameters for the various authentication
61 schemes supported by Squid.
63 format: auth_param scheme parameter [setting]
65 The order in which authentication schemes are presented to the client is
66 dependent on the order the scheme first appears in config file. IE
67 has a bug (it's not RFC 2617 compliant) in that it will use the basic
68 scheme if basic is the first entry presented, even if more secure
69 schemes are presented. For now use the order in the recommended
70 settings section below. If other browsers have difficulties (don't
71 recognize the schemes offered even if you are using basic) either
72 put basic first, or disable the other schemes (by commenting out their
75 Once an authentication scheme is fully configured, it can only be
76 shutdown by shutting squid down and restarting. Changes can be made on
77 the fly and activated with a reconfigure. I.E. You can change to a
78 different helper, but not unconfigure the helper completely.
80 Please note that while this directive defines how Squid processes
81 authentication it does not automatically activate authentication.
82 To use authentication you must in addition make use of ACLs based
83 on login name in http_access (proxy_auth, proxy_auth_regex or
84 external with %LOGIN used in the format tag). The browser will be
85 challenged for authentication on the first such acl encountered
86 in http_access processing and will also be re-challenged for new
87 login credentials if the request is being denied by a proxy_auth
90 WARNING: authentication can't be used in a transparently intercepting
91 proxy as the client then thinks it is talking to an origin server and
92 not the proxy. This is a limitation of bending the TCP/IP protocol to
93 transparently intercepting port 80, not a limitation in Squid.
95 === Parameters for the basic scheme follow. ===
98 Specify the command for the external authenticator. Such a program
99 reads a line containing "username password" and replies "OK" or
100 "ERR" in an endless loop. "ERR" responses may optionally be followed
101 by a error description available as %m in the returned error page.
102 If you use an authenticator, make sure you have 1 acl of type proxy_auth.
104 By default, the basic authentication scheme is not used unless a
105 program is specified.
107 If you want to use the traditional NCSA proxy authentication, set
108 this line to something like
110 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
112 "children" numberofchildren
113 The number of authenticator processes to spawn. If you start too few
114 Squid will have to wait for them to process a backlog of credential
115 verifications, slowing it down. When password verifications are
116 done via a (slow) network you are likely to need lots of
117 authenticator processes.
118 auth_param basic children 5
120 "concurrency" concurrency
121 The number of concurrent requests the helper can process.
122 The default of 0 is used for helpers who only supports
123 one request at a time. Setting this changes the protocol used to
124 include a channel number first on the request/response line, allowing
125 multiple requests to be sent to the same helper in parallell without
126 wating for the response.
127 Must not be set unless it's known the helper supports this.
128 auth_param basic concurrency 0
131 Specifies the realm name which is to be reported to the
132 client for the basic proxy authentication scheme (part of
133 the text the user will see when prompted their username and
134 password). There is no default.
135 auth_param basic realm Squid proxy-caching web server
137 "credentialsttl" timetolive
138 Specifies how long squid assumes an externally validated
139 username:password pair is valid for - in other words how
140 often the helper program is called for that user. Set this
141 low to force revalidation with short lived passwords. Note
142 setting this high does not impact your susceptibility
143 to replay attacks unless you are using an one-time password
144 system (such as SecureID). If you are using such a system,
145 you will be vulnerable to replay attacks unless you also
146 use the max_user_ip ACL in an http_access rule.
148 "casesensitive" on|off
149 Specifies if usernames are case sensitive. Most user databases are
150 case insensitive allowing the same username to be spelled using both
151 lower and upper case letters, but some are case sensitive. This
152 makes a big difference for user_max_ip ACL processing and similar.
153 auth_param basic casesensitive off
155 === Parameters for the digest scheme follow ===
158 Specify the command for the external authenticator. Such
159 a program reads a line containing "username":"realm" and
160 replies with the appropriate H(A1) value hex encoded or
161 ERR if the user (or his H(A1) hash) does not exists.
162 See rfc 2616 for the definition of H(A1).
163 "ERR" responses may optionally be followed by a error description
164 available as %m in the returned error page.
166 By default, the digest authentication scheme is not used unless a
167 program is specified.
169 If you want to use a digest authenticator, set this line to
172 auth_param digest program @DEFAULT_PREFIX@/bin/digest_auth_pw @DEFAULT_PREFIX@/etc/digpass
174 "children" numberofchildren
175 The number of authenticator processes to spawn (no default).
176 If you start too few Squid will have to wait for them to
177 process a backlog of H(A1) calculations, slowing it down.
178 When the H(A1) calculations are done via a (slow) network
179 you are likely to need lots of authenticator processes.
180 auth_param digest children 5
183 Specifies the realm name which is to be reported to the
184 client for the digest proxy authentication scheme (part of
185 the text the user will see when prompted their username and
186 password). There is no default.
187 auth_param digest realm Squid proxy-caching web server
189 "nonce_garbage_interval" timeinterval
190 Specifies the interval that nonces that have been issued
191 to client_agent's are checked for validity.
193 "nonce_max_duration" timeinterval
194 Specifies the maximum length of time a given nonce will be
197 "nonce_max_count" number
198 Specifies the maximum number of times a given nonce can be
201 "nonce_strictness" on|off
202 Determines if squid requires strict increment-by-1 behavior
203 for nonce counts, or just incrementing (off - for use when
204 useragents generate nonce counts that occasionally miss 1
205 (ie, 1,2,4,6)). Default off.
207 "check_nonce_count" on|off
208 This directive if set to off can disable the nonce count check
209 completely to work around buggy digest qop implementations in
210 certain mainstream browser versions. Default on to check the
211 nonce count to protect from authentication replay attacks.
213 "post_workaround" on|off
214 This is a workaround to certain buggy browsers who sends
215 an incorrect request digest in POST requests when reusing
216 the same nonce as acquired earlier on a GET request.
218 === NTLM scheme options follow ===
221 Specify the command for the external NTLM authenticator.
222 Such a program reads exchanged NTLMSSP packets with
223 the browser via Squid until authentication is completed.
224 If you use an NTLM authenticator, make sure you have 1 acl
225 of type proxy_auth. By default, the NTLM authenticator_program
228 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
230 "children" numberofchildren
231 The number of authenticator processes to spawn (no default).
232 If you start too few Squid will have to wait for them to
233 process a backlog of credential verifications, slowing it
234 down. When credential verifications are done via a (slow)
235 network you are likely to need lots of authenticator
238 auth_param ntlm children 5
241 If you experience problems with PUT/POST requests when using the
242 Negotiate authentication scheme then you can try setting this to
243 off. This will cause Squid to forcibly close the connection on
244 the initial requests where the browser asks which schemes are
245 supported by the proxy.
247 auth_param ntlm keep_alive on
249 === Options for configuring the NEGOTIATE auth-scheme follow ===
252 Specify the command for the external Negotiate authenticator.
253 This protocol is used in Microsoft Active-Directory enabled setups with
254 the Microsoft Internet Explorer or Mozilla Firefox browsers.
255 Its main purpose is to exchange credentials with the Squid proxy
256 using the Kerberos mechanisms.
257 If you use a Negotiate authenticator, make sure you have at least one acl
258 of type proxy_auth active. By default, the negotiate authenticator_program
260 The only supported program for this role is the ntlm_auth
261 program distributed as part of Samba, version 4 or later.
263 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
265 "children" numberofchildren
266 The number of authenticator processes to spawn (no default).
267 If you start too few Squid will have to wait for them to
268 process a backlog of credential verifications, slowing it
269 down. When crendential verifications are done via a (slow)
270 network you are likely to need lots of authenticator
272 auth_param negotiate children 5
275 If you experience problems with PUT/POST requests when using the
276 Negotiate authentication scheme then you can try setting this to
277 off. This will cause Squid to forcibly close the connection on
278 the initial requests where the browser asks which schemes are
279 supported by the proxy.
281 auth_param negotiate keep_alive on
284 #Recommended minimum configuration per scheme:
285 #auth_param negotiate program <uncomment and complete this line to activate>
286 #auth_param negotiate children 5
287 #auth_param negotiate keep_alive on
288 #auth_param ntlm program <uncomment and complete this line to activate>
289 #auth_param ntlm children 5
290 #auth_param ntlm keep_alive on
291 #auth_param digest program <uncomment and complete this line>
292 #auth_param digest children 5
293 #auth_param digest realm Squid proxy-caching web server
294 #auth_param digest nonce_garbage_interval 5 minutes
295 #auth_param digest nonce_max_duration 30 minutes
296 #auth_param digest nonce_max_count 50
297 #auth_param basic program <uncomment and complete this line>
298 #auth_param basic children 5
299 #auth_param basic realm Squid proxy-caching web server
300 #auth_param basic credentialsttl 2 hours
304 NAME: authenticate_cache_garbage_interval
307 LOC: Config.authenticateGCInterval
309 The time period between garbage collection across the username cache.
310 This is a tradeoff between memory utilization (long intervals - say
311 2 days) and CPU (short intervals - say 1 minute). Only change if you
315 NAME: authenticate_ttl
318 LOC: Config.authenticateTTL
320 The time a user & their credentials stay in the logged in
321 user cache since their last request. When the garbage
322 interval passes, all user credentials that have passed their
323 TTL are removed from memory.
326 NAME: authenticate_ip_ttl
328 LOC: Config.authenticateIpTTL
331 If you use proxy authentication and the 'max_user_ip' ACL,
332 this directive controls how long Squid remembers the IP
333 addresses associated with each user. Use a small value
334 (e.g., 60 seconds) if your users might change addresses
335 quickly, as is the case with dialups. You might be safe
336 using a larger value (e.g., 2 hours) in a corporate LAN
337 environment with relatively static address assignments.
342 -----------------------------------------------------------------------------
345 NAME: external_acl_type
346 TYPE: externalAclHelper
347 LOC: Config.externalAclHelperList
350 This option defines external acl classes using a helper program
351 to look up the status
353 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
357 ttl=n TTL in seconds for cached results (defaults to 3600
360 TTL for cached negative lookups (default same
362 children=n Number of acl helper processes spawn to service
363 external acl lookups of this type. (default 5)
364 concurrency=n concurrency level per process. Only used with helpers
365 capable of processing more than one query at a time.
366 cache=n result cache size, 0 is unbounded (default)
367 grace=n Percentage remaining of TTL where a refresh of a
368 cached entry should be initiated without needing to
369 wait for a new reply. (default 0 for no grace period)
370 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
372 FORMAT specifications
374 %LOGIN Authenticated user login name
375 %EXT_USER Username from external acl
376 %IDENT Ident user name
378 %SRCPORT Client source port
381 %PROTO Requested protocol
383 %PATH Requested URL path
384 %METHOD Request method
385 %MYADDR Squid interface address
386 %MYPORT Squid http_port number
387 %PATH Requested URL-path (including query-string if any)
388 %USER_CERT SSL User certificate in PEM format
389 %USER_CERTCHAIN SSL User certificate chain in PEM format
390 %USER_CERT_xx SSL User certificate subject attribute xx
391 %USER_CA_xx SSL User certificate issuer attribute xx
392 %{Header} HTTP request header
393 %{Hdr:member} HTTP request header list member
395 HTTP request header list member using ; as
396 list separator. ; can be any non-alphanumeric
399 In addition to the above, any string specified in the referencing
400 acl will also be included in the helper request line, after the
401 specified formats (see the "acl external" directive)
403 The helper receives lines per the above format specification,
404 and returns lines starting with OK or ERR indicating the validity
405 of the request and optionally followed by additional keywords with
408 General result syntax:
410 OK/ERR keyword=value ...
414 user= The users name (login)
415 password= The users password (for login= cache_peer option)
416 message= Message describing the reason. Available as %o
418 tag= Apply a tag to a request (for both ERR and OK results)
419 Only sets a tag, does not alter existing tags.
420 log= String to be logged in access.log. Available as
421 %ea in logformat specifications
423 If protocol=3.0 (the default) then URL escaping is used to protect
424 each value in both requests and responses.
426 If using protocol=2.5 then all values need to be enclosed in quotes
427 if they may contain whitespace, or the whitespace escaped using \.
428 And quotes or \ characters within the keyword value must be \ escaped.
430 When using the concurrency= option the protocol is changed by
431 introducing a query channel tag infront of the request/response.
432 The query channel tag is a number between 0 and concurrency-1.
440 Defining an Access List
442 acl aclname acltype string1 ...
443 acl aclname acltype "file" ...
445 when using "file", the file should contain one item per line
447 acltype is one of the types described below
449 By default, regular expressions are CASE-SENSITIVE. To make
450 them case-insensitive, use the -i option.
452 acl aclname src ip-address/netmask ... (clients IP address)
453 acl aclname src addr1-addr2/netmask ... (range of addresses)
454 acl aclname dst ip-address/netmask ... (URL host's IP address)
455 acl aclname myip ip-address/netmask ... (local socket IP address)
457 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
458 # The arp ACL requires the special configure option --enable-arp-acl.
459 # Furthermore, the ARP ACL code is not portable to all operating systems.
460 # It works on Linux, Solaris, Windows, FreeBSD, and some other *BSD variants.
462 # NOTE: Squid can only determine the MAC address for clients that are on
463 # the same subnet. If the client is on a different subnet, then Squid cannot
464 # find out its MAC address.
466 acl aclname srcdomain .foo.com ... # reverse lookup, client IP
467 acl aclname dstdomain .foo.com ... # Destination server from URL
468 acl aclname srcdom_regex [-i] xxx ... # regex matching client name
469 acl aclname dstdom_regex [-i] xxx ... # regex matching server
470 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
471 # based URL is used and no match is found. The name "none" is used
472 # if the reverse lookup fails.
474 acl aclname http_status 200 301 500- 400-403 ... # status code in reply
476 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
485 h1:m1 must be less than h2:m2
486 acl aclname url_regex [-i] ^http:// ... # regex matching on whole URL
487 acl aclname urlpath_regex [-i] \.gif$ ... # regex matching on URL path
488 acl aclname port 80 70 21 ...
489 acl aclname port 0-1024 ... # ranges allowed
490 acl aclname myport 3128 ... # (local socket TCP port)
491 acl aclname proto HTTP FTP ...
492 acl aclname method GET POST ...
493 acl aclname browser [-i] regexp ...
494 # pattern match on User-Agent header (see also req_header below)
495 acl aclname referer_regex [-i] regexp ...
496 # pattern match on Referer header
497 # Referer is highly unreliable, so use with care
498 acl aclname ident username ...
499 acl aclname ident_regex [-i] pattern ...
500 # string match on ident output.
501 # use REQUIRED to accept any non-null ident.
502 acl aclname src_as number ...
503 acl aclname dst_as number ...
504 # Except for access control, AS numbers can be used for
505 # routing of requests to specific caches. Here's an
506 # example for routing all requests for AS#1241 and only
507 # those to mycache.mydomain.net:
508 # acl asexample dst_as 1241
509 # cache_peer_access mycache.mydomain.net allow asexample
510 # cache_peer_access mycache_mydomain.net deny all
512 acl aclname proxy_auth [-i] username ...
513 acl aclname proxy_auth_regex [-i] pattern ...
514 # list of valid usernames
515 # use REQUIRED to accept any valid username.
517 # NOTE: when a Proxy-Authentication header is sent but it is not
518 # needed during ACL checking the username is NOT logged
521 # NOTE: proxy_auth requires a EXTERNAL authentication program
522 # to check username/password combinations (see
523 # auth_param directive).
525 # NOTE: proxy_auth can't be used in a transparent proxy as
526 # the browser needs to be configured for using a proxy in order
527 # to respond to proxy authentication.
529 acl aclname snmp_community string ...
530 # A community string to limit access to your SNMP Agent
533 # acl snmppublic snmp_community public
535 acl aclname maxconn number
536 # This will be matched when the client's IP address has
537 # more than <number> HTTP connections established.
539 acl aclname max_user_ip [-s] number
540 # This will be matched when the user attempts to log in from more
541 # than <number> different ip addresses. The authenticate_ip_ttl
542 # parameter controls the timeout on the ip entries.
543 # If -s is specified the limit is strict, denying browsing
544 # from any further IP addresses until the ttl has expired. Without
545 # -s Squid will just annoy the user by "randomly" denying requests.
546 # (the counter is reset each time the limit is reached and a
548 # NOTE: in acceleration mode or where there is mesh of child proxies,
549 # clients may appear to come from multiple addresses if they are
550 # going through proxy farms, so a limit of 1 may cause user problems.
552 acl aclname req_mime_type mime-type1 ...
553 # regex match against the mime type of the request generated
554 # by the client. Can be used to detect file upload or some
555 # types HTTP tunneling requests.
556 # NOTE: This does NOT match the reply. You cannot use this
557 # to match the returned file type.
559 acl aclname req_header header-name [-i] any\.regex\.here
560 # regex match against any of the known request headers. May be
561 # thought of as a superset of "browser", "referer" and "mime-type"
564 acl aclname rep_mime_type mime-type1 ...
565 # regex match against the mime type of the reply received by
566 # squid. Can be used to detect file download or some
567 # types HTTP tunneling requests.
568 # NOTE: This has no effect in http_access rules. It only has
569 # effect in rules that affect the reply data stream such as
572 acl aclname rep_header header-name [-i] any\.regex\.here
573 # regex match against any of the known reply headers. May be
574 # thought of as a superset of "browser", "referer" and "mime-type"
577 acl acl_name external class_name [arguments...]
578 # external ACL lookup via a helper class defined by the
579 # external_acl_type directive.
581 acl aclname user_cert attribute values...
582 # match against attributes in a user SSL certificate
583 # attribute is one of DN/C/O/CN/L/ST
585 acl aclname ca_cert attribute values...
586 # match against attributes a users issuing CA SSL certificate
587 # attribute is one of DN/C/O/CN/L/ST
589 acl aclname ext_user username ...
590 acl aclname ext_user_regex [-i] pattern ...
591 # string match on username returned by external acl helper
592 # use REQUIRED to accept any non-null user name.
595 acl macaddress arp 09:00:2b:23:45:67
596 acl myexample dst_as 1241
597 acl password proxy_auth REQUIRED
598 acl fileupload req_mime_type -i ^multipart/form-data$
599 acl javascript rep_mime_type -i ^application/x-javascript$
602 #Recommended minimum configuration:
603 acl all src 0.0.0.0/0.0.0.0
604 acl manager proto cache_object
605 acl localhost src 127.0.0.1/255.255.255.255
606 acl to_localhost dst 127.0.0.0/8
607 acl SSL_ports port 443
608 acl Safe_ports port 80 # http
609 acl Safe_ports port 21 # ftp
610 acl Safe_ports port 443 # https
611 acl Safe_ports port 70 # gopher
612 acl Safe_ports port 210 # wais
613 acl Safe_ports port 1025-65535 # unregistered ports
614 acl Safe_ports port 280 # http-mgmt
615 acl Safe_ports port 488 # gss-http
616 acl Safe_ports port 591 # filemaker
617 acl Safe_ports port 777 # multiling http
618 acl CONNECT method CONNECT
624 LOC: Config.accessList.http
626 DEFAULT_IF_NONE: deny all
628 Allowing or Denying access based on defined access lists
630 Access to the HTTP port:
631 http_access allow|deny [!]aclname ...
633 NOTE on default values:
635 If there are no "access" lines present, the default is to deny
638 If none of the "access" lines cause a match, the default is the
639 opposite of the last line in the list. If the last line was
640 deny, the default is allow. Conversely, if the last line
641 is allow, the default will be deny. For these reasons, it is a
642 good idea to have an "deny all" or "allow all" entry at the end
643 of your access lists to avoid potential confusion.
646 #Recommended minimum configuration:
648 # Only allow cachemgr access from localhost
649 http_access allow manager localhost
650 http_access deny manager
651 # Deny requests to unknown ports
652 http_access deny !Safe_ports
653 # Deny CONNECT to other than SSL ports
654 http_access deny CONNECT !SSL_ports
656 # We strongly recommend the following be uncommented to protect innocent
657 # web applications running on the proxy server who think the only
658 # one who can access services on "localhost" is a local user
659 #http_access deny to_localhost
661 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
663 # Example rule allowing access from your local networks. Adapt
664 # to list your (internal) IP networks from where browsing should
666 #acl our_networks src 192.168.1.0/24 192.168.2.0/24
667 #http_access allow our_networks
669 # And finally deny all other access to this proxy
674 NAME: http_reply_access
676 LOC: Config.accessList.reply
679 Allow replies to client requests. This is complementary to http_access.
681 http_reply_access allow|deny [!] aclname ...
683 NOTE: if there are no access lines present, the default is to allow
686 If none of the access lines cause a match the opposite of the
687 last line will apply. Thus it is good practice to end the rules
688 with an "allow all" or "deny all" entry.
693 LOC: Config.accessList.icp
695 DEFAULT_IF_NONE: deny all
697 Allowing or Denying access to the ICP port based on defined
700 icp_access allow|deny [!]aclname ...
702 See http_access for details
705 #Allow ICP queries from everyone
713 LOC: Config.accessList.htcp
715 DEFAULT_IF_NONE: deny all
717 Allowing or Denying access to the HTCP port based on defined
720 htcp_access allow|deny [!]aclname ...
722 See http_access for details
724 #Allow HTCP queries from everyone
725 htcp_access allow all
728 NAME: htcp_clr_access
731 LOC: Config.accessList.htcp_clr
733 DEFAULT_IF_NONE: deny all
735 Allowing or Denying access to purge content using HTCP based
736 on defined access lists
738 htcp_clr_access allow|deny [!]aclname ...
740 See http_access for details
742 #Allow HTCP CLR requests from trusted peers
743 acl htcp_clr_peer src 172.16.1.2
744 htcp_clr_access allow htcp_clr_peer
749 LOC: Config.accessList.miss
752 Use to force your neighbors to use you as a sibling instead of
753 a parent. For example:
755 acl localclients src 172.16.0.0/16
756 miss_access allow localclients
757 miss_access deny !localclients
759 This means only your local clients are allowed to fetch
760 MISSES and all other clients can only fetch HITS.
762 By default, allow all clients who passed the http_access rules
763 to fetch MISSES from us.
767 # miss_access allow all
771 NAME: ident_lookup_access
775 DEFAULT_IF_NONE: deny all
776 LOC: Config.accessList.identLookup
778 A list of ACL elements which, if matched, cause an ident
779 (RFC 931) lookup to be performed for this request. For
780 example, you might choose to always perform ident lookups
781 for your main multi-user Unix boxes, but not for your Macs
782 and PCs. By default, ident lookups are not performed for
785 To enable ident lookups for specific client addresses, you
786 can follow this example:
788 acl ident_aware_hosts src 198.168.1.0/255.255.255.0
789 ident_lookup_access allow ident_aware_hosts
790 ident_lookup_access deny all
792 Only src type ACL checks are fully supported. A src_domain
793 ACL might work at times, but it will not always provide
797 NAME: reply_body_max_size
798 COMMENT: size [acl acl...]
801 LOC: Config.ReplyBodySize
803 This option specifies the maximum size of a reply body. It can be
804 used to prevent users from downloading very large files, such as
805 MP3's and movies. When the reply headers are received, the
806 reply_body_max_size lines are processed, and the first line where
807 all (if any) listed ACLs are true is used as the maximum body size
810 This size is checked twice. First when we get the reply headers,
811 we check the content-length value. If the content length value exists
812 and is larger than the allowed size, the request is denied and the
813 user receives an error message that says "the request or reply
814 is too large." If there is no content-length, and the reply
815 size exceeds this limit, the client's connection is just closed
816 and they will receive a partial reply.
818 WARNING: downstream caches probably can not detect a partial reply
819 if there is no content-length header, so they will cache
820 partial responses and give them out as hits. You should NOT
821 use this option if you have downstream caches.
823 WARNING: A maximum size smaller than the size of squid's error messages
824 will cause an infinite loop and crash squid. Ensure that the smallest
825 non-zero value you use is greater that the maximum header size plus
826 the size of your largest error page.
828 If you set this parameter none (the default), there will be
834 -----------------------------------------------------------------------------
837 NAME: http_port ascii_port
840 LOC: Config.Sockaddr.http
842 Usage: port [options]
843 hostname:port [options]
844 1.2.3.4:port [options]
846 The socket addresses where Squid will listen for HTTP client
847 requests. You may specify multiple socket addresses.
848 There are three forms: port alone, hostname with port, and
849 IP address with port. If you specify a hostname or IP
850 address, Squid binds the socket to that specific
851 address. This replaces the old 'tcp_incoming_address'
852 option. Most likely, you do not need to bind to a specific
853 address, so you can use the port number alone.
855 If you are running Squid in accelerator mode, you
856 probably want to listen on port 80 also, or instead.
858 The -a command line option may be used to specify additional
859 port(s) where Squid listens for proxy request. Such ports will
860 be plain proxy ports with no options.
862 You may specify multiple socket addresses on multiple lines.
866 transparent Support for transparent interception of
867 outgoing requests without browser settings.
869 tproxy Support Linux TPROXY for spoofing outgoing
870 connections using the client IP address.
872 accel Accelerator mode. Also needs at least one of
873 vhost / vport / defaultsite.
875 defaultsite=domainname
876 What to use for the Host: header if it is not present
877 in a request. Determines what site (not origin server)
878 accelerators should consider the default.
881 vhost Accelerator mode using Host header for virtual
882 domain support. Implies accel.
884 vport Accelerator with IP based virtual host support.
887 vport=NN As above, but uses specified port number rather
888 than the http_port number. Implies accel.
890 protocol= Protocol to reconstruct accelerated requests with.
893 disable-pmtu-discovery=
894 Control Path-MTU discovery usage:
895 off lets OS decide on what to do (default).
896 transparent disable PMTU discovery when transparent
898 always disable always PMTU discovery.
900 In many setups of transparently intercepting proxies
901 Path-MTU discovery can not work on traffic towards the
902 clients. This is the case when the intercepting device
903 does not fully track connections and fails to forward
904 ICMP must fragment messages to the cache server. If you
905 have such setup and experience that certain clients
906 sporadically hang or never complete requests set
907 disable-pmtu-discovery option to 'transparent'.
909 If you run Squid on a dual-homed machine with an internal
910 and an external interface we recommend you to specify the
911 internal address:port in http_port. This way Squid will only be
912 visible on the internal address.
915 # Squid normally listens to port 3128
916 http_port @DEFAULT_HTTP_PORT@
922 TYPE: https_port_list
924 LOC: Config.Sockaddr.https
926 Usage: [ip:]port cert=certificate.pem [key=key.pem] [options...]
928 The socket address where Squid will listen for HTTPS client
931 This is really only useful for situations where you are running
932 squid in accelerator mode and you want to do the SSL work at the
935 You may specify multiple socket addresses on multiple lines,
936 each with their own SSL certificate and/or options.
940 accel Accelerator mode. Also needs at least one of
941 defaultsite or vhost.
943 defaultsite= The name of the https site presented on
944 this port. Implies accel.
946 vhost Accelerator mode using Host header for virtual
947 domain support. Requires a wildcard certificate
948 or other certificate valid for more than one domain.
951 protocol= Protocol to reconstruct accelerated requests with.
954 cert= Path to SSL certificate (PEM format).
956 key= Path to SSL private key file (PEM format)
957 if not specified, the certificate file is
958 assumed to be a combined certificate and
961 version= The version of SSL/TLS supported
962 1 automatic (default)
967 cipher= Colon separated list of supported ciphers.
969 options= Various SSL engine options. The most important
971 NO_SSLv2 Disallow the use of SSLv2
972 NO_SSLv3 Disallow the use of SSLv3
973 NO_TLSv1 Disallow the use of TLSv1
974 SINGLE_DH_USE Always create a new key when using
975 temporary/ephemeral DH key exchanges
976 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
977 documentation for a complete list of options.
979 clientca= File containing the list of CAs to use when
980 requesting a client certificate.
982 cafile= File containing additional CA certificates to
983 use when verifying client certificates. If unset
984 clientca will be used.
986 capath= Directory containing additional CA certificates
987 and CRL lists to use when verifying client certificates.
989 crlfile= File of additional CRL lists to use when verifying
990 the client certificate, in addition to CRLs stored in
991 the capath. Implies VERIFY_CRL flag below.
993 dhparams= File containing DH parameters for temporary/ephemeral
996 sslflags= Various flags modifying the use of SSL:
998 Don't request client certificates
999 immediately, but wait until acl processing
1000 requires a certificate (not yet implemented).
1002 Don't use the default CA lists built in
1005 Don't allow for session reuse. Each connection
1006 will result in a new SSL session.
1008 Verify CRL lists when accepting client
1011 Verify CRL lists for all certificates in the
1012 client certificate chain.
1014 sslcontext= SSL session ID context identifier.
1016 vport Accelerator with IP based virtual host support.
1018 vport=NN As above, but uses specified port number rather
1019 than the https_port number. Implies accel.
1023 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1026 LOC: Config.accessList.outgoing_tos
1028 Allows you to select a TOS/Diffserv value to mark outgoing
1029 connections with, based on the username or source address
1032 tcp_outgoing_tos ds-field [!]aclname ...
1034 Example where normal_service_net uses the TOS value 0x00
1035 and normal_service_net uses 0x20
1037 acl normal_service_net src 10.0.0.0/255.255.255.0
1038 acl good_service_net src 10.0.1.0/255.255.255.0
1039 tcp_outgoing_tos 0x00 normal_service_net 0x00
1040 tcp_outgoing_tos 0x20 good_service_net
1042 TOS/DSCP values really only have local significance - so you should
1043 know what you're specifying. For more information, see RFC2474 and
1046 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1047 "default" to use whatever default your host has. Note that in
1048 practice often only values 0 - 63 is usable as the two highest bits
1049 have been redefined for use by ECN (RFC3168).
1051 Processing proceeds in the order specified, and stops at first fully
1054 Note: The use of this directive using client dependent ACLs is
1055 incompatible with the use of server side persistent connections. To
1056 ensure correct results it is best to set server_persisten_connections
1057 to off when using this directive in such configurations.
1060 NAME: clientside_tos
1063 LOC: Config.accessList.clientside_tos
1065 Allows you to select a TOS/Diffserv value to mark client-side
1066 connections with, based on the username or source address
1070 NAME: tcp_outgoing_address
1073 LOC: Config.accessList.outgoing_address
1075 Allows you to map requests to different outgoing IP addresses
1076 based on the username or source address of the user making
1079 tcp_outgoing_address ipaddr [[!]aclname] ...
1081 Example where requests from 10.0.0.0/24 will be forwarded
1082 with source address 10.1.0.1, 10.0.2.0/24 forwarded with
1083 source address 10.1.0.2 and the rest will be forwarded with
1084 source address 10.1.0.3.
1086 acl normal_service_net src 10.0.0.0/255.255.255.0
1087 acl good_service_net src 10.0.1.0/255.255.255.0
1088 tcp_outgoing_address 10.0.0.1 normal_service_net
1089 tcp_outgoing_address 10.0.0.2 good_service_net
1090 tcp_outgoing_address 10.0.0.3
1092 Processing proceeds in the order specified, and stops at first fully
1095 Note: The use of this directive using client dependent ACLs is
1096 incompatible with the use of server side persistent connections. To
1097 ensure correct results it is best to set server_persistent_connections
1098 to off when using this directive in such configurations.
1103 -----------------------------------------------------------------------------
1106 NAME: ssl_unclean_shutdown
1110 LOC: Config.SSL.unclean_shutdown
1112 Some browsers (especially MSIE) bugs out on SSL shutdown
1119 LOC: Config.SSL.ssl_engine
1122 The OpenSSL engine to use. You will need to set this if you
1123 would like to use hardware SSL acceleration for example.
1126 NAME: sslproxy_client_certificate
1129 LOC: Config.ssl_client.cert
1132 Client SSL Certificate to use when proxying https:// URLs
1135 NAME: sslproxy_client_key
1138 LOC: Config.ssl_client.key
1141 Client SSL Key to use when proxying https:// URLs
1144 NAME: sslproxy_version
1147 LOC: Config.ssl_client.version
1150 SSL version level to use when proxying https:// URLs
1153 NAME: sslproxy_options
1156 LOC: Config.ssl_client.options
1159 SSL engine options to use when proxying https:// URLs
1162 NAME: sslproxy_cipher
1165 LOC: Config.ssl_client.cipher
1168 SSL cipher list to use when proxying https:// URLs
1171 NAME: sslproxy_cafile
1174 LOC: Config.ssl_client.cafile
1177 file containing CA certificates to use when verifying server
1178 certificates while proxying https:// URLs
1181 NAME: sslproxy_capath
1184 LOC: Config.ssl_client.capath
1187 directory containing CA certificates to use when verifying
1188 server certificates while proxying https:// URLs
1191 NAME: sslproxy_flags
1194 LOC: Config.ssl_client.flags
1197 Various flags modifying the use of SSL while proxying https:// URLs:
1198 DONT_VERIFY_PEER Accept certificates even if they fail to
1200 NO_DEFAULT_CA Don't use the default CA list built in
1204 NAME: sslpassword_program
1207 LOC: Config.Program.ssl_password
1210 Specify a program used for entering SSL key passphrases
1211 when using encrypted SSL certificate keys. If not specified
1212 keys must either be unencrypted, or Squid started with the -N
1213 option to allow it to query interactively for the passphrase.
1217 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
1218 -----------------------------------------------------------------------------
1226 To specify other caches in a hierarchy, use the format:
1228 cache_peer hostname type http-port icp-port [options]
1233 # hostname type port port options
1234 # -------------------- -------- ----- ----- -----------
1235 cache_peer parent.foo.net parent 3128 3130 proxy-only default
1236 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
1237 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
1239 type: either 'parent', 'sibling', or 'multicast'.
1241 proxy-port: The port number where the cache listens for proxy
1244 icp-port: Used for querying neighbor caches about
1245 objects. To have a non-ICP neighbor
1246 specify '7' for the ICP port and make sure the
1247 neighbor machine has the UDP echo port
1248 enabled in its /etc/inetd.conf file.
1249 NOTE: Also requires icp_port option enabled to send/receive
1250 requests via this method.
1260 weighted-round-robin
1267 login=user:password | PASS | *:password
1278 sslcert=/path/to/ssl/certificate
1279 sslkey=/path/to/ssl/key
1283 front-end-https[=on|auto]
1285 use 'proxy-only' to specify objects fetched
1286 from this cache should not be saved locally.
1288 use 'weight=n' to affect the selection of a peer
1289 during any weighted peer-selection mechanisms.
1290 The weight must be an integer; default is 1,
1291 larger weights are favored more.
1292 This option does not affect parent selection if a peering
1293 protocol is not in use.
1295 use 'basetime=n' to specify a base amount to
1296 be subtracted from round trip times of parents.
1297 It is subtracted before division by weight in calculating
1298 which parent to fectch from. If the rtt is less than the
1299 base time the rtt is set to a minimal value.
1301 use 'ttl=n' to specify a IP multicast TTL to use
1302 when sending an ICP queries to this address.
1303 Only useful when sending to a multicast group.
1304 Because we don't accept ICP replies from random
1305 hosts, you must configure other group members as
1306 peers with the 'multicast-responder' option below.
1308 use 'no-query' to NOT send ICP queries to this
1311 use 'background-ping' to only send ICP queries to this
1312 neighbor infrequently. This is used to keep the neighbor
1313 round trip time updated and is usually used in
1314 conjunction with weighted-round-robin.
1316 use 'default' if this is a parent cache which can
1317 be used as a "last-resort" if a peer cannot be located
1318 by any of the peer-selection mechanisms.
1319 If specified more than once, only the first is used.
1321 use 'round-robin' to define a set of parents which
1322 should be used in a round-robin fashion in the
1323 absence of any ICP queries.
1325 use 'weighted-round-robin' to define a set of parents
1326 which should be used in a round-robin fashion with the
1327 frequency of each parent being based on the round trip
1328 time. Closer parents are used more often.
1329 Usually used for background-ping parents.
1331 use 'carp' to define a set of parents which should
1332 be used as a CARP array. The requests will be
1333 distributed among the parents based on the CARP load
1334 balancing hash function based on their weigth.
1336 'multicast-responder' indicates the named peer
1337 is a member of a multicast group. ICP queries will
1338 not be sent directly to the peer, but ICP replies
1339 will be accepted from it.
1341 'closest-only' indicates that, for ICP_OP_MISS
1342 replies, we'll only forward CLOSEST_PARENT_MISSes
1343 and never FIRST_PARENT_MISSes.
1345 use 'no-digest' to NOT request cache digests from
1348 'no-netdb-exchange' disables requesting ICMP
1349 RTT database (NetDB) from the neighbor.
1351 use 'no-delay' to prevent access to this neighbor
1352 from influencing the delay pools.
1354 use 'login=user:password' if this is a personal/workgroup
1355 proxy and your parent requires proxy authentication.
1356 Note: The string can include URL escapes (i.e. %20 for
1357 spaces). This also means % must be written as %%.
1359 use 'login=PASS' if users must authenticate against
1360 the upstream proxy or in the case of a reverse proxy
1361 configuration, the origin web server. This will pass
1362 the users credentials as they are to the peer.
1363 This only works for the Basic HTTP authentication scheme.
1364 Note: To combine this with proxy_auth both proxies must
1365 share the same user database as HTTP only allows for
1366 a single login (one for proxy, one for origin server).
1367 Also be warned this will expose your users proxy
1368 password to the peer. USE WITH CAUTION
1370 use 'login=*:password' to pass the username to the
1371 upstream cache, but with a fixed password. This is meant
1372 to be used when the peer is in another administrative
1373 domain, but it is still needed to identify each user.
1374 The star can optionally be followed by some extra
1375 information which is added to the username. This can
1376 be used to identify this proxy to the peer, similar to
1377 the login=username:password option above.
1379 use 'connect-timeout=nn' to specify a peer
1380 specific connect timeout (also see the
1381 peer_connect_timeout directive)
1383 use 'digest-url=url' to tell Squid to fetch the cache
1384 digest (if digests are enabled) for this host from
1385 the specified URL rather than the Squid default
1388 use 'allow-miss' to disable Squid's use of only-if-cached
1389 when forwarding requests to siblings. This is primarily
1390 useful when icp_hit_stale is used by the sibling. To
1391 extensive use of this option may result in forwarding
1392 loops, and you should avoid having two-way peerings
1393 with this option. (for example to deny peer usage on
1394 requests from peer by denying cache_peer_access if the
1397 use 'max-conn=n' to limit the amount of connections Squid
1398 may open to this peer.
1400 use 'htcp' to send HTCP, instead of ICP, queries
1401 to the neighbor. You probably also want to
1402 set the "icp port" to 4827 instead of 3130.
1404 use 'htcp-oldsquid' to send HTCP to old Squid versions
1406 'originserver' causes this parent peer to be contacted as
1407 a origin server. Meant to be used in accelerator setups.
1409 use 'name=xxx' if you have multiple peers on the same
1410 host but different ports. This name can be used to
1411 differentiate the peers in cache_peer_access and similar
1414 use 'forceddomain=name' to forcibly set the Host header
1415 of requests forwarded to this peer. Useful in accelerator
1416 setups where the server (peer) expects a certain domain
1417 name and using redirectors to feed this domain name
1420 use 'ssl' to indicate connections to this peer should
1421 be SSL/TLS encrypted.
1423 use 'sslcert=/path/to/ssl/certificate' to specify a client
1424 SSL certificate to use when connecting to this peer.
1426 use 'sslkey=/path/to/ssl/key' to specify the private SSL
1427 key corresponding to sslcert above. If 'sslkey' is not
1428 specified 'sslcert' is assumed to reference a
1429 combined file containing both the certificate and the key.
1431 use sslversion=1|2|3|4 to specify the SSL version to use
1432 when connecting to this peer
1433 1 = automatic (default)
1438 use sslcipher=... to specify the list of valid SSL ciphers
1439 to use when connecting to this peer.
1441 use ssloptions=... to specify various SSL engine options:
1442 NO_SSLv2 Disallow the use of SSLv2
1443 NO_SSLv3 Disallow the use of SSLv3
1444 NO_TLSv1 Disallow the use of TLSv1
1445 See src/ssl_support.c or the OpenSSL documentation for
1446 a more complete list.
1448 use sslcafile=... to specify a file containing
1449 additional CA certificates to use when verifying the
1452 use sslcapath=... to specify a directory containing
1453 additional CA certificates to use when verifying the
1456 use sslcrlfile=... to specify a certificate revocation
1457 list file to use when verifying the peer certificate.
1459 use sslflags=... to specify various flags modifying the
1462 Accept certificates even if they fail to
1465 Don't use the default CA list built in
1468 Don't verify the peer certificate
1469 matches the server name
1471 use ssldomain= to specify the peer name as advertised
1472 in it's certificate. Used for verifying the correctness
1473 of the received peer certificate. If not specified the
1474 peer hostname will be used.
1476 use front-end-https to enable the "Front-End-Https: On"
1477 header needed when using Squid as a SSL frontend in front
1478 of Microsoft OWA. See MS KB document Q307347 for details
1479 on this header. If set to auto the header will
1480 only be added if the request is forwarded as a https://
1484 NAME: cache_peer_domain cache_host_domain
1489 Use to limit the domains for which a neighbor cache will be
1492 cache_peer_domain cache-host domain [domain ...]
1493 cache_peer_domain cache-host !domain
1495 For example, specifying
1497 cache_peer_domain parent.foo.net .edu
1499 has the effect such that UDP query packets are sent to
1500 'bigserver' only when the requested object exists on a
1501 server in the .edu domain. Prefixing the domainname
1502 with '!' means the cache will be queried for objects
1505 NOTE: * Any number of domains may be given for a cache-host,
1506 either on the same or separate lines.
1507 * When multiple domains are given for a particular
1508 cache-host, the first matched domain is applied.
1509 * Cache hosts with no domain restrictions are queried
1511 * There are no defaults.
1512 * There is also a 'cache_peer_access' tag in the ACL
1516 NAME: cache_peer_access
1521 Similar to 'cache_peer_domain' but provides more flexibility by
1524 cache_peer_access cache-host allow|deny [!]aclname ...
1526 The syntax is identical to 'http_access' and the other lists of
1527 ACL elements. See the comments for 'http_access' below, or
1528 the Squid FAQ (http://www.squid-cache.org/FAQ/FAQ-10.html).
1531 NAME: neighbor_type_domain
1532 TYPE: hostdomaintype
1536 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
1538 Modifying the neighbor type for specific domains is now
1539 possible. You can treat some domains differently than the the
1540 default neighbor type specified on the 'cache_peer' line.
1541 Normally it should only be necessary to list domains which
1542 should be treated differently because the default neighbor type
1543 applies for hostnames which do not match domains listed here.
1546 cache_peer cache.foo.org parent 3128 3130
1547 neighbor_type_domain cache.foo.org sibling .com .net
1548 neighbor_type_domain cache.foo.org sibling .au .de
1551 NAME: dead_peer_timeout
1555 LOC: Config.Timeout.deadPeer
1557 This controls how long Squid waits to declare a peer cache
1558 as "dead." If there are no ICP replies received in this
1559 amount of time, Squid will declare the peer dead and not
1560 expect to receive any further ICP replies. However, it
1561 continues to send ICP queries, and will mark the peer as
1562 alive upon receipt of the first subsequent ICP reply.
1564 This timeout also affects when Squid expects to receive ICP
1565 replies from peers. If more than 'dead_peer' seconds have
1566 passed since the last ICP reply was received, Squid will not
1567 expect to receive an ICP reply on the next query. Thus, if
1568 your time between requests is greater than this timeout, you
1569 will see a lot of requests sent DIRECT to origin servers
1570 instead of to your parents.
1573 NAME: hierarchy_stoplist
1576 LOC: Config.hierarchy_stoplist
1578 A list of words which, if found in a URL, cause the object to
1579 be handled directly by this cache. In other words, use this
1580 to not query neighbor caches for certain objects. You may
1581 list this option multiple times.
1582 Note: never_direct overrides this option.
1584 #We recommend you to use at least the following line.
1585 hierarchy_stoplist cgi-bin ?
1590 MEMORY CACHE OPTIONS
1591 -----------------------------------------------------------------------------
1598 LOC: Config.memMaxSize
1600 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
1601 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
1602 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
1603 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
1605 'cache_mem' specifies the ideal amount of memory to be used
1607 * In-Transit objects
1609 * Negative-Cached objects
1611 Data for these objects are stored in 4 KB blocks. This
1612 parameter specifies the ideal upper limit on the total size of
1613 4 KB blocks allocated. In-Transit objects take the highest
1616 In-transit objects have priority over the others. When
1617 additional space is needed for incoming data, negative-cached
1618 and hot objects will be released. In other words, the
1619 negative-cached and hot objects will fill up any unused space
1620 not needed for in-transit objects.
1622 If circumstances require, this limit will be exceeded.
1623 Specifically, if your incoming request rate requires more than
1624 'cache_mem' of memory to hold in-transit objects, Squid will
1625 exceed this limit to satisfy the new requests. When the load
1626 decreases, blocks will be freed until the high-water mark is
1627 reached. Thereafter, blocks will be used to store hot
1631 NAME: maximum_object_size_in_memory
1635 LOC: Config.Store.maxInMemObjSize
1637 Objects greater than this size will not be attempted to kept in
1638 the memory cache. This should be set high enough to keep objects
1639 accessed frequently in memory to improve performance whilst low
1640 enough to keep larger objects from hoarding cache_mem.
1643 NAME: memory_replacement_policy
1645 LOC: Config.memPolicy
1648 The memory replacement policy parameter determines which
1649 objects are purged from memory when memory space is needed.
1651 See cache_replacement_policy for details.
1656 -----------------------------------------------------------------------------
1659 NAME: cache_replacement_policy
1661 LOC: Config.replPolicy
1664 The cache replacement policy parameter determines which
1665 objects are evicted (replaced) when disk space is needed.
1667 lru : Squid's original list based LRU policy
1668 heap GDSF : Greedy-Dual Size Frequency
1669 heap LFUDA: Least Frequently Used with Dynamic Aging
1670 heap LRU : LRU policy implemented using a heap
1672 Applies to any cache_dir lines listed below this.
1674 The LRU policies keeps recently referenced objects.
1676 The heap GDSF policy optimizes object hit rate by keeping smaller
1677 popular objects in cache so it has a better chance of getting a
1678 hit. It achieves a lower byte hit rate than LFUDA though since
1679 it evicts larger (possibly popular) objects.
1681 The heap LFUDA policy keeps popular objects in cache regardless of
1682 their size and thus optimizes byte hit rate at the expense of
1683 hit rate since one large, popular object will prevent many
1684 smaller, slightly less popular objects from being cached.
1686 Both policies utilize a dynamic aging mechanism that prevents
1687 cache pollution that can otherwise occur with frequency-based
1688 replacement policies.
1690 NOTE: if using the LFUDA replacement policy you should increase
1691 the value of maximum_object_size above its default of 4096 KB to
1692 to maximize the potential byte hit rate improvement of LFUDA.
1694 For more information about the GDSF and LFUDA cache replacement
1695 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
1696 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
1702 DEFAULT_IF_NONE: ufs @DEFAULT_SWAP_DIR@ 100 16 256
1703 LOC: Config.cacheSwap
1707 cache_dir Type Directory-Name Fs-specific-data [options]
1709 You can specify multiple cache_dir lines to spread the
1710 cache among different disk partitions.
1712 Type specifies the kind of storage system to use. Only "ufs"
1713 is built by default. To enable any of the other storage systems
1714 see the --enable-storeio configure option.
1716 'Directory' is a top-level directory where cache swap
1717 files will be stored. If you want to use an entire disk
1718 for caching, this can be the mount-point directory.
1719 The directory must exist and be writable by the Squid
1720 process. Squid will NOT create this directory for you.
1724 "ufs" is the old well-known Squid storage format that has always
1727 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
1729 'Mbytes' is the amount of disk space (MB) to use under this
1730 directory. The default is 100 MB. Change this to suit your
1731 configuration. Do NOT put the size of your disk drive here.
1732 Instead, if you want Squid to use the entire disk drive,
1733 subtract 20% and use that value.
1735 'Level-1' is the number of first-level subdirectories which
1736 will be created under the 'Directory'. The default is 16.
1738 'Level-2' is the number of second-level subdirectories which
1739 will be created under each first-level directory. The default
1742 The aufs store type:
1744 "aufs" uses the same storage format as "ufs", utilizing
1745 POSIX-threads to avoid blocking the main Squid process on
1746 disk-I/O. This was formerly known in Squid as async-io.
1748 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
1750 see argument descriptions under ufs above
1752 The diskd store type:
1754 "diskd" uses the same storage format as "ufs", utilizing a
1755 separate process to avoid blocking the main Squid process on
1758 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
1760 see argument descriptions under ufs above
1762 Q1 specifies the number of unacknowledged I/O requests when Squid
1763 stops opening new files. If this many messages are in the queues,
1764 Squid won't open new files. Default is 64
1766 Q2 specifies the number of unacknowledged messages when Squid
1767 starts blocking. If this many messages are in the queues,
1768 Squid blocks until it receives some replies. Default is 72
1770 When Q1 < Q2 (the default), the cache directory is optimized
1771 for lower response time at the expense of a decrease in hit
1772 ratio. If Q1 > Q2, the cache directory is optimized for
1773 higher hit ratio at the expense of an increase in response
1776 The coss store type:
1778 block-size=n defines the "block size" for COSS cache_dir's.
1779 Squid uses file numbers as block numbers. Since file numbers
1780 are limited to 24 bits, the block size determines the maximum
1781 size of the COSS partition. The default is 512 bytes, which
1782 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
1783 you should not change the coss block size after Squid
1784 has written some objects to the cache_dir.
1786 The coss file store has changed from 2.5. Now it uses a file
1787 called 'stripe' in the directory names in the config - and
1788 this will be created by squid -z.
1790 The null store type:
1792 no options are allowed or required
1796 no-store, no new objects should be stored to this cache_dir
1798 max-size=n, refers to the max object size this storedir supports.
1799 It is used to initially choose the storedir to dump the object.
1800 Note: To make optimal use of the max-size limits you should order
1801 the cache_dir lines with the smallest max-size value first and the
1802 ones with no max-size specification last.
1804 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
1805 which can be changed with the --with-coss-membuf-size=N configure
1809 NAME: store_dir_select_algorithm
1811 LOC: Config.store_dir_select_algorithm
1814 Set this to 'round-robin' as an alternative.
1817 NAME: max_open_disk_fds
1819 LOC: Config.max_open_disk_fds
1822 To avoid having disk as the I/O bottleneck Squid can optionally
1823 bypass the on-disk cache if more than this amount of disk file
1824 descriptors are open.
1826 A value of 0 indicates no limit.
1829 NAME: minimum_object_size
1833 LOC: Config.Store.minObjectSize
1835 Objects smaller than this size will NOT be saved on disk. The
1836 value is specified in kilobytes, and the default is 0 KB, which
1837 means there is no minimum.
1840 NAME: maximum_object_size
1844 LOC: Config.Store.maxObjectSize
1846 Objects larger than this size will NOT be saved on disk. The
1847 value is specified in kilobytes, and the default is 4MB. If
1848 you wish to get a high BYTES hit ratio, you should probably
1849 increase this (one 32 MB object hit counts for 3200 10KB
1850 hits). If you wish to increase speed more than your want to
1851 save bandwidth you should leave this low.
1853 NOTE: if using the LFUDA replacement policy you should increase
1854 this value to maximize the byte hit rate improvement of LFUDA!
1855 See replacement_policy below for a discussion of this policy.
1858 NAME: cache_swap_low
1859 COMMENT: (percent, 0-100)
1862 LOC: Config.Swap.lowWaterMark
1865 NAME: cache_swap_high
1866 COMMENT: (percent, 0-100)
1869 LOC: Config.Swap.highWaterMark
1872 The low- and high-water marks for cache object replacement.
1873 Replacement begins when the swap (disk) usage is above the
1874 low-water mark and attempts to maintain utilization near the
1875 low-water mark. As swap utilization gets close to high-water
1876 mark object eviction becomes more aggressive. If utilization is
1877 close to the low-water mark less replacement is done each time.
1879 Defaults are 90% and 95%. If you have a large cache, 5% could be
1880 hundreds of MB. If this is the case you may wish to set these
1881 numbers closer together.
1886 -----------------------------------------------------------------------------
1891 LOC: Config.Log.logformats
1896 logformat <name> <format specification>
1898 Defines an access log format.
1900 The <format specification> is a string with embedded % format codes
1902 % format codes all follow the same basic structure where all but
1903 the formatcode is optional. Output strings are automatically escaped
1904 as required according to their context and the output format
1905 modifiers are usually not needed, but can be specified if an explicit
1906 output format is desired.
1908 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
1910 " output in quoted string format
1911 [ output in squid text log format as used by log_mime_hdrs
1912 # output in URL quoted format
1916 width field width. If starting with 0 the
1917 output is zero padded
1918 {arg} argument such as header name etc
1922 >a Client source IP address
1924 >p Client source port
1925 <A Server IP address or peer name
1926 la Local IP address (http_port)
1927 lp Local port number (http_port)
1928 ts Seconds since epoch
1929 tu subsecond time (milliseconds)
1930 tl Local time. Optional strftime format argument
1931 default %d/%b/%Y:%H:%M:%S %z
1932 tg GMT time. Optional strftime format argument
1933 default %d/%b/%Y:%H:%M:%S %z
1934 tr Response time (milliseconds)
1935 >h Request header. Optional header name argument
1936 on the format header[:[separator]element]
1937 <h Reply header. Optional header name argument
1940 ul User name from authentication
1941 ui User name from ident
1942 us User name from SSL
1943 ue User name from external acl helper
1945 Ss Squid request status (TCP_MISS etc)
1946 Sh Squid hierarchy status (DEFAULT_PARENT etc)
1947 mt MIME content type
1948 rm Request method (GET/POST etc)
1950 rp Request URL-Path excluding hostname
1951 rv Request protocol version
1952 et Tag returned by external acl
1953 ea Log string returned by external acl
1954 <st Reply size including HTTP headers
1955 <sH Reply high offset sent
1956 <sS Upstream object size
1957 % a literal % character
1959 logformat squid %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt
1960 logformat squidmime %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt [%>h] [%<h]
1961 logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st %Ss:%Sh
1962 logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
1965 NAME: access_log cache_access_log
1967 LOC: Config.Log.accesslogs
1970 These files log client request activities. Has a line every HTTP or
1971 ICP request. The format is:
1972 access_log <filepath> [<logformat name> [acl acl ...]]
1973 access_log none [acl acl ...]]
1975 Will log to the specified file using the specified format (which
1976 must be defined in a logformat directive) those entries which match
1977 ALL the acl's specified (which must be defined in acl clauses).
1978 If no acl is specified, all requests will be logged to this file.
1980 To disable logging of a request use the filepath "none", in which case
1981 a logformat name should not be specified.
1983 To log the request via syslog specify a filepath of "syslog":
1985 access_log syslog[:facility.priority] [format [acl1 [acl2 ....]]]
1986 where facility could be any of:
1987 authpriv, daemon, local0 .. local7 or user.
1989 And priority could be any of:
1990 err, warning, notice, info, debug.
1992 access_log @DEFAULT_ACCESS_LOG@ squid
1998 LOC: Config.accessList.log
2000 COMMENT: allow|deny acl acl...
2002 This options allows you to control which requests gets logged
2003 to access.log (see access_log directive). Requests denied for
2004 logging will also not be accounted for in performance counters.
2009 DEFAULT: @DEFAULT_CACHE_LOG@
2012 Cache logging file. This is where general information about
2013 your cache's behavior goes. You can increase the amount of data
2014 logged to this file with the "debug_options" tag below.
2017 NAME: cache_store_log
2019 DEFAULT: @DEFAULT_STORE_LOG@
2020 LOC: Config.Log.store
2022 Logs the activities of the storage manager. Shows which
2023 objects are ejected from the cache, and which objects are
2024 saved and for how long. To disable, enter "none". There are
2025 not really utilities to analyze this data, so you can safely
2029 NAME: cache_swap_state cache_swap_log
2031 LOC: Config.Log.swap
2034 Location for the cache "swap.state" file. This index file holds
2035 the metadata of objects saved on disk. It is used to rebuild
2036 the cache during startup. Normally this file resides in each
2037 'cache_dir' directory, but you may specify an alternate
2038 pathname here. Note you must give a full filename, not just
2039 a directory. Since this is the index for the whole object
2040 list you CANNOT periodically rotate it!
2042 If %s can be used in the file name it will be replaced with a
2043 a representation of the cache_dir name where each / is replaced
2044 with '.'. This is needed to allow adding/removing cache_dir
2045 lines when cache_swap_log is being used.
2047 If have more than one 'cache_dir', and %s is not used in the name
2048 these swap logs will have names such as:
2054 The numbered extension (which is added automatically)
2055 corresponds to the order of the 'cache_dir' lines in this
2056 configuration file. If you change the order of the 'cache_dir'
2057 lines in this file, these index files will NOT correspond to
2058 the correct 'cache_dir' entry (unless you manually rename
2059 them). We recommend you do NOT use this option. It is
2060 better to keep these index files in each 'cache_dir' directory.
2063 NAME: logfile_rotate
2066 LOC: Config.Log.rotateNumber
2068 Specifies the number of logfile rotations to make when you
2069 type 'squid -k rotate'. The default is 10, which will rotate
2070 with extensions 0 through 9. Setting logfile_rotate to 0 will
2071 disable the file name rotation, but the logfiles are still closed
2072 and re-opened. This will enable you to rename the logfiles
2073 yourself just before sending the rotate signal.
2075 Note, the 'squid -k rotate' command normally sends a USR1
2076 signal to the running squid process. In certain situations
2077 (e.g. on Linux with Async I/O), USR1 is used for other
2078 purposes, so -k rotate uses another signal. It is best to get
2079 in the habit of using 'squid -k rotate' instead of 'kill -USR1
2083 NAME: emulate_httpd_log
2087 LOC: Config.onoff.common_log
2089 The Cache can emulate the log file format which many 'httpd'
2090 programs use. To disable/enable this emulation, set
2091 emulate_httpd_log to 'off' or 'on'. The default
2092 is to use the native log format since it includes useful
2093 information Squid-specific log analyzers use.
2096 NAME: log_ip_on_direct
2100 LOC: Config.onoff.log_ip_on_direct
2102 Log the destination IP address in the hierarchy log tag when going
2103 direct. Earlier Squid versions logged the hostname here. If you
2104 prefer the old way set this to off.
2109 DEFAULT: @DEFAULT_MIME_TABLE@
2110 LOC: Config.mimeTablePathname
2112 Pathname to Squid's MIME table. You shouldn't need to change
2113 this, but the default file contains examples and formatting
2114 information if you do.
2120 LOC: Config.onoff.log_mime_hdrs
2123 The Cache can record both the request and the response MIME
2124 headers for each HTTP transaction. The headers are encoded
2125 safely and will appear as two bracketed fields at the end of
2126 the access log (for either the native or httpd-emulated log
2127 formats). To enable this logging set log_mime_hdrs to 'on'.
2132 LOC: Config.Log.useragent
2134 IFDEF: USE_USERAGENT_LOG
2136 Squid will write the User-Agent field from HTTP requests
2137 to the filename specified here. By default useragent_log
2141 NAME: referer_log referrer_log
2143 LOC: Config.Log.referer
2145 IFDEF: USE_REFERER_LOG
2147 Squid will write the Referer field from HTTP requests to the
2148 filename specified here. By default referer_log is disabled.
2149 Note that "referer" is actually a misspelling of "referrer"
2150 however the misspelt version has been accepted into the HTTP RFCs
2156 DEFAULT: @DEFAULT_PID_FILE@
2157 LOC: Config.pidFilename
2159 A filename to write the process-id to. To disable, enter "none".
2165 LOC: Config.debugOptions
2167 Logging options are set as section,level where each source file
2168 is assigned a unique section. Lower levels result in less
2169 output, Full debugging (level 9) can result in a very large
2170 log file, so be careful. The magic word "ALL" sets debugging
2171 levels for all sections. We recommend normally running with
2179 LOC: Config.onoff.log_fqdn
2181 Turn this on if you wish to log fully qualified domain names
2182 in the access.log. To do this Squid does a DNS lookup of all
2183 IP's connecting to it. This can (in some situations) increase
2184 latency, which makes your cache seem slower for interactive
2188 NAME: client_netmask
2190 LOC: Config.Addrs.client_netmask
2191 DEFAULT: 255.255.255.255
2193 A netmask for client addresses in logfiles and cachemgr output.
2194 Change this to protect the privacy of your cache clients.
2195 A netmask of 255.255.255.0 will log all IP's in that range with
2196 the last digit set to '0'.
2203 LOC: Config.Log.forward
2205 Logs the server-side requests.
2207 This is currently work in progress.
2210 NAME: strip_query_terms
2212 LOC: Config.onoff.strip_query_terms
2215 By default, Squid strips query terms from requested URLs before
2216 logging. This protects your user's privacy.
2223 LOC: Config.onoff.buffered_logs
2225 cache.log log file is written with stdio functions, and as such
2226 it can be buffered or unbuffered. By default it will be unbuffered.
2227 Buffering it can speed up the writing slightly (though you are
2228 unlikely to need to worry unless you run with tons of debugging
2229 enabled in which case performance will suffer badly anyway..).
2233 OPTIONS FOR FTP GATEWAYING
2234 -----------------------------------------------------------------------------
2240 LOC: Config.Ftp.anon_user
2242 If you want the anonymous login password to be more informative
2243 (and enable the use of picky ftp servers), set this to something
2244 reasonable for your domain, like wwwuser@somewhere.net
2246 The reason why this is domainless by default is the
2247 request can be made on the behalf of a user in any domain,
2248 depending on how the cache is used.
2249 Some ftp server also validate the email address is valid
2250 (for example perl.com).
2253 NAME: ftp_list_width
2256 LOC: Config.Ftp.list_width
2258 Sets the width of ftp listings. This should be set to fit in
2259 the width of a standard browser. Setting this too small
2260 can cut off long filenames when browsing ftp sites.
2266 LOC: Config.Ftp.passive
2268 If your firewall does not allow Squid to use passive
2269 connections, turn off this option.
2272 NAME: ftp_sanitycheck
2275 LOC: Config.Ftp.sanitycheck
2277 For security and data integrity reasons Squid by default performs
2278 sanity checks of the addresses of FTP data connections ensure the
2279 data connection is to the requested server. If you need to allow
2280 FTP connections to servers using another IP address for the data
2281 connection turn this off.
2284 NAME: ftp_telnet_protocol
2287 LOC: Config.Ftp.telnet
2289 The FTP protocol is officially defined to use the telnet protocol
2290 as transport channel for the control connection. However, many
2291 implementations are broken and does not respect this aspect of
2294 If you have trouble accessing files with ASCII code 255 in the
2295 path or similar problems involving this ASCII code you can
2296 try setting this directive to off. If that helps, report to the
2297 operator of the FTP server in question that their FTP server
2298 is broken and does not follow the FTP standard.
2302 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
2303 -----------------------------------------------------------------------------
2308 DEFAULT: @DEFAULT_DISKD@
2309 LOC: Config.Program.diskd
2311 Specify the location of the diskd executable.
2312 Note this is only useful if you have compiled in
2313 diskd as one of the store io modules.
2316 NAME: unlinkd_program
2319 DEFAULT: @DEFAULT_UNLINKD@
2320 LOC: Config.Program.unlinkd
2322 Specify the location of the executable for file deletion process.
2325 NAME: pinger_program
2327 DEFAULT: @DEFAULT_PINGER@
2328 LOC: Config.Program.pinger
2331 Specify the location of the executable for the pinger process.
2335 OPTIONS FOR URL REWRITING
2336 -----------------------------------------------------------------------------
2339 NAME: url_rewrite_program redirect_program
2341 LOC: Config.Program.redirect
2344 Specify the location of the executable for the URL rewriter.
2345 Since they can perform almost any function there isn't one included.
2347 For each requested URL rewriter will receive on line with the format
2349 URL <SP> client_ip "/" fqdn <SP> user <SP> method <NL>
2351 And the rewriter may return a rewritten URL. The other components of
2352 the request line does not need to be returned (ignored if they are).
2354 The rewriter can also indicate that a client-side redirect should
2355 be performed to the new URL. This is done by prefixing the returned
2356 URL with "301:" (moved permanently) or 302: (moved temporarily).
2358 By default, a URL rewriter is not used.
2361 NAME: url_rewrite_children redirect_children
2364 LOC: Config.redirectChildren
2366 The number of redirector processes to spawn. If you start
2367 too few Squid will have to wait for them to process a backlog of
2368 URLs, slowing it down. If you start too many they will use RAM
2369 and other system resources.
2372 NAME: url_rewrite_concurrency redirect_concurrency
2375 LOC: Config.redirectConcurrency
2377 The number of requests each redirector helper can handle in
2378 parallel. Defaults to 0 which indicates the redirector
2379 is a old-style single threaded redirector.
2382 NAME: url_rewrite_host_header redirect_rewrites_host_header
2385 LOC: Config.onoff.redir_rewrites_host
2387 By default Squid rewrites any Host: header in redirected
2388 requests. If you are running an accelerator this may
2389 not be a wanted effect of a redirector.
2391 WARNING: Entries are cached on the result of the URL rewriting
2392 process, so be careful if you have domain-virtual hosts.
2395 NAME: url_rewrite_access redirector_access
2398 LOC: Config.accessList.redirector
2400 If defined, this access list specifies which requests are
2401 sent to the redirector processes. By default all requests
2405 NAME: url_rewrite_bypass redirector_bypass
2407 LOC: Config.onoff.redirector_bypass
2410 When this is 'on', a request will not go through the
2411 redirector if all redirectors are busy. If this is 'off'
2412 and the redirector queue grows too large, Squid will exit
2413 with a FATAL error and ask you to increase the number of
2414 redirectors. You should only enable this if the redirectors
2415 are not critical to your caching system. If you use
2416 redirectors for access control, and you enable this option,
2417 users may have access to pages they should not
2418 be allowed to request.
2422 OPTIONS FOR TUNING THE CACHE
2423 -----------------------------------------------------------------------------
2426 NAME: cache no_cache
2429 LOC: Config.accessList.noCache
2431 A list of ACL elements which, if matched, cause the request to
2432 not be satisfied from the cache and the reply to not be cached.
2433 In other words, use this to force certain objects to never be cached.
2435 You must use the word 'DENY' to indicate the ACL names which should
2438 Default is to allow all to be cached
2440 #We recommend you to use the following two lines.
2441 acl QUERY urlpath_regex cgi-bin \?
2446 NAME: refresh_pattern
2447 TYPE: refreshpattern
2451 usage: refresh_pattern [-i] regex min percent max [options]
2453 By default, regular expressions are CASE-SENSITIVE. To make
2454 them case-insensitive, use the -i option.
2456 'Min' is the time (in minutes) an object without an explicit
2457 expiry time should be considered fresh. The recommended
2458 value is 0, any higher values may cause dynamic applications
2459 to be erroneously cached unless the application designer
2460 has taken the appropriate actions.
2462 'Percent' is a percentage of the objects age (time since last
2463 modification age) an object without explicit expiry time
2464 will be considered fresh.
2466 'Max' is an upper limit on how long objects without an explicit
2467 expiry time will be considered fresh.
2469 options: override-expire
2479 override-expire enforces min age even if the server
2480 sent a Expires: header. Doing this VIOLATES the HTTP
2481 standard. Enabling this feature could make you liable
2482 for problems which it causes.
2484 override-lastmod enforces min age even on objects
2485 that were modified recently.
2487 reload-into-ims changes client no-cache or ``reload''
2488 to If-Modified-Since requests. Doing this VIOLATES the
2489 HTTP standard. Enabling this feature could make you
2490 liable for problems which it causes.
2492 ignore-reload ignores a client no-cache or ``reload''
2493 header. Doing this VIOLATES the HTTP standard. Enabling
2494 this feature could make you liable for problems which
2497 ignore-no-cache ignores any ``Pragma: no-cache'' and
2498 ``Cache-control: no-cache'' headers received from a server.
2499 The HTTP RFC never allows the use of this (Pragma) header
2500 from a server, only a client, though plenty of servers
2503 ignore-no-store ignores any ``Cache-control: no-store''
2504 headers received from a server. Doing this VIOLATES
2505 the HTTP standard. Enabling this feature could make you
2506 liable for problems which it causes.
2508 ignore-private ignores any ``Cache-control: private''
2509 headers received from a server. Doing this VIOLATES
2510 the HTTP standard. Enabling this feature could make you
2511 liable for problems which it causes.
2513 ignore-auth caches responses to requests with authorization,
2514 as if the originserver had sent ``Cache-control: public''
2515 in the response header. Doing this VIOLATES the HTTP standard.
2516 Enabling this feature could make you liable for problems which
2519 refresh-ims causes squid to contact the origin server
2520 when a client issues an If-Modified-Since request. This
2521 ensures that the client will receive an updated version
2522 if one is available.
2524 Basically a cached object is:
2526 FRESH if expires < now, else STALE
2528 FRESH if lm-factor < percent, else STALE
2532 The refresh_pattern lines are checked in the order listed here.
2533 The first entry which matches is used. If none of the entries
2534 match the default will be used.
2536 Note, you must uncomment all the default lines if you want
2537 to change one. The default setting is only active if none is
2542 refresh_pattern ^ftp: 1440 20% 10080
2543 refresh_pattern ^gopher: 1440 0% 1440
2544 refresh_pattern . 0 20% 4320
2548 NAME: quick_abort_min
2552 LOC: Config.quickAbort.min
2555 NAME: quick_abort_max
2559 LOC: Config.quickAbort.max
2562 NAME: quick_abort_pct
2566 LOC: Config.quickAbort.pct
2568 The cache by default continues downloading aborted requests
2569 which are almost completed (less than 16 KB remaining). This
2570 may be undesirable on slow (e.g. SLIP) links and/or very busy
2571 caches. Impatient users may tie up file descriptors and
2572 bandwidth by repeatedly requesting and immediately aborting
2575 When the user aborts a request, Squid will check the
2576 quick_abort values to the amount of data transfered until
2579 If the transfer has less than 'quick_abort_min' KB remaining,
2580 it will finish the retrieval.
2582 If the transfer has more than 'quick_abort_max' KB remaining,
2583 it will abort the retrieval.
2585 If more than 'quick_abort_pct' of the transfer has completed,
2586 it will finish the retrieval.
2588 If you do not want any retrieval to continue after the client
2589 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
2592 If you want retrievals to always continue if they are being
2593 cached set 'quick_abort_min' to '-1 KB'.
2596 NAME: read_ahead_gap
2597 COMMENT: buffer-size
2599 LOC: Config.readAheadGap
2602 The amount of data the cache will buffer ahead of what has been
2603 sent to the client when retrieving an object from another server.
2609 LOC: Config.negativeTtl
2612 Time-to-Live (TTL) for failed requests. Certain types of
2613 failures (such as "connection refused" and "404 Not Found") are
2614 negatively-cached for a configurable amount of time. The
2615 default is 5 minutes. Note that this is different from
2616 negative caching of DNS lookups.
2619 NAME: positive_dns_ttl
2622 LOC: Config.positiveDnsTtl
2625 Upper limit on how long Squid will cache positive DNS responses.
2626 Default is 6 hours (360 minutes). This directive must be set
2627 larger than negative_dns_ttl.
2630 NAME: negative_dns_ttl
2633 LOC: Config.negativeDnsTtl
2636 Time-to-Live (TTL) for negative caching of failed DNS lookups.
2637 This also sets the lower cache limit on positive lookups.
2638 Minimum value is 1 second, and it is not recommendable to go
2639 much below 10 seconds.
2642 NAME: range_offset_limit
2645 LOC: Config.rangeOffsetLimit
2648 Sets a upper limit on how far into the the file a Range request
2649 may be to cause Squid to prefetch the whole file. If beyond this
2650 limit Squid forwards the Range request as it is and the result
2653 This is to stop a far ahead range request (lets say start at 17MB)
2654 from making Squid fetch the whole object up to that point before
2655 sending anything to the client.
2657 A value of -1 causes Squid to always fetch the object from the
2658 beginning so it may cache the result. (2.0 style)
2660 A value of 0 causes Squid to never fetch more than the
2661 client requested. (default)
2664 NAME: minimum_expiry_time
2667 LOC: Config.minimum_expiry_time
2670 The minimum caching time according to (Expires - Date)
2671 Headers Squid honors if the object can't be revalidated
2672 defaults to 60 seconds. In reverse proxy enorinments it
2673 might be desirable to honor shorter object lifetimes. It
2674 is most likely better to make your server return a
2675 meaningful Last-Modified header however. In ESI environments
2676 where page fragments often have short lifetimes, this will
2677 often be best set to 0.
2680 NAME: store_avg_object_size
2684 LOC: Config.Store.avgObjectSize
2686 Average object size, used to estimate number of objects your
2687 cache can hold. The default is 13 KB.
2690 NAME: store_objects_per_bucket
2693 LOC: Config.Store.objectsPerBucket
2695 Target number of objects per bucket in the store hash table.
2696 Lowering this value increases the total number of buckets and
2697 also the storage maintenance rate. The default is 20.
2702 -----------------------------------------------------------------------------
2705 NAME: request_header_max_size
2709 LOC: Config.maxRequestHeaderSize
2711 This specifies the maximum size for HTTP headers in a request.
2712 Request headers are usually relatively small (about 512 bytes).
2713 Placing a limit on the request header size will catch certain
2714 bugs (for example with persistent connections) and possibly
2715 buffer-overflow or denial-of-service attacks.
2718 NAME: reply_header_max_size
2722 LOC: Config.maxReplyHeaderSize
2724 This specifies the maximum size for HTTP headers in a reply.
2725 Reply headers are usually relatively small (about 512 bytes).
2726 Placing a limit on the reply header size will catch certain
2727 bugs (for example with persistent connections) and possibly
2728 buffer-overflow or denial-of-service attacks.
2731 NAME: request_body_max_size
2735 LOC: Config.maxRequestBodySize
2737 This specifies the maximum size for an HTTP request body.
2738 In other words, the maximum size of a PUT/POST request.
2739 A user who attempts to send a request with a body larger
2740 than this limit receives an "Invalid Request" error message.
2741 If you set this parameter to a zero (the default), there will
2742 be no limit imposed.
2748 LOC: Config.accessList.brokenPosts
2750 A list of ACL elements which, if matched, causes Squid to send
2751 an extra CRLF pair after the body of a PUT/POST request.
2753 Some HTTP servers has broken implementations of PUT/POST,
2754 and rely on an extra CRLF pair sent by some WWW clients.
2756 Quote from RFC2616 section 4.1 on this matter:
2758 Note: certain buggy HTTP/1.0 client implementations generate an
2759 extra CRLF's after a POST request. To restate what is explicitly
2760 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
2761 a request with an extra CRLF.
2764 acl buggy_server url_regex ^http://....
2765 broken_posts allow buggy_server
2769 IFDEF: HTTP_VIOLATIONS
2773 LOC: Config.onoff.via
2775 If set (default), Squid will include a Via header in requests and
2776 replies as required by RFC2616.
2782 LOC: Config.onoff.ie_refresh
2785 Microsoft Internet Explorer up until version 5.5 Service
2786 Pack 1 has an issue with transparent proxies, wherein it
2787 is impossible to force a refresh. Turning this on provides
2788 a partial fix to the problem, by causing all IMS-REFRESH
2789 requests from older IE versions to check the origin server
2790 for fresh content. This reduces hit ratio by some amount
2791 (~10% in my experience), but allows users to actually get
2792 fresh content when they want it. Note because Squid
2793 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
2794 of 5.5 is unchanged from old versions of Squid (i.e. a
2795 forced refresh is impossible). Newer versions of IE will,
2796 hopefully, continue to have the new behavior and will be
2797 handled based on that assumption. This option defaults to
2798 the old Squid behavior, which is better for hit ratios but
2799 worse for clients using IE, if they need to be able to
2800 force fresh content.
2803 NAME: vary_ignore_expire
2806 LOC: Config.onoff.vary_ignore_expire
2809 Many HTTP servers supporting Vary gives such objects
2810 immediate expiry time with no cache-control header
2811 when requested by a HTTP/1.0 client. This option
2812 enables Squid to ignore such expiry times until
2813 HTTP/1.1 is fully implemented.
2814 WARNING: This may eventually cause some varying
2815 objects not intended for caching to get cached.
2818 NAME: extension_methods
2820 LOC: Config.ext_methods
2823 Squid only knows about standardized HTTP request methods.
2824 You can add up to 20 additional "extension" methods here.
2827 NAME: request_entities
2829 LOC: Config.onoff.request_entities
2832 Squid defaults to deny GET and HEAD requests with request entities,
2833 as the meaning of such requests are undefined in the HTTP standard
2834 even if not explicitly forbidden.
2836 Set this directive to on if you have clients which insists
2837 on sending request entities in GET or HEAD requests. But be warned
2838 that there is server software (both proxies and web servers) which
2839 can fail to properly process this kind of request which may make you
2840 vulnerable to cache pollution attacks if enabled.
2843 NAME: request_header_access
2844 IFDEF: HTTP_VIOLATIONS
2845 TYPE: http_header_access[]
2846 LOC: Config.request_header_access
2849 Usage: request_header_access header_name allow|deny [!]aclname ...
2851 WARNING: Doing this VIOLATES the HTTP standard. Enabling
2852 this feature could make you liable for problems which it
2855 This option replaces the old 'anonymize_headers' and the
2856 older 'http_anonymizer' option with something that is much
2857 more configurable. This new method creates a list of ACLs
2858 for each header, allowing you very fine-tuned header
2861 This option only applies to request headers, i.e., from the
2862 client to the server.
2864 You can only specify known headers for the header name.
2865 Other headers are reclassified as 'Other'. You can also
2866 refer to all the headers with 'All'.
2868 For example, to achieve the same behavior as the old
2869 'http_anonymizer standard' option, you should use:
2871 request_header_access From deny all
2872 request_header_access Referer deny all
2873 request_header_access Server deny all
2874 request_header_access User-Agent deny all
2875 request_header_access WWW-Authenticate deny all
2876 request_header_access Link deny all
2878 Or, to reproduce the old 'http_anonymizer paranoid' feature
2881 request_header_access Allow allow all
2882 request_header_access Authorization allow all
2883 request_header_access WWW-Authenticate allow all
2884 request_header_access Proxy-Authorization allow all
2885 request_header_access Proxy-Authenticate allow all
2886 request_header_access Cache-Control allow all
2887 request_header_access Content-Encoding allow all
2888 request_header_access Content-Length allow all
2889 request_header_access Content-Type allow all
2890 request_header_access Date allow all
2891 request_header_access Expires allow all
2892 request_header_access Host allow all
2893 request_header_access If-Modified-Since allow all
2894 request_header_access Last-Modified allow all
2895 request_header_access Location allow all
2896 request_header_access Pragma allow all
2897 request_header_access Accept allow all
2898 request_header_access Accept-Charset allow all
2899 request_header_access Accept-Encoding allow all
2900 request_header_access Accept-Language allow all
2901 request_header_access Content-Language allow all
2902 request_header_access Mime-Version allow all
2903 request_header_access Retry-After allow all
2904 request_header_access Title allow all
2905 request_header_access Connection allow all
2906 request_header_access Proxy-Connection allow all
2907 request_header_access All deny all
2909 although many of those are HTTP reply headers, and so should be
2910 controlled with the reply_header_access directive.
2912 By default, all headers are allowed (no anonymizing is
2916 NAME: reply_header_access
2917 IFDEF: HTTP_VIOLATIONS
2918 TYPE: http_header_access[]
2919 LOC: Config.reply_header_access
2922 Usage: reply_header_access header_name allow|deny [!]aclname ...
2924 WARNING: Doing this VIOLATES the HTTP standard. Enabling
2925 this feature could make you liable for problems which it
2928 This option only applies to reply headers, i.e., from the
2929 server to the client.
2931 This is the same as request_header_access, but in the other
2934 This option replaces the old 'anonymize_headers' and the
2935 older 'http_anonymizer' option with something that is much
2936 more configurable. This new method creates a list of ACLs
2937 for each header, allowing you very fine-tuned header
2940 You can only specify known headers for the header name.
2941 Other headers are reclassified as 'Other'. You can also
2942 refer to all the headers with 'All'.
2944 For example, to achieve the same behavior as the old
2945 'http_anonymizer standard' option, you should use:
2947 reply_header_access From deny all
2948 reply_header_access Referer deny all
2949 reply_header_access Server deny all
2950 reply_header_access User-Agent deny all
2951 reply_header_access WWW-Authenticate deny all
2952 reply_header_access Link deny all
2954 Or, to reproduce the old 'http_anonymizer paranoid' feature
2957 reply_header_access Allow allow all
2958 reply_header_access Authorization allow all
2959 reply_header_access WWW-Authenticate allow all
2960 reply_header_access Proxy-Authorization allow all
2961 reply_header_access Proxy-Authenticate allow all
2962 reply_header_access Cache-Control allow all
2963 reply_header_access Content-Encoding allow all
2964 reply_header_access Content-Length allow all
2965 reply_header_access Content-Type allow all
2966 reply_header_access Date allow all
2967 reply_header_access Expires allow all
2968 reply_header_access Host allow all
2969 reply_header_access If-Modified-Since allow all
2970 reply_header_access Last-Modified allow all
2971 reply_header_access Location allow all
2972 reply_header_access Pragma allow all
2973 reply_header_access Accept allow all
2974 reply_header_access Accept-Charset allow all
2975 reply_header_access Accept-Encoding allow all
2976 reply_header_access Accept-Language allow all
2977 reply_header_access Content-Language allow all
2978 reply_header_access Mime-Version allow all
2979 reply_header_access Retry-After allow all
2980 reply_header_access Title allow all
2981 reply_header_access Connection allow all
2982 reply_header_access Proxy-Connection allow all
2983 reply_header_access All deny all
2985 although the HTTP request headers won't be usefully controlled
2986 by this directive -- see request_header_access for details.
2988 By default, all headers are allowed (no anonymizing is
2992 NAME: header_replace
2993 IFDEF: HTTP_VIOLATIONS
2994 TYPE: http_header_replace[]
2995 LOC: Config.request_header_access
2998 Usage: header_replace header_name message
2999 Example: header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
3001 This option allows you to change the contents of headers
3002 denied with header_access above, by replacing them with
3003 some fixed string. This replaces the old fake_user_agent
3006 This only applies to request headers, not reply headers.
3008 By default, headers are removed if denied.
3011 NAME: relaxed_header_parser
3012 COMMENT: on|off|warn
3014 LOC: Config.onoff.relaxed_header_parser
3017 In the default "on" setting Squid accepts certain forms
3018 of non-compliant HTTP messages where it is unambiguous
3019 what the sending application intended even if the message
3020 is not correctly formatted. The messages is then normalized
3021 to the correct form when forwarded by Squid.
3023 If set to "warn" then a warning will be emitted in cache.log
3024 each time such HTTP error is encountered.
3026 If set to "off" then such HTTP errors will cause the request
3027 or response to be rejected.
3032 -----------------------------------------------------------------------------
3035 NAME: forward_timeout
3038 LOC: Config.Timeout.forward
3041 This parameter specifies how long Squid should at most attempt in
3042 finding a forwarding path for the request before giving up.
3045 NAME: connect_timeout
3048 LOC: Config.Timeout.connect
3051 This parameter specifies how long to wait for the TCP connect to
3052 the requested server or peer to complete before Squid should
3053 attempt to find another path where to forward the request.
3056 NAME: peer_connect_timeout
3059 LOC: Config.Timeout.peer_connect
3062 This parameter specifies how long to wait for a pending TCP
3063 connection to a peer cache. The default is 30 seconds. You
3064 may also set different timeout values for individual neighbors
3065 with the 'connect-timeout' option on a 'cache_peer' line.
3071 LOC: Config.Timeout.read
3074 The read_timeout is applied on server-side connections. After
3075 each successful read(), the timeout will be extended by this
3076 amount. If no data is read again after this amount of time,
3077 the request is aborted and logged with ERR_READ_TIMEOUT. The
3078 default is 15 minutes.
3081 NAME: request_timeout
3083 LOC: Config.Timeout.request
3086 How long to wait for an HTTP request after initial
3087 connection establishment.
3090 NAME: persistent_request_timeout
3092 LOC: Config.Timeout.persistent_request
3095 How long to wait for the next HTTP request on a persistent
3096 connection after the previous request completes.
3099 NAME: client_lifetime
3102 LOC: Config.Timeout.lifetime
3105 The maximum amount of time a client (browser) is allowed to
3106 remain connected to the cache process. This protects the Cache
3107 from having a lot of sockets (and hence file descriptors) tied up
3108 in a CLOSE_WAIT state from remote clients that go away without
3109 properly shutting down (either because of a network failure or
3110 because of a poor client implementation). The default is one
3113 NOTE: The default value is intended to be much larger than any
3114 client would ever need to be connected to your cache. You
3115 should probably change client_lifetime only as a last resort.
3116 If you seem to have many client connections tying up
3117 filedescriptors, we recommend first tuning the read_timeout,
3118 request_timeout, persistent_request_timeout and quick_abort values.
3121 NAME: half_closed_clients
3123 LOC: Config.onoff.half_closed_clients
3126 Some clients may shutdown the sending side of their TCP
3127 connections, while leaving their receiving sides open. Sometimes,
3128 Squid can not tell the difference between a half-closed and a
3129 fully-closed TCP connection. By default, half-closed client
3130 connections are kept open until a read(2) or write(2) on the
3131 socket returns an error. Change this option to 'off' and Squid
3132 will immediately close client connections when read(2) returns
3133 "no more data to read."
3138 LOC: Config.Timeout.pconn
3141 Timeout for idle persistent connections to servers and other
3148 LOC: Config.Timeout.ident
3151 Maximum time to wait for IDENT lookups to complete.
3153 If this is too high, and you enabled IDENT lookups from untrusted
3154 users, you might be susceptible to denial-of-service by having
3155 many ident requests going at once.
3158 NAME: shutdown_lifetime
3161 LOC: Config.shutdownLifetime
3164 When SIGTERM or SIGHUP is received, the cache is put into
3165 "shutdown pending" mode until all active sockets are closed.
3166 This value is the lifetime to set for all open descriptors
3167 during shutdown mode. Any active clients after this many
3168 seconds will receive a 'timeout' message.
3172 ADMINISTRATIVE PARAMETERS
3173 -----------------------------------------------------------------------------
3179 LOC: Config.adminEmail
3181 Email-address of local cache manager who will receive
3182 mail if the cache dies. The default is "webmaster."
3188 LOC: Config.EmailFrom
3190 From: email-address for mail sent when the cache dies.
3191 The default is to use 'appname@unique_hostname'.
3192 Default appname value is "squid", can be changed into
3193 src/globals.h before building squid.
3199 LOC: Config.EmailProgram
3201 Email program used to send mail if the cache dies.
3202 The default is "mail". The specified program must complain
3203 with the standard Unix mail syntax:
3204 mail_program recipient < mailfile
3205 Optional command line options can be specified.
3208 NAME: cache_effective_user
3210 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
3211 LOC: Config.effectiveUser
3213 If you start Squid as root, it will change its effective/real
3214 UID/GID to the user specified below. The default is to change
3215 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
3216 see also; cache_effective_group
3219 NAME: cache_effective_group
3222 LOC: Config.effectiveGroup
3224 Squid sets the GID to the effective user's default group ID
3225 (taken from the password file) and supplementary group list
3226 from the groups membership.
3228 If you want Squid to run with a specific GID regardless of
3229 the group memberships of the effective user then set this
3230 to the group (or GID) you want Squid to run as. When set
3231 all other group privileges of the effective user are ignored
3232 and only this GID is effective. If Squid is not started as
3233 root the user starting Squid MUST be member of the specified
3236 This option is not recommended by the Squid Team.
3237 Our preference is for administrators to configure a secure
3238 user account for squid with UID/GID matching system policies.
3241 NAME: httpd_suppress_version_string
3245 LOC: Config.onoff.httpd_suppress_version_string
3247 Suppress Squid version string info in HTTP headers and HTML error pages.
3250 NAME: visible_hostname
3252 LOC: Config.visibleHostname
3255 If you want to present a special hostname in error messages, etc,
3256 define this. Otherwise, the return value of gethostname()
3257 will be used. If you have multiple caches in a cluster and
3258 get errors about IP-forwarding you must set them to have individual
3259 names with this setting.
3262 NAME: unique_hostname
3264 LOC: Config.uniqueHostname
3267 If you want to have multiple machines with the same
3268 'visible_hostname' you must give each machine a different
3269 'unique_hostname' so forwarding loops can be detected.
3272 NAME: hostname_aliases
3274 LOC: Config.hostnameAliases
3277 A list of other DNS names your cache has.
3281 OPTIONS FOR THE CACHE REGISTRATION SERVICE
3282 -----------------------------------------------------------------------------
3284 This section contains parameters for the (optional) cache
3285 announcement service. This service is provided to help
3286 cache administrators locate one another in order to join or
3287 create cache hierarchies.
3289 An 'announcement' message is sent (via UDP) to the registration
3290 service by Squid. By default, the announcement message is NOT
3291 SENT unless you enable it with 'announce_period' below.
3293 The announcement message includes your hostname, plus the
3294 following information from this configuration file:
3300 All current information is processed regularly and made
3301 available on the Web at http://www.ircache.net/Cache/Tracker/.
3304 NAME: announce_period
3306 LOC: Config.Announce.period
3309 This is how frequently to send cache announcements. The
3310 default is `0' which disables sending the announcement
3313 To enable announcing your cache, just uncomment the line
3317 #To enable announcing your cache, just uncomment the line below.
3318 #announce_period 1 day
3324 DEFAULT: tracker.ircache.net
3325 LOC: Config.Announce.host
3331 LOC: Config.Announce.file
3337 LOC: Config.Announce.port
3339 announce_host and announce_port set the hostname and port
3340 number where the registration message will be sent.
3342 Hostname will default to 'tracker.ircache.net' and port will
3343 default default to 3131. If the 'filename' argument is given,
3344 the contents of that file will be included in the announce
3349 HTTPD-ACCELERATOR OPTIONS
3350 -----------------------------------------------------------------------------
3353 NAME: httpd_accel_surrogate_id
3356 LOC: Config.Accel.surrogate_id
3359 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
3360 need an identification token to allow control targeting. Because
3361 a farm of surrogates may all perform the same tasks, they may share
3362 an identification token.
3365 NAME: http_accel_surrogate_remote
3370 LOC: Config.onoff.surrogate_is_remote
3372 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
3373 Set this to on to have squid behave as a remote surrogate.
3378 COMMENT: libxml2|expat|custom
3380 LOC: ESIParser::Type
3383 ESI markup is not strictly XML compatible. The custom ESI parser
3384 will give higher performance, but cannot handle non ASCII character
3389 DELAY POOL PARAMETERS
3390 -----------------------------------------------------------------------------
3394 TYPE: delay_pool_count
3399 This represents the number of delay pools to be used. For example,
3400 if you have one class 2 delay pool and one class 3 delays pool, you
3401 have a total of 2 delay pools.
3405 TYPE: delay_pool_class
3410 This defines the class of each delay pool. There must be exactly one
3411 delay_class line for each delay pool. For example, to define two
3412 delay pools, one of class 2 and one of class 3, the settings above
3416 delay_pools 4 # 4 delay pools
3417 delay_class 1 2 # pool 1 is a class 2 pool
3418 delay_class 2 3 # pool 2 is a class 3 pool
3419 delay_class 3 4 # pool 3 is a class 4 pool
3420 delay_class 4 5 # pool 4 is a class 5 pool
3422 The delay pool classes are:
3424 class 1 Everything is limited by a single aggregate
3427 class 2 Everything is limited by a single aggregate
3428 bucket as well as an "individual" bucket chosen
3429 from bits 25 through 32 of the IP address.
3431 class 3 Everything is limited by a single aggregate
3432 bucket as well as a "network" bucket chosen
3433 from bits 17 through 24 of the IP address and a
3434 "individual" bucket chosen from bits 17 through
3435 32 of the IP address.
3437 class 4 Everything in a class 3 delay pool, with an
3438 additional limit on a per user basis. This
3439 only takes effect if the username is established
3440 in advance - by forcing authentication in your
3443 class 5 Requests are grouped according their tag (see
3444 external_acl's tag= reply).
3446 NOTE: If an IP address is a.b.c.d
3447 -> bits 25 through 32 are "d"
3448 -> bits 17 through 24 are "c"
3449 -> bits 17 through 32 are "c * 256 + d"
3453 TYPE: delay_pool_access
3458 This is used to determine which delay pool a request falls into.
3460 delay_access is sorted per pool and the matching starts with pool 1,
3461 then pool 2, ..., and finally pool N. The first delay pool where the
3462 request is allowed is selected for the request. If it does not allow
3463 the request to any pool then the request is not delayed (default).
3465 For example, if you want some_big_clients in delay
3466 pool 1 and lotsa_little_clients in delay pool 2:
3469 delay_access 1 allow some_big_clients
3470 delay_access 1 deny all
3471 delay_access 2 allow lotsa_little_clients
3472 delay_access 2 deny all
3473 delay_access 3 allow authenticated_clients
3476 NAME: delay_parameters
3477 TYPE: delay_pool_rates
3482 This defines the parameters for a delay pool. Each delay pool has
3483 a number of "buckets" associated with it, as explained in the
3484 description of delay_class. For a class 1 delay pool, the syntax is:
3486 delay_parameters pool aggregate
3488 For a class 2 delay pool:
3490 delay_parameters pool aggregate individual
3492 For a class 3 delay pool:
3494 delay_parameters pool aggregate network individual
3496 For a class 4 delay pool:
3498 delay_parameters pool aggregate network individual user
3500 For a class 5 delay pool:
3502 delay_parameters pool tag
3504 The variables here are:
3506 pool a pool number - ie, a number between 1 and the
3507 number specified in delay_pools as used in
3510 aggregate the "delay parameters" for the aggregate bucket
3513 individual the "delay parameters" for the individual
3514 buckets (class 2, 3).
3516 network the "delay parameters" for the network buckets
3519 user the delay parameters for the user buckets
3522 tag the delay parameters for the tag buckets
3525 A pair of delay parameters is written restore/maximum, where restore is
3526 the number of bytes (not bits - modem and network speeds are usually
3527 quoted in bits) per second placed into the bucket, and maximum is the
3528 maximum number of bytes which can be in the bucket at any time.
3530 For example, if delay pool number 1 is a class 2 delay pool as in the
3531 above example, and is being used to strictly limit each host to 64kbps
3532 (plus overheads), with no overall limit, the line is:
3534 delay_parameters 1 -1/-1 8000/8000
3536 Note that the figure -1 is used to represent "unlimited".
3538 And, if delay pool number 2 is a class 3 delay pool as in the above
3539 example, and you want to limit it to a total of 256kbps (strict limit)
3540 with each 8-bit network permitted 64kbps (strict limit) and each
3541 individual host permitted 4800bps with a bucket maximum size of 64kb
3542 to permit a decent web page to be downloaded at a decent speed
3543 (if the network is not being limited due to overuse) but slow down
3544 large downloads more significantly:
3546 delay_parameters 2 32000/32000 8000/8000 600/8000
3548 There must be one delay_parameters line for each delay pool.
3550 Finally, for a class 4 delay pool as in the example - each user will
3551 be limited to 128Kb no matter how many workstations they are logged into.:
3553 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
3556 NAME: delay_initial_bucket_level
3557 COMMENT: (percent, 0-100)
3561 LOC: Config.Delay.initial
3563 The initial bucket percentage is used to determine how much is put
3564 in each bucket when squid starts, is reconfigured, or first notices
3565 a host accessing it (in class 2 and class 3, individual hosts and
3566 networks only have buckets associated with them once they have been
3571 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
3572 -----------------------------------------------------------------------------
3577 LOC: Config.Wccp.router
3582 TYPE: sockaddr_in_list
3583 LOC: Config.Wccp2.router
3587 Use this option to define your WCCP ``home'' router for
3590 wccp_router supports a single WCCP(v1) router
3592 wccp2_router supports multiple WCCPv2 routers
3594 only one of the two may be used at the same time and defines
3595 which version of WCCP to use.
3600 LOC: Config.Wccp.version
3604 This directive is only relevant if you need to set up WCCP(v1)
3605 to some very old and end-of-life Cisco routers. In all other
3606 setups it must be left unset or at the default setting.
3607 It defines an internal version in the WCCP(v1) protocol,
3608 with version 4 being the officially documented protocol.
3610 According to some users, Cisco IOS 11.2 and earlier only
3611 support WCCP version 3. If you're using that or an earlier
3612 version of IOS, you may need to change this value to 3, otherwise
3613 do not specify this parameter.
3616 NAME: wccp2_rebuild_wait
3618 LOC: Config.Wccp2.rebuildwait
3622 If this is enabled Squid will wait for the cache dir rebuild to finish
3623 before sending the first wccp2 HereIAm packet
3626 NAME: wccp2_forwarding_method
3628 LOC: Config.Wccp2.forwarding_method
3632 WCCP2 allows the setting of forwarding methods between the
3633 router/switch and the cache. Valid values are as follows:
3635 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
3636 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
3638 Currently (as of IOS 12.4) cisco routers only support GRE.
3639 Cisco switches only support the L2 redirect assignment method.
3642 NAME: wccp2_return_method
3644 LOC: Config.Wccp2.return_method
3648 WCCP2 allows the setting of return methods between the
3649 router/switch and the cache for packets that the cache
3650 decides not to handle. Valid values are as follows:
3652 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
3653 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
3655 Currently (as of IOS 12.4) cisco routers only support GRE.
3656 Cisco switches only support the L2 redirect assignment.
3658 If the "ip wccp redirect exclude in" command has been
3659 enabled on the cache interface, then it is still safe for
3660 the proxy server to use a l2 redirect method even if this
3661 option is set to GRE.
3664 NAME: wccp2_assignment_method
3666 LOC: Config.Wccp2.assignment_method
3670 WCCP2 allows the setting of methods to assign the WCCP hash
3671 Valid values are as follows:
3676 As a general rule, cisco routers support the hash assignment method
3677 and cisco switches support the mask assignment method.
3682 LOC: Config.Wccp2.info
3684 DEFAULT_IF_NONE: standard 0
3687 WCCP2 allows for multiple traffic services. There are two
3688 types: "standard" and "dynamic". The standard type defines
3689 one service id - http (id 0). The dynamic service ids can be from
3690 51 to 255 inclusive. In order to use a dynamic service id
3691 one must define the type of traffic to be redirected; this is done
3692 using the wccp2_service_info option.
3694 The "standard" type does not require a wccp2_service_info option,
3695 just specifying the service id will suffice.
3697 MD5 service authentication can be enabled by adding
3698 "password=<password>" to the end of this service declaration.
3702 wccp2_service standard 0 # for the 'web-cache' standard service
3703 wccp2_service dynamic 80 # a dynamic service type which will be
3704 # fleshed out with subsequent options.
3705 wccp2_service standard 0 password=foo
3709 NAME: wccp2_service_info
3710 TYPE: wccp2_service_info
3711 LOC: Config.Wccp2.info
3715 Dynamic WCCPv2 services require further information to define the
3716 traffic you wish to have diverted.
3720 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
3721 priority=<priority> ports=<port>,<port>..
3723 The relevant WCCPv2 flags:
3724 + src_ip_hash, dst_ip_hash
3725 + source_port_hash, dest_port_hash
3726 + src_ip_alt_hash, dst_ip_alt_hash
3727 + src_port_alt_hash, dst_port_alt_hash
3730 The port list can be one to eight entries.
3734 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
3735 priority=240 ports=80
3737 Note: the service id must have been defined by a previous
3738 'wccp2_service dynamic <id>' entry.
3743 LOC: Config.Wccp2.weight
3747 Each cache server gets assigned a set of the destination
3748 hash proportional to their weight.
3753 LOC: Config.Wccp.address
3759 LOC: Config.Wccp2.address
3763 Use this option if you require WCCP to use a specific
3766 The default behavior is to not bind to any specific address.
3770 PERSISTENT CONNECTION HANDLING
3771 -----------------------------------------------------------------------------
3773 Also see "pconn_timeout" in the TIMEOUTS section
3776 NAME: client_persistent_connections
3778 LOC: Config.onoff.client_pconns
3782 NAME: server_persistent_connections
3784 LOC: Config.onoff.server_pconns
3787 Persistent connection support for clients and servers. By
3788 default, Squid uses persistent connections (when allowed)
3789 with its clients and servers. You can use these options to
3790 disable persistent connections with clients and/or servers.
3793 NAME: persistent_connection_after_error
3795 LOC: Config.onoff.error_pconns
3798 With this directive the use of persistent connections after
3799 HTTP errors can be disabled. Useful if you have clients
3800 who fail to handle errors on persistent connections proper.
3803 NAME: detect_broken_pconn
3805 LOC: Config.onoff.detect_broken_server_pconns
3808 Some servers have been found to incorrectly signal the use
3809 of HTTP/1.0 persistent connections even on replies not
3810 compatible, causing significant delays. This server problem
3811 has mostly been seen on redirects.
3813 By enabling this directive Squid attempts to detect such
3814 broken replies and automatically assume the reply is finished
3815 after 10 seconds timeout.
3819 CACHE DIGEST OPTIONS
3820 -----------------------------------------------------------------------------
3823 NAME: digest_generation
3824 IFDEF: USE_CACHE_DIGESTS
3826 LOC: Config.onoff.digest_generation
3829 This controls whether the server will generate a Cache Digest
3830 of its contents. By default, Cache Digest generation is
3831 enabled if Squid is compiled with USE_CACHE_DIGESTS defined.
3834 NAME: digest_bits_per_entry
3835 IFDEF: USE_CACHE_DIGESTS
3837 LOC: Config.digest.bits_per_entry
3840 This is the number of bits of the server's Cache Digest which
3841 will be associated with the Digest entry for a given HTTP
3842 Method and URL (public key) combination. The default is 5.
3845 NAME: digest_rebuild_period
3846 IFDEF: USE_CACHE_DIGESTS
3849 LOC: Config.digest.rebuild_period
3852 This is the number of seconds between Cache Digest rebuilds.
3855 NAME: digest_rewrite_period
3857 IFDEF: USE_CACHE_DIGESTS
3859 LOC: Config.digest.rewrite_period
3862 This is the number of seconds between Cache Digest writes to
3866 NAME: digest_swapout_chunk_size
3869 IFDEF: USE_CACHE_DIGESTS
3870 LOC: Config.digest.swapout_chunk_size
3873 This is the number of bytes of the Cache Digest to write to
3874 disk at a time. It defaults to 4096 bytes (4KB), the Squid
3878 NAME: digest_rebuild_chunk_percentage
3879 COMMENT: (percent, 0-100)
3880 IFDEF: USE_CACHE_DIGESTS
3882 LOC: Config.digest.rebuild_chunk_percentage
3885 This is the percentage of the Cache Digest to be scanned at a
3886 time. By default it is set to 10% of the Cache Digest.
3891 -----------------------------------------------------------------------------
3896 LOC: Config.Port.snmp
3900 The port number where Squid listens for SNMP requests. To enable
3901 SNMP support set this to a suitable port number. Port number
3902 3401 is often used for the Squid SNMP agent. By default it's
3903 set to "0" (disabled)
3911 LOC: Config.accessList.snmp
3913 DEFAULT_IF_NONE: deny all
3916 Allowing or denying access to the SNMP port.
3918 All access to the agent is denied by default.
3921 snmp_access allow|deny [!]aclname ...
3924 snmp_access allow snmppublic localhost
3925 snmp_access deny all
3928 NAME: snmp_incoming_address
3930 LOC: Config.Addrs.snmp_incoming
3934 NAME: snmp_outgoing_address
3936 LOC: Config.Addrs.snmp_outgoing
3937 DEFAULT: 255.255.255.255
3940 Just like 'udp_incoming_address' above, but for the SNMP port.
3942 snmp_incoming_address is used for the SNMP socket receiving
3943 messages from SNMP agents.
3944 snmp_outgoing_address is used for SNMP packets returned to SNMP
3947 The default snmp_incoming_address (0.0.0.0) is to listen on all
3948 available network interfaces.
3950 If snmp_outgoing_address is set to 255.255.255.255 (the default)
3951 it will use the same socket as snmp_incoming_address. Only
3952 change this if you want to have SNMP replies sent using another
3953 address than where this Squid listens for SNMP queries.
3955 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
3956 the same value since they both use port 3401.
3961 -----------------------------------------------------------------------------
3964 NAME: icp_port udp_port
3967 LOC: Config.Port.icp
3969 The port number where Squid sends and receives ICP queries to
3970 and from neighbor caches. The standard UDP port for ICP is 3130.
3971 Default is disabled (0).
3973 icp_port @DEFAULT_ICP_PORT@
3981 LOC: Config.Port.htcp
3983 The port number where Squid sends and receives HTCP queries to
3984 and from neighbor caches. To turn it on you want to set it to
3985 4827. By default it is set to "0" (disabled).
3991 NAME: log_icp_queries
3995 LOC: Config.onoff.log_udp
3997 If set, ICP queries are logged to access.log. You may wish
3998 do disable this if your ICP load is VERY high to speed things
3999 up or to simplify log analysis.
4002 NAME: udp_incoming_address
4004 LOC:Config.Addrs.udp_incoming
4008 NAME: udp_outgoing_address
4010 LOC: Config.Addrs.udp_outgoing
4011 DEFAULT: 255.255.255.255
4013 udp_incoming_address is used for the ICP socket receiving packets
4015 udp_outgoing_address is used for ICP packets sent out to other
4018 The default behavior is to not bind to any specific address.
4020 A udp_incoming_address value of 0.0.0.0 indicates Squid
4021 should listen for UDP messages on all available interfaces.
4023 If udp_outgoing_address is set to 255.255.255.255 (the default)
4024 it will use the same socket as udp_incoming_address. Only
4025 change this if you want to have ICP queries sent using another
4026 address than where this Squid listens for ICP queries from other
4029 NOTE, udp_incoming_address and udp_outgoing_address can not
4030 have the same value since they both use port 3130.
4037 LOC: Config.onoff.icp_hit_stale
4039 If you want to return ICP_HIT for stale cache objects, set this
4040 option to 'on'. If you have sibling relationships with caches
4041 in other administrative domains, this should be 'off'. If you only
4042 have sibling relationships with caches under your control,
4043 it is probably okay to set this to 'on'.
4044 If set to 'on', your siblings should use the option "allow-miss"
4045 on their cache_peer lines for connecting to you.
4048 NAME: minimum_direct_hops
4051 LOC: Config.minDirectHops
4053 If using the ICMP pinging stuff, do direct fetches for sites
4054 which are no more than this many hops away.
4057 NAME: minimum_direct_rtt
4060 LOC: Config.minDirectRtt
4062 If using the ICMP pinging stuff, do direct fetches for sites
4063 which are no more than this many rtt milliseconds away.
4069 LOC: Config.Netdb.low
4075 LOC: Config.Netdb.high
4077 The low and high water marks for the ICMP measurement
4078 database. These are counts, not percents. The defaults are
4079 900 and 1000. When the high water mark is reached, database
4080 entries will be deleted until the low mark is reached.
4083 NAME: netdb_ping_period
4085 LOC: Config.Netdb.period
4088 The minimum period for measuring a site. There will be at
4089 least this much delay between successive pings to the same
4090 network. The default is five minutes.
4097 LOC: Config.onoff.query_icmp
4099 If you want to ask your peers to include ICMP data in their ICP
4100 replies, enable this option.
4102 If your peer has configured Squid (during compilation) with
4103 '--enable-icmp' that peer will send ICMP pings to origin server
4104 sites of the URLs it receives. If you enable this option the
4105 ICP replies from that peer will include the ICMP data (if available).
4106 Then, when choosing a parent cache, Squid will choose the parent with
4107 the minimal RTT to the origin server. When this happens, the
4108 hierarchy field of the access.log will be
4109 "CLOSEST_PARENT_MISS". This option is off by default.
4112 NAME: test_reachability
4116 LOC: Config.onoff.test_reachability
4118 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
4119 instead of ICP_MISS if the target host is NOT in the ICMP
4120 database, or has a zero RTT.
4123 NAME: icp_query_timeout
4127 LOC: Config.Timeout.icp_query
4129 Normally Squid will automatically determine an optimal ICP
4130 query timeout value based on the round-trip-time of recent ICP
4131 queries. If you want to override the value determined by
4132 Squid, set this 'icp_query_timeout' to a non-zero value. This
4133 value is specified in MILLISECONDS, so, to use a 2-second
4134 timeout (the old default), you would write:
4136 icp_query_timeout 2000
4139 NAME: maximum_icp_query_timeout
4143 LOC: Config.Timeout.icp_query_max
4145 Normally the ICP query timeout is determined dynamically. But
4146 sometimes it can lead to very large values (say 5 seconds).
4147 Use this option to put an upper limit on the dynamic timeout
4148 value. Do NOT use this option to always use a fixed (instead
4149 of a dynamic) timeout value. To set a fixed timeout see the
4150 'icp_query_timeout' directive.
4153 NAME: minimum_icp_query_timeout
4157 LOC: Config.Timeout.icp_query_min
4159 Normally the ICP query timeout is determined dynamically. But
4160 sometimes it can lead to very small timeouts, even lower than
4161 the normal latency variance on your link due to traffic.
4162 Use this option to put an lower limit on the dynamic timeout
4163 value. Do NOT use this option to always use a fixed (instead
4164 of a dynamic) timeout value. To set a fixed timeout see the
4165 'icp_query_timeout' directive.
4168 NAME: background_ping_rate
4172 LOC: Config.backgroundPingRate
4174 Controls how often the ICP pings are sent to siblings that
4175 have background-ping set.
4179 MULTICAST ICP OPTIONS
4180 -----------------------------------------------------------------------------
4185 LOC: Config.mcast_group_list
4188 This tag specifies a list of multicast groups which your server
4189 should join to receive multicasted ICP queries.
4191 NOTE! Be very careful what you put here! Be sure you
4192 understand the difference between an ICP _query_ and an ICP
4193 _reply_. This option is to be set only if you want to RECEIVE
4194 multicast queries. Do NOT set this option to SEND multicast
4195 ICP (use cache_peer for that). ICP replies are always sent via
4196 unicast, so this option does not affect whether or not you will
4197 receive replies from multicast group members.
4199 You must be very careful to NOT use a multicast address which
4200 is already in use by another group of caches.
4202 If you are unsure about multicast, please read the Multicast
4203 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
4205 Usage: mcast_groups 239.128.16.128 224.0.1.20
4207 By default, Squid doesn't listen on any multicast groups.
4210 NAME: mcast_miss_addr
4211 IFDEF: MULTICAST_MISS_STREAM
4213 LOC: Config.mcast_miss.addr
4214 DEFAULT: 255.255.255.255
4216 If you enable this option, every "cache miss" URL will
4217 be sent out on the specified multicast address.
4219 Do not enable this option unless you are are absolutely
4220 certain you understand what you are doing.
4223 NAME: mcast_miss_ttl
4224 IFDEF: MULTICAST_MISS_STREAM
4226 LOC: Config.mcast_miss.ttl
4229 This is the time-to-live value for packets multicasted
4230 when multicasting off cache miss URLs is enabled. By
4231 default this is set to 'site scope', i.e. 16.
4234 NAME: mcast_miss_port
4235 IFDEF: MULTICAST_MISS_STREAM
4237 LOC: Config.mcast_miss.port
4240 This is the port number to be used in conjunction with
4244 NAME: mcast_miss_encode_key
4245 IFDEF: MULTICAST_MISS_STREAM
4247 LOC: Config.mcast_miss.encode_key
4248 DEFAULT: XXXXXXXXXXXXXXXX
4250 The URLs that are sent in the multicast miss stream are
4251 encrypted. This is the encryption key.
4254 NAME: mcast_icp_query_timeout
4258 LOC: Config.Timeout.mcast_icp_query
4260 For multicast peers, Squid regularly sends out ICP "probes" to
4261 count how many other peers are listening on the given multicast
4262 address. This value specifies how long Squid should wait to
4263 count all the replies. The default is 2000 msec, or 2
4268 INTERNAL ICON OPTIONS
4269 -----------------------------------------------------------------------------
4272 NAME: icon_directory
4274 LOC: Config.icons.directory
4275 DEFAULT: @DEFAULT_ICON_DIR@
4277 Where the icons are stored. These are normally kept in
4281 NAME: global_internal_static
4283 LOC: Config.onoff.global_internal_static
4286 This directive controls is Squid should intercept all requests for
4287 /squid-internal-static/ no matter which host the URL is requesting
4288 (default on setting), or if nothing special should be done for
4289 such URLs (off setting). The purpose of this directive is to make
4290 icons etc work better in complex cache hierarchies where it may
4291 not always be possible for all corners in the cache mesh to reach
4292 the server generating a directory listing.
4295 NAME: short_icon_urls
4297 LOC: Config.icons.use_short_names
4300 If this is enabled Squid will use short URLs for icons.
4301 If disabled it will revert to the old behavior of including
4302 it's own name and port in the URL.
4304 If you run a complex cache hierarchy with a mix of Squid and
4305 other proxies you may need to disable this directive.
4310 -----------------------------------------------------------------------------
4313 NAME: error_directory
4315 LOC: Config.errorDirectory
4316 DEFAULT: @DEFAULT_ERROR_DIR@
4318 If you wish to create your own versions of the default
4319 (English) error files, either to customize them to suit your
4320 language or company copy the template English files to another
4321 directory and point this tag at them.
4323 The squid developers are interested in making squid available in
4324 a wide variety of languages. If you are making translations for a
4325 langauge that Squid does not currently provide please consider
4326 contributing your translation back to the project.
4331 LOC: Config.errHtmlText
4334 HTML text to include in error messages. Make this a "mailto"
4335 URL to your admin address, or maybe just a link to your
4336 organizations Web page.
4338 To include this in your error messages, you must rewrite
4339 the error template files (found in the "errors" directory).
4340 Wherever you want the 'err_html_text' line to appear,
4341 insert a %L tag in the error template file.
4344 NAME: email_err_data
4347 LOC: Config.onoff.emailErrData
4350 If enabled, information about the occurred error will be
4351 included in the mailto links of the ERR pages (if %W is set)
4352 so that the email body contains the data.
4353 Syntax is <A HREF="mailto:%w%W">%w</A>
4358 LOC: Config.denyInfoList
4361 Usage: deny_info err_page_name acl
4362 or deny_info http://... acl
4363 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
4365 This can be used to return a ERR_ page for requests which
4366 do not pass the 'http_access' rules. Squid remembers the last
4367 acl it evaluated in http_access, and if a 'deny_info' line exists
4368 for that ACL Squid returns a corresponding error page.
4370 The acl is typically the last acl on the http_access deny line which
4371 denied access. The exceptions to this rule are:
4372 - When Squid needs to request authentication credentials. It's then
4373 the first authentication related acl encountered
4374 - When none of the http_access lines matches. It's then the last
4375 acl processed on the last http_access line.
4377 You may use ERR_ pages that come with Squid or create your own pages
4378 and put them into the configured errors/ directory.
4380 Alternatively you can specify an error URL. The browsers will
4381 get redirected (302) to the specified URL. %s in the redirection
4382 URL will be replaced by the requested URL.
4384 Alternatively you can tell Squid to reset the TCP connection
4385 by specifying TCP_RESET.
4389 OPTIONS INFLUENCING REQUEST FORWARDING
4390 -----------------------------------------------------------------------------
4393 NAME: nonhierarchical_direct
4395 LOC: Config.onoff.nonhierarchical_direct
4398 By default, Squid will send any non-hierarchical requests
4399 (matching hierarchy_stoplist or not cacheable request type) direct
4402 If you set this to off, Squid will prefer to send these
4403 requests to parents.
4405 Note that in most configurations, by turning this off you will only
4406 add latency to these request without any improvement in global hit
4409 If you are inside an firewall see never_direct instead of
4415 LOC: Config.onoff.prefer_direct
4418 Normally Squid tries to use parents for most requests. If you for some
4419 reason like it to first try going direct and only use a parent if
4420 going direct fails set this to on.
4422 By combining nonhierarchical_direct off and prefer_direct on you
4423 can set up Squid to use a parent as a backup path if going direct
4426 Note: If you want Squid to use parents for all requests see
4427 the never_direct directive. prefer_direct only modifies how Squid
4428 acts on cacheable requests.
4433 LOC: Config.accessList.AlwaysDirect
4436 Usage: always_direct allow|deny [!]aclname ...
4438 Here you can use ACL elements to specify requests which should
4439 ALWAYS be forwarded by Squid to the origin servers without using
4440 any peers. For example, to always directly forward requests for
4441 local servers ignoring any parents or siblings you may have use
4444 acl local-servers dstdomain my.domain.net
4445 always_direct allow local-servers
4447 To always forward FTP requests directly, use
4450 always_direct allow FTP
4452 NOTE: There is a similar, but opposite option named
4453 'never_direct'. You need to be aware that "always_direct deny
4454 foo" is NOT the same thing as "never_direct allow foo". You
4455 may need to use a deny rule to exclude a more-specific case of
4456 some other rule. Example:
4458 acl local-external dstdomain external.foo.net
4459 acl local-servers dstdomain .foo.net
4460 always_direct deny local-external
4461 always_direct allow local-servers
4463 NOTE: If your goal is to make the client forward the request
4464 directly to the origin server bypassing Squid then this needs
4465 to be done in the client configuration. Squid configuration
4466 can only tell Squid how Squid should fetch the object.
4468 NOTE: This directive is not related to caching. The replies
4469 is cached as usual even if you use always_direct. To not cache
4470 the replies see no_cache.
4472 This option replaces some v1.1 options such as local_domain
4478 LOC: Config.accessList.NeverDirect
4481 Usage: never_direct allow|deny [!]aclname ...
4483 never_direct is the opposite of always_direct. Please read
4484 the description for always_direct if you have not already.
4486 With 'never_direct' you can use ACL elements to specify
4487 requests which should NEVER be forwarded directly to origin
4488 servers. For example, to force the use of a proxy for all
4489 requests, except those in your local domain use something like:
4491 acl local-servers dstdomain .foo.net
4492 acl all src 0.0.0.0/0.0.0.0
4493 never_direct deny local-servers
4494 never_direct allow all
4496 or if Squid is inside a firewall and there are local intranet
4497 servers inside the firewall use something like:
4499 acl local-intranet dstdomain .foo.net
4500 acl local-external dstdomain external.foo.net
4501 always_direct deny local-external
4502 always_direct allow local-intranet
4503 never_direct allow all
4505 This option replaces some v1.1 options such as inside_firewall
4510 ADVANCED NETWORKING OPTIONS
4511 -----------------------------------------------------------------------------
4514 NAME: incoming_icp_average
4517 LOC: Config.comm_incoming.icp_average
4520 NAME: incoming_http_average
4523 LOC: Config.comm_incoming.http_average
4526 NAME: incoming_dns_average
4529 LOC: Config.comm_incoming.dns_average
4532 NAME: min_icp_poll_cnt
4535 LOC: Config.comm_incoming.icp_min_poll
4538 NAME: min_dns_poll_cnt
4541 LOC: Config.comm_incoming.dns_min_poll
4544 NAME: min_http_poll_cnt
4547 LOC: Config.comm_incoming.http_min_poll
4549 Heavy voodoo here. I can't even believe you are reading this.
4550 Are you crazy? Don't even think about adjusting these unless
4551 you understand the algorithms in comm_select.c first!
4557 LOC: Config.accept_filter
4561 The name of an accept(2) filter to install on Squid's
4562 listen socket(s). This feature is perhaps specific to
4563 FreeBSD and requires support in the kernel.
4565 The 'httpready' filter delays delivering new connections
4566 to Squid until a full HTTP request has been received.
4567 See the accf_http(9) man page for details.
4569 The 'dataready' filter delays delivering new connections
4570 to Squid until there is some data to process.
4571 See the accf_dataready(9) man page for details.
4575 The 'data' filter delays delivering of new connections
4576 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
4577 You may optionally specify a number of seconds to wait by
4578 'data=N' where N is the number of seconds. Defaults to 30
4579 if not specified. See the tcp(7) man page for details.
4582 accept_filter httpready
4587 NAME: tcp_recv_bufsize
4591 LOC: Config.tcpRcvBufsz
4593 Size of receive buffer to set for TCP sockets. Probably just
4594 as easy to change your kernel's default. Set to zero to use
4595 the default buffer size.
4600 -----------------------------------------------------------------------------
4607 LOC: TheICAPConfig.onoff
4610 If you want to enable the ICAP module support, set this to on.
4613 NAME: icap_connect_timeout
4616 LOC: TheICAPConfig.connect_timeout_raw
4619 This parameter specifies how long to wait for the TCP connect to
4620 the requested ICAP server to complete before giving up and either
4621 terminating the HTTP transaction or bypassing the failure.
4623 The default for optional services is peer_connect_timeout.
4624 The default for essential services is connect_timeout.
4625 If this option is explicitly set, its value applies to all services.
4628 NAME: icap_io_timeout
4632 LOC: TheICAPConfig.io_timeout_raw
4635 This parameter specifies how long to wait for an I/O activity on
4636 an established, active ICAP connection before giving up and
4637 either terminating the HTTP transaction or bypassing the
4640 The default is read_timeout.
4643 NAME: icap_service_failure_limit
4646 LOC: TheICAPConfig.service_failure_limit
4649 The limit specifies the number of failures that Squid tolerates
4650 when establishing a new TCP connection with an ICAP service. If
4651 the number of failures exceeds the limit, the ICAP service is
4652 not used for new ICAP requests until it is time to refresh its
4653 OPTIONS. The per-service failure counter is reset to zero each
4654 time Squid fetches new service OPTIONS.
4656 A negative value disables the limit. Without the limit, an ICAP
4657 service will not be considered down due to connectivity failures
4658 between ICAP OPTIONS requests.
4661 NAME: icap_service_revival_delay
4664 LOC: TheICAPConfig.service_revival_delay
4667 The delay specifies the number of seconds to wait after an ICAP
4668 OPTIONS request failure before requesting the options again. The
4669 failed ICAP service is considered "down" until fresh OPTIONS are
4672 The actual delay cannot be smaller than the hardcoded minimum
4673 delay of 30 seconds.
4676 NAME: icap_preview_enable
4680 LOC: TheICAPConfig.preview_enable
4683 Set this to 'on' if you want to enable the ICAP preview
4687 NAME: icap_preview_size
4690 LOC: TheICAPConfig.preview_size
4693 The default size of preview data to be sent to the ICAP server.
4694 -1 means no preview. This value might be overwritten on a per server
4695 basis by OPTIONS requests.
4698 NAME: icap_default_options_ttl
4701 LOC: TheICAPConfig.default_options_ttl
4704 The default TTL value for ICAP OPTIONS responses that don't have
4705 an Options-TTL header.
4708 NAME: icap_persistent_connections
4712 LOC: TheICAPConfig.reuse_connections
4715 Whether or not Squid should use persistent connections to
4719 NAME: icap_send_client_ip
4723 LOC: TheICAPConfig.send_client_ip
4726 This adds the header "X-Client-IP" to ICAP requests.
4729 NAME: icap_send_client_username
4733 LOC: TheICAPConfig.send_client_username
4736 This sends authenticated HTTP client username (if available) to
4737 the ICAP service. The username value is encoded based on the
4738 icap_client_username_encode option and is sent using the header
4739 specified by the icap_client_username_header option.
4742 NAME: icap_client_username_header
4745 LOC: TheICAPConfig.client_username_header
4746 DEFAULT: X-Client-Username
4748 ICAP request header name to use for send_client_username.
4751 NAME: icap_client_username_encode
4755 LOC: TheICAPConfig.client_username_encode
4758 Whether to base64 encode the authenticated client username.
4762 TYPE: icap_service_type
4767 Defines a single ICAP service
4769 icap_service servicename vectoring_point bypass service_url
4771 vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
4772 This specifies at which point of transaction processing the
4773 ICAP service should be activated. *_postcache vectoring points
4774 are not yet supported.
4776 If set to 1, the ICAP service is treated as optional. If the
4777 service cannot be reached or malfunctions, Squid will try to
4778 ignore any errors and process the message as if the service
4779 was not enabled. No all ICAP errors can be bypassed.
4780 If set to 0, the ICAP service is treated as essential and all
4781 ICAP errors will result in an error page returned to the
4783 service_url = icap://servername:port/service
4786 icap_service service_1 reqmod_precache 0 icap://icap1.mydomain.net:1344/reqmod
4787 icap_service service_2 respmod_precache 0 icap://icap2.mydomain.net:1344/respmod
4791 TYPE: icap_class_type
4796 Defines an ICAP service chain. If there are multiple services per
4797 vectoring point, they are processed in the specified order.
4799 icap_class classname servicename...
4802 icap_class class_1 service_1 service_2
4803 icap class class_2 service_1 service_3
4807 TYPE: icap_access_type
4812 Redirects a request through an ICAP service class, depending
4815 icap_access classname allow|deny [!]aclname...
4817 The icap_access statements are processed in the order they appear in
4818 this configuration file. If an access list matches, the processing stops.
4819 For an "allow" rule, the specified class is used for the request. A "deny"
4820 rule simply stops processing without using the class. You can also use the
4821 special classname "None".
4823 For backward compatibility, it is also possible to use services
4826 icap_access class_1 allow all
4831 -----------------------------------------------------------------------------
4834 NAME: check_hostnames
4837 LOC: Config.onoff.check_hostnames
4839 For security and stability reasons Squid can check
4840 hostnames for Internet standard RFC compliance. If you want
4841 Squid to perform these checks turn this directive on.
4844 NAME: allow_underscore
4847 LOC: Config.onoff.allow_underscore
4849 Underscore characters is not strictly allowed in Internet hostnames
4850 but nevertheless used by many sites. Set this to off if you want
4851 Squid to be strict about the standard.
4852 This check is performed only when check_hostnames is set to on.
4855 NAME: cache_dns_program
4857 IFDEF: USE_DNSSERVERS
4858 DEFAULT: @DEFAULT_DNSSERVER@
4859 LOC: Config.Program.dnsserver
4861 Specify the location of the executable for dnslookup process.
4866 IFDEF: USE_DNSSERVERS
4868 LOC: Config.dnsChildren
4870 The number of processes spawn to service DNS name lookups.
4871 For heavily loaded caches on large servers, you should
4872 probably increase this value to at least 10. The maximum
4873 is 32. The default is 5.
4875 You must have at least one dnsserver process.
4878 NAME: dns_retransmit_interval
4881 LOC: Config.Timeout.idns_retransmit
4882 IFDEF: !USE_DNSSERVERS
4884 Initial retransmit interval for DNS queries. The interval is
4885 doubled each time all configured DNS servers have been tried.
4892 LOC: Config.Timeout.idns_query
4893 IFDEF: !USE_DNSSERVERS
4895 DNS Query timeout. If no response is received to a DNS query
4896 within this time all DNS servers for the queried domain
4897 are assumed to be unavailable.
4904 LOC: Config.onoff.res_defnames
4906 Normally the RES_DEFNAMES resolver option is disabled
4907 (see res_init(3)). This prevents caches in a hierarchy
4908 from interpreting single-component hostnames locally. To allow
4909 Squid to handle single-component names, enable this option.
4912 NAME: dns_nameservers
4915 LOC: Config.dns_nameservers
4917 Use this if you want to specify a list of DNS name servers
4918 (IP addresses) to use instead of those given in your
4919 /etc/resolv.conf file.
4920 On Windows platforms, if no value is specified here or in
4921 the /etc/resolv.conf file, the list of DNS name servers are
4922 taken from the Windows registry, both static and dynamic DHCP
4923 configurations are supported.
4925 Example: dns_nameservers 10.0.0.1 192.172.0.4
4930 DEFAULT: @DEFAULT_HOSTS@
4931 LOC: Config.etcHostsPath
4933 Location of the host-local IP name-address associations
4934 database. Most Operating Systems have such a file on different
4936 - Un*X & Linux: /etc/hosts
4937 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
4938 (%SystemRoot% value install default is c:\winnt)
4939 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
4940 (%SystemRoot% value install default is c:\windows)
4941 - Windows 9x/Me: %windir%\hosts
4942 (%windir% value is usually c:\windows)
4943 - Cygwin: /etc/hosts
4945 The file contains newline-separated definitions, in the
4946 form ip_address_in_dotted_form name [name ...] names are
4947 whitespace-separated. Lines beginning with an hash (#)
4948 character are comments.
4950 The file is checked at startup and upon configuration.
4951 If set to 'none', it won't be checked.
4952 If append_domain is used, that domain will be added to
4953 domain-local (i.e. not containing any dot character) host
4959 LOC: Config.dns_testname_list
4961 DEFAULT_IF_NONE: netscape.com internic.net nlanr.net microsoft.com
4963 The DNS tests exit as soon as the first site is successfully looked up
4965 This test can be disabled with the -D command line option.
4970 LOC: Config.appendDomain
4973 Appends local domain name to hostnames without any dots in
4974 them. append_domain must begin with a period.
4976 Be warned there are now Internet names with no dots in
4977 them using only top-domain names, so setting this may
4978 cause some Internet sites to become unavailable.
4981 append_domain .yourdomain.com
4984 NAME: ignore_unknown_nameservers
4986 LOC: Config.onoff.ignore_unknown_nameservers
4989 By default Squid checks that DNS responses are received
4990 from the same IP addresses they are sent to. If they
4991 don't match, Squid ignores the response and writes a warning
4992 message to cache.log. You can allow responses from unknown
4993 nameservers by setting this option to 'off'.
4997 COMMENT: (number of entries)
5000 LOC: Config.ipcache.size
5007 LOC: Config.ipcache.low
5014 LOC: Config.ipcache.high
5016 The size, low-, and high-water marks for the IP cache.
5019 NAME: fqdncache_size
5020 COMMENT: (number of entries)
5023 LOC: Config.fqdncache.size
5025 Maximum number of FQDN cache entries.
5030 -----------------------------------------------------------------------------
5037 LOC: Config.onoff.mem_pools
5039 If set, Squid will keep pools of allocated (but unused) memory
5040 available for future use. If memory is a premium on your
5041 system and you believe your malloc library outperforms Squid
5042 routines, disable this.
5045 NAME: memory_pools_limit
5049 LOC: Config.MemPools.limit
5051 Used only with memory_pools on:
5052 memory_pools_limit 50 MB
5054 If set to a non-zero value, Squid will keep at most the specified
5055 limit of allocated (but unused) memory in memory pools. All free()
5056 requests that exceed this limit will be handled by your malloc
5057 library. Squid does not pre-allocate any memory, just safe-keeps
5058 objects that otherwise would be free()d. Thus, it is safe to set
5059 memory_pools_limit to a reasonably high value even if your
5060 configuration will use less memory.
5062 If set to zero, Squid will keep all memory it can. That is, there
5063 will be no limit on the total amount of memory used for safe-keeping.
5065 To disable memory allocation optimization, do not set
5066 memory_pools_limit to 0. Set memory_pools to "off" instead.
5068 An overhead for maintaining memory pools is not taken into account
5069 when the limit is checked. This overhead is close to four bytes per
5070 object kept. However, pools may actually _save_ memory because of
5071 reduced memory thrashing in your malloc library.
5078 LOC: opt_forwarded_for
5080 If set, Squid will include your system's IP address or name
5081 in the HTTP requests it forwards. By default it looks like
5084 X-Forwarded-For: 192.1.2.3
5086 If you disable this, it will appear as
5088 X-Forwarded-For: unknown
5091 NAME: cachemgr_passwd
5092 TYPE: cachemgrpasswd
5094 LOC: Config.passwd_list
5096 Specify passwords for cachemgr operations.
5098 Usage: cachemgr_passwd password action action ...
5100 Some valid actions are (see cache manager menu for a full list):
5139 * Indicates actions which will not be performed without a
5140 valid password, others can be performed if not listed here.
5142 To disable an action, set the password to "disable".
5143 To allow performing an action without a password, set the
5146 Use the keyword "all" to set the same password for all actions.
5149 cachemgr_passwd secret shutdown
5150 cachemgr_passwd lesssssssecret info stats/objects
5151 cachemgr_passwd disable all
5158 LOC: Config.onoff.client_db
5160 If you want to disable collecting per-client statistics,
5161 turn off client_db here.
5164 NAME: refresh_all_ims
5168 LOC: Config.onoff.refresh_all_ims
5170 When you enable this option, squid will always check
5171 the origin server for an update when a client sends an
5172 If-Modified-Since request. Many browsers use IMS
5173 requests when the user requests a reload, and this
5174 ensures those clients receive the latest version.
5176 By default (off), squid may return a Not Modified response
5177 based on the age of the cached version.
5180 NAME: reload_into_ims
5181 IFDEF: HTTP_VIOLATIONS
5185 LOC: Config.onoff.reload_into_ims
5187 When you enable this option, client no-cache or ``reload''
5188 requests will be changed to If-Modified-Since requests.
5189 Doing this VIOLATES the HTTP standard. Enabling this
5190 feature could make you liable for problems which it
5193 see also refresh_pattern for a more selective approach.
5196 NAME: maximum_single_addr_tries
5198 LOC: Config.retry.maxtries
5201 This sets the maximum number of connection attempts for a
5202 host that only has one address (for multiple-address hosts,
5203 each address is tried once).
5205 The default value is one attempt, the (not recommended)
5206 maximum is 255 tries. A warning message will be generated
5207 if it is set to a value greater than ten.
5209 Note: This is in addition to the request re-forwarding which
5210 takes place if Squid fails to get a satisfying response.
5213 NAME: retry_on_error
5215 LOC: Config.retry.onerror
5218 If set to on Squid will automatically retry requests when
5219 receiving an error response. This is mainly useful if you
5220 are in a complex cache hierarchy to work around access
5224 NAME: as_whois_server
5226 LOC: Config.as_whois_server
5227 DEFAULT: whois.ra.net
5228 DEFAULT_IF_NONE: whois.ra.net
5230 WHOIS server to query for AS numbers. NOTE: AS numbers are
5231 queried only when Squid starts up, not for every request.
5236 LOC: Config.onoff.offline
5239 Enable this option and Squid will never try to validate cached
5243 NAME: uri_whitespace
5244 TYPE: uri_whitespace
5245 LOC: Config.uri_whitespace
5248 What to do with requests that have whitespace characters in the
5251 strip: The whitespace characters are stripped out of the URL.
5252 This is the behavior recommended by RFC2396.
5253 deny: The request is denied. The user receives an "Invalid
5255 allow: The request is allowed and the URI is not changed. The
5256 whitespace characters remain in the URI. Note the
5257 whitespace is passed to redirector processes if they
5259 encode: The request is allowed and the whitespace characters are
5260 encoded according to RFC1738. This could be considered
5261 a violation of the HTTP/1.1
5262 RFC because proxies are not allowed to rewrite URI's.
5263 chop: The request is allowed and the URI is chopped at the
5264 first whitespace. This might also be considered a
5270 LOC: Config.coredump_dir
5272 DEFAULT_IF_NONE: none
5274 By default Squid leaves core files in the directory from where
5275 it was started. If you set 'coredump_dir' to a directory
5276 that exists, Squid will chdir() to that directory at startup
5277 and coredump files will be left there.
5280 # Leave coredumps in the first cache dir
5281 coredump_dir @DEFAULT_SWAP_DIR@
5287 LOC: Config.chroot_dir
5290 Use this to have Squid do a chroot() while initializing. This
5291 also causes Squid to fully drop root privileges after
5292 initializing. This means, for example, if you use a HTTP
5293 port less than 1024 and try to reconfigure, you will may get an
5294 error saying that Squid can not open the port.
5297 NAME: balance_on_multiple_ip
5299 LOC: Config.onoff.balance_on_multiple_ip
5302 Some load balancing servers based on round robin DNS have been
5303 found not to preserve user session state across requests
5304 to different IP addresses.
5306 By default Squid rotates IP's per request. By disabling
5307 this directive only connection failure triggers rotation.
5310 NAME: pipeline_prefetch
5312 LOC: Config.onoff.pipeline_prefetch
5315 To boost the performance of pipelined requests to closer
5316 match that of a non-proxied environment Squid can try to fetch
5317 up to two requests in parallel from a pipeline.
5319 Defaults to off for bandwidth management and access logging
5323 NAME: high_response_time_warning
5326 LOC: Config.warnings.high_rptm
5329 If the one-minute median response time exceeds this value,
5330 Squid prints a WARNING with debug level 0 to get the
5331 administrators attention. The value is in milliseconds.
5334 NAME: high_page_fault_warning
5336 LOC: Config.warnings.high_pf
5339 If the one-minute average page fault rate exceeds this
5340 value, Squid prints a WARNING with debug level 0 to get
5341 the administrators attention. The value is in page faults
5345 NAME: high_memory_warning
5347 LOC: Config.warnings.high_memory
5350 If the memory usage (as determined by mallinfo) exceeds
5351 this amount, Squid prints a WARNING with debug level 0 to get
5352 the administrators attention.
5355 NAME: sleep_after_fork
5356 COMMENT: (microseconds)
5358 LOC: Config.sleep_after_fork
5361 When this is set to a non-zero value, the main Squid process
5362 sleeps the specified number of microseconds after a fork()
5363 system call. This sleep may help the situation where your
5364 system reports fork() failures due to lack of (virtual)
5365 memory. Note, however, if you have a lot of child
5366 processes, these sleep delays will add up and your
5367 Squid will not service requests for some amount of time
5368 until all the child processes have been started.
5369 On Windows value less then 1000 (1 milliseconds) are