2 # SQUID Web Proxy Cache http://www.squid-cache.org/
3 # ----------------------------------------------------------
5 # Squid is the result of efforts by numerous individuals from
6 # the Internet community; see the CONTRIBUTORS file for full
7 # details. Many organizations have provided support for Squid's
8 # development; see the SPONSORS file for full details. Squid is
9 # Copyrighted (C) 2000 by the Regents of the University of
10 # California; see the COPYRIGHT file for full details. Squid
11 # incorporates software developed and/or copyrighted by other
12 # sources; see the CREDITS file for full details.
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
31 ----------------------------
33 This is the default Squid configuration file. You may wish
34 to look at the Squid home page (http://www.squid-cache.org/)
35 for the FAQ and other documentation.
37 The default Squid config file shows what the defaults for
38 various options happen to be. If you don't need to change the
39 default, you shouldn't uncomment the line. Doing so may cause
40 run-time problems. In some cases "none" refers to no default
41 setting at all, while in other cases it refers to a valid
42 option - the comments for that keyword indicate if this is the
48 Configuration options can be included using the "include" directive.
49 Include takes a list of files to include. Quoting and wildcards is
54 include /path/to/included/file/squid.acl.config
56 Includes can be nested up to a hard-coded depth of 16 levels.
57 This arbitrary restriction is to prevent recursive include references
58 from causing Squid entering an infinite loop whilst trying to load
62 Conditional configuration
64 If-statements can be used to make configuration directives
68 ... regular configuration directives ...
70 ... regular configuration directives ...]
73 The else part is optional. The keywords "if", "else", and "endif"
74 must be typed on their own lines, as if they were regular
75 configuration directives.
77 These individual conditions types are supported:
80 Always evaluates to true.
82 Always evaluates to false.
84 Equality comparison of two integer numbers.
89 The following SMP-related preprocessor macros can be used.
91 ${process_name} expands to the current Squid process "name"
92 (e.g., squid1, squid2, or cache1).
94 ${process_number} expands to the current Squid process
95 identifier, which is an integer number (e.g., 1, 2, 3) unique
96 across all Squid processes.
100 OPTIONS FOR AUTHENTICATION
101 -----------------------------------------------------------------------------
109 This is used to define parameters for the various authentication
110 schemes supported by Squid.
112 format: auth_param scheme parameter [setting]
114 The order in which authentication schemes are presented to the client is
115 dependent on the order the scheme first appears in config file. IE
116 has a bug (it's not RFC 2617 compliant) in that it will use the basic
117 scheme if basic is the first entry presented, even if more secure
118 schemes are presented. For now use the order in the recommended
119 settings section below. If other browsers have difficulties (don't
120 recognize the schemes offered even if you are using basic) either
121 put basic first, or disable the other schemes (by commenting out their
124 Once an authentication scheme is fully configured, it can only be
125 shutdown by shutting squid down and restarting. Changes can be made on
126 the fly and activated with a reconfigure. I.E. You can change to a
127 different helper, but not unconfigure the helper completely.
129 Please note that while this directive defines how Squid processes
130 authentication it does not automatically activate authentication.
131 To use authentication you must in addition make use of ACLs based
132 on login name in http_access (proxy_auth, proxy_auth_regex or
133 external with %LOGIN used in the format tag). The browser will be
134 challenged for authentication on the first such acl encountered
135 in http_access processing and will also be re-challenged for new
136 login credentials if the request is being denied by a proxy_auth
139 WARNING: authentication can't be used in a transparently intercepting
140 proxy as the client then thinks it is talking to an origin server and
141 not the proxy. This is a limitation of bending the TCP/IP protocol to
142 transparently intercepting port 80, not a limitation in Squid.
143 Ports flagged 'transparent', 'intercept', or 'tproxy' have
144 authentication disabled.
146 === Parameters for the basic scheme follow. ===
149 Specify the command for the external authenticator. Such a program
150 reads a line containing "username password" and replies "OK" or
151 "ERR" in an endless loop. "ERR" responses may optionally be followed
152 by a error description available as %m in the returned error page.
153 If you use an authenticator, make sure you have 1 acl of type
156 By default, the basic authentication scheme is not used unless a
157 program is specified.
159 If you want to use the traditional NCSA proxy authentication, set
160 this line to something like
162 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
165 HTTP uses iso-latin-1 as characterset, while some authentication
166 backends such as LDAP expects UTF-8. If this is set to on Squid will
167 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
168 username & password to the helper.
170 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
171 The maximum number of authenticator processes to spawn. If you start too few
172 Squid will have to wait for them to process a backlog of credential
173 verifications, slowing it down. When password verifications are
174 done via a (slow) network you are likely to need lots of
175 authenticator processes.
177 The startup= and idle= options permit some skew in the exact amount
178 run. A minimum of startup=N will begin during startup and reconfigure
179 and Squid will start more in groups of up to idle=N in an attempt to meet
180 traffic needs and to keep idle=N free above those traffic needs up to
183 The concurrency= option sets the number of concurrent requests the
184 helper can process. The default of 0 is used for helpers who only
185 supports one request at a time. Setting this to a number greater than
186 0 changes the protocol used to include a channel number first on the
187 request/response line, allowing multiple requests to be sent to the
188 same helper in parallell without wating for the response.
189 Must not be set unless it's known the helper supports this.
191 auth_param basic children 20 startup=0 idle=1
194 Specifies the realm name which is to be reported to the
195 client for the basic proxy authentication scheme (part of
196 the text the user will see when prompted their username and
197 password). There is no default.
198 auth_param basic realm Squid proxy-caching web server
200 "credentialsttl" timetolive
201 Specifies how long squid assumes an externally validated
202 username:password pair is valid for - in other words how
203 often the helper program is called for that user. Set this
204 low to force revalidation with short lived passwords. Note
205 setting this high does not impact your susceptibility
206 to replay attacks unless you are using an one-time password
207 system (such as SecureID). If you are using such a system,
208 you will be vulnerable to replay attacks unless you also
209 use the max_user_ip ACL in an http_access rule.
211 "casesensitive" on|off
212 Specifies if usernames are case sensitive. Most user databases are
213 case insensitive allowing the same username to be spelled using both
214 lower and upper case letters, but some are case sensitive. This
215 makes a big difference for user_max_ip ACL processing and similar.
216 auth_param basic casesensitive off
218 === Parameters for the digest scheme follow ===
221 Specify the command for the external authenticator. Such
222 a program reads a line containing "username":"realm" and
223 replies with the appropriate H(A1) value hex encoded or
224 ERR if the user (or his H(A1) hash) does not exists.
225 See rfc 2616 for the definition of H(A1).
226 "ERR" responses may optionally be followed by a error description
227 available as %m in the returned error page.
229 By default, the digest authentication scheme is not used unless a
230 program is specified.
232 If you want to use a digest authenticator, set this line to
235 auth_param digest program @DEFAULT_PREFIX@/bin/digest_pw_auth @DEFAULT_PREFIX@/etc/digpass
238 HTTP uses iso-latin-1 as characterset, while some authentication
239 backends such as LDAP expects UTF-8. If this is set to on Squid will
240 translate the HTTP iso-latin-1 charset to UTF-8 before sending the
241 username & password to the helper.
243 "children" numberofchildren [startup=N] [idle=N] [concurrency=N]
244 The maximum number of authenticator processes to spawn (default 5).
245 If you start too few Squid will have to wait for them to
246 process a backlog of H(A1) calculations, slowing it down.
247 When the H(A1) calculations are done via a (slow) network
248 you are likely to need lots of authenticator processes.
250 The startup= and idle= options permit some skew in the exact amount
251 run. A minimum of startup=N will begin during startup and reconfigure
252 and Squid will start more in groups of up to idle=N in an attempt to meet
253 traffic needs and to keep idle=N free above those traffic needs up to
256 The concurrency= option sets the number of concurrent requests the
257 helper can process. The default of 0 is used for helpers who only
258 supports one request at a time. Setting this to a number greater than
259 0 changes the protocol used to include a channel number first on the
260 request/response line, allowing multiple requests to be sent to the
261 same helper in parallell without wating for the response.
262 Must not be set unless it's known the helper supports this.
264 auth_param digest children 20 startup=0 idle=1
267 Specifies the realm name which is to be reported to the
268 client for the digest proxy authentication scheme (part of
269 the text the user will see when prompted their username and
270 password). There is no default.
271 auth_param digest realm Squid proxy-caching web server
273 "nonce_garbage_interval" timeinterval
274 Specifies the interval that nonces that have been issued
275 to client_agent's are checked for validity.
277 "nonce_max_duration" timeinterval
278 Specifies the maximum length of time a given nonce will be
281 "nonce_max_count" number
282 Specifies the maximum number of times a given nonce can be
285 "nonce_strictness" on|off
286 Determines if squid requires strict increment-by-1 behavior
287 for nonce counts, or just incrementing (off - for use when
288 useragents generate nonce counts that occasionally miss 1
289 (ie, 1,2,4,6)). Default off.
291 "check_nonce_count" on|off
292 This directive if set to off can disable the nonce count check
293 completely to work around buggy digest qop implementations in
294 certain mainstream browser versions. Default on to check the
295 nonce count to protect from authentication replay attacks.
297 "post_workaround" on|off
298 This is a workaround to certain buggy browsers who sends
299 an incorrect request digest in POST requests when reusing
300 the same nonce as acquired earlier on a GET request.
302 === NTLM scheme options follow ===
305 Specify the command for the external NTLM authenticator.
306 Such a program reads exchanged NTLMSSP packets with
307 the browser via Squid until authentication is completed.
308 If you use an NTLM authenticator, make sure you have 1 acl
309 of type proxy_auth. By default, the NTLM authenticator_program
312 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
314 "children" numberofchildren [startup=N] [idle=N]
315 The maximum number of authenticator processes to spawn (default 5).
316 If you start too few Squid will have to wait for them to
317 process a backlog of credential verifications, slowing it
318 down. When credential verifications are done via a (slow)
319 network you are likely to need lots of authenticator
322 The startup= and idle= options permit some skew in the exact amount
323 run. A minimum of startup=N will begin during startup and reconfigure
324 and Squid will start more in groups of up to idle=N in an attempt to meet
325 traffic needs and to keep idle=N free above those traffic needs up to
328 auth_param ntlm children 20 startup=0 idle=1
331 If you experience problems with PUT/POST requests when using the
332 Negotiate authentication scheme then you can try setting this to
333 off. This will cause Squid to forcibly close the connection on
334 the initial requests where the browser asks which schemes are
335 supported by the proxy.
337 auth_param ntlm keep_alive on
339 === Options for configuring the NEGOTIATE auth-scheme follow ===
342 Specify the command for the external Negotiate authenticator.
343 This protocol is used in Microsoft Active-Directory enabled setups with
344 the Microsoft Internet Explorer or Mozilla Firefox browsers.
345 Its main purpose is to exchange credentials with the Squid proxy
346 using the Kerberos mechanisms.
347 If you use a Negotiate authenticator, make sure you have at least
348 one acl of type proxy_auth active. By default, the negotiate
349 authenticator_program is not used.
350 The only supported program for this role is the ntlm_auth
351 program distributed as part of Samba, version 4 or later.
353 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
355 "children" numberofchildren [startup=N] [idle=N]
356 The maximum number of authenticator processes to spawn (default 5).
357 If you start too few Squid will have to wait for them to
358 process a backlog of credential verifications, slowing it
359 down. When crendential verifications are done via a (slow)
360 network you are likely to need lots of authenticator
363 The startup= and idle= options permit some skew in the exact amount
364 run. A minimum of startup=N will begin during startup and reconfigure
365 and Squid will start more in groups of up to idle=N in an attempt to meet
366 traffic needs and to keep idle=N free above those traffic needs up to
369 auth_param negotiate children 20 startup=0 idle=1
372 If you experience problems with PUT/POST requests when using the
373 Negotiate authentication scheme then you can try setting this to
374 off. This will cause Squid to forcibly close the connection on
375 the initial requests where the browser asks which schemes are
376 supported by the proxy.
378 auth_param negotiate keep_alive on
383 #Recommended minimum configuration per scheme:
384 #auth_param negotiate program <uncomment and complete this line to activate>
385 #auth_param negotiate children 20 startup=0 idle=1
386 #auth_param negotiate keep_alive on
388 #auth_param ntlm program <uncomment and complete this line to activate>
389 #auth_param ntlm children 20 startup=0 idle=1
390 #auth_param ntlm keep_alive on
392 #auth_param digest program <uncomment and complete this line>
393 #auth_param digest children 20 startup=0 idle=1
394 #auth_param digest realm Squid proxy-caching web server
395 #auth_param digest nonce_garbage_interval 5 minutes
396 #auth_param digest nonce_max_duration 30 minutes
397 #auth_param digest nonce_max_count 50
399 #auth_param basic program <uncomment and complete this line>
400 #auth_param basic children 5 stratup=5 idle=1
401 #auth_param basic realm Squid proxy-caching web server
402 #auth_param basic credentialsttl 2 hours
405 NAME: authenticate_cache_garbage_interval
408 LOC: Config.authenticateGCInterval
410 The time period between garbage collection across the username cache.
411 This is a tradeoff between memory utilization (long intervals - say
412 2 days) and CPU (short intervals - say 1 minute). Only change if you
416 NAME: authenticate_ttl
419 LOC: Config.authenticateTTL
421 The time a user & their credentials stay in the logged in
422 user cache since their last request. When the garbage
423 interval passes, all user credentials that have passed their
424 TTL are removed from memory.
427 NAME: authenticate_ip_ttl
429 LOC: Config.authenticateIpTTL
432 If you use proxy authentication and the 'max_user_ip' ACL,
433 this directive controls how long Squid remembers the IP
434 addresses associated with each user. Use a small value
435 (e.g., 60 seconds) if your users might change addresses
436 quickly, as is the case with dialups. You might be safe
437 using a larger value (e.g., 2 hours) in a corporate LAN
438 environment with relatively static address assignments.
443 -----------------------------------------------------------------------------
446 NAME: external_acl_type
447 TYPE: externalAclHelper
448 LOC: Config.externalAclHelperList
451 This option defines external acl classes using a helper program
452 to look up the status
454 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
458 ttl=n TTL in seconds for cached results (defaults to 3600
461 TTL for cached negative lookups (default same
464 Maximum number of acl helper processes spawned to service
465 external acl lookups of this type. (default 20)
467 Minimum number of acl helper processes to spawn during
468 startup and reconfigure to service external acl lookups
469 of this type. (default 0)
471 Number of acl helper processes to keep ahead of traffic
472 loads. Squid will spawn this many at once whenever load
473 rises above the capabilities of existing processes.
474 Up to the value of children-max. (default 1)
475 concurrency=n concurrency level per process. Only used with helpers
476 capable of processing more than one query at a time.
477 cache=n limit the result cache size, default is unbounded.
478 grace=n Percentage remaining of TTL where a refresh of a
479 cached entry should be initiated without needing to
480 wait for a new reply. (default is for no grace period)
481 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
482 ipv4 / ipv6 IP-mode used to communicate to this helper.
483 For compatability with older configurations and helpers
484 the default is 'ipv4'.
486 FORMAT specifications
488 %LOGIN Authenticated user login name
489 %EXT_USER Username from external acl
490 %IDENT Ident user name
492 %SRCPORT Client source port
495 %PROTO Requested protocol
497 %PATH Requested URL path
498 %METHOD Request method
499 %MYADDR Squid interface address
500 %MYPORT Squid http_port number
501 %PATH Requested URL-path (including query-string if any)
502 %USER_CERT SSL User certificate in PEM format
503 %USER_CERTCHAIN SSL User certificate chain in PEM format
504 %USER_CERT_xx SSL User certificate subject attribute xx
505 %USER_CA_xx SSL User certificate issuer attribute xx
507 %>{Header} HTTP request header "Header"
509 HTTP request header "Hdr" list member "member"
511 HTTP request header list member using ; as
512 list separator. ; can be any non-alphanumeric
515 %<{Header} HTTP reply header "Header"
517 HTTP reply header "Hdr" list member "member"
519 HTTP reply header list member using ; as
520 list separator. ; can be any non-alphanumeric
523 In addition to the above, any string specified in the referencing
524 acl will also be included in the helper request line, after the
525 specified formats (see the "acl external" directive)
527 The helper receives lines per the above format specification,
528 and returns lines starting with OK or ERR indicating the validity
529 of the request and optionally followed by additional keywords with
532 General result syntax:
534 OK/ERR keyword=value ...
538 user= The users name (login)
539 password= The users password (for login= cache_peer option)
540 message= Message describing the reason. Available as %o
542 tag= Apply a tag to a request (for both ERR and OK results)
543 Only sets a tag, does not alter existing tags.
544 log= String to be logged in access.log. Available as
545 %ea in logformat specifications
547 If protocol=3.0 (the default) then URL escaping is used to protect
548 each value in both requests and responses.
550 If using protocol=2.5 then all values need to be enclosed in quotes
551 if they may contain whitespace, or the whitespace escaped using \.
552 And quotes or \ characters within the keyword value must be \ escaped.
554 When using the concurrency= option the protocol is changed by
555 introducing a query channel tag infront of the request/response.
556 The query channel tag is a number between 0 and concurrency-1.
564 Defining an Access List
566 Every access list definition must begin with an aclname and acltype,
567 followed by either type-specific arguments or a quoted filename that
570 acl aclname acltype argument ...
571 acl aclname acltype "file" ...
573 When using "file", the file should contain one item per line.
575 By default, regular expressions are CASE-SENSITIVE. To make
576 them case-insensitive, use the -i option.
578 Some acl types require suspending the current request in order
579 to access some external data source.
580 Those which do are marked with the tag [slow], those which
581 don't are marked as [fast].
582 See http://wiki.squid-cache.org/SquidFaq/SquidAcl
583 for further information
585 ***** ACL TYPES AVAILABLE *****
587 acl aclname src ip-address/netmask ... # clients IP address [fast]
588 acl aclname src addr1-addr2/netmask ... # range of addresses [fast]
589 acl aclname dst ip-address/netmask ... # URL host's IP address [slow]
590 acl aclname myip ip-address/netmask ... # local socket IP address [fast]
592 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
593 # The arp ACL requires the special configure option --enable-arp-acl.
594 # Furthermore, the ARP ACL code is not portable to all operating systems.
595 # It works on Linux, Solaris, Windows, FreeBSD, and some
596 # other *BSD variants.
599 # NOTE: Squid can only determine the MAC address for clients that are on
600 # the same subnet. If the client is on a different subnet,
601 # then Squid cannot find out its MAC address.
603 acl aclname srcdomain .foo.com ...
604 # reverse lookup, from client IP [slow]
605 acl aclname dstdomain .foo.com ...
606 # Destination server from URL [fast]
607 acl aclname srcdom_regex [-i] \.foo\.com ...
608 # regex matching client name [slow]
609 acl aclname dstdom_regex [-i] \.foo\.com ...
610 # regex matching server [fast]
612 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
613 # based URL is used and no match is found. The name "none" is used
614 # if the reverse lookup fails.
616 acl aclname src_as number ...
617 acl aclname dst_as number ...
619 # Except for access control, AS numbers can be used for
620 # routing of requests to specific caches. Here's an
621 # example for routing all requests for AS#1241 and only
622 # those to mycache.mydomain.net:
623 # acl asexample dst_as 1241
624 # cache_peer_access mycache.mydomain.net allow asexample
625 # cache_peer_access mycache_mydomain.net deny all
627 acl aclname peername myPeer ...
629 # match against a named cache_peer entry
630 # set unique name= on cache_peer lines for reliable use.
632 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
642 # h1:m1 must be less than h2:m2
644 acl aclname url_regex [-i] ^http:// ...
645 # regex matching on whole URL [fast]
646 acl aclname urlpath_regex [-i] \.gif$ ...
647 # regex matching on URL path [fast]
649 acl aclname port 80 70 21 0-1024... # destination TCP port [fast]
651 acl aclname myport 3128 ... # local socket TCP port [fast]
652 acl aclname myportname 3128 ... # http(s)_port name [fast]
654 acl aclname proto HTTP FTP ... # request protocol [fast]
656 acl aclname method GET POST ... # HTTP request method [fast]
658 acl aclname http_status 200 301 500- 400-403 ...
659 # status code in reply [fast]
661 acl aclname browser [-i] regexp ...
662 # pattern match on User-Agent header (see also req_header below) [fast]
664 acl aclname referer_regex [-i] regexp ...
665 # pattern match on Referer header [fast]
666 # Referer is highly unreliable, so use with care
668 acl aclname ident username ...
669 acl aclname ident_regex [-i] pattern ...
670 # string match on ident output [slow]
671 # use REQUIRED to accept any non-null ident.
673 acl aclname proxy_auth [-i] username ...
674 acl aclname proxy_auth_regex [-i] pattern ...
675 # perform http authentication challenge to the client and match against
676 # supplied credentials [slow]
678 # takes a list of allowed usernames.
679 # use REQUIRED to accept any valid username.
681 # Will use proxy authentication in forward-proxy scenarios, and plain
682 # http authenticaiton in reverse-proxy scenarios
684 # NOTE: when a Proxy-Authentication header is sent but it is not
685 # needed during ACL checking the username is NOT logged
688 # NOTE: proxy_auth requires a EXTERNAL authentication program
689 # to check username/password combinations (see
690 # auth_param directive).
692 # NOTE: proxy_auth can't be used in a transparent/intercepting proxy
693 # as the browser needs to be configured for using a proxy in order
694 # to respond to proxy authentication.
696 acl aclname snmp_community string ...
697 # A community string to limit access to your SNMP Agent [fast]
700 # acl snmppublic snmp_community public
702 acl aclname maxconn number
703 # This will be matched when the client's IP address has
704 # more than <number> HTTP connections established. [fast]
706 acl aclname max_user_ip [-s] number
707 # This will be matched when the user attempts to log in from more
708 # than <number> different ip addresses. The authenticate_ip_ttl
709 # parameter controls the timeout on the ip entries. [fast]
710 # If -s is specified the limit is strict, denying browsing
711 # from any further IP addresses until the ttl has expired. Without
712 # -s Squid will just annoy the user by "randomly" denying requests.
713 # (the counter is reset each time the limit is reached and a
715 # NOTE: in acceleration mode or where there is mesh of child proxies,
716 # clients may appear to come from multiple addresses if they are
717 # going through proxy farms, so a limit of 1 may cause user problems.
719 acl aclname random probability
720 # Pseudo-randomly match requests. Based on the probability given.
721 # Probability may be written as a decimal (0.333), fraction (1/3)
722 # or ratio of matches:non-matches (3:5).
724 acl aclname req_mime_type [-i] mime-type ...
725 # regex match against the mime type of the request generated
726 # by the client. Can be used to detect file upload or some
727 # types HTTP tunneling requests [fast]
728 # NOTE: This does NOT match the reply. You cannot use this
729 # to match the returned file type.
731 acl aclname req_header header-name [-i] any\.regex\.here
732 # regex match against any of the known request headers. May be
733 # thought of as a superset of "browser", "referer" and "mime-type"
736 acl aclname rep_mime_type [-i] mime-type ...
737 # regex match against the mime type of the reply received by
738 # squid. Can be used to detect file download or some
739 # types HTTP tunneling requests. [fast]
740 # NOTE: This has no effect in http_access rules. It only has
741 # effect in rules that affect the reply data stream such as
744 acl aclname rep_header header-name [-i] any\.regex\.here
745 # regex match against any of the known reply headers. May be
746 # thought of as a superset of "browser", "referer" and "mime-type"
749 acl aclname external class_name [arguments...]
750 # external ACL lookup via a helper class defined by the
751 # external_acl_type directive [slow]
753 acl aclname user_cert attribute values...
754 # match against attributes in a user SSL certificate
755 # attribute is one of DN/C/O/CN/L/ST [fast]
757 acl aclname ca_cert attribute values...
758 # match against attributes a users issuing CA SSL certificate
759 # attribute is one of DN/C/O/CN/L/ST [fast]
761 acl aclname ext_user username ...
762 acl aclname ext_user_regex [-i] pattern ...
763 # string match on username returned by external acl helper [slow]
764 # use REQUIRED to accept any non-null user name.
766 acl aclname tag tagvalue ...
767 # string match on tag returned by external acl helper [slow]
769 acl aclname hier_code codename ...
770 # string match against squid hierarchy code(s); [fast]
771 # e.g., DIRECT, PARENT_HIT, NONE, etc.
773 # NOTE: This has no effect in http_access rules. It only has
774 # effect in rules that affect the reply data stream such as
778 acl macaddress arp 09:00:2b:23:45:67
779 acl myexample dst_as 1241
780 acl password proxy_auth REQUIRED
781 acl fileupload req_mime_type -i ^multipart/form-data$
782 acl javascript rep_mime_type -i ^application/x-javascript$
786 # Recommended minimum configuration:
788 acl manager proto cache_object
789 acl localhost src 127.0.0.1/32
790 @IPV6_ONLY_SETTING@acl localhost src ::1/128
791 acl to_localhost dst 127.0.0.0/8 0.0.0.0/32
792 @IPV6_ONLY_SETTING@acl to_localhost dst ::1/128
794 # Example rule allowing access from your local networks.
795 # Adapt to list your (internal) IP networks from where browsing
797 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network
798 acl localnet src 172.16.0.0/12 # RFC1918 possible internal network
799 acl localnet src 192.168.0.0/16 # RFC1918 possible internal network
800 @IPV6_ONLY_SETTING@acl localnet src fc00::/7 # RFC 4193 local private network range
801 @IPV6_ONLY_SETTING@acl localnet src fe80::/10 # RFC 4291 link-local (directly plugged) machines
803 acl SSL_ports port 443
804 acl Safe_ports port 80 # http
805 acl Safe_ports port 21 # ftp
806 acl Safe_ports port 443 # https
807 acl Safe_ports port 70 # gopher
808 acl Safe_ports port 210 # wais
809 acl Safe_ports port 1025-65535 # unregistered ports
810 acl Safe_ports port 280 # http-mgmt
811 acl Safe_ports port 488 # gss-http
812 acl Safe_ports port 591 # filemaker
813 acl Safe_ports port 777 # multiling http
814 acl CONNECT method CONNECT
818 NAME: follow_x_forwarded_for
820 IFDEF: FOLLOW_X_FORWARDED_FOR
821 LOC: Config.accessList.followXFF
823 DEFAULT_IF_NONE: deny all
825 Allowing or Denying the X-Forwarded-For header to be followed to
826 find the original source of a request.
828 Requests may pass through a chain of several other proxies
829 before reaching us. The X-Forwarded-For header will contain a
830 comma-separated list of the IP addresses in the chain, with the
831 rightmost address being the most recent.
833 If a request reaches us from a source that is allowed by this
834 configuration item, then we consult the X-Forwarded-For header
835 to see where that host received the request from. If the
836 X-Forwarded-For header contains multiple addresses, we continue
837 backtracking until we reach an address for which we are not allowed
838 to follow the X-Forwarded-For header, or until we reach the first
839 address in the list. For the purpose of ACL used in the
840 follow_x_forwarded_for directive the src ACL type always matches
841 the address we are testing and srcdomain matches its rDNS.
843 The end result of this process is an IP address that we will
844 refer to as the indirect client address. This address may
845 be treated as the client address for access control, ICAP, delay
846 pools and logging, depending on the acl_uses_indirect_client,
847 icap_uses_indirect_client, delay_pool_uses_indirect_client,
848 log_uses_indirect_client and tproxy_uses_indirect_client options.
850 This clause only supports fast acl types.
851 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
853 SECURITY CONSIDERATIONS:
855 Any host for which we follow the X-Forwarded-For header
856 can place incorrect information in the header, and Squid
857 will use the incorrect information as if it were the
858 source address of the request. This may enable remote
859 hosts to bypass any access control restrictions that are
860 based on the client's source addresses.
864 acl localhost src 127.0.0.1
865 acl my_other_proxy srcdomain .proxy.example.com
866 follow_x_forwarded_for allow localhost
867 follow_x_forwarded_for allow my_other_proxy
870 NAME: acl_uses_indirect_client
873 IFDEF: FOLLOW_X_FORWARDED_FOR
875 LOC: Config.onoff.acl_uses_indirect_client
877 Controls whether the indirect client address
878 (see follow_x_forwarded_for) is used instead of the
879 direct client address in acl matching.
882 NAME: delay_pool_uses_indirect_client
885 IFDEF: FOLLOW_X_FORWARDED_FOR&&DELAY_POOLS
887 LOC: Config.onoff.delay_pool_uses_indirect_client
889 Controls whether the indirect client address
890 (see follow_x_forwarded_for) is used instead of the
891 direct client address in delay pools.
894 NAME: log_uses_indirect_client
897 IFDEF: FOLLOW_X_FORWARDED_FOR
899 LOC: Config.onoff.log_uses_indirect_client
901 Controls whether the indirect client address
902 (see follow_x_forwarded_for) is used instead of the
903 direct client address in the access log.
906 NAME: tproxy_uses_indirect_client
909 IFDEF: FOLLOW_X_FORWARDED_FOR&&LINUX_NETFILTER
911 LOC: Config.onoff.tproxy_uses_indirect_client
913 Controls whether the indirect client address
914 (see follow_x_forwarded_for) is used instead of the
915 direct client address when spoofing the outgoing client.
917 This has no effect on requests arriving in non-tproxy
920 SECURITY WARNING: Usage of this option is dangerous
921 and should not be used trivially. Correct configuration
922 of folow_x_forewarded_for with a limited set of trusted
923 sources is required to prevent abuse of your proxy.
928 LOC: Config.accessList.http
930 DEFAULT_IF_NONE: deny all
932 Allowing or Denying access based on defined access lists
934 Access to the HTTP port:
935 http_access allow|deny [!]aclname ...
937 NOTE on default values:
939 If there are no "access" lines present, the default is to deny
942 If none of the "access" lines cause a match, the default is the
943 opposite of the last line in the list. If the last line was
944 deny, the default is allow. Conversely, if the last line
945 is allow, the default will be deny. For these reasons, it is a
946 good idea to have an "deny all" entry at the end of your access
947 lists to avoid potential confusion.
949 This clause supports both fast and slow acl types.
950 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
955 # Recommended minimum Access Permission configuration:
957 # Only allow cachemgr access from localhost
958 http_access allow manager localhost
959 http_access deny manager
961 # Deny requests to certain unsafe ports
962 http_access deny !Safe_ports
964 # Deny CONNECT to other than secure SSL ports
965 http_access deny CONNECT !SSL_ports
967 # We strongly recommend the following be uncommented to protect innocent
968 # web applications running on the proxy server who think the only
969 # one who can access services on "localhost" is a local user
970 #http_access deny to_localhost
973 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
976 # Example rule allowing access from your local networks.
977 # Adapt localnet in the ACL section to list your (internal) IP networks
978 # from where browsing should be allowed
979 http_access allow localnet
980 http_access allow localhost
982 # And finally deny all other access to this proxy
987 NAME: adapted_http_access http_access2
989 LOC: Config.accessList.adapted_http
992 Allowing or Denying access based on defined access lists
994 Essentially identical to http_access, but runs after redirectors
995 and ICAP/eCAP adaptation. Allowing access control based on their
998 If not set then only http_access is used.
1001 NAME: http_reply_access
1003 LOC: Config.accessList.reply
1006 Allow replies to client requests. This is complementary to http_access.
1008 http_reply_access allow|deny [!] aclname ...
1010 NOTE: if there are no access lines present, the default is to allow
1013 If none of the access lines cause a match the opposite of the
1014 last line will apply. Thus it is good practice to end the rules
1015 with an "allow all" or "deny all" entry.
1017 This clause supports both fast and slow acl types.
1018 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1023 LOC: Config.accessList.icp
1025 DEFAULT_IF_NONE: deny all
1027 Allowing or Denying access to the ICP port based on defined
1030 icp_access allow|deny [!]aclname ...
1032 See http_access for details
1034 This clause only supports fast acl types.
1035 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1037 # Allow ICP queries from local networks only
1038 #icp_access allow localnet
1039 #icp_access deny all
1045 LOC: Config.accessList.htcp
1047 DEFAULT_IF_NONE: deny all
1049 Allowing or Denying access to the HTCP port based on defined
1052 htcp_access allow|deny [!]aclname ...
1054 See http_access for details
1056 NOTE: The default if no htcp_access lines are present is to
1057 deny all traffic. This default may cause problems with peers
1058 using the htcp or htcp-oldsquid options.
1060 This clause only supports fast acl types.
1061 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1063 # Allow HTCP queries from local networks only
1064 #htcp_access allow localnet
1065 #htcp_access deny all
1068 NAME: htcp_clr_access
1071 LOC: Config.accessList.htcp_clr
1073 DEFAULT_IF_NONE: deny all
1075 Allowing or Denying access to purge content using HTCP based
1076 on defined access lists
1078 htcp_clr_access allow|deny [!]aclname ...
1080 See http_access for details
1082 This clause only supports fast acl types.
1083 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1085 # Allow HTCP CLR requests from trusted peers
1086 acl htcp_clr_peer src 172.16.1.2
1087 htcp_clr_access allow htcp_clr_peer
1092 LOC: Config.accessList.miss
1095 Use to force your neighbors to use you as a sibling instead of
1096 a parent. For example:
1098 acl localclients src 172.16.0.0/16
1099 miss_access allow localclients
1100 miss_access deny !localclients
1102 This means only your local clients are allowed to fetch
1103 MISSES and all other clients can only fetch HITS.
1105 By default, allow all clients who passed the http_access rules
1106 to fetch MISSES from us.
1108 This clause only supports fast acl types.
1109 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1112 NAME: ident_lookup_access
1116 DEFAULT_IF_NONE: deny all
1117 LOC: Ident::TheConfig.identLookup
1119 A list of ACL elements which, if matched, cause an ident
1120 (RFC 931) lookup to be performed for this request. For
1121 example, you might choose to always perform ident lookups
1122 for your main multi-user Unix boxes, but not for your Macs
1123 and PCs. By default, ident lookups are not performed for
1126 To enable ident lookups for specific client addresses, you
1127 can follow this example:
1129 acl ident_aware_hosts src 198.168.1.0/24
1130 ident_lookup_access allow ident_aware_hosts
1131 ident_lookup_access deny all
1133 Only src type ACL checks are fully supported. A srcdomain
1134 ACL might work at times, but it will not always provide
1137 This clause only supports fast acl types.
1138 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1141 NAME: reply_body_max_size
1142 COMMENT: size [acl acl...]
1145 LOC: Config.ReplyBodySize
1147 This option specifies the maximum size of a reply body. It can be
1148 used to prevent users from downloading very large files, such as
1149 MP3's and movies. When the reply headers are received, the
1150 reply_body_max_size lines are processed, and the first line where
1151 all (if any) listed ACLs are true is used as the maximum body size
1154 This size is checked twice. First when we get the reply headers,
1155 we check the content-length value. If the content length value exists
1156 and is larger than the allowed size, the request is denied and the
1157 user receives an error message that says "the request or reply
1158 is too large." If there is no content-length, and the reply
1159 size exceeds this limit, the client's connection is just closed
1160 and they will receive a partial reply.
1162 WARNING: downstream caches probably can not detect a partial reply
1163 if there is no content-length header, so they will cache
1164 partial responses and give them out as hits. You should NOT
1165 use this option if you have downstream caches.
1167 WARNING: A maximum size smaller than the size of squid's error messages
1168 will cause an infinite loop and crash squid. Ensure that the smallest
1169 non-zero value you use is greater that the maximum header size plus
1170 the size of your largest error page.
1172 If you set this parameter none (the default), there will be
1175 Configuration Format is:
1176 reply_body_max_size SIZE UNITS [acl ...]
1178 reply_body_max_size 10 MB
1184 -----------------------------------------------------------------------------
1187 NAME: http_port ascii_port
1188 TYPE: http_port_list
1190 LOC: Config.Sockaddr.http
1192 Usage: port [mode] [options]
1193 hostname:port [mode] [options]
1194 1.2.3.4:port [mode] [options]
1196 The socket addresses where Squid will listen for HTTP client
1197 requests. You may specify multiple socket addresses.
1198 There are three forms: port alone, hostname with port, and
1199 IP address with port. If you specify a hostname or IP
1200 address, Squid binds the socket to that specific
1201 address. Most likely, you do not need to bind to a specific
1202 address, so you can use the port number alone.
1204 If you are running Squid in accelerator mode, you
1205 probably want to listen on port 80 also, or instead.
1207 The -a command line option may be used to specify additional
1208 port(s) where Squid listens for proxy request. Such ports will
1209 be plain proxy ports with no options.
1211 You may specify multiple socket addresses on multiple lines.
1215 intercept Support for IP-Layer interception of
1216 outgoing requests without browser settings.
1217 NP: disables authentication and IPv6 on the port.
1219 tproxy Support Linux TPROXY for spoofing outgoing
1220 connections using the client IP address.
1221 NP: disables authentication and maybe IPv6 on the port.
1223 accel Accelerator mode. Also needs at least one of
1224 vhost / vport / defaultsite.
1226 sslbump Intercept each CONNECT request matching ssl_bump ACL,
1227 establish secure connection with the client and with
1228 the server, decrypt HTTP messages as they pass through
1229 Squid, and treat them as unencrypted HTTP messages,
1230 becoming the man-in-the-middle.
1232 The ssl_bump option is required to fully enable
1233 the SslBump feature.
1235 Omitting the mode flag causes default forward proxy mode to be used.
1238 Accelerator Mode Options:
1240 allow-direct Allow direct forwarding in accelerator mode. Normally
1241 accelerated requests are denied direct forwarding as if
1242 never_direct was used.
1244 defaultsite=domainname
1245 What to use for the Host: header if it is not present
1246 in a request. Determines what site (not origin server)
1247 accelerators should consider the default.
1250 vhost Using the Host header for virtual domain support.
1251 Also uses the port as specified in Host: header.
1253 vport IP based virtual host support. Using the http_port number
1254 in passed on Host: headers.
1256 vport=NN Uses the specified port number rather than the
1259 protocol= Protocol to reconstruct accelerated requests with.
1260 Defaults to http://.
1262 ignore-cc Ignore request Cache-Control headers.
1264 Warning: This option violates HTTP specifications if
1265 used in non-accelerator setups.
1268 SSL Bump Mode Options:
1270 cert= Path to SSL certificate (PEM format).
1272 key= Path to SSL private key file (PEM format)
1273 if not specified, the certificate file is
1274 assumed to be a combined certificate and
1277 version= The version of SSL/TLS supported
1278 1 automatic (default)
1283 cipher= Colon separated list of supported ciphers.
1285 options= Various SSL engine options. The most important
1287 NO_SSLv2 Disallow the use of SSLv2
1288 NO_SSLv3 Disallow the use of SSLv3
1289 NO_TLSv1 Disallow the use of TLSv1
1290 SINGLE_DH_USE Always create a new key when using
1291 temporary/ephemeral DH key exchanges
1292 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1293 documentation for a complete list of options.
1295 clientca= File containing the list of CAs to use when
1296 requesting a client certificate.
1298 cafile= File containing additional CA certificates to
1299 use when verifying client certificates. If unset
1300 clientca will be used.
1302 capath= Directory containing additional CA certificates
1303 and CRL lists to use when verifying client certificates.
1305 crlfile= File of additional CRL lists to use when verifying
1306 the client certificate, in addition to CRLs stored in
1307 the capath. Implies VERIFY_CRL flag below.
1309 dhparams= File containing DH parameters for temporary/ephemeral
1312 sslflags= Various flags modifying the use of SSL:
1314 Don't request client certificates
1315 immediately, but wait until acl processing
1316 requires a certificate (not yet implemented).
1318 Don't use the default CA lists built in
1321 Don't allow for session reuse. Each connection
1322 will result in a new SSL session.
1324 Verify CRL lists when accepting client
1327 Verify CRL lists for all certificates in the
1328 client certificate chain.
1330 sslcontext= SSL session ID context identifier.
1335 connection-auth[=on|off]
1336 use connection-auth=off to tell Squid to prevent
1337 forwarding Microsoft connection oriented authentication
1338 (NTLM, Negotiate and Kerberos)
1340 disable-pmtu-discovery=
1341 Control Path-MTU discovery usage:
1342 off lets OS decide on what to do (default).
1343 transparent disable PMTU discovery when transparent
1345 always disable always PMTU discovery.
1347 In many setups of transparently intercepting proxies
1348 Path-MTU discovery can not work on traffic towards the
1349 clients. This is the case when the intercepting device
1350 does not fully track connections and fails to forward
1351 ICMP must fragment messages to the cache server. If you
1352 have such setup and experience that certain clients
1353 sporadically hang or never complete requests set
1354 disable-pmtu-discovery option to 'transparent'.
1356 name= Specifies a internal name for the port. Defaults to
1357 the port specification (port or addr:port)
1359 tcpkeepalive[=idle,interval,timeout]
1360 Enable TCP keepalive probes of idle connections
1361 idle is the initial time before TCP starts probing
1362 the connection, interval how often to probe, and
1363 timeout the time before giving up.
1365 If you run Squid on a dual-homed machine with an internal
1366 and an external interface we recommend you to specify the
1367 internal address:port in http_port. This way Squid will only be
1368 visible on the internal address.
1372 # Squid normally listens to port 3128
1373 http_port @DEFAULT_HTTP_PORT@
1379 TYPE: https_port_list
1381 LOC: Config.Sockaddr.https
1383 Usage: [ip:]port cert=certificate.pem [key=key.pem] [options...]
1385 The socket address where Squid will listen for HTTPS client
1388 This is really only useful for situations where you are running
1389 squid in accelerator mode and you want to do the SSL work at the
1392 You may specify multiple socket addresses on multiple lines,
1393 each with their own SSL certificate and/or options.
1397 accel Accelerator mode. Also needs at least one of
1398 defaultsite or vhost.
1400 defaultsite= The name of the https site presented on
1401 this port. Implies accel.
1403 vhost Accelerator mode using Host header for virtual
1404 domain support. Requires a wildcard certificate
1405 or other certificate valid for more than one domain.
1408 protocol= Protocol to reconstruct accelerated requests with.
1411 cert= Path to SSL certificate (PEM format).
1413 key= Path to SSL private key file (PEM format)
1414 if not specified, the certificate file is
1415 assumed to be a combined certificate and
1418 version= The version of SSL/TLS supported
1419 1 automatic (default)
1424 cipher= Colon separated list of supported ciphers.
1426 options= Various SSL engine options. The most important
1428 NO_SSLv2 Disallow the use of SSLv2
1429 NO_SSLv3 Disallow the use of SSLv3
1430 NO_TLSv1 Disallow the use of TLSv1
1431 SINGLE_DH_USE Always create a new key when using
1432 temporary/ephemeral DH key exchanges
1433 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
1434 documentation for a complete list of options.
1436 clientca= File containing the list of CAs to use when
1437 requesting a client certificate.
1439 cafile= File containing additional CA certificates to
1440 use when verifying client certificates. If unset
1441 clientca will be used.
1443 capath= Directory containing additional CA certificates
1444 and CRL lists to use when verifying client certificates.
1446 crlfile= File of additional CRL lists to use when verifying
1447 the client certificate, in addition to CRLs stored in
1448 the capath. Implies VERIFY_CRL flag below.
1450 dhparams= File containing DH parameters for temporary/ephemeral
1453 sslflags= Various flags modifying the use of SSL:
1455 Don't request client certificates
1456 immediately, but wait until acl processing
1457 requires a certificate (not yet implemented).
1459 Don't use the default CA lists built in
1462 Don't allow for session reuse. Each connection
1463 will result in a new SSL session.
1465 Verify CRL lists when accepting client
1468 Verify CRL lists for all certificates in the
1469 client certificate chain.
1471 sslcontext= SSL session ID context identifier.
1473 vport Accelerator with IP based virtual host support.
1475 vport=NN As above, but uses specified port number rather
1476 than the https_port number. Implies accel.
1478 name= Specifies a internal name for the port. Defaults to
1479 the port specification (port or addr:port)
1483 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
1486 LOC: Config.accessList.outgoing_tos
1488 Allows you to select a TOS/Diffserv value to mark outgoing
1489 connections with, based on the username or source address
1492 tcp_outgoing_tos ds-field [!]aclname ...
1494 Example where normal_service_net uses the TOS value 0x00
1495 and good_service_net uses 0x20
1497 acl normal_service_net src 10.0.0.0/255.255.255.0
1498 acl good_service_net src 10.0.1.0/255.255.255.0
1499 tcp_outgoing_tos 0x00 normal_service_net
1500 tcp_outgoing_tos 0x20 good_service_net
1502 TOS/DSCP values really only have local significance - so you should
1503 know what you're specifying. For more information, see RFC2474,
1504 RFC2475, and RFC3260.
1506 The TOS/DSCP byte must be exactly that - a octet value 0 - 255, or
1507 "default" to use whatever default your host has. Note that in
1508 practice often only values 0 - 63 is usable as the two highest bits
1509 have been redefined for use by ECN (RFC3168).
1511 Processing proceeds in the order specified, and stops at first fully
1514 Note: The use of this directive using client dependent ACLs is
1515 incompatible with the use of server side persistent connections. To
1516 ensure correct results it is best to set server_persisten_connections
1517 to off when using this directive in such configurations.
1520 NAME: clientside_tos
1523 LOC: Config.accessList.clientside_tos
1525 Allows you to select a TOS/Diffserv value to mark client-side
1526 connections with, based on the username or source address
1534 LOC: Ip::Qos::TheConfig
1536 Allows you to select a TOS/DSCP value to mark outgoing
1537 connections with, based on where the reply was sourced.
1539 TOS values really only have local significance - so you should
1540 know what you're specifying. For more information, see RFC2474,
1541 RFC2475, and RFC3260.
1543 The TOS/DSCP byte must be exactly that - octet value 0x00-0xFF.
1544 Note that in practice often only values up to 0x3F are usable
1545 as the two highest bits have been redefined for use by ECN
1548 This setting is configured by setting the source TOS values:
1550 local-hit=0xFF Value to mark local cache hits.
1552 sibling-hit=0xFF Value to mark hits from sibling peers.
1554 parent-hit=0xFF Value to mark hits from parent peers.
1557 NOTE: 'miss' preserve feature is only possible on Linux at this time.
1559 For the following to work correctly, you will need to patch your
1560 linux kernel with the TOS preserving ZPH patch.
1561 The kernel patch can be downloaded from http://zph.bratcheda.org
1563 disable-preserve-miss
1564 If set, any HTTP response towards clients will
1565 have the TOS value of the response comming from the
1566 remote server masked with the value of miss-mask.
1569 Allows you to mask certain bits in the TOS received from the
1570 remote server, before copying the value to the TOS sent
1572 Default: 0xFF (TOS from server is not changed).
1576 NAME: tcp_outgoing_address
1579 LOC: Config.accessList.outgoing_address
1581 Allows you to map requests to different outgoing IP addresses
1582 based on the username or source address of the user making
1585 tcp_outgoing_address ipaddr [[!]aclname] ...
1587 Example where requests from 10.0.0.0/24 will be forwarded
1588 with source address 10.1.0.1, 10.0.2.0/24 forwarded with
1589 source address 10.1.0.2 and the rest will be forwarded with
1590 source address 10.1.0.3.
1592 acl normal_service_net src 10.0.0.0/24
1593 acl good_service_net src 10.0.2.0/24
1594 tcp_outgoing_address 10.1.0.1 normal_service_net
1595 tcp_outgoing_address 10.1.0.2 good_service_net
1596 tcp_outgoing_address 10.1.0.3
1598 Processing proceeds in the order specified, and stops at first fully
1601 Note: The use of this directive using client dependent ACLs is
1602 incompatible with the use of server side persistent connections. To
1603 ensure correct results it is best to set server_persistent_connections
1604 to off when using this directive in such configurations.
1606 Note: The use of this directive to set a local IP on outgoing TCP links
1607 is incompatible with using TPROXY to set client IP out outbound TCP links.
1608 When needing to contact peers use the no-tproxy cache_peer option to
1609 re-enable normal forwarding such as this.
1613 Squid is built with a capability of bridging the IPv4 and IPv6
1615 tcp_outgoing_address as exampled above breaks this bridging by forcing
1616 all outbound traffic through a certain IPv4 which may be on the wrong
1617 side of the IPv4/IPv6 boundary.
1619 To operate with tcp_outgoing_address and keep the bridging benefits
1620 an additional ACL needs to be used which ensures the IPv6-bound traffic
1621 is never forced or permitted out the IPv4 interface.
1623 acl to_ipv6 dst ipv6
1624 tcp_outgoing_address 2002::c001 good_service_net to_ipv6
1625 tcp_outgoing_address 10.1.0.2 good_service_net !to_ipv6
1627 tcp_outgoing_address 2002::beef normal_service_net to_ipv6
1628 tcp_outgoing_address 10.1.0.1 normal_service_net !to_ipv6
1630 tcp_outgoing_address 2002::1 to_ipv6
1631 tcp_outgoing_address 10.1.0.3 !to_ipv6
1634 'dst ipv6' bases its selection assuming DIRECT access.
1635 If peers are used the peername ACL are needed to select outgoing
1636 address which can link to the peer.
1638 'dst ipv6' is a slow ACL. It will only work here if 'dst' is used
1639 previously in the http_access rules to locate the destination IP.
1640 Some more magic may be needed for that:
1641 http_access allow to_ipv6 !all
1642 (meaning, allow if to IPv6 but not from anywhere ;)
1648 -----------------------------------------------------------------------------
1651 NAME: ssl_unclean_shutdown
1655 LOC: Config.SSL.unclean_shutdown
1657 Some browsers (especially MSIE) bugs out on SSL shutdown
1664 LOC: Config.SSL.ssl_engine
1667 The OpenSSL engine to use. You will need to set this if you
1668 would like to use hardware SSL acceleration for example.
1671 NAME: sslproxy_client_certificate
1674 LOC: Config.ssl_client.cert
1677 Client SSL Certificate to use when proxying https:// URLs
1680 NAME: sslproxy_client_key
1683 LOC: Config.ssl_client.key
1686 Client SSL Key to use when proxying https:// URLs
1689 NAME: sslproxy_version
1692 LOC: Config.ssl_client.version
1695 SSL version level to use when proxying https:// URLs
1698 NAME: sslproxy_options
1701 LOC: Config.ssl_client.options
1704 SSL engine options to use when proxying https:// URLs
1706 The most important being:
1708 NO_SSLv2 Disallow the use of SSLv2
1709 NO_SSLv3 Disallow the use of SSLv3
1710 NO_TLSv1 Disallow the use of TLSv1
1712 Always create a new key when using
1713 temporary/ephemeral DH key exchanges
1715 These options vary depending on your SSL engine.
1716 See the OpenSSL SSL_CTX_set_options documentation for a
1717 complete list of possible options.
1720 NAME: sslproxy_cipher
1723 LOC: Config.ssl_client.cipher
1726 SSL cipher list to use when proxying https:// URLs
1728 Colon separated list of supported ciphers.
1731 NAME: sslproxy_cafile
1734 LOC: Config.ssl_client.cafile
1737 file containing CA certificates to use when verifying server
1738 certificates while proxying https:// URLs
1741 NAME: sslproxy_capath
1744 LOC: Config.ssl_client.capath
1747 directory containing CA certificates to use when verifying
1748 server certificates while proxying https:// URLs
1754 LOC: Config.accessList.ssl_bump
1757 This ACL controls which CONNECT requests to an http_port
1758 marked with an sslBump flag are actually "bumped". Please
1759 see the sslBump flag of an http_port option for more details
1760 about decoding proxied SSL connections.
1762 By default, no requests are bumped.
1764 See also: http_port sslBump
1766 This clause only supports fast acl types.
1767 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1770 # Example: Bump all requests except those originating from localhost and
1771 # those going to webax.com or example.com sites.
1773 acl localhost src 127.0.0.1/32
1774 acl broken_sites dstdomain .webax.com
1775 acl broken_sites dstdomain .example.com
1776 ssl_bump deny localhost
1777 ssl_bump deny broken_sites
1781 NAME: sslproxy_flags
1784 LOC: Config.ssl_client.flags
1787 Various flags modifying the use of SSL while proxying https:// URLs:
1788 DONT_VERIFY_PEER Accept certificates that fail verification.
1789 For refined control, see sslproxy_cert_error.
1790 NO_DEFAULT_CA Don't use the default CA list built in
1795 NAME: sslproxy_cert_error
1798 LOC: Config.ssl_client.cert_error
1801 Use this ACL to bypass server certificate validation errors.
1803 For example, the following lines will bypass all validation errors
1804 when talking to servers located at 172.16.0.0/16. All other
1805 validation errors will result in ERR_SECURE_CONNECT_FAIL error.
1807 acl BrokenServersAtTrustedIP dst 172.16.0.0/16
1808 sslproxy_cert_error allow BrokenServersAtTrustedIP
1809 sslproxy_cert_error deny all
1811 This clause only supports fast acl types.
1812 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
1813 Using slow acl types may result in server crashes
1815 Without this option, all server certificate validation errors
1816 terminate the transaction. Bypassing validation errors is dangerous
1817 because an error usually implies that the server cannot be trusted and
1818 the connection may be insecure.
1820 See also: sslproxy_flags and DONT_VERIFY_PEER.
1822 Default setting: sslproxy_cert_error deny all
1827 NAME: sslpassword_program
1830 LOC: Config.Program.ssl_password
1833 Specify a program used for entering SSL key passphrases
1834 when using encrypted SSL certificate keys. If not specified
1835 keys must either be unencrypted, or Squid started with the -N
1836 option to allow it to query interactively for the passphrase.
1840 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
1841 -----------------------------------------------------------------------------
1849 To specify other caches in a hierarchy, use the format:
1851 cache_peer hostname type http-port icp-port [options]
1856 # hostname type port port options
1857 # -------------------- -------- ----- ----- -----------
1858 cache_peer parent.foo.net parent 3128 3130 default
1859 cache_peer sib1.foo.net sibling 3128 3130 proxy-only
1860 cache_peer sib2.foo.net sibling 3128 3130 proxy-only
1861 cache_peer example.com parent 80 0 no-query default
1862 cache_peer cdn.example.com sibling 3128 0
1864 type: either 'parent', 'sibling', or 'multicast'.
1866 proxy-port: The port number where the peer accept HTTP requests.
1867 For other Squid proxies this is usually 3128
1868 For web servers this is usually 80
1870 icp-port: Used for querying neighbor caches about objects.
1871 Set to 0 if the peer does not support ICP or HTCP.
1872 See ICP and HTCP options below for additional details.
1875 ==== ICP OPTIONS ====
1877 You MUST also set icp_port and icp_access explicitly when using these options.
1878 The defaults will prevent peer traffic using ICP.
1881 no-query Disable ICP queries to this neighbor.
1884 Indicates the named peer is a member of a multicast group.
1885 ICP queries will not be sent directly to the peer, but ICP
1886 replies will be accepted from it.
1888 closest-only Indicates that, for ICP_OP_MISS replies, we'll only forward
1889 CLOSEST_PARENT_MISSes and never FIRST_PARENT_MISSes.
1892 To only send ICP queries to this neighbor infrequently.
1893 This is used to keep the neighbor round trip time updated
1894 and is usually used in conjunction with weighted-round-robin.
1897 ==== HTCP OPTIONS ====
1899 You MUST also set htcp_port and htcp_access explicitly when using these options.
1900 The defaults will prevent peer traffic using HTCP.
1903 htcp Send HTCP, instead of ICP, queries to the neighbor.
1904 You probably also want to set the "icp-port" to 4827
1907 htcp-oldsquid Send HTCP to old Squid versions.
1909 htcp-no-clr Send HTCP to the neighbor but without
1910 sending any CLR requests. This cannot be used with
1913 htcp-only-clr Send HTCP to the neighbor but ONLY CLR requests.
1914 This cannot be used with htcp-no-clr.
1917 Send HTCP to the neighbor including CLRs but only when
1918 they do not result from PURGE requests.
1921 Forward any HTCP CLR requests this proxy receives to the peer.
1924 ==== PEER SELECTION METHODS ====
1926 The default peer selection method is ICP, with the first responding peer
1927 being used as source. These options can be used for better load balancing.
1930 default This is a parent cache which can be used as a "last-resort"
1931 if a peer cannot be located by any of the peer-selection methods.
1932 If specified more than once, only the first is used.
1934 round-robin Load-Balance parents which should be used in a round-robin
1935 fashion in the absence of any ICP queries.
1936 weight=N can be used to add bias.
1938 weighted-round-robin
1939 Load-Balance parents which should be used in a round-robin
1940 fashion with the frequency of each parent being based on the
1941 round trip time. Closer parents are used more often.
1942 Usually used for background-ping parents.
1943 weight=N can be used to add bias.
1945 carp Load-Balance parents which should be used as a CARP array.
1946 The requests will be distributed among the parents based on the
1947 CARP load balancing hash function based on their weight.
1949 userhash Load-balance parents based on the client proxy_auth or ident username.
1951 sourcehash Load-balance parents based on the client source IP.
1954 To be used only for cache peers of type "multicast".
1955 ALL members of this multicast group have "sibling"
1956 relationship with it, not "parent". This is to a mulicast
1957 group when the requested object would be fetched only from
1958 a "parent" cache, anyway. It's useful, e.g., when
1959 configuring a pool of redundant Squid proxies, being
1960 members of the same multicast group.
1963 ==== PEER SELECTION OPTIONS ====
1965 weight=N use to affect the selection of a peer during any weighted
1966 peer-selection mechanisms.
1967 The weight must be an integer; default is 1,
1968 larger weights are favored more.
1969 This option does not affect parent selection if a peering
1970 protocol is not in use.
1972 basetime=N Specify a base amount to be subtracted from round trip
1974 It is subtracted before division by weight in calculating
1975 which parent to fectch from. If the rtt is less than the
1976 base time the rtt is set to a minimal value.
1978 ttl=N Specify a IP multicast TTL to use when sending an ICP
1979 queries to this address.
1980 Only useful when sending to a multicast group.
1981 Because we don't accept ICP replies from random
1982 hosts, you must configure other group members as
1983 peers with the 'multicast-responder' option.
1985 no-delay To prevent access to this neighbor from influencing the
1988 digest-url=URL Tell Squid to fetch the cache digest (if digests are
1989 enabled) for this host from the specified URL rather
1990 than the Squid default location.
1993 ==== ACCELERATOR / REVERSE-PROXY OPTIONS ====
1995 originserver Causes this parent to be contacted as an origin server.
1996 Meant to be used in accelerator setups when the peer
2000 Set the Host header of requests forwarded to this peer.
2001 Useful in accelerator setups where the server (peer)
2002 expects a certain domain name but clients may request
2003 others. ie example.com or www.example.com
2005 no-digest Disable request of cache digests.
2008 Disables requesting ICMP RTT database (NetDB).
2011 ==== AUTHENTICATION OPTIONS ====
2014 If this is a personal/workgroup proxy and your parent
2015 requires proxy authentication.
2017 Note: The string can include URL escapes (i.e. %20 for
2018 spaces). This also means % must be written as %%.
2021 Send login details received from client to this peer.
2022 Both Proxy- and WWW-Authorization headers are passed
2023 without alteration to the peer.
2024 Authentication is not required by Squid for this to work.
2026 Note: This will pass any form of authentication but
2027 only Basic auth will work through a proxy unless the
2028 connection-auth options are also used.
2030 login=PASS Send login details received from client to this peer.
2031 Authentication is not required by this option.
2033 If there are no client-provided authentication headers
2034 to pass on, but username and password are available
2035 from an external ACL user= and password= result tags
2036 they may be sent instead.
2038 Note: To combine this with proxy_auth both proxies must
2039 share the same user database as HTTP only allows for
2040 a single login (one for proxy, one for origin server).
2041 Also be warned this will expose your users proxy
2042 password to the peer. USE WITH CAUTION
2045 Send the username to the upstream cache, but with a
2046 fixed password. This is meant to be used when the peer
2047 is in another administrative domain, but it is still
2048 needed to identify each user.
2049 The star can optionally be followed by some extra
2050 information which is added to the username. This can
2051 be used to identify this proxy to the peer, similar to
2052 the login=username:password option above.
2055 If this is a personal/workgroup proxy and your parent
2056 requires a secure proxy authentication.
2057 The first principal from the default keytab or defined by
2058 the environment variable KRB5_KTNAME will be used.
2060 login=NEGOTIATE:principal_name
2061 If this is a personal/workgroup proxy and your parent
2062 requires a secure proxy authentication.
2063 The principal principal_name from the default keytab or
2064 defined by the environment variable KRB5_KTNAME will be
2067 connection-auth=on|off
2068 Tell Squid that this peer does or not support Microsoft
2069 connection oriented authentication, and any such
2070 challenges received from there should be ignored.
2071 Default is auto to automatically determine the status
2075 ==== SSL / HTTPS / TLS OPTIONS ====
2077 ssl Encrypt connections to this peer with SSL/TLS.
2079 sslcert=/path/to/ssl/certificate
2080 A client SSL certificate to use when connecting to
2083 sslkey=/path/to/ssl/key
2084 The private SSL key corresponding to sslcert above.
2085 If 'sslkey' is not specified 'sslcert' is assumed to
2086 reference a combined file containing both the
2087 certificate and the key.
2090 The SSL version to use when connecting to this peer
2091 1 = automatic (default)
2096 sslcipher=... The list of valid SSL ciphers to use when connecting
2099 ssloptions=... Specify various SSL engine options:
2100 NO_SSLv2 Disallow the use of SSLv2
2101 NO_SSLv3 Disallow the use of SSLv3
2102 NO_TLSv1 Disallow the use of TLSv1
2103 See src/ssl_support.c or the OpenSSL documentation for
2104 a more complete list.
2106 sslcafile=... A file containing additional CA certificates to use
2107 when verifying the peer certificate.
2109 sslcapath=... A directory containing additional CA certificates to
2110 use when verifying the peer certificate.
2112 sslcrlfile=... A certificate revocation list file to use when
2113 verifying the peer certificate.
2115 sslflags=... Specify various flags modifying the SSL implementation:
2118 Accept certificates even if they fail to
2121 Don't use the default CA list built in
2124 Don't verify the peer certificate
2125 matches the server name
2127 ssldomain= The peer name as advertised in it's certificate.
2128 Used for verifying the correctness of the received peer
2129 certificate. If not specified the peer hostname will be
2133 Enable the "Front-End-Https: On" header needed when
2134 using Squid as a SSL frontend in front of Microsoft OWA.
2135 See MS KB document Q307347 for details on this header.
2136 If set to auto the header will only be added if the
2137 request is forwarded as a https:// URL.
2140 ==== GENERAL OPTIONS ====
2143 A peer-specific connect timeout.
2144 Also see the peer_connect_timeout directive.
2146 connect-fail-limit=N
2147 How many times connecting to a peer must fail before
2148 it is marked as down. Default is 10.
2150 allow-miss Disable Squid's use of only-if-cached when forwarding
2151 requests to siblings. This is primarily useful when
2152 icp_hit_stale is used by the sibling. To extensive use
2153 of this option may result in forwarding loops, and you
2154 should avoid having two-way peerings with this option.
2155 For example to deny peer usage on requests from peer
2156 by denying cache_peer_access if the source is a peer.
2158 max-conn=N Limit the amount of connections Squid may open to this
2161 name=xxx Unique name for the peer.
2162 Required if you have multiple peers on the same host
2163 but different ports.
2164 This name can be used in cache_peer_access and similar
2165 directives to dentify the peer.
2166 Can be used by outgoing access controls through the
2169 no-tproxy Do not use the client-spoof TPROXY support when forwarding
2170 requests to this peer. Use normal address selection instead.
2172 proxy-only objects fetched from the peer will not be stored locally.
2176 NAME: cache_peer_domain cache_host_domain
2181 Use to limit the domains for which a neighbor cache will be
2184 cache_peer_domain cache-host domain [domain ...]
2185 cache_peer_domain cache-host !domain
2187 For example, specifying
2189 cache_peer_domain parent.foo.net .edu
2191 has the effect such that UDP query packets are sent to
2192 'bigserver' only when the requested object exists on a
2193 server in the .edu domain. Prefixing the domainname
2194 with '!' means the cache will be queried for objects
2197 NOTE: * Any number of domains may be given for a cache-host,
2198 either on the same or separate lines.
2199 * When multiple domains are given for a particular
2200 cache-host, the first matched domain is applied.
2201 * Cache hosts with no domain restrictions are queried
2203 * There are no defaults.
2204 * There is also a 'cache_peer_access' tag in the ACL
2208 NAME: cache_peer_access
2213 Similar to 'cache_peer_domain' but provides more flexibility by
2216 cache_peer_access cache-host allow|deny [!]aclname ...
2218 The syntax is identical to 'http_access' and the other lists of
2219 ACL elements. See the comments for 'http_access' below, or
2220 the Squid FAQ (http://wiki.squid-cache.org/SquidFaq/SquidAcl).
2223 NAME: neighbor_type_domain
2224 TYPE: hostdomaintype
2228 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
2230 Modifying the neighbor type for specific domains is now
2231 possible. You can treat some domains differently than the
2232 default neighbor type specified on the 'cache_peer' line.
2233 Normally it should only be necessary to list domains which
2234 should be treated differently because the default neighbor type
2235 applies for hostnames which do not match domains listed here.
2238 cache_peer cache.foo.org parent 3128 3130
2239 neighbor_type_domain cache.foo.org sibling .com .net
2240 neighbor_type_domain cache.foo.org sibling .au .de
2243 NAME: dead_peer_timeout
2247 LOC: Config.Timeout.deadPeer
2249 This controls how long Squid waits to declare a peer cache
2250 as "dead." If there are no ICP replies received in this
2251 amount of time, Squid will declare the peer dead and not
2252 expect to receive any further ICP replies. However, it
2253 continues to send ICP queries, and will mark the peer as
2254 alive upon receipt of the first subsequent ICP reply.
2256 This timeout also affects when Squid expects to receive ICP
2257 replies from peers. If more than 'dead_peer' seconds have
2258 passed since the last ICP reply was received, Squid will not
2259 expect to receive an ICP reply on the next query. Thus, if
2260 your time between requests is greater than this timeout, you
2261 will see a lot of requests sent DIRECT to origin servers
2262 instead of to your parents.
2265 NAME: forward_max_tries
2268 LOC: Config.forward_max_tries
2270 Controls how many different forward paths Squid will try
2271 before giving up. See also forward_timeout.
2274 NAME: hierarchy_stoplist
2277 LOC: Config.hierarchy_stoplist
2279 A list of words which, if found in a URL, cause the object to
2280 be handled directly by this cache. In other words, use this
2281 to not query neighbor caches for certain objects. You may
2282 list this option multiple times.
2283 Note: never_direct overrides this option.
2286 # We recommend you to use at least the following line.
2287 hierarchy_stoplist cgi-bin ?
2292 MEMORY CACHE OPTIONS
2293 -----------------------------------------------------------------------------
2300 LOC: Config.memMaxSize
2302 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
2303 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
2304 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
2305 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
2307 'cache_mem' specifies the ideal amount of memory to be used
2309 * In-Transit objects
2311 * Negative-Cached objects
2313 Data for these objects are stored in 4 KB blocks. This
2314 parameter specifies the ideal upper limit on the total size of
2315 4 KB blocks allocated. In-Transit objects take the highest
2318 In-transit objects have priority over the others. When
2319 additional space is needed for incoming data, negative-cached
2320 and hot objects will be released. In other words, the
2321 negative-cached and hot objects will fill up any unused space
2322 not needed for in-transit objects.
2324 If circumstances require, this limit will be exceeded.
2325 Specifically, if your incoming request rate requires more than
2326 'cache_mem' of memory to hold in-transit objects, Squid will
2327 exceed this limit to satisfy the new requests. When the load
2328 decreases, blocks will be freed until the high-water mark is
2329 reached. Thereafter, blocks will be used to store hot
2333 NAME: maximum_object_size_in_memory
2337 LOC: Config.Store.maxInMemObjSize
2339 Objects greater than this size will not be attempted to kept in
2340 the memory cache. This should be set high enough to keep objects
2341 accessed frequently in memory to improve performance whilst low
2342 enough to keep larger objects from hoarding cache_mem.
2345 NAME: memory_cache_mode
2350 Controls which objects to keep in the memory cache (cache_mem)
2352 always Keep most recently fetched objects in memory (default)
2354 disk Only disk cache hits are kept in memory, which means
2355 an object must first be cached on disk and then hit
2356 a second time before cached in memory.
2358 network Only objects fetched from network is kept in memory
2361 NAME: memory_replacement_policy
2363 LOC: Config.memPolicy
2366 The memory replacement policy parameter determines which
2367 objects are purged from memory when memory space is needed.
2369 See cache_replacement_policy for details.
2374 -----------------------------------------------------------------------------
2377 NAME: cache_replacement_policy
2379 LOC: Config.replPolicy
2382 The cache replacement policy parameter determines which
2383 objects are evicted (replaced) when disk space is needed.
2385 lru : Squid's original list based LRU policy
2386 heap GDSF : Greedy-Dual Size Frequency
2387 heap LFUDA: Least Frequently Used with Dynamic Aging
2388 heap LRU : LRU policy implemented using a heap
2390 Applies to any cache_dir lines listed below this.
2392 The LRU policies keeps recently referenced objects.
2394 The heap GDSF policy optimizes object hit rate by keeping smaller
2395 popular objects in cache so it has a better chance of getting a
2396 hit. It achieves a lower byte hit rate than LFUDA though since
2397 it evicts larger (possibly popular) objects.
2399 The heap LFUDA policy keeps popular objects in cache regardless of
2400 their size and thus optimizes byte hit rate at the expense of
2401 hit rate since one large, popular object will prevent many
2402 smaller, slightly less popular objects from being cached.
2404 Both policies utilize a dynamic aging mechanism that prevents
2405 cache pollution that can otherwise occur with frequency-based
2406 replacement policies.
2408 NOTE: if using the LFUDA replacement policy you should increase
2409 the value of maximum_object_size above its default of 4096 KB to
2410 to maximize the potential byte hit rate improvement of LFUDA.
2412 For more information about the GDSF and LFUDA cache replacement
2413 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
2414 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
2420 LOC: Config.cacheSwap
2424 cache_dir Type Directory-Name Fs-specific-data [options]
2426 You can specify multiple cache_dir lines to spread the
2427 cache among different disk partitions.
2429 Type specifies the kind of storage system to use. Only "ufs"
2430 is built by default. To enable any of the other storage systems
2431 see the --enable-storeio configure option.
2433 'Directory' is a top-level directory where cache swap
2434 files will be stored. If you want to use an entire disk
2435 for caching, this can be the mount-point directory.
2436 The directory must exist and be writable by the Squid
2437 process. Squid will NOT create this directory for you.
2441 "ufs" is the old well-known Squid storage format that has always
2444 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
2446 'Mbytes' is the amount of disk space (MB) to use under this
2447 directory. The default is 100 MB. Change this to suit your
2448 configuration. Do NOT put the size of your disk drive here.
2449 Instead, if you want Squid to use the entire disk drive,
2450 subtract 20% and use that value.
2452 'Level-1' is the number of first-level subdirectories which
2453 will be created under the 'Directory'. The default is 16.
2455 'Level-2' is the number of second-level subdirectories which
2456 will be created under each first-level directory. The default
2459 The aufs store type:
2461 "aufs" uses the same storage format as "ufs", utilizing
2462 POSIX-threads to avoid blocking the main Squid process on
2463 disk-I/O. This was formerly known in Squid as async-io.
2465 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
2467 see argument descriptions under ufs above
2469 The diskd store type:
2471 "diskd" uses the same storage format as "ufs", utilizing a
2472 separate process to avoid blocking the main Squid process on
2475 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
2477 see argument descriptions under ufs above
2479 Q1 specifies the number of unacknowledged I/O requests when Squid
2480 stops opening new files. If this many messages are in the queues,
2481 Squid won't open new files. Default is 64
2483 Q2 specifies the number of unacknowledged messages when Squid
2484 starts blocking. If this many messages are in the queues,
2485 Squid blocks until it receives some replies. Default is 72
2487 When Q1 < Q2 (the default), the cache directory is optimized
2488 for lower response time at the expense of a decrease in hit
2489 ratio. If Q1 > Q2, the cache directory is optimized for
2490 higher hit ratio at the expense of an increase in response
2493 The coss store type:
2495 NP: COSS filesystem in Squid-3 has been deemed too unstable for
2496 production use and has thus been removed from this release.
2497 We hope that it can be made usable again soon.
2499 block-size=n defines the "block size" for COSS cache_dir's.
2500 Squid uses file numbers as block numbers. Since file numbers
2501 are limited to 24 bits, the block size determines the maximum
2502 size of the COSS partition. The default is 512 bytes, which
2503 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
2504 you should not change the coss block size after Squid
2505 has written some objects to the cache_dir.
2507 The coss file store has changed from 2.5. Now it uses a file
2508 called 'stripe' in the directory names in the config - and
2509 this will be created by squid -z.
2513 no-store, no new objects should be stored to this cache_dir
2515 max-size=n, refers to the max object size this storedir supports.
2516 It is used to initially choose the storedir to dump the object.
2517 Note: To make optimal use of the max-size limits you should order
2518 the cache_dir lines with the smallest max-size value first and the
2519 ones with no max-size specification last.
2521 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
2522 which can be changed with the --with-coss-membuf-size=N configure
2526 # Uncomment and adjust the following to add a disk cache directory.
2527 #cache_dir ufs @DEFAULT_SWAP_DIR@ 100 16 256
2531 NAME: store_dir_select_algorithm
2533 LOC: Config.store_dir_select_algorithm
2536 Set this to 'round-robin' as an alternative.
2539 NAME: max_open_disk_fds
2541 LOC: Config.max_open_disk_fds
2544 To avoid having disk as the I/O bottleneck Squid can optionally
2545 bypass the on-disk cache if more than this amount of disk file
2546 descriptors are open.
2548 A value of 0 indicates no limit.
2551 NAME: minimum_object_size
2555 LOC: Config.Store.minObjectSize
2557 Objects smaller than this size will NOT be saved on disk. The
2558 value is specified in kilobytes, and the default is 0 KB, which
2559 means there is no minimum.
2562 NAME: maximum_object_size
2566 LOC: Config.Store.maxObjectSize
2568 Objects larger than this size will NOT be saved on disk. The
2569 value is specified in kilobytes, and the default is 4MB. If
2570 you wish to get a high BYTES hit ratio, you should probably
2571 increase this (one 32 MB object hit counts for 3200 10KB
2572 hits). If you wish to increase speed more than your want to
2573 save bandwidth you should leave this low.
2575 NOTE: if using the LFUDA replacement policy you should increase
2576 this value to maximize the byte hit rate improvement of LFUDA!
2577 See replacement_policy below for a discussion of this policy.
2580 NAME: cache_swap_low
2581 COMMENT: (percent, 0-100)
2584 LOC: Config.Swap.lowWaterMark
2587 NAME: cache_swap_high
2588 COMMENT: (percent, 0-100)
2591 LOC: Config.Swap.highWaterMark
2594 The low- and high-water marks for cache object replacement.
2595 Replacement begins when the swap (disk) usage is above the
2596 low-water mark and attempts to maintain utilization near the
2597 low-water mark. As swap utilization gets close to high-water
2598 mark object eviction becomes more aggressive. If utilization is
2599 close to the low-water mark less replacement is done each time.
2601 Defaults are 90% and 95%. If you have a large cache, 5% could be
2602 hundreds of MB. If this is the case you may wish to set these
2603 numbers closer together.
2608 -----------------------------------------------------------------------------
2613 LOC: Config.Log.logformats
2618 logformat <name> <format specification>
2620 Defines an access log format.
2622 The <format specification> is a string with embedded % format codes
2624 % format codes all follow the same basic structure where all but
2625 the formatcode is optional. Output strings are automatically escaped
2626 as required according to their context and the output format
2627 modifiers are usually not needed, but can be specified if an explicit
2628 output format is desired.
2630 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
2632 " output in quoted string format
2633 [ output in squid text log format as used by log_mime_hdrs
2634 # output in URL quoted format
2638 width field width. If starting with 0 the
2639 output is zero padded
2640 {arg} argument such as header name etc
2644 % a literal % character
2645 >a Client source IP address
2647 >p Client source port
2648 <A Server IP address or peer name
2649 la Local IP address (http_port)
2650 lp Local port number (http_port)
2651 <lp Local port number of the last server or peer connection
2652 sn Unique sequence number per log line entry
2653 ts Seconds since epoch
2654 tu subsecond time (milliseconds)
2655 tl Local time. Optional strftime format argument
2656 default %d/%b/%Y:%H:%M:%S %z
2657 tg GMT time. Optional strftime format argument
2658 default %d/%b/%Y:%H:%M:%S %z
2659 tr Response time (milliseconds)
2660 dt Total time spent making DNS lookups (milliseconds)
2662 HTTP cache related format codes:
2664 [http::]>h Original request header. Optional header name argument
2665 on the format header[:[separator]element]
2666 [http::]>ha The HTTP request headers after adaptation and redirection.
2667 Optional header name argument as for >h
2668 [http::]<h Reply header. Optional header name argument
2670 [http::]un User name
2671 [http::]ul User name from authentication
2672 [http::]ui User name from ident
2673 [http::]us User name from SSL
2674 [http::]ue User name from external acl helper
2675 [http::]>Hs HTTP status code sent to the client
2676 [http::]<Hs HTTP status code received from the next hop
2677 [http::]Ss Squid request status (TCP_MISS etc)
2678 [http::]Sh Squid hierarchy status (DEFAULT_PARENT etc)
2679 [http::]mt MIME content type
2680 [http::]rm Request method (GET/POST etc)
2681 [http::]ru Request URL
2682 [http::]rp Request URL-Path excluding hostname
2683 [http::]rv Request protocol version
2684 [http::]et Tag returned by external acl
2685 [http::]ea Log string returned by external acl
2686 [http::]<st Sent reply size including HTTP headers
2687 [http::]>st Received request size including HTTP headers. In the
2688 case of chunked requests the chunked encoding metadata
2690 [http::]>sh Received HTTP request headers size
2691 [http::]<sh Sent HTTP reply headers size
2692 [http::]st Request+Reply size including HTTP headers
2693 [http::]<sH Reply high offset sent
2694 [http::]<sS Upstream object size
2695 [http::]<pt Peer response time in milliseconds. The timer starts
2696 when the last request byte is sent to the next hop
2697 and stops when the last response byte is received.
2698 [http::]<tt Total server-side time in milliseconds. The timer
2699 starts with the first connect request (or write I/O)
2700 sent to the first selected peer. The timer stops
2701 with the last I/O with the last peer.
2703 If ICAP is enabled, the following two codes become available (as
2704 well as ICAP log codes documented with the icap_log option):
2706 icap::tt Total ICAP processing time for the HTTP
2707 transaction. The timer ticks when ICAP
2708 ACLs are checked and when ICAP
2709 transaction is in progress.
2711 icap::<last_h The header of the last ICAP response
2712 related to the HTTP transaction. Like
2713 <h, accepts an optional header name
2714 argument. Will not change semantics
2715 when multiple ICAP transactions per HTTP
2716 transaction are supported.
2718 If adaptation is enabled the following two codes become available:
2720 adapt::sum_trs Summed adaptation transaction response
2721 times recorded as a comma-separated list in
2722 the order of transaction start time. Each time
2723 value is recorded as an integer number,
2724 representing response time of one or more
2725 adaptation (ICAP or eCAP) transaction in
2726 milliseconds. When a failed transaction is
2727 being retried or repeated, its time is not
2728 logged individually but added to the
2729 replacement (next) transaction. See also:
2732 adapt::all_trs All adaptation transaction response times.
2733 Same as adaptation_strs but response times of
2734 individual transactions are never added
2735 together. Instead, all transaction response
2736 times are recorded individually.
2738 You can prefix adapt::*_trs format codes with adaptation
2739 service name in curly braces to record response time(s) specific
2740 to that service. For example: %{my_service}adapt::sum_trs
2742 The default formats available (which do not need re-defining) are:
2744 logformat squid %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %un %Sh/%<A %mt
2745 logformat squidmime %ts.%03tu %6tr %>a %Ss/%03>Hs %<st %rm %ru %un %Sh/%<A %mt [%>h] [%<h]
2746 logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st %Ss:%Sh
2747 logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %>Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
2750 NAME: access_log cache_access_log
2752 LOC: Config.Log.accesslogs
2754 DEFAULT_IF_NONE: daemon:@DEFAULT_ACCESS_LOG@ squid
2756 These files log client request activities. Has a line every HTTP or
2757 ICP request. The format is:
2758 access_log <module>:<place> [<logformat name> [acl acl ...]]
2759 access_log none [acl acl ...]]
2761 Will log to the specified module:place using the specified format (which
2762 must be defined in a logformat directive) those entries which match
2763 ALL the acl's specified (which must be defined in acl clauses).
2764 If no acl is specified, all requests will be logged to this destination.
2766 ===== Modules Currently available =====
2768 none Do not log any requests matchign these ACL.
2769 Do not specify Place or logformat name.
2771 stdio Write each log line to disk immediately at the completion of
2773 Place: the filename and path to be written.
2775 daemon Very similar to stdio. But instead of writing to disk the log
2776 line is passed to a daemon helper for asychronous handling instead.
2777 Place: varies depending on the daemon.
2779 log_file_daemon Place: the file name and path to be written.
2781 syslog To log each request via syslog facility.
2782 Place: The syslog facility and priority level for these entries.
2783 Place Format: facility.priority
2785 where facility could be any of:
2786 authpriv, daemon, local0 ... local7 or user.
2788 And priority could be any of:
2789 err, warning, notice, info, debug.
2791 udp To send each log line as text data to a UDP receiver.
2792 Place: The destination host name or IP and port.
2793 Place Format: \\host:port
2795 tcp To send each log line as text data to a TCP receiver.
2796 Place: The destination host name or IP and port.
2797 Place Format: \\host:port
2800 access_log daemon:@DEFAULT_ACCESS_LOG@ squid
2806 LOC: Config.Log.icaplogs
2809 ICAP log files record ICAP transaction summaries, one line per
2812 The icap_log option format is:
2813 icap_log <filepath> [<logformat name> [acl acl ...]]
2814 icap_log none [acl acl ...]]
2816 Please see access_log option documentation for details. The two
2817 kinds of logs share the overall configuration approach and many
2820 ICAP processing of a single HTTP message or transaction may
2821 require multiple ICAP transactions. In such cases, multiple
2822 ICAP transaction log lines will correspond to a single access
2825 ICAP log uses logformat codes that make sense for an ICAP
2826 transaction. Header-related codes are applied to the HTTP header
2827 embedded in an ICAP server response, with the following caveats:
2828 For REQMOD, there is no HTTP response header unless the ICAP
2829 server performed request satisfaction. For RESPMOD, the HTTP
2830 request header is the header sent to the ICAP server. For
2831 OPTIONS, there are no HTTP headers.
2833 The following format codes are also available for ICAP logs:
2835 icap::<A ICAP server IP address. Similar to <A.
2837 icap::<service_name ICAP service name from the icap_service
2838 option in Squid configuration file.
2840 icap::ru ICAP Request-URI. Similar to ru.
2842 icap::rm ICAP request method (REQMOD, RESPMOD, or
2843 OPTIONS). Similar to existing rm.
2845 icap::>st Bytes sent to the ICAP server (TCP payload
2846 only; i.e., what Squid writes to the socket).
2848 icap::<st Bytes received from the ICAP server (TCP
2849 payload only; i.e., what Squid reads from
2852 icap::tr Transaction response time (in
2853 milliseconds). The timer starts when
2854 the ICAP transaction is created and
2855 stops when the transaction is completed.
2858 icap::tio Transaction I/O time (in milliseconds). The
2859 timer starts when the first ICAP request
2860 byte is scheduled for sending. The timers
2861 stops when the last byte of the ICAP response
2864 icap::to Transaction outcome: ICAP_ERR* for all
2865 transaction errors, ICAP_OPT for OPTION
2866 transactions, ICAP_ECHO for 204
2867 responses, ICAP_MOD for message
2868 modification, and ICAP_SAT for request
2869 satisfaction. Similar to Ss.
2871 icap::Hs ICAP response status code. Similar to Hs.
2873 icap::>h ICAP request header(s). Similar to >h.
2875 icap::<h ICAP response header(s). Similar to <h.
2877 The default ICAP log format, which can be used without an explicit
2878 definition, is called icap_squid:
2880 logformat icap_squid %ts.%03tu %6icap::tr %>a %icap::to/%03icap::Hs %icap::<size %icap::rm %icap::ru% %un -/%icap::<A -
2882 See also: logformat, log_icap, and %icap::<last_h
2885 NAME: logfile_daemon
2887 DEFAULT: @DEFAULT_LOGFILED@
2888 LOC: Log::TheConfig.logfile_daemon
2890 Specify the path to the logfile-writing daemon. This daemon is
2891 used to write the access and store logs, if configured.
2893 Squid sends a number of commands to the log daemon:
2894 L<data>\n - logfile data
2899 r<n>\n - set rotate count to <n>
2900 b<n>\n - 1 = buffer output, 0 = don't buffer output
2902 No responses is expected.
2907 LOC: Config.accessList.log
2909 COMMENT: allow|deny acl acl...
2911 This options allows you to control which requests gets logged
2912 to access.log (see access_log directive). Requests denied for
2913 logging will also not be accounted for in performance counters.
2915 This clause only supports fast acl types.
2916 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
2922 LOC: Config.accessList.icap
2925 This options allows you to control which requests get logged
2926 to icap.log. See the icap_log directive for ICAP log details.
2929 NAME: cache_store_log
2932 LOC: Config.Log.store
2934 Logs the activities of the storage manager. Shows which
2935 objects are ejected from the cache, and which objects are
2936 saved and for how long. To disable, enter "none" or remove the line.
2937 There are not really utilities to analyze this data, so you can safely
2941 cache_store_log @DEFAULT_STORE_LOG@
2944 NAME: cache_swap_state cache_swap_log
2946 LOC: Config.Log.swap
2949 Location for the cache "swap.state" file. This index file holds
2950 the metadata of objects saved on disk. It is used to rebuild
2951 the cache during startup. Normally this file resides in each
2952 'cache_dir' directory, but you may specify an alternate
2953 pathname here. Note you must give a full filename, not just
2954 a directory. Since this is the index for the whole object
2955 list you CANNOT periodically rotate it!
2957 If %s can be used in the file name it will be replaced with a
2958 a representation of the cache_dir name where each / is replaced
2959 with '.'. This is needed to allow adding/removing cache_dir
2960 lines when cache_swap_log is being used.
2962 If have more than one 'cache_dir', and %s is not used in the name
2963 these swap logs will have names such as:
2969 The numbered extension (which is added automatically)
2970 corresponds to the order of the 'cache_dir' lines in this
2971 configuration file. If you change the order of the 'cache_dir'
2972 lines in this file, these index files will NOT correspond to
2973 the correct 'cache_dir' entry (unless you manually rename
2974 them). We recommend you do NOT use this option. It is
2975 better to keep these index files in each 'cache_dir' directory.
2978 NAME: logfile_rotate
2981 LOC: Config.Log.rotateNumber
2983 Specifies the number of logfile rotations to make when you
2984 type 'squid -k rotate'. The default is 10, which will rotate
2985 with extensions 0 through 9. Setting logfile_rotate to 0 will
2986 disable the file name rotation, but the logfiles are still closed
2987 and re-opened. This will enable you to rename the logfiles
2988 yourself just before sending the rotate signal.
2990 Note, the 'squid -k rotate' command normally sends a USR1
2991 signal to the running squid process. In certain situations
2992 (e.g. on Linux with Async I/O), USR1 is used for other
2993 purposes, so -k rotate uses another signal. It is best to get
2994 in the habit of using 'squid -k rotate' instead of 'kill -USR1
2997 Note, from Squid-3.1 this option has no effect on the cache.log,
2998 that log can be rotated separately by using debug_options
3001 NAME: emulate_httpd_log
3005 LOC: Config.onoff.common_log
3007 The Cache can emulate the log file format which many 'httpd'
3008 programs use. To disable/enable this emulation, set
3009 emulate_httpd_log to 'off' or 'on'. The default
3010 is to use the native log format since it includes useful
3011 information Squid-specific log analyzers use.
3014 NAME: log_ip_on_direct
3018 LOC: Config.onoff.log_ip_on_direct
3020 Log the destination IP address in the hierarchy log tag when going
3021 direct. Earlier Squid versions logged the hostname here. If you
3022 prefer the old way set this to off.
3027 DEFAULT: @DEFAULT_MIME_TABLE@
3028 LOC: Config.mimeTablePathname
3030 Pathname to Squid's MIME table. You shouldn't need to change
3031 this, but the default file contains examples and formatting
3032 information if you do.
3038 LOC: Config.onoff.log_mime_hdrs
3041 The Cache can record both the request and the response MIME
3042 headers for each HTTP transaction. The headers are encoded
3043 safely and will appear as two bracketed fields at the end of
3044 the access log (for either the native or httpd-emulated log
3045 formats). To enable this logging set log_mime_hdrs to 'on'.
3050 LOC: Config.Log.useragent
3052 IFDEF: USE_USERAGENT_LOG
3054 Squid will write the User-Agent field from HTTP requests
3055 to the filename specified here. By default useragent_log
3059 NAME: referer_log referrer_log
3061 LOC: Config.Log.referer
3063 IFDEF: USE_REFERER_LOG
3065 Squid will write the Referer field from HTTP requests to the
3066 filename specified here. By default referer_log is disabled.
3067 Note that "referer" is actually a misspelling of "referrer"
3068 however the misspelt version has been accepted into the HTTP RFCs
3074 DEFAULT: @DEFAULT_PID_FILE@
3075 LOC: Config.pidFilename
3077 A filename to write the process-id to. To disable, enter "none".
3084 LOC: Config.onoff.log_fqdn
3086 Turn this on if you wish to log fully qualified domain names
3087 in the access.log. To do this Squid does a DNS lookup of all
3088 IP's connecting to it. This can (in some situations) increase
3089 latency, which makes your cache seem slower for interactive
3093 NAME: client_netmask
3095 LOC: Config.Addrs.client_netmask
3098 A netmask for client addresses in logfiles and cachemgr output.
3099 Change this to protect the privacy of your cache clients.
3100 A netmask of 255.255.255.0 will log all IP's in that range with
3101 the last digit set to '0'.
3108 LOC: Config.Log.forward
3110 Logs the server-side requests.
3112 This is currently work in progress.
3115 NAME: strip_query_terms
3117 LOC: Config.onoff.strip_query_terms
3120 By default, Squid strips query terms from requested URLs before
3121 logging. This protects your user's privacy.
3128 LOC: Config.onoff.buffered_logs
3130 cache.log log file is written with stdio functions, and as such
3131 it can be buffered or unbuffered. By default it will be unbuffered.
3132 Buffering it can speed up the writing slightly (though you are
3133 unlikely to need to worry unless you run with tons of debugging
3134 enabled in which case performance will suffer badly anyway..).
3137 NAME: netdb_filename
3139 DEFAULT: @DEFAULT_NETDB_FILE@
3140 LOC: Config.netdbFilename
3143 A filename where Squid stores it's netdb state between restarts.
3144 To disable, enter "none".
3148 OPTIONS FOR TROUBLESHOOTING
3149 -----------------------------------------------------------------------------
3155 DEFAULT_IF_NONE: @DEFAULT_CACHE_LOG@
3156 LOC: Debug::cache_log
3158 Cache logging file. This is where general information about
3159 your cache's behavior goes. You can increase the amount of data
3160 logged to this file and how often its rotated with "debug_options"
3166 LOC: Debug::debugOptions
3168 Logging options are set as section,level where each source file
3169 is assigned a unique section. Lower levels result in less
3170 output, Full debugging (level 9) can result in a very large
3171 log file, so be careful.
3173 The magic word "ALL" sets debugging levels for all sections.
3174 We recommend normally running with "ALL,1".
3176 The rotate=N option can be used to keep more or less of these logs
3177 than would otherwise be kept by logfile_rotate.
3178 For most uses a single log should be enough to monitor current
3179 events affecting Squid.
3184 LOC: Config.coredump_dir
3186 DEFAULT_IF_NONE: none
3188 By default Squid leaves core files in the directory from where
3189 it was started. If you set 'coredump_dir' to a directory
3190 that exists, Squid will chdir() to that directory at startup
3191 and coredump files will be left there.
3195 # Leave coredumps in the first cache dir
3196 coredump_dir @DEFAULT_SWAP_DIR@
3202 OPTIONS FOR FTP GATEWAYING
3203 -----------------------------------------------------------------------------
3209 LOC: Config.Ftp.anon_user
3211 If you want the anonymous login password to be more informative
3212 (and enable the use of picky ftp servers), set this to something
3213 reasonable for your domain, like wwwuser@somewhere.net
3215 The reason why this is domainless by default is the
3216 request can be made on the behalf of a user in any domain,
3217 depending on how the cache is used.
3218 Some ftp server also validate the email address is valid
3219 (for example perl.com).
3225 LOC: Config.Ftp.passive
3227 If your firewall does not allow Squid to use passive
3228 connections, turn off this option.
3230 Use of ftp_epsv_all option requires this to be ON.
3236 LOC: Config.Ftp.epsv_all
3238 FTP Protocol extensions permit the use of a special "EPSV ALL" command.
3240 NATs may be able to put the connection on a "fast path" through the
3241 translator, as the EPRT command will never be used and therefore,
3242 translation of the data portion of the segments will never be needed.
3244 When a client only expects to do two-way FTP transfers this may be
3246 If squid finds that it must do a three-way FTP transfer after issuing
3247 an EPSV ALL command, the FTP session will fail.
3249 If you have any doubts about this option do not use it.
3250 Squid will nicely attempt all other connection methods.
3252 Requires ftp_passive to be ON (default) for any effect.
3258 LOC: Config.Ftp.epsv
3260 FTP Protocol extensions permit the use of a special "EPSV" command.
3262 NATs may be able to put the connection on a "fast path" through the
3263 translator using EPSV, as the EPRT command will never be used
3264 and therefore, translation of the data portion of the segments
3265 will never be needed.
3267 Turning this OFF will prevent EPSV being attempted.
3268 WARNING: Doing so will convert Squid back to the old behavior with all
3269 the related problems with external NAT devices/layers.
3271 Requires ftp_passive to be ON (default) for any effect.
3274 NAME: ftp_sanitycheck
3277 LOC: Config.Ftp.sanitycheck
3279 For security and data integrity reasons Squid by default performs
3280 sanity checks of the addresses of FTP data connections ensure the
3281 data connection is to the requested server. If you need to allow
3282 FTP connections to servers using another IP address for the data
3283 connection turn this off.
3286 NAME: ftp_telnet_protocol
3289 LOC: Config.Ftp.telnet
3291 The FTP protocol is officially defined to use the telnet protocol
3292 as transport channel for the control connection. However, many
3293 implementations are broken and does not respect this aspect of
3296 If you have trouble accessing files with ASCII code 255 in the
3297 path or similar problems involving this ASCII code you can
3298 try setting this directive to off. If that helps, report to the
3299 operator of the FTP server in question that their FTP server
3300 is broken and does not follow the FTP standard.
3304 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
3305 -----------------------------------------------------------------------------
3310 DEFAULT: @DEFAULT_DISKD@
3311 LOC: Config.Program.diskd
3313 Specify the location of the diskd executable.
3314 Note this is only useful if you have compiled in
3315 diskd as one of the store io modules.
3318 NAME: unlinkd_program
3321 DEFAULT: @DEFAULT_UNLINKD@
3322 LOC: Config.Program.unlinkd
3324 Specify the location of the executable for file deletion process.
3327 NAME: pinger_program
3329 DEFAULT: @DEFAULT_PINGER@
3330 LOC: Config.pinger.program
3333 Specify the location of the executable for the pinger process.
3339 LOC: Config.pinger.enable
3342 Control whether the pinger is active at run-time.
3343 Enables turning ICMP pinger on and off with a simple
3344 squid -k reconfigure.
3349 OPTIONS FOR URL REWRITING
3350 -----------------------------------------------------------------------------
3353 NAME: url_rewrite_program redirect_program
3355 LOC: Config.Program.redirect
3358 Specify the location of the executable for the URL rewriter.
3359 Since they can perform almost any function there isn't one included.
3361 For each requested URL rewriter will receive on line with the format
3363 URL <SP> client_ip "/" fqdn <SP> user <SP> method [<SP> kvpairs]<NL>
3365 In the future, the rewriter interface will be extended with
3366 key=value pairs ("kvpairs" shown above). Rewriter programs
3367 should be prepared to receive and possibly ignore additional
3368 whitespace-separated tokens on each input line.
3370 And the rewriter may return a rewritten URL. The other components of
3371 the request line does not need to be returned (ignored if they are).
3373 The rewriter can also indicate that a client-side redirect should
3374 be performed to the new URL. This is done by prefixing the returned
3375 URL with "301:" (moved permanently) or 302: (moved temporarily).
3377 By default, a URL rewriter is not used.
3380 NAME: url_rewrite_children redirect_children
3381 TYPE: HelperChildConfig
3382 DEFAULT: 20 startup=0 idle=1 concurrency=0
3383 LOC: Config.redirectChildren
3385 The maximum number of redirector processes to spawn. If you limit
3386 it too few Squid will have to wait for them to process a backlog of
3387 URLs, slowing it down. If you allow too many they will use RAM
3388 and other system resources noticably.
3390 The startup= and idle= options allow some measure of skew in your
3395 Sets a minimum of how many processes are to be spawned when Squid
3396 starts or reconfigures. When set to zero the first request will
3397 cause spawning of the first child process to handle it.
3399 Starting too few will cause an initial slowdown in traffic as Squid
3400 attempts to simultaneously spawn enough processes to cope.
3404 Sets a minimum of how many processes Squid is to try and keep available
3405 at all times. When traffic begins to rise above what the existing
3406 processes can handle this many more will be spawned up to the maximum
3407 configured. A minimum setting of 1 is required.
3411 The number of requests each redirector helper can handle in
3412 parallel. Defaults to 0 which indicates the redirector
3413 is a old-style single threaded redirector.
3415 When this directive is set to a value >= 1 then the protocol
3416 used to communicate with the helper is modified to include
3417 a request ID in front of the request/response. The request
3418 ID from the request must be echoed back with the response
3422 NAME: url_rewrite_host_header redirect_rewrites_host_header
3425 LOC: Config.onoff.redir_rewrites_host
3427 By default Squid rewrites any Host: header in redirected
3428 requests. If you are running an accelerator this may
3429 not be a wanted effect of a redirector.
3431 WARNING: Entries are cached on the result of the URL rewriting
3432 process, so be careful if you have domain-virtual hosts.
3435 NAME: url_rewrite_access redirector_access
3438 LOC: Config.accessList.redirector
3440 If defined, this access list specifies which requests are
3441 sent to the redirector processes. By default all requests
3444 This clause supports both fast and slow acl types.
3445 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3448 NAME: url_rewrite_bypass redirector_bypass
3450 LOC: Config.onoff.redirector_bypass
3453 When this is 'on', a request will not go through the
3454 redirector if all redirectors are busy. If this is 'off'
3455 and the redirector queue grows too large, Squid will exit
3456 with a FATAL error and ask you to increase the number of
3457 redirectors. You should only enable this if the redirectors
3458 are not critical to your caching system. If you use
3459 redirectors for access control, and you enable this option,
3460 users may have access to pages they should not
3461 be allowed to request.
3465 OPTIONS FOR TUNING THE CACHE
3466 -----------------------------------------------------------------------------
3469 NAME: cache no_cache
3472 LOC: Config.accessList.noCache
3474 A list of ACL elements which, if matched and denied, cause the request to
3475 not be satisfied from the cache and the reply to not be cached.
3476 In other words, use this to force certain objects to never be cached.
3478 You must use the words 'allow' or 'deny' to indicate whether items
3479 matching the ACL should be allowed or denied into the cache.
3481 Default is to allow all to be cached.
3483 This clause supports both fast and slow acl types.
3484 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3487 NAME: refresh_pattern
3488 TYPE: refreshpattern
3492 usage: refresh_pattern [-i] regex min percent max [options]
3494 By default, regular expressions are CASE-SENSITIVE. To make
3495 them case-insensitive, use the -i option.
3497 'Min' is the time (in minutes) an object without an explicit
3498 expiry time should be considered fresh. The recommended
3499 value is 0, any higher values may cause dynamic applications
3500 to be erroneously cached unless the application designer
3501 has taken the appropriate actions.
3503 'Percent' is a percentage of the objects age (time since last
3504 modification age) an object without explicit expiry time
3505 will be considered fresh.
3507 'Max' is an upper limit on how long objects without an explicit
3508 expiry time will be considered fresh.
3510 options: override-expire
3516 ignore-must-revalidate
3522 override-expire enforces min age even if the server
3523 sent an explicit expiry time (e.g., with the
3524 Expires: header or Cache-Control: max-age). Doing this
3525 VIOLATES the HTTP standard. Enabling this feature
3526 could make you liable for problems which it causes.
3528 Note: override-expire does not enforce staleness - it only extends
3529 freshness / min. If the server returns a Expires time which
3530 is longer than your max time, Squid will still consider
3531 the object fresh for that period of time.
3533 override-lastmod enforces min age even on objects
3534 that were modified recently.
3536 reload-into-ims changes client no-cache or ``reload''
3537 to If-Modified-Since requests. Doing this VIOLATES the
3538 HTTP standard. Enabling this feature could make you
3539 liable for problems which it causes.
3541 ignore-reload ignores a client no-cache or ``reload''
3542 header. Doing this VIOLATES the HTTP standard. Enabling
3543 this feature could make you liable for problems which
3546 ignore-no-cache ignores any ``Pragma: no-cache'' and
3547 ``Cache-control: no-cache'' headers received from a server.
3548 The HTTP RFC never allows the use of this (Pragma) header
3549 from a server, only a client, though plenty of servers
3552 ignore-no-store ignores any ``Cache-control: no-store''
3553 headers received from a server. Doing this VIOLATES
3554 the HTTP standard. Enabling this feature could make you
3555 liable for problems which it causes.
3557 ignore-must-revalidate ignores any ``Cache-Control: must-revalidate``
3558 headers received from a server. Doing this VIOLATES
3559 the HTTP standard. Enabling this feature could make you
3560 liable for problems which it causes.
3562 ignore-private ignores any ``Cache-control: private''
3563 headers received from a server. Doing this VIOLATES
3564 the HTTP standard. Enabling this feature could make you
3565 liable for problems which it causes.
3567 ignore-auth caches responses to requests with authorization,
3568 as if the originserver had sent ``Cache-control: public''
3569 in the response header. Doing this VIOLATES the HTTP standard.
3570 Enabling this feature could make you liable for problems which
3573 refresh-ims causes squid to contact the origin server
3574 when a client issues an If-Modified-Since request. This
3575 ensures that the client will receive an updated version
3576 if one is available.
3578 store-stale stores responses even if they don't have explicit
3579 freshness or a validator (i.e., Last-Modified or an ETag)
3580 present, or if they're already stale. By default, Squid will
3581 not cache such responses because they usually can't be
3582 reused. Note that such responses will be stale by default.
3584 Basically a cached object is:
3586 FRESH if expires < now, else STALE
3588 FRESH if lm-factor < percent, else STALE
3592 The refresh_pattern lines are checked in the order listed here.
3593 The first entry which matches is used. If none of the entries
3594 match the default will be used.
3596 Note, you must uncomment all the default lines if you want
3597 to change one. The default setting is only active if none is
3602 # Add any of your own refresh_pattern entries above these.
3603 refresh_pattern ^ftp: 1440 20% 10080
3604 refresh_pattern ^gopher: 1440 0% 1440
3605 refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
3606 refresh_pattern . 0 20% 4320
3610 NAME: quick_abort_min
3614 LOC: Config.quickAbort.min
3617 NAME: quick_abort_max
3621 LOC: Config.quickAbort.max
3624 NAME: quick_abort_pct
3628 LOC: Config.quickAbort.pct
3630 The cache by default continues downloading aborted requests
3631 which are almost completed (less than 16 KB remaining). This
3632 may be undesirable on slow (e.g. SLIP) links and/or very busy
3633 caches. Impatient users may tie up file descriptors and
3634 bandwidth by repeatedly requesting and immediately aborting
3637 When the user aborts a request, Squid will check the
3638 quick_abort values to the amount of data transfered until
3641 If the transfer has less than 'quick_abort_min' KB remaining,
3642 it will finish the retrieval.
3644 If the transfer has more than 'quick_abort_max' KB remaining,
3645 it will abort the retrieval.
3647 If more than 'quick_abort_pct' of the transfer has completed,
3648 it will finish the retrieval.
3650 If you do not want any retrieval to continue after the client
3651 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
3654 If you want retrievals to always continue if they are being
3655 cached set 'quick_abort_min' to '-1 KB'.
3658 NAME: read_ahead_gap
3659 COMMENT: buffer-size
3661 LOC: Config.readAheadGap
3664 The amount of data the cache will buffer ahead of what has been
3665 sent to the client when retrieving an object from another server.
3669 IFDEF: HTTP_VIOLATIONS
3672 LOC: Config.negativeTtl
3675 Set the Default Time-to-Live (TTL) for failed requests.
3676 Certain types of failures (such as "connection refused" and
3677 "404 Not Found") are able to be negatively-cached for a short time.
3678 Modern web servers should provide Expires: header, however if they
3679 do not this can provide a minimum TTL.
3680 The default is not to cache errors with unknown expiry details.
3682 Note that this is different from negative caching of DNS lookups.
3684 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3685 this feature could make you liable for problems which it
3689 NAME: positive_dns_ttl
3692 LOC: Config.positiveDnsTtl
3695 Upper limit on how long Squid will cache positive DNS responses.
3696 Default is 6 hours (360 minutes). This directive must be set
3697 larger than negative_dns_ttl.
3700 NAME: negative_dns_ttl
3703 LOC: Config.negativeDnsTtl
3706 Time-to-Live (TTL) for negative caching of failed DNS lookups.
3707 This also sets the lower cache limit on positive lookups.
3708 Minimum value is 1 second, and it is not recommendable to go
3709 much below 10 seconds.
3712 NAME: range_offset_limit
3713 COMMENT: size [acl acl...]
3715 LOC: Config.rangeOffsetLimit
3718 usage: (size) [units] [[!]aclname]
3720 Sets an upper limit on how far (number of bytes) into the file
3721 a Range request may be to cause Squid to prefetch the whole file.
3722 If beyond this limit, Squid forwards the Range request as it is and
3723 the result is NOT cached.
3725 This is to stop a far ahead range request (lets say start at 17MB)
3726 from making Squid fetch the whole object up to that point before
3727 sending anything to the client.
3729 Multiple range_offset_limit lines may be specified, and they will
3730 be searched from top to bottom on each request until a match is found.
3731 The first match found will be used. If no line matches a request, the
3732 default limit of 0 bytes will be used.
3734 'size' is the limit specified as a number of units.
3736 'units' specifies whether to use bytes, KB, MB, etc.
3737 If no units are specified bytes are assumed.
3739 A size of 0 causes Squid to never fetch more than the
3740 client requested. (default)
3742 A size of 'none' causes Squid to always fetch the object from the
3743 beginning so it may cache the result. (2.0 style)
3745 'aclname' is the name of a defined ACL.
3747 NP: Using 'none' as the byte value here will override any quick_abort settings
3748 that may otherwise apply to the range request. The range request will
3749 be fully fetched from start to finish regardless of the client
3750 actions. This affects bandwidth usage.
3753 NAME: minimum_expiry_time
3756 LOC: Config.minimum_expiry_time
3759 The minimum caching time according to (Expires - Date)
3760 Headers Squid honors if the object can't be revalidated
3761 defaults to 60 seconds. In reverse proxy environments it
3762 might be desirable to honor shorter object lifetimes. It
3763 is most likely better to make your server return a
3764 meaningful Last-Modified header however. In ESI environments
3765 where page fragments often have short lifetimes, this will
3766 often be best set to 0.
3769 NAME: store_avg_object_size
3773 LOC: Config.Store.avgObjectSize
3775 Average object size, used to estimate number of objects your
3776 cache can hold. The default is 13 KB.
3779 NAME: store_objects_per_bucket
3782 LOC: Config.Store.objectsPerBucket
3784 Target number of objects per bucket in the store hash table.
3785 Lowering this value increases the total number of buckets and
3786 also the storage maintenance rate. The default is 20.
3791 -----------------------------------------------------------------------------
3794 NAME: request_header_max_size
3798 LOC: Config.maxRequestHeaderSize
3800 This specifies the maximum size for HTTP headers in a request.
3801 Request headers are usually relatively small (about 512 bytes).
3802 Placing a limit on the request header size will catch certain
3803 bugs (for example with persistent connections) and possibly
3804 buffer-overflow or denial-of-service attacks.
3807 NAME: reply_header_max_size
3811 LOC: Config.maxReplyHeaderSize
3813 This specifies the maximum size for HTTP headers in a reply.
3814 Reply headers are usually relatively small (about 512 bytes).
3815 Placing a limit on the reply header size will catch certain
3816 bugs (for example with persistent connections) and possibly
3817 buffer-overflow or denial-of-service attacks.
3820 NAME: request_body_max_size
3824 LOC: Config.maxRequestBodySize
3826 This specifies the maximum size for an HTTP request body.
3827 In other words, the maximum size of a PUT/POST request.
3828 A user who attempts to send a request with a body larger
3829 than this limit receives an "Invalid Request" error message.
3830 If you set this parameter to a zero (the default), there will
3831 be no limit imposed.
3834 NAME: chunked_request_body_max_size
3838 LOC: Config.maxChunkedRequestBodySize
3840 A broken or confused HTTP/1.1 client may send a chunked HTTP
3841 request to Squid. Squid does not have full support for that
3842 feature yet. To cope with such requests, Squid buffers the
3843 entire request and then dechunks request body to create a
3844 plain HTTP/1.0 request with a known content length. The plain
3845 request is then used by the rest of Squid code as usual.
3847 The option value specifies the maximum size of the buffer used
3848 to hold the request before the conversion. If the chunked
3849 request size exceeds the specified limit, the conversion
3850 fails, and the client receives an "unsupported request" error,
3851 as if dechunking was disabled.
3853 Dechunking is enabled by default. To disable conversion of
3854 chunked requests, set the maximum to zero.
3856 Request dechunking feature and this option in particular are a
3857 temporary hack. When chunking requests and responses are fully
3858 supported, there will be no need to buffer a chunked request.
3862 IFDEF: HTTP_VIOLATIONS
3865 LOC: Config.accessList.brokenPosts
3867 A list of ACL elements which, if matched, causes Squid to send
3868 an extra CRLF pair after the body of a PUT/POST request.
3870 Some HTTP servers has broken implementations of PUT/POST,
3871 and rely on an extra CRLF pair sent by some WWW clients.
3873 Quote from RFC2616 section 4.1 on this matter:
3875 Note: certain buggy HTTP/1.0 client implementations generate an
3876 extra CRLF's after a POST request. To restate what is explicitly
3877 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
3878 a request with an extra CRLF.
3880 This clause only supports fast acl types.
3881 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
3884 acl buggy_server url_regex ^http://....
3885 broken_posts allow buggy_server
3888 NAME: icap_uses_indirect_client
3891 IFDEF: FOLLOW_X_FORWARDED_FOR&&ICAP_CLIENT
3893 LOC: Adaptation::Icap::TheConfig.icap_uses_indirect_client
3895 Controls whether the indirect client address
3896 (see follow_x_forwarded_for) instead of the
3897 direct client address is passed to an ICAP
3898 server as "X-Client-IP".
3902 IFDEF: HTTP_VIOLATIONS
3906 LOC: Config.onoff.via
3908 If set (default), Squid will include a Via header in requests and
3909 replies as required by RFC2616.
3915 LOC: Config.onoff.ie_refresh
3918 Microsoft Internet Explorer up until version 5.5 Service
3919 Pack 1 has an issue with transparent proxies, wherein it
3920 is impossible to force a refresh. Turning this on provides
3921 a partial fix to the problem, by causing all IMS-REFRESH
3922 requests from older IE versions to check the origin server
3923 for fresh content. This reduces hit ratio by some amount
3924 (~10% in my experience), but allows users to actually get
3925 fresh content when they want it. Note because Squid
3926 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
3927 of 5.5 is unchanged from old versions of Squid (i.e. a
3928 forced refresh is impossible). Newer versions of IE will,
3929 hopefully, continue to have the new behavior and will be
3930 handled based on that assumption. This option defaults to
3931 the old Squid behavior, which is better for hit ratios but
3932 worse for clients using IE, if they need to be able to
3933 force fresh content.
3936 NAME: vary_ignore_expire
3939 LOC: Config.onoff.vary_ignore_expire
3942 Many HTTP servers supporting Vary gives such objects
3943 immediate expiry time with no cache-control header
3944 when requested by a HTTP/1.0 client. This option
3945 enables Squid to ignore such expiry times until
3946 HTTP/1.1 is fully implemented.
3948 WARNING: If turned on this may eventually cause some
3949 varying objects not intended for caching to get cached.
3952 NAME: request_entities
3954 LOC: Config.onoff.request_entities
3957 Squid defaults to deny GET and HEAD requests with request entities,
3958 as the meaning of such requests are undefined in the HTTP standard
3959 even if not explicitly forbidden.
3961 Set this directive to on if you have clients which insists
3962 on sending request entities in GET or HEAD requests. But be warned
3963 that there is server software (both proxies and web servers) which
3964 can fail to properly process this kind of request which may make you
3965 vulnerable to cache pollution attacks if enabled.
3968 NAME: request_header_access
3969 IFDEF: HTTP_VIOLATIONS
3970 TYPE: http_header_access[]
3971 LOC: Config.request_header_access
3974 Usage: request_header_access header_name allow|deny [!]aclname ...
3976 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3977 this feature could make you liable for problems which it
3980 This option replaces the old 'anonymize_headers' and the
3981 older 'http_anonymizer' option with something that is much
3982 more configurable. This new method creates a list of ACLs
3983 for each header, allowing you very fine-tuned header
3986 This option only applies to request headers, i.e., from the
3987 client to the server.
3989 You can only specify known headers for the header name.
3990 Other headers are reclassified as 'Other'. You can also
3991 refer to all the headers with 'All'.
3993 For example, to achieve the same behavior as the old
3994 'http_anonymizer standard' option, you should use:
3996 request_header_access From deny all
3997 request_header_access Referer deny all
3998 request_header_access Server deny all
3999 request_header_access User-Agent deny all
4000 request_header_access WWW-Authenticate deny all
4001 request_header_access Link deny all
4003 Or, to reproduce the old 'http_anonymizer paranoid' feature
4006 request_header_access Allow allow all
4007 request_header_access Authorization allow all
4008 request_header_access WWW-Authenticate allow all
4009 request_header_access Proxy-Authorization allow all
4010 request_header_access Proxy-Authenticate allow all
4011 request_header_access Cache-Control allow all
4012 request_header_access Content-Encoding allow all
4013 request_header_access Content-Length allow all
4014 request_header_access Content-Type allow all
4015 request_header_access Date allow all
4016 request_header_access Expires allow all
4017 request_header_access Host allow all
4018 request_header_access If-Modified-Since allow all
4019 request_header_access Last-Modified allow all
4020 request_header_access Location allow all
4021 request_header_access Pragma allow all
4022 request_header_access Accept allow all
4023 request_header_access Accept-Charset allow all
4024 request_header_access Accept-Encoding allow all
4025 request_header_access Accept-Language allow all
4026 request_header_access Content-Language allow all
4027 request_header_access Mime-Version allow all
4028 request_header_access Retry-After allow all
4029 request_header_access Title allow all
4030 request_header_access Connection allow all
4031 request_header_access Proxy-Connection allow all
4032 request_header_access All deny all
4034 although many of those are HTTP reply headers, and so should be
4035 controlled with the reply_header_access directive.
4037 By default, all headers are allowed (no anonymizing is
4041 NAME: reply_header_access
4042 IFDEF: HTTP_VIOLATIONS
4043 TYPE: http_header_access[]
4044 LOC: Config.reply_header_access
4047 Usage: reply_header_access header_name allow|deny [!]aclname ...
4049 WARNING: Doing this VIOLATES the HTTP standard. Enabling
4050 this feature could make you liable for problems which it
4053 This option only applies to reply headers, i.e., from the
4054 server to the client.
4056 This is the same as request_header_access, but in the other
4059 This option replaces the old 'anonymize_headers' and the
4060 older 'http_anonymizer' option with something that is much
4061 more configurable. This new method creates a list of ACLs
4062 for each header, allowing you very fine-tuned header
4065 You can only specify known headers for the header name.
4066 Other headers are reclassified as 'Other'. You can also
4067 refer to all the headers with 'All'.
4069 For example, to achieve the same behavior as the old
4070 'http_anonymizer standard' option, you should use:
4072 reply_header_access From deny all
4073 reply_header_access Referer deny all
4074 reply_header_access Server deny all
4075 reply_header_access User-Agent deny all
4076 reply_header_access WWW-Authenticate deny all
4077 reply_header_access Link deny all
4079 Or, to reproduce the old 'http_anonymizer paranoid' feature
4082 reply_header_access Allow allow all
4083 reply_header_access Authorization allow all
4084 reply_header_access WWW-Authenticate allow all
4085 reply_header_access Proxy-Authorization allow all
4086 reply_header_access Proxy-Authenticate allow all
4087 reply_header_access Cache-Control allow all
4088 reply_header_access Content-Encoding allow all
4089 reply_header_access Content-Length allow all
4090 reply_header_access Content-Type allow all
4091 reply_header_access Date allow all
4092 reply_header_access Expires allow all
4093 reply_header_access Host allow all
4094 reply_header_access If-Modified-Since allow all
4095 reply_header_access Last-Modified allow all
4096 reply_header_access Location allow all
4097 reply_header_access Pragma allow all
4098 reply_header_access Accept allow all
4099 reply_header_access Accept-Charset allow all
4100 reply_header_access Accept-Encoding allow all
4101 reply_header_access Accept-Language allow all
4102 reply_header_access Content-Language allow all
4103 reply_header_access Mime-Version allow all
4104 reply_header_access Retry-After allow all
4105 reply_header_access Title allow all
4106 reply_header_access Connection allow all
4107 reply_header_access Proxy-Connection allow all
4108 reply_header_access All deny all
4110 although the HTTP request headers won't be usefully controlled
4111 by this directive -- see request_header_access for details.
4113 By default, all headers are allowed (no anonymizing is
4117 NAME: header_replace
4118 IFDEF: HTTP_VIOLATIONS
4119 TYPE: http_header_replace[]
4120 LOC: Config.request_header_access
4123 Usage: header_replace header_name message
4124 Example: header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
4126 This option allows you to change the contents of headers
4127 denied with header_access above, by replacing them with
4128 some fixed string. This replaces the old fake_user_agent
4131 This only applies to request headers, not reply headers.
4133 By default, headers are removed if denied.
4136 NAME: relaxed_header_parser
4137 COMMENT: on|off|warn
4139 LOC: Config.onoff.relaxed_header_parser
4142 In the default "on" setting Squid accepts certain forms
4143 of non-compliant HTTP messages where it is unambiguous
4144 what the sending application intended even if the message
4145 is not correctly formatted. The messages is then normalized
4146 to the correct form when forwarded by Squid.
4148 If set to "warn" then a warning will be emitted in cache.log
4149 each time such HTTP error is encountered.
4151 If set to "off" then such HTTP errors will cause the request
4152 or response to be rejected.
4155 NAME: ignore_expect_100
4157 IFDEF: HTTP_VIOLATIONS
4159 LOC: Config.onoff.ignore_expect_100
4162 This option makes Squid ignore any Expect: 100-continue header present
4163 in the request. RFC 2616 requires that Squid being unable to satisfy
4164 the response expectation MUST return a 417 error.
4166 Note: Enabling this is a HTTP protocol violation, but some clients may
4167 not handle it well..
4172 -----------------------------------------------------------------------------
4175 NAME: forward_timeout
4178 LOC: Config.Timeout.forward
4181 This parameter specifies how long Squid should at most attempt in
4182 finding a forwarding path for the request before giving up.
4185 NAME: connect_timeout
4188 LOC: Config.Timeout.connect
4191 This parameter specifies how long to wait for the TCP connect to
4192 the requested server or peer to complete before Squid should
4193 attempt to find another path where to forward the request.
4196 NAME: peer_connect_timeout
4199 LOC: Config.Timeout.peer_connect
4202 This parameter specifies how long to wait for a pending TCP
4203 connection to a peer cache. The default is 30 seconds. You
4204 may also set different timeout values for individual neighbors
4205 with the 'connect-timeout' option on a 'cache_peer' line.
4211 LOC: Config.Timeout.read
4214 The read_timeout is applied on server-side connections. After
4215 each successful read(), the timeout will be extended by this
4216 amount. If no data is read again after this amount of time,
4217 the request is aborted and logged with ERR_READ_TIMEOUT. The
4218 default is 15 minutes.
4224 LOC: Config.Timeout.write
4227 This timeout is tracked for all connections that have data
4228 available for writing and are waiting for the socket to become
4229 ready. After each successful write, the timeout is extended by
4230 the configured amount. If Squid has data to write but the
4231 connection is not ready for the configured duration, the
4232 transaction associated with the connection is terminated. The
4233 default is 15 minutes.
4236 NAME: request_timeout
4238 LOC: Config.Timeout.request
4241 How long to wait for an HTTP request after initial
4242 connection establishment.
4245 NAME: persistent_request_timeout
4247 LOC: Config.Timeout.persistent_request
4250 How long to wait for the next HTTP request on a persistent
4251 connection after the previous request completes.
4254 NAME: client_lifetime
4257 LOC: Config.Timeout.lifetime
4260 The maximum amount of time a client (browser) is allowed to
4261 remain connected to the cache process. This protects the Cache
4262 from having a lot of sockets (and hence file descriptors) tied up
4263 in a CLOSE_WAIT state from remote clients that go away without
4264 properly shutting down (either because of a network failure or
4265 because of a poor client implementation). The default is one
4268 NOTE: The default value is intended to be much larger than any
4269 client would ever need to be connected to your cache. You
4270 should probably change client_lifetime only as a last resort.
4271 If you seem to have many client connections tying up
4272 filedescriptors, we recommend first tuning the read_timeout,
4273 request_timeout, persistent_request_timeout and quick_abort values.
4276 NAME: half_closed_clients
4278 LOC: Config.onoff.half_closed_clients
4281 Some clients may shutdown the sending side of their TCP
4282 connections, while leaving their receiving sides open. Sometimes,
4283 Squid can not tell the difference between a half-closed and a
4284 fully-closed TCP connection.
4286 By default, Squid will immediately close client connections when
4287 read(2) returns "no more data to read."
4289 Change this option to 'on' and Squid will keep open connections
4290 until a read(2) or write(2) on the socket returns an error.
4291 This may show some benefits for reverse proxies. But if not
4292 it is recommended to leave OFF.
4297 LOC: Config.Timeout.pconn
4300 Timeout for idle persistent connections to servers and other
4307 LOC: Ident::TheConfig.timeout
4310 Maximum time to wait for IDENT lookups to complete.
4312 If this is too high, and you enabled IDENT lookups from untrusted
4313 users, you might be susceptible to denial-of-service by having
4314 many ident requests going at once.
4317 NAME: shutdown_lifetime
4320 LOC: Config.shutdownLifetime
4323 When SIGTERM or SIGHUP is received, the cache is put into
4324 "shutdown pending" mode until all active sockets are closed.
4325 This value is the lifetime to set for all open descriptors
4326 during shutdown mode. Any active clients after this many
4327 seconds will receive a 'timeout' message.
4331 ADMINISTRATIVE PARAMETERS
4332 -----------------------------------------------------------------------------
4338 LOC: Config.adminEmail
4340 Email-address of local cache manager who will receive
4341 mail if the cache dies. The default is "webmaster."
4347 LOC: Config.EmailFrom
4349 From: email-address for mail sent when the cache dies.
4350 The default is to use 'appname@unique_hostname'.
4351 Default appname value is "squid", can be changed into
4352 src/globals.h before building squid.
4358 LOC: Config.EmailProgram
4360 Email program used to send mail if the cache dies.
4361 The default is "mail". The specified program must comply
4362 with the standard Unix mail syntax:
4363 mail-program recipient < mailfile
4365 Optional command line options can be specified.
4368 NAME: cache_effective_user
4370 DEFAULT: @DEFAULT_CACHE_EFFECTIVE_USER@
4371 LOC: Config.effectiveUser
4373 If you start Squid as root, it will change its effective/real
4374 UID/GID to the user specified below. The default is to change
4375 to UID of @DEFAULT_CACHE_EFFECTIVE_USER@.
4376 see also; cache_effective_group
4379 NAME: cache_effective_group
4382 LOC: Config.effectiveGroup
4384 Squid sets the GID to the effective user's default group ID
4385 (taken from the password file) and supplementary group list
4386 from the groups membership.
4388 If you want Squid to run with a specific GID regardless of
4389 the group memberships of the effective user then set this
4390 to the group (or GID) you want Squid to run as. When set
4391 all other group privileges of the effective user are ignored
4392 and only this GID is effective. If Squid is not started as
4393 root the user starting Squid MUST be member of the specified
4396 This option is not recommended by the Squid Team.
4397 Our preference is for administrators to configure a secure
4398 user account for squid with UID/GID matching system policies.
4401 NAME: httpd_suppress_version_string
4405 LOC: Config.onoff.httpd_suppress_version_string
4407 Suppress Squid version string info in HTTP headers and HTML error pages.
4410 NAME: visible_hostname
4412 LOC: Config.visibleHostname
4415 If you want to present a special hostname in error messages, etc,
4416 define this. Otherwise, the return value of gethostname()
4417 will be used. If you have multiple caches in a cluster and
4418 get errors about IP-forwarding you must set them to have individual
4419 names with this setting.
4422 NAME: unique_hostname
4424 LOC: Config.uniqueHostname
4427 If you want to have multiple machines with the same
4428 'visible_hostname' you must give each machine a different
4429 'unique_hostname' so forwarding loops can be detected.
4432 NAME: hostname_aliases
4434 LOC: Config.hostnameAliases
4437 A list of other DNS names your cache has.
4445 Minimum umask which should be enforced while the proxy
4446 is running, in addition to the umask set at startup.
4448 For a traditional octal representation of umasks, start
4453 OPTIONS FOR THE CACHE REGISTRATION SERVICE
4454 -----------------------------------------------------------------------------
4456 This section contains parameters for the (optional) cache
4457 announcement service. This service is provided to help
4458 cache administrators locate one another in order to join or
4459 create cache hierarchies.
4461 An 'announcement' message is sent (via UDP) to the registration
4462 service by Squid. By default, the announcement message is NOT
4463 SENT unless you enable it with 'announce_period' below.
4465 The announcement message includes your hostname, plus the
4466 following information from this configuration file:
4472 All current information is processed regularly and made
4473 available on the Web at http://www.ircache.net/Cache/Tracker/.
4476 NAME: announce_period
4478 LOC: Config.Announce.period
4481 This is how frequently to send cache announcements. The
4482 default is `0' which disables sending the announcement
4485 To enable announcing your cache, just set an announce period.
4488 announce_period 1 day
4493 DEFAULT: tracker.ircache.net
4494 LOC: Config.Announce.host
4500 LOC: Config.Announce.file
4506 LOC: Config.Announce.port
4508 announce_host and announce_port set the hostname and port
4509 number where the registration message will be sent.
4511 Hostname will default to 'tracker.ircache.net' and port will
4512 default default to 3131. If the 'filename' argument is given,
4513 the contents of that file will be included in the announce
4518 HTTPD-ACCELERATOR OPTIONS
4519 -----------------------------------------------------------------------------
4522 NAME: httpd_accel_surrogate_id
4525 LOC: Config.Accel.surrogate_id
4527 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
4528 need an identification token to allow control targeting. Because
4529 a farm of surrogates may all perform the same tasks, they may share
4530 an identification token.
4532 The default ID is the visible_hostname
4535 NAME: http_accel_surrogate_remote
4539 LOC: Config.onoff.surrogate_is_remote
4541 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
4542 Set this to on to have squid behave as a remote surrogate.
4546 IFDEF: USE_SQUID_ESI
4547 COMMENT: libxml2|expat|custom
4549 LOC: ESIParser::Type
4552 ESI markup is not strictly XML compatible. The custom ESI parser
4553 will give higher performance, but cannot handle non ASCII character
4558 DELAY POOL PARAMETERS
4559 -----------------------------------------------------------------------------
4563 TYPE: delay_pool_count
4568 This represents the number of delay pools to be used. For example,
4569 if you have one class 2 delay pool and one class 3 delays pool, you
4570 have a total of 2 delay pools.
4574 TYPE: delay_pool_class
4579 This defines the class of each delay pool. There must be exactly one
4580 delay_class line for each delay pool. For example, to define two
4581 delay pools, one of class 2 and one of class 3, the settings above
4585 delay_pools 4 # 4 delay pools
4586 delay_class 1 2 # pool 1 is a class 2 pool
4587 delay_class 2 3 # pool 2 is a class 3 pool
4588 delay_class 3 4 # pool 3 is a class 4 pool
4589 delay_class 4 5 # pool 4 is a class 5 pool
4591 The delay pool classes are:
4593 class 1 Everything is limited by a single aggregate
4596 class 2 Everything is limited by a single aggregate
4597 bucket as well as an "individual" bucket chosen
4598 from bits 25 through 32 of the IPv4 address.
4600 class 3 Everything is limited by a single aggregate
4601 bucket as well as a "network" bucket chosen
4602 from bits 17 through 24 of the IP address and a
4603 "individual" bucket chosen from bits 17 through
4604 32 of the IPv4 address.
4606 class 4 Everything in a class 3 delay pool, with an
4607 additional limit on a per user basis. This
4608 only takes effect if the username is established
4609 in advance - by forcing authentication in your
4612 class 5 Requests are grouped according their tag (see
4613 external_acl's tag= reply).
4615 NOTE: If an IP address is a.b.c.d
4616 -> bits 25 through 32 are "d"
4617 -> bits 17 through 24 are "c"
4618 -> bits 17 through 32 are "c * 256 + d"
4620 NOTE-2: Due to the use of bitmasks in class 2,3,4 pools they only apply to
4621 IPv4 traffic. Class 1 and 5 pools may be used with IPv6 traffic.
4625 TYPE: delay_pool_access
4630 This is used to determine which delay pool a request falls into.
4632 delay_access is sorted per pool and the matching starts with pool 1,
4633 then pool 2, ..., and finally pool N. The first delay pool where the
4634 request is allowed is selected for the request. If it does not allow
4635 the request to any pool then the request is not delayed (default).
4637 For example, if you want some_big_clients in delay
4638 pool 1 and lotsa_little_clients in delay pool 2:
4641 delay_access 1 allow some_big_clients
4642 delay_access 1 deny all
4643 delay_access 2 allow lotsa_little_clients
4644 delay_access 2 deny all
4645 delay_access 3 allow authenticated_clients
4648 NAME: delay_parameters
4649 TYPE: delay_pool_rates
4654 This defines the parameters for a delay pool. Each delay pool has
4655 a number of "buckets" associated with it, as explained in the
4656 description of delay_class. For a class 1 delay pool, the syntax is:
4658 delay_parameters pool aggregate
4660 For a class 2 delay pool:
4662 delay_parameters pool aggregate individual
4664 For a class 3 delay pool:
4666 delay_parameters pool aggregate network individual
4668 For a class 4 delay pool:
4670 delay_parameters pool aggregate network individual user
4672 For a class 5 delay pool:
4674 delay_parameters pool tag
4676 The variables here are:
4678 pool a pool number - ie, a number between 1 and the
4679 number specified in delay_pools as used in
4682 aggregate the "delay parameters" for the aggregate bucket
4685 individual the "delay parameters" for the individual
4686 buckets (class 2, 3).
4688 network the "delay parameters" for the network buckets
4691 user the delay parameters for the user buckets
4694 tag the delay parameters for the tag buckets
4697 A pair of delay parameters is written restore/maximum, where restore is
4698 the number of bytes (not bits - modem and network speeds are usually
4699 quoted in bits) per second placed into the bucket, and maximum is the
4700 maximum number of bytes which can be in the bucket at any time.
4702 For example, if delay pool number 1 is a class 2 delay pool as in the
4703 above example, and is being used to strictly limit each host to 64kbps
4704 (plus overheads), with no overall limit, the line is:
4706 delay_parameters 1 -1/-1 8000/8000
4708 Note that the figure -1 is used to represent "unlimited".
4710 And, if delay pool number 2 is a class 3 delay pool as in the above
4711 example, and you want to limit it to a total of 256kbps (strict limit)
4712 with each 8-bit network permitted 64kbps (strict limit) and each
4713 individual host permitted 4800bps with a bucket maximum size of 64kb
4714 to permit a decent web page to be downloaded at a decent speed
4715 (if the network is not being limited due to overuse) but slow down
4716 large downloads more significantly:
4718 delay_parameters 2 32000/32000 8000/8000 600/8000
4720 There must be one delay_parameters line for each delay pool.
4722 Finally, for a class 4 delay pool as in the example - each user will
4723 be limited to 128Kb no matter how many workstations they are logged into.:
4725 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
4728 NAME: delay_initial_bucket_level
4729 COMMENT: (percent, 0-100)
4733 LOC: Config.Delay.initial
4735 The initial bucket percentage is used to determine how much is put
4736 in each bucket when squid starts, is reconfigured, or first notices
4737 a host accessing it (in class 2 and class 3, individual hosts and
4738 networks only have buckets associated with them once they have been
4743 WCCPv1 AND WCCPv2 CONFIGURATION OPTIONS
4744 -----------------------------------------------------------------------------
4749 LOC: Config.Wccp.router
4753 Use this option to define your WCCP ``home'' router for
4756 wccp_router supports a single WCCP(v1) router
4758 wccp2_router supports multiple WCCPv2 routers
4760 only one of the two may be used at the same time and defines
4761 which version of WCCP to use.
4765 TYPE: IpAddress_list
4766 LOC: Config.Wccp2.router
4770 Use this option to define your WCCP ``home'' router for
4773 wccp_router supports a single WCCP(v1) router
4775 wccp2_router supports multiple WCCPv2 routers
4777 only one of the two may be used at the same time and defines
4778 which version of WCCP to use.
4783 LOC: Config.Wccp.version
4787 This directive is only relevant if you need to set up WCCP(v1)
4788 to some very old and end-of-life Cisco routers. In all other
4789 setups it must be left unset or at the default setting.
4790 It defines an internal version in the WCCP(v1) protocol,
4791 with version 4 being the officially documented protocol.
4793 According to some users, Cisco IOS 11.2 and earlier only
4794 support WCCP version 3. If you're using that or an earlier
4795 version of IOS, you may need to change this value to 3, otherwise
4796 do not specify this parameter.
4799 NAME: wccp2_rebuild_wait
4801 LOC: Config.Wccp2.rebuildwait
4805 If this is enabled Squid will wait for the cache dir rebuild to finish
4806 before sending the first wccp2 HereIAm packet
4809 NAME: wccp2_forwarding_method
4811 LOC: Config.Wccp2.forwarding_method
4815 WCCP2 allows the setting of forwarding methods between the
4816 router/switch and the cache. Valid values are as follows:
4818 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4819 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4821 Currently (as of IOS 12.4) cisco routers only support GRE.
4822 Cisco switches only support the L2 redirect assignment method.
4825 NAME: wccp2_return_method
4827 LOC: Config.Wccp2.return_method
4831 WCCP2 allows the setting of return methods between the
4832 router/switch and the cache for packets that the cache
4833 decides not to handle. Valid values are as follows:
4835 gre - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4836 l2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4838 Currently (as of IOS 12.4) cisco routers only support GRE.
4839 Cisco switches only support the L2 redirect assignment.
4841 If the "ip wccp redirect exclude in" command has been
4842 enabled on the cache interface, then it is still safe for
4843 the proxy server to use a l2 redirect method even if this
4844 option is set to GRE.
4847 NAME: wccp2_assignment_method
4849 LOC: Config.Wccp2.assignment_method
4853 WCCP2 allows the setting of methods to assign the WCCP hash
4854 Valid values are as follows:
4856 hash - Hash assignment
4857 mask - Mask assignment
4859 As a general rule, cisco routers support the hash assignment method
4860 and cisco switches support the mask assignment method.
4865 LOC: Config.Wccp2.info
4867 DEFAULT_IF_NONE: standard 0
4870 WCCP2 allows for multiple traffic services. There are two
4871 types: "standard" and "dynamic". The standard type defines
4872 one service id - http (id 0). The dynamic service ids can be from
4873 51 to 255 inclusive. In order to use a dynamic service id
4874 one must define the type of traffic to be redirected; this is done
4875 using the wccp2_service_info option.
4877 The "standard" type does not require a wccp2_service_info option,
4878 just specifying the service id will suffice.
4880 MD5 service authentication can be enabled by adding
4881 "password=<password>" to the end of this service declaration.
4885 wccp2_service standard 0 # for the 'web-cache' standard service
4886 wccp2_service dynamic 80 # a dynamic service type which will be
4887 # fleshed out with subsequent options.
4888 wccp2_service standard 0 password=foo
4891 NAME: wccp2_service_info
4892 TYPE: wccp2_service_info
4893 LOC: Config.Wccp2.info
4897 Dynamic WCCPv2 services require further information to define the
4898 traffic you wish to have diverted.
4902 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
4903 priority=<priority> ports=<port>,<port>..
4905 The relevant WCCPv2 flags:
4906 + src_ip_hash, dst_ip_hash
4907 + source_port_hash, dst_port_hash
4908 + src_ip_alt_hash, dst_ip_alt_hash
4909 + src_port_alt_hash, dst_port_alt_hash
4912 The port list can be one to eight entries.
4916 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
4917 priority=240 ports=80
4919 Note: the service id must have been defined by a previous
4920 'wccp2_service dynamic <id>' entry.
4925 LOC: Config.Wccp2.weight
4929 Each cache server gets assigned a set of the destination
4930 hash proportional to their weight.
4935 LOC: Config.Wccp.address
4942 LOC: Config.Wccp2.address
4946 Use this option if you require WCCP to use a specific
4949 The default behavior is to not bind to any specific address.
4953 PERSISTENT CONNECTION HANDLING
4954 -----------------------------------------------------------------------------
4956 Also see "pconn_timeout" in the TIMEOUTS section
4959 NAME: client_persistent_connections
4961 LOC: Config.onoff.client_pconns
4965 NAME: server_persistent_connections
4967 LOC: Config.onoff.server_pconns
4970 Persistent connection support for clients and servers. By
4971 default, Squid uses persistent connections (when allowed)
4972 with its clients and servers. You can use these options to
4973 disable persistent connections with clients and/or servers.
4976 NAME: persistent_connection_after_error
4978 LOC: Config.onoff.error_pconns
4981 With this directive the use of persistent connections after
4982 HTTP errors can be disabled. Useful if you have clients
4983 who fail to handle errors on persistent connections proper.
4986 NAME: detect_broken_pconn
4988 LOC: Config.onoff.detect_broken_server_pconns
4991 Some servers have been found to incorrectly signal the use
4992 of HTTP/1.0 persistent connections even on replies not
4993 compatible, causing significant delays. This server problem
4994 has mostly been seen on redirects.
4996 By enabling this directive Squid attempts to detect such
4997 broken replies and automatically assume the reply is finished
4998 after 10 seconds timeout.
5002 CACHE DIGEST OPTIONS
5003 -----------------------------------------------------------------------------
5006 NAME: digest_generation
5007 IFDEF: USE_CACHE_DIGESTS
5009 LOC: Config.onoff.digest_generation
5012 This controls whether the server will generate a Cache Digest
5013 of its contents. By default, Cache Digest generation is
5014 enabled if Squid is compiled with --enable-cache-digests defined.
5017 NAME: digest_bits_per_entry
5018 IFDEF: USE_CACHE_DIGESTS
5020 LOC: Config.digest.bits_per_entry
5023 This is the number of bits of the server's Cache Digest which
5024 will be associated with the Digest entry for a given HTTP
5025 Method and URL (public key) combination. The default is 5.
5028 NAME: digest_rebuild_period
5029 IFDEF: USE_CACHE_DIGESTS
5032 LOC: Config.digest.rebuild_period
5035 This is the wait time between Cache Digest rebuilds.
5038 NAME: digest_rewrite_period
5040 IFDEF: USE_CACHE_DIGESTS
5042 LOC: Config.digest.rewrite_period
5045 This is the wait time between Cache Digest writes to
5049 NAME: digest_swapout_chunk_size
5052 IFDEF: USE_CACHE_DIGESTS
5053 LOC: Config.digest.swapout_chunk_size
5056 This is the number of bytes of the Cache Digest to write to
5057 disk at a time. It defaults to 4096 bytes (4KB), the Squid
5061 NAME: digest_rebuild_chunk_percentage
5062 COMMENT: (percent, 0-100)
5063 IFDEF: USE_CACHE_DIGESTS
5065 LOC: Config.digest.rebuild_chunk_percentage
5068 This is the percentage of the Cache Digest to be scanned at a
5069 time. By default it is set to 10% of the Cache Digest.
5074 -----------------------------------------------------------------------------
5079 LOC: Config.Port.snmp
5083 The port number where Squid listens for SNMP requests. To enable
5084 SNMP support set this to a suitable port number. Port number
5085 3401 is often used for the Squid SNMP agent. By default it's
5086 set to "0" (disabled)
5094 LOC: Config.accessList.snmp
5096 DEFAULT_IF_NONE: deny all
5099 Allowing or denying access to the SNMP port.
5101 All access to the agent is denied by default.
5104 snmp_access allow|deny [!]aclname ...
5106 This clause only supports fast acl types.
5107 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5109 snmp_access allow snmppublic localhost
5110 snmp_access deny all
5113 NAME: snmp_incoming_address
5115 LOC: Config.Addrs.snmp_incoming
5120 NAME: snmp_outgoing_address
5122 LOC: Config.Addrs.snmp_outgoing
5126 Just like 'udp_incoming_address', but for the SNMP port.
5128 snmp_incoming_address is used for the SNMP socket receiving
5129 messages from SNMP agents.
5130 snmp_outgoing_address is used for SNMP packets returned to SNMP
5133 The default snmp_incoming_address is to listen on all
5134 available network interfaces.
5136 If snmp_outgoing_address is not set it will use the same socket
5137 as snmp_incoming_address. Only change this if you want to have
5138 SNMP replies sent using another address than where this Squid
5139 listens for SNMP queries.
5141 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
5142 the same value since they both use port 3401.
5147 -----------------------------------------------------------------------------
5150 NAME: icp_port udp_port
5153 LOC: Config.Port.icp
5155 The port number where Squid sends and receives ICP queries to
5156 and from neighbor caches. The standard UDP port for ICP is 3130.
5157 Default is disabled (0).
5160 icp_port @DEFAULT_ICP_PORT@
5167 LOC: Config.Port.htcp
5169 The port number where Squid sends and receives HTCP queries to
5170 and from neighbor caches. To turn it on you want to set it to
5171 4827. By default it is set to "0" (disabled).
5177 NAME: log_icp_queries
5181 LOC: Config.onoff.log_udp
5183 If set, ICP queries are logged to access.log. You may wish
5184 do disable this if your ICP load is VERY high to speed things
5185 up or to simplify log analysis.
5188 NAME: udp_incoming_address
5190 LOC:Config.Addrs.udp_incoming
5193 udp_incoming_address is used for UDP packets received from other
5196 The default behavior is to not bind to any specific address.
5198 Only change this if you want to have all UDP queries received on
5199 a specific interface/address.
5201 NOTE: udp_incoming_address is used by the ICP, HTCP, and DNS
5202 modules. Altering it will affect all of them in the same manner.
5204 see also; udp_outgoing_address
5206 NOTE, udp_incoming_address and udp_outgoing_address can not
5207 have the same value since they both use the same port.
5210 NAME: udp_outgoing_address
5212 LOC: Config.Addrs.udp_outgoing
5215 udp_outgoing_address is used for UDP packets sent out to other
5218 The default behavior is to not bind to any specific address.
5220 Instead it will use the same socket as udp_incoming_address.
5221 Only change this if you want to have UDP queries sent using another
5222 address than where this Squid listens for UDP queries from other
5225 NOTE: udp_outgoing_address is used by the ICP, HTCP, and DNS
5226 modules. Altering it will affect all of them in the same manner.
5228 see also; udp_incoming_address
5230 NOTE, udp_incoming_address and udp_outgoing_address can not
5231 have the same value since they both use the same port.
5238 LOC: Config.onoff.icp_hit_stale
5240 If you want to return ICP_HIT for stale cache objects, set this
5241 option to 'on'. If you have sibling relationships with caches
5242 in other administrative domains, this should be 'off'. If you only
5243 have sibling relationships with caches under your control,
5244 it is probably okay to set this to 'on'.
5245 If set to 'on', your siblings should use the option "allow-miss"
5246 on their cache_peer lines for connecting to you.
5249 NAME: minimum_direct_hops
5252 LOC: Config.minDirectHops
5254 If using the ICMP pinging stuff, do direct fetches for sites
5255 which are no more than this many hops away.
5258 NAME: minimum_direct_rtt
5261 LOC: Config.minDirectRtt
5263 If using the ICMP pinging stuff, do direct fetches for sites
5264 which are no more than this many rtt milliseconds away.
5270 LOC: Config.Netdb.low
5276 LOC: Config.Netdb.high
5278 The low and high water marks for the ICMP measurement
5279 database. These are counts, not percents. The defaults are
5280 900 and 1000. When the high water mark is reached, database
5281 entries will be deleted until the low mark is reached.
5284 NAME: netdb_ping_period
5286 LOC: Config.Netdb.period
5289 The minimum period for measuring a site. There will be at
5290 least this much delay between successive pings to the same
5291 network. The default is five minutes.
5298 LOC: Config.onoff.query_icmp
5300 If you want to ask your peers to include ICMP data in their ICP
5301 replies, enable this option.
5303 If your peer has configured Squid (during compilation) with
5304 '--enable-icmp' that peer will send ICMP pings to origin server
5305 sites of the URLs it receives. If you enable this option the
5306 ICP replies from that peer will include the ICMP data (if available).
5307 Then, when choosing a parent cache, Squid will choose the parent with
5308 the minimal RTT to the origin server. When this happens, the
5309 hierarchy field of the access.log will be
5310 "CLOSEST_PARENT_MISS". This option is off by default.
5313 NAME: test_reachability
5317 LOC: Config.onoff.test_reachability
5319 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
5320 instead of ICP_MISS if the target host is NOT in the ICMP
5321 database, or has a zero RTT.
5324 NAME: icp_query_timeout
5328 LOC: Config.Timeout.icp_query
5330 Normally Squid will automatically determine an optimal ICP
5331 query timeout value based on the round-trip-time of recent ICP
5332 queries. If you want to override the value determined by
5333 Squid, set this 'icp_query_timeout' to a non-zero value. This
5334 value is specified in MILLISECONDS, so, to use a 2-second
5335 timeout (the old default), you would write:
5337 icp_query_timeout 2000
5340 NAME: maximum_icp_query_timeout
5344 LOC: Config.Timeout.icp_query_max
5346 Normally the ICP query timeout is determined dynamically. But
5347 sometimes it can lead to very large values (say 5 seconds).
5348 Use this option to put an upper limit on the dynamic timeout
5349 value. Do NOT use this option to always use a fixed (instead
5350 of a dynamic) timeout value. To set a fixed timeout see the
5351 'icp_query_timeout' directive.
5354 NAME: minimum_icp_query_timeout
5358 LOC: Config.Timeout.icp_query_min
5360 Normally the ICP query timeout is determined dynamically. But
5361 sometimes it can lead to very small timeouts, even lower than
5362 the normal latency variance on your link due to traffic.
5363 Use this option to put an lower limit on the dynamic timeout
5364 value. Do NOT use this option to always use a fixed (instead
5365 of a dynamic) timeout value. To set a fixed timeout see the
5366 'icp_query_timeout' directive.
5369 NAME: background_ping_rate
5373 LOC: Config.backgroundPingRate
5375 Controls how often the ICP pings are sent to siblings that
5376 have background-ping set.
5380 MULTICAST ICP OPTIONS
5381 -----------------------------------------------------------------------------
5386 LOC: Config.mcast_group_list
5389 This tag specifies a list of multicast groups which your server
5390 should join to receive multicasted ICP queries.
5392 NOTE! Be very careful what you put here! Be sure you
5393 understand the difference between an ICP _query_ and an ICP
5394 _reply_. This option is to be set only if you want to RECEIVE
5395 multicast queries. Do NOT set this option to SEND multicast
5396 ICP (use cache_peer for that). ICP replies are always sent via
5397 unicast, so this option does not affect whether or not you will
5398 receive replies from multicast group members.
5400 You must be very careful to NOT use a multicast address which
5401 is already in use by another group of caches.
5403 If you are unsure about multicast, please read the Multicast
5404 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
5406 Usage: mcast_groups 239.128.16.128 224.0.1.20
5408 By default, Squid doesn't listen on any multicast groups.
5411 NAME: mcast_miss_addr
5412 IFDEF: MULTICAST_MISS_STREAM
5414 LOC: Config.mcast_miss.addr
5417 If you enable this option, every "cache miss" URL will
5418 be sent out on the specified multicast address.
5420 Do not enable this option unless you are are absolutely
5421 certain you understand what you are doing.
5424 NAME: mcast_miss_ttl
5425 IFDEF: MULTICAST_MISS_STREAM
5427 LOC: Config.mcast_miss.ttl
5430 This is the time-to-live value for packets multicasted
5431 when multicasting off cache miss URLs is enabled. By
5432 default this is set to 'site scope', i.e. 16.
5435 NAME: mcast_miss_port
5436 IFDEF: MULTICAST_MISS_STREAM
5438 LOC: Config.mcast_miss.port
5441 This is the port number to be used in conjunction with
5445 NAME: mcast_miss_encode_key
5446 IFDEF: MULTICAST_MISS_STREAM
5448 LOC: Config.mcast_miss.encode_key
5449 DEFAULT: XXXXXXXXXXXXXXXX
5451 The URLs that are sent in the multicast miss stream are
5452 encrypted. This is the encryption key.
5455 NAME: mcast_icp_query_timeout
5459 LOC: Config.Timeout.mcast_icp_query
5461 For multicast peers, Squid regularly sends out ICP "probes" to
5462 count how many other peers are listening on the given multicast
5463 address. This value specifies how long Squid should wait to
5464 count all the replies. The default is 2000 msec, or 2
5469 INTERNAL ICON OPTIONS
5470 -----------------------------------------------------------------------------
5473 NAME: icon_directory
5475 LOC: Config.icons.directory
5476 DEFAULT: @DEFAULT_ICON_DIR@
5478 Where the icons are stored. These are normally kept in
5482 NAME: global_internal_static
5484 LOC: Config.onoff.global_internal_static
5487 This directive controls is Squid should intercept all requests for
5488 /squid-internal-static/ no matter which host the URL is requesting
5489 (default on setting), or if nothing special should be done for
5490 such URLs (off setting). The purpose of this directive is to make
5491 icons etc work better in complex cache hierarchies where it may
5492 not always be possible for all corners in the cache mesh to reach
5493 the server generating a directory listing.
5496 NAME: short_icon_urls
5498 LOC: Config.icons.use_short_names
5501 If this is enabled Squid will use short URLs for icons.
5502 If disabled it will revert to the old behavior of including
5503 it's own name and port in the URL.
5505 If you run a complex cache hierarchy with a mix of Squid and
5506 other proxies you may need to disable this directive.
5511 -----------------------------------------------------------------------------
5514 NAME: error_directory
5516 LOC: Config.errorDirectory
5519 If you wish to create your own versions of the default
5520 error files to customize them to suit your company copy
5521 the error/template files to another directory and point
5524 WARNING: This option will disable multi-language support
5525 on error pages if used.
5527 The squid developers are interested in making squid available in
5528 a wide variety of languages. If you are making translations for a
5529 language that Squid does not currently provide please consider
5530 contributing your translation back to the project.
5531 http://wiki.squid-cache.org/Translations
5533 The squid developers working on translations are happy to supply drop-in
5534 translated error files in exchange for any new language contributions.
5537 NAME: error_default_language
5538 IFDEF: USE_ERR_LOCALES
5540 LOC: Config.errorDefaultLanguage
5543 Set the default language which squid will send error pages in
5544 if no existing translation matches the clients language
5547 If unset (default) generic English will be used.
5549 The squid developers are interested in making squid available in
5550 a wide variety of languages. If you are interested in making
5551 translations for any language see the squid wiki for details.
5552 http://wiki.squid-cache.org/Translations
5555 NAME: error_log_languages
5556 IFDEF: USE_ERR_LOCALES
5558 LOC: Config.errorLogMissingLanguages
5561 Log to cache.log what languages users are attempting to
5562 auto-negotiate for translations.
5564 Successful negotiations are not logged. Only failures
5565 have meaning to indicate that Squid may need an upgrade
5566 of its error page translations.
5569 NAME: err_page_stylesheet
5571 LOC: Config.errorStylesheet
5572 DEFAULT: @DEFAULT_CONFIG_DIR@/errorpage.css
5574 CSS Stylesheet to pattern the display of Squid default error pages.
5576 For information on CSS see http://www.w3.org/Style/CSS/
5581 LOC: Config.errHtmlText
5584 HTML text to include in error messages. Make this a "mailto"
5585 URL to your admin address, or maybe just a link to your
5586 organizations Web page.
5588 To include this in your error messages, you must rewrite
5589 the error template files (found in the "errors" directory).
5590 Wherever you want the 'err_html_text' line to appear,
5591 insert a %L tag in the error template file.
5594 NAME: email_err_data
5597 LOC: Config.onoff.emailErrData
5600 If enabled, information about the occurred error will be
5601 included in the mailto links of the ERR pages (if %W is set)
5602 so that the email body contains the data.
5603 Syntax is <A HREF="mailto:%w%W">%w</A>
5608 LOC: Config.denyInfoList
5611 Usage: deny_info err_page_name acl
5612 or deny_info http://... acl
5613 or deny_info TCP_RESET acl
5615 This can be used to return a ERR_ page for requests which
5616 do not pass the 'http_access' rules. Squid remembers the last
5617 acl it evaluated in http_access, and if a 'deny_info' line exists
5618 for that ACL Squid returns a corresponding error page.
5620 The acl is typically the last acl on the http_access deny line which
5621 denied access. The exceptions to this rule are:
5622 - When Squid needs to request authentication credentials. It's then
5623 the first authentication related acl encountered
5624 - When none of the http_access lines matches. It's then the last
5625 acl processed on the last http_access line.
5627 NP: If providing your own custom error pages with error_directory
5628 you may also specify them by your custom file name:
5629 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
5631 Alternatively you can tell Squid to reset the TCP connection
5632 by specifying TCP_RESET.
5634 Or you can specify an error URL or URL pattern. The browsers will
5635 get redirected (302) to the specified URL after formattgin tags have
5639 %a - username (if available. Password NOT included)
5642 %E - Error description
5644 %H - Request domain name
5645 %i - Client IP Address
5647 %o - Message result from external ACL helper
5648 %p - Request Port number
5649 %P - Request Protocol name
5650 %R - Request URL path
5651 %T - Timestamp in RFC 1123 format
5652 %U - Full canonical URL from client
5653 (HTTPS URLs terminate with *)
5654 %u - Full canonical URL from client
5655 %w - Admin email from squid.conf
5656 %% - Literal percent (%) code
5661 OPTIONS INFLUENCING REQUEST FORWARDING
5662 -----------------------------------------------------------------------------
5665 NAME: nonhierarchical_direct
5667 LOC: Config.onoff.nonhierarchical_direct
5670 By default, Squid will send any non-hierarchical requests
5671 (matching hierarchy_stoplist or not cacheable request type) direct
5674 If you set this to off, Squid will prefer to send these
5675 requests to parents.
5677 Note that in most configurations, by turning this off you will only
5678 add latency to these request without any improvement in global hit
5681 If you are inside an firewall see never_direct instead of
5687 LOC: Config.onoff.prefer_direct
5690 Normally Squid tries to use parents for most requests. If you for some
5691 reason like it to first try going direct and only use a parent if
5692 going direct fails set this to on.
5694 By combining nonhierarchical_direct off and prefer_direct on you
5695 can set up Squid to use a parent as a backup path if going direct
5698 Note: If you want Squid to use parents for all requests see
5699 the never_direct directive. prefer_direct only modifies how Squid
5700 acts on cacheable requests.
5705 LOC: Config.accessList.AlwaysDirect
5708 Usage: always_direct allow|deny [!]aclname ...
5710 Here you can use ACL elements to specify requests which should
5711 ALWAYS be forwarded by Squid to the origin servers without using
5712 any peers. For example, to always directly forward requests for
5713 local servers ignoring any parents or siblings you may have use
5716 acl local-servers dstdomain my.domain.net
5717 always_direct allow local-servers
5719 To always forward FTP requests directly, use
5722 always_direct allow FTP
5724 NOTE: There is a similar, but opposite option named
5725 'never_direct'. You need to be aware that "always_direct deny
5726 foo" is NOT the same thing as "never_direct allow foo". You
5727 may need to use a deny rule to exclude a more-specific case of
5728 some other rule. Example:
5730 acl local-external dstdomain external.foo.net
5731 acl local-servers dstdomain .foo.net
5732 always_direct deny local-external
5733 always_direct allow local-servers
5735 NOTE: If your goal is to make the client forward the request
5736 directly to the origin server bypassing Squid then this needs
5737 to be done in the client configuration. Squid configuration
5738 can only tell Squid how Squid should fetch the object.
5740 NOTE: This directive is not related to caching. The replies
5741 is cached as usual even if you use always_direct. To not cache
5742 the replies see the 'cache' directive.
5744 This clause supports both fast and slow acl types.
5745 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5750 LOC: Config.accessList.NeverDirect
5753 Usage: never_direct allow|deny [!]aclname ...
5755 never_direct is the opposite of always_direct. Please read
5756 the description for always_direct if you have not already.
5758 With 'never_direct' you can use ACL elements to specify
5759 requests which should NEVER be forwarded directly to origin
5760 servers. For example, to force the use of a proxy for all
5761 requests, except those in your local domain use something like:
5763 acl local-servers dstdomain .foo.net
5764 never_direct deny local-servers
5765 never_direct allow all
5767 or if Squid is inside a firewall and there are local intranet
5768 servers inside the firewall use something like:
5770 acl local-intranet dstdomain .foo.net
5771 acl local-external dstdomain external.foo.net
5772 always_direct deny local-external
5773 always_direct allow local-intranet
5774 never_direct allow all
5776 This clause supports both fast and slow acl types.
5777 See http://wiki.squid-cache.org/SquidFaq/SquidAcl for details.
5781 ADVANCED NETWORKING OPTIONS
5782 -----------------------------------------------------------------------------
5785 NAME: incoming_icp_average
5788 LOC: Config.comm_incoming.icp_average
5791 NAME: incoming_http_average
5794 LOC: Config.comm_incoming.http_average
5797 NAME: incoming_dns_average
5800 LOC: Config.comm_incoming.dns_average
5803 NAME: min_icp_poll_cnt
5806 LOC: Config.comm_incoming.icp_min_poll
5809 NAME: min_dns_poll_cnt
5812 LOC: Config.comm_incoming.dns_min_poll
5815 NAME: min_http_poll_cnt
5818 LOC: Config.comm_incoming.http_min_poll
5820 Heavy voodoo here. I can't even believe you are reading this.
5821 Are you crazy? Don't even think about adjusting these unless
5822 you understand the algorithms in comm_select.c first!
5828 LOC: Config.accept_filter
5832 The name of an accept(2) filter to install on Squid's
5833 listen socket(s). This feature is perhaps specific to
5834 FreeBSD and requires support in the kernel.
5836 The 'httpready' filter delays delivering new connections
5837 to Squid until a full HTTP request has been received.
5838 See the accf_http(9) man page for details.
5840 The 'dataready' filter delays delivering new connections
5841 to Squid until there is some data to process.
5842 See the accf_dataready(9) man page for details.
5846 The 'data' filter delays delivering of new connections
5847 to Squid until there is some data to process by TCP_ACCEPT_DEFER.
5848 You may optionally specify a number of seconds to wait by
5849 'data=N' where N is the number of seconds. Defaults to 30
5850 if not specified. See the tcp(7) man page for details.
5853 accept_filter httpready
5858 NAME: client_ip_max_connections
5860 LOC: Config.client_ip_max_connections
5863 Set an absolute limit on the number of connections a single
5864 client IP can use. Any more than this and Squid will begin to drop
5865 new connections from the client until it closes some links.
5867 Note that this is a global limit. It affects all HTTP, HTCP, Gopher and FTP
5868 connections from the client. For finer control use the ACL access controls.
5870 Requires client_db to be enabled (the default).
5872 WARNING: This may noticably slow down traffic received via external proxies
5873 or NAT devices and cause them to rebound error messages back to their clients.
5876 NAME: tcp_recv_bufsize
5880 LOC: Config.tcpRcvBufsz
5882 Size of receive buffer to set for TCP sockets. Probably just
5883 as easy to change your kernel's default. Set to zero to use
5884 the default buffer size.
5889 -----------------------------------------------------------------------------
5896 LOC: Adaptation::Icap::TheConfig.onoff
5899 If you want to enable the ICAP module support, set this to on.
5902 NAME: icap_connect_timeout
5905 LOC: Adaptation::Icap::TheConfig.connect_timeout_raw
5908 This parameter specifies how long to wait for the TCP connect to
5909 the requested ICAP server to complete before giving up and either
5910 terminating the HTTP transaction or bypassing the failure.
5912 The default for optional services is peer_connect_timeout.
5913 The default for essential services is connect_timeout.
5914 If this option is explicitly set, its value applies to all services.
5917 NAME: icap_io_timeout
5921 LOC: Adaptation::Icap::TheConfig.io_timeout_raw
5924 This parameter specifies how long to wait for an I/O activity on
5925 an established, active ICAP connection before giving up and
5926 either terminating the HTTP transaction or bypassing the
5929 The default is read_timeout.
5932 NAME: icap_service_failure_limit
5933 COMMENT: limit [in memory-depth time-units]
5934 TYPE: icap_service_failure_limit
5936 LOC: Adaptation::Icap::TheConfig
5939 The limit specifies the number of failures that Squid tolerates
5940 when establishing a new TCP connection with an ICAP service. If
5941 the number of failures exceeds the limit, the ICAP service is
5942 not used for new ICAP requests until it is time to refresh its
5945 A negative value disables the limit. Without the limit, an ICAP
5946 service will not be considered down due to connectivity failures
5947 between ICAP OPTIONS requests.
5949 Squid forgets ICAP service failures older than the specified
5950 value of memory-depth. The memory fading algorithm
5951 is approximate because Squid does not remember individual
5952 errors but groups them instead, splitting the option
5953 value into ten time slots of equal length.
5955 When memory-depth is 0 and by default this option has no
5956 effect on service failure expiration.
5958 Squid always forgets failures when updating service settings
5959 using an ICAP OPTIONS transaction, regardless of this option
5963 # suspend service usage after 10 failures in 5 seconds:
5964 icap_service_failure_limit 10 in 5 seconds
5967 NAME: icap_service_revival_delay
5970 LOC: Adaptation::Icap::TheConfig.service_revival_delay
5973 The delay specifies the number of seconds to wait after an ICAP
5974 OPTIONS request failure before requesting the options again. The
5975 failed ICAP service is considered "down" until fresh OPTIONS are
5978 The actual delay cannot be smaller than the hardcoded minimum
5979 delay of 30 seconds.
5982 NAME: icap_preview_enable
5986 LOC: Adaptation::Icap::TheConfig.preview_enable
5989 The ICAP Preview feature allows the ICAP server to handle the
5990 HTTP message by looking only at the beginning of the message body
5991 or even without receiving the body at all. In some environments,
5992 previews greatly speedup ICAP processing.
5994 During an ICAP OPTIONS transaction, the server may tell Squid what
5995 HTTP messages should be previewed and how big the preview should be.
5996 Squid will not use Preview if the server did not request one.
5998 To disable ICAP Preview for all ICAP services, regardless of
5999 individual ICAP server OPTIONS responses, set this option to "off".
6001 icap_preview_enable off
6004 NAME: icap_preview_size
6007 LOC: Adaptation::Icap::TheConfig.preview_size
6010 The default size of preview data to be sent to the ICAP server.
6011 -1 means no preview. This value might be overwritten on a per server
6012 basis by OPTIONS requests.
6015 NAME: icap_206_enable
6019 LOC: Adaptation::Icap::TheConfig.allow206_enable
6022 206 (Partial Content) responses is an ICAP extension that allows the
6023 ICAP agents to optionally combine adapted and original HTTP message
6024 content. The decision to combine is postponed until the end of the
6025 ICAP response. Squid supports Partial Content extension by default.
6027 Activation of the Partial Content extension is negotiated with each
6028 ICAP service during OPTIONS exchange. Most ICAP servers should handle
6029 negotation correctly even if they do not support the extension, but
6030 some might fail. To disable Partial Content support for all ICAP
6031 services and to avoid any negotiation, set this option to "off".
6037 NAME: icap_default_options_ttl
6040 LOC: Adaptation::Icap::TheConfig.default_options_ttl
6043 The default TTL value for ICAP OPTIONS responses that don't have
6044 an Options-TTL header.
6047 NAME: icap_persistent_connections
6051 LOC: Adaptation::Icap::TheConfig.reuse_connections
6054 Whether or not Squid should use persistent connections to
6058 NAME: icap_send_client_ip
6062 LOC: Adaptation::Icap::TheConfig.send_client_ip
6065 This adds the header "X-Client-IP" to ICAP requests.
6068 NAME: icap_send_client_username
6072 LOC: Adaptation::Icap::TheConfig.send_client_username
6075 This sends authenticated HTTP client username (if available) to
6076 the ICAP service. The username value is encoded based on the
6077 icap_client_username_encode option and is sent using the header
6078 specified by the icap_client_username_header option.
6081 NAME: icap_client_username_header
6084 LOC: Adaptation::Icap::TheConfig.client_username_header
6085 DEFAULT: X-Client-Username
6087 ICAP request header name to use for send_client_username.
6090 NAME: icap_client_username_encode
6094 LOC: Adaptation::Icap::TheConfig.client_username_encode
6097 Whether to base64 encode the authenticated client username.
6101 TYPE: icap_service_type
6103 LOC: Adaptation::Icap::TheConfig
6106 Defines a single ICAP service using the following format:
6108 icap_service service_name vectoring_point [options] service_url
6111 an opaque identifier which must be unique in squid.conf
6113 vectoring_point: reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
6114 This specifies at which point of transaction processing the
6115 ICAP service should be activated. *_postcache vectoring points
6116 are not yet supported.
6118 service_url: icap://servername:port/servicepath
6119 ICAP server and service location.
6121 ICAP does not allow a single service to handle both REQMOD and RESPMOD
6122 transactions. Squid does not enforce that requirement. You can specify
6123 services with the same service_url and different vectoring_points. You
6124 can even specify multiple identical services as long as their
6125 service_names differ.
6128 Service options are separated by white space. ICAP services support
6129 the following name=value options:
6132 If set to 'on' or '1', the ICAP service is treated as
6133 optional. If the service cannot be reached or malfunctions,
6134 Squid will try to ignore any errors and process the message as
6135 if the service was not enabled. No all ICAP errors can be
6136 bypassed. If set to 0, the ICAP service is treated as
6137 essential and all ICAP errors will result in an error page
6138 returned to the HTTP client.
6140 Bypass is off by default: services are treated as essential.
6143 If set to 'on' or '1', the ICAP service is allowed to
6144 dynamically change the current message adaptation plan by
6145 returning a chain of services to be used next. The services
6146 are specified using the X-Next-Services ICAP response header
6147 value, formatted as a comma-separated list of service names.
6148 Each named service should be configured in squid.conf and
6149 should have the same method and vectoring point as the current
6150 ICAP transaction. Services violating these rules are ignored.
6151 An empty X-Next-Services value results in an empty plan which
6152 ends the current adaptation.
6154 Routing is not allowed by default: the ICAP X-Next-Services
6155 response header is ignored.
6157 Older icap_service format without optional named parameters is
6158 deprecated but supported for backward compatibility.
6161 icap_service svcBlocker reqmod_precache bypass=0 icap://icap1.mydomain.net:1344/reqmod
6162 icap_service svcLogger reqmod_precache routing=on icap://icap2.mydomain.net:1344/respmod
6166 TYPE: icap_class_type
6171 This deprecated option was documented to define an ICAP service
6172 chain, even though it actually defined a set of similar, redundant
6173 services, and the chains were not supported.
6175 To define a set of redundant services, please use the
6176 adaptation_service_set directive. For service chains, use
6177 adaptation_service_chain.
6181 TYPE: icap_access_type
6186 This option is deprecated. Please use adaptation_access, which
6187 has the same ICAP functionality, but comes with better
6188 documentation, and eCAP support.
6193 -----------------------------------------------------------------------------
6200 LOC: Adaptation::Ecap::TheConfig.onoff
6203 Controls whether eCAP support is enabled.
6207 TYPE: ecap_service_type
6209 LOC: Adaptation::Ecap::TheConfig
6212 Defines a single eCAP service
6214 ecap_service servicename vectoring_point bypass service_url
6216 vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
6217 This specifies at which point of transaction processing the
6218 eCAP service should be activated. *_postcache vectoring points
6219 are not yet supported.
6221 If set to 1, the eCAP service is treated as optional. If the
6222 service cannot be reached or malfunctions, Squid will try to
6223 ignore any errors and process the message as if the service
6224 was not enabled. No all eCAP errors can be bypassed.
6225 If set to 0, the eCAP service is treated as essential and all
6226 eCAP errors will result in an error page returned to the
6228 service_url = ecap://vendor/service_name?custom&cgi=style¶meters=optional
6231 ecap_service service_1 reqmod_precache 0 ecap://filters-R-us/leakDetector?on_error=block
6232 ecap_service service_2 respmod_precache 1 icap://filters-R-us/virusFilter?config=/etc/vf.cfg
6235 NAME: loadable_modules
6237 IFDEF: USE_LOADABLE_MODULES
6238 LOC: Config.loadable_module_names
6241 Instructs Squid to load the specified dynamic module(s) or activate
6242 preloaded module(s).
6244 loadable_modules @DEFAULT_PREFIX@/lib/MinimalAdapter.so
6248 MESSAGE ADAPTATION OPTIONS
6249 -----------------------------------------------------------------------------
6252 NAME: adaptation_service_set
6253 TYPE: adaptation_service_set_type
6254 IFDEF: USE_ADAPTATION
6259 Configures an ordered set of similar, redundant services. This is
6260 useful when hot standby or backup adaptation servers are available.
6262 adaptation_service_set set_name service_name1 service_name2 ...
6264 The named services are used in the set declaration order. The first
6265 applicable adaptation service from the set is used first. The next
6266 applicable service is tried if and only if the transaction with the
6267 previous service fails and the message waiting to be adapted is still
6270 When adaptation starts, broken services are ignored as if they were
6271 not a part of the set. A broken service is a down optional service.
6273 The services in a set must be attached to the same vectoring point
6274 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
6276 If all services in a set are optional then adaptation failures are
6277 bypassable. If all services in the set are essential, then a
6278 transaction failure with one service may still be retried using
6279 another service from the set, but when all services fail, the master
6280 transaction fails as well.
6282 A set may contain a mix of optional and essential services, but that
6283 is likely to lead to surprising results because broken services become
6284 ignored (see above), making previously bypassable failures fatal.
6285 Technically, it is the bypassability of the last failed service that
6288 See also: adaptation_access adaptation_service_chain
6291 adaptation_service_set svcBlocker urlFilterPrimary urlFilterBackup
6292 adaptation service_set svcLogger loggerLocal loggerRemote
6295 NAME: adaptation_service_chain
6296 TYPE: adaptation_service_chain_type
6297 IFDEF: USE_ADAPTATION
6302 Configures a list of complementary services that will be applied
6303 one-by-one, forming an adaptation chain or pipeline. This is useful
6304 when Squid must perform different adaptations on the same message.
6306 adaptation_service_chain chain_name service_name1 svc_name2 ...
6308 The named services are used in the chain declaration order. The first
6309 applicable adaptation service from the chain is used first. The next
6310 applicable service is applied to the successful adaptation results of
6311 the previous service in the chain.
6313 When adaptation starts, broken services are ignored as if they were
6314 not a part of the chain. A broken service is a down optional service.
6316 Request satisfaction terminates the adaptation chain because Squid
6317 does not currently allow declaration of RESPMOD services at the
6318 "reqmod_precache" vectoring point (see icap_service or ecap_service).
6320 The services in a chain must be attached to the same vectoring point
6321 (e.g., pre-cache) and use the same adaptation method (e.g., REQMOD).
6323 A chain may contain a mix of optional and essential services. If an
6324 essential adaptation fails (or the failure cannot be bypassed for
6325 other reasons), the master transaction fails. Otherwise, the failure
6326 is bypassed as if the failed adaptation service was not in the chain.
6328 See also: adaptation_access adaptation_service_set
6331 adaptation_service_chain svcRequest requestLogger urlFilter leakDetector
6334 NAME: adaptation_access
6335 TYPE: adaptation_access_type
6336 IFDEF: USE_ADAPTATION
6340 Sends an HTTP transaction to an ICAP or eCAP adaptation service.
6342 adaptation_access service_name allow|deny [!]aclname...
6343 adaptation_access set_name allow|deny [!]aclname...
6345 At each supported vectoring point, the adaptation_access
6346 statements are processed in the order they appear in this
6347 configuration file. Statements pointing to the following services
6348 are ignored (i.e., skipped without checking their ACL):
6350 - services serving different vectoring points
6351 - "broken-but-bypassable" services
6352 - "up" services configured to ignore such transactions
6353 (e.g., based on the ICAP Transfer-Ignore header).
6355 When a set_name is used, all services in the set are checked
6356 using the same rules, to find the first applicable one. See
6357 adaptation_service_set for details.
6359 If an access list is checked and there is a match, the
6360 processing stops: For an "allow" rule, the corresponding
6361 adaptation service is used for the transaction. For a "deny"
6362 rule, no adaptation service is activated.
6364 It is currently not possible to apply more than one adaptation
6365 service at the same vectoring point to the same HTTP transaction.
6367 See also: icap_service and ecap_service
6370 adaptation_access service_1 allow all
6373 NAME: adaptation_service_iteration_limit
6375 IFDEF: USE_ADAPTATION
6376 LOC: Adaptation::Config::service_iteration_limit
6379 Limits the number of iterations allowed when applying adaptation
6380 services to a message. If your longest adaptation set or chain
6381 may have more than 16 services, increase the limit beyond its
6382 default value of 16. If detecting infinite iteration loops sooner
6383 is critical, make the iteration limit match the actual number
6384 of services in your longest adaptation set or chain.
6386 Infinite adaptation loops are most likely with routing services.
6388 See also: icap_service routing=1
6391 NAME: adaptation_masterx_shared_names
6393 IFDEF: USE_ADAPTATION
6394 LOC: Adaptation::Config::masterx_shared_name
6397 For each master transaction (i.e., the HTTP request and response
6398 sequence, including all related ICAP and eCAP exchanges), Squid
6399 maintains a table of metadata. The table entries are (name, value)
6400 pairs shared among eCAP and ICAP exchanges. The table is destroyed
6401 with the master transaction.
6403 This option specifies the table entry names that Squid must accept
6404 from and forward to the adaptation transactions.
6406 An ICAP REQMOD or RESPMOD transaction may set an entry in the
6407 shared table by returning an ICAP header field with a name
6408 specified in adaptation_masterx_shared_names. Squid will store
6409 and forward that ICAP header field to subsequent ICAP
6410 transactions within the same master transaction scope.
6412 Only one shared entry name is supported at this time.
6415 # share authentication information among ICAP services
6416 adaptation_masterx_shared_names X-Subscriber-ID
6422 LOC: Adaptation::Icap::TheConfig.repeat
6424 DEFAULT_IF_NONE: deny all
6426 This ACL determines which retriable ICAP transactions are
6427 retried. Transactions that received a complete ICAP response
6428 and did not have to consume or produce HTTP bodies to receive
6429 that response are usually retriable.
6431 icap_retry allow|deny [!]aclname ...
6433 Squid automatically retries some ICAP I/O timeouts and errors
6434 due to persistent connection race conditions.
6436 See also: icap_retry_limit
6439 NAME: icap_retry_limit
6442 LOC: Adaptation::Icap::TheConfig.repeat_limit
6445 Limits the number of retries allowed. When set to zero (default),
6446 no retries are allowed.
6448 Communication errors due to persistent connection race
6449 conditions are unavoidable, automatically retried, and do not
6450 count against this limit.
6452 See also: icap_retry
6458 -----------------------------------------------------------------------------
6461 NAME: check_hostnames
6464 LOC: Config.onoff.check_hostnames
6466 For security and stability reasons Squid can check
6467 hostnames for Internet standard RFC compliance. If you want
6468 Squid to perform these checks turn this directive on.
6471 NAME: allow_underscore
6474 LOC: Config.onoff.allow_underscore
6476 Underscore characters is not strictly allowed in Internet hostnames
6477 but nevertheless used by many sites. Set this to off if you want
6478 Squid to be strict about the standard.
6479 This check is performed only when check_hostnames is set to on.
6482 NAME: cache_dns_program
6484 IFDEF: USE_DNSSERVERS
6485 DEFAULT: @DEFAULT_DNSSERVER@
6486 LOC: Config.Program.dnsserver
6488 Specify the location of the executable for dnslookup process.
6492 TYPE: HelperChildConfig
6493 IFDEF: USE_DNSSERVERS
6494 DEFAULT: 32 startup=1 idle=1
6495 LOC: Config.dnsChildren
6497 The maximum number of processes spawn to service DNS name lookups.
6498 If you limit it too few Squid will have to wait for them to process
6499 a backlog of requests, slowing it down. If you allow too many they
6500 will use RAM and other system resources noticably.
6501 The maximum this may be safely set to is 32.
6503 The startup= and idle= options allow some measure of skew in your
6508 Sets a minimum of how many processes are to be spawned when Squid
6509 starts or reconfigures. When set to zero the first request will
6510 cause spawning of the first child process to handle it.
6512 Starting too few will cause an initial slowdown in traffic as Squid
6513 attempts to simultaneously spawn enough processes to cope.
6517 Sets a minimum of how many processes Squid is to try and keep available
6518 at all times. When traffic begins to rise above what the existing
6519 processes can handle this many more will be spawned up to the maximum
6520 configured. A minimum setting of 1 is required.
6523 NAME: dns_retransmit_interval
6526 LOC: Config.Timeout.idns_retransmit
6527 IFDEF: !USE_DNSSERVERS
6529 Initial retransmit interval for DNS queries. The interval is
6530 doubled each time all configured DNS servers have been tried.
6536 LOC: Config.Timeout.idns_query
6537 IFDEF: !USE_DNSSERVERS
6539 DNS Query timeout. If no response is received to a DNS query
6540 within this time all DNS servers for the queried domain
6541 are assumed to be unavailable.
6548 LOC: Config.onoff.res_defnames
6550 Normally the RES_DEFNAMES resolver option is disabled
6551 (see res_init(3)). This prevents caches in a hierarchy
6552 from interpreting single-component hostnames locally. To allow
6553 Squid to handle single-component names, enable this option.
6556 NAME: dns_nameservers
6559 LOC: Config.dns_nameservers
6561 Use this if you want to specify a list of DNS name servers
6562 (IP addresses) to use instead of those given in your
6563 /etc/resolv.conf file.
6564 On Windows platforms, if no value is specified here or in
6565 the /etc/resolv.conf file, the list of DNS name servers are
6566 taken from the Windows registry, both static and dynamic DHCP
6567 configurations are supported.
6569 Example: dns_nameservers 10.0.0.1 192.172.0.4
6574 DEFAULT: @DEFAULT_HOSTS@
6575 LOC: Config.etcHostsPath
6577 Location of the host-local IP name-address associations
6578 database. Most Operating Systems have such a file on different
6580 - Un*X & Linux: /etc/hosts
6581 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
6582 (%SystemRoot% value install default is c:\winnt)
6583 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
6584 (%SystemRoot% value install default is c:\windows)
6585 - Windows 9x/Me: %windir%\hosts
6586 (%windir% value is usually c:\windows)
6587 - Cygwin: /etc/hosts
6589 The file contains newline-separated definitions, in the
6590 form ip_address_in_dotted_form name [name ...] names are
6591 whitespace-separated. Lines beginning with an hash (#)
6592 character are comments.
6594 The file is checked at startup and upon configuration.
6595 If set to 'none', it won't be checked.
6596 If append_domain is used, that domain will be added to
6597 domain-local (i.e. not containing any dot character) host
6603 LOC: Config.appendDomain
6606 Appends local domain name to hostnames without any dots in
6607 them. append_domain must begin with a period.
6609 Be warned there are now Internet names with no dots in
6610 them using only top-domain names, so setting this may
6611 cause some Internet sites to become unavailable.
6614 append_domain .yourdomain.com
6617 NAME: ignore_unknown_nameservers
6619 LOC: Config.onoff.ignore_unknown_nameservers
6622 By default Squid checks that DNS responses are received
6623 from the same IP addresses they are sent to. If they
6624 don't match, Squid ignores the response and writes a warning
6625 message to cache.log. You can allow responses from unknown
6626 nameservers by setting this option to 'off'.
6629 NAME: dns_v4_fallback
6632 LOC: Config.onoff.dns_require_A
6634 Standard practice with DNS is to lookup either A or AAAA records
6635 and use the results if it succeeds. Only looking up the other if
6636 the first attempt fails or otherwise produces no results.
6638 That policy however will cause squid to produce error pages for some
6639 servers that advertise AAAA but are unreachable over IPv6.
6641 If this is ON squid will always lookup both AAAA and A, using both.
6642 If this is OFF squid will lookup AAAA and only try A if none found.
6644 WARNING: There are some possibly unwanted side-effects with this on:
6645 *) Doubles the load placed by squid on the DNS network.
6646 *) May negatively impact connection delay times.
6650 COMMENT: (number of entries)
6653 LOC: Config.ipcache.size
6660 LOC: Config.ipcache.low
6667 LOC: Config.ipcache.high
6669 The size, low-, and high-water marks for the IP cache.
6672 NAME: fqdncache_size
6673 COMMENT: (number of entries)
6676 LOC: Config.fqdncache.size
6678 Maximum number of FQDN cache entries.
6683 -----------------------------------------------------------------------------
6690 LOC: Config.onoff.mem_pools
6692 If set, Squid will keep pools of allocated (but unused) memory
6693 available for future use. If memory is a premium on your
6694 system and you believe your malloc library outperforms Squid
6695 routines, disable this.
6698 NAME: memory_pools_limit
6702 LOC: Config.MemPools.limit
6704 Used only with memory_pools on:
6705 memory_pools_limit 50 MB
6707 If set to a non-zero value, Squid will keep at most the specified
6708 limit of allocated (but unused) memory in memory pools. All free()
6709 requests that exceed this limit will be handled by your malloc
6710 library. Squid does not pre-allocate any memory, just safe-keeps
6711 objects that otherwise would be free()d. Thus, it is safe to set
6712 memory_pools_limit to a reasonably high value even if your
6713 configuration will use less memory.
6715 If set to none, Squid will keep all memory it can. That is, there
6716 will be no limit on the total amount of memory used for safe-keeping.
6718 To disable memory allocation optimization, do not set
6719 memory_pools_limit to 0. Set memory_pools to "off" instead.
6721 An overhead for maintaining memory pools is not taken into account
6722 when the limit is checked. This overhead is close to four bytes per
6723 object kept. However, pools may actually _save_ memory because of
6724 reduced memory thrashing in your malloc library.
6728 COMMENT: on|off|transparent|truncate|delete
6731 LOC: opt_forwarded_for
6733 If set to "on", Squid will append your client's IP address
6734 in the HTTP requests it forwards. By default it looks like:
6736 X-Forwarded-For: 192.1.2.3
6738 If set to "off", it will appear as
6740 X-Forwarded-For: unknown
6742 If set to "transparent", Squid will not alter the
6743 X-Forwarded-For header in any way.
6745 If set to "delete", Squid will delete the entire
6746 X-Forwarded-For header.
6748 If set to "truncate", Squid will remove all existing
6749 X-Forwarded-For entries, and place itself as the sole entry.
6752 NAME: cachemgr_passwd
6753 TYPE: cachemgrpasswd
6755 LOC: Config.passwd_list
6757 Specify passwords for cachemgr operations.
6759 Usage: cachemgr_passwd password action action ...
6761 Some valid actions are (see cache manager menu for a full list):
6801 * Indicates actions which will not be performed without a
6802 valid password, others can be performed if not listed here.
6804 To disable an action, set the password to "disable".
6805 To allow performing an action without a password, set the
6808 Use the keyword "all" to set the same password for all actions.
6811 cachemgr_passwd secret shutdown
6812 cachemgr_passwd lesssssssecret info stats/objects
6813 cachemgr_passwd disable all
6820 LOC: Config.onoff.client_db
6822 If you want to disable collecting per-client statistics,
6823 turn off client_db here.
6826 NAME: refresh_all_ims
6830 LOC: Config.onoff.refresh_all_ims
6832 When you enable this option, squid will always check
6833 the origin server for an update when a client sends an
6834 If-Modified-Since request. Many browsers use IMS
6835 requests when the user requests a reload, and this
6836 ensures those clients receive the latest version.
6838 By default (off), squid may return a Not Modified response
6839 based on the age of the cached version.
6842 NAME: reload_into_ims
6843 IFDEF: HTTP_VIOLATIONS
6847 LOC: Config.onoff.reload_into_ims
6849 When you enable this option, client no-cache or ``reload''
6850 requests will be changed to If-Modified-Since requests.
6851 Doing this VIOLATES the HTTP standard. Enabling this
6852 feature could make you liable for problems which it
6855 see also refresh_pattern for a more selective approach.
6858 NAME: maximum_single_addr_tries
6860 LOC: Config.retry.maxtries
6863 This sets the maximum number of connection attempts for a
6864 host that only has one address (for multiple-address hosts,
6865 each address is tried once).
6867 The default value is one attempt, the (not recommended)
6868 maximum is 255 tries. A warning message will be generated
6869 if it is set to a value greater than ten.
6871 Note: This is in addition to the request re-forwarding which
6872 takes place if Squid fails to get a satisfying response.
6875 NAME: retry_on_error
6877 LOC: Config.retry.onerror
6880 If set to on Squid will automatically retry requests when
6881 receiving an error response. This is mainly useful if you
6882 are in a complex cache hierarchy to work around access
6886 NAME: as_whois_server
6888 LOC: Config.as_whois_server
6889 DEFAULT: whois.ra.net
6890 DEFAULT_IF_NONE: whois.ra.net
6892 WHOIS server to query for AS numbers. NOTE: AS numbers are
6893 queried only when Squid starts up, not for every request.
6898 LOC: Config.onoff.offline
6901 Enable this option and Squid will never try to validate cached
6905 NAME: uri_whitespace
6906 TYPE: uri_whitespace
6907 LOC: Config.uri_whitespace
6910 What to do with requests that have whitespace characters in the
6913 strip: The whitespace characters are stripped out of the URL.
6914 This is the behavior recommended by RFC2396.
6915 deny: The request is denied. The user receives an "Invalid
6917 allow: The request is allowed and the URI is not changed. The
6918 whitespace characters remain in the URI. Note the
6919 whitespace is passed to redirector processes if they
6921 encode: The request is allowed and the whitespace characters are
6922 encoded according to RFC1738. This could be considered
6923 a violation of the HTTP/1.1
6924 RFC because proxies are not allowed to rewrite URI's.
6925 chop: The request is allowed and the URI is chopped at the
6926 first whitespace. This might also be considered a
6932 LOC: Config.chroot_dir
6935 Specifies a directory where Squid should do a chroot() while
6936 initializing. This also causes Squid to fully drop root
6937 privileges after initializing. This means, for example, if you
6938 use a HTTP port less than 1024 and try to reconfigure, you may
6939 get an error saying that Squid can not open the port.
6942 NAME: balance_on_multiple_ip
6944 LOC: Config.onoff.balance_on_multiple_ip
6947 Modern IP resolvers in squid sort lookup results by preferred access.
6948 By default squid will use these IP in order and only rotates to
6949 the next listed when the most preffered fails.
6951 Some load balancing servers based on round robin DNS have been
6952 found not to preserve user session state across requests
6953 to different IP addresses.
6955 Enabling this directive Squid rotates IP's per request.
6958 NAME: pipeline_prefetch
6960 LOC: Config.onoff.pipeline_prefetch
6963 To boost the performance of pipelined requests to closer
6964 match that of a non-proxied environment Squid can try to fetch
6965 up to two requests in parallel from a pipeline.
6967 Defaults to off for bandwidth management and access logging
6971 NAME: high_response_time_warning
6974 LOC: Config.warnings.high_rptm
6977 If the one-minute median response time exceeds this value,
6978 Squid prints a WARNING with debug level 0 to get the
6979 administrators attention. The value is in milliseconds.
6982 NAME: high_page_fault_warning
6984 LOC: Config.warnings.high_pf
6987 If the one-minute average page fault rate exceeds this
6988 value, Squid prints a WARNING with debug level 0 to get
6989 the administrators attention. The value is in page faults
6993 NAME: high_memory_warning
6995 LOC: Config.warnings.high_memory
6998 If the memory usage (as determined by mallinfo) exceeds
6999 this amount, Squid prints a WARNING with debug level 0 to get
7000 the administrators attention.
7003 NAME: sleep_after_fork
7004 COMMENT: (microseconds)
7006 LOC: Config.sleep_after_fork
7009 When this is set to a non-zero value, the main Squid process
7010 sleeps the specified number of microseconds after a fork()
7011 system call. This sleep may help the situation where your
7012 system reports fork() failures due to lack of (virtual)
7013 memory. Note, however, if you have a lot of child
7014 processes, these sleep delays will add up and your
7015 Squid will not service requests for some amount of time
7016 until all the child processes have been started.
7017 On Windows value less then 1000 (1 milliseconds) are
7021 NAME: windows_ipaddrchangemonitor
7022 IFDEF: _SQUID_MSWIN_
7026 LOC: Config.onoff.WIN32_IpAddrChangeMonitor
7028 On Windows Squid by default will monitor IP address changes and will
7029 reconfigure itself after any detected event. This is very useful for
7030 proxies connected to internet with dial-up interfaces.
7031 In some cases (a Proxy server acting as VPN gateway is one) it could be
7032 desiderable to disable this behaviour setting this to 'off'.
7033 Note: after changing this, Squid service must be restarted.
7038 IFDEF: USE_SQUID_EUI
7040 LOC: Eui::TheConfig.euiLookup
7042 Whether to lookup the EUI or MAC address of a connected client.
7045 NAME: max_filedescriptors max_filedesc
7048 LOC: Config.max_filedescriptors
7050 The maximum number of filedescriptors supported.
7052 The default "0" means Squid inherits the current ulimit setting.
7054 Note: Changing this requires a restart of Squid. Also
7055 not all comm loops supports large values.
7063 Number of main Squid processes or "workers" to fork and maintain.
7064 0: "no daemon" mode, like running "squid -N ..."
7065 1: "no SMP" mode, start one main Squid process daemon (default)
7066 N: start N main Squid process daemons (i.e., SMP mode)
7068 In SMP mode, each worker does nearly all what a single Squid daemon
7069 does (e.g., listen on http_port and forward HTTP requests).