3 # $Id: cf.data.pre,v 1.431 2007/04/07 09:35:38 serassio Exp $
6 # SQUID Web Proxy Cache http://www.squid-cache.org/
7 # ----------------------------------------------------------
9 # Squid is the result of efforts by numerous individuals from
10 # the Internet community; see the CONTRIBUTORS file for full
11 # details. Many organizations have provided support for Squid's
12 # development; see the SPONSORS file for full details. Squid is
13 # Copyrighted (C) 2000 by the Regents of the University of
14 # California; see the COPYRIGHT file for full details. Squid
15 # incorporates software developed and/or copyrighted by other
16 # sources; see the CREDITS file for full details.
18 # This program is free software; you can redistribute it and/or modify
19 # it under the terms of the GNU General Public License as published by
20 # the Free Software Foundation; either version 2 of the License, or
21 # (at your option) any later version.
23 # This program is distributed in the hope that it will be useful,
24 # but WITHOUT ANY WARRANTY; without even the implied warranty of
25 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
26 # GNU General Public License for more details.
28 # You should have received a copy of the GNU General Public License
29 # along with this program; if not, write to the Free Software
30 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
34 WELCOME TO SQUID @VERSION@
35 ----------------------------
37 This is the default Squid configuration file. You may wish
38 to look at the Squid home page (http://www.squid-cache.org/)
39 for the FAQ and other documentation.
41 The default Squid config file shows what the defaults for
42 various options happen to be. If you don't need to change the
43 default, you shouldn't uncomment the line. Doing so may cause
44 run-time problems. In some cases "none" refers to no default
45 setting at all, while in other cases it refers to a valid
46 option - the comments for that keyword indicate if this is the
53 -----------------------------------------------------------------------------
56 NAME: http_port ascii_port
59 LOC: Config.Sockaddr.http
62 hostname:port [options]
63 1.2.3.4:port [options]
65 The socket addresses where Squid will listen for HTTP client
66 requests. You may specify multiple socket addresses.
67 There are three forms: port alone, hostname with port, and
68 IP address with port. If you specify a hostname or IP
69 address, Squid binds the socket to that specific
70 address. This replaces the old 'tcp_incoming_address'
71 option. Most likely, you do not need to bind to a specific
72 address, so you can use the port number alone.
74 If you are running Squid in accelerator mode, you
75 probably want to listen on port 80 also, or instead.
77 The -a command line option will override the *first* port
78 number listed here. That option will NOT override an IP
81 You may specify multiple socket addresses on multiple lines.
85 transparent Support for transparent proxies
87 accel Accelerator mode. Also set implicit by the other
88 accelerator directives
90 vhost Accelerator mode using Host header for virtual
93 vport Accelerator with IP based virtual host support
95 vport=NN As above, but uses specified port number rather
96 than the http_port number
98 defaultsite= Main web site name for accelerators
100 protocol= Protocol to reconstruct accelerated requests with.
103 disable-pmtu-discovery=
104 Control Path-MTU discovery usage:
105 off lets OS decide on what to do (default).
106 transparent disable PMTU discovery when transparent
108 always disable always PMTU discovery.
110 In many setups of transparently intercepting proxies Path-MTU
111 discovery can not work on traffic towards the clients. This is
112 the case when the intercepting device does not fully track
113 connections and fails to forward ICMP must fragment messages
114 to the cache server. If you have such setup and experience that
115 certain clients sporadically hang or never complete requests set
116 disable-pmtu-discovery option to 'transparent'.
118 If you run Squid on a dual-homed machine with an internal
119 and an external interface we recommend you to specify the
120 internal address:port in http_port. This way Squid will only be
121 visible on the internal address.
123 # Squid normally listens to port 3128
124 http_port @DEFAULT_HTTP_PORT@
130 TYPE: https_port_list
132 LOC: Config.Sockaddr.https
134 Usage: [ip:]port cert=certificate.pem [key=key.pem] [options...]
136 The socket address where Squid will listen for HTTPS client
139 This is really only useful for situations where you are running
140 squid in accelerator mode and you want to do the SSL work at the
143 You may specify multiple socket addresses on multiple lines,
144 each with their own SSL certificate and/or options.
148 defaultsite= The name of the https site presented on
151 protocol= Protocol to reconstruct accelerated requests
152 with. Defaults to https
154 cert= Path to SSL certificate (PEM format)
156 key= Path to SSL private key file (PEM format)
157 if not specified, the certificate file is
158 assumed to be a combined certificate and
161 version= The version of SSL/TLS supported
162 1 automatic (default)
167 cipher= Colon separated list of supported ciphers
169 options= Various SSL engine options. The most important
171 NO_SSLv2 Disallow the use of SSLv2
172 NO_SSLv3 Disallow the use of SSLv3
173 NO_TLSv1 Disallow the use of TLSv1
174 SINGLE_DH_USE Always create a new key when using
175 temporary/ephemeral DH key exchanges
176 See src/ssl_support.c or OpenSSL SSL_CTX_set_options
177 documentation for a complete list of options
179 clientca= File containing the list of CAs to use when
180 requesting a client certificate
182 cafile= File containing additional CA certificates to
183 use when verifying client certificates. If unset
184 clientca will be used
186 capath= Directory containing additional CA certificates
187 and CRL lists to use when verifying client certificates
189 crlfile= File of additional CRL lists to use when verifying
190 the client certificate, in addition to CRLs stored in
191 the capath. Implies VERIFY_CRL flag below.
193 dhparams= File containing DH parameters for temporary/ephemeral
196 sslflags= Various flags modifying the use of SSL:
198 Don't request client certificates
199 immediately, but wait until acl processing
200 requires a certificate (not yet implemented)
202 Don't use the default CA lists built in
205 Don't allow for session reuse. Each connection
206 will result in a new SSL session.
208 Verify CRL lists when accepting client
211 Verify CRL lists for all certificates in the
212 client certificate chain
214 sslcontext= SSL session ID context identifier.
216 accel Accelerator mode. Also set implicit by the other
217 accelerator directives
219 vhost Accelerator mode using Host header for virtual
222 vport Accelerator with IP based virtual host support
224 vport=NN As above, but uses specified port number rather
225 than the https_port number
229 NAME: ssl_unclean_shutdown
233 LOC: Config.SSL.unclean_shutdown
235 Some browsers (especially MSIE) bugs out on SSL shutdown
242 LOC: Config.SSL.ssl_engine
245 The openssl engine to use. You will need to set this if you
246 would like to use hardware SSL acceleration for example.
249 NAME: sslproxy_client_certificate
252 LOC: Config.ssl_client.cert
255 Client SSL Certificate to use when proxying https:// URLs
258 NAME: sslproxy_client_key
261 LOC: Config.ssl_client.key
264 Client SSL Key to use when proxying https:// URLs
267 NAME: sslproxy_version
270 LOC: Config.ssl_client.version
273 SSL version level to use when proxying https:// URLs
276 NAME: sslproxy_options
279 LOC: Config.ssl_client.options
282 SSL engine options to use when proxying https:// URLs
285 NAME: sslproxy_cipher
288 LOC: Config.ssl_client.cipher
291 SSL cipher list to use when proxying https:// URLs
294 NAME: sslproxy_cafile
297 LOC: Config.ssl_client.cafile
300 file containing CA certificates to use when verifying server
301 certificates while proxying https:// URLs
304 NAME: sslproxy_capath
307 LOC: Config.ssl_client.capath
310 directory containing CA certificates to use when verifying
311 server certificates while proxying https:// URLs
317 LOC: Config.ssl_client.flags
320 Various flags modifying the use of SSL while proxying https:// URLs:
321 DONT_VERIFY_PEER Accept certificates even if they fail to
323 NO_DEFAULT_CA Don't use the default CA list built in
327 NAME: sslpassword_program
330 LOC: Config.Program.ssl_password
333 Specify a program used for entering SSL key passphrases
334 when using encrypted SSL certificate keys. If not specified
335 keys must either be unencrypted, or Squid started with the -N
336 option to allow it to query interactively for the passphrase.
339 NAME: icp_port udp_port
344 The port number where Squid sends and receives ICP queries to
345 and from neighbor caches. The standard UDP port for ICP is 3130.
346 Default is disabled (0).
348 icp_port @DEFAULT_ICP_PORT@
356 LOC: Config.Port.htcp
358 The port number where Squid sends and receives HTCP queries to
359 and from neighbor caches. Default is 4827. To disable use
366 LOC: Config.mcast_group_list
369 This tag specifies a list of multicast groups which your server
370 should join to receive multicasted ICP queries.
372 NOTE! Be very careful what you put here! Be sure you
373 understand the difference between an ICP _query_ and an ICP
374 _reply_. This option is to be set only if you want to RECEIVE
375 multicast queries. Do NOT set this option to SEND multicast
376 ICP (use cache_peer for that). ICP replies are always sent via
377 unicast, so this option does not affect whether or not you will
378 receive replies from multicast group members.
380 You must be very careful to NOT use a multicast address which
381 is already in use by another group of caches.
383 If you are unsure about multicast, please read the Multicast
384 chapter in the Squid FAQ (http://www.squid-cache.org/FAQ/).
386 Usage: mcast_groups 239.128.16.128 224.0.1.20
388 By default, Squid doesn't listen on any multicast groups.
392 NAME: udp_incoming_address
394 LOC:Config.Addrs.udp_incoming
398 NAME: udp_outgoing_address
400 LOC: Config.Addrs.udp_outgoing
401 DEFAULT: 255.255.255.255
403 udp_incoming_address is used for the ICP socket receiving packets
405 udp_outgoing_address is used for ICP packets sent out to other
408 The default behavior is to not bind to any specific address.
410 A udp_incoming_address value of 0.0.0.0 indicates Squid
411 should listen for UDP messages on all available interfaces.
413 If udp_outgoing_address is set to 255.255.255.255 (the default)
414 it will use the same socket as udp_incoming_address. Only
415 change this if you want to have ICP queries sent using another
416 address than where this Squid listens for ICP queries from other
419 NOTE, udp_incoming_address and udp_outgoing_address can not
420 have the same value since they both use port 3130.
424 OPTIONS WHICH AFFECT THE NEIGHBOR SELECTION ALGORITHM
425 -----------------------------------------------------------------------------
433 To specify other caches in a hierarchy, use the format:
435 cache_peer hostname type http_port icp_port [options]
440 # hostname type port port options
441 # -------------------- -------- ----- ----- -----------
442 cache_peer parent.foo.net parent 3128 3130 [proxy-only]
443 cache_peer sib1.foo.net sibling 3128 3130 [proxy-only]
444 cache_peer sib2.foo.net sibling 3128 3130 [proxy-only]
446 type: either 'parent', 'sibling', or 'multicast'.
448 proxy_port: The port number where the cache listens for proxy
451 icp_port: Used for querying neighbor caches about
452 objects. To have a non-ICP neighbor
453 specify '7' for the ICP port and make sure the
454 neighbor machine has the UDP echo port
455 enabled in its /etc/inetd.conf file.
472 login=user:password | PASS | *:password
483 sslcert=/path/to/ssl/certificate
484 sslkey=/path/to/ssl/key
488 front-end-https[=on|auto]
490 use 'proxy-only' to specify objects fetched
491 from this cache should not be saved locally.
493 use 'weight=n' to specify a weighted parent.
494 The weight must be an integer. The default weight
495 is 1, larger weights are favored more.
497 use 'basetime=n' to specify a base amount to
498 be subtracted from round trip times of parents.
499 It is subtracted before division by weight in calculating
500 which parent to fectch from. If the rtt is less than the
501 base time the rtt is set to a minimal value.
503 use 'ttl=n' to specify a IP multicast TTL to use
504 when sending an ICP queries to this address.
505 Only useful when sending to a multicast group.
506 Because we don't accept ICP replies from random
507 hosts, you must configure other group members as
508 peers with the 'multicast-responder' option below.
510 use 'no-query' to NOT send ICP queries to this
513 use 'background-ping' to only send ICP queries to this
514 neighbor infrequently. This is used to keep the neighbor
515 round trip time updated and is usually used in
516 conjunction with weighted-round-robin.
518 use 'default' if this is a parent cache which can
519 be used as a "last-resort." You should probably
520 only use 'default' in situations where you cannot
521 use ICP with your parent cache(s).
523 use 'round-robin' to define a set of parents which
524 should be used in a round-robin fashion in the
525 absence of any ICP queries.
527 use 'weighted-round-robin' to define a set of parents
528 which should be used in a round-robin fashion with the
529 frequency of each parent being based on the round trip
530 time. Closer parents are used more often.
531 Usually used for background-ping parents.
533 use 'carp' to define a set of parents which should
534 be used as a CARP array. The requests will be
535 distributed among the parents based on the CARP load
536 balancing hash function based on their weigth.
538 'multicast-responder' indicates the named peer
539 is a member of a multicast group. ICP queries will
540 not be sent directly to the peer, but ICP replies
541 will be accepted from it.
543 'closest-only' indicates that, for ICP_OP_MISS
544 replies, we'll only forward CLOSEST_PARENT_MISSes
545 and never FIRST_PARENT_MISSes.
547 use 'no-digest' to NOT request cache digests from
550 'no-netdb-exchange' disables requesting ICMP
551 RTT database (NetDB) from the neighbor.
553 use 'no-delay' to prevent access to this neighbor
554 from influencing the delay pools.
556 use 'login=user:password' if this is a personal/workgroup
557 proxy and your parent requires proxy authentication.
558 Note: The string can include URL escapes (i.e. %20 for
559 spaces). This also means % must be written as %%.
561 use 'login=PASS' if users must authenticate against
562 the upstream proxy or in the case of a reverse proxy
563 configuration, the origin web server. This will pass
564 the users credentials as they are to the peer.
565 This only works for the Basic HTTP authentication scheme.
566 Note: To combine this with proxy_auth both proxies must
567 share the same user database as HTTP only allows for
569 Also be warned this will expose your users proxy
570 password to the peer. USE WITH CAUTION
572 use 'login=*:password' to pass the username to the
573 upstream cache, but with a fixed password. This is meant
574 to be used when the peer is in another administrative
575 domain, but it is still needed to identify each user.
576 The star can optionally be followed by some extra
577 information which is added to the username. This can
578 be used to identify this proxy to the peer, similar to
579 the login=username:password option above.
581 use 'connect-timeout=nn' to specify a peer
582 specific connect timeout (also see the
583 peer_connect_timeout directive)
585 use 'digest-url=url' to tell Squid to fetch the cache
586 digest (if digests are enabled) for this host from
587 the specified URL rather than the Squid default
590 use 'allow-miss' to disable Squid's use of only-if-cached
591 when forwarding requests to siblings. This is primarily
592 useful when icp_hit_stale is used by the sibling. To
593 extensive use of this option may result in forwarding
594 loops, and you should avoid having two-way peerings
595 with this option. (for example to deny peer usage on
596 requests from peer by denying cache_peer_access if the
599 use 'max-conn' to limit the amount of connections Squid
600 may open to this peer.
602 use 'htcp' to send HTCP, instead of ICP, queries
603 to the neighbor. You probably also want to
604 set the "icp port" to 4827 instead of 3130.
606 use 'htcp-oldsquid' to send HTCP to old Squid versions
608 'originserver' causes this parent peer to be contacted as
609 a origin server. Meant to be used in accelerator setups.
611 use 'name=xxx' if you have multiple peers on the same
612 host but different ports. This name can be used to
613 differentiate the peers in cache_peer_access and similar
616 use 'forceddomain=name' to forcibly set the Host header
617 of requests forwarded to this peer. Useful in accelerator
618 setups where the server (peer) expects a certain domain
619 name and using redirectors to feed this domainname
622 use 'ssl' to indicate connections to this peer should
623 bs SSL/TLS encrypted.
625 use 'sslcert=/path/to/ssl/certificate' to specify a client
626 SSL certificate to use when connecting to this peer.
628 use 'sslkey=/path/to/ssl/key' to specify the private SSL
629 key corresponding to sslcert above. If 'sslkey' is not
630 specified 'sslcert' is assumed to reference a
631 combined file containing both the certificate and the key.
633 use sslversion=1|2|3|4 to specify the SSL version to use
634 when connecting to this peer
635 1 = automatic (default)
640 use sslcipher=... to specify the list of valid SSL chipers
641 to use when connecting to this peer
643 use ssloptions=... to specify various SSL engine options:
644 NO_SSLv2 Disallow the use of SSLv2
645 NO_SSLv3 Disallow the use of SSLv3
646 NO_TLSv1 Disallow the use of TLSv1
647 See src/ssl_support.c or the OpenSSL documentation for
648 a more complete list.
650 use cafile=... to specify a file containing additional
651 CA certificates to use when verifying the peer certificate
653 use capath=... to specify a directory containing additional
654 CA certificates to use when verifying the peer certificate
656 use sslflags=... to specify various flags modifying the
659 Accept certificates even if they fail to
662 Don't use the default CA list built in
665 Don't verify the peer certificate
666 matches the server name
668 use sslname= to specify the peer name as advertised
669 in it's certificate. Used for verifying the correctness
670 of the received peer certificate. If not specified the
671 peer hostname will be used.
673 use front-end-https to enable the "Front-End-Https: On"
674 header needed when using Squid as a SSL frontend infront
675 of Microsoft OWA. See MS KB document Q307347 for details
676 on this header. If set to auto the header will
677 only be added if the request is forwarded as a https://
680 NOTE: non-ICP neighbors must be specified as 'parent'.
683 NAME: cache_peer_domain cache_host_domain
688 Use to limit the domains for which a neighbor cache will be
691 cache_peer_domain cache-host domain [domain ...]
692 cache_peer_domain cache-host !domain
694 For example, specifying
696 cache_peer_domain parent.foo.net .edu
698 has the effect such that UDP query packets are sent to
699 'bigserver' only when the requested object exists on a
700 server in the .edu domain. Prefixing the domainname
701 with '!' means the cache will be queried for objects
704 NOTE: * Any number of domains may be given for a cache-host,
705 either on the same or separate lines.
706 * When multiple domains are given for a particular
707 cache-host, the first matched domain is applied.
708 * Cache hosts with no domain restrictions are queried
710 * There are no defaults.
711 * There is also a 'cache_peer_access' tag in the ACL
716 NAME: neighbor_type_domain
721 usage: neighbor_type_domain neighbor parent|sibling domain domain ...
723 Modifying the neighbor type for specific domains is now
724 possible. You can treat some domains differently than the the
725 default neighbor type specified on the 'cache_peer' line.
726 Normally it should only be necessary to list domains which
727 should be treated differently because the default neighbor type
728 applies for hostnames which do not match domains listed here.
731 cache_peer parent cache.foo.org 3128 3130
732 neighbor_type_domain cache.foo.org sibling .com .net
733 neighbor_type_domain cache.foo.org sibling .au .de
736 NAME: icp_query_timeout
740 LOC: Config.Timeout.icp_query
742 Normally Squid will automatically determine an optimal ICP
743 query timeout value based on the round-trip-time of recent ICP
744 queries. If you want to override the value determined by
745 Squid, set this 'icp_query_timeout' to a non-zero value. This
746 value is specified in MILLISECONDS, so, to use a 2-second
747 timeout (the old default), you would write:
749 icp_query_timeout 2000
752 NAME: maximum_icp_query_timeout
756 LOC: Config.Timeout.icp_query_max
758 Normally the ICP query timeout is determined dynamically. But
759 sometimes it can lead to very large values (say 5 seconds).
760 Use this option to put an upper limit on the dynamic timeout
761 value. Do NOT use this option to always use a fixed (instead
762 of a dynamic) timeout value. To set a fixed timeout see the
763 'icp_query_timeout' directive.
766 NAME: minimum_icp_query_timeout
770 LOC: Config.Timeout.icp_query_min
772 Normally the ICP query timeout is determined dynamically. But
773 sometimes it can lead to very small timeouts, even lower than
774 the normal latency variance on your link due to traffic.
775 Use this option to put an lower limit on the dynamic timeout
776 value. Do NOT use this option to always use a fixed (instead
777 of a dynamic) timeout value. To set a fixed timeout see the
778 'icp_query_timeout' directive.
781 NAME: mcast_icp_query_timeout
785 LOC: Config.Timeout.mcast_icp_query
787 For Multicast peers, Squid regularly sends out ICP "probes" to
788 count how many other peers are listening on the given multicast
789 address. This value specifies how long Squid should wait to
790 count all the replies. The default is 2000 msec, or 2
794 NAME: dead_peer_timeout
798 LOC: Config.Timeout.deadPeer
800 This controls how long Squid waits to declare a peer cache
801 as "dead." If there are no ICP replies received in this
802 amount of time, Squid will declare the peer dead and not
803 expect to receive any further ICP replies. However, it
804 continues to send ICP queries, and will mark the peer as
805 alive upon receipt of the first subsequent ICP reply.
807 This timeout also affects when Squid expects to receive ICP
808 replies from peers. If more than 'dead_peer' seconds have
809 passed since the last ICP reply was received, Squid will not
810 expect to receive an ICP reply on the next query. Thus, if
811 your time between requests is greater than this timeout, you
812 will see a lot of requests sent DIRECT to origin servers
813 instead of to your parents.
817 NAME: hierarchy_stoplist
820 LOC: Config.hierarchy_stoplist
822 A list of words which, if found in a URL, cause the object to
823 be handled directly by this cache. In other words, use this
824 to not query neighbor caches for certain objects. You may
825 list this option multiple times.
827 #We recommend you to use at least the following line.
828 hierarchy_stoplist cgi-bin ?
836 LOC: Config.accessList.noCache
838 A list of ACL elements which, if matched, cause the request to
839 not be satisfied from the cache and the reply to not be cached.
840 In other words, use this to force certain objects to never be cached.
842 You must use the word 'DENY' to indicate the ACL names which should
845 Default is to allow all to be cached
847 #We recommend you to use the following two lines.
848 acl QUERY urlpath_regex cgi-bin \?
853 NAME: background_ping_rate
857 LOC: Config.backgroundPingRate
859 Controls how often the ICP pings are sent to siblings that
860 have background-ping set.
865 OPTIONS WHICH AFFECT THE CACHE SIZE
866 -----------------------------------------------------------------------------
873 LOC: Config.memMaxSize
875 NOTE: THIS PARAMETER DOES NOT SPECIFY THE MAXIMUM PROCESS SIZE.
876 IT ONLY PLACES A LIMIT ON HOW MUCH ADDITIONAL MEMORY SQUID WILL
877 USE AS A MEMORY CACHE OF OBJECTS. SQUID USES MEMORY FOR OTHER
878 THINGS AS WELL. SEE THE SQUID FAQ SECTION 8 FOR DETAILS.
880 'cache_mem' specifies the ideal amount of memory to be used
884 * Negative-Cached objects
886 Data for these objects are stored in 4 KB blocks. This
887 parameter specifies the ideal upper limit on the total size of
888 4 KB blocks allocated. In-Transit objects take the highest
891 In-transit objects have priority over the others. When
892 additional space is needed for incoming data, negative-cached
893 and hot objects will be released. In other words, the
894 negative-cached and hot objects will fill up any unused space
895 not needed for in-transit objects.
897 If circumstances require, this limit will be exceeded.
898 Specifically, if your incoming request rate requires more than
899 'cache_mem' of memory to hold in-transit objects, Squid will
900 exceed this limit to satisfy the new requests. When the load
901 decreases, blocks will be freed until the high-water mark is
902 reached. Thereafter, blocks will be used to store hot
908 COMMENT: (percent, 0-100)
911 LOC: Config.Swap.lowWaterMark
914 NAME: cache_swap_high
915 COMMENT: (percent, 0-100)
918 LOC: Config.Swap.highWaterMark
921 The low- and high-water marks for cache object replacement.
922 Replacement begins when the swap (disk) usage is above the
923 low-water mark and attempts to maintain utilization near the
924 low-water mark. As swap utilization gets close to high-water
925 mark object eviction becomes more aggressive. If utilization is
926 close to the low-water mark less replacement is done each time.
928 Defaults are 90% and 95%. If you have a large cache, 5% could be
929 hundreds of MB. If this is the case you may wish to set these
930 numbers closer together.
933 NAME: maximum_object_size
937 LOC: Config.Store.maxObjectSize
939 Objects larger than this size will NOT be saved on disk. The
940 value is specified in kilobytes, and the default is 4MB. If
941 you wish to get a high BYTES hit ratio, you should probably
942 increase this (one 32 MB object hit counts for 3200 10KB
943 hits). If you wish to increase speed more than your want to
944 save bandwidth you should leave this low.
946 NOTE: if using the LFUDA replacement policy you should increase
947 this value to maximize the byte hit rate improvement of LFUDA!
948 See replacement_policy below for a discussion of this policy.
951 NAME: minimum_object_size
955 LOC: Config.Store.minObjectSize
957 Objects smaller than this size will NOT be saved on disk. The
958 value is specified in kilobytes, and the default is 0 KB, which
959 means there is no minimum.
962 NAME: maximum_object_size_in_memory
966 LOC: Config.Store.maxInMemObjSize
968 Objects greater than this size will not be attempted to kept in
969 the memory cache. This should be set high enough to keep objects
970 accessed frequently in memory to improve performance whilst low
971 enough to keep larger objects from hoarding cache_mem .
975 COMMENT: (number of entries)
978 LOC: Config.ipcache.size
985 LOC: Config.ipcache.low
992 LOC: Config.ipcache.high
994 The size, low-, and high-water marks for the IP cache.
998 COMMENT: (number of entries)
1001 LOC: Config.fqdncache.size
1003 Maximum number of FQDN cache entries.
1006 NAME: cache_replacement_policy
1008 LOC: Config.replPolicy
1011 The cache replacement policy parameter determines which
1012 objects are evicted (replaced) when disk space is needed.
1014 lru : Squid's original list based LRU policy
1015 heap GDSF : Greedy-Dual Size Frequency
1016 heap LFUDA: Least Frequently Used with Dynamic Aging
1017 heap LRU : LRU policy implemented using a heap
1019 Applies to any cache_dir lines listed below this.
1021 The LRU policies keeps recently referenced objects.
1023 The heap GDSF policy optimizes object hit rate by keeping smaller
1024 popular objects in cache so it has a better chance of getting a
1025 hit. It achieves a lower byte hit rate than LFUDA though since
1026 it evicts larger (possibly popular) objects.
1028 The heap LFUDA policy keeps popular objects in cache regardless of
1029 their size and thus optimizes byte hit rate at the expense of
1030 hit rate since one large, popular object will prevent many
1031 smaller, slightly less popular objects from being cached.
1033 Both policies utilize a dynamic aging mechanism that prevents
1034 cache pollution that can otherwise occur with frequency-based
1035 replacement policies.
1037 NOTE: if using the LFUDA replacement policy you should increase
1038 the value of maximum_object_size above its default of 4096 KB to
1039 to maximize the potential byte hit rate improvement of LFUDA.
1041 For more information about the GDSF and LFUDA cache replacement
1042 policies see http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html
1043 and http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.
1046 NAME: memory_replacement_policy
1048 LOC: Config.memPolicy
1051 The memory replacement policy parameter determines which
1052 objects are purged from memory when memory space is needed.
1054 See cache_replacement_policy for details.
1059 LOGFILE PATHNAMES AND CACHE DIRECTORIES
1060 -----------------------------------------------------------------------------
1066 DEFAULT_IF_NONE: ufs @DEFAULT_SWAP_DIR@ 100 16 256
1067 LOC: Config.cacheSwap
1071 cache_dir Type Directory-Name Fs-specific-data [options]
1073 You can specify multiple cache_dir lines to spread the
1074 cache among different disk partitions.
1076 Type specifies the kind of storage system to use. Only "ufs"
1077 is built by default. To enable any of the other storage systems
1078 see the --enable-storeio configure option.
1080 'Directory' is a top-level directory where cache swap
1081 files will be stored. If you want to use an entire disk
1082 for caching, this can be the mount-point directory.
1083 The directory must exist and be writable by the Squid
1084 process. Squid will NOT create this directory for you.
1088 "ufs" is the old well-known Squid storage format that has always
1091 cache_dir ufs Directory-Name Mbytes L1 L2 [options]
1093 'Mbytes' is the amount of disk space (MB) to use under this
1094 directory. The default is 100 MB. Change this to suit your
1095 configuration. Do NOT put the size of your disk drive here.
1096 Instead, if you want Squid to use the entire disk drive,
1097 subtract 20% and use that value.
1099 'Level-1' is the number of first-level subdirectories which
1100 will be created under the 'Directory'. The default is 16.
1102 'Level-2' is the number of second-level subdirectories which
1103 will be created under each first-level directory. The default
1106 The aufs store type:
1108 "aufs" uses the same storage format as "ufs", utilizing
1109 POSIX-threads to avoid blocking the main Squid process on
1110 disk-I/O. This was formerly known in Squid as async-io.
1112 cache_dir aufs Directory-Name Mbytes L1 L2 [options]
1114 see argument descriptions under ufs above
1116 The diskd store type:
1118 "diskd" uses the same storage format as "ufs", utilizing a
1119 separate process to avoid blocking the main Squid process on
1122 cache_dir diskd Directory-Name Mbytes L1 L2 [options] [Q1=n] [Q2=n]
1124 see argument descriptions under ufs above
1126 Q1 specifies the number of unacknowledged I/O requests when Squid
1127 stops opening new files. If this many messages are in the queues,
1128 Squid won't open new files. Default is 64
1130 Q2 specifies the number of unacknowledged messages when Squid
1131 starts blocking. If this many messages are in the queues,
1132 Squid blocks until it receives some replies. Default is 72
1134 When Q1 < Q2 (the default), the cache directory is optimized
1135 for lower response time at the expense of a decrease in hit
1136 ratio. If Q1 > Q2, the cache directory is optimized for
1137 higher hit ratio at the expense of an increase in response
1140 The coss store type:
1142 block-size=n defines the "block size" for COSS cache_dir's.
1143 Squid uses file numbers as block numbers. Since file numbers
1144 are limited to 24 bits, the block size determines the maximum
1145 size of the COSS partition. The default is 512 bytes, which
1146 leads to a maximum cache_dir size of 512<<24, or 8 GB. Note
1147 you should not change the coss block size after Squid
1148 has written some objects to the cache_dir.
1150 The coss file store has changed from 2.5. Now it uses a file
1151 called 'stripe' in the directory names in the config - and
1152 this will be created by squid -z.
1156 read-only, this cache_dir is read only.
1158 max-size=n, refers to the max object size this storedir supports.
1159 It is used to initially choose the storedir to dump the object.
1160 Note: To make optimal use of the max-size limits you should order
1161 the cache_dir lines with the smallest max-size value first and the
1162 ones with no max-size specification last.
1164 Note for coss, max-size must be less than COSS_MEMBUF_SZ,
1165 which can be changed with the --with-coss-membuf-size=N configure
1168 The null store type:
1170 no options are allowed or required
1176 LOC: Config.Log.logformats
1181 logformat <name> <format specification>
1183 Defines an access log format.
1185 The <format specification> is a string with embedded % format codes
1187 % format codes all follow the same basic structure where all but
1188 the formatcode is optional. Output strings are automatically escaped
1189 as required according to their context and the output format
1190 modifiers are usually not needed, but can be specified if an explicit
1191 output format is desired.
1193 % ["|[|'|#] [-] [[0]width] [{argument}] formatcode
1195 " output in quoted string format
1196 [ output in squid text log format as used by log_mime_hdrs
1197 # output in URL quoted format
1201 width field width. If starting with 0 the
1202 output is zero padded
1203 {arg} argument such as header name etc
1207 >a Client source IP address
1209 >p Client source port
1210 <A Server IP address or peer name
1211 la Local IP address (http_port)
1212 lp Local port number (http_port)
1213 ts Seconds since epoch
1214 tu subsecond time (milliseconds)
1215 tl Local time. Optional strftime format argument
1216 default %d/%b/%Y:%H:%M:S %z
1217 tg GMT time. Optional strftime format argument
1218 default %d/%b/%Y:%H:%M:S %z
1219 tr Response time (milliseconds)
1220 >h Request header. Optional header name argument
1221 on the format header[:[separator]element]
1222 <h Reply header. Optional header name argument
1227 ue User from external acl
1229 Ss Squid request status (TCP_MISS etc)
1230 Sh Squid hierarchy status (DEFAULT_PARENT etc)
1231 mt MIME content type
1232 rm Request method (GET/POST etc)
1234 rv Request protocol version
1235 et Tag returned by external acl
1236 ea Log string returned by external acl
1237 <st Reply size including HTTP headers
1238 <sH Reply high offset sent
1239 <sS Upstream object size
1240 % a literal % character
1242 logformat squid %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt
1243 logformat squidmime %ts.%03tu %6tr %>a %Ss/%03Hs %<st %rm %ru %un %Sh/%<A %mt [%>h] [%<h]
1244 logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st %Ss:%Sh
1245 logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
1248 NAME: access_log cache_access_log
1250 LOC: Config.Log.accesslogs
1252 DEFAULT_IF_NONE: @DEFAULT_ACCESS_LOG@
1254 These files log client request activities. Has a line every HTTP or
1255 ICP request. The format is:
1256 access_log <filepath> [<logformat name> [acl acl ...]]
1257 access_log none [acl acl ...]]
1259 Will log to the specified file using the specified format (which
1260 must be defined in a logformat directive) those entries which match
1261 ALL the acl's specified (which must be defined in acl clauses).
1262 If no acl is specified, all requests will be logged to this file.
1264 To disable logging of a request use the filepath "none", in which case
1265 a logformat name should not be specified.
1267 To log the request via syslog specify a filepath of "syslog":
1269 access_log syslog[:facility|priority] [format [acl1 [acl2 ....]]]
1270 where facility could be any of:
1271 LOG_AUTHPRIV, LOG_DAEMON, LOG_LOCAL0 .. LOG_LOCAL7 or LOG_USER.
1273 And priority could be any of:
1274 LOG_ERR, LOG_WARNING, LOG_NOTICE, LOG_INFO, LOG_DEBUG.
1280 DEFAULT: @DEFAULT_CACHE_LOG@
1283 Cache logging file. This is where general information about
1284 your cache's behavior goes. You can increase the amount of data
1285 logged to this file with the "debug_options" tag below.
1289 NAME: cache_store_log
1291 DEFAULT: @DEFAULT_STORE_LOG@
1292 LOC: Config.Log.store
1294 Logs the activities of the storage manager. Shows which
1295 objects are ejected from the cache, and which objects are
1296 saved and for how long. To disable, enter "none". There are
1297 not really utilities to analyze this data, so you can safely
1302 NAME: cache_swap_state cache_swap_log
1304 LOC: Config.Log.swap
1307 Location for the cache "swap.state" file. This index file holds
1308 the metadata of objects saved on disk. It is used to rebuild
1309 the cache during startup. Normally this file resides in each
1310 'cache_dir' directory, but you may specify an alternate
1311 pathname here. Note you must give a full filename, not just
1312 a directory. Since this is the index for the whole object
1313 list you CANNOT periodically rotate it!
1315 If %s can be used in the file name it will be replaced with a
1316 a representation of the cache_dir name where each / is replaced
1317 with '.'. This is needed to allow adding/removing cache_dir
1318 lines when cache_swap_log is being used.
1320 If have more than one 'cache_dir', and %s is not used in the name
1321 these swap logs will have names such as:
1327 The numbered extension (which is added automatically)
1328 corresponds to the order of the 'cache_dir' lines in this
1329 configuration file. If you change the order of the 'cache_dir'
1330 lines in this file, these log files will NOT correspond to
1331 the correct 'cache_dir' entry (unless you manually rename
1332 them). We recommend you do NOT use this option. It is
1333 better to keep these log files in each 'cache_dir' directory.
1337 NAME: emulate_httpd_log
1341 LOC: Config.onoff.common_log
1343 The Cache can emulate the log file format which many 'httpd'
1344 programs use. To disable/enable this emulation, set
1345 emulate_httpd_log to 'off' or 'on'. The default
1346 is to use the native log format since it includes useful
1347 information Squid-specific log analyzers use.
1350 NAME: log_ip_on_direct
1354 LOC: Config.onoff.log_ip_on_direct
1356 Log the destination IP address in the hierarchy log tag when going
1357 direct. Earlier Squid versions logged the hostname here. If you
1358 prefer the old way set this to off.
1363 DEFAULT: @DEFAULT_MIME_TABLE@
1364 LOC: Config.mimeTablePathname
1366 Pathname to Squid's MIME table. You shouldn't need to change
1367 this, but the default file contains examples and formatting
1368 information if you do.
1375 LOC: Config.onoff.log_mime_hdrs
1378 The Cache can record both the request and the response MIME
1379 headers for each HTTP transaction. The headers are encoded
1380 safely and will appear as two bracketed fields at the end of
1381 the access log (for either the native or httpd-emulated log
1382 formats). To enable this logging set log_mime_hdrs to 'on'.
1388 LOC: Config.Log.useragent
1390 IFDEF: USE_USERAGENT_LOG
1392 Squid will write the User-Agent field from HTTP requests
1393 to the filename specified here. By default useragent_log
1400 LOC: Config.Log.referer
1402 IFDEF: USE_REFERER_LOG
1404 Squid will write the Referer field from HTTP requests to the
1405 filename specified here. By default referer_log is disabled.
1411 DEFAULT: @DEFAULT_PID_FILE@
1412 LOC: Config.pidFilename
1414 A filename to write the process-id to. To disable, enter "none".
1421 LOC: Config.debugOptions
1423 Logging options are set as section,level where each source file
1424 is assigned a unique section. Lower levels result in less
1425 output, Full debugging (level 9) can result in a very large
1426 log file, so be careful. The magic word "ALL" sets debugging
1427 levels for all sections. We recommend normally running with
1436 LOC: Config.onoff.log_fqdn
1438 Turn this on if you wish to log fully qualified domain names
1439 in the access.log. To do this Squid does a DNS lookup of all
1440 IP's connecting to it. This can (in some situations) increase
1441 latency, which makes your cache seem slower for interactive
1446 NAME: client_netmask
1448 LOC: Config.Addrs.client_netmask
1449 DEFAULT: 255.255.255.255
1451 A netmask for client addresses in logfiles and cachemgr output.
1452 Change this to protect the privacy of your cache clients.
1453 A netmask of 255.255.255.0 will log all IP's in that range with
1454 the last digit set to '0'.
1459 OPTIONS FOR EXTERNAL SUPPORT PROGRAMS
1460 -----------------------------------------------------------------------------
1466 LOC: Config.Ftp.anon_user
1468 If you want the anonymous login password to be more informative
1469 (and enable the use of picky ftp servers), set this to something
1470 reasonable for your domain, like wwwuser@somewhere.net
1472 The reason why this is domainless by default is the
1473 request can be made on the behalf of a user in any domain,
1474 depending on how the cache is used.
1475 Some ftp server also validate the email address is valid
1476 (for example perl.com).
1479 NAME: ftp_list_width
1482 LOC: Config.Ftp.list_width
1484 Sets the width of ftp listings. This should be set to fit in
1485 the width of a standard browser. Setting this too small
1486 can cut off long filenames when browsing ftp sites.
1492 LOC: Config.Ftp.passive
1494 If your firewall does not allow Squid to use passive
1495 connections, turn off this option.
1498 NAME: ftp_sanitycheck
1501 LOC: Config.Ftp.sanitycheck
1503 For security and data integrity reasons Squid by default performs
1504 sanity checks of the addresses of FTP data connections ensure the
1505 data connection is to the requested server. If you need to allow
1506 FTP connections to servers using another IP address for the data
1507 connection turn this off.
1510 NAME: check_hostnames
1513 LOC: Config.onoff.check_hostnames
1515 For security and stability reasons Squid can check
1516 hostnames for Internet standard RFC compliance. If you want
1517 Squid to perform these checks turn this directive on.
1520 NAME: allow_underscore
1523 LOC: Config.onoff.allow_underscore
1525 Underscore characers is not strictly allowed in Internet hostnames
1526 but nevertheless used by many sites. Set this to off if you want
1527 Squid to be strict about the standard.
1528 This check is performed only when check_hostnames is set to on.
1531 NAME: ftp_telnet_protocol
1534 LOC: Config.Ftp.telnet
1536 The FTP protocol is officially defined to use the telnet protocol
1537 as transport channel for the control connection. However, many
1538 implementations are broken and does not respect this aspect of
1541 If you have trouble accessing files with ASCII code 255 in the
1542 path or similar problems involving this ASCII code you can
1543 try setting this directive to off. If that helps, report to the
1544 operator of the FTP server in question that their FTP server
1545 is broken and does not follow the FTP standard.
1548 NAME: cache_dns_program
1550 IFDEF: USE_DNSSERVERS
1551 DEFAULT: @DEFAULT_DNSSERVER@
1552 LOC: Config.Program.dnsserver
1554 Specify the location of the executable for dnslookup process.
1559 IFDEF: USE_DNSSERVERS
1561 LOC: Config.dnsChildren
1563 The number of processes spawn to service DNS name lookups.
1564 For heavily loaded caches on large servers, you should
1565 probably increase this value to at least 10. The maximum
1566 is 32. The default is 5.
1568 You must have at least one dnsserver process.
1571 NAME: dns_retransmit_interval
1574 LOC: Config.Timeout.idns_retransmit
1575 IFDEF: !USE_DNSSERVERS
1577 Initial retransmit interval for DNS queries. The interval is
1578 doubled each time all configured DNS servers have been tried.
1585 LOC: Config.Timeout.idns_query
1586 IFDEF: !USE_DNSSERVERS
1588 DNS Query timeout. If no response is received to a DNS query
1589 within this time all DNS servers for the queried domain
1590 are assumed to be unavailable.
1597 LOC: Config.onoff.res_defnames
1599 Normally the RES_DEFNAMES resolver option is disabled
1600 (see res_init(3)). This prevents caches in a hierarchy
1601 from interpreting single-component hostnames locally. To allow
1602 Squid to handle single-component names, enable this option.
1605 NAME: dns_nameservers
1608 LOC: Config.dns_nameservers
1610 Use this if you want to specify a list of DNS name servers
1611 (IP addresses) to use instead of those given in your
1612 /etc/resolv.conf file.
1613 On Windows platforms, if no value is specified here or in
1614 the /etc/resolv.conf file, the list of DNS name servers are
1615 taken from the Windows registry, both static and dynamic DHCP
1616 configurations are supported.
1618 Example: dns_nameservers 10.0.0.1 192.172.0.4
1623 DEFAULT: @DEFAULT_HOSTS@
1624 LOC: Config.etcHostsPath
1626 Location of the host-local IP name-address associations
1627 database. Most Operating Systems have such a file on different
1629 - Un*X & Linux: /etc/hosts
1630 - Windows NT/2000: %SystemRoot%\system32\drivers\etc\hosts
1631 (%SystemRoot% value install default is c:\winnt)
1632 - Windows XP/2003: %SystemRoot%\system32\drivers\etc\hosts
1633 (%SystemRoot% value install default is c:\windows)
1634 - Windows 9x/Me: %windir%\hosts
1635 (%windir% value is usually c:\windows)
1636 - Cygwin: /etc/hosts
1638 The file contains newline-separated definitions, in the
1639 form ip_address_in_dotted_form name [name ...] names are
1640 whitespace-separated. Lines beginning with an hash (#)
1641 character are comments.
1643 The file is checked at startup and upon configuration.
1644 If set to 'none', it won't be checked.
1645 If append_domain is used, that domain will be added to
1646 domain-local (i.e. not containing any dot character) host
1652 DEFAULT: @DEFAULT_DISKD@
1653 LOC: Config.Program.diskd
1655 Specify the location of the diskd executable.
1656 Note this is only useful if you have compiled in
1657 diskd as one of the store io modules.
1660 NAME: unlinkd_program
1663 DEFAULT: @DEFAULT_UNLINKD@
1664 LOC: Config.Program.unlinkd
1666 Specify the location of the executable for file deletion process.
1669 NAME: pinger_program
1671 DEFAULT: @DEFAULT_PINGER@
1672 LOC: Config.Program.pinger
1675 Specify the location of the executable for the pinger process.
1679 NAME: url_rewrite_program redirect_program
1681 LOC: Config.Program.redirect
1684 Specify the location of the executable for the URL redirector.
1685 Since they can perform almost any function there isn't one included.
1686 See the FAQ (section 15) for information on how to write one.
1687 By default, a redirector is not used.
1691 NAME: url_rewrite_children redirect_children
1694 LOC: Config.redirectChildren
1696 The number of redirector processes to spawn. If you start
1697 too few Squid will have to wait for them to process a backlog of
1698 URLs, slowing it down. If you start too many they will use RAM
1699 and other system resources.
1702 NAME: url_rewrite_concurrency redirect_concurrency
1705 LOC: Config.redirectConcurrency
1707 The number of requests each redirector helper can handle in
1708 parallell. Defaults to 0 which indicates the redirector
1709 is a old-style singlethreaded redirector.
1712 NAME: url_rewrite_host_header redirect_rewrites_host_header
1715 LOC: Config.onoff.redir_rewrites_host
1717 By default Squid rewrites any Host: header in redirected
1718 requests. If you are running an accelerator this may
1719 not be a wanted effect of a redirector.
1721 WARNING: Entries are cached on the result of the URL rewriting
1722 process, so be careful if you have domain-virtual hosts.
1725 NAME: redirector_access
1728 LOC: Config.accessList.redirector
1730 If defined, this access list specifies which requests are
1731 sent to the redirector processes. By default all requests
1737 LOC: Config.authConfiguration
1740 This is used to pass parameters to the various authentication
1742 format: auth_param scheme parameter [setting]
1744 auth_param basic program @DEFAULT_PREFIX@/bin/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
1745 would tell the basic authentication scheme it's program parameter.
1747 The order authentication prompts are presented to the client_agent
1748 is dependent on the order the scheme first appears in config file.
1749 IE has a bug (it's not rfc 2617 compliant) in that it will use the basic
1750 scheme if basic is the first entry presented, even if more secure schemes
1751 are presented. For now use the order in the file below. If other browsers
1752 have difficulties (don't recognize the schemes offered even if you are using
1753 basic) either put basic first, or disable the other schemes (by commenting
1754 out their program entry).
1756 Once an authentication scheme is fully configured, it can only be shutdown
1757 by shutting squid down and restarting. Changes can be made on the fly and
1758 activated with a reconfigure. I.E. You can change to a different helper,
1759 but not unconfigure the helper completely.
1761 === Parameters for the basic scheme follow. ===
1764 Specify the command for the external authenticator. Such a
1765 program reads a line containing "username password" and replies
1766 "ERR" in an endless loop. "ERR" responses may optionally be followed
1767 by a error description available as %m in the returned error page.
1768 If you use an authenticator, make sure you have 1 acl of type proxy_auth.
1769 By default, the basic authentication scheme is not used unless a program
1772 If you want to use the traditional proxy authentication,
1773 jump over to the ../helpers/basic_auth/NCSA directory and
1778 Then, set this line to something like
1780 auth_param basic program @DEFAULT_PREFIX@/libexec/ncsa_auth @DEFAULT_PREFIX@/etc/passwd
1782 "children" numberofchildren
1783 The number of authenticator processes to spawn (no default).
1784 If you start too few Squid will have to wait for them to
1785 process a backlog of usercode/password verifications, slowing
1786 it down. When password verifications are done via a (slow)
1787 network you are likely to need lots of authenticator
1789 auth_param basic children 5
1791 "concurrency" concurrency
1792 The number of concurrent requests the helper can process.
1793 The default of 0 is used for helpers who only supports
1794 one request at a time.
1795 auth_param basic concurrency 0
1798 Specifies the realm name which is to be reported to the
1799 client for the basic proxy authentication scheme (part of
1800 the text the user will see when prompted their username and
1801 password). There is no default.
1802 auth_param basic realm Squid proxy-caching web server
1804 "credentialsttl" timetolive
1805 Specifies how long squid assumes an externally validated
1806 username:password pair is valid for - in other words how
1807 often the helper program is called for that user. Set this
1808 low to force revalidation with short lived passwords. Note
1809 setting this high does not impact your susceptibility
1810 to replay attacks unless you are using an one-time password
1811 system (such as SecureID). If you are using such a system,
1812 you will be vulnerable to replay attacks unless you also
1813 use the max_user_ip ACL in an http_access rule.
1815 "casesensitive" on|off
1816 Specifies if usernames are case sensitive. Most user databases are
1817 case insensitive allowing the same username to be spelled using both
1818 lower and upper case letters, but some are case sensitive. This
1819 makes a big difference for user_max_ip ACL processing and similar.
1820 auth_param basic casesensitive off
1822 === Parameters for the digest scheme follow ===
1825 Specify the command for the external authenticator. Such
1826 a program reads a line containing "username":"realm" and
1827 replies with the appropriate H(A1) value hex encoded or
1828 ERR if the user (or his H(A1) hash) does not exists.
1829 See rfc 2616 for the definition of H(A1).
1830 "ERR" responses may optionally be followed by a error description
1831 available as %m in the returned error page.
1833 By default, the digest authentication is not used unless a
1834 program is specified.
1836 If you want to use a digest authenticator, jump over to the
1837 helpers/digest_auth/ directory and choose the authenticator
1838 to use. In it's directory type
1842 Then, set this line to something like
1844 auth_param digest program @DEFAULT_PREFIX@/bin/digest_auth_pw @DEFAULT_PREFIX@/etc/digpass
1847 "children" numberofchildren
1848 The number of authenticator processes to spawn (no default).
1849 If you start too few Squid will have to wait for them to
1850 process a backlog of H(A1) calculations, slowing it down.
1851 When the H(A1) calculations are done via a (slow) network
1852 you are likely to need lots of authenticator processes.
1853 auth_param digest children 5
1856 Specifies the realm name which is to be reported to the
1857 client for the digest proxy authentication scheme (part of
1858 the text the user will see when prompted their username and
1859 password). There is no default.
1860 auth_param digest realm Squid proxy-caching web server
1862 "nonce_garbage_interval" timeinterval
1863 Specifies the interval that nonces that have been issued
1864 to client_agent's are checked for validity.
1866 "nonce_max_duration" timeinterval
1867 Specifies the maximum length of time a given nonce will be
1870 "nonce_max_count" number
1871 Specifies the maximum number of times a given nonce can be
1874 "nonce_strictness" on|off
1875 Determines if squid requires strict increment-by-1 behavior
1876 for nonce counts, or just incrementing (off - for use when
1877 useragents generate nonce counts that occasionally miss 1
1878 (ie, 1,2,4,6)). Default off.
1880 "check_nonce_count" on|off
1881 This directive if set to off can disable the nonce count check
1882 completely to work around buggy digest qop implementations in
1883 certain mainstream browser versions. Default on to check the
1884 nonce count to protect from authentication replay attacks.
1886 "post_workaround" on|off
1887 This is a workaround to certain buggy browsers who sends
1888 an incorrect request digest in POST requests when reusing
1889 the same nonce as acquired earlier on a GET request.
1891 === NTLM scheme options follow ===
1894 Specify the command for the external NTLM authenticator.
1895 Such a program reads exchanged NTLMSSP packets with
1896 the browser via Squid until authentication is completed.
1897 If you use an NTLM authenticator, make sure you have 1 acl
1898 of type proxy_auth. By default, the NTLM authenticator_program
1901 auth_param ntlm program @DEFAULT_PREFIX@/bin/ntlm_auth
1903 "children" numberofchildren
1904 The number of authenticator processes to spawn (no default).
1905 If you start too few Squid will have to wait for them to
1906 process a backlog of credential verifications, slowing it
1907 down. When credential verifications are done via a (slow)
1908 network you are likely to need lots of authenticator
1911 auth_param ntlm children 5
1914 If you experience problems with PUT/POST requests when using the
1915 Negotiate authentication scheme then you can try setting this to
1916 off. This will cause Squid to forcibly close the connection on
1917 the initial requests where the browser asks which schemes are
1918 supported by the proxy.
1920 auth_param ntlm keep_alive on
1922 === Options for configuring the NEGOTIATE auth-scheme follow ===
1925 Specify the command for the external Negotiate authenticator.
1926 This protocol is used in Microsoft Active-Directory enabled setups with
1927 the Microsoft Internet Explorer or Mozilla Firefox browsers.
1928 Its main purpose is to exchange credentials with the Squid proxy
1929 using the Kerberos mechanisms.
1930 If you use a Negotiate authenticator, make sure you have at least one acl
1931 of type proxy_auth active. By default, the negotiate authenticator_program
1933 The only supported program for this role is the ntlm_auth
1934 program distributed as part of Samba, version 4 or later.
1936 auth_param negotiate program @DEFAULT_PREFIX@/bin/ntlm_auth --helper-protocol=gss-spnego
1938 "children" numberofchildren
1939 The number of authenticator processes to spawn (no default).
1940 If you start too few Squid will have to wait for them to
1941 process a backlog of credential verifications, slowing it
1942 down. When crendential verifications are done via a (slow)
1943 network you are likely to need lots of authenticator
1945 auth_param negotiate children 5
1948 If you experience problems with PUT/POST requests when using the
1949 Negotiate authentication scheme then you can try setting this to
1950 off. This will cause Squid to forcibly close the connection on
1951 the initial requests where the browser asks which schemes are
1952 supported by the proxy.
1954 auth_param negotiate keep_alive on
1957 #Recommended minimum configuration per scheme:
1958 #auth_param negotiate program <uncomment and complete this line to activate>
1959 #auth_param negotiate children 5
1960 #auth_param negotiate keep_alive on
1961 #auth_param ntlm program <uncomment and complete this line to activate>
1962 #auth_param ntlm children 5
1963 #auth_param ntlm keep_alive on
1964 #auth_param digest program <uncomment and complete this line>
1965 #auth_param digest children 5
1966 #auth_param digest realm Squid proxy-caching web server
1967 #auth_param digest nonce_garbage_interval 5 minutes
1968 #auth_param digest nonce_max_duration 30 minutes
1969 #auth_param digest nonce_max_count 50
1970 #auth_param basic program <uncomment and complete this line>
1971 #auth_param basic children 5
1972 #auth_param basic realm Squid proxy-caching web server
1973 #auth_param basic credentialsttl 2 hours
1977 NAME: authenticate_cache_garbage_interval
1980 LOC: Config.authenticateGCInterval
1982 The time period between garbage collection across the
1983 username cache. This is a tradeoff between memory utilization
1984 (long intervals - say 2 days) and CPU (short intervals -
1985 say 1 minute). Only change if you have good reason to.
1988 NAME: authenticate_ttl
1991 LOC: Config.authenticateTTL
1993 The time a user & their credentials stay in the logged in
1994 user cache since their last request. When the garbage
1995 interval passes, all user credentials that have passed their
1996 TTL are removed from memory.
1999 NAME: authenticate_ip_ttl
2001 LOC: Config.authenticateIpTTL
2004 If you use proxy authentication and the 'max_user_ip' ACL,
2005 this directive controls how long Squid remembers the IP
2006 addresses associated with each user. Use a small value
2007 (e.g., 60 seconds) if your users might change addresses
2008 quickly, as is the case with dialups. You might be safe
2009 using a larger value (e.g., 2 hours) in a corporate LAN
2010 environment with relatively static address assignments.
2013 NAME: external_acl_type
2014 TYPE: externalAclHelper
2015 LOC: Config.externalAclHelperList
2018 This option defines external acl classes using a helper program
2019 to look up the status
2021 external_acl_type name [options] FORMAT.. /path/to/helper [helper arguments..]
2025 ttl=n TTL in seconds for cached results (defaults to 3600
2028 TTL for cached negative lookups (default same
2030 children=n Number of acl helper processes spawn to service
2031 external acl lookups of this type.
2032 concurrency=n concurrency level per process. Use 0 for old style
2033 helpers who can only process a single request at a
2035 cache=n result cache size, 0 is unbounded (default)
2036 grace=n Percentage remaining of TTL where a refresh of a
2037 cached entry should be initiated without needing to
2038 wait for a new reply. (default 0 for no grace period)
2039 protocol=2.5 Compatibility mode for Squid-2.5 external acl helpers
2041 FORMAT specifications
2043 %LOGIN Authenticated user login name
2044 %EXT_USER Username from external acl
2045 %IDENT Ident user name
2047 %SRCPORT Client source port
2049 %PROTO Requested protocol
2050 %PORT Requested port
2051 %PATH Requested URL path
2052 %METHOD Request method
2053 %MYADDR Squid interface address
2054 %MYPORT Squid http_port number
2055 %USER_CERT SSL User certificate in PEM format
2056 %USER_CERTCHAIN SSL User certificate chain in PEM format
2057 %USER_CERT_xx SSL User certificate subject attribute xx
2058 %USER_CA_xx SSL User certificate issuer attribute xx
2059 %{Header} HTTP request header
2060 %{Hdr:member} HTTP request header list member
2062 HTTP request header list member using ; as
2063 list separator. ; can be any non-alphanumeric
2066 In addition, any string specified in the referencing acl will
2067 also be included in the helper request line, after the specified
2068 formats (see the "acl external" directive)
2070 The helper receives lines per the above format specification,
2071 and returns lines starting with OK or ERR indicating the validity
2072 of the request and optionally followed by additional keywords with
2073 more details. To protect from odd characters the data is URL
2076 General result syntax:
2078 OK/ERR keyword=value ...
2082 user= The users name (login)
2083 password= The users password (for login= cache_peer option)
2084 message= Message describing the reason. Available as %o
2086 tag= Apply a tag to a request (for both ERR and OK results)
2087 Only sets a tag, does not alter existing tags.
2088 log= String to be logged in access.log. Available as
2089 %ea in logformat specifications
2091 Keyword values need to be URL escaped if they may contain
2092 contain whitespace or quotes.
2094 In Squid-2.5 compatibility mode quoting using " and \ is used
2095 instead of URL escaping.
2099 OPTIONS FOR TUNING THE CACHE
2100 -----------------------------------------------------------------------------
2103 NAME: request_header_max_size
2107 LOC: Config.maxRequestHeaderSize
2109 This specifies the maximum size for HTTP headers in a request.
2110 Request headers are usually relatively small (about 512 bytes).
2111 Placing a limit on the request header size will catch certain
2112 bugs (for example with persistent connections) and possibly
2113 buffer-overflow or denial-of-service attacks.
2116 NAME: request_body_max_size
2120 LOC: Config.maxRequestBodySize
2122 This specifies the maximum size for an HTTP request body.
2123 In other words, the maximum size of a PUT/POST request.
2124 A user who attempts to send a request with a body larger
2125 than this limit receives an "Invalid Request" error message.
2126 If you set this parameter to a zero (the default), there will
2127 be no limit imposed.
2130 NAME: refresh_pattern
2131 TYPE: refreshpattern
2135 usage: refresh_pattern [-i] regex min percent max [options]
2137 By default, regular expressions are CASE-SENSITIVE. To make
2138 them case-insensitive, use the -i option.
2140 'Min' is the time (in minutes) an object without an explicit
2141 expiry time should be considered fresh. The recommended
2142 value is 0, any higher values may cause dynamic applications
2143 to be erroneously cached unless the application designer
2144 has taken the appropriate actions.
2146 'Percent' is a percentage of the objects age (time since last
2147 modification age) an object without explicit expiry time
2148 will be considered fresh.
2150 'Max' is an upper limit on how long objects without an explicit
2151 expiry time will be considered fresh.
2153 options: override-expire
2163 override-expire enforces min age even if the server
2164 sent a Expires: header. Doing this VIOLATES the HTTP
2165 standard. Enabling this feature could make you liable
2166 for problems which it causes.
2168 override-lastmod enforces min age even on objects
2169 that were modified recently.
2171 reload-into-ims changes client no-cache or ``reload''
2172 to If-Modified-Since requests. Doing this VIOLATES the
2173 HTTP standard. Enabling this feature could make you
2174 liable for problems which it causes.
2176 ignore-reload ignores a client no-cache or ``reload''
2177 header. Doing this VIOLATES the HTTP standard. Enabling
2178 this feature could make you liable for problems which
2181 ignore-no-cache ignores any ``Pragma: no-cache'' and
2182 ``Cache-control: no-cache'' headers received from a server.
2183 The HTTP RFC never allows the use of this (Pragma) header
2184 from a server, only a client, though plenty of servers
2187 ignore-no-store ignores any ``Cache-control: no-store''
2188 headers received from a server. Doing this VIOLATES
2189 the HTTP standard. Enabling this feature could make you
2190 liable for problems which it causes.
2192 ignore-private ignores any ``Cache-control: private''
2193 headers received from a server. Doing this VIOLATES
2194 the HTTP standard. Enabling this feature could make you
2195 liable for problems which it causes.
2197 ignore-auth caches responses to requests with authorization,
2198 irrespective of ``Cache-control'' headers received from
2199 a server. Doing this VIOLATES the HTTP standard. Enabling
2200 this feature could make you liable for problems which
2203 refresh-ims causes squid to contact the origin server
2204 when a client issues an If-Modified-Since request. This
2205 ensures that the client will receive an updated version
2206 if one is available.
2208 Basically a cached object is:
2210 FRESH if expires < now, else STALE
2212 FRESH if lm-factor < percent, else STALE
2216 The refresh_pattern lines are checked in the order listed here.
2217 The first entry which matches is used. If none of the entries
2218 match the default will be used.
2220 Note, you must uncomment all the default lines if you want
2221 to change one. The default setting is only active if none is
2226 refresh_pattern ^ftp: 1440 20% 10080
2227 refresh_pattern ^gopher: 1440 0% 1440
2228 refresh_pattern . 0 20% 4320
2232 NAME: quick_abort_min
2236 LOC: Config.quickAbort.min
2239 NAME: quick_abort_max
2243 LOC: Config.quickAbort.max
2246 NAME: quick_abort_pct
2250 LOC: Config.quickAbort.pct
2252 The cache by default continues downloading aborted requests
2253 which are almost completed (less than 16 KB remaining). This
2254 may be undesirable on slow (e.g. SLIP) links and/or very busy
2255 caches. Impatient users may tie up file descriptors and
2256 bandwidth by repeatedly requesting and immediately aborting
2259 When the user aborts a request, Squid will check the
2260 quick_abort values to the amount of data transfered until
2263 If the transfer has less than 'quick_abort_min' KB remaining,
2264 it will finish the retrieval.
2266 If the transfer has more than 'quick_abort_max' KB remaining,
2267 it will abort the retrieval.
2269 If more than 'quick_abort_pct' of the transfer has completed,
2270 it will finish the retrieval.
2272 If you do not want any retrieval to continue after the client
2273 has aborted, set both 'quick_abort_min' and 'quick_abort_max'
2276 If you want retrievals to always continue if they are being
2277 cached set 'quick_abort_min' to '-1 KB'.
2280 NAME: read_ahead_gap
2281 COMMENT: buffer-size
2283 LOC: Config.readAheadGap
2286 The amount of data the cache will buffer ahead of what has been
2287 sent to the client when retrieving an object from another server.
2293 LOC: Config.negativeTtl
2296 Time-to-Live (TTL) for failed requests. Certain types of
2297 failures (such as "connection refused" and "404 Not Found") are
2298 negatively-cached for a configurable amount of time. The
2299 default is 5 minutes. Note that this is different from
2300 negative caching of DNS lookups.
2304 NAME: positive_dns_ttl
2307 LOC: Config.positiveDnsTtl
2310 Time-to-Live (TTL) for positive caching of successful DNS lookups.
2311 Default is 6 hours (360 minutes). If you want to minimize the
2312 use of Squid's ipcache, set this to 1, not 0.
2316 NAME: negative_dns_ttl
2319 LOC: Config.negativeDnsTtl
2322 Time-to-Live (TTL) for negative caching of failed DNS lookups.
2325 NAME: range_offset_limit
2328 LOC: Config.rangeOffsetLimit
2331 Sets a upper limit on how far into the the file a Range request
2332 may be to cause Squid to prefetch the whole file. If beyond this
2333 limit Squid forwards the Range request as it is and the result
2336 This is to stop a far ahead range request (lets say start at 17MB)
2337 from making Squid fetch the whole object up to that point before
2338 sending anything to the client.
2340 A value of -1 causes Squid to always fetch the object from the
2341 beginning so it may cache the result. (2.0 style)
2343 A value of 0 causes Squid to never fetch more than the
2344 client requested. (default)
2350 -----------------------------------------------------------------------------
2353 NAME: forward_timeout
2356 LOC: Config.Timeout.forward
2359 This parameter specifies how long Squid should at most attempt in
2360 finding a forwarding path for the request before giving up.
2363 NAME: connect_timeout
2366 LOC: Config.Timeout.connect
2369 This parameter specifies how long to wait for the TCP connect to
2370 the requested server or peer to complete before Squid should
2371 attempt to find another path where to forward the request.
2374 NAME: peer_connect_timeout
2377 LOC: Config.Timeout.peer_connect
2380 This parameter specifies how long to wait for a pending TCP
2381 connection to a peer cache. The default is 30 seconds. You
2382 may also set different timeout values for individual neighbors
2383 with the 'connect-timeout' option on a 'cache_peer' line.
2389 LOC: Config.Timeout.read
2392 The read_timeout is applied on server-side connections. After
2393 each successful read(), the timeout will be extended by this
2394 amount. If no data is read again after this amount of time,
2395 the request is aborted and logged with ERR_READ_TIMEOUT. The
2396 default is 15 minutes.
2400 NAME: request_timeout
2402 LOC: Config.Timeout.request
2405 How long to wait for an HTTP request after initial
2406 connection establishment.
2410 NAME: persistent_request_timeout
2412 LOC: Config.Timeout.persistent_request
2415 How long to wait for the next HTTP request on a persistent
2416 connection after the previous request completes.
2420 NAME: client_lifetime
2423 LOC: Config.Timeout.lifetime
2426 The maximum amount of time a client (browser) is allowed to
2427 remain connected to the cache process. This protects the Cache
2428 from having a lot of sockets (and hence file descriptors) tied up
2429 in a CLOSE_WAIT state from remote clients that go away without
2430 properly shutting down (either because of a network failure or
2431 because of a poor client implementation). The default is one
2434 NOTE: The default value is intended to be much larger than any
2435 client would ever need to be connected to your cache. You
2436 should probably change client_lifetime only as a last resort.
2437 If you seem to have many client connections tying up
2438 filedescriptors, we recommend first tuning the read_timeout,
2439 request_timeout, persistent_request_timeout and quick_abort values.
2442 NAME: half_closed_clients
2444 LOC: Config.onoff.half_closed_clients
2447 Some clients may shutdown the sending side of their TCP
2448 connections, while leaving their receiving sides open. Sometimes,
2449 Squid can not tell the difference between a half-closed and a
2450 fully-closed TCP connection. By default, half-closed client
2451 connections are kept open until a read(2) or write(2) on the
2452 socket returns an error. Change this option to 'off' and Squid
2453 will immediately close client connections when read(2) returns
2454 "no more data to read."
2459 LOC: Config.Timeout.pconn
2460 DEFAULT: 120 seconds
2462 Timeout for idle persistent connections to servers and other
2469 LOC: Config.Timeout.ident
2472 Maximum time to wait for IDENT lookups to complete.
2474 If this is too high, and you enabled IDENT lookups from untrusted
2475 users, you might be susceptible to denial-of-service by having
2476 many ident requests going at once.
2480 NAME: shutdown_lifetime
2483 LOC: Config.shutdownLifetime
2486 When SIGTERM or SIGHUP is received, the cache is put into
2487 "shutdown pending" mode until all active sockets are closed.
2488 This value is the lifetime to set for all open descriptors
2489 during shutdown mode. Any active clients after this many
2490 seconds will receive a 'timeout' message.
2495 -----------------------------------------------------------------------------
2503 Defining an Access List
2505 acl aclname acltype string1 ...
2506 acl aclname acltype "file" ...
2508 when using "file", the file should contain one item per line
2510 acltype is one of the types described below
2512 By default, regular expressions are CASE-SENSITIVE. To make
2513 them case-insensitive, use the -i option.
2515 acl aclname src ip-address/netmask ... (clients IP address)
2516 acl aclname src addr1-addr2/netmask ... (range of addresses)
2517 acl aclname dst ip-address/netmask ... (URL host's IP address)
2518 acl aclname myip ip-address/netmask ... (local socket IP address)
2520 acl aclname arp mac-address ... (xx:xx:xx:xx:xx:xx notation)
2521 # The arp ACL requires the special configure option --enable-arp-acl.
2522 # Furthermore, the ARP ACL code is not portable to all operating systems.
2523 # It works on Linux, Solaris, Windows, FreeBSD, and some other *BSD variants.
2525 # NOTE: Squid can only determine the MAC address for clients that are on
2526 # the same subnet. If the client is on a different subnet, then Squid cannot
2527 # find out its MAC address.
2529 acl aclname srcdomain .foo.com ... # reverse lookup, client IP
2530 acl aclname dstdomain .foo.com ... # Destination server from URL
2531 acl aclname srcdom_regex [-i] xxx ... # regex matching client name
2532 acl aclname dstdom_regex [-i] xxx ... # regex matching server
2533 # For dstdomain and dstdom_regex a reverse lookup is tried if a IP
2534 # based URL is used and no match is found. The name "none" is used
2535 # if the reverse lookup fails.
2537 acl aclname http_status 200 301 500- 400-403 ... # status code in reply
2539 acl aclname time [day-abbrevs] [h1:m1-h2:m2]
2548 h1:m1 must be less than h2:m2
2549 acl aclname url_regex [-i] ^http:// ... # regex matching on whole URL
2550 acl aclname urlpath_regex [-i] \.gif$ ... # regex matching on URL path
2551 acl aclname port 80 70 21 ...
2552 acl aclname port 0-1024 ... # ranges allowed
2553 acl aclname myport 3128 ... # (local socket TCP port)
2554 acl aclname proto HTTP FTP ...
2555 acl aclname method GET POST ...
2556 acl aclname browser [-i] regexp ...
2557 # pattern match on User-Agent header (see also req_header below)
2558 acl aclname referer_regex [-i] regexp ...
2559 # pattern match on Referer header
2560 # Referer is highly unreliable, so use with care
2561 acl aclname ident username ...
2562 acl aclname ident_regex [-i] pattern ...
2563 # string match on ident output.
2564 # use REQUIRED to accept any non-null ident.
2565 acl aclname src_as number ...
2566 acl aclname dst_as number ...
2567 # Except for access control, AS numbers can be used for
2568 # routing of requests to specific caches. Here's an
2569 # example for routing all requests for AS#1241 and only
2570 # those to mycache.mydomain.net:
2571 # acl asexample dst_as 1241
2572 # cache_peer_access mycache.mydomain.net allow asexample
2573 # cache_peer_access mycache_mydomain.net deny all
2575 acl aclname proxy_auth [-i] username ...
2576 acl aclname proxy_auth_regex [-i] pattern ...
2577 # list of valid usernames
2578 # use REQUIRED to accept any valid username.
2580 # NOTE: when a Proxy-Authentication header is sent but it is not
2581 # needed during ACL checking the username is NOT logged
2584 # NOTE: proxy_auth requires a EXTERNAL authentication program
2585 # to check username/password combinations (see
2586 # auth_param directive).
2588 # NOTE: proxy_auth can't be used in a transparent proxy as
2589 # the browser needs to be configured for using a proxy in order
2590 # to respond to proxy authentication.
2592 acl aclname snmp_community string ...
2593 # A community string to limit access to your SNMP Agent
2596 # acl snmppublic snmp_community public
2598 acl aclname maxconn number
2599 # This will be matched when the client's IP address has
2600 # more than <number> HTTP connections established.
2602 acl aclname max_user_ip [-s] number
2603 # This will be matched when the user attempts to log in from more
2604 # than <number> different ip addresses. The authenticate_ip_ttl
2605 # parameter controls the timeout on the ip entries.
2606 # If -s is specified the limit is strict, denying browsing
2607 # from any further IP addresses until the ttl has expired. Without
2608 # -s Squid will just annoy the user by "randomly" denying requests.
2609 # (the counter is reset each time the limit is reached and a
2610 # request is denied)
2611 # NOTE: in acceleration mode or where there is mesh of child proxies,
2612 # clients may appear to come from multiple addresses if they are
2613 # going through proxy farms, so a limit of 1 may cause user problems.
2615 acl aclname rep_mime_type mime-type1 ...
2616 # regex match against the mime type of the reply received by
2617 # squid. Can be used to detect file download or some
2618 # types HTTP tunneling requests.
2619 # NOTE: This has no effect in http_access rules. It only has
2620 # effect in rules that affect the reply data stream such as
2621 # http_reply_access.
2623 acl aclname rep_header header-name [-i] any\.regex\.here
2624 # regex match against any of the known reply headers. May be
2625 # thought of as a superset of "browser", "referer" and "mime-type"
2628 acl aclname req_mime_type mime-type1 ...
2629 # regex match agains the mime type of the request generated
2630 # by the client. Can be used to detect file upload or some
2631 # types HTTP tunneling requests.
2632 # NOTE: This does NOT match the reply. You cannot use this
2633 # to match the returned file type.
2635 acl aclname req_header header-name [-i] any\.regex\.here
2636 # regex match against any of the known request headers. May be
2637 # thought of as a superset of "browser", "referer" and "mime-type"
2640 acl acl_name external class_name [arguments...]
2641 # external ACL lookup via a helper class defined by the
2642 # external_acl_type directive.
2644 acl aclname user_cert attribute values...
2645 # match against attributes in a user SSL certificate
2646 # attribute is one of DN/C/O/CN/L/ST
2648 acl aclname ca_cert attribute values...
2649 # match against attributes a users issuing CA SSL certificate
2650 # attribute is one of DN/C/O/CN/L/ST
2652 acl aclname ext_user username ...
2653 acl aclname ext_user_regex [-i] pattern ...
2654 # string match on username returned by external acl processing
2655 # use REQUIRED to accept any non-null user name.
2658 acl macaddress arp 09:00:2b:23:45:67
2659 acl myexample dst_as 1241
2660 acl password proxy_auth REQUIRED
2661 acl fileupload req_mime_type -i ^multipart/form-data$
2662 acl javascript rep_mime_type -i ^application/x-javascript$
2665 #Recommended minimum configuration:
2666 acl all src 0.0.0.0/0.0.0.0
2667 acl manager proto cache_object
2668 acl localhost src 127.0.0.1/255.255.255.255
2669 acl to_localhost dst 127.0.0.0/8
2670 acl SSL_ports port 443
2671 acl Safe_ports port 80 # http
2672 acl Safe_ports port 21 # ftp
2673 acl Safe_ports port 443 # https
2674 acl Safe_ports port 70 # gopher
2675 acl Safe_ports port 210 # wais
2676 acl Safe_ports port 1025-65535 # unregistered ports
2677 acl Safe_ports port 280 # http-mgmt
2678 acl Safe_ports port 488 # gss-http
2679 acl Safe_ports port 591 # filemaker
2680 acl Safe_ports port 777 # multiling http
2681 acl CONNECT method CONNECT
2687 LOC: Config.accessList.http
2689 DEFAULT_IF_NONE: deny all
2691 Allowing or Denying access based on defined access lists
2693 Access to the HTTP port:
2694 http_access allow|deny [!]aclname ...
2696 NOTE on default values:
2698 If there are no "access" lines present, the default is to deny
2701 If none of the "access" lines cause a match, the default is the
2702 opposite of the last line in the list. If the last line was
2703 deny, the default is allow. Conversely, if the last line
2704 is allow, the default will be deny. For these reasons, it is a
2705 good idea to have an "deny all" or "allow all" entry at the end
2706 of your access lists to avoid potential confusion.
2709 #Recommended minimum configuration:
2711 # Only allow cachemgr access from localhost
2712 http_access allow manager localhost
2713 http_access deny manager
2714 # Deny requests to unknown ports
2715 http_access deny !Safe_ports
2716 # Deny CONNECT to other than SSL ports
2717 http_access deny CONNECT !SSL_ports
2719 # We strongly recommend the following be uncommented to protect innocent
2720 # web applications running on the proxy server who think the only
2721 # one who can access services on "localhost" is a local user
2722 #http_access deny to_localhost
2724 # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
2726 # Example rule allowing access from your local networks. Adapt
2727 # to list your (internal) IP networks from where browsing should
2729 #acl our_networks src 192.168.1.0/24 192.168.2.0/24
2730 #http_access allow our_networks
2732 # And finally deny all other access to this proxy
2733 http_access deny all
2737 NAME: http_reply_access
2739 LOC: Config.accessList.reply
2741 DEFAULT_IF_NONE: allow all
2743 Allow replies to client requests. This is complementary to http_access.
2745 http_reply_access allow|deny [!] aclname ...
2747 NOTE: if there are no access lines present, the default is to allow
2750 If none of the access lines cause a match the opposite of the
2751 last line will apply. Thus it is good practice to end the rules
2752 with an "allow all" or "deny all" entry.
2755 #Recommended minimum configuration:
2757 # Insert your own rules here.
2760 # and finally allow by default
2761 http_reply_access allow all
2768 LOC: Config.accessList.icp
2770 DEFAULT_IF_NONE: deny all
2772 Allowing or Denying access to the ICP port based on defined
2775 icp_access allow|deny [!]aclname ...
2777 See http_access for details
2780 #Allow ICP queries from everyone
2781 icp_access allow all
2789 LOC: Config.accessList.htcp
2791 DEFAULT_IF_NONE: deny all
2793 Allowing or Denying access to the HTCP port based on defined
2796 htcp_access allow|deny [!]aclname ...
2798 See http_access for details
2800 #Allow HTCP queries from everyone
2801 htcp_access allow all
2804 NAME: htcp_clr_access
2807 LOC: Config.accessList.htcp_clr
2809 DEFAULT_IF_NONE: deny all
2811 Allowing or Denying access to purge content using HTCP based
2812 on defined access lists
2814 htcp_clr_access allow|deny [!]aclname ...
2816 See http_access for details
2818 #Allow HTCP CLR requests from trusted peers
2819 acl htcp_clr_peer src 172.16.1.2
2820 htcp_clr_access allow htcp_clr_peer
2826 LOC: Config.accessList.miss
2829 Use to force your neighbors to use you as a sibling instead of
2830 a parent. For example:
2832 acl localclients src 172.16.0.0/16
2833 miss_access allow localclients
2834 miss_access deny !localclients
2836 This means only your local clients are allowed to fetch
2837 MISSES and all other clients can only fetch HITS.
2839 By default, allow all clients who passed the http_access rules
2840 to fetch MISSES from us.
2844 # miss_access allow all
2849 NAME: cache_peer_access
2854 Similar to 'cache_peer_domain' but provides more flexibility by
2857 cache_peer_access cache-host allow|deny [!]aclname ...
2859 The syntax is identical to 'http_access' and the other lists of
2860 ACL elements. See the comments for 'http_access' below, or
2861 the Squid FAQ (http://www.squid-cache.org/FAQ/FAQ-10.html).
2864 NAME: ident_lookup_access
2868 DEFAULT_IF_NONE: deny all
2869 LOC: Config.accessList.identLookup
2871 A list of ACL elements which, if matched, cause an ident
2872 (RFC 931) lookup to be performed for this request. For
2873 example, you might choose to always perform ident lookups
2874 for your main multi-user Unix boxes, but not for your Macs
2875 and PCs. By default, ident lookups are not performed for
2878 To enable ident lookups for specific client addresses, you
2879 can follow this example:
2881 acl ident_aware_hosts src 198.168.1.0/255.255.255.0
2882 ident_lookup_access allow ident_aware_hosts
2883 ident_lookup_access deny all
2885 Only src type ACL checks are fully supported. A src_domain
2886 ACL might work at times, but it will not always provide
2890 NAME: tcp_outgoing_tos tcp_outgoing_ds tcp_outgoing_dscp
2893 LOC: Config.accessList.outgoing_tos
2895 Allows you to select a TOS/Diffserv value to mark outgoing
2896 connections with, based on the username or source address
2899 tcp_outgoing_tos ds-field [!]aclname ...
2901 Example where normal_service_net uses the TOS value 0x00
2902 and normal_service_net uses 0x20
2904 acl normal_service_net src 10.0.0.0/255.255.255.0
2905 acl good_service_net src 10.0.1.0/255.255.255.0
2906 tcp_outgoing_tos 0x00 normal_service_net 0x00
2907 tcp_outgoing_tos 0x20 good_service_net
2909 TOS/DSCP values really only have local significance - so you should
2910 know what you're specifying. For more, see RFC 2474
2912 The TOS/DSCP byte must be exactly that - a byte, value 0 - 255, or
2913 "default" to use whatever default your host has.
2915 Processing proceeds in the order specified, and stops at first fully
2919 NAME: tcp_outgoing_address
2922 LOC: Config.accessList.outgoing_address
2924 Allows you to map requests to different outgoing IP addresses
2925 based on the username or sourceaddress of the user making
2928 tcp_outgoing_address ipaddr [[!]aclname] ...
2930 Example where requests from 10.0.0.0/24 will be forwarded
2931 with source address 10.1.0.1, 10.0.2.0/24 forwarded with
2932 source address 10.1.0.2 and the rest will be forwarded with
2933 source address 10.1.0.3.
2935 acl normal_service_net src 10.0.0.0/255.255.255.0
2936 acl good_service_net src 10.0.1.0/255.255.255.0
2937 tcp_outgoing_address 10.0.0.1 normal_service_net
2938 tcp_outgoing_address 10.0.0.2 good_service_net
2939 tcp_outgoing_address 10.0.0.3
2941 Processing proceeds in the order specified, and stops at first fully
2945 NAME: reply_header_max_size
2949 LOC: Config.maxReplyHeaderSize
2951 This specifies the maximum size for HTTP headers in a reply.
2952 Reply headers are usually relatively small (about 512 bytes).
2953 Placing a limit on the reply header size will catch certain
2954 bugs (for example with persistent connections) and possibly
2955 buffer-overflow or denial-of-service attacks.
2958 NAME: reply_body_max_size
2959 COMMENT: size [acl acl...]
2962 LOC: Config.ReplyBodySize
2964 This option specifies the maximum size of a reply body. It can be
2965 used to prevent users from downloading very large files, such as
2966 MP3's and movies. When the reply headers are received, the
2967 reply_body_max_size lines are processed, and the first line where
2968 all (if any) listed ACLs are true is used as the maximum body size
2971 This size is checked twice. First when we get the reply headers,
2972 we check the content-length value. If the content length value exists
2973 and is larger than the allowed size, the request is denied and the
2974 user receives an error message that says "the request or reply
2975 is too large." If there is no content-length, and the reply
2976 size exceeds this limit, the client's connection is just closed
2977 and they will receive a partial reply.
2979 WARNING: downstream caches probably can not detect a partial reply
2980 if there is no content-length header, so they will cache
2981 partial responses and give them out as hits. You should NOT
2982 use this option if you have downstream caches.
2984 WARNING: A maximum size smaller than the size of squid's error messages
2985 will cause an infinite loop and crash squid. Ensure that the smallest
2986 non-zero value you use is greater that the maximum header size plus
2987 the size of your largest error page.
2989 If you set this parameter none (the default), there will be
2995 LOC: Config.accessList.log
2997 COMMENT: allow|deny acl acl...
2999 This options allows you to control which requests gets logged
3000 to access.log (see access_log directive). Requests denied for
3001 logging will also not be accounted for in performance counters.
3005 ADMINISTRATIVE PARAMETERS
3006 -----------------------------------------------------------------------------
3012 LOC: Config.adminEmail
3014 Email-address of local cache manager who will receive
3015 mail if the cache dies. The default is "webmaster."
3022 LOC: Config.EmailFrom
3024 From: email-address for mail sent when the cache dies.
3025 The default is to use 'appname@unique_hostname'.
3026 Default appname value is "squid", can be changed into
3027 src/globals.h before building squid.
3034 LOC: Config.EmailProgram
3036 Email program used to send mail if the cache dies.
3037 The default is "mail". The specified program must complain
3038 with the standard Unix mail syntax:
3039 mail_program recipient < mailfile
3040 Optional command line options can be specified.
3044 NAME: cache_effective_user
3047 LOC: Config.effectiveUser
3049 If you start Squid as root, it will change its effective/real
3050 UID/GID to the user specified below. The default is to change
3051 to UID to nobody. If you define cache_effective_user, but not
3052 cache_effective_group, Squid sets the GID to the effective
3053 user's default group ID (taken from the password file) and
3054 supplementary group list from the from groups membership of
3055 cache_effective_user.
3059 NAME: cache_effective_group
3062 LOC: Config.effectiveGroup
3064 If you want Squid to run with a specific GID regardless of
3065 the group memberships of the effective user then set this
3066 to the group (or GID) you want Squid to run as. When set
3067 all other group privileges of the effective user is ignored
3068 and only this GID is effective. If Squid is not started as
3069 root the user starting Squid must be member of the specified
3074 NAME: httpd_suppress_version_string
3078 LOC: Config.onoff.httpd_suppress_version_string
3080 Suppress Squid version string info in HTTP headers and HTML error pages.
3084 NAME: visible_hostname
3086 LOC: Config.visibleHostname
3089 If you want to present a special hostname in error messages, etc,
3090 define this. Otherwise, the return value of gethostname()
3091 will be used. If you have multiple caches in a cluster and
3092 get errors about IP-forwarding you must set them to have individual
3093 names with this setting.
3097 NAME: unique_hostname
3099 LOC: Config.uniqueHostname
3102 If you want to have multiple machines with the same
3103 'visible_hostname' you must give each machine a different
3104 'unique_hostname' so forwarding loops can be detected.
3108 NAME: hostname_aliases
3110 LOC: Config.hostnameAliases
3113 A list of other DNS names your cache has.
3117 OPTIONS FOR THE CACHE REGISTRATION SERVICE
3118 -----------------------------------------------------------------------------
3120 This section contains parameters for the (optional) cache
3121 announcement service. This service is provided to help
3122 cache administrators locate one another in order to join or
3123 create cache hierarchies.
3125 An 'announcement' message is sent (via UDP) to the registration
3126 service by Squid. By default, the announcement message is NOT
3127 SENT unless you enable it with 'announce_period' below.
3129 The announcement message includes your hostname, plus the
3130 following information from this configuration file:
3136 All current information is processed regularly and made
3137 available on the Web at http://www.ircache.net/Cache/Tracker/.
3140 NAME: announce_period
3142 LOC: Config.Announce.period
3145 This is how frequently to send cache announcements. The
3146 default is `0' which disables sending the announcement
3149 To enable announcing your cache, just uncomment the line
3153 #To enable announcing your cache, just uncomment the line below.
3154 #announce_period 1 day
3161 DEFAULT: tracker.ircache.net
3162 LOC: Config.Announce.host
3168 LOC: Config.Announce.file
3174 LOC: Config.Announce.port
3176 announce_host and announce_port set the hostname and port
3177 number where the registration message will be sent.
3179 Hostname will default to 'tracker.ircache.net' and port will
3180 default default to 3131. If the 'filename' argument is given,
3181 the contents of that file will be included in the announce
3185 NAME: httpd_accel_surrogate_id
3188 LOC: Config.Accel.surrogate_id
3191 Surrogates (http://www.esi.org/architecture_spec_1.0.html)
3192 need an identification token to allow control targeting. Because
3193 a farm of surrogates may all perform the same tasks, they may share
3194 an identification token.
3197 NAME: http_accel_surrogate_remote
3202 LOC: Config.onoff.surrogate_is_remote
3204 Remote surrogates (such as those in a CDN) honour Surrogate-Control: no-store-remote.
3205 Set this to on to have squid behave as a remote surrogate.
3210 COMMENT: libxml2|expat|custom
3212 LOC: ESIParser::Type
3215 ESI markup is not strictly XML compatible. The custom ESI parser
3216 will give higher performance, but cannot handle non ASCII character
3222 -----------------------------------------------------------------------------
3227 LOC: Config.dns_testname_list
3229 DEFAULT_IF_NONE: netscape.com internic.net nlanr.net microsoft.com
3231 The DNS tests exit as soon as the first site is successfully looked up
3233 This test can be disabled with the -D command line option.
3237 NAME: logfile_rotate
3240 LOC: Config.Log.rotateNumber
3242 Specifies the number of logfile rotations to make when you
3243 type 'squid -k rotate'. The default is 10, which will rotate
3244 with extensions 0 through 9. Setting logfile_rotate to 0 will
3245 disable the file name rotation, but the logfiles are still closed
3246 and re-opened. This will enable you to rename the logfiles
3247 yourself just before sending the rotate signal.
3249 Note, the 'squid -k rotate' command normally sends a USR1
3250 signal to the running squid process. In certain situations
3251 (e.g. on Linux with Async I/O), USR1 is used for other
3252 purposes, so -k rotate uses another signal. It is best to get
3253 in the habit of using 'squid -k rotate' instead of 'kill -USR1
3260 LOC: Config.appendDomain
3263 Appends local domain name to hostnames without any dots in
3264 them. append_domain must begin with a period.
3266 Be warned there are now Internet names with no dots in
3267 them using only top-domain names, so setting this may
3268 cause some Internet sites to become unavailable.
3271 append_domain .yourdomain.com
3275 NAME: tcp_recv_bufsize
3279 LOC: Config.tcpRcvBufsz
3281 Size of receive buffer to set for TCP sockets. Probably just
3282 as easy to change your kernel's default. Set to zero to use
3283 the default buffer size.
3288 LOC: Config.errHtmlText
3291 HTML text to include in error messages. Make this a "mailto"
3292 URL to your admin address, or maybe just a link to your
3293 organizations Web page.
3295 To include this in your error messages, you must rewrite
3296 the error template files (found in the "errors" directory).
3297 Wherever you want the 'err_html_text' line to appear,
3298 insert a %L tag in the error template file.
3301 NAME: email_err_data
3304 LOC: Config.onoff.emailErrData
3307 If enabled, information about the occurred error will be
3308 included in the mailto links of the ERR pages (if %W is set)
3309 so that the email body contains the data.
3310 Syntax is <A HREF="mailto:%w%W">%w</A>
3316 LOC: Config.denyInfoList
3319 Usage: deny_info err_page_name acl
3320 or deny_info http://... acl
3321 Example: deny_info ERR_CUSTOM_ACCESS_DENIED bad_guys
3323 This can be used to return a ERR_ page for requests which
3324 do not pass the 'http_access' rules. A single ACL will cause
3325 the http_access check to fail. If a 'deny_info' line exists
3326 for that ACL Squid returns a corresponding error page.
3328 You may use ERR_ pages that come with Squid or create your own pages
3329 and put them into the configured errors/ directory.
3331 Alternatively you can specify an error URL. The browsers will
3332 get redirected (302) to the specified URL. %s in the redirection
3333 URL will be replaced by the requested URL.
3335 Alternatively you can tell Squid to reset the TCP connection
3336 by specifying TCP_RESET.
3343 LOC: Config.onoff.mem_pools
3345 If set, Squid will keep pools of allocated (but unused) memory
3346 available for future use. If memory is a premium on your
3347 system and you believe your malloc library outperforms Squid
3348 routines, disable this.
3351 NAME: memory_pools_limit
3355 LOC: Config.MemPools.limit
3357 Used only with memory_pools on:
3358 memory_pools_limit 50 MB
3360 If set to a non-zero value, Squid will keep at most the specified
3361 limit of allocated (but unused) memory in memory pools. All free()
3362 requests that exceed this limit will be handled by your malloc
3363 library. Squid does not pre-allocate any memory, just safe-keeps
3364 objects that otherwise would be free()d. Thus, it is safe to set
3365 memory_pools_limit to a reasonably high value even if your
3366 configuration will use less memory.
3368 If set to zero, Squid will keep all memory it can. That is, there
3369 will be no limit on the total amount of memory used for safe-keeping.
3371 To disable memory allocation optimization, do not set
3372 memory_pools_limit to 0. Set memory_pools to "off" instead.
3374 An overhead for maintaining memory pools is not taken into account
3375 when the limit is checked. This overhead is close to four bytes per
3376 object kept. However, pools may actually _save_ memory because of
3377 reduced memory thrashing in your malloc library.
3381 IFDEF: HTTP_VIOLATIONS
3385 LOC: Config.onoff.via
3387 If set (default), Squid will include a Via header in requests and
3388 replies as required by RFC2616.
3395 LOC: opt_forwarded_for
3397 If set, Squid will include your system's IP address or name
3398 in the HTTP requests it forwards. By default it looks like
3401 X-Forwarded-For: 192.1.2.3
3403 If you disable this, it will appear as
3405 X-Forwarded-For: unknown
3408 NAME: log_icp_queries
3412 LOC: Config.onoff.log_udp
3414 If set, ICP queries are logged to access.log. You may wish
3415 do disable this if your ICP load is VERY high to speed things
3416 up or to simplify log analysis.
3423 LOC: Config.onoff.icp_hit_stale
3425 If you want to return ICP_HIT for stale cache objects, set this
3426 option to 'on'. If you have sibling relationships with caches
3427 in other administrative domains, this should be 'off'. If you only
3428 have sibling relationships with caches under your control,
3429 it is probably okay to set this to 'on'.
3430 If set to 'on', your siblings should use the option "allow-miss"
3431 on their cache_peer lines for connecting to you.
3435 NAME: minimum_direct_hops
3438 LOC: Config.minDirectHops
3440 If using the ICMP pinging stuff, do direct fetches for sites
3441 which are no more than this many hops away.
3444 NAME: minimum_direct_rtt
3447 LOC: Config.minDirectRtt
3449 If using the ICMP pinging stuff, do direct fetches for sites
3450 which are no more than this many rtt milliseconds away.
3453 NAME: cachemgr_passwd
3454 TYPE: cachemgrpasswd
3456 LOC: Config.passwd_list
3458 Specify passwords for cachemgr operations.
3460 Usage: cachemgr_passwd password action action ...
3462 Some valid actions are (see cache manager menu for a full list):
3501 * Indicates actions which will not be performed without a
3502 valid password, others can be performed if not listed here.
3504 To disable an action, set the password to "disable".
3505 To allow performing an action without a password, set the
3508 Use the keyword "all" to set the same password for all actions.
3511 cachemgr_passwd secret shutdown
3512 cachemgr_passwd lesssssssecret info stats/objects
3513 cachemgr_passwd disable all
3516 NAME: store_avg_object_size
3520 LOC: Config.Store.avgObjectSize
3522 Average object size, used to estimate number of objects your
3523 cache can hold. See doc/Release-Notes-1.1.txt. The default is
3527 NAME: store_objects_per_bucket
3530 LOC: Config.Store.objectsPerBucket
3532 Target number of objects per bucket in the store hash table.
3533 Lowering this value increases the total number of buckets and
3534 also the storage maintenance rate. The default is 20.
3541 LOC: Config.onoff.client_db
3543 If you want to disable collecting per-client statistics,
3544 turn off client_db here.
3551 LOC: Config.Netdb.low
3557 LOC: Config.Netdb.high
3559 The low and high water marks for the ICMP measurement
3560 database. These are counts, not percents. The defaults are
3561 900 and 1000. When the high water mark is reached, database
3562 entries will be deleted until the low mark is reached.
3566 NAME: netdb_ping_period
3568 LOC: Config.Netdb.period
3571 The minimum period for measuring a site. There will be at
3572 least this much delay between successive pings to the same
3573 network. The default is five minutes.
3581 LOC: Config.onoff.query_icmp
3583 If you want to ask your peers to include ICMP data in their ICP
3584 replies, enable this option.
3586 If your peer has configured Squid (during compilation) with
3587 '--enable-icmp' that peer will send ICMP pings to origin server
3588 sites of the URLs it receives. If you enable this option the
3589 ICP replies from that peer will include the ICMP data (if available).
3590 Then, when choosing a parent cache, Squid will choose the parent with
3591 the minimal RTT to the origin server. When this happens, the
3592 hierarchy field of the access.log will be
3593 "CLOSEST_PARENT_MISS". This option is off by default.
3596 NAME: test_reachability
3600 LOC: Config.onoff.test_reachability
3602 When this is 'on', ICP MISS replies will be ICP_MISS_NOFETCH
3603 instead of ICP_MISS if the target host is NOT in the ICMP
3604 database, or has a zero RTT.
3611 LOC: Config.onoff.buffered_logs
3613 cache.log log file is written with stdio functions, and as such
3614 it can be buffered or unbuffered. By default it will be unbuffered.
3615 Buffering it can speed up the writing slightly (though you are
3616 unlikely to need to worry unless you run with tons of debugging
3617 enabled in which case performance will suffer badly anyway..).
3620 NAME: refresh_all_ims
3624 LOC: Config.onoff.refresh_all_ims
3626 When you enable this option, squid will always check
3627 the origin server for an update when a client sends an
3628 If-Modified-Since request. Many browsers use IMS
3629 requests when the user requests a reload, and this
3630 ensures those clients receive the latest version.
3632 By default (off), squid may return a Not Modified response
3633 based on the age of the cached version.
3636 NAME: reload_into_ims
3637 IFDEF: HTTP_VIOLATIONS
3641 LOC: Config.onoff.reload_into_ims
3643 When you enable this option, client no-cache or ``reload''
3644 requests will be changed to If-Modified-Since requests.
3645 Doing this VIOLATES the HTTP standard. Enabling this
3646 feature could make you liable for problems which it
3649 see also refresh_pattern for a more selective approach.
3654 LOC: Config.accessList.AlwaysDirect
3657 Usage: always_direct allow|deny [!]aclname ...
3659 Here you can use ACL elements to specify requests which should
3660 ALWAYS be forwarded directly to origin servers. For example,
3661 to always directly forward requests for local servers use
3664 acl local-servers dstdomain my.domain.net
3665 always_direct allow local-servers
3667 To always forward FTP requests directly, use
3670 always_direct allow FTP
3672 NOTE: There is a similar, but opposite option named
3673 'never_direct'. You need to be aware that "always_direct deny
3674 foo" is NOT the same thing as "never_direct allow foo". You
3675 may need to use a deny rule to exclude a more-specific case of
3676 some other rule. Example:
3678 acl local-external dstdomain external.foo.net
3679 acl local-servers dstdomain .foo.net
3680 always_direct deny local-external
3681 always_direct allow local-servers
3683 This option replaces some v1.1 options such as local_domain
3689 LOC: Config.accessList.NeverDirect
3692 Usage: never_direct allow|deny [!]aclname ...
3694 never_direct is the opposite of always_direct. Please read
3695 the description for always_direct if you have not already.
3697 With 'never_direct' you can use ACL elements to specify
3698 requests which should NEVER be forwarded directly to origin
3699 servers. For example, to force the use of a proxy for all
3700 requests, except those in your local domain use something like:
3702 acl local-servers dstdomain .foo.net
3703 acl all src 0.0.0.0/0.0.0.0
3704 never_direct deny local-servers
3705 never_direct allow all
3707 or if Squid is inside a firewall and there are local intranet
3708 servers inside the firewall use something like:
3710 acl local-intranet dstdomain .foo.net
3711 acl local-external dstdomain external.foo.net
3712 always_direct deny local-external
3713 always_direct allow local-intranet
3714 never_direct allow all
3716 This option replaces some v1.1 options such as inside_firewall
3720 NAME: request_header_access
3721 IFDEF: HTTP_VIOLATIONS
3722 TYPE: http_header_access[]
3723 LOC: Config.request_header_access
3726 Usage: request_header_access header_name allow|deny [!]aclname ...
3728 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3729 this feature could make you liable for problems which it
3732 This option replaces the old 'anonymize_headers' and the
3733 older 'http_anonymizer' option with something that is much
3734 more configurable. This new method creates a list of ACLs
3735 for each header, allowing you very fine-tuned header
3738 This option only applies to request headers, i.e., from the
3739 client to the server.
3741 You can only specify known headers for the header name.
3742 Other headers are reclassified as 'Other'. You can also
3743 refer to all the headers with 'All'.
3745 For example, to achieve the same behavior as the old
3746 'http_anonymizer standard' option, you should use:
3748 request_header_access From deny all
3749 request_header_access Referer deny all
3750 request_header_access Server deny all
3751 request_header_access User-Agent deny all
3752 request_header_access WWW-Authenticate deny all
3753 request_header_access Link deny all
3755 Or, to reproduce the old 'http_anonymizer paranoid' feature
3758 request_header_access Allow allow all
3759 request_header_access Authorization allow all
3760 request_header_access WWW-Authenticate allow all
3761 request_header_access Proxy-Authorization allow all
3762 request_header_access Proxy-Authenticate allow all
3763 request_header_access Cache-Control allow all
3764 request_header_access Content-Encoding allow all
3765 request_header_access Content-Length allow all
3766 request_header_access Content-Type allow all
3767 request_header_access Date allow all
3768 request_header_access Expires allow all
3769 request_header_access Host allow all
3770 request_header_access If-Modified-Since allow all
3771 request_header_access Last-Modified allow all
3772 request_header_access Location allow all
3773 request_header_access Pragma allow all
3774 request_header_access Accept allow all
3775 request_header_access Accept-Charset allow all
3776 request_header_access Accept-Encoding allow all
3777 request_header_access Accept-Language allow all
3778 request_header_access Content-Language allow all
3779 request_header_access Mime-Version allow all
3780 request_header_access Retry-After allow all
3781 request_header_access Title allow all
3782 request_header_access Connection allow all
3783 request_header_access Proxy-Connection allow all
3784 request_header_access All deny all
3786 although many of those are HTTP reply headers, and so should be
3787 controlled with the reply_header_access directive.
3789 By default, all headers are allowed (no anonymizing is
3793 NAME: reply_header_access
3794 IFDEF: HTTP_VIOLATIONS
3795 TYPE: http_header_access[]
3796 LOC: Config.reply_header_access
3799 Usage: reply_header_access header_name allow|deny [!]aclname ...
3801 WARNING: Doing this VIOLATES the HTTP standard. Enabling
3802 this feature could make you liable for problems which it
3805 This option only applies to reply headers, i.e., from the
3806 server to the client.
3808 This is the same as request_header_access, but in the other
3811 This option replaces the old 'anonymize_headers' and the
3812 older 'http_anonymizer' option with something that is much
3813 more configurable. This new method creates a list of ACLs
3814 for each header, allowing you very fine-tuned header
3817 You can only specify known headers for the header name.
3818 Other headers are reclassified as 'Other'. You can also
3819 refer to all the headers with 'All'.
3821 For example, to achieve the same behavior as the old
3822 'http_anonymizer standard' option, you should use:
3824 reply_header_access From deny all
3825 reply_header_access Referer deny all
3826 reply_header_access Server deny all
3827 reply_header_access User-Agent deny all
3828 reply_header_access WWW-Authenticate deny all
3829 reply_header_access Link deny all
3831 Or, to reproduce the old 'http_anonymizer paranoid' feature
3834 reply_header_access Allow allow all
3835 reply_header_access Authorization allow all
3836 reply_header_access WWW-Authenticate allow all
3837 reply_header_access Proxy-Authorization allow all
3838 reply_header_access Proxy-Authenticate allow all
3839 reply_header_access Cache-Control allow all
3840 reply_header_access Content-Encoding allow all
3841 reply_header_access Content-Length allow all
3842 reply_header_access Content-Type allow all
3843 reply_header_access Date allow all
3844 reply_header_access Expires allow all
3845 reply_header_access Host allow all
3846 reply_header_access If-Modified-Since allow all
3847 reply_header_access Last-Modified allow all
3848 reply_header_access Location allow all
3849 reply_header_access Pragma allow all
3850 reply_header_access Accept allow all
3851 reply_header_access Accept-Charset allow all
3852 reply_header_access Accept-Encoding allow all
3853 reply_header_access Accept-Language allow all
3854 reply_header_access Content-Language allow all
3855 reply_header_access Mime-Version allow all
3856 reply_header_access Retry-After allow all
3857 reply_header_access Title allow all
3858 reply_header_access Connection allow all
3859 reply_header_access Proxy-Connection allow all
3860 reply_header_access All deny all
3862 although the HTTP request headers won't be usefully controlled
3863 by this directive -- see request_header_access for details.
3865 By default, all headers are allowed (no anonymizing is
3869 NAME: header_replace
3870 IFDEF: HTTP_VIOLATIONS
3871 TYPE: http_header_replace[]
3872 LOC: Config.request_header_access
3875 Usage: header_replace header_name message
3876 Example: header_replace User-Agent Nutscrape/1.0 (CP/M; 8-bit)
3878 This option allows you to change the contents of headers
3879 denied with header_access above, by replacing them with
3880 some fixed string. This replaces the old fake_user_agent
3883 This only applies to request headers, not reply headers.
3885 By default, headers are removed if denied.
3888 NAME: icon_directory
3890 LOC: Config.icons.directory
3891 DEFAULT: @DEFAULT_ICON_DIR@
3893 Where the icons are stored. These are normally kept in
3897 NAME: global_internal_static
3899 LOC: Config.onoff.global_internal_static
3902 This directive controls is Squid should intercept all requests for
3903 /squid-internal-static/ no matter which host the URL is requesting
3904 (default on setting), or if nothing special should be done for
3905 such URLs (off setting). The purpose of this directive is to make
3906 icons etc work better in complex cache hierarchies where it may
3907 not always be possible for all corners in the cache mesh to reach
3908 the server generating a directory listing.
3911 NAME: short_icon_urls
3913 LOC: Config.icons.use_short_names
3916 If this is enabled Squid will use short URLs for icons.
3917 If disabled it will revert to the old behavior of including
3918 it's own name and port in the URL.
3920 If you run a complex cache hierarchy with a mix of Squid and
3921 other proxies you may need to disable this directive.
3924 NAME: error_directory
3926 LOC: Config.errorDirectory
3927 DEFAULT: @DEFAULT_ERROR_DIR@
3929 If you wish to create your own versions of the default
3930 (English) error files, either to customize them to suit your
3931 language or company copy the template English files to another
3932 directory and point this tag at them.
3935 NAME: maximum_single_addr_tries
3937 LOC: Config.retry.maxtries
3940 This sets the maximum number of connection attempts for a
3941 host that only has one address (for multiple-address hosts,
3942 each address is tried once).
3944 The default value is one attempt, the (not recommended)
3945 maximum is 255 tries. A warning message will be generated
3946 if it is set to a value greater than ten.
3948 Note: This is in addition to the request re-forwarding which
3949 takes place if Squid fails to get a satisfying response.
3952 NAME: retry_on_error
3954 LOC: Config.retry.onerror
3957 If set to on Squid will automatically retry requests when
3958 receiving an error response. This is mainly useful if you
3959 are in a complex cache hierarchy to work around access
3965 LOC: Config.Port.snmp
3969 Squid can now serve statistics and status information via SNMP.
3970 By default it listens to port 3401 on the machine. If you don't
3971 wish to use SNMP, set this to "0".
3973 Note: If you want Squid to use parents for all requests see
3974 the never_direct directive. prefer_direct only modifies how Squid
3975 acts on cachable requests.
3980 LOC: Config.accessList.snmp
3982 DEFAULT_IF_NONE: deny all
3985 Allowing or denying access to the SNMP port.
3987 All access to the agent is denied by default.
3990 snmp_access allow|deny [!]aclname ...
3993 snmp_access allow snmppublic localhost
3994 snmp_access deny all
3997 NAME: snmp_incoming_address
3999 LOC: Config.Addrs.snmp_incoming
4003 NAME: snmp_outgoing_address
4005 LOC: Config.Addrs.snmp_outgoing
4006 DEFAULT: 255.255.255.255
4009 Just like 'udp_incoming_address' above, but for the SNMP port.
4011 snmp_incoming_address is used for the SNMP socket receiving
4012 messages from SNMP agents.
4013 snmp_outgoing_address is used for SNMP packets returned to SNMP
4016 The default snmp_incoming_address (0.0.0.0) is to listen on all
4017 available network interfaces.
4019 If snmp_outgoing_address is set to 255.255.255.255 (the default)
4020 it will use the same socket as snmp_incoming_address. Only
4021 change this if you want to have SNMP replies sent using another
4022 address than where this Squid listens for SNMP queries.
4024 NOTE, snmp_incoming_address and snmp_outgoing_address can not have
4025 the same value since they both use port 3401.
4028 NAME: as_whois_server
4030 LOC: Config.as_whois_server
4031 DEFAULT: whois.ra.net
4032 DEFAULT_IF_NONE: whois.ra.net
4034 WHOIS server to query for AS numbers. NOTE: AS numbers are
4035 queried only when Squid starts up, not for every request.
4040 LOC: Config.Wccp.router
4045 TYPE: sockaddr_in_list
4046 LOC: Config.Wccp2.router
4050 Use this option to define your WCCP ``home'' router for
4053 wccp_router supports a single WCCP(v1) router
4055 wccp2_router supports multiple WCCPv2 routers
4057 only one of the two may be used at the same time and defines
4058 which version of WCCP to use.
4063 LOC: Config.Wccp.version
4067 This directive is only relevant if you need to set up WCCP(v1)
4068 to some very old and end-of-life Cisco routers. In all other
4069 setups it must be left unset or at the default setting.
4070 It defines an internal version in the WCCP(v1) protocol,
4071 with version 4 being the officially documented protocol.
4073 According to some users, Cisco IOS 11.2 and earlier only
4074 support WCCP version 3. If you're using that or an earlier
4075 version of IOS, you may need to change this value to 3, otherwise
4076 do not specify this parameter.
4079 NAME: wccp2_rebuild_wait
4081 LOC: Config.Wccp2.rebuildwait
4085 If this is enabled Squid will wait for the cache dir rebuild to finish
4086 before sending the first wccp2 HereIAm packet
4089 NAME: wccp2_forwarding_method
4091 LOC: Config.Wccp2.forwarding_method
4095 WCCP2 allows the setting of forwarding methods between the
4096 router/switch and the cache. Valid values are as follows:
4098 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4099 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4101 Currently (as of IOS 12.4) cisco routers only support GRE.
4102 Cisco switches only support the L2 redirect assignment method.
4105 NAME: wccp2_return_method
4107 LOC: Config.Wccp2.return_method
4111 WCCP2 allows the setting of return methods between the
4112 router/switch and the cache for packets that the cache
4113 decides not to handle. Valid values are as follows:
4115 1 - GRE encapsulation (forward the packet in a GRE/WCCP tunnel)
4116 2 - L2 redirect (forward the packet using Layer 2/MAC rewriting)
4118 Currently (as of IOS 12.4) cisco routers only support GRE.
4119 Cisco switches only support the L2 redirect assignment.
4121 If the "ip wccp redirect exclude in" command has been
4122 enabled on the cache interface, then it is still safe for
4123 the proxy server to use a l2 redirect method even if this
4124 option is set to GRE.
4127 NAME: wccp2_assignment_method
4129 LOC: Config.Wccp2.assignment_method
4133 WCCP2 allows the setting of methods to assign the WCCP hash
4134 Valid values are as follows:
4139 As a general rule, cisco routers support the hash assignment method
4140 and cisco switches support the mask assignment method.
4145 LOC: Config.Wccp2.info
4147 DEFAULT_IF_NONE: standard 0
4150 WCCP2 allows for multiple traffic services. There are two
4151 types: "standard" and "dynamic". The standard type defines
4152 one service id - http (id 0). The dynamic service ids can be from
4153 51 to 255 inclusive. In order to use a dynamic service id
4154 one must define the type of traffic to be redirected; this is done
4155 using the wccp2_service_info option.
4157 The "standard" type does not require a wccp2_service_info option,
4158 just specifying the service id will suffice.
4160 MD5 service authentication can be enabled by adding
4161 "password=<password>" to the end of this service declaration.
4165 wccp2_service standard 0 # for the 'web-cache' standard service
4166 wccp2_service dynamic 80 # a dynamic service type which will be
4167 # fleshed out with subsequent options.
4168 wccp2_service standard 0 password=foo
4172 NAME: wccp2_service_info
4173 TYPE: wccp2_service_info
4174 LOC: Config.Wccp2.info
4178 Dynamic WCCPv2 services require further information to define the
4179 traffic you wish to have diverted.
4183 wccp2_service_info <id> protocol=<protocol> flags=<flag>,<flag>..
4184 priority=<priority> ports=<port>,<port>..
4186 The relevant WCCPv2 flags:
4187 + src_ip_hash, dst_ip_hash
4188 + source_port_hash, dest_port_hash
4189 + src_ip_alt_hash, dst_ip_alt_hash
4190 + src_port_alt_hash, dst_port_alt_hash
4193 The port list can be one to eight entries.
4197 wccp2_service_info 80 protocol=tcp flags=src_ip_hash,ports_source
4198 priority=240 ports=80
4200 Note: the service id must have been defined by a previous
4201 'wccp2_service dynamic <id>' entry.
4206 LOC: Config.Wccp2.weight
4210 Each cache server gets assigned a set of the destination
4211 hash proportional to their weight.
4216 LOC: Config.Wccp.address
4222 LOC: Config.Wccp2.address
4226 Use this option if you require WCCP to use a specific
4229 The default behavior is to not bind to any specific address.
4234 DELAY POOL PARAMETERS (all require DELAY_POOLS compilation option)
4235 -----------------------------------------------------------------------------
4239 TYPE: delay_pool_count
4244 This represents the number of delay pools to be used. For example,
4245 if you have one class 2 delay pool and one class 3 delays pool, you
4246 have a total of 2 delay pools.
4250 TYPE: delay_pool_class
4255 This defines the class of each delay pool. There must be exactly one
4256 delay_class line for each delay pool. For example, to define two
4257 delay pools, one of class 2 and one of class 3, the settings above
4261 delay_pools 4 # 4 delay pools
4262 delay_class 1 2 # pool 1 is a class 2 pool
4263 delay_class 2 3 # pool 2 is a class 3 pool
4264 delay_class 3 4 # pool 3 is a class 4 pool
4265 delay_class 4 5 # pool 4 is a class 5 pool
4267 The delay pool classes are:
4269 class 1 Everything is limited by a single aggregate
4272 class 2 Everything is limited by a single aggregate
4273 bucket as well as an "individual" bucket chosen
4274 from bits 25 through 32 of the IP address.
4276 class 3 Everything is limited by a single aggregate
4277 bucket as well as a "network" bucket chosen
4278 from bits 17 through 24 of the IP address and a
4279 "individual" bucket chosen from bits 17 through
4280 32 of the IP address.
4282 class 4 Everything in a class 3 delay pool, with an
4283 additional limit on a per user basis. This
4284 only takes effect if the username is established
4285 in advance - by forcing authentication in your
4288 class 5 Requests are grouped according their tag (see
4289 external_acl's tag= reply).
4291 NOTE: If an IP address is a.b.c.d
4292 -> bits 25 through 32 are "d"
4293 -> bits 17 through 24 are "c"
4294 -> bits 17 through 32 are "c * 256 + d"
4298 TYPE: delay_pool_access
4303 This is used to determine which delay pool a request falls into.
4305 delay_access is sorted per pool and the matching starts with pool 1,
4306 then pool 2, ..., and finally pool N. The first delay pool where the
4307 request is allowed is selected for the request. If it does not allow
4308 the request to any pool then the request is not delayed (default).
4310 For example, if you want some_big_clients in delay
4311 pool 1 and lotsa_little_clients in delay pool 2:
4314 delay_access 1 allow some_big_clients
4315 delay_access 1 deny all
4316 delay_access 2 allow lotsa_little_clients
4317 delay_access 2 deny all
4318 delay_access 3 allow authenticated_clients
4321 NAME: delay_parameters
4322 TYPE: delay_pool_rates
4327 This defines the parameters for a delay pool. Each delay pool has
4328 a number of "buckets" associated with it, as explained in the
4329 description of delay_class. For a class 1 delay pool, the syntax is:
4331 delay_parameters pool aggregate
4333 For a class 2 delay pool:
4335 delay_parameters pool aggregate individual
4337 For a class 3 delay pool:
4339 delay_parameters pool aggregate network individual
4341 For a class 4 delay pool:
4343 delay_parameters pool aggregate network individual user
4345 For a class 5 delay pool:
4347 delay_parameters pool tag
4349 The variables here are:
4351 pool a pool number - ie, a number between 1 and the
4352 number specified in delay_pools as used in
4355 aggregate the "delay parameters" for the aggregate bucket
4358 individual the "delay parameters" for the individual
4359 buckets (class 2, 3).
4361 network the "delay parameters" for the network buckets
4364 user the delay parameters for the user buckets
4367 tag the delay parameters for the tag buckets
4370 A pair of delay parameters is written restore/maximum, where restore is
4371 the number of bytes (not bits - modem and network speeds are usually
4372 quoted in bits) per second placed into the bucket, and maximum is the
4373 maximum number of bytes which can be in the bucket at any time.
4375 For example, if delay pool number 1 is a class 2 delay pool as in the
4376 above example, and is being used to strictly limit each host to 64kbps
4377 (plus overheads), with no overall limit, the line is:
4379 delay_parameters 1 -1/-1 8000/8000
4381 Note that the figure -1 is used to represent "unlimited".
4383 And, if delay pool number 2 is a class 3 delay pool as in the above
4384 example, and you want to limit it to a total of 256kbps (strict limit)
4385 with each 8-bit network permitted 64kbps (strict limit) and each
4386 individual host permitted 4800bps with a bucket maximum size of 64kb
4387 to permit a decent web page to be downloaded at a decent speed
4388 (if the network is not being limited due to overuse) but slow down
4389 large downloads more significantly:
4391 delay_parameters 2 32000/32000 8000/8000 600/8000
4393 There must be one delay_parameters line for each delay pool.
4395 Finally, for a class 4 delay pool as in the example - each user will
4396 be limited to 128Kb no matter how many workstations they are logged into.:
4398 delay_parameters 4 32000/32000 8000/8000 600/64000 16000/16000
4401 NAME: delay_initial_bucket_level
4402 COMMENT: (percent, 0-100)
4406 LOC: Config.Delay.initial
4408 The initial bucket percentage is used to determine how much is put
4409 in each bucket when squid starts, is reconfigured, or first notices
4410 a host accessing it (in class 2 and class 3, individual hosts and
4411 networks only have buckets associated with them once they have been
4415 NAME: incoming_icp_average
4418 LOC: Config.comm_incoming.icp_average
4421 NAME: incoming_http_average
4424 LOC: Config.comm_incoming.http_average
4427 NAME: incoming_dns_average
4430 LOC: Config.comm_incoming.dns_average
4433 NAME: min_icp_poll_cnt
4436 LOC: Config.comm_incoming.icp_min_poll
4439 NAME: min_dns_poll_cnt
4442 LOC: Config.comm_incoming.dns_min_poll
4445 NAME: min_http_poll_cnt
4448 LOC: Config.comm_incoming.http_min_poll
4450 Heavy voodoo here. I can't even believe you are reading this.
4451 Are you crazy? Don't even think about adjusting these unless
4452 you understand the algorithms in comm_select.c first!
4455 NAME: max_open_disk_fds
4457 LOC: Config.max_open_disk_fds
4460 To avoid having disk as the I/O bottleneck Squid can optionally
4461 bypass the on-disk cache if more than this amount of disk file
4462 descriptors are open.
4464 A value of 0 indicates no limit.
4469 LOC: Config.onoff.offline
4472 Enable this option and Squid will never try to validate cached
4476 NAME: uri_whitespace
4477 TYPE: uri_whitespace
4478 LOC: Config.uri_whitespace
4481 What to do with requests that have whitespace characters in the
4484 strip: The whitespace characters are stripped out of the URL.
4485 This is the behavior recommended by RFC2396.
4486 deny: The request is denied. The user receives an "Invalid
4488 allow: The request is allowed and the URI is not changed. The
4489 whitespace characters remain in the URI. Note the
4490 whitespace is passed to redirector processes if they
4492 encode: The request is allowed and the whitespace characters are
4493 encoded according to RFC1738. This could be considered
4494 a violation of the HTTP/1.1
4495 RFC because proxies are not allowed to rewrite URI's.
4496 chop: The request is allowed and the URI is chopped at the
4497 first whitespace. This might also be considered a
4504 LOC: Config.accessList.brokenPosts
4506 A list of ACL elements which, if matched, causes Squid to send
4507 an extra CRLF pair after the body of a PUT/POST request.
4509 Some HTTP servers has broken implementations of PUT/POST,
4510 and rely on an extra CRLF pair sent by some WWW clients.
4512 Quote from RFC 2068 section 4.1 on this matter:
4514 Note: certain buggy HTTP/1.0 client implementations generate an
4515 extra CRLF's after a POST request. To restate what is explicitly
4516 forbidden by the BNF, an HTTP/1.1 client must not preface or follow
4517 a request with an extra CRLF.
4520 acl buggy_server url_regex ^http://....
4521 broken_posts allow buggy_server
4524 NAME: mcast_miss_addr
4525 IFDEF: MULTICAST_MISS_STREAM
4527 LOC: Config.mcast_miss.addr
4528 DEFAULT: 255.255.255.255
4530 If you enable this option, every "cache miss" URL will
4531 be sent out on the specified multicast address.
4533 Do not enable this option unless you are are absolutely
4534 certain you understand what you are doing.
4537 NAME: mcast_miss_ttl
4538 IFDEF: MULTICAST_MISS_STREAM
4540 LOC: Config.mcast_miss.ttl
4543 This is the time-to-live value for packets multicasted
4544 when multicasting off cache miss URLs is enabled. By
4545 default this is set to 'site scope', i.e. 16.
4548 NAME: mcast_miss_port
4549 IFDEF: MULTICAST_MISS_STREAM
4551 LOC: Config.mcast_miss.port
4554 This is the port number to be used in conjunction with
4558 NAME: mcast_miss_encode_key
4559 IFDEF: MULTICAST_MISS_STREAM
4561 LOC: Config.mcast_miss.encode_key
4562 DEFAULT: XXXXXXXXXXXXXXXX
4564 The URLs that are sent in the multicast miss stream are
4565 encrypted. This is the encryption key.
4568 NAME: nonhierarchical_direct
4570 LOC: Config.onoff.nonhierarchical_direct
4573 By default, Squid will send any non-hierarchical requests
4574 (matching hierarchy_stoplist or not cachable request type) direct
4577 If you set this to off, Squid will prefer to send these
4578 requests to parents.
4580 Note that in most configurations, by turning this off you will only
4581 add latency to these request without any improvement in global hit
4584 If you are inside an firewall see never_direct instead of
4590 LOC: Config.onoff.prefer_direct
4593 Normally Squid tries to use parents for most requests. If you for some
4594 reason like it to first try going direct and only use a parent if
4595 going direct fails set this to on.
4597 By combining nonhierarchical_direct off and prefer_direct on you
4598 can set up Squid to use a parent as a backup path if going direct
4602 NAME: strip_query_terms
4604 LOC: Config.onoff.strip_query_terms
4607 By default, Squid strips query terms from requested URLs before
4608 logging. This protects your user's privacy.
4613 LOC: Config.coredump_dir
4615 DEFAULT_IF_NONE: none
4617 By default Squid leaves core files in the directory from where
4618 it was started. If you set 'coredump_dir' to a directory
4619 that exists, Squid will chdir() to that directory at startup
4620 and coredump files will be left there.
4623 # Leave coredumps in the first cache dir
4624 coredump_dir @DEFAULT_SWAP_DIR@
4628 NAME: redirector_bypass
4630 LOC: Config.onoff.redirector_bypass
4633 When this is 'on', a request will not go through the
4634 redirector if all redirectors are busy. If this is 'off'
4635 and the redirector queue grows too large, Squid will exit
4636 with a FATAL error and ask you to increase the number of
4637 redirectors. You should only enable this if the redirectors
4638 are not critical to your caching system. If you use
4639 redirectors for access control, and you enable this option,
4640 users may have access to pages they should not
4641 be allowed to request.
4644 NAME: ignore_unknown_nameservers
4646 LOC: Config.onoff.ignore_unknown_nameservers
4649 By default Squid checks that DNS responses are received
4650 from the same IP addresses they are sent to. If they
4651 don't match, Squid ignores the response and writes a warning
4652 message to cache.log. You can allow responses from unknown
4653 nameservers by setting this option to 'off'.
4656 NAME: digest_generation
4657 IFDEF: USE_CACHE_DIGESTS
4659 LOC: Config.onoff.digest_generation
4662 This controls whether the server will generate a Cache Digest
4663 of its contents. By default, Cache Digest generation is
4664 enabled if Squid is compiled with USE_CACHE_DIGESTS defined.
4667 NAME: digest_bits_per_entry
4668 IFDEF: USE_CACHE_DIGESTS
4670 LOC: Config.digest.bits_per_entry
4673 This is the number of bits of the server's Cache Digest which
4674 will be associated with the Digest entry for a given HTTP
4675 Method and URL (public key) combination. The default is 5.
4678 NAME: digest_rebuild_period
4679 IFDEF: USE_CACHE_DIGESTS
4682 LOC: Config.digest.rebuild_period
4685 This is the number of seconds between Cache Digest rebuilds.
4688 NAME: digest_rewrite_period
4690 IFDEF: USE_CACHE_DIGESTS
4692 LOC: Config.digest.rewrite_period
4695 This is the number of seconds between Cache Digest writes to
4699 NAME: digest_swapout_chunk_size
4702 IFDEF: USE_CACHE_DIGESTS
4703 LOC: Config.digest.swapout_chunk_size
4706 This is the number of bytes of the Cache Digest to write to
4707 disk at a time. It defaults to 4096 bytes (4KB), the Squid
4711 NAME: digest_rebuild_chunk_percentage
4712 COMMENT: (percent, 0-100)
4713 IFDEF: USE_CACHE_DIGESTS
4715 LOC: Config.digest.rebuild_chunk_percentage
4718 This is the percentage of the Cache Digest to be scanned at a
4719 time. By default it is set to 10% of the Cache Digest.
4724 LOC: Config.chroot_dir
4727 Use this to have Squid do a chroot() while initializing. This
4728 also causes Squid to fully drop root privileges after
4729 initializing. This means, for example, if you use a HTTP
4730 port less than 1024 and try to reconfigure, you will get an
4734 NAME: client_persistent_connections
4736 LOC: Config.onoff.client_pconns
4740 NAME: server_persistent_connections
4742 LOC: Config.onoff.server_pconns
4745 Persistent connection support for clients and servers. By
4746 default, Squid uses persistent connections (when allowed)
4747 with its clients and servers. You can use these options to
4748 disable persistent connections with clients and/or servers.
4751 NAME: persistent_connection_after_error
4753 LOC: Config.onoff.error_pconns
4756 With this directive the use of persistent connections after
4757 HTTP errors can be disabled. Useful if you have clients
4758 who fail to handle errors on persistent connections proper.
4761 NAME: detect_broken_pconn
4763 LOC: Config.onoff.detect_broken_server_pconns
4766 Some servers have been found to incorrectly signal the use
4767 of HTTP/1.0 persistent connections even on replies not
4768 compatible, causing significant delays. This server problem
4769 has mostly been seen on redirects.
4771 By enabling this directive Squid attempts to detect such
4772 broken replies and automatically assume the reply is finished
4773 after 10 seconds timeout.
4776 NAME: balance_on_multiple_ip
4778 LOC: Config.onoff.balance_on_multiple_ip
4781 Some load balancing servers based on round robin DNS have been
4782 found not to preserve user session state across requests
4783 to different IP addresses.
4785 By default Squid rotates IP's per request. By disabling
4786 this directive only connection failure triggers rotation.
4789 NAME: pipeline_prefetch
4791 LOC: Config.onoff.pipeline_prefetch
4794 To boost the performance of pipelined requests to closer
4795 match that of a non-proxied environment Squid can try to fetch
4796 up to two requests in parallel from a pipeline.
4798 Defaults to off for bandwidth management and access logging
4802 NAME: extension_methods
4804 LOC: Config.ext_methods
4807 Squid only knows about standardized HTTP request methods.
4808 You can add up to 20 additional "extension" methods here.
4811 NAME: request_entities
4813 LOC: Config.onoff.request_entities
4816 Squid defaults to deny GET and HEAD requests with request entities,
4817 as the meaning of such requests are undefined in the HTTP standard
4818 even if not explicitly forbidden.
4820 Set this directive to on if you have clients which insists
4821 on sending request entities in GET or HEAD requests.
4824 NAME: high_response_time_warning
4827 LOC: Config.warnings.high_rptm
4830 If the one-minute median response time exceeds this value,
4831 Squid prints a WARNING with debug level 0 to get the
4832 administrators attention. The value is in milliseconds.
4835 NAME: high_page_fault_warning
4837 LOC: Config.warnings.high_pf
4840 If the one-minute average page fault rate exceeds this
4841 value, Squid prints a WARNING with debug level 0 to get
4842 the administrators attention. The value is in page faults
4846 NAME: high_memory_warning
4848 LOC: Config.warnings.high_memory
4851 If the memory usage (as determined by mallinfo) exceeds
4852 value, Squid prints a WARNING with debug level 0 to get
4853 the administrators attention.
4856 NAME: store_dir_select_algorithm
4858 LOC: Config.store_dir_select_algorithm
4861 Set this to 'round-robin' as an alternative.
4868 LOC: Config.Log.forward
4870 Logs the server-side requests.
4872 This is currently work in progress.
4878 LOC: Config.onoff.ie_refresh
4881 Microsoft Internet Explorer up until version 5.5 Service
4882 Pack 1 has an issue with transparent proxies, wherein it
4883 is impossible to force a refresh. Turning this on provides
4884 a partial fix to the problem, by causing all IMS-REFRESH
4885 requests from older IE versions to check the origin server
4886 for fresh content. This reduces hit ratio by some amount
4887 (~10% in my experience), but allows users to actually get
4888 fresh content when they want it. Note because Squid
4889 cannot tell if the user is using 5.5 or 5.5SP1, the behavior
4890 of 5.5 is unchanged from old versions of Squid (i.e. a
4891 forced refresh is impossible). Newer versions of IE will,
4892 hopefully, continue to have the new behavior and will be
4893 handled based on that assumption. This option defaults to
4894 the old Squid behavior, which is better for hit ratios but
4895 worse for clients using IE, if they need to be able to
4896 force fresh content.
4899 NAME: vary_ignore_expire
4902 LOC: Config.onoff.vary_ignore_expire
4905 Many HTTP servers supporting Vary gives such objects
4906 immediate expiry time with no cache-control header
4907 when requested by a HTTP/1.0 client. This option
4908 enables Squid to ignore such expiry times until
4909 HTTP/1.1 is fully implemented.
4910 WARNING: This may eventually cause some varying
4911 objects not intended for caching to get cached.
4914 NAME: sleep_after_fork
4915 COMMENT: (microseconds)
4917 LOC: Config.sleep_after_fork
4920 When this is set to a non-zero value, the main Squid process
4921 sleeps the specified number of microseconds after a fork()
4922 system call. This sleep may help the situation where your
4923 system reports fork() failures due to lack of (virtual)
4924 memory. Note, however, if you have a lot of child
4925 processes, these sleep delays will add up and your
4926 Squid will not service requests for some amount of time
4927 until all the child processes have been started.
4928 On Windows value less then 1000 (1 millisencond) are
4932 NAME: minimum_expiry_time
4935 LOC: Config.minimum_expiry_time
4938 The minimum caching time according to (Expires - Date)
4939 Headers Squid honors if the object can't be revalidated
4940 defaults to 60 seconds. In reverse proxy enorinments it
4941 might be desirable to honor shorter object lifetimes. It
4942 is most likely better to make your server return a
4943 meaningful Last-Modified header however. In ESI environments
4944 where page fragments often have short lifetimes, this will
4945 often be best set to 0.
4948 NAME: relaxed_header_parser
4949 COMMENT: on|off|warn
4951 LOC: Config.onoff.relaxed_header_parser
4954 In the default "on" setting Squid accepts certain forms
4955 of non-compliant HTTP messages where it is unambiguous
4956 what the sending application intended even if the message
4957 is not correctly formatted. The messages is then normalized
4958 to the correct form when forwarded by Squid.
4960 If set to "warn" then a warning will be emitted in cache.log
4961 each time such HTTP error is encountered.
4963 If set to "off" then such HTTP errors will cause the request
4964 or response to be rejected.
4969 -----------------------------------------------------------------------------
4976 LOC: TheICAPConfig.onoff
4979 If you want to enable the ICAP module support, set this to on.
4982 NAME: icap_service_failure_limit
4985 LOC: TheICAPConfig.service_failure_limit
4988 The limit specifies the number of failures that Squid tolerates
4989 when establishing a new TCP connection with an ICAP service. If
4990 the number of failures exceeds the limit, the ICAP service is
4991 not used for new ICAP requests until it is time to refresh its
4992 OPTIONS. The per-service failure counter is reset to zero each
4993 time Squid fetches new service OPTIONS.
4995 A negative value disables the limit. Without the limit, an ICAP
4996 service will not be considered down due to connectivity failures
4997 between ICAP OPTIONS requests.
5000 NAME: icap_service_revival_delay
5003 LOC: TheICAPConfig.service_revival_delay
5006 The delay specifies the number of seconds to wait after an ICAP
5007 OPTIONS request failure before requesting the options again. The
5008 failed ICAP service is considered "down" until fresh OPTIONS are
5011 The actual delay cannot be smaller than the hardcoded minimum
5012 delay of 60 seconds.
5015 NAME: icap_preview_enable
5019 LOC: TheICAPConfig.preview_enable
5022 Set this to 'on' if you want to enable the ICAP preview
5026 NAME: icap_preview_size
5029 LOC: TheICAPConfig.preview_size
5032 The default size of preview data to be sent to the ICAP server.
5033 -1 means no preview. This value might be overwritten on a per server
5034 basis by OPTIONS requests.
5037 NAME: icap_default_options_ttl
5040 LOC: TheICAPConfig.default_options_ttl
5043 The default TTL value for ICAP OPTIONS responses that don't have
5044 an Options-TTL header.
5047 NAME: icap_persistent_connections
5051 LOC: TheICAPConfig.reuse_connections
5054 Whether or not Squid should use persistent connections to
5058 NAME: icap_send_client_ip
5062 LOC: TheICAPConfig.send_client_ip
5065 This adds the header "X-Client-IP" to ICAP requests.
5068 NAME: icap_send_client_username
5072 LOC: TheICAPConfig.send_client_username
5075 This sends authenticated HTTP client username (if available) to
5076 the ICAP service. The username value is encoded based on the
5077 icap_client_username_encode option and is sent using the header
5078 specified by the icap_client_username_header option.
5081 NAME: icap_client_username_header
5084 LOC: TheICAPConfig.client_username_header
5085 DEFAULT: X-Client-Username
5087 ICAP request header name to use for send_client_username.
5090 NAME: icap_client_username_encode
5094 LOC: TheICAPConfig.client_username_encode
5097 Whether to base64 encode the authenticated client username.
5101 TYPE: icap_service_type
5106 Defines a single ICAP service
5108 icap_service servicename vectoring_point bypass service_url
5110 vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache
5111 This specifies at which point of request processing the ICAP
5112 service should be plugged in.
5114 If set to 1 and the ICAP server cannot be reached, the request will go
5115 through without being processed by an ICAP server
5116 service_url = icap://servername:port/service
5118 Note: reqmod_precache and respmod_postcache is not yet implemented
5121 icap_service service_1 reqmod_precache 0 icap://icap1.mydomain.net:1344/reqmod
5122 icap_service service_2 respmod_precache 0 icap://icap2.mydomain.net:1344/respmod
5126 TYPE: icap_class_type
5131 Defines an ICAP service chain. If there are multiple services per
5132 vectoring point, they are processed in the specified order.
5134 icap_class classname servicename...
5137 icap_class class_1 service_1 service_2
5138 icap class class_2 service_1 service_3
5142 TYPE: icap_access_type
5147 Redirects a request through an ICAP service class, depending
5150 icap_access classname allow|deny [!]aclname...
5152 The icap_access statements are processed in the order they appear in
5153 this configuration file. If an access list matches, the processing stops.
5154 For an "allow" rule, the specified class is used for the request. A "deny"
5155 rule simply stops processing without using the class. You can also use the
5156 special classname "None".
5158 For backward compatibility, it is also possible to use services
5161 icap_access class_1 allow all